linux/fs/zonefs/super.c

1450 lines
37 KiB
C
Raw Normal View History

fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
// SPDX-License-Identifier: GPL-2.0
/*
* Simple file system for zoned block devices exposing zones as files.
*
* Copyright (C) 2019 Western Digital Corporation or its affiliates.
*/
#include <linux/module.h>
#include <linux/pagemap.h>
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
#include <linux/magic.h>
#include <linux/iomap.h>
#include <linux/init.h>
#include <linux/slab.h>
#include <linux/blkdev.h>
#include <linux/statfs.h>
#include <linux/writeback.h>
#include <linux/quotaops.h>
#include <linux/seq_file.h>
#include <linux/parser.h>
#include <linux/uio.h>
#include <linux/mman.h>
#include <linux/sched/mm.h>
#include <linux/crc32.h>
#include <linux/task_io_accounting_ops.h>
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
#include "zonefs.h"
#define CREATE_TRACE_POINTS
#include "trace.h"
/*
* Get the name of a zone group directory.
*/
static const char *zonefs_zgroup_name(enum zonefs_ztype ztype)
{
switch (ztype) {
case ZONEFS_ZTYPE_CNV:
return "cnv";
case ZONEFS_ZTYPE_SEQ:
return "seq";
default:
WARN_ON_ONCE(1);
return "???";
}
}
/*
* Manage the active zone count.
*/
static void zonefs_account_active(struct super_block *sb,
struct zonefs_zone *z)
{
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
if (zonefs_zone_is_cnv(z))
return;
/*
* For zones that transitioned to the offline or readonly condition,
* we only need to clear the active state.
*/
if (z->z_flags & (ZONEFS_ZONE_OFFLINE | ZONEFS_ZONE_READONLY))
goto out;
/*
* If the zone is active, that is, if it is explicitly open or
* partially written, check if it was already accounted as active.
*/
if ((z->z_flags & ZONEFS_ZONE_OPEN) ||
(z->z_wpoffset > 0 && z->z_wpoffset < z->z_capacity)) {
if (!(z->z_flags & ZONEFS_ZONE_ACTIVE)) {
z->z_flags |= ZONEFS_ZONE_ACTIVE;
atomic_inc(&sbi->s_active_seq_files);
}
return;
}
out:
/* The zone is not active. If it was, update the active count */
if (z->z_flags & ZONEFS_ZONE_ACTIVE) {
z->z_flags &= ~ZONEFS_ZONE_ACTIVE;
atomic_dec(&sbi->s_active_seq_files);
}
}
/*
* Manage the active zone count. Called with zi->i_truncate_mutex held.
*/
void zonefs_inode_account_active(struct inode *inode)
{
lockdep_assert_held(&ZONEFS_I(inode)->i_truncate_mutex);
return zonefs_account_active(inode->i_sb, zonefs_inode_zone(inode));
}
/*
* Execute a zone management operation.
*/
static int zonefs_zone_mgmt(struct super_block *sb,
struct zonefs_zone *z, enum req_op op)
{
int ret;
/*
* With ZNS drives, closing an explicitly open zone that has not been
* written will change the zone state to "closed", that is, the zone
* will remain active. Since this can then cause failure of explicit
* open operation on other zones if the drive active zone resources
* are exceeded, make sure that the zone does not remain active by
* resetting it.
*/
if (op == REQ_OP_ZONE_CLOSE && !z->z_wpoffset)
op = REQ_OP_ZONE_RESET;
trace_zonefs_zone_mgmt(sb, z, op);
ret = blkdev_zone_mgmt(sb->s_bdev, op, z->z_sector,
z->z_size >> SECTOR_SHIFT, GFP_NOFS);
if (ret) {
zonefs_err(sb,
"Zone management operation %s at %llu failed %d\n",
blk_op_str(op), z->z_sector, ret);
return ret;
}
return 0;
}
int zonefs_inode_zone_mgmt(struct inode *inode, enum req_op op)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
lockdep_assert_held(&ZONEFS_I(inode)->i_truncate_mutex);
zonefs: fix zonefs_iomap_begin() for reads If a readahead is issued to a sequential zone file with an offset exactly equal to the current file size, the iomap type is set to IOMAP_UNWRITTEN, which will prevent an IO, but the iomap length is calculated as 0. This causes a WARN_ON() in iomap_iter(): [17309.548939] WARNING: CPU: 3 PID: 2137 at fs/iomap/iter.c:34 iomap_iter+0x9cf/0xe80 [...] [17309.650907] RIP: 0010:iomap_iter+0x9cf/0xe80 [...] [17309.754560] Call Trace: [17309.757078] <TASK> [17309.759240] ? lock_is_held_type+0xd8/0x130 [17309.763531] iomap_readahead+0x1a8/0x870 [17309.767550] ? iomap_read_folio+0x4c0/0x4c0 [17309.771817] ? lockdep_hardirqs_on_prepare+0x400/0x400 [17309.778848] ? lock_release+0x370/0x750 [17309.784462] ? folio_add_lru+0x217/0x3f0 [17309.790220] ? reacquire_held_locks+0x4e0/0x4e0 [17309.796543] read_pages+0x17d/0xb60 [17309.801854] ? folio_add_lru+0x238/0x3f0 [17309.807573] ? readahead_expand+0x5f0/0x5f0 [17309.813554] ? policy_node+0xb5/0x140 [17309.819018] page_cache_ra_unbounded+0x27d/0x450 [17309.825439] filemap_get_pages+0x500/0x1450 [17309.831444] ? filemap_add_folio+0x140/0x140 [17309.837519] ? lock_is_held_type+0xd8/0x130 [17309.843509] filemap_read+0x28c/0x9f0 [17309.848953] ? zonefs_file_read_iter+0x1ea/0x4d0 [zonefs] [17309.856162] ? trace_contention_end+0xd6/0x130 [17309.862416] ? __mutex_lock+0x221/0x1480 [17309.868151] ? zonefs_file_read_iter+0x166/0x4d0 [zonefs] [17309.875364] ? filemap_get_pages+0x1450/0x1450 [17309.881647] ? __mutex_unlock_slowpath+0x15e/0x620 [17309.888248] ? wait_for_completion_io_timeout+0x20/0x20 [17309.895231] ? lock_is_held_type+0xd8/0x130 [17309.901115] ? lock_is_held_type+0xd8/0x130 [17309.906934] zonefs_file_read_iter+0x356/0x4d0 [zonefs] [17309.913750] new_sync_read+0x2d8/0x520 [17309.919035] ? __x64_sys_lseek+0x1d0/0x1d0 Furthermore, this causes iomap_readahead() to loop forever as iomap_readahead_iter() always returns 0, making no progress. Fix this by treating reads after the file size as access to holes, setting the iomap type to IOMAP_HOLE, the iomap addr to IOMAP_NULL_ADDR and using the length argument as is for the iomap length. To simplify the code with this change, zonefs_iomap_begin() is split into the read variant, zonefs_read_iomap_begin() and zonefs_read_iomap_ops, and the write variant, zonefs_write_iomap_begin() and zonefs_write_iomap_ops. Reported-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Fixes: 8dcc1a9d90c1 ("fs: New zonefs file system") Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com>
2022-05-23 15:29:10 +08:00
return zonefs_zone_mgmt(inode->i_sb, zonefs_inode_zone(inode), op);
zonefs: fix zonefs_iomap_begin() for reads If a readahead is issued to a sequential zone file with an offset exactly equal to the current file size, the iomap type is set to IOMAP_UNWRITTEN, which will prevent an IO, but the iomap length is calculated as 0. This causes a WARN_ON() in iomap_iter(): [17309.548939] WARNING: CPU: 3 PID: 2137 at fs/iomap/iter.c:34 iomap_iter+0x9cf/0xe80 [...] [17309.650907] RIP: 0010:iomap_iter+0x9cf/0xe80 [...] [17309.754560] Call Trace: [17309.757078] <TASK> [17309.759240] ? lock_is_held_type+0xd8/0x130 [17309.763531] iomap_readahead+0x1a8/0x870 [17309.767550] ? iomap_read_folio+0x4c0/0x4c0 [17309.771817] ? lockdep_hardirqs_on_prepare+0x400/0x400 [17309.778848] ? lock_release+0x370/0x750 [17309.784462] ? folio_add_lru+0x217/0x3f0 [17309.790220] ? reacquire_held_locks+0x4e0/0x4e0 [17309.796543] read_pages+0x17d/0xb60 [17309.801854] ? folio_add_lru+0x238/0x3f0 [17309.807573] ? readahead_expand+0x5f0/0x5f0 [17309.813554] ? policy_node+0xb5/0x140 [17309.819018] page_cache_ra_unbounded+0x27d/0x450 [17309.825439] filemap_get_pages+0x500/0x1450 [17309.831444] ? filemap_add_folio+0x140/0x140 [17309.837519] ? lock_is_held_type+0xd8/0x130 [17309.843509] filemap_read+0x28c/0x9f0 [17309.848953] ? zonefs_file_read_iter+0x1ea/0x4d0 [zonefs] [17309.856162] ? trace_contention_end+0xd6/0x130 [17309.862416] ? __mutex_lock+0x221/0x1480 [17309.868151] ? zonefs_file_read_iter+0x166/0x4d0 [zonefs] [17309.875364] ? filemap_get_pages+0x1450/0x1450 [17309.881647] ? __mutex_unlock_slowpath+0x15e/0x620 [17309.888248] ? wait_for_completion_io_timeout+0x20/0x20 [17309.895231] ? lock_is_held_type+0xd8/0x130 [17309.901115] ? lock_is_held_type+0xd8/0x130 [17309.906934] zonefs_file_read_iter+0x356/0x4d0 [zonefs] [17309.913750] new_sync_read+0x2d8/0x520 [17309.919035] ? __x64_sys_lseek+0x1d0/0x1d0 Furthermore, this causes iomap_readahead() to loop forever as iomap_readahead_iter() always returns 0, making no progress. Fix this by treating reads after the file size as access to holes, setting the iomap type to IOMAP_HOLE, the iomap addr to IOMAP_NULL_ADDR and using the length argument as is for the iomap length. To simplify the code with this change, zonefs_iomap_begin() is split into the read variant, zonefs_read_iomap_begin() and zonefs_read_iomap_ops, and the write variant, zonefs_write_iomap_begin() and zonefs_write_iomap_ops. Reported-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Fixes: 8dcc1a9d90c1 ("fs: New zonefs file system") Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com>
2022-05-23 15:29:10 +08:00
}
void zonefs_i_size_write(struct inode *inode, loff_t isize)
zonefs: fix zonefs_iomap_begin() for reads If a readahead is issued to a sequential zone file with an offset exactly equal to the current file size, the iomap type is set to IOMAP_UNWRITTEN, which will prevent an IO, but the iomap length is calculated as 0. This causes a WARN_ON() in iomap_iter(): [17309.548939] WARNING: CPU: 3 PID: 2137 at fs/iomap/iter.c:34 iomap_iter+0x9cf/0xe80 [...] [17309.650907] RIP: 0010:iomap_iter+0x9cf/0xe80 [...] [17309.754560] Call Trace: [17309.757078] <TASK> [17309.759240] ? lock_is_held_type+0xd8/0x130 [17309.763531] iomap_readahead+0x1a8/0x870 [17309.767550] ? iomap_read_folio+0x4c0/0x4c0 [17309.771817] ? lockdep_hardirqs_on_prepare+0x400/0x400 [17309.778848] ? lock_release+0x370/0x750 [17309.784462] ? folio_add_lru+0x217/0x3f0 [17309.790220] ? reacquire_held_locks+0x4e0/0x4e0 [17309.796543] read_pages+0x17d/0xb60 [17309.801854] ? folio_add_lru+0x238/0x3f0 [17309.807573] ? readahead_expand+0x5f0/0x5f0 [17309.813554] ? policy_node+0xb5/0x140 [17309.819018] page_cache_ra_unbounded+0x27d/0x450 [17309.825439] filemap_get_pages+0x500/0x1450 [17309.831444] ? filemap_add_folio+0x140/0x140 [17309.837519] ? lock_is_held_type+0xd8/0x130 [17309.843509] filemap_read+0x28c/0x9f0 [17309.848953] ? zonefs_file_read_iter+0x1ea/0x4d0 [zonefs] [17309.856162] ? trace_contention_end+0xd6/0x130 [17309.862416] ? __mutex_lock+0x221/0x1480 [17309.868151] ? zonefs_file_read_iter+0x166/0x4d0 [zonefs] [17309.875364] ? filemap_get_pages+0x1450/0x1450 [17309.881647] ? __mutex_unlock_slowpath+0x15e/0x620 [17309.888248] ? wait_for_completion_io_timeout+0x20/0x20 [17309.895231] ? lock_is_held_type+0xd8/0x130 [17309.901115] ? lock_is_held_type+0xd8/0x130 [17309.906934] zonefs_file_read_iter+0x356/0x4d0 [zonefs] [17309.913750] new_sync_read+0x2d8/0x520 [17309.919035] ? __x64_sys_lseek+0x1d0/0x1d0 Furthermore, this causes iomap_readahead() to loop forever as iomap_readahead_iter() always returns 0, making no progress. Fix this by treating reads after the file size as access to holes, setting the iomap type to IOMAP_HOLE, the iomap addr to IOMAP_NULL_ADDR and using the length argument as is for the iomap length. To simplify the code with this change, zonefs_iomap_begin() is split into the read variant, zonefs_read_iomap_begin() and zonefs_read_iomap_ops, and the write variant, zonefs_write_iomap_begin() and zonefs_write_iomap_ops. Reported-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Fixes: 8dcc1a9d90c1 ("fs: New zonefs file system") Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com>
2022-05-23 15:29:10 +08:00
{
struct zonefs_zone *z = zonefs_inode_zone(inode);
zonefs: fix zonefs_iomap_begin() for reads If a readahead is issued to a sequential zone file with an offset exactly equal to the current file size, the iomap type is set to IOMAP_UNWRITTEN, which will prevent an IO, but the iomap length is calculated as 0. This causes a WARN_ON() in iomap_iter(): [17309.548939] WARNING: CPU: 3 PID: 2137 at fs/iomap/iter.c:34 iomap_iter+0x9cf/0xe80 [...] [17309.650907] RIP: 0010:iomap_iter+0x9cf/0xe80 [...] [17309.754560] Call Trace: [17309.757078] <TASK> [17309.759240] ? lock_is_held_type+0xd8/0x130 [17309.763531] iomap_readahead+0x1a8/0x870 [17309.767550] ? iomap_read_folio+0x4c0/0x4c0 [17309.771817] ? lockdep_hardirqs_on_prepare+0x400/0x400 [17309.778848] ? lock_release+0x370/0x750 [17309.784462] ? folio_add_lru+0x217/0x3f0 [17309.790220] ? reacquire_held_locks+0x4e0/0x4e0 [17309.796543] read_pages+0x17d/0xb60 [17309.801854] ? folio_add_lru+0x238/0x3f0 [17309.807573] ? readahead_expand+0x5f0/0x5f0 [17309.813554] ? policy_node+0xb5/0x140 [17309.819018] page_cache_ra_unbounded+0x27d/0x450 [17309.825439] filemap_get_pages+0x500/0x1450 [17309.831444] ? filemap_add_folio+0x140/0x140 [17309.837519] ? lock_is_held_type+0xd8/0x130 [17309.843509] filemap_read+0x28c/0x9f0 [17309.848953] ? zonefs_file_read_iter+0x1ea/0x4d0 [zonefs] [17309.856162] ? trace_contention_end+0xd6/0x130 [17309.862416] ? __mutex_lock+0x221/0x1480 [17309.868151] ? zonefs_file_read_iter+0x166/0x4d0 [zonefs] [17309.875364] ? filemap_get_pages+0x1450/0x1450 [17309.881647] ? __mutex_unlock_slowpath+0x15e/0x620 [17309.888248] ? wait_for_completion_io_timeout+0x20/0x20 [17309.895231] ? lock_is_held_type+0xd8/0x130 [17309.901115] ? lock_is_held_type+0xd8/0x130 [17309.906934] zonefs_file_read_iter+0x356/0x4d0 [zonefs] [17309.913750] new_sync_read+0x2d8/0x520 [17309.919035] ? __x64_sys_lseek+0x1d0/0x1d0 Furthermore, this causes iomap_readahead() to loop forever as iomap_readahead_iter() always returns 0, making no progress. Fix this by treating reads after the file size as access to holes, setting the iomap type to IOMAP_HOLE, the iomap addr to IOMAP_NULL_ADDR and using the length argument as is for the iomap length. To simplify the code with this change, zonefs_iomap_begin() is split into the read variant, zonefs_read_iomap_begin() and zonefs_read_iomap_ops, and the write variant, zonefs_write_iomap_begin() and zonefs_write_iomap_ops. Reported-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Fixes: 8dcc1a9d90c1 ("fs: New zonefs file system") Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com>
2022-05-23 15:29:10 +08:00
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
i_size_write(inode, isize);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/*
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
* A full zone is no longer open/active and does not need
* explicit closing.
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
*/
if (isize >= z->z_capacity) {
struct zonefs_sb_info *sbi = ZONEFS_SB(inode->i_sb);
if (z->z_flags & ZONEFS_ZONE_ACTIVE)
atomic_dec(&sbi->s_active_seq_files);
z->z_flags &= ~(ZONEFS_ZONE_OPEN | ZONEFS_ZONE_ACTIVE);
}
}
void zonefs_update_stats(struct inode *inode, loff_t new_isize)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
struct super_block *sb = inode->i_sb;
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
loff_t old_isize = i_size_read(inode);
loff_t nr_blocks;
if (new_isize == old_isize)
return;
spin_lock(&sbi->s_lock);
/*
* This may be called for an update after an IO error.
* So beware of the values seen.
*/
if (new_isize < old_isize) {
nr_blocks = (old_isize - new_isize) >> sb->s_blocksize_bits;
if (sbi->s_used_blocks > nr_blocks)
sbi->s_used_blocks -= nr_blocks;
else
sbi->s_used_blocks = 0;
} else {
sbi->s_used_blocks +=
(new_isize - old_isize) >> sb->s_blocksize_bits;
if (sbi->s_used_blocks > sbi->s_blocks)
sbi->s_used_blocks = sbi->s_blocks;
}
spin_unlock(&sbi->s_lock);
}
/*
* Check a zone condition. Return the amount of written (and still readable)
* data in the zone.
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
*/
static loff_t zonefs_check_zone_condition(struct super_block *sb,
struct zonefs_zone *z,
struct blk_zone *zone)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
switch (zone->cond) {
case BLK_ZONE_COND_OFFLINE:
zonefs_warn(sb, "Zone %llu: offline zone\n",
z->z_sector);
z->z_flags |= ZONEFS_ZONE_OFFLINE;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
case BLK_ZONE_COND_READONLY:
/*
* The write pointer of read-only zones is invalid, so we cannot
* determine the zone wpoffset (inode size). We thus keep the
* zone wpoffset as is, which leads to an empty file
* (wpoffset == 0) on mount. For a runtime error, this keeps
* the inode size as it was when last updated so that the user
* can recover data.
*/
zonefs_warn(sb, "Zone %llu: read-only zone\n",
z->z_sector);
z->z_flags |= ZONEFS_ZONE_READONLY;
if (zonefs_zone_is_cnv(z))
return z->z_capacity;
return z->z_wpoffset;
case BLK_ZONE_COND_FULL:
/* The write pointer of full zones is invalid. */
return z->z_capacity;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
default:
if (zonefs_zone_is_cnv(z))
return z->z_capacity;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return (zone->wp - zone->start) << SECTOR_SHIFT;
}
}
/*
* Check a zone condition and adjust its inode access permissions for
* offline and readonly zones.
*/
static void zonefs_inode_update_mode(struct inode *inode)
{
struct zonefs_zone *z = zonefs_inode_zone(inode);
if (z->z_flags & ZONEFS_ZONE_OFFLINE) {
/* Offline zones cannot be read nor written */
inode->i_flags |= S_IMMUTABLE;
inode->i_mode &= ~0777;
} else if (z->z_flags & ZONEFS_ZONE_READONLY) {
/* Readonly zones cannot be written */
inode->i_flags |= S_IMMUTABLE;
if (z->z_flags & ZONEFS_ZONE_INIT_MODE)
inode->i_mode &= ~0777;
else
inode->i_mode &= ~0222;
}
z->z_flags &= ~ZONEFS_ZONE_INIT_MODE;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
z->z_mode = inode->i_mode;
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct zonefs_ioerr_data {
struct inode *inode;
bool write;
};
static int zonefs_io_error_cb(struct blk_zone *zone, unsigned int idx,
void *data)
{
struct zonefs_ioerr_data *err = data;
struct inode *inode = err->inode;
struct zonefs_zone *z = zonefs_inode_zone(inode);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct super_block *sb = inode->i_sb;
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
loff_t isize, data_size;
/*
* Check the zone condition: if the zone is not "bad" (offline or
* read-only), read errors are simply signaled to the IO issuer as long
* as there is no inconsistency between the inode size and the amount of
* data writen in the zone (data_size).
*/
data_size = zonefs_check_zone_condition(sb, z, zone);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
isize = i_size_read(inode);
if (!(z->z_flags & (ZONEFS_ZONE_READONLY | ZONEFS_ZONE_OFFLINE)) &&
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
!err->write && isize == data_size)
return 0;
/*
* At this point, we detected either a bad zone or an inconsistency
* between the inode size and the amount of data written in the zone.
* For the latter case, the cause may be a write IO error or an external
* action on the device. Two error patterns exist:
* 1) The inode size is lower than the amount of data in the zone:
* a write operation partially failed and data was writen at the end
* of the file. This can happen in the case of a large direct IO
* needing several BIOs and/or write requests to be processed.
* 2) The inode size is larger than the amount of data in the zone:
* this can happen with a deferred write error with the use of the
* device side write cache after getting successful write IO
* completions. Other possibilities are (a) an external corruption,
* e.g. an application reset the zone directly, or (b) the device
* has a serious problem (e.g. firmware bug).
*
* In all cases, warn about inode size inconsistency and handle the
* IO error according to the zone condition and to the mount options.
*/
if (zonefs_zone_is_seq(z) && isize != data_size)
zonefs_warn(sb,
"inode %lu: invalid size %lld (should be %lld)\n",
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
inode->i_ino, isize, data_size);
/*
* First handle bad zones signaled by hardware. The mount options
* errors=zone-ro and errors=zone-offline result in changing the
* zone condition to read-only and offline respectively, as if the
* condition was signaled by the hardware.
*/
if ((z->z_flags & ZONEFS_ZONE_OFFLINE) ||
(sbi->s_mount_opts & ZONEFS_MNTOPT_ERRORS_ZOL)) {
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs_warn(sb, "inode %lu: read/write access disabled\n",
inode->i_ino);
if (!(z->z_flags & ZONEFS_ZONE_OFFLINE))
z->z_flags |= ZONEFS_ZONE_OFFLINE;
zonefs_inode_update_mode(inode);
data_size = 0;
} else if ((z->z_flags & ZONEFS_ZONE_READONLY) ||
(sbi->s_mount_opts & ZONEFS_MNTOPT_ERRORS_ZRO)) {
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs_warn(sb, "inode %lu: write access disabled\n",
inode->i_ino);
if (!(z->z_flags & ZONEFS_ZONE_READONLY))
z->z_flags |= ZONEFS_ZONE_READONLY;
zonefs_inode_update_mode(inode);
data_size = isize;
} else if (sbi->s_mount_opts & ZONEFS_MNTOPT_ERRORS_RO &&
data_size > isize) {
/* Do not expose garbage data */
data_size = isize;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
/*
* If the filesystem is mounted with the explicit-open mount option, we
* need to clear the ZONEFS_ZONE_OPEN flag if the zone transitioned to
* the read-only or offline condition, to avoid attempting an explicit
* close of the zone when the inode file is closed.
*/
if ((sbi->s_mount_opts & ZONEFS_MNTOPT_EXPLICIT_OPEN) &&
(z->z_flags & (ZONEFS_ZONE_READONLY | ZONEFS_ZONE_OFFLINE)))
z->z_flags &= ~ZONEFS_ZONE_OPEN;
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/*
* If error=remount-ro was specified, any error result in remounting
* the volume as read-only.
*/
if ((sbi->s_mount_opts & ZONEFS_MNTOPT_ERRORS_RO) && !sb_rdonly(sb)) {
zonefs_warn(sb, "remounting filesystem read-only\n");
sb->s_flags |= SB_RDONLY;
}
/*
* Update block usage stats and the inode size to prevent access to
* invalid data.
*/
zonefs_update_stats(inode, data_size);
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
zonefs_i_size_write(inode, data_size);
z->z_wpoffset = data_size;
zonefs_inode_account_active(inode);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
}
/*
* When an file IO error occurs, check the file zone to see if there is a change
* in the zone condition (e.g. offline or read-only). For a failed write to a
* sequential zone, the zone write pointer position must also be checked to
* eventually correct the file size and zonefs inode write pointer offset
* (which can be out of sync with the drive due to partial write failures).
*/
void __zonefs_io_error(struct inode *inode, bool write)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
struct zonefs_zone *z = zonefs_inode_zone(inode);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct super_block *sb = inode->i_sb;
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
unsigned int noio_flag;
unsigned int nr_zones = 1;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct zonefs_ioerr_data err = {
.inode = inode,
.write = write,
};
int ret;
/*
* The only files that have more than one zone are conventional zone
* files with aggregated conventional zones, for which the inode zone
* size is always larger than the device zone size.
*/
if (z->z_size > bdev_zone_sectors(sb->s_bdev))
nr_zones = z->z_size >>
(sbi->s_zone_sectors_shift + SECTOR_SHIFT);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/*
* Memory allocations in blkdev_report_zones() can trigger a memory
* reclaim which may in turn cause a recursion into zonefs as well as
* struct request allocations for the same device. The former case may
* end up in a deadlock on the inode truncate mutex, while the latter
* may prevent IO forward progress. Executing the report zones under
* the GFP_NOIO context avoids both problems.
*/
noio_flag = memalloc_noio_save();
ret = blkdev_report_zones(sb->s_bdev, z->z_sector, nr_zones,
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs_io_error_cb, &err);
if (ret != nr_zones)
zonefs_err(sb, "Get inode %lu zone information failed %d\n",
inode->i_ino, ret);
memalloc_noio_restore(noio_flag);
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
static struct kmem_cache *zonefs_inode_cachep;
static struct inode *zonefs_alloc_inode(struct super_block *sb)
{
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct zonefs_inode_info *zi;
zi = alloc_inode_sb(sb, zonefs_inode_cachep, GFP_KERNEL);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
if (!zi)
return NULL;
inode_init_once(&zi->i_vnode);
mutex_init(&zi->i_truncate_mutex);
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
zi->i_wr_refcnt = 0;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return &zi->i_vnode;
}
static void zonefs_free_inode(struct inode *inode)
{
kmem_cache_free(zonefs_inode_cachep, ZONEFS_I(inode));
}
/*
* File system stat.
*/
static int zonefs_statfs(struct dentry *dentry, struct kstatfs *buf)
{
struct super_block *sb = dentry->d_sb;
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
enum zonefs_ztype t;
buf->f_type = ZONEFS_MAGIC;
buf->f_bsize = sb->s_blocksize;
buf->f_namelen = ZONEFS_NAME_MAX;
spin_lock(&sbi->s_lock);
buf->f_blocks = sbi->s_blocks;
if (WARN_ON(sbi->s_used_blocks > sbi->s_blocks))
buf->f_bfree = 0;
else
buf->f_bfree = buf->f_blocks - sbi->s_used_blocks;
buf->f_bavail = buf->f_bfree;
for (t = 0; t < ZONEFS_ZTYPE_MAX; t++) {
if (sbi->s_zgroup[t].g_nr_zones)
buf->f_files += sbi->s_zgroup[t].g_nr_zones + 1;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
buf->f_ffree = 0;
spin_unlock(&sbi->s_lock);
buf->f_fsid = uuid_to_fsid(sbi->s_uuid.b);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
}
enum {
Opt_errors_ro, Opt_errors_zro, Opt_errors_zol, Opt_errors_repair,
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
Opt_explicit_open, Opt_err,
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
};
static const match_table_t tokens = {
{ Opt_errors_ro, "errors=remount-ro"},
{ Opt_errors_zro, "errors=zone-ro"},
{ Opt_errors_zol, "errors=zone-offline"},
{ Opt_errors_repair, "errors=repair"},
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
{ Opt_explicit_open, "explicit-open" },
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{ Opt_err, NULL}
};
static int zonefs_parse_options(struct super_block *sb, char *options)
{
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
substring_t args[MAX_OPT_ARGS];
char *p;
if (!options)
return 0;
while ((p = strsep(&options, ",")) != NULL) {
int token;
if (!*p)
continue;
token = match_token(p, tokens, args);
switch (token) {
case Opt_errors_ro:
sbi->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
sbi->s_mount_opts |= ZONEFS_MNTOPT_ERRORS_RO;
break;
case Opt_errors_zro:
sbi->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
sbi->s_mount_opts |= ZONEFS_MNTOPT_ERRORS_ZRO;
break;
case Opt_errors_zol:
sbi->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
sbi->s_mount_opts |= ZONEFS_MNTOPT_ERRORS_ZOL;
break;
case Opt_errors_repair:
sbi->s_mount_opts &= ~ZONEFS_MNTOPT_ERRORS_MASK;
sbi->s_mount_opts |= ZONEFS_MNTOPT_ERRORS_REPAIR;
break;
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
case Opt_explicit_open:
sbi->s_mount_opts |= ZONEFS_MNTOPT_EXPLICIT_OPEN;
break;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
default:
return -EINVAL;
}
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
}
static int zonefs_show_options(struct seq_file *seq, struct dentry *root)
{
struct zonefs_sb_info *sbi = ZONEFS_SB(root->d_sb);
if (sbi->s_mount_opts & ZONEFS_MNTOPT_ERRORS_RO)
seq_puts(seq, ",errors=remount-ro");
if (sbi->s_mount_opts & ZONEFS_MNTOPT_ERRORS_ZRO)
seq_puts(seq, ",errors=zone-ro");
if (sbi->s_mount_opts & ZONEFS_MNTOPT_ERRORS_ZOL)
seq_puts(seq, ",errors=zone-offline");
if (sbi->s_mount_opts & ZONEFS_MNTOPT_ERRORS_REPAIR)
seq_puts(seq, ",errors=repair");
return 0;
}
static int zonefs_remount(struct super_block *sb, int *flags, char *data)
{
sync_filesystem(sb);
return zonefs_parse_options(sb, data);
}
static int zonefs_inode_setattr(struct mnt_idmap *idmap,
struct dentry *dentry, struct iattr *iattr)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
struct inode *inode = d_inode(dentry);
int ret;
if (unlikely(IS_IMMUTABLE(inode)))
return -EPERM;
ret = setattr_prepare(&nop_mnt_idmap, dentry, iattr);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
if (ret)
return ret;
/*
* Since files and directories cannot be created nor deleted, do not
* allow setting any write attributes on the sub-directories grouping
* files by zone type.
*/
if ((iattr->ia_valid & ATTR_MODE) && S_ISDIR(inode->i_mode) &&
(iattr->ia_mode & 0222))
return -EPERM;
if (((iattr->ia_valid & ATTR_UID) &&
!uid_eq(iattr->ia_uid, inode->i_uid)) ||
((iattr->ia_valid & ATTR_GID) &&
!gid_eq(iattr->ia_gid, inode->i_gid))) {
ret = dquot_transfer(&nop_mnt_idmap, inode, iattr);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
if (ret)
return ret;
}
if (iattr->ia_valid & ATTR_SIZE) {
ret = zonefs_file_truncate(inode, iattr->ia_size);
if (ret)
return ret;
}
setattr_copy(&nop_mnt_idmap, inode, iattr);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (S_ISREG(inode->i_mode)) {
struct zonefs_zone *z = zonefs_inode_zone(inode);
z->z_mode = inode->i_mode;
z->z_uid = inode->i_uid;
z->z_gid = inode->i_gid;
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
}
static const struct inode_operations zonefs_file_inode_operations = {
.setattr = zonefs_inode_setattr,
};
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
static long zonefs_fname_to_fno(const struct qstr *fname)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
const char *name = fname->name;
unsigned int len = fname->len;
long fno = 0, shift = 1;
const char *rname;
char c = *name;
unsigned int i;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/*
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
* File names are always a base-10 number string without any
* leading 0s.
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
*/
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (!isdigit(c))
return -ENOENT;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (len > 1 && c == '0')
return -ENOENT;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (len == 1)
return c - '0';
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
for (i = 0, rname = name + len - 1; i < len; i++, rname--) {
c = *rname;
if (!isdigit(c))
return -ENOENT;
fno += (c - '0') * shift;
shift *= 10;
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
return fno;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
static struct inode *zonefs_get_file_inode(struct inode *dir,
struct dentry *dentry)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct zonefs_zone_group *zgroup = dir->i_private;
struct super_block *sb = dir->i_sb;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct zonefs_zone *z;
struct inode *inode;
ino_t ino;
long fno;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
/* Get the file number from the file name */
fno = zonefs_fname_to_fno(&dentry->d_name);
if (fno < 0)
return ERR_PTR(fno);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (!zgroup->g_nr_zones || fno >= zgroup->g_nr_zones)
return ERR_PTR(-ENOENT);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
z = &zgroup->g_zones[fno];
ino = z->z_sector >> sbi->s_zone_sectors_shift;
inode = iget_locked(sb, ino);
if (!inode)
return ERR_PTR(-ENOMEM);
if (!(inode->i_state & I_NEW)) {
WARN_ON_ONCE(inode->i_private != z);
return inode;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
inode->i_ino = ino;
inode->i_mode = z->z_mode;
inode->i_ctime = inode->i_mtime = inode->i_atime = dir->i_ctime;
inode->i_uid = z->z_uid;
inode->i_gid = z->z_gid;
inode->i_size = z->z_wpoffset;
inode->i_blocks = z->z_capacity >> SECTOR_SHIFT;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
inode->i_private = z;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
inode->i_op = &zonefs_file_inode_operations;
inode->i_fop = &zonefs_file_operations;
inode->i_mapping->a_ops = &zonefs_file_aops;
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
/* Update the inode access rights depending on the zone condition */
zonefs_inode_update_mode(inode);
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
unlock_new_inode(inode);
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
return inode;
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
static struct inode *zonefs_get_zgroup_inode(struct super_block *sb,
enum zonefs_ztype ztype)
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
{
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct inode *root = d_inode(sb->s_root);
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct inode *inode;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
ino_t ino = bdev_nr_zones(sb->s_bdev) + ztype + 1;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
inode = iget_locked(sb, ino);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
if (!inode)
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
return ERR_PTR(-ENOMEM);
if (!(inode->i_state & I_NEW))
return inode;
inode->i_ino = ino;
zonefs changes for 6.3-rc1 * Reorganize zonefs code to split file related operations to a new fs/zonefs/file.c file. From me. * Modify zonefs to use dynamically allocated inodes and dentries (using the inode and dentry caches) instead of statically allocating everything on mount. This saves a significant amount of memory for very large zoned block devices with 10s of thousands of zones. From me. * Make zonefs_sb_ktype a const struct kobj_type, from Thomas. -----BEGIN PGP SIGNATURE----- iHUEABYKAB0WIQSRPv8tYSvhwAzJdzjdoc3SxdoYdgUCY/VokQAKCRDdoc3SxdoY dgCVAP427Q2wBrphU1//iW6R/7OZMz24nwmJcYPBAVLTngQO5gD+IVX+ru4xhD4Q W1VKFN9QIDbjEiEc+Lmt9iQuDz+roAI= =e/FD -----END PGP SIGNATURE----- Merge tag 'zonefs-6.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs Pull zonefs updates from Damien Le Moal: - Reorganize zonefs code to split file related operations to a new fs/zonefs/file.c file (me) - Modify zonefs to use dynamically allocated inodes and dentries (using the inode and dentry caches) instead of statically allocating everything on mount. This saves a significant amount of memory for very large zoned block devices with 10s of thousands of zones (me) - Make zonefs_sb_ktype a const struct kobj_type (Thomas) * tag 'zonefs-6.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs: zonefs: make kobj_type structure constant zonefs: Cache zone group directory inodes zonefs: Dynamically create file inodes when needed zonefs: Separate zone information from inode information zonefs: Reduce struct zonefs_inode_info size zonefs: Simplify IO error handling zonefs: Reorganize code
2023-02-23 06:11:54 +08:00
inode_init_owner(&nop_mnt_idmap, inode, root, S_IFDIR | 0555);
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
inode->i_size = sbi->s_zgroup[ztype].g_nr_zones;
inode->i_ctime = inode->i_mtime = inode->i_atime = root->i_ctime;
inode->i_private = &sbi->s_zgroup[ztype];
set_nlink(inode, 2);
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
inode->i_op = &zonefs_dir_inode_operations;
inode->i_fop = &zonefs_dir_operations;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
unlock_new_inode(inode);
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
return inode;
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
static struct inode *zonefs_get_dir_inode(struct inode *dir,
struct dentry *dentry)
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
{
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct super_block *sb = dir->i_sb;
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
const char *name = dentry->d_name.name;
enum zonefs_ztype ztype;
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
/*
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
* We only need to check for the "seq" directory and
* the "cnv" directory if we have conventional zones.
*/
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (dentry->d_name.len != 3)
return ERR_PTR(-ENOENT);
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
for (ztype = 0; ztype < ZONEFS_ZTYPE_MAX; ztype++) {
if (sbi->s_zgroup[ztype].g_nr_zones &&
memcmp(name, zonefs_zgroup_name(ztype), 3) == 0)
break;
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (ztype == ZONEFS_ZTYPE_MAX)
return ERR_PTR(-ENOENT);
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
return zonefs_get_zgroup_inode(sb, ztype);
zonefs: open/close zone on file open/close NVMe Zoned Namespace introduced the concept of active zones, which are zones in the implicit open, explicit open or closed condition. Drives may have a limit on the number of zones that can be simultaneously active. This potential limitation translate into a risk for applications to see write IO errors due to this limit if the zone of a file being written to is not already active when a write request is issued. To avoid these potential errors, the zone of a file can explicitly be made active using an open zone command when the file is open for the first time. If the zone open command succeeds, the application is then guaranteed that write requests can be processed. This indirect management of active zones relies on the maximum number of open zones of a drive, which is always lower or equal to the maximum number of active zones. On the first open of a sequential zone file, send a REQ_OP_ZONE_OPEN command to the block device. Conversely, on the last release of a zone file and send a REQ_OP_ZONE_CLOSE to the device if the zone is not full or empty. As truncating a zone file to 0 or max can deactivate a zone as well, we need to serialize against truncates and also be careful not to close a zone as the file may still be open for writing, e.g. the user called ftruncate(). If the zone file is not open and a process does a truncate(), then no close operation is needed. Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
2020-09-11 16:56:50 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
static struct dentry *zonefs_lookup(struct inode *dir, struct dentry *dentry,
unsigned int flags)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct inode *inode;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (dentry->d_name.len > ZONEFS_NAME_MAX)
return ERR_PTR(-ENAMETOOLONG);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (dir == d_inode(dir->i_sb->s_root))
inode = zonefs_get_dir_inode(dir, dentry);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
else
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
inode = zonefs_get_file_inode(dir, dentry);
if (IS_ERR(inode))
return ERR_CAST(inode);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
return d_splice_alias(inode, dentry);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
static int zonefs_readdir_root(struct file *file, struct dir_context *ctx)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct inode *inode = file_inode(file);
struct super_block *sb = inode->i_sb;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
enum zonefs_ztype ztype = ZONEFS_ZTYPE_CNV;
ino_t base_ino = bdev_nr_zones(sb->s_bdev) + 1;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (ctx->pos >= inode->i_size)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (!dir_emit_dots(file, ctx))
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (ctx->pos == 2) {
if (!sbi->s_zgroup[ZONEFS_ZTYPE_CNV].g_nr_zones)
ztype = ZONEFS_ZTYPE_SEQ;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (!dir_emit(ctx, zonefs_zgroup_name(ztype), 3,
base_ino + ztype, DT_DIR))
return 0;
ctx->pos++;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (ctx->pos == 3 && ztype != ZONEFS_ZTYPE_SEQ) {
ztype = ZONEFS_ZTYPE_SEQ;
if (!dir_emit(ctx, zonefs_zgroup_name(ztype), 3,
base_ino + ztype, DT_DIR))
return 0;
ctx->pos++;
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
static int zonefs_readdir_zgroup(struct file *file,
struct dir_context *ctx)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct inode *inode = file_inode(file);
struct zonefs_zone_group *zgroup = inode->i_private;
struct super_block *sb = inode->i_sb;
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
struct zonefs_zone *z;
int fname_len;
char *fname;
ino_t ino;
int f;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
/*
* The size of zone group directories is equal to the number
* of zone files in the group and does note include the "." and
* ".." entries. Hence the "+ 2" here.
*/
if (ctx->pos >= inode->i_size + 2)
return 0;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (!dir_emit_dots(file, ctx))
return 0;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
fname = kmalloc(ZONEFS_NAME_MAX, GFP_KERNEL);
if (!fname)
return -ENOMEM;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
for (f = ctx->pos - 2; f < zgroup->g_nr_zones; f++) {
z = &zgroup->g_zones[f];
ino = z->z_sector >> sbi->s_zone_sectors_shift;
fname_len = snprintf(fname, ZONEFS_NAME_MAX - 1, "%u", f);
if (!dir_emit(ctx, fname, fname_len, ino, DT_REG))
break;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
ctx->pos++;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
kfree(fname);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
return 0;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
static int zonefs_readdir(struct file *file, struct dir_context *ctx)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct inode *inode = file_inode(file);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
if (inode == d_inode(inode->i_sb->s_root))
return zonefs_readdir_root(file, ctx);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
return zonefs_readdir_zgroup(file, ctx);
}
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
const struct inode_operations zonefs_dir_inode_operations = {
.lookup = zonefs_lookup,
.setattr = zonefs_inode_setattr,
};
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
const struct file_operations zonefs_dir_operations = {
.llseek = generic_file_llseek,
.read = generic_read_dir,
.iterate_shared = zonefs_readdir,
};
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
struct zonefs_zone_data {
struct super_block *sb;
unsigned int nr_zones[ZONEFS_ZTYPE_MAX];
sector_t cnv_zone_start;
struct blk_zone *zones;
};
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
static int zonefs_get_zone_info_cb(struct blk_zone *zone, unsigned int idx,
void *data)
{
struct zonefs_zone_data *zd = data;
struct super_block *sb = zd->sb;
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
/*
* We do not care about the first zone: it contains the super block
* and not exposed as a file.
*/
if (!idx)
return 0;
/*
* Count the number of zones that will be exposed as files.
* For sequential zones, we always have as many files as zones.
* FOr conventional zones, the number of files depends on if we have
* conventional zones aggregation enabled.
*/
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
switch (zone->type) {
case BLK_ZONE_TYPE_CONVENTIONAL:
if (sbi->s_features & ZONEFS_F_AGGRCNV) {
/* One file per set of contiguous conventional zones */
if (!(sbi->s_zgroup[ZONEFS_ZTYPE_CNV].g_nr_zones) ||
zone->start != zd->cnv_zone_start)
sbi->s_zgroup[ZONEFS_ZTYPE_CNV].g_nr_zones++;
zd->cnv_zone_start = zone->start + zone->len;
} else {
/* One file per zone */
sbi->s_zgroup[ZONEFS_ZTYPE_CNV].g_nr_zones++;
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
break;
case BLK_ZONE_TYPE_SEQWRITE_REQ:
case BLK_ZONE_TYPE_SEQWRITE_PREF:
sbi->s_zgroup[ZONEFS_ZTYPE_SEQ].g_nr_zones++;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
break;
default:
zonefs_err(zd->sb, "Unsupported zone type 0x%x\n",
zone->type);
return -EIO;
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
memcpy(&zd->zones[idx], zone, sizeof(struct blk_zone));
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
}
static int zonefs_get_zone_info(struct zonefs_zone_data *zd)
{
struct block_device *bdev = zd->sb->s_bdev;
int ret;
zd->zones = kvcalloc(bdev_nr_zones(bdev), sizeof(struct blk_zone),
GFP_KERNEL);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
if (!zd->zones)
return -ENOMEM;
/* Get zones information from the device */
ret = blkdev_report_zones(bdev, 0, BLK_ALL_ZONES,
zonefs_get_zone_info_cb, zd);
if (ret < 0) {
zonefs_err(zd->sb, "Zone report failed %d\n", ret);
return ret;
}
if (ret != bdev_nr_zones(bdev)) {
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs_err(zd->sb, "Invalid zone report (%d/%u zones)\n",
ret, bdev_nr_zones(bdev));
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return -EIO;
}
return 0;
}
static inline void zonefs_free_zone_info(struct zonefs_zone_data *zd)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
kvfree(zd->zones);
}
/*
* Create a zone group and populate it with zone files.
*/
static int zonefs_init_zgroup(struct super_block *sb,
struct zonefs_zone_data *zd,
enum zonefs_ztype ztype)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
struct zonefs_zone_group *zgroup = &sbi->s_zgroup[ztype];
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
struct blk_zone *zone, *next, *end;
struct zonefs_zone *z;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
unsigned int n = 0;
int ret;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/* Allocate the zone group. If it is empty, we have nothing to do. */
if (!zgroup->g_nr_zones)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
zgroup->g_zones = kvcalloc(zgroup->g_nr_zones,
sizeof(struct zonefs_zone), GFP_KERNEL);
if (!zgroup->g_zones)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return -ENOMEM;
/*
* Initialize the zone groups using the device zone information.
* We always skip the first zone as it contains the super block
* and is not use to back a file.
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
*/
end = zd->zones + bdev_nr_zones(sb->s_bdev);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
for (zone = &zd->zones[1]; zone < end; zone = next) {
next = zone + 1;
if (zonefs_zone_type(zone) != ztype)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
continue;
if (WARN_ON_ONCE(n >= zgroup->g_nr_zones))
return -EINVAL;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/*
* For conventional zones, contiguous zones can be aggregated
* together to form larger files. Note that this overwrites the
* length of the first zone of the set of contiguous zones
* aggregated together. If one offline or read-only zone is
* found, assume that all zones aggregated have the same
* condition.
*/
if (ztype == ZONEFS_ZTYPE_CNV &&
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
(sbi->s_features & ZONEFS_F_AGGRCNV)) {
for (; next < end; next++) {
if (zonefs_zone_type(next) != ztype)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
break;
zone->len += next->len;
zone->capacity += next->capacity;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
if (next->cond == BLK_ZONE_COND_READONLY &&
zone->cond != BLK_ZONE_COND_OFFLINE)
zone->cond = BLK_ZONE_COND_READONLY;
else if (next->cond == BLK_ZONE_COND_OFFLINE)
zone->cond = BLK_ZONE_COND_OFFLINE;
}
}
z = &zgroup->g_zones[n];
if (ztype == ZONEFS_ZTYPE_CNV)
z->z_flags |= ZONEFS_ZONE_CNV;
z->z_sector = zone->start;
z->z_size = zone->len << SECTOR_SHIFT;
if (z->z_size > bdev_zone_sectors(sb->s_bdev) << SECTOR_SHIFT &&
!(sbi->s_features & ZONEFS_F_AGGRCNV)) {
zonefs_err(sb,
"Invalid zone size %llu (device zone sectors %llu)\n",
z->z_size,
bdev_zone_sectors(sb->s_bdev) << SECTOR_SHIFT);
return -EINVAL;
}
z->z_capacity = min_t(loff_t, MAX_LFS_FILESIZE,
zone->capacity << SECTOR_SHIFT);
z->z_wpoffset = zonefs_check_zone_condition(sb, z, zone);
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
z->z_mode = S_IFREG | sbi->s_perm;
z->z_uid = sbi->s_uid;
z->z_gid = sbi->s_gid;
/*
* Let zonefs_inode_update_mode() know that we will need
* special initialization of the inode mode the first time
* it is accessed.
*/
z->z_flags |= ZONEFS_ZONE_INIT_MODE;
sb->s_maxbytes = max(z->z_capacity, sb->s_maxbytes);
sbi->s_blocks += z->z_capacity >> sb->s_blocksize_bits;
sbi->s_used_blocks += z->z_wpoffset >> sb->s_blocksize_bits;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/*
* For sequential zones, make sure that any open zone is closed
* first to ensure that the initial number of open zones is 0,
* in sync with the open zone accounting done when the mount
* option ZONEFS_MNTOPT_EXPLICIT_OPEN is used.
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
*/
if (ztype == ZONEFS_ZTYPE_SEQ &&
(zone->cond == BLK_ZONE_COND_IMP_OPEN ||
zone->cond == BLK_ZONE_COND_EXP_OPEN)) {
ret = zonefs_zone_mgmt(sb, z, REQ_OP_ZONE_CLOSE);
if (ret)
return ret;
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs_account_active(sb, z);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
n++;
}
if (WARN_ON_ONCE(n != zgroup->g_nr_zones))
return -EINVAL;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs_info(sb, "Zone group \"%s\" has %u file%s\n",
zonefs_zgroup_name(ztype),
zgroup->g_nr_zones,
zgroup->g_nr_zones > 1 ? "s" : "");
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
static void zonefs_free_zgroups(struct super_block *sb)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
enum zonefs_ztype ztype;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
if (!sbi)
return;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
for (ztype = 0; ztype < ZONEFS_ZTYPE_MAX; ztype++) {
kvfree(sbi->s_zgroup[ztype].g_zones);
sbi->s_zgroup[ztype].g_zones = NULL;
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
/*
* Create a zone group and populate it with zone files.
*/
static int zonefs_init_zgroups(struct super_block *sb)
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
{
struct zonefs_zone_data zd;
enum zonefs_ztype ztype;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
int ret;
/* First get the device zone information */
memset(&zd, 0, sizeof(struct zonefs_zone_data));
zd.sb = sb;
ret = zonefs_get_zone_info(&zd);
if (ret)
goto cleanup;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/* Allocate and initialize the zone groups */
for (ztype = 0; ztype < ZONEFS_ZTYPE_MAX; ztype++) {
ret = zonefs_init_zgroup(sb, &zd, ztype);
if (ret) {
zonefs_info(sb,
"Zone group \"%s\" initialization failed\n",
zonefs_zgroup_name(ztype));
break;
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
cleanup:
zonefs_free_zone_info(&zd);
if (ret)
zonefs_free_zgroups(sb);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return ret;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
/*
* Read super block information from the device.
*/
static int zonefs_read_super(struct super_block *sb)
{
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
struct zonefs_super *super;
u32 crc, stored_crc;
struct page *page;
struct bio_vec bio_vec;
struct bio bio;
int ret;
page = alloc_page(GFP_KERNEL);
if (!page)
return -ENOMEM;
bio_init(&bio, sb->s_bdev, &bio_vec, 1, REQ_OP_READ);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
bio.bi_iter.bi_sector = 0;
bio_add_page(&bio, page, PAGE_SIZE, 0);
ret = submit_bio_wait(&bio);
if (ret)
goto free_page;
super = page_address(page);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
ret = -EINVAL;
if (le32_to_cpu(super->s_magic) != ZONEFS_MAGIC)
goto free_page;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
stored_crc = le32_to_cpu(super->s_crc);
super->s_crc = 0;
crc = crc32(~0U, (unsigned char *)super, sizeof(struct zonefs_super));
if (crc != stored_crc) {
zonefs_err(sb, "Invalid checksum (Expected 0x%08x, got 0x%08x)",
crc, stored_crc);
goto free_page;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
sbi->s_features = le64_to_cpu(super->s_features);
if (sbi->s_features & ~ZONEFS_F_DEFINED_FEATURES) {
zonefs_err(sb, "Unknown features set 0x%llx\n",
sbi->s_features);
goto free_page;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
if (sbi->s_features & ZONEFS_F_UID) {
sbi->s_uid = make_kuid(current_user_ns(),
le32_to_cpu(super->s_uid));
if (!uid_valid(sbi->s_uid)) {
zonefs_err(sb, "Invalid UID feature\n");
goto free_page;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
}
if (sbi->s_features & ZONEFS_F_GID) {
sbi->s_gid = make_kgid(current_user_ns(),
le32_to_cpu(super->s_gid));
if (!gid_valid(sbi->s_gid)) {
zonefs_err(sb, "Invalid GID feature\n");
goto free_page;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
}
if (sbi->s_features & ZONEFS_F_PERM)
sbi->s_perm = le32_to_cpu(super->s_perm);
if (memchr_inv(super->s_reserved, 0, sizeof(super->s_reserved))) {
zonefs_err(sb, "Reserved area is being used\n");
goto free_page;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
import_uuid(&sbi->s_uuid, super->s_uuid);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
ret = 0;
free_page:
__free_page(page);
return ret;
}
static const struct super_operations zonefs_sops = {
.alloc_inode = zonefs_alloc_inode,
.free_inode = zonefs_free_inode,
.statfs = zonefs_statfs,
.remount_fs = zonefs_remount,
.show_options = zonefs_show_options,
};
static int zonefs_get_zgroup_inodes(struct super_block *sb)
{
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
struct inode *dir_inode;
enum zonefs_ztype ztype;
for (ztype = 0; ztype < ZONEFS_ZTYPE_MAX; ztype++) {
if (!sbi->s_zgroup[ztype].g_nr_zones)
continue;
dir_inode = zonefs_get_zgroup_inode(sb, ztype);
if (IS_ERR(dir_inode))
return PTR_ERR(dir_inode);
sbi->s_zgroup[ztype].g_inode = dir_inode;
}
return 0;
}
static void zonefs_release_zgroup_inodes(struct super_block *sb)
{
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
enum zonefs_ztype ztype;
if (!sbi)
return;
for (ztype = 0; ztype < ZONEFS_ZTYPE_MAX; ztype++) {
if (sbi->s_zgroup[ztype].g_inode) {
iput(sbi->s_zgroup[ztype].g_inode);
sbi->s_zgroup[ztype].g_inode = NULL;
}
}
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
/*
* Check that the device is zoned. If it is, get the list of zones and create
* sub-directories and files according to the device zone configuration and
* format options.
*/
static int zonefs_fill_super(struct super_block *sb, void *data, int silent)
{
struct zonefs_sb_info *sbi;
struct inode *inode;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
enum zonefs_ztype ztype;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
int ret;
if (!bdev_is_zoned(sb->s_bdev)) {
zonefs_err(sb, "Not a zoned block device\n");
return -EINVAL;
}
/*
* Initialize super block information: the maximum file size is updated
* when the zone files are created so that the format option
* ZONEFS_F_AGGRCNV which increases the maximum file size of a file
* beyond the zone size is taken into account.
*/
sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
if (!sbi)
return -ENOMEM;
spin_lock_init(&sbi->s_lock);
sb->s_fs_info = sbi;
sb->s_magic = ZONEFS_MAGIC;
sb->s_maxbytes = 0;
sb->s_op = &zonefs_sops;
sb->s_time_gran = 1;
/*
* The block size is set to the device zone write granularity to ensure
* that write operations are always aligned according to the device
* interface constraints.
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
*/
sb_set_blocksize(sb, bdev_zone_write_granularity(sb->s_bdev));
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
sbi->s_zone_sectors_shift = ilog2(bdev_zone_sectors(sb->s_bdev));
sbi->s_uid = GLOBAL_ROOT_UID;
sbi->s_gid = GLOBAL_ROOT_GID;
sbi->s_perm = 0640;
sbi->s_mount_opts = ZONEFS_MNTOPT_ERRORS_RO;
atomic_set(&sbi->s_wro_seq_files, 0);
sbi->s_max_wro_seq_files = bdev_max_open_zones(sb->s_bdev);
atomic_set(&sbi->s_active_seq_files, 0);
sbi->s_max_active_seq_files = bdev_max_active_zones(sb->s_bdev);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
ret = zonefs_read_super(sb);
if (ret)
return ret;
ret = zonefs_parse_options(sb, data);
if (ret)
return ret;
zonefs_info(sb, "Mounting %u zones", bdev_nr_zones(sb->s_bdev));
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
if (!sbi->s_max_wro_seq_files &&
!sbi->s_max_active_seq_files &&
sbi->s_mount_opts & ZONEFS_MNTOPT_EXPLICIT_OPEN) {
zonefs_info(sb,
"No open and active zone limits. Ignoring explicit_open mount option\n");
sbi->s_mount_opts &= ~ZONEFS_MNTOPT_EXPLICIT_OPEN;
}
/* Initialize the zone groups */
ret = zonefs_init_zgroups(sb);
if (ret)
goto cleanup;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
/* Create the root directory inode */
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
ret = -ENOMEM;
inode = new_inode(sb);
if (!inode)
goto cleanup;
inode->i_ino = bdev_nr_zones(sb->s_bdev);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
inode->i_mode = S_IFDIR | 0555;
inode->i_ctime = inode->i_mtime = inode->i_atime = current_time(inode);
inode->i_op = &zonefs_dir_inode_operations;
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
inode->i_fop = &zonefs_dir_operations;
inode->i_size = 2;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
set_nlink(inode, 2);
zonefs: Dynamically create file inodes when needed Allocating and initializing all inodes and dentries for all files results in a very large memory usage with high capacity zoned block devices. For instance, with a 26 TB SMR HDD with over 96000 zones, mounting the disk with zonefs results in about 130 MB of memory used, the vast majority of this space being used for vfs inodes and dentries. However, since a user will rarely access all zones at the same time, dynamically creating file inodes and dentries on demand, similarly to regular file systems, can significantly reduce memory usage. This patch modifies mount processing to not create the inodes and dentries for zone files. Instead, the directory inode operation zonefs_lookup() and directory file operation zonefs_readdir() are introduced to allocate and initialize inodes on-demand using the helper functions zonefs_get_dir_inode() and zonefs_get_zgroup_inode(). Implementation of these functions is simple, relying on the static nature of zonefs directories and files. Directory inodes are linked to the volume zone groups (struct zonefs_zone_group) they represent by using the directory inode i_private field. This simplifies the implementation of the lookup and readdir operations. Unreferenced zone file inodes can be evicted from the inode cache at any time. In such case, the only inode information that cannot be recreated from the zone information that is saved in the zone group data structures attached to the volume super block is the inode uid, gid and access rights. These values may have been changed by the user. To keep these attributes for the life time of the mount, as before, the inode mode, uid and gid are saved in the inode zone information and the saved values are used to initialize regular file inodes when an inode lookup happens. The zone information mode, uid and gid are initialized in zonefs_init_zgroup() using the default values. With these changes, the static minimal memory usage of a zonefs volume is mostly reduced to the array of zone information for each zone group. For the 26 TB SMR hard-disk mentioned above, the memory usage after mount becomes about 5.4 MB, a reduction by a factor of 24 from the initial 130 MB memory use. Co-developed-by: Jorgen Hansen <Jorgen.Hansen@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
2022-11-30 10:01:09 +08:00
for (ztype = 0; ztype < ZONEFS_ZTYPE_MAX; ztype++) {
if (sbi->s_zgroup[ztype].g_nr_zones) {
inc_nlink(inode);
inode->i_size++;
}
}
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
sb->s_root = d_make_root(inode);
if (!sb->s_root)
goto cleanup;
/*
* Take a reference on the zone groups directory inodes
* to keep them in the inode cache.
*/
ret = zonefs_get_zgroup_inodes(sb);
if (ret)
goto cleanup;
ret = zonefs_sysfs_register(sb);
if (ret)
goto cleanup;
return 0;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
cleanup:
zonefs_release_zgroup_inodes(sb);
zonefs_free_zgroups(sb);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return ret;
}
static struct dentry *zonefs_mount(struct file_system_type *fs_type,
int flags, const char *dev_name, void *data)
{
return mount_bdev(fs_type, flags, dev_name, data, zonefs_fill_super);
}
static void zonefs_kill_super(struct super_block *sb)
{
struct zonefs_sb_info *sbi = ZONEFS_SB(sb);
/* Release the reference on the zone group directory inodes */
zonefs_release_zgroup_inodes(sb);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
kill_block_super(sb);
zonefs_sysfs_unregister(sb);
zonefs_free_zgroups(sb);
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
kfree(sbi);
}
/*
* File system definition and registration.
*/
static struct file_system_type zonefs_type = {
.owner = THIS_MODULE,
.name = "zonefs",
.mount = zonefs_mount,
.kill_sb = zonefs_kill_super,
.fs_flags = FS_REQUIRES_DEV,
};
static int __init zonefs_init_inodecache(void)
{
zonefs_inode_cachep = kmem_cache_create("zonefs_inode_cache",
sizeof(struct zonefs_inode_info), 0,
(SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD | SLAB_ACCOUNT),
NULL);
if (zonefs_inode_cachep == NULL)
return -ENOMEM;
return 0;
}
static void zonefs_destroy_inodecache(void)
{
/*
* Make sure all delayed rcu free inodes are flushed before we
* destroy the inode cache.
*/
rcu_barrier();
kmem_cache_destroy(zonefs_inode_cachep);
}
static int __init zonefs_init(void)
{
int ret;
BUILD_BUG_ON(sizeof(struct zonefs_super) != ZONEFS_SUPER_SIZE);
ret = zonefs_init_inodecache();
if (ret)
return ret;
ret = zonefs_sysfs_init();
if (ret)
goto destroy_inodecache;
ret = register_filesystem(&zonefs_type);
if (ret)
goto sysfs_exit;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
return 0;
sysfs_exit:
zonefs_sysfs_exit();
destroy_inodecache:
zonefs_destroy_inodecache();
return ret;
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
}
static void __exit zonefs_exit(void)
{
unregister_filesystem(&zonefs_type);
zonefs_sysfs_exit();
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
zonefs_destroy_inodecache();
}
MODULE_AUTHOR("Damien Le Moal");
MODULE_DESCRIPTION("Zone file system for zoned block devices");
MODULE_LICENSE("GPL");
MODULE_ALIAS_FS("zonefs");
fs: New zonefs file system zonefs is a very simple file system exposing each zone of a zoned block device as a file. Unlike a regular file system with zoned block device support (e.g. f2fs), zonefs does not hide the sequential write constraint of zoned block devices to the user. Files representing sequential write zones of the device must be written sequentially starting from the end of the file (append only writes). As such, zonefs is in essence closer to a raw block device access interface than to a full featured POSIX file system. The goal of zonefs is to simplify the implementation of zoned block device support in applications by replacing raw block device file accesses with a richer file API, avoiding relying on direct block device file ioctls which may be more obscure to developers. One example of this approach is the implementation of LSM (log-structured merge) tree structures (such as used in RocksDB and LevelDB) on zoned block devices by allowing SSTables to be stored in a zone file similarly to a regular file system rather than as a range of sectors of a zoned device. The introduction of the higher level construct "one file is one zone" can help reducing the amount of changes needed in the application as well as introducing support for different application programming languages. Zonefs on-disk metadata is reduced to an immutable super block to persistently store a magic number and optional feature flags and values. On mount, zonefs uses blkdev_report_zones() to obtain the device zone configuration and populates the mount point with a static file tree solely based on this information. E.g. file sizes come from the device zone type and write pointer offset managed by the device itself. The zone files created on mount have the following characteristics. 1) Files representing zones of the same type are grouped together under a common sub-directory: * For conventional zones, the sub-directory "cnv" is used. * For sequential write zones, the sub-directory "seq" is used. These two directories are the only directories that exist in zonefs. Users cannot create other directories and cannot rename nor delete the "cnv" and "seq" sub-directories. 2) The name of zone files is the number of the file within the zone type sub-directory, in order of increasing zone start sector. 3) The size of conventional zone files is fixed to the device zone size. Conventional zone files cannot be truncated. 4) The size of sequential zone files represent the file's zone write pointer position relative to the zone start sector. Truncating these files is allowed only down to 0, in which case, the zone is reset to rewind the zone write pointer position to the start of the zone, or up to the zone size, in which case the file's zone is transitioned to the FULL state (finish zone operation). 5) All read and write operations to files are not allowed beyond the file zone size. Any access exceeding the zone size is failed with the -EFBIG error. 6) Creating, deleting, renaming or modifying any attribute of files and sub-directories is not allowed. 7) There are no restrictions on the type of read and write operations that can be issued to conventional zone files. Buffered, direct and mmap read & write operations are accepted. For sequential zone files, there are no restrictions on read operations, but all write operations must be direct IO append writes. mmap write of sequential files is not allowed. Several optional features of zonefs can be enabled at format time. * Conventional zone aggregation: ranges of contiguous conventional zones can be aggregated into a single larger file instead of the default one file per zone. * File ownership: The owner UID and GID of zone files is by default 0 (root) but can be changed to any valid UID/GID. * File access permissions: the default 640 access permissions can be changed. The mkzonefs tool is used to format zoned block devices for use with zonefs. This tool is available on Github at: git@github.com:damien-lemoal/zonefs-tools.git. zonefs-tools also includes a test suite which can be run against any zoned block device, including null_blk block device created with zoned mode. Example: the following formats a 15TB host-managed SMR HDD with 256 MB zones with the conventional zones aggregation feature enabled. $ sudo mkzonefs -o aggr_cnv /dev/sdX $ sudo mount -t zonefs /dev/sdX /mnt $ ls -l /mnt/ total 0 dr-xr-xr-x 2 root root 1 Nov 25 13:23 cnv dr-xr-xr-x 2 root root 55356 Nov 25 13:23 seq The size of the zone files sub-directories indicate the number of files existing for each type of zones. In this example, there is only one conventional zone file (all conventional zones are aggregated under a single file). $ ls -l /mnt/cnv total 137101312 -rw-r----- 1 root root 140391743488 Nov 25 13:23 0 This aggregated conventional zone file can be used as a regular file. $ sudo mkfs.ext4 /mnt/cnv/0 $ sudo mount -o loop /mnt/cnv/0 /data The "seq" sub-directory grouping files for sequential write zones has in this example 55356 zones. $ ls -lv /mnt/seq total 14511243264 -rw-r----- 1 root root 0 Nov 25 13:23 0 -rw-r----- 1 root root 0 Nov 25 13:23 1 -rw-r----- 1 root root 0 Nov 25 13:23 2 ... -rw-r----- 1 root root 0 Nov 25 13:23 55354 -rw-r----- 1 root root 0 Nov 25 13:23 55355 For sequential write zone files, the file size changes as data is appended at the end of the file, similarly to any regular file system. $ dd if=/dev/zero of=/mnt/seq/0 bs=4K count=1 conv=notrunc oflag=direct 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000452219 s, 9.1 MB/s $ ls -l /mnt/seq/0 -rw-r----- 1 root root 4096 Nov 25 13:23 /mnt/seq/0 The written file can be truncated to the zone size, preventing any further write operation. $ truncate -s 268435456 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 268435456 Nov 25 13:49 /mnt/seq/0 Truncation to 0 size allows freeing the file zone storage space and restart append-writes to the file. $ truncate -s 0 /mnt/seq/0 $ ls -l /mnt/seq/0 -rw-r----- 1 root root 0 Nov 25 13:49 /mnt/seq/0 Since files are statically mapped to zones on the disk, the number of blocks of a file as reported by stat() and fstat() indicates the size of the file zone. $ stat /mnt/seq/0 File: /mnt/seq/0 Size: 0 Blocks: 524288 IO Block: 4096 regular empty file Device: 870h/2160d Inode: 50431 Links: 1 Access: (0640/-rw-r-----) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2019-11-25 13:23:57.048971997 +0900 Modify: 2019-11-25 13:52:25.553805765 +0900 Change: 2019-11-25 13:52:25.553805765 +0900 Birth: - The number of blocks of the file ("Blocks") in units of 512B blocks gives the maximum file size of 524288 * 512 B = 256 MB, corresponding to the device zone size in this example. Of note is that the "IO block" field always indicates the minimum IO size for writes and corresponds to the device physical sector size. This code contains contributions from: * Johannes Thumshirn <jthumshirn@suse.de>, * Darrick J. Wong <darrick.wong@oracle.com>, * Christoph Hellwig <hch@lst.de>, * Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> and * Ting Yao <tingyao@hust.edu.cn>. Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-12-25 15:07:44 +08:00
module_init(zonefs_init);
module_exit(zonefs_exit);