renegade-project/linux-next
2
0
Fork 0
mirror of https://github.com/edk2-porting/linux-next.git synced 2024-12-27 14:43:58 +08:00
Code
316b003540
linux-next/fs/ceph/file.c

2517 lines
65 KiB
C
Raw Normal View History

License cleanup: add SPDX GPL-2.0 license identifier to files with no license Many source files in the tree are missing licensing information, which makes it harder for compliance tools to determine the correct license. By default all files without license information are under the default license of the kernel, which is GPL version 2. Update the files which contain no license information with the 'GPL-2.0' SPDX license identifier. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This patch is based on work done by Thomas Gleixner and Kate Stewart and Philippe Ombredanne. How this work was done: Patches were generated and checked against linux-4.14-rc6 for a subset of the use cases: - file had no licensing information it it. - file was a */uapi/* one with no licensing information in it, - file was a */uapi/* one with existing licensing information, Further patches will be generated in subsequent months to fix up cases where non-standard license headers were used, and references to license had to be inferred by heuristics based on keywords. The analysis to determine which SPDX License Identifier to be applied to a file was done in a spreadsheet of side by side results from of the output of two independent scanners (ScanCode & Windriver) producing SPDX tag:value files created by Philippe Ombredanne. Philippe prepared the base worksheet, and did an initial spot review of a few 1000 files. The 4.13 kernel was the starting point of the analysis with 60,537 files assessed. Kate Stewart did a file by file comparison of the scanner results in the spreadsheet to determine which SPDX license identifier(s) to be applied to the file. She confirmed any determination that was not immediately clear with lawyers working with the Linux Foundation. Criteria used to select files for SPDX license identifier tagging was: - Files considered eligible had to be source code files. - Make and config files were included as candidates if they contained >5 lines of source - File already had some variant of a license header in it (even if <5 lines). All documentation files were explicitly excluded. The following heuristics were used to determine which SPDX license identifiers to apply. - when both scanners couldn't find any license traces, file was considered to have no license information in it, and the top level COPYING file license applied. For non */uapi/* files that summary was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 11139 and resulted in the first patch in this series. If that file was a */uapi/* path one, it was "GPL-2.0 WITH Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 WITH Linux-syscall-note 930 and resulted in the second patch in this series. - if a file had some form of licensing information in it, and was one of the */uapi/* ones, it was denoted with the Linux-syscall-note if any GPL family license was found in the file or had no licensing in it (per prior point). Results summary: SPDX license identifier # files ---------------------------------------------------|------ GPL-2.0 WITH Linux-syscall-note 270 GPL-2.0+ WITH Linux-syscall-note 169 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17 LGPL-2.1+ WITH Linux-syscall-note 15 GPL-1.0+ WITH Linux-syscall-note 14 ((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5 LGPL-2.0+ WITH Linux-syscall-note 4 LGPL-2.1 WITH Linux-syscall-note 3 ((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3 ((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1 and that resulted in the third patch in this series. - when the two scanners agreed on the detected license(s), that became the concluded license(s). - when there was disagreement between the two scanners (one detected a license but the other didn't, or they both detected different licenses) a manual inspection of the file occurred. - In most cases a manual inspection of the information in the file resulted in a clear resolution of the license that should apply (and which scanner probably needed to revisit its heuristics). - When it was not immediately clear, the license identifier was confirmed with lawyers working with the Linux Foundation. - If there was any question as to the appropriate license identifier, the file was flagged for further research and to be revisited later in time. In total, over 70 hours of logged manual review was done on the spreadsheet to determine the SPDX license identifiers to apply to the source files by Kate, Philippe, Thomas and, in some cases, confirmation by lawyers working with the Linux Foundation. Kate also obtained a third independent scan of the 4.13 code base from FOSSology, and compared selected files where the other two scanners disagreed against that SPDX file, to see if there was new insights. The Windriver scanner is based on an older version of FOSSology in part, so they are related. Thomas did random spot checks in about 500 files from the spreadsheets for the uapi headers and agreed with SPDX license identifier in the files he inspected. For the non-uapi files Thomas did random spot checks in about 15000 files. In initial set of patches against 4.14-rc6, 3 files were found to have copy/paste license identifier errors, and have been fixed to reflect the correct identifier. Additionally Philippe spent 10 hours this week doing a detailed manual inspection and review of the 12,461 patched files from the initial patch version early this week with: - a full scancode scan run, collecting the matched texts, detected license ids and scores - reviewing anything where there was a license detected (about 500+ files) to ensure that the applied SPDX license was correct - reviewing anything where there was no detection but the patch license was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied SPDX license was correct This produced a worksheet with 20 files needing minor correction. This worksheet was then exported into 3 different .csv files for the different types of files to be modified. These .csv files were then reviewed by Greg. Thomas wrote a script to parse the csv files and add the proper SPDX tag to the file, in the format that the file expected. This script was further refined by Greg based on the output to detect more types of files automatically and to distinguish between header and source .c files (which need different comment types.) Finally Greg ran the script using the .csv files to generate the patches. Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 22:07:57 +08:00
// SPDX-License-Identifier: GPL-2.0
ceph: factor out libceph from Ceph file system This factors out protocol and low-level storage parts of ceph into a separate libceph module living in net/ceph and include/linux/ceph. This is mostly a matter of moving files around. However, a few key pieces of the interface change as well: - ceph_client becomes ceph_fs_client and ceph_client, where the latter captures the mon and osd clients, and the fs_client gets the mds client and file system specific pieces. - Mount option parsing and debugfs setup is correspondingly broken into two pieces. - The mon client gets a generic handler callback for otherwise unknown messages (mds map, in this case). - The basic supported/required feature bits can be expanded (and are by ceph_fs_client). No functional change, aside from some subtle error handling cases that got cleaned up in the refactoring process. Signed-off-by: Sage Weil <sage@newdream.net>
2010-04-07 06:14:15 +08:00
#include <linux/ceph/ceph_debug.h>
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
#include <linux/ceph/striper.h>
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: factor out libceph from Ceph file system This factors out protocol and low-level storage parts of ceph into a separate libceph module living in net/ceph and include/linux/ceph. This is mostly a matter of moving files around. However, a few key pieces of the interface change as well: - ceph_client becomes ceph_fs_client and ceph_client, where the latter captures the mon and osd clients, and the fs_client gets the mds client and file system specific pieces. - Mount option parsing and debugfs setup is correspondingly broken into two pieces. - The mon client gets a generic handler callback for otherwise unknown messages (mds map, in this case). - The basic supported/required feature bits can be expanded (and are by ceph_fs_client). No functional change, aside from some subtle error handling cases that got cleaned up in the refactoring process. Signed-off-by: Sage Weil <sage@newdream.net>
2010-04-07 06:14:15 +08:00
#include <linux/module.h>
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
#include <linux/sched.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
#include <linux/slab.h>
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
#include <linux/file.h>
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
#include <linux/mount.h>
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
#include <linux/namei.h>
#include <linux/writeback.h>
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
#include <linux/falloc.h>
ceph: increment change_attribute on local changes We don't set SB_I_VERSION on ceph since we need to manage it ourselves, so we must increment it whenever we update the file times. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-06-06 20:57:27 +08:00
#include <linux/iversion.h>
ceph: add read/write latency metric support Calculate the latency for OSD read requests. Add a new r_end_stamp field to struct ceph_osd_request that will hold the time of that the reply was received. Use that to calculate the RTT for each call, and divide the sum of those by number of calls to get averate RTT. Keep a tally of RTT for OSD writes and number of calls to track average latency of OSD writes. URL: https://tracker.ceph.com/issues/43215 Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-20 11:45:01 +08:00
#include <linux/ktime.h>
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
#include "super.h"
#include "mds_client.h"
ceph: use fscache as a local presisent cache Adding support for fscache to the Ceph filesystem. This would bring it to on par with some of the other network filesystems in Linux (like NFS, AFS, etc...) In order to mount the filesystem with fscache the 'fsc' mount option must be passed. Signed-off-by: Milosz Tanski <milosz@adfin.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-08-22 05:29:54 +08:00
#include "cache.h"
ceph: add buffered/direct exclusionary locking for reads and writes xfstest generic/451 intermittently fails. The test does O_DIRECT writes to a file, and then reads back the result using buffered I/O, while running a separate set of tasks that are also doing buffered reads. The client will invalidate the cache prior to a direct write, but it's easy for one of the other readers' replies to race in and reinstantiate the invalidated range with stale data. To fix this, we must to serialize direct I/O writes and buffered reads. We could just sprinkle in some shared locks on the i_rwsem for reads, and increase the exclusive footprint on the write side, but that would cause O_DIRECT writes to end up serialized vs. other direct requests. Instead, borrow the scheme used by nfs.ko. Buffered writes take the i_rwsem exclusively, but buffered reads take a shared lock, allowing them to run in parallel. O_DIRECT requests also take a shared lock, but we need for them to not run in parallel with buffered reads. A flag on the ceph_inode_info is used to indicate whether it's in direct or buffered I/O mode. When a conflicting request is submitted, it will block until the inode can be flipped to the necessary mode. Link: https://tracker.ceph.com/issues/40985 Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-08-03 01:15:39 +08:00
#include "io.h"
ceph: add read/write latency metric support Calculate the latency for OSD read requests. Add a new r_end_stamp field to struct ceph_osd_request that will hold the time of that the reply was received. Use that to calculate the RTT for each call, and divide the sum of those by number of calls to get averate RTT. Keep a tally of RTT for OSD writes and number of calls to track average latency of OSD writes. URL: https://tracker.ceph.com/issues/43215 Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-20 11:45:01 +08:00
#include "metric.h"
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: fix file open flags on ppc64 The file open flags (O_foo) are platform specific and should never go out to an interface that is not local to the system. Unfortunately these flags have leaked out onto the wire in the cephfs implementation. That lead to bogus flags getting transmitted on ppc64. This patch converts the kernel view of flags to the ceph view of file open flags. Fixes: 124e68e74 ("ceph: file operations") Signed-off-by: Alexander Graf <agraf@suse.de> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-04-28 00:34:00 +08:00
static __le32 ceph_flags_sys2wire(u32 flags)
{
u32 wire_flags = 0;
switch (flags & O_ACCMODE) {
case O_RDONLY:
wire_flags |= CEPH_O_RDONLY;
break;
case O_WRONLY:
wire_flags |= CEPH_O_WRONLY;
break;
case O_RDWR:
wire_flags |= CEPH_O_RDWR;
break;
}
ceph: filter out used flags when printing unused open flags Filter out used access mode flags when printing unused open flags. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-09 15:12:40 +08:00
flags &= ~O_ACCMODE;
ceph: fix file open flags on ppc64 The file open flags (O_foo) are platform specific and should never go out to an interface that is not local to the system. Unfortunately these flags have leaked out onto the wire in the cephfs implementation. That lead to bogus flags getting transmitted on ppc64. This patch converts the kernel view of flags to the ceph view of file open flags. Fixes: 124e68e74 ("ceph: file operations") Signed-off-by: Alexander Graf <agraf@suse.de> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-04-28 00:34:00 +08:00
#define ceph_sys2wire(a) if (flags & a) { wire_flags |= CEPH_##a; flags &= ~a; }
ceph_sys2wire(O_CREAT);
ceph_sys2wire(O_EXCL);
ceph_sys2wire(O_TRUNC);
ceph_sys2wire(O_DIRECTORY);
ceph_sys2wire(O_NOFOLLOW);
#undef ceph_sys2wire
if (flags)
ceph: add newline to end of debug message format Some of dout format do not include newline in the end, fix for the files which are in fs/ceph and net/ceph directories, and changing printk to dout for printing debug info in super.c Signed-off-by: Chengguang Xu <cgxu519@icloud.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-01-30 16:29:17 +08:00
dout("unused open flags: %x\n", flags);
ceph: fix file open flags on ppc64 The file open flags (O_foo) are platform specific and should never go out to an interface that is not local to the system. Unfortunately these flags have leaked out onto the wire in the cephfs implementation. That lead to bogus flags getting transmitted on ppc64. This patch converts the kernel view of flags to the ceph view of file open flags. Fixes: 124e68e74 ("ceph: file operations") Signed-off-by: Alexander Graf <agraf@suse.de> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-04-28 00:34:00 +08:00
return cpu_to_le32(wire_flags);
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/*
* Ceph file operations
*
* Implement basic open/close functionality, and implement
* read/write.
*
* We implement three modes of file I/O:
* - buffered uses the generic_file_aio_{read,write} helpers
*
* - synchronous is used when there is multi-client read/write
* sharing, avoids the page cache, and synchronously waits for an
* ack from the OSD.
*
* - direct io takes the variant of the sync path that references
* user pages directly.
*
* fsync() flushes and waits on dirty pages, but just queues metadata
* for writeback: since the MDS can recover size and mtime there is no
* need to wait for MDS acknowledgement.
*/
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
/*
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
* How many pages to get in one call to iov_iter_get_pages(). This
* determines the size of the on-stack array used as a buffer.
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
*/
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
#define ITER_GET_BVECS_PAGES 64
static ssize_t __iter_get_bvecs(struct iov_iter *iter, size_t maxsize,
struct bio_vec *bvecs)
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
{
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
size_t size = 0;
int bvec_idx = 0;
if (maxsize > iov_iter_count(iter))
maxsize = iov_iter_count(iter);
while (size < maxsize) {
struct page *pages[ITER_GET_BVECS_PAGES];
ssize_t bytes;
size_t start;
int idx = 0;
bytes = iov_iter_get_pages(iter, pages, maxsize - size,
ITER_GET_BVECS_PAGES, &start);
if (bytes < 0)
return size ?: bytes;
iov_iter_advance(iter, bytes);
size += bytes;
for ( ; bytes; idx++, bvec_idx++) {
struct bio_vec bv = {
.bv_page = pages[idx],
.bv_len = min_t(int, bytes, PAGE_SIZE - start),
.bv_offset = start,
};
bvecs[bvec_idx] = bv;
bytes -= bv.bv_len;
start = 0;
}
}
return size;
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
}
/*
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
* iov_iter_get_pages() only considers one iov_iter segment, no matter
* what maxsize or maxpages are given. For ITER_BVEC that is a single
* page.
*
* Attempt to get up to @maxsize bytes worth of pages from @iter.
* Return the number of bytes in the created bio_vec array, or an error.
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
*/
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
static ssize_t iter_get_bvecs_alloc(struct iov_iter *iter, size_t maxsize,
struct bio_vec **bvecs, int *num_bvecs)
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
{
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
struct bio_vec *bv;
size_t orig_count = iov_iter_count(iter);
ssize_t bytes;
int npages;
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
iov_iter_truncate(iter, maxsize);
npages = iov_iter_npages(iter, INT_MAX);
iov_iter_reexpand(iter, orig_count);
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
/*
* __iter_get_bvecs() may populate only part of the array -- zero it
* out.
*/
bv = kvmalloc_array(npages, sizeof(*bv), GFP_KERNEL | __GFP_ZERO);
if (!bv)
return -ENOMEM;
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
bytes = __iter_get_bvecs(iter, maxsize, bv);
if (bytes < 0) {
/*
* No pages were pinned -- just free the array.
*/
kvfree(bv);
return bytes;
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
}
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
*bvecs = bv;
*num_bvecs = npages;
return bytes;
}
static void put_bvecs(struct bio_vec *bvecs, int num_bvecs, bool should_dirty)
{
int i;
for (i = 0; i < num_bvecs; i++) {
if (bvecs[i].bv_page) {
if (should_dirty)
set_page_dirty_lock(bvecs[i].bv_page);
put_page(bvecs[i].bv_page);
}
}
kvfree(bvecs);
ceph: combine as many iovec as possile into one OSD request Both ceph_sync_direct_write and ceph_sync_read iterate iovec elements one by one, send one OSD request for each iovec. This is sub-optimal, We can combine serveral iovec into one page vector, and send an OSD request for the whole page vector. Signed-off-by: Zhu, Caifeng <zhucaifeng@unissoft-nj.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-10-08 15:26:15 +08:00
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/*
* Prepare an open request. Preallocate ceph_cap to avoid an
* inopportune ENOMEM later.
*/
static struct ceph_mds_request *
prepare_open_request(struct super_block *sb, int flags, int create_mode)
{
ceph: factor out libceph from Ceph file system This factors out protocol and low-level storage parts of ceph into a separate libceph module living in net/ceph and include/linux/ceph. This is mostly a matter of moving files around. However, a few key pieces of the interface change as well: - ceph_client becomes ceph_fs_client and ceph_client, where the latter captures the mon and osd clients, and the fs_client gets the mds client and file system specific pieces. - Mount option parsing and debugfs setup is correspondingly broken into two pieces. - The mon client gets a generic handler callback for otherwise unknown messages (mds map, in this case). - The basic supported/required feature bits can be expanded (and are by ceph_fs_client). No functional change, aside from some subtle error handling cases that got cleaned up in the refactoring process. Signed-off-by: Sage Weil <sage@newdream.net>
2010-04-07 06:14:15 +08:00
struct ceph_fs_client *fsc = ceph_sb_to_client(sb);
struct ceph_mds_client *mdsc = fsc->mdsc;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
struct ceph_mds_request *req;
int want_auth = USE_ANY_MDS;
int op = (flags & O_CREAT) ? CEPH_MDS_OP_CREATE : CEPH_MDS_OP_OPEN;
if (flags & (O_WRONLY|O_RDWR|O_CREAT|O_TRUNC))
want_auth = USE_AUTH_MDS;
req = ceph_mdsc_create_request(mdsc, op, want_auth);
if (IS_ERR(req))
goto out;
req->r_fmode = ceph_flags_to_mode(flags);
ceph: fix file open flags on ppc64 The file open flags (O_foo) are platform specific and should never go out to an interface that is not local to the system. Unfortunately these flags have leaked out onto the wire in the cephfs implementation. That lead to bogus flags getting transmitted on ppc64. This patch converts the kernel view of flags to the ceph view of file open flags. Fixes: 124e68e74 ("ceph: file operations") Signed-off-by: Alexander Graf <agraf@suse.de> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-04-28 00:34:00 +08:00
req->r_args.open.flags = ceph_flags_sys2wire(flags);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
req->r_args.open.mode = cpu_to_le32(create_mode);
out:
return req;
}
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
static int ceph_init_file_info(struct inode *inode, struct file *file,
int fmode, bool isdir)
{
ceph: track and report error of async metadata operation Use errseq_t to track and report errors of async metadata operations, similar to how kernel handles errors during writeback. If any dirty caps or any unsafe request gets dropped during session eviction, record -EIO in corresponding inode's i_meta_err. The error will be reported by subsequent fsync, Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:42 +08:00
struct ceph_inode_info *ci = ceph_inode(inode);
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
struct ceph_file_info *fi;
dout("%s %p %p 0%o (%s)\n", __func__, inode, file,
inode->i_mode, isdir ? "dir" : "regular");
BUG_ON(inode->i_fop->release != ceph_release);
if (isdir) {
struct ceph_dir_file_info *dfi =
kmem_cache_zalloc(ceph_dir_file_cachep, GFP_KERNEL);
ceph: simplify calling of ceph_get_fmode() Originally, calling ceph_get_fmode() for open files is by thread that handles request reply. There is a small window between updating caps and and waking the request initiator. We need to prevent ceph_check_caps() from releasing wanted caps in the window. Previous patches made fill_inode() call __ceph_touch_fmode() for open file requests. This prevented ceph_check_caps() from releasing wanted caps for 'caps_wanted_delay_min' seconds, enough for request initiator to get woken up and call ceph_get_fmode(). This allows us to now call ceph_get_fmode() in ceph_open() instead. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:02 +08:00
if (!dfi)
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
return -ENOMEM;
file->private_data = dfi;
fi = &dfi->file_info;
dfi->next_offset = 2;
dfi->readdir_cache_idx = -1;
} else {
fi = kmem_cache_zalloc(ceph_file_cachep, GFP_KERNEL);
ceph: simplify calling of ceph_get_fmode() Originally, calling ceph_get_fmode() for open files is by thread that handles request reply. There is a small window between updating caps and and waking the request initiator. We need to prevent ceph_check_caps() from releasing wanted caps in the window. Previous patches made fill_inode() call __ceph_touch_fmode() for open file requests. This prevented ceph_check_caps() from releasing wanted caps for 'caps_wanted_delay_min' seconds, enough for request initiator to get woken up and call ceph_get_fmode(). This allows us to now call ceph_get_fmode() in ceph_open() instead. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:02 +08:00
if (!fi)
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
return -ENOMEM;
file->private_data = fi;
}
ceph: simplify calling of ceph_get_fmode() Originally, calling ceph_get_fmode() for open files is by thread that handles request reply. There is a small window between updating caps and and waking the request initiator. We need to prevent ceph_check_caps() from releasing wanted caps in the window. Previous patches made fill_inode() call __ceph_touch_fmode() for open file requests. This prevented ceph_check_caps() from releasing wanted caps for 'caps_wanted_delay_min' seconds, enough for request initiator to get woken up and call ceph_get_fmode(). This allows us to now call ceph_get_fmode() in ceph_open() instead. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:02 +08:00
ceph_get_fmode(ci, fmode, 1);
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
fi->fmode = fmode;
ceph: simplify calling of ceph_get_fmode() Originally, calling ceph_get_fmode() for open files is by thread that handles request reply. There is a small window between updating caps and and waking the request initiator. We need to prevent ceph_check_caps() from releasing wanted caps in the window. Previous patches made fill_inode() call __ceph_touch_fmode() for open file requests. This prevented ceph_check_caps() from releasing wanted caps for 'caps_wanted_delay_min' seconds, enough for request initiator to get woken up and call ceph_get_fmode(). This allows us to now call ceph_get_fmode() in ceph_open() instead. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:02 +08:00
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
spin_lock_init(&fi->rw_contexts_lock);
INIT_LIST_HEAD(&fi->rw_contexts);
ceph: track and report error of async metadata operation Use errseq_t to track and report errors of async metadata operations, similar to how kernel handles errors during writeback. If any dirty caps or any unsafe request gets dropped during session eviction, record -EIO in corresponding inode's i_meta_err. The error will be reported by subsequent fsync, Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:42 +08:00
fi->meta_err = errseq_sample(&ci->i_meta_err);
ceph: invalidate all write mode filp after reconnect Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:46 +08:00
fi->filp_gen = READ_ONCE(ceph_inode_to_client(inode)->filp_gen);
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
return 0;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/*
* initialize private struct file data.
* if we fail, clean up by dropping fmode reference on the ceph_inode
*/
static int ceph_init_file(struct inode *inode, struct file *file, int fmode)
{
int ret = 0;
switch (inode->i_mode & S_IFMT) {
case S_IFREG:
ceph: disable fscache when inode is opened for write All other filesystems do not add dirty pages to fscache. They all disable fscache when inode is opened for write. Only ceph adds dirty pages to fscache, but the code is buggy. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-05-18 15:25:03 +08:00
ceph_fscache_register_inode_cookie(inode);
ceph_fscache_file_set_cookie(inode, file);
fs: mark expected switch fall-throughs In preparation to enabling -Wimplicit-fallthrough, mark switch cases where we are expecting to fall through. This patch fixes the following warnings: fs/affs/affs.h:124:38: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/configfs/dir.c:1692:11: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/configfs/dir.c:1694:7: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ceph/file.c:249:3: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ext4/hash.c:233:15: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ext4/hash.c:246:15: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ext2/inode.c:1237:7: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ext2/inode.c:1244:7: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ext4/indirect.c:1182:6: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ext4/indirect.c:1188:6: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ext4/indirect.c:1432:6: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ext4/indirect.c:1440:6: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/f2fs/node.c:618:8: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/f2fs/node.c:620:8: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/btrfs/ref-verify.c:522:15: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/gfs2/bmap.c:711:7: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/gfs2/bmap.c:722:7: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/jffs2/fs.c:339:6: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/nfsd/nfs4proc.c:429:12: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ufs/util.h:62:6: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/ufs/util.h:43:6: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/fcntl.c:770:7: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/seq_file.c:319:10: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/libfs.c:148:11: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/libfs.c:150:7: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/signalfd.c:178:7: warning: this statement may fall through [-Wimplicit-fallthrough=] fs/locks.c:1473:16: warning: this statement may fall through [-Wimplicit-fallthrough=] Warning level 3 was used: -Wimplicit-fallthrough=3 This patch is part of the ongoing efforts to enabling -Wimplicit-fallthrough. Reviewed-by: Kees Cook <keescook@chromium.org> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
2019-01-23 16:48:28 +08:00
/* fall through */
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
case S_IFDIR:
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
ret = ceph_init_file_info(inode, file, fmode,
S_ISDIR(inode->i_mode));
if (ret)
return ret;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
break;
case S_IFLNK:
dout("init_file %p %p 0%o (symlink)\n", inode, file,
inode->i_mode);
break;
default:
dout("init_file %p %p 0%o (special)\n", inode, file,
inode->i_mode);
/*
* we need to drop the open ref now, since we don't
* have .release set to ceph_release.
*/
BUG_ON(inode->i_fop->release == ceph_release);
/* call the proper open fop */
ret = inode->i_fop->open(inode, file);
}
return ret;
}
ceph: renew caps for read/write if mds session got killed. When mds session gets killed, read/write operation may hang. Client waits for Frw caps, but mds does not know what caps client wants. To recover this, client sends an open request to mds. The request will tell mds what caps client wants. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-04-08 15:27:16 +08:00
/*
* try renew caps after session gets killed.
*/
ceph: consider inode's last read/write when calculating wanted caps Add i_last_rd and i_last_wr to ceph_inode_info. These fields are used to track the last time the client acquired read/write caps for the inode. If there is no read/write on an inode for 'caps_wanted_delay_max' seconds, __ceph_caps_file_wanted() does not request caps for read/write even there are open files. Call __ceph_touch_fmode() for dir operations. __ceph_caps_file_wanted() calculates dir's wanted caps according to last dir read/modification. If there is recent dir read, dir inode wants CEPH_CAP_ANY_SHARED caps. If there is recent dir modification, also wants CEPH_CAP_FILE_EXCL. Readdir is a special case. Dir inode wants CEPH_CAP_FILE_EXCL after readdir, as with that, modifications do not need to release CEPH_CAP_FILE_SHARED or invalidate all dentry leases issued by readdir. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:00 +08:00
int ceph_renew_caps(struct inode *inode, int fmode)
ceph: renew caps for read/write if mds session got killed. When mds session gets killed, read/write operation may hang. Client waits for Frw caps, but mds does not know what caps client wants. To recover this, client sends an open request to mds. The request will tell mds what caps client wants. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-04-08 15:27:16 +08:00
{
ceph: consider inode's last read/write when calculating wanted caps Add i_last_rd and i_last_wr to ceph_inode_info. These fields are used to track the last time the client acquired read/write caps for the inode. If there is no read/write on an inode for 'caps_wanted_delay_max' seconds, __ceph_caps_file_wanted() does not request caps for read/write even there are open files. Call __ceph_touch_fmode() for dir operations. __ceph_caps_file_wanted() calculates dir's wanted caps according to last dir read/modification. If there is recent dir read, dir inode wants CEPH_CAP_ANY_SHARED caps. If there is recent dir modification, also wants CEPH_CAP_FILE_EXCL. Readdir is a special case. Dir inode wants CEPH_CAP_FILE_EXCL after readdir, as with that, modifications do not need to release CEPH_CAP_FILE_SHARED or invalidate all dentry leases issued by readdir. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:00 +08:00
struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
ceph: renew caps for read/write if mds session got killed. When mds session gets killed, read/write operation may hang. Client waits for Frw caps, but mds does not know what caps client wants. To recover this, client sends an open request to mds. The request will tell mds what caps client wants. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-04-08 15:27:16 +08:00
struct ceph_inode_info *ci = ceph_inode(inode);
struct ceph_mds_request *req;
int err, flags, wanted;
spin_lock(&ci->i_ceph_lock);
ceph: consider inode's last read/write when calculating wanted caps Add i_last_rd and i_last_wr to ceph_inode_info. These fields are used to track the last time the client acquired read/write caps for the inode. If there is no read/write on an inode for 'caps_wanted_delay_max' seconds, __ceph_caps_file_wanted() does not request caps for read/write even there are open files. Call __ceph_touch_fmode() for dir operations. __ceph_caps_file_wanted() calculates dir's wanted caps according to last dir read/modification. If there is recent dir read, dir inode wants CEPH_CAP_ANY_SHARED caps. If there is recent dir modification, also wants CEPH_CAP_FILE_EXCL. Readdir is a special case. Dir inode wants CEPH_CAP_FILE_EXCL after readdir, as with that, modifications do not need to release CEPH_CAP_FILE_SHARED or invalidate all dentry leases issued by readdir. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:00 +08:00
__ceph_touch_fmode(ci, mdsc, fmode);
ceph: renew caps for read/write if mds session got killed. When mds session gets killed, read/write operation may hang. Client waits for Frw caps, but mds does not know what caps client wants. To recover this, client sends an open request to mds. The request will tell mds what caps client wants. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-04-08 15:27:16 +08:00
wanted = __ceph_caps_file_wanted(ci);
if (__ceph_is_any_real_caps(ci) &&
ceph: fix wrong check in ceph_renew_caps() Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-03-25 17:28:10 +08:00
(!(wanted & CEPH_CAP_ANY_WR) || ci->i_auth_cap)) {
ceph: renew caps for read/write if mds session got killed. When mds session gets killed, read/write operation may hang. Client waits for Frw caps, but mds does not know what caps client wants. To recover this, client sends an open request to mds. The request will tell mds what caps client wants. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-04-08 15:27:16 +08:00
int issued = __ceph_caps_issued(ci, NULL);
spin_unlock(&ci->i_ceph_lock);
dout("renew caps %p want %s issued %s updating mds_wanted\n",
inode, ceph_cap_string(wanted), ceph_cap_string(issued));
ceph_check_caps(ci, 0, NULL);
return 0;
}
spin_unlock(&ci->i_ceph_lock);
flags = 0;
if ((wanted & CEPH_CAP_FILE_RD) && (wanted & CEPH_CAP_FILE_WR))
flags = O_RDWR;
else if (wanted & CEPH_CAP_FILE_RD)
flags = O_RDONLY;
else if (wanted & CEPH_CAP_FILE_WR)
flags = O_WRONLY;
#ifdef O_LAZY
if (wanted & CEPH_CAP_FILE_LAZYIO)
flags |= O_LAZY;
#endif
req = prepare_open_request(inode->i_sb, flags, 0);
if (IS_ERR(req)) {
err = PTR_ERR(req);
goto out;
}
req->r_inode = inode;
ihold(inode);
req->r_num_caps = 1;
err = ceph_mdsc_do_request(mdsc, NULL, req);
ceph_mdsc_put_request(req);
out:
dout("renew caps %p open result=%d\n", inode, err);
return err < 0 ? err : 0;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/*
* If we already have the requisite capabilities, we can satisfy
* the open request locally (no need to request new caps from the
* MDS). We do, however, need to inform the MDS (asynchronously)
* if our wanted caps set expands.
*/
int ceph_open(struct inode *inode, struct file *file)
{
struct ceph_inode_info *ci = ceph_inode(inode);
ceph: factor out libceph from Ceph file system This factors out protocol and low-level storage parts of ceph into a separate libceph module living in net/ceph and include/linux/ceph. This is mostly a matter of moving files around. However, a few key pieces of the interface change as well: - ceph_client becomes ceph_fs_client and ceph_client, where the latter captures the mon and osd clients, and the fs_client gets the mds client and file system specific pieces. - Mount option parsing and debugfs setup is correspondingly broken into two pieces. - The mon client gets a generic handler callback for otherwise unknown messages (mds map, in this case). - The basic supported/required feature bits can be expanded (and are by ceph_fs_client). No functional change, aside from some subtle error handling cases that got cleaned up in the refactoring process. Signed-off-by: Sage Weil <sage@newdream.net>
2010-04-07 06:14:15 +08:00
struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb);
struct ceph_mds_client *mdsc = fsc->mdsc;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
struct ceph_mds_request *req;
ceph: change variable name to follow common rule Variable name ci is mostly used for ceph_inode_info. Variable name fi is mostly used for ceph_file_info. Variable name cf is mostly used for ceph_cap_flush. Change variable name to follow above common rules in case of confusing. Signed-off-by: Chengguang Xu <cgxu519@icloud.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-02-28 19:43:47 +08:00
struct ceph_file_info *fi = file->private_data;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
int err;
int flags, fmode, wanted;
ceph: change variable name to follow common rule Variable name ci is mostly used for ceph_inode_info. Variable name fi is mostly used for ceph_file_info. Variable name cf is mostly used for ceph_cap_flush. Change variable name to follow above common rules in case of confusing. Signed-off-by: Chengguang Xu <cgxu519@icloud.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-02-28 19:43:47 +08:00
if (fi) {
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
dout("open file %p is already opened\n", file);
return 0;
}
/* filter out O_CREAT|O_EXCL; vfs did that already. yuck. */
flags = file->f_flags & ~(O_CREAT|O_EXCL);
if (S_ISDIR(inode->i_mode))
flags = O_DIRECTORY; /* mds likes to know */
dout("open inode %p ino %llx.%llx file %p flags %d (%d)\n", inode,
ceph_vinop(inode), file, flags, file->f_flags);
fmode = ceph_flags_to_mode(flags);
wanted = ceph_caps_for_mode(fmode);
/* snapped files are read-only */
if (ceph_snap(inode) != CEPH_NOSNAP && (file->f_mode & FMODE_WRITE))
return -EROFS;
/* trivially open snapdir */
if (ceph_snap(inode) == CEPH_SNAPDIR) {
return ceph_init_file(inode, file, fmode);
}
/*
ceph: fix open for write on clustered mds Normally when we open a file we already have a cap, and simply update the wanted set. However, if we open a file for write, but don't have an auth cap, that doesn't work; we need to open a new cap with the auth MDS. Only reuse existing caps if we are opening for read or the existing cap is auth. Signed-off-by: Sage Weil <sage@newdream.net>
2010-11-08 01:07:15 +08:00
* No need to block if we have caps on the auth MDS (for
* write) or any MDS (for read). Update wanted set
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
* asynchronously.
*/
ceph: use i_ceph_lock instead of i_lock We have been using i_lock to protect all kinds of data structures in the ceph_inode_info struct, including lists of inodes that we need to iterate over while avoiding races with inode destruction. That requires grabbing a reference to the inode with the list lock protected, but igrab() now takes i_lock to check the inode flags. Changing the list lock ordering would be a painful process. However, using a ceph-specific i_ceph_lock in the ceph inode instead of i_lock is a simple mechanical change and avoids the ordering constraints imposed by igrab(). Reported-by: Amon Ott <a.ott@m-privacy.de> Signed-off-by: Sage Weil <sage@newdream.net>
2011-12-01 01:47:09 +08:00
spin_lock(&ci->i_ceph_lock);
ceph: fix open for write on clustered mds Normally when we open a file we already have a cap, and simply update the wanted set. However, if we open a file for write, but don't have an auth cap, that doesn't work; we need to open a new cap with the auth MDS. Only reuse existing caps if we are opening for read or the existing cap is auth. Signed-off-by: Sage Weil <sage@newdream.net>
2010-11-08 01:07:15 +08:00
if (__ceph_is_any_real_caps(ci) &&
(((fmode & CEPH_FILE_MODE_WR) == 0) || ci->i_auth_cap)) {
ceph: avoid calling ceph_renew_caps() infinitely __ceph_caps_mds_wanted() ignores caps from stale session. So the return value of __ceph_caps_mds_wanted() can keep the same across ceph_renew_caps(). This causes try_get_cap_refs() to keep calling ceph_renew_caps(). The fix is ignore the session valid check for the try_get_cap_refs() case. If session is stale, just let the caps requester sleep. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2017-01-29 22:15:47 +08:00
int mds_wanted = __ceph_caps_mds_wanted(ci, true);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
int issued = __ceph_caps_issued(ci, NULL);
dout("open %p fmode %d want %s issued %s using existing\n",
inode, fmode, ceph_cap_string(wanted),
ceph_cap_string(issued));
ceph: simplify calling of ceph_get_fmode() Originally, calling ceph_get_fmode() for open files is by thread that handles request reply. There is a small window between updating caps and and waking the request initiator. We need to prevent ceph_check_caps() from releasing wanted caps in the window. Previous patches made fill_inode() call __ceph_touch_fmode() for open file requests. This prevented ceph_check_caps() from releasing wanted caps for 'caps_wanted_delay_min' seconds, enough for request initiator to get woken up and call ceph_get_fmode(). This allows us to now call ceph_get_fmode() in ceph_open() instead. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:02 +08:00
__ceph_touch_fmode(ci, mdsc, fmode);
ceph: use i_ceph_lock instead of i_lock We have been using i_lock to protect all kinds of data structures in the ceph_inode_info struct, including lists of inodes that we need to iterate over while avoiding races with inode destruction. That requires grabbing a reference to the inode with the list lock protected, but igrab() now takes i_lock to check the inode flags. Changing the list lock ordering would be a painful process. However, using a ceph-specific i_ceph_lock in the ceph inode instead of i_lock is a simple mechanical change and avoids the ordering constraints imposed by igrab(). Reported-by: Amon Ott <a.ott@m-privacy.de> Signed-off-by: Sage Weil <sage@newdream.net>
2011-12-01 01:47:09 +08:00
spin_unlock(&ci->i_ceph_lock);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/* adjust wanted? */
if ((issued & wanted) != wanted &&
(mds_wanted & wanted) != wanted &&
ceph_snap(inode) != CEPH_SNAPDIR)
ceph_check_caps(ci, 0, NULL);
return ceph_init_file(inode, file, fmode);
} else if (ceph_snap(inode) != CEPH_NOSNAP &&
(ci->i_snap_caps & wanted) == wanted) {
ceph: consider inode's last read/write when calculating wanted caps Add i_last_rd and i_last_wr to ceph_inode_info. These fields are used to track the last time the client acquired read/write caps for the inode. If there is no read/write on an inode for 'caps_wanted_delay_max' seconds, __ceph_caps_file_wanted() does not request caps for read/write even there are open files. Call __ceph_touch_fmode() for dir operations. __ceph_caps_file_wanted() calculates dir's wanted caps according to last dir read/modification. If there is recent dir read, dir inode wants CEPH_CAP_ANY_SHARED caps. If there is recent dir modification, also wants CEPH_CAP_FILE_EXCL. Readdir is a special case. Dir inode wants CEPH_CAP_FILE_EXCL after readdir, as with that, modifications do not need to release CEPH_CAP_FILE_SHARED or invalidate all dentry leases issued by readdir. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:00 +08:00
__ceph_touch_fmode(ci, mdsc, fmode);
ceph: use i_ceph_lock instead of i_lock We have been using i_lock to protect all kinds of data structures in the ceph_inode_info struct, including lists of inodes that we need to iterate over while avoiding races with inode destruction. That requires grabbing a reference to the inode with the list lock protected, but igrab() now takes i_lock to check the inode flags. Changing the list lock ordering would be a painful process. However, using a ceph-specific i_ceph_lock in the ceph inode instead of i_lock is a simple mechanical change and avoids the ordering constraints imposed by igrab(). Reported-by: Amon Ott <a.ott@m-privacy.de> Signed-off-by: Sage Weil <sage@newdream.net>
2011-12-01 01:47:09 +08:00
spin_unlock(&ci->i_ceph_lock);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
return ceph_init_file(inode, file, fmode);
}
ceph: use fscache as a local presisent cache Adding support for fscache to the Ceph filesystem. This would bring it to on par with some of the other network filesystems in Linux (like NFS, AFS, etc...) In order to mount the filesystem with fscache the 'fsc' mount option must be passed. Signed-off-by: Milosz Tanski <milosz@adfin.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-08-22 05:29:54 +08:00
ceph: use i_ceph_lock instead of i_lock We have been using i_lock to protect all kinds of data structures in the ceph_inode_info struct, including lists of inodes that we need to iterate over while avoiding races with inode destruction. That requires grabbing a reference to the inode with the list lock protected, but igrab() now takes i_lock to check the inode flags. Changing the list lock ordering would be a painful process. However, using a ceph-specific i_ceph_lock in the ceph inode instead of i_lock is a simple mechanical change and avoids the ordering constraints imposed by igrab(). Reported-by: Amon Ott <a.ott@m-privacy.de> Signed-off-by: Sage Weil <sage@newdream.net>
2011-12-01 01:47:09 +08:00
spin_unlock(&ci->i_ceph_lock);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
dout("open fmode %d wants %s\n", fmode, ceph_cap_string(wanted));
req = prepare_open_request(inode->i_sb, flags, 0);
if (IS_ERR(req)) {
err = PTR_ERR(req);
goto out;
}
ceph: use ihold when we already have an inode ref We should use ihold whenever we already have a stable inode ref, even when we aren't holding i_lock. This avoids adding new and unnecessary locking dependencies. Signed-off-by: Sage Weil <sage@newdream.net>
2011-05-28 00:24:26 +08:00
req->r_inode = inode;
ihold(inode);
ceph: use fscache as a local presisent cache Adding support for fscache to the Ceph filesystem. This would bring it to on par with some of the other network filesystems in Linux (like NFS, AFS, etc...) In order to mount the filesystem with fscache the 'fsc' mount option must be passed. Signed-off-by: Milosz Tanski <milosz@adfin.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-08-22 05:29:54 +08:00
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
req->r_num_caps = 1;
ceph: no need to get parent inode in ceph_open parent inode is needed in creating new inode case. For ceph_open, the target inode already exists. Signed-off-by: Jianpeng Ma <jianpeng.ma@intel.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-08-18 10:25:35 +08:00
err = ceph_mdsc_do_request(mdsc, NULL, req);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
if (!err)
err = ceph_init_file(inode, file, req->r_fmode);
ceph_mdsc_put_request(req);
dout("open result=%d on %llx.%llx\n", err, ceph_vinop(inode));
out:
return err;
}
ceph: cache layout in parent dir on first sync create If a create is done, then typically we'll end up writing to the file soon afterward. We don't want to wait for the reply before doing that when doing an async create, so that means we need the layout for the new file before we've gotten the response from the MDS. All files created in a directory will initially inherit the same layout, so copy off the requisite info from the first synchronous create in the directory, and save it in a new i_cached_layout field. Zero out the layout when we lose Dc caps in the dir. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-01-02 20:11:38 +08:00
/* Clone the layout from a synchronous create, if the dir now has Dc caps */
static void
cache_file_layout(struct inode *dst, struct inode *src)
{
struct ceph_inode_info *cdst = ceph_inode(dst);
struct ceph_inode_info *csrc = ceph_inode(src);
spin_lock(&cdst->i_ceph_lock);
if ((__ceph_caps_issued(cdst, NULL) & CEPH_CAP_DIR_CREATE) &&
!ceph_file_layout_is_valid(&cdst->i_cached_layout)) {
memcpy(&cdst->i_cached_layout, &csrc->i_layout,
sizeof(cdst->i_cached_layout));
rcu_assign_pointer(cdst->i_cached_layout.pool_ns,
ceph_try_get_string(csrc->i_layout.pool_ns));
}
spin_unlock(&cdst->i_ceph_lock);
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: attempt to do async create when possible With the Octopus release, the MDS will hand out directory create caps. If we have Fxc caps on the directory, and complete directory information or a known negative dentry, then we can return without waiting on the reply, allowing the open() call to return very quickly to userland. We use the normal ceph_fill_inode() routine to fill in the inode, so we have to gin up some reply inode information with what we'd expect the newly-created inode to have. The client assumes that it has a full set of caps on the new inode, and that the MDS will revoke them when there is conflicting access. This functionality is gated on the wsync/nowsync mount options. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-28 01:06:14 +08:00
/*
* Try to set up an async create. We need caps, a file layout, and inode number,
* and either a lease on the dentry or complete dir info. If any of those
* criteria are not satisfied, then return false and the caller can go
* synchronous.
*/
static int try_prep_async_create(struct inode *dir, struct dentry *dentry,
struct ceph_file_layout *lo, u64 *pino)
{
struct ceph_inode_info *ci = ceph_inode(dir);
struct ceph_dentry_info *di = ceph_dentry(dentry);
int got = 0, want = CEPH_CAP_FILE_EXCL | CEPH_CAP_DIR_CREATE;
u64 ino;
spin_lock(&ci->i_ceph_lock);
/* No auth cap means no chance for Dc caps */
if (!ci->i_auth_cap)
goto no_async;
/* Any delegated inos? */
if (xa_empty(&ci->i_auth_cap->session->s_delegated_inos))
goto no_async;
if (!ceph_file_layout_is_valid(&ci->i_cached_layout))
goto no_async;
if ((__ceph_caps_issued(ci, NULL) & want) != want)
goto no_async;
if (d_in_lookup(dentry)) {
if (!__ceph_dir_is_complete(ci))
goto no_async;
ceph: update dentry lease for async create Otherwise ceph_d_delete() may return 1 for the dentry, which makes dput() prune the dentry and clear parent dir's complete flag. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-04 21:22:20 +08:00
spin_lock(&dentry->d_lock);
di->lease_shared_gen = atomic_read(&ci->i_shared_gen);
spin_unlock(&dentry->d_lock);
ceph: attempt to do async create when possible With the Octopus release, the MDS will hand out directory create caps. If we have Fxc caps on the directory, and complete directory information or a known negative dentry, then we can return without waiting on the reply, allowing the open() call to return very quickly to userland. We use the normal ceph_fill_inode() routine to fill in the inode, so we have to gin up some reply inode information with what we'd expect the newly-created inode to have. The client assumes that it has a full set of caps on the new inode, and that the MDS will revoke them when there is conflicting access. This functionality is gated on the wsync/nowsync mount options. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-28 01:06:14 +08:00
} else if (atomic_read(&ci->i_shared_gen) !=
READ_ONCE(di->lease_shared_gen)) {
goto no_async;
}
ino = ceph_get_deleg_ino(ci->i_auth_cap->session);
if (!ino)
goto no_async;
*pino = ino;
ceph_take_cap_refs(ci, want, false);
memcpy(lo, &ci->i_cached_layout, sizeof(*lo));
rcu_assign_pointer(lo->pool_ns,
ceph_try_get_string(ci->i_cached_layout.pool_ns));
got = want;
no_async:
spin_unlock(&ci->i_ceph_lock);
return got;
}
static void restore_deleg_ino(struct inode *dir, u64 ino)
{
struct ceph_inode_info *ci = ceph_inode(dir);
struct ceph_mds_session *s = NULL;
spin_lock(&ci->i_ceph_lock);
if (ci->i_auth_cap)
s = ceph_get_mds_session(ci->i_auth_cap->session);
spin_unlock(&ci->i_ceph_lock);
if (s) {
int err = ceph_restore_deleg_ino(s, ino);
if (err)
pr_warn("ceph: unable to restore delegated ino 0x%llx to session: %d\n",
ino, err);
ceph_put_mds_session(s);
}
}
static void ceph_async_create_cb(struct ceph_mds_client *mdsc,
struct ceph_mds_request *req)
{
int result = req->r_err ? req->r_err :
le32_to_cpu(req->r_reply_info.head->result);
if (result == -EJUKEBOX)
goto out;
mapping_set_error(req->r_parent->i_mapping, result);
if (result) {
struct dentry *dentry = req->r_dentry;
ceph: fix potential bad pointer deref in async dirops cb's The new async dirops callback routines can pass ERR_PTR values to ceph_mdsc_free_path, which could cause an oops. Make ceph_mdsc_free_path ignore ERR_PTR values. Also, ensure that the pr_warn messages look sane even if ceph_mdsc_build_path fails. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-04-08 20:41:38 +08:00
int pathlen = 0;
u64 base = 0;
ceph: attempt to do async create when possible With the Octopus release, the MDS will hand out directory create caps. If we have Fxc caps on the directory, and complete directory information or a known negative dentry, then we can return without waiting on the reply, allowing the open() call to return very quickly to userland. We use the normal ceph_fill_inode() routine to fill in the inode, so we have to gin up some reply inode information with what we'd expect the newly-created inode to have. The client assumes that it has a full set of caps on the new inode, and that the MDS will revoke them when there is conflicting access. This functionality is gated on the wsync/nowsync mount options. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-28 01:06:14 +08:00
char *path = ceph_mdsc_build_path(req->r_dentry, &pathlen,
&base, 0);
ceph_dir_clear_complete(req->r_parent);
if (!d_unhashed(dentry))
d_drop(dentry);
/* FIXME: start returning I/O errors on all accesses? */
pr_warn("ceph: async create failure path=(%llx)%s result=%d!\n",
base, IS_ERR(path) ? "<<bad>>" : path, result);
ceph_mdsc_free_path(path, pathlen);
}
if (req->r_target_inode) {
struct ceph_inode_info *ci = ceph_inode(req->r_target_inode);
u64 ino = ceph_vino(req->r_target_inode).ino;
if (req->r_deleg_ino != ino)
pr_warn("%s: inode number mismatch! err=%d deleg_ino=0x%llx target=0x%llx\n",
__func__, req->r_err, req->r_deleg_ino, ino);
mapping_set_error(req->r_target_inode->i_mapping, result);
spin_lock(&ci->i_ceph_lock);
if (ci->i_ceph_flags & CEPH_I_ASYNC_CREATE) {
ci->i_ceph_flags &= ~CEPH_I_ASYNC_CREATE;
wake_up_bit(&ci->i_ceph_flags, CEPH_ASYNC_CREATE_BIT);
}
ceph_kick_flushing_inode_caps(req->r_session, ci);
spin_unlock(&ci->i_ceph_lock);
} else {
pr_warn("%s: no req->r_target_inode for 0x%llx\n", __func__,
req->r_deleg_ino);
}
out:
ceph_mdsc_release_dir_caps(req);
}
static int ceph_finish_async_create(struct inode *dir, struct dentry *dentry,
struct file *file, umode_t mode,
struct ceph_mds_request *req,
struct ceph_acl_sec_ctx *as_ctx,
struct ceph_file_layout *lo)
{
int ret;
char xattr_buf[4];
struct ceph_mds_reply_inode in = { };
struct ceph_mds_reply_info_in iinfo = { .in = &in };
struct ceph_inode_info *ci = ceph_inode(dir);
struct inode *inode;
struct timespec64 now;
struct ceph_vino vino = { .ino = req->r_deleg_ino,
.snap = CEPH_NOSNAP };
ktime_get_real_ts64(&now);
inode = ceph_get_inode(dentry->d_sb, vino);
if (IS_ERR(inode))
return PTR_ERR(inode);
iinfo.inline_version = CEPH_INLINE_NONE;
iinfo.change_attr = 1;
ceph_encode_timespec64(&iinfo.btime, &now);
iinfo.xattr_len = ARRAY_SIZE(xattr_buf);
iinfo.xattr_data = xattr_buf;
memset(iinfo.xattr_data, 0, iinfo.xattr_len);
in.ino = cpu_to_le64(vino.ino);
in.snapid = cpu_to_le64(CEPH_NOSNAP);
in.version = cpu_to_le64(1); // ???
in.cap.caps = in.cap.wanted = cpu_to_le32(CEPH_CAP_ALL_FILE);
in.cap.cap_id = cpu_to_le64(1);
in.cap.realm = cpu_to_le64(ci->i_snap_realm->ino);
in.cap.flags = CEPH_CAP_FLAG_AUTH;
in.ctime = in.mtime = in.atime = iinfo.btime;
in.mode = cpu_to_le32((u32)mode);
in.truncate_seq = cpu_to_le32(1);
in.truncate_size = cpu_to_le64(-1ULL);
in.xattr_version = cpu_to_le64(1);
in.uid = cpu_to_le32(from_kuid(&init_user_ns, current_fsuid()));
in.gid = cpu_to_le32(from_kgid(&init_user_ns, dir->i_mode & S_ISGID ?
dir->i_gid : current_fsgid()));
in.nlink = cpu_to_le32(1);
in.max_size = cpu_to_le64(lo->stripe_unit);
ceph_file_layout_to_legacy(lo, &in.layout);
ret = ceph_fill_inode(inode, NULL, &iinfo, NULL, req->r_session,
req->r_fmode, NULL);
if (ret) {
dout("%s failed to fill inode: %d\n", __func__, ret);
ceph_dir_clear_complete(dir);
if (!d_unhashed(dentry))
d_drop(dentry);
if (inode->i_state & I_NEW)
discard_new_inode(inode);
} else {
struct dentry *dn;
dout("%s d_adding new inode 0x%llx to 0x%lx/%s\n", __func__,
vino.ino, dir->i_ino, dentry->d_name.name);
ceph_dir_clear_ordered(dir);
ceph_init_inode_acls(inode, as_ctx);
if (inode->i_state & I_NEW) {
/*
* If it's not I_NEW, then someone created this before
* we got here. Assume the server is aware of it at
* that point and don't worry about setting
* CEPH_I_ASYNC_CREATE.
*/
ceph_inode(inode)->i_ceph_flags = CEPH_I_ASYNC_CREATE;
unlock_new_inode(inode);
}
if (d_in_lookup(dentry) || d_really_is_negative(dentry)) {
if (!d_unhashed(dentry))
d_drop(dentry);
dn = d_splice_alias(inode, dentry);
WARN_ON_ONCE(dn && dn != dentry);
}
file->f_mode |= FMODE_CREATED;
ret = finish_open(file, dentry, ceph_open);
}
return ret;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/*
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
* Do a lookup + open with a single request. If we get a non-existent
* file or symlink, return 1 so the VFS can retry.
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
*/
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
get rid of 'opened' argument of ->atomic_open() - part 3 now it can be done... Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-06-09 01:32:02 +08:00
struct file *file, unsigned flags, umode_t mode)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
{
ceph: factor out libceph from Ceph file system This factors out protocol and low-level storage parts of ceph into a separate libceph module living in net/ceph and include/linux/ceph. This is mostly a matter of moving files around. However, a few key pieces of the interface change as well: - ceph_client becomes ceph_fs_client and ceph_client, where the latter captures the mon and osd clients, and the fs_client gets the mds client and file system specific pieces. - Mount option parsing and debugfs setup is correspondingly broken into two pieces. - The mon client gets a generic handler callback for otherwise unknown messages (mds map, in this case). - The basic supported/required feature bits can be expanded (and are by ceph_fs_client). No functional change, aside from some subtle error handling cases that got cleaned up in the refactoring process. Signed-off-by: Sage Weil <sage@newdream.net>
2010-04-07 06:14:15 +08:00
struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
struct ceph_mds_client *mdsc = fsc->mdsc;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
struct ceph_mds_request *req;
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
struct dentry *dn;
ceph: rename struct ceph_acls_info to ceph_acl_sec_ctx Also rename ceph_release_acls_info() to ceph_release_acl_sec_ctx(). And move their definitions to different files. This is preparation for security label support. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-05-26 15:35:39 +08:00
struct ceph_acl_sec_ctx as_ctx = {};
ceph: attempt to do async create when possible With the Octopus release, the MDS will hand out directory create caps. If we have Fxc caps on the directory, and complete directory information or a known negative dentry, then we can return without waiting on the reply, allowing the open() call to return very quickly to userland. We use the normal ceph_fill_inode() routine to fill in the inode, so we have to gin up some reply inode information with what we'd expect the newly-created inode to have. The client assumes that it has a full set of caps on the new inode, and that the MDS will revoke them when there is conflicting access. This functionality is gated on the wsync/nowsync mount options. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-28 01:06:14 +08:00
bool try_async = ceph_test_mount_opt(fsc, ASYNC_DIROPS);
ceph: quota: support for ceph.quota.max_files This patch adds support for the max_files quota. It hooks into all the ceph functions that add new filesystem objects that need to be checked against the quota limits. When these limits are hit, -EDQUOT is returned. Note that we're not checking quotas on ceph_link(). ceph_link doesn't really create a new inode, and since the MDS doesn't update the directory statistics when a new (hard) link is created (only with symlinks), they are not accounted as a new file. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-01-05 18:47:19 +08:00
int mask;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
int err;
assorted conversions to %p[dD] Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-10-22 08:11:25 +08:00
dout("atomic_open %p dentry %p '%pd' %s flags %d mode 0%o\n",
dir, dentry, dentry,
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
d_unhashed(dentry) ? "unhashed" : "hashed", flags, mode);
if (dentry->d_name.len > NAME_MAX)
return -ENAMETOOLONG;
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
if (flags & O_CREAT) {
ceph: quota: support for ceph.quota.max_files This patch adds support for the max_files quota. It hooks into all the ceph functions that add new filesystem objects that need to be checked against the quota limits. When these limits are hit, -EDQUOT is returned. Note that we're not checking quotas on ceph_link(). ceph_link doesn't really create a new inode, and since the MDS doesn't update the directory statistics when a new (hard) link is created (only with symlinks), they are not accounted as a new file. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-01-05 18:47:19 +08:00
if (ceph_quota_is_max_files_exceeded(dir))
return -EDQUOT;
ceph: rename struct ceph_acls_info to ceph_acl_sec_ctx Also rename ceph_release_acls_info() to ceph_release_acl_sec_ctx(). And move their definitions to different files. This is preparation for security label support. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-05-26 15:35:39 +08:00
err = ceph_pre_init_acls(dir, &mode, &as_ctx);
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
if (err < 0)
return err;
ceph: add selinux support When creating new file/directory, use security_dentry_init_security() to prepare selinux context for the new inode, then send openc/mkdir request to MDS, together with selinux xattr. security_dentry_init_security() only supports single security module and only selinux has dentry_init_security hook. So only selinux is supported for now. We can add support for other security modules once kernel has a generic version of dentry_init_security() Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-05-26 16:27:56 +08:00
err = ceph_security_init_secctx(dentry, mode, &as_ctx);
if (err < 0)
goto out_ctx;
ceph: don't try to handle hashed dentries in non-O_CREAT atomic_open If ceph_atomic_open is handed a !d_in_lookup dentry, then that means that it already passed d_revalidate so we *know* that it's negative (or at least was very recently). Just return -ENOENT in that case. This also addresses a subtle bug in dentry handling. Non-O_CREAT opens call atomic_open with the parent's i_rwsem shared, but calling d_splice_alias on a hashed dentry requires the exclusive lock. If ceph_atomic_open receives a hashed, negative dentry on a non-O_CREAT open, and another client were to race in and create the file before we issue our OPEN, ceph_fill_trace could end up calling d_splice_alias on the dentry with the new inode with insufficient locks. Cc: stable@vger.kernel.org Reported-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-10-31 00:15:10 +08:00
} else if (!d_in_lookup(dentry)) {
/* If it's not being looked up, it's negative */
return -ENOENT;
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
}
ceph: attempt to do async create when possible With the Octopus release, the MDS will hand out directory create caps. If we have Fxc caps on the directory, and complete directory information or a known negative dentry, then we can return without waiting on the reply, allowing the open() call to return very quickly to userland. We use the normal ceph_fill_inode() routine to fill in the inode, so we have to gin up some reply inode information with what we'd expect the newly-created inode to have. The client assumes that it has a full set of caps on the new inode, and that the MDS will revoke them when there is conflicting access. This functionality is gated on the wsync/nowsync mount options. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-28 01:06:14 +08:00
retry:
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/* do the open */
req = prepare_open_request(dir->i_sb, flags, mode);
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
if (IS_ERR(req)) {
err = PTR_ERR(req);
ceph: rename struct ceph_acls_info to ceph_acl_sec_ctx Also rename ceph_release_acls_info() to ceph_release_acl_sec_ctx(). And move their definitions to different files. This is preparation for security label support. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-05-26 15:35:39 +08:00
goto out_ctx;
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
req->r_dentry = dget(dentry);
req->r_num_caps = 2;
ceph: attempt to do async create when possible With the Octopus release, the MDS will hand out directory create caps. If we have Fxc caps on the directory, and complete directory information or a known negative dentry, then we can return without waiting on the reply, allowing the open() call to return very quickly to userland. We use the normal ceph_fill_inode() routine to fill in the inode, so we have to gin up some reply inode information with what we'd expect the newly-created inode to have. The client assumes that it has a full set of caps on the new inode, and that the MDS will revoke them when there is conflicting access. This functionality is gated on the wsync/nowsync mount options. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-28 01:06:14 +08:00
mask = CEPH_STAT_CAP_INODE | CEPH_CAP_AUTH_SHARED;
if (ceph_security_xattr_wanted(dir))
mask |= CEPH_CAP_XATTR_SHARED;
req->r_args.open.mask = cpu_to_le32(mask);
req->r_parent = dir;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
if (flags & O_CREAT) {
ceph: attempt to do async create when possible With the Octopus release, the MDS will hand out directory create caps. If we have Fxc caps on the directory, and complete directory information or a known negative dentry, then we can return without waiting on the reply, allowing the open() call to return very quickly to userland. We use the normal ceph_fill_inode() routine to fill in the inode, so we have to gin up some reply inode information with what we'd expect the newly-created inode to have. The client assumes that it has a full set of caps on the new inode, and that the MDS will revoke them when there is conflicting access. This functionality is gated on the wsync/nowsync mount options. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-28 01:06:14 +08:00
struct ceph_file_layout lo;
ceph: voluntarily drop Ax cap for requests that create new inode MDS need to rdlock directory inode's authlock when handling these requests. Voluntarily dropping CEPH_CAP_AUTH_EXCL avoids a cap revoke message. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-11-23 17:47:15 +08:00
req->r_dentry_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_AUTH_EXCL;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
ceph: rename struct ceph_acls_info to ceph_acl_sec_ctx Also rename ceph_release_acls_info() to ceph_release_acl_sec_ctx(). And move their definitions to different files. This is preparation for security label support. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-05-26 15:35:39 +08:00
if (as_ctx.pagelist) {
req->r_pagelist = as_ctx.pagelist;
as_ctx.pagelist = NULL;
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
}
ceph: attempt to do async create when possible With the Octopus release, the MDS will hand out directory create caps. If we have Fxc caps on the directory, and complete directory information or a known negative dentry, then we can return without waiting on the reply, allowing the open() call to return very quickly to userland. We use the normal ceph_fill_inode() routine to fill in the inode, so we have to gin up some reply inode information with what we'd expect the newly-created inode to have. The client assumes that it has a full set of caps on the new inode, and that the MDS will revoke them when there is conflicting access. This functionality is gated on the wsync/nowsync mount options. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-28 01:06:14 +08:00
if (try_async &&
(req->r_dir_caps =
try_prep_async_create(dir, dentry, &lo,
&req->r_deleg_ino))) {
set_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags);
req->r_args.open.flags |= cpu_to_le32(CEPH_O_EXCL);
req->r_callback = ceph_async_create_cb;
err = ceph_mdsc_submit_request(mdsc, dir, req);
if (!err) {
err = ceph_finish_async_create(dir, dentry,
file, mode, req,
&as_ctx, &lo);
} else if (err == -EJUKEBOX) {
restore_deleg_ino(dir, req->r_deleg_ino);
ceph_mdsc_put_request(req);
try_async = false;
goto retry;
}
goto out_req;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
}
ceph: fix security xattr deadlock When security is enabled, security module can call filesystem's getxattr/setxattr callbacks during d_instantiate(). For cephfs, d_instantiate() is usually called by MDS' dispatch thread, while handling MDS reply. If the MDS reply does not include xattrs and corresponding caps, getxattr/setxattr need to send a new request to MDS and waits for the reply. This makes MDS' dispatch sleep, nobody handles later MDS replies. The fix is make sure lookup/atomic_open reply include xattrs and corresponding caps. So getxattr can be handled by cached xattrs. This requires some modification to both MDS and request message. (Client tells MDS what caps it wants; MDS encodes proper caps in the reply) Smack security module may call setxattr during d_instantiate(). Unlike getxattr, we can't force MDS to issue CEPH_CAP_XATTR_EXCL to us. So just make setxattr return error when called by MDS' dispatch thread. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-03-07 10:34:50 +08:00
ceph: add a new flag to indicate whether parent is locked struct ceph_mds_request has an r_locked_dir pointer, which is set to indicate the parent inode and that its i_rwsem is locked. In some critical places, we need to be able to indicate the parent inode to the request handling code, even when its i_rwsem may not be locked. Most of the code that operates on r_locked_dir doesn't require that the i_rwsem be locked. We only really need it to handle manipulation of the dcache. The rest (filling of the inode, updating dentry leases, etc.) already has its own locking. Add a new r_req_flags bit that indicates whether the parent is locked when doing the request, and rename the pointer to "r_parent". For now, all the places that set r_parent also set this flag, but that will change in a later patch. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Yan, Zheng <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-01-31 23:28:26 +08:00
set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags);
ceph: fix bad parent_inode calc in ceph_lookup_open We were always getting NULL here because the intent file f_dentry is always NULL at this point, which means we were always passing NULL to ceph_mdsc_do_request. In reality, this was fine, since this isn't currently ever a write operation that needs to get strung on the dir's unsafe list. Use the dir explicitly, and only pass it if this open has side-effects that a dir fsync should flush. Reviewed-by: Yehuda Sadeh <yehuda@hq.newdream.net> Signed-off-by: Sage Weil <sage@newdream.net>
2011-07-27 02:27:48 +08:00
err = ceph_mdsc_do_request(mdsc,
(flags & (O_CREAT|O_TRUNC)) ? dir : NULL,
req);
ceph: fix atomic_open snapdir ceph_handle_snapdir() checks ceph_mdsc_do_request()'s return value and creates snapdir inode if it's -ENOENT Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-01-19 13:23:20 +08:00
err = ceph_handle_snapdir(req, dentry, err);
ceph: Check for err on mds request in atomic_open The error returned by ceph_mdsc_do_request includes errors sending the request, errors on timeout, or any errors coming from the mds. If ceph_mdsc_do_request returns an error, the reply struct will most likely be bogus. We need to bail out and propogate the error instead of overwriting it. Signed-off-by: Sam Lang <sam.lang@inktank.com> Reviewed-by: Sage Weil <sage@inktank.com>
2012-12-20 03:44:23 +08:00
if (err)
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
goto out_req;
ceph: Check for err on mds request in atomic_open The error returned by ceph_mdsc_do_request includes errors sending the request, errors on timeout, or any errors coming from the mds. If ceph_mdsc_do_request returns an error, the reply struct will most likely be bogus. We need to bail out and propogate the error instead of overwriting it. Signed-off-by: Sam Lang <sam.lang@inktank.com> Reviewed-by: Sage Weil <sage@inktank.com>
2012-12-20 03:44:23 +08:00
ceph: remove the useless judgement err != 0 is already handled. So skip this. Signed-off-by: Jianpeng Ma <jianpeng.ma@intel.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-08-18 10:23:50 +08:00
if ((flags & O_CREAT) && !req->r_reply_info.head->is_dentry)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
err = ceph_handle_notrace_create(dir, dentry);
ceph: implement i_op->atomic_open() Add an ->atomic_open implementation which replaces the atomic lookup+open+create operation implemented via ->lookup and ->create operations. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> CC: Sage Weil <sage@newdream.net> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2012-06-05 21:10:25 +08:00
Use the right predicate in ->atomic_open() instances ->atomic_open() can be given an in-lookup dentry *or* a negative one found in dcache. Use d_in_lookup() to tell one from another, rather than d_unhashed(). Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-07-05 21:44:53 +08:00
if (d_in_lookup(dentry)) {
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
dn = ceph_finish_lookup(req, dentry, err);
if (IS_ERR(dn))
err = PTR_ERR(dn);
} else {
/* we were given a hashed negative dentry */
dn = NULL;
}
if (err)
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
goto out_req;
VFS: normal filesystems (and lustre): d_inode() annotations that's the bulk of filesystem drivers dealing with inodes of their own Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2015-03-18 06:25:59 +08:00
if (dn || d_really_is_negative(dentry) || d_is_symlink(dentry)) {
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
/* make vfs retry on splice, ENOENT, or symlink */
dout("atomic_open finish_no_open on dn %p\n", dn);
err = finish_no_open(file, dn);
} else {
dout("atomic_open finish_open on dn %p\n", dn);
ceph: Check for created flag in response from mds The mds now sends back a created inode if the create request performed the create. If the file already existed, no inode is returned in the reply. This allows ceph to set the created flag in atomic_open so that permissions are properly checked in the case that the file wasn't created by the create call to the mds. To ensure compability with previous kernels, a feature for sending back the inode in the create reply was added, so that the mds will only send back the inode if the client indicates it supports the feature. Signed-off-by: Sam Lang <sam.lang@inktank.com> Reviewed-by: Sage Weil <sage@inktank.com>
2012-12-29 01:56:46 +08:00
if (req->r_op == CEPH_MDS_OP_CREATE && req->r_reply_info.has_create_ino) {
ceph: cache layout in parent dir on first sync create If a create is done, then typically we'll end up writing to the file soon afterward. We don't want to wait for the reply before doing that when doing an async create, so that means we need the layout for the new file before we've gotten the response from the MDS. All files created in a directory will initially inherit the same layout, so copy off the requisite info from the first synchronous create in the directory, and save it in a new i_cached_layout field. Zero out the layout when we lose Dc caps in the dir. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-01-02 20:11:38 +08:00
struct inode *newino = d_inode(dentry);
cache_file_layout(dir, newino);
ceph_init_inode_acls(newino, &as_ctx);
introduce FMODE_CREATED and switch to it Parallel to FILE_CREATED, goes into ->f_mode instead of *opened. NFS is a bit of a wart here - it doesn't have file at the point where FILE_CREATED used to be set, so we need to propagate it there (for now). IMA is another one (here and everywhere)... Note that this needs do_dentry_open() to leave old bits in ->f_mode alone - we want it to preserve FMODE_CREATED if it had been already set (no other bit can be there). Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-06-09 01:22:02 +08:00
file->f_mode |= FMODE_CREATED;
ceph: Check for created flag in response from mds The mds now sends back a created inode if the create request performed the create. If the file already existed, no inode is returned in the reply. This allows ceph to set the created flag in atomic_open so that permissions are properly checked in the case that the file wasn't created by the create call to the mds. To ensure compability with previous kernels, a feature for sending back the inode in the create reply was added, so that the mds will only send back the inode if the client indicates it supports the feature. Signed-off-by: Sam Lang <sam.lang@inktank.com> Reviewed-by: Sage Weil <sage@inktank.com>
2012-12-29 01:56:46 +08:00
}
getting rid of 'opened' argument of ->atomic_open() - part 1 'opened' argument of finish_open() is unused. Kill it. Signed-off-by Al Viro <viro@zeniv.linux.org.uk>
2018-06-08 23:44:56 +08:00
err = finish_open(file, dentry, ceph_open);
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
}
ceph: include the initial ACL in create/mkdir/mknod MDS requests Current code set new file/directory's initial ACL in a non-atomic manner. Client first sends request to MDS to create new file/directory, then set the initial ACL after the new file/directory is successfully created. The fix is include the initial ACL in create/mkdir/mknod MDS requests. So MDS can handle creating file/directory and setting the initial ACL in one request. Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2014-09-16 20:35:17 +08:00
out_req:
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
ceph_mdsc_put_request(req);
ceph: rename struct ceph_acls_info to ceph_acl_sec_ctx Also rename ceph_release_acls_info() to ceph_release_acl_sec_ctx(). And move their definitions to different files. This is preparation for security label support. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-05-26 15:35:39 +08:00
out_ctx:
ceph_release_acl_sec_ctx(&as_ctx);
ceph: simplify+fix atomic_open The initial ->atomic_open op was carried over from the old intent code, which was incomplete and didn't really work. Replace it with a fresh method. In particular: * always attempt to do an atomic open+lookup, both for the create case and for lookups of existing files. * fix symlink handling by returning 1 to the VFS so that we can follow the link to its destination. This fixes a longstanding ceph bug (#2392). Signed-off-by: Sage Weil <sage@inktank.com>
2012-08-01 02:27:36 +08:00
dout("atomic_open result=%d\n", err);
make ->atomic_open() return int Change of calling conventions: old new NULL 1 file 0 ERR_PTR(-ve) -ve Caller *knows* that struct file *; no need to return it. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2012-06-22 16:39:14 +08:00
return err;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
}
int ceph_release(struct inode *inode, struct file *file)
{
struct ceph_inode_info *ci = ceph_inode(inode);
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
if (S_ISDIR(inode->i_mode)) {
struct ceph_dir_file_info *dfi = file->private_data;
dout("release inode %p dir file %p\n", inode, file);
WARN_ON(!list_empty(&dfi->file_info.rw_contexts));
ceph: consider inode's last read/write when calculating wanted caps Add i_last_rd and i_last_wr to ceph_inode_info. These fields are used to track the last time the client acquired read/write caps for the inode. If there is no read/write on an inode for 'caps_wanted_delay_max' seconds, __ceph_caps_file_wanted() does not request caps for read/write even there are open files. Call __ceph_touch_fmode() for dir operations. __ceph_caps_file_wanted() calculates dir's wanted caps according to last dir read/modification. If there is recent dir read, dir inode wants CEPH_CAP_ANY_SHARED caps. If there is recent dir modification, also wants CEPH_CAP_FILE_EXCL. Readdir is a special case. Dir inode wants CEPH_CAP_FILE_EXCL after readdir, as with that, modifications do not need to release CEPH_CAP_FILE_SHARED or invalidate all dentry leases issued by readdir. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:00 +08:00
ceph_put_fmode(ci, dfi->file_info.fmode, 1);
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
if (dfi->last_readdir)
ceph_mdsc_put_request(dfi->last_readdir);
kfree(dfi->last_name);
kfree(dfi->dir_info);
kmem_cache_free(ceph_dir_file_cachep, dfi);
} else {
struct ceph_file_info *fi = file->private_data;
dout("release inode %p regular file %p\n", inode, file);
WARN_ON(!list_empty(&fi->rw_contexts));
ceph: consider inode's last read/write when calculating wanted caps Add i_last_rd and i_last_wr to ceph_inode_info. These fields are used to track the last time the client acquired read/write caps for the inode. If there is no read/write on an inode for 'caps_wanted_delay_max' seconds, __ceph_caps_file_wanted() does not request caps for read/write even there are open files. Call __ceph_touch_fmode() for dir operations. __ceph_caps_file_wanted() calculates dir's wanted caps according to last dir read/modification. If there is recent dir read, dir inode wants CEPH_CAP_ANY_SHARED caps. If there is recent dir modification, also wants CEPH_CAP_FILE_EXCL. Readdir is a special case. Dir inode wants CEPH_CAP_FILE_EXCL after readdir, as with that, modifications do not need to release CEPH_CAP_FILE_SHARED or invalidate all dentry leases issued by readdir. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:00 +08:00
ceph_put_fmode(ci, fi->fmode, 1);
ceph: optimize memory usage In current code, regular file and directory use same struct ceph_file_info to store fs specific data so the struct has to include some fields which are only used for directory (e.g., readdir related info), when having plenty of regular files, it will lead to memory waste. This patch introduces dedicated ceph_dir_file_info cache for readdir related thins. So that regular file does not include those unused fields anymore. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-13 10:42:44 +08:00
kmem_cache_free(ceph_file_cachep, fi);
}
ceph: return EBADF if waiting for caps on closed file Verify the file is actually open for the given caps when we are waiting for caps. This ensures we will wake up and return EBADF if another thread closes the file out from under us. Note that EBADF is also the correct return code from write(2) when called on a file handle opened for reading (although the vfs should catch that). Signed-off-by: Sage Weil <sage@newdream.net>
2010-03-02 01:57:54 +08:00
/* wake up anyone waiting for caps on this inode */
ceph: use complete_all and wake_up_all This fixes an issue triggered by running concurrent syncs. One of the syncs would go through while the other would just hang indefinitely. In any case, we never actually want to wake a single waiter, so the *_all functions should be used. Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Signed-off-by: Sage Weil <sage@newdream.net>
2010-07-28 04:11:08 +08:00
wake_up_all(&ci->i_cap_wq);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
return 0;
}
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
enum {
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
HAVE_RETRIED = 1,
CHECK_EOF = 2,
READ_INLINE = 3,
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
};
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/*
* Completely synchronous read and write methods. Direct from __user
* buffer to osd, or directly to user pages (if O_DIRECT).
*
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
* If the read spans object boundary, just do multiple reads. (That's not
* atomic, but good enough for now.)
*
* If we get a short result from the OSD, check against i_size; we need to
* only return a short read to the caller if we hit EOF.
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
*/
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to,
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
int *retry_op)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
{
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
struct file *file = iocb->ki_filp;
new helper: file_inode(file) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-01-24 06:07:38 +08:00
struct inode *inode = file_inode(file);
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
struct ceph_inode_info *ci = ceph_inode(inode);
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
struct ceph_osd_client *osdc = &fsc->client->osdc;
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
ssize_t ret;
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
u64 off = iocb->ki_pos;
u64 len = iov_iter_count(to);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: include snapc in debug message of write Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-08-16 17:24:58 +08:00
dout("sync_read on file %p %llu~%u %s\n", file, off, (unsigned)len,
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
(file->f_flags & O_DIRECT) ? "O_DIRECT" : "");
ceph: check zero length in ceph_sync_read() Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
2014-07-21 10:15:48 +08:00
if (!len)
return 0;
ceph: fix comment, remove extraneous args The offset/length arguments aren't used. Signed-off-by: Sage Weil <sage@newdream.net>
2010-11-10 04:24:53 +08:00
/*
* flush any page cache pages in this range. this
* will make concurrent normal and sync io slow,
* but it will at least behave sensibly when they are
* in sequence.
*/
ceph: pass inclusive lend parameter to filemap_write_and_wait_range() The 'lend' parameter of filemap_write_and_wait_range is required to be inclusive, so follow the rule. Same for invalidate_inode_pages2_range. Signed-off-by: zhengbin <zhengbin13@huawei.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-02-01 19:19:15 +08:00
ret = filemap_write_and_wait_range(inode->i_mapping,
off, off + len - 1);
ceph: sync read/write considers page cache In the cases where we either do a sync read or a write, we need to make sure that everything in the page cache is flushed. In the case of a sync write we invalidate the relevant pages, so that subsequent read/write reflects the new data written. Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 03:14:41 +08:00
if (ret < 0)
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
return ret;
ceph: sync read/write considers page cache In the cases where we either do a sync read or a write, we need to make sure that everything in the page cache is flushed. In the case of a sync write we invalidate the relevant pages, so that subsequent read/write reflects the new data written. Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 03:14:41 +08:00
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
ret = 0;
while ((len = iov_iter_count(to)) > 0) {
struct ceph_osd_request *req;
struct page **pages;
int num_pages;
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
size_t page_off;
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
u64 i_size;
bool more;
req = ceph_osdc_new_request(osdc, &ci->i_layout,
ci->i_vino, off, &len, 0, 1,
CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
NULL, ci->i_truncate_seq,
ci->i_truncate_size, false);
if (IS_ERR(req)) {
ret = PTR_ERR(req);
break;
}
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
more = len < iov_iter_count(to);
Merge branch 'work.afs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs Pull AFS updates from Al Viro: "AFS series, with some iov_iter bits included" * 'work.afs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (26 commits) missing bits of "iov_iter: Separate type from direction and use accessor functions" afs: Probe multiple fileservers simultaneously afs: Fix callback handling afs: Eliminate the address pointer from the address list cursor afs: Allow dumping of server cursor on operation failure afs: Implement YFS support in the fs client afs: Expand data structure fields to support YFS afs: Get the target vnode in afs_rmdir() and get a callback on it afs: Calc callback expiry in op reply delivery afs: Fix FS.FetchStatus delivery from updating wrong vnode afs: Implement the YFS cache manager service afs: Remove callback details from afs_callback_break struct afs: Commit the status on a new file/dir/symlink afs: Increase to 64-bit volume ID and 96-bit vnode ID for YFS afs: Don't invoke the server to read data beyond EOF afs: Add a couple of tracepoints to log I/O errors afs: Handle EIO from delivery function afs: Fix TTL on VL server and address lists afs: Implement VL server rotation afs: Improve FS server rotation error handling ...
2018-11-02 10:58:52 +08:00
if (unlikely(iov_iter_is_pipe(to))) {
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
ret = iov_iter_get_pages_alloc(to, &pages, len,
&page_off);
if (ret <= 0) {
ceph_osdc_put_request(req);
ret = -ENOMEM;
break;
}
num_pages = DIV_ROUND_UP(ret + page_off, PAGE_SIZE);
if (ret < len) {
len = ret;
osd_req_op_extent_update(req, 0, len);
more = false;
}
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
} else {
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
num_pages = calc_pages_for(off, len);
page_off = off & ~PAGE_MASK;
pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
if (IS_ERR(pages)) {
ceph_osdc_put_request(req);
ret = PTR_ERR(pages);
break;
}
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
}
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_off,
false, false);
ret = ceph_osdc_start_request(osdc, req, false);
if (!ret)
ret = ceph_osdc_wait_request(osdc, req);
ceph: add read/write latency metric support Calculate the latency for OSD read requests. Add a new r_end_stamp field to struct ceph_osd_request that will hold the time of that the reply was received. Use that to calculate the RTT for each call, and divide the sum of those by number of calls to get averate RTT. Keep a tally of RTT for OSD writes and number of calls to track average latency of OSD writes. URL: https://tracker.ceph.com/issues/43215 Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-20 11:45:01 +08:00
ceph_update_read_latency(&fsc->mdsc->metric,
req->r_start_latency,
req->r_end_latency,
ret);
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
ceph_osdc_put_request(req);
i_size = i_size_read(inode);
dout("sync_read %llu~%llu got %zd i_size %llu%s\n",
off, len, ret, i_size, (more ? " MORE" : ""));
if (ret == -ENOENT)
ret = 0;
if (ret >= 0 && ret < len && (off + ret < i_size)) {
int zlen = min(len - ret, i_size - off - ret);
int zoff = page_off + ret;
dout("sync_read zero gap %llu~%llu\n",
off + ret, off + ret + zlen);
ceph_zero_page_vector_range(zoff, zlen, pages);
ret += zlen;
}
Merge branch 'work.afs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs Pull AFS updates from Al Viro: "AFS series, with some iov_iter bits included" * 'work.afs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (26 commits) missing bits of "iov_iter: Separate type from direction and use accessor functions" afs: Probe multiple fileservers simultaneously afs: Fix callback handling afs: Eliminate the address pointer from the address list cursor afs: Allow dumping of server cursor on operation failure afs: Implement YFS support in the fs client afs: Expand data structure fields to support YFS afs: Get the target vnode in afs_rmdir() and get a callback on it afs: Calc callback expiry in op reply delivery afs: Fix FS.FetchStatus delivery from updating wrong vnode afs: Implement the YFS cache manager service afs: Remove callback details from afs_callback_break struct afs: Commit the status on a new file/dir/symlink afs: Increase to 64-bit volume ID and 96-bit vnode ID for YFS afs: Don't invoke the server to read data beyond EOF afs: Add a couple of tracepoints to log I/O errors afs: Handle EIO from delivery function afs: Fix TTL on VL server and address lists afs: Implement VL server rotation afs: Improve FS server rotation error handling ...
2018-11-02 10:58:52 +08:00
if (unlikely(iov_iter_is_pipe(to))) {
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
if (ret > 0) {
iov_iter_advance(to, ret);
off += ret;
} else {
iov_iter_advance(to, 0);
}
ceph_put_page_vector(pages, num_pages, false);
} else {
int idx = 0;
size_t left = ret > 0 ? ret : 0;
while (left > 0) {
size_t len, copied;
page_off = off & ~PAGE_MASK;
len = min_t(size_t, left, PAGE_SIZE - page_off);
copied = copy_page_to_iter(pages[idx++],
page_off, len, to);
off += copied;
left -= copied;
if (copied < len) {
ret = -EFAULT;
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
break;
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
}
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
}
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
ceph_release_page_vector(pages, num_pages);
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
}
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
ceph: auto reconnect after blacklisted Make client use osd reply and session message to infer if itself is blacklisted. Client reconnect to cluster using new entity addr if it is blacklisted. Auto reconnect is limited to once every 30 minutes. Auto reconnect is disabled by default. It can be enabled/disabled by recover_session=<no|clean> mount option. In 'clean' mode, client drops any dirty data/metadata, invalidates page caches and invalidates all writable file handles. After reconnect, file locks become stale because MDS loses track of them. If an inode contains any stale file locks, read/write on the indoe are not allowed until applications release all stale file locks. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:47 +08:00
if (ret < 0) {
if (ret == -EBLACKLISTED)
fsc->blacklisted = true;
break;
}
if (off >= i_size || !more)
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
break;
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
if (off > iocb->ki_pos) {
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
if (ret >= 0 &&
iov_iter_count(to) > 0 && off >= i_size_read(inode))
*retry_op = CHECK_EOF;
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
ret = off - iocb->ki_pos;
iocb->ki_pos = off;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: refactor ceph_sync_read() Avoid allocating memory for the entire user request: striped_read() does a synchronous OSD request per object, so it doesn't need more than object size worth of pages at a time. [ Preserve the comment, changelog. ] Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-09-29 16:02:19 +08:00
dout("sync_read result %zd retry_op %d\n", ret, *retry_op);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
return ret;
}
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
struct ceph_aio_request {
struct kiocb *iocb;
size_t total_len;
ceph: only dirty ITER_IOVEC pages for direct read If a page is already locked, attempting to dirty it leads to a deadlock in lock_page(). This is what currently happens to ITER_BVEC pages when a dio-enabled loop device is backed by ceph: $ losetup --direct-io /dev/loop0 /mnt/cephfs/img $ xfs_io -c 'pread 0 4k' /dev/loop0 Follow other file systems and only dirty ITER_IOVEC pages. Cc: stable@kernel.org Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-16 11:22:29 +08:00
bool write;
bool should_dirty;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
int error;
struct list_head osd_reqs;
unsigned num_reqs;
atomic_t pending_reqs;
libceph: use timespec64 for r_mtime The request mtime field is used all over ceph, and is currently represented as a 'timespec' structure in Linux. This changes it to timespec64 to allow times beyond 2038, modifying all users at the same time. [ Remove now redundant ts variable in writepage_nounlock(). ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-14 04:18:37 +08:00
struct timespec64 mtime;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
struct ceph_cap_flush *prealloc_cf;
};
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
struct ceph_aio_work {
struct work_struct work;
struct ceph_osd_request *req;
};
static void ceph_aio_retry_work(struct work_struct *work);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
static void ceph_aio_complete(struct inode *inode,
struct ceph_aio_request *aio_req)
{
struct ceph_inode_info *ci = ceph_inode(inode);
int ret;
if (!atomic_dec_and_test(&aio_req->pending_reqs))
return;
ceph: increment/decrement dio counter on async requests Ceph can in some cases issue an async DIO request, in which case we can end up calling ceph_end_io_direct before the I/O is actually complete. That may allow buffered operations to proceed while DIO requests are still in flight. Fix this by incrementing the i_dio_count when issuing an async DIO request, and decrement it when tearing down the aio_req. Fixes: 321fe13c9398 ("ceph: add buffered/direct exclusionary locking for reads and writes") Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-13 22:56:06 +08:00
if (aio_req->iocb->ki_flags & IOCB_DIRECT)
inode_dio_end(inode);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
ret = aio_req->error;
if (!ret)
ret = aio_req->total_len;
dout("ceph_aio_complete %p rc %d\n", inode, ret);
if (ret >= 0 && aio_req->write) {
int dirty;
loff_t endoff = aio_req->iocb->ki_pos + aio_req->total_len;
if (endoff > i_size_read(inode)) {
if (ceph_inode_set_size(inode, endoff))
ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL);
}
spin_lock(&ci->i_ceph_lock);
ci->i_inline_version = CEPH_INLINE_NONE;
dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR,
&aio_req->prealloc_cf);
spin_unlock(&ci->i_ceph_lock);
if (dirty)
__mark_inode_dirty(inode, dirty);
}
ceph_put_cap_refs(ci, (aio_req->write ? CEPH_CAP_FILE_WR :
CEPH_CAP_FILE_RD));
aio_req->iocb->ki_complete(aio_req->iocb, ret, 0);
ceph_free_cap_flush(aio_req->prealloc_cf);
kfree(aio_req);
}
libceph: drop msg argument from ceph_osdc_callback_t finish_read(), its only user, uses it to get to hdr.data_len, which is what ->r_result is set to on success. This gains us the ability to safely call callbacks from contexts other than reply, e.g. map check. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-04-28 22:07:24 +08:00
static void ceph_aio_complete_req(struct ceph_osd_request *req)
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
{
int rc = req->r_result;
struct inode *inode = req->r_inode;
struct ceph_aio_request *aio_req = req->r_priv;
struct ceph_osd_data *osd_data = osd_req_op_extent_osd_data(req, 0);
ceph: add read/write latency metric support Calculate the latency for OSD read requests. Add a new r_end_stamp field to struct ceph_osd_request that will hold the time of that the reply was received. Use that to calculate the RTT for each call, and divide the sum of those by number of calls to get averate RTT. Keep a tally of RTT for OSD writes and number of calls to track average latency of OSD writes. URL: https://tracker.ceph.com/issues/43215 Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-20 11:45:01 +08:00
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
struct ceph_client_metric *metric = &fsc->mdsc->metric;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_BVECS);
BUG_ON(!osd_data->num_bvecs);
dout("ceph_aio_complete_req %p rc %d bytes %u\n",
inode, rc, osd_data->bvec_pos.iter.bi_size);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
ceph: add read/write latency metric support Calculate the latency for OSD read requests. Add a new r_end_stamp field to struct ceph_osd_request that will hold the time of that the reply was received. Use that to calculate the RTT for each call, and divide the sum of those by number of calls to get averate RTT. Keep a tally of RTT for OSD writes and number of calls to track average latency of OSD writes. URL: https://tracker.ceph.com/issues/43215 Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-20 11:45:01 +08:00
/* r_start_latency == 0 means the request was not submitted */
if (req->r_start_latency) {
if (aio_req->write)
ceph_update_write_latency(metric, req->r_start_latency,
req->r_end_latency, rc);
else
ceph_update_read_latency(metric, req->r_start_latency,
req->r_end_latency, rc);
}
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (rc == -EOLDSNAPC) {
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
struct ceph_aio_work *aio_work;
BUG_ON(!aio_req->write);
aio_work = kmalloc(sizeof(*aio_work), GFP_NOFS);
if (aio_work) {
INIT_WORK(&aio_work->work, ceph_aio_retry_work);
aio_work->req = req;
ceph: single workqueue for inode related works We have three workqueue for inode works. Later patch will introduce one more work for inode. It's not good to introcuce more workqueue and add more 'struct work_struct' to 'struct ceph_inode_info'. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-05-18 11:18:44 +08:00
queue_work(ceph_inode_to_client(inode)->inode_wq,
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
&aio_work->work);
return;
}
rc = -ENOMEM;
} else if (!aio_req->write) {
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (rc == -ENOENT)
rc = 0;
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
if (rc >= 0 && osd_data->bvec_pos.iter.bi_size > rc) {
struct iov_iter i;
int zlen = osd_data->bvec_pos.iter.bi_size - rc;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
/*
* If read is satisfied by single OSD request,
* it can pass EOF. Otherwise read is within
* i_size.
*/
if (aio_req->num_reqs == 1) {
loff_t i_size = i_size_read(inode);
loff_t endoff = aio_req->iocb->ki_pos + rc;
if (endoff < i_size)
zlen = min_t(size_t, zlen,
i_size - endoff);
aio_req->total_len = rc + zlen;
}
iov_iter: Separate type from direction and use accessor functions In the iov_iter struct, separate the iterator type from the iterator direction and use accessor functions to access them in most places. Convert a bunch of places to use switch-statements to access them rather then chains of bitwise-AND statements. This makes it easier to add further iterator types. Also, this can be more efficient as to implement a switch of small contiguous integers, the compiler can use ~50% fewer compare instructions than it has to use bitwise-and instructions. Further, cease passing the iterator type into the iterator setup function. The iterator function can set that itself. Only the direction is required. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
iov_iter_bvec(&i, READ, osd_data->bvec_pos.bvecs,
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
osd_data->num_bvecs,
osd_data->bvec_pos.iter.bi_size);
iov_iter_advance(&i, rc);
iov_iter_zero(zlen, &i);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
}
}
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
put_bvecs(osd_data->bvec_pos.bvecs, osd_data->num_bvecs,
aio_req->should_dirty);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
ceph_osdc_put_request(req);
if (rc < 0)
cmpxchg(&aio_req->error, 0, rc);
ceph_aio_complete(inode, aio_req);
return;
}
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
static void ceph_aio_retry_work(struct work_struct *work)
{
struct ceph_aio_work *aio_work =
container_of(work, struct ceph_aio_work, work);
struct ceph_osd_request *orig_req = aio_work->req;
struct ceph_aio_request *aio_req = orig_req->r_priv;
struct inode *inode = orig_req->r_inode;
struct ceph_inode_info *ci = ceph_inode(inode);
struct ceph_snap_context *snapc;
struct ceph_osd_request *req;
int ret;
spin_lock(&ci->i_ceph_lock);
if (__ceph_have_pending_cap_snap(ci)) {
struct ceph_cap_snap *capsnap =
list_last_entry(&ci->i_cap_snaps,
struct ceph_cap_snap,
ci_item);
snapc = ceph_get_snap_context(capsnap->context);
} else {
BUG_ON(!ci->i_head_snapc);
snapc = ceph_get_snap_context(ci->i_head_snapc);
}
spin_unlock(&ci->i_ceph_lock);
ceph: num_ops is off by one in ceph_aio_retry_work() Two OSD op slots are allocated, but only one is ever used. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-11 22:15:38 +08:00
req = ceph_osdc_alloc_request(orig_req->r_osdc, snapc, 1,
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
false, GFP_NOFS);
ceph: checking for IS_ERR instead of NULL ceph_osdc_alloc_request() returns NULL on error, it never returns error pointers. Fixes: 5be0389dac66 ('ceph: re-send AIO write request when getting -EOLDSNAP error') Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-01-26 17:24:44 +08:00
if (!req) {
ret = -ENOMEM;
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
req = orig_req;
goto out;
}
ceph: don't use CEPH_OSD_FLAG_ORDERSNAP Inode can be moved between snap realms. It's possible inode is moved into a snap realm whose seq number is smaller than old snap realm's. So there is no guarantee that seq number inode's snap context always increases. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-08-16 17:27:05 +08:00
req->r_flags = /* CEPH_OSD_FLAG_ORDERSNAP | */ CEPH_OSD_FLAG_WRITE;
libceph: introduce ceph_osd_request_target, calc_target() Introduce ceph_osd_request_target, containing all mapping-related fields of ceph_osd_request and calc_target() for calculating mappings and populating it. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-04-28 22:07:23 +08:00
ceph_oloc_copy(&req->r_base_oloc, &orig_req->r_base_oloc);
libceph: variable-sized ceph_object_id Currently ceph_object_id can hold object names of up to 100 (CEPH_MAX_OID_NAME_LEN) characters. This is enough for all use cases, expect one - long rbd image names: - a format 1 header is named "<imgname>.rbd" - an object that points to a format 2 header is named "rbd_id.<imgname>" We operate on these potentially long-named objects during rbd map, and, for format 1 images, during header refresh. (A format 2 header name is a small system-generated string.) Lift this 100 character limit by making ceph_object_id be able to point to an externally-allocated string. Apart from being able to work with almost arbitrarily-long named objects, this allows us to reduce the size of ceph_object_id from >100 bytes to 64 bytes. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-04-30 01:54:20 +08:00
ceph_oid_copy(&req->r_base_oid, &orig_req->r_base_oid);
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
libceph, rbd, ceph: move ceph_osdc_alloc_messages() calls The current requirement is that ceph_osdc_alloc_messages() should be called after oid and oloc are known. In preparation for preallocating message data items, move ceph_osdc_alloc_messages() further down, so that it is called when OSD op codes are known. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 22:11:37 +08:00
req->r_ops[0] = orig_req->r_ops[0];
req->r_mtime = aio_req->mtime;
req->r_data_offset = req->r_ops[0].extent.offset;
libceph: move message allocation out of ceph_osdc_alloc_request() The size of ->r_request and ->r_reply messages depends on the size of the object name (ceph_object_id), while the size of ceph_osd_request is fixed. Move message allocation into a separate function that would have to be called after ceph_object_id and ceph_object_locator (which is also going to become variable in size with RADOS namespaces) have been filled in: req = ceph_osdc_alloc_request(...); <fill in req->r_base_oid> <fill in req->r_base_oloc> ceph_osdc_alloc_messages(req); Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-04-27 20:15:51 +08:00
ret = ceph_osdc_alloc_messages(req, GFP_NOFS);
if (ret) {
ceph_osdc_put_request(req);
req = orig_req;
goto out;
}
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
ceph_osdc_put_request(orig_req);
req->r_callback = ceph_aio_complete_req;
req->r_inode = inode;
req->r_priv = aio_req;
ret = ceph_osdc_start_request(req->r_osdc, req, false);
out:
if (ret < 0) {
req->r_result = ret;
libceph: drop msg argument from ceph_osdc_callback_t finish_read(), its only user, uses it to get to hdr.data_len, which is what ->r_result is set to on success. This gains us the ability to safely call callbacks from contexts other than reply, e.g. map check. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-04-28 22:07:24 +08:00
ceph_aio_complete_req(req);
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
}
ceph: fix snap context leak in error path Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-01-26 23:05:37 +08:00
ceph_put_snap_context(snapc);
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
kfree(aio_work);
}
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
static ssize_t
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter,
struct ceph_snap_context *snapc,
struct ceph_cap_flush **pcf)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
{
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
struct file *file = iocb->ki_filp;
new helper: file_inode(file) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-01-24 06:07:38 +08:00
struct inode *inode = file_inode(file);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
struct ceph_inode_info *ci = ceph_inode(inode);
ceph: factor out libceph from Ceph file system This factors out protocol and low-level storage parts of ceph into a separate libceph module living in net/ceph and include/linux/ceph. This is mostly a matter of moving files around. However, a few key pieces of the interface change as well: - ceph_client becomes ceph_fs_client and ceph_client, where the latter captures the mon and osd clients, and the fs_client gets the mds client and file system specific pieces. - Mount option parsing and debugfs setup is correspondingly broken into two pieces. - The mon client gets a generic handler callback for otherwise unknown messages (mds map, in this case). - The basic supported/required feature bits can be expanded (and are by ceph_fs_client). No functional change, aside from some subtle error handling cases that got cleaned up in the refactoring process. Signed-off-by: Sage Weil <sage@newdream.net>
2010-04-07 06:14:15 +08:00
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
ceph: add read/write latency metric support Calculate the latency for OSD read requests. Add a new r_end_stamp field to struct ceph_osd_request that will hold the time of that the reply was received. Use that to calculate the RTT for each call, and divide the sum of those by number of calls to get averate RTT. Keep a tally of RTT for OSD writes and number of calls to track average latency of OSD writes. URL: https://tracker.ceph.com/issues/43215 Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-20 11:45:01 +08:00
struct ceph_client_metric *metric = &fsc->mdsc->metric;
libceph: don't build request in ceph_osdc_new_request() This patch moves the call to ceph_osdc_build_request() out of ceph_osdc_new_request() and into its caller. This is in order to defer formatting osd operation information into the request message until just before request is started. The only unusual (ab)user of ceph_osdc_build_request() is ceph_writepages_start(), where the final length of write request may change (downward) based on the current inode size or the oldest snapshot context with dirty data for the inode. The remaining callers don't change anything in the request after has been built. This means the ops array is now supplied by the caller. It also means there is no need to pass the mtime to ceph_osdc_new_request() (it gets provided to ceph_osdc_build_request()). And rather than passing a do_sync flag, have the number of ops in the ops array supplied imply adding a second STARTSYNC operation after the READ or WRITE requested. This and some of the patches that follow are related to having the messenger (only) be responsible for filling the content of the message header, as described here: http://tracker.ceph.com/issues/4589 Signed-off-by: Alex Elder <elder@inktank.com> Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
2013-03-15 03:09:05 +08:00
struct ceph_vino vino;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
struct ceph_osd_request *req;
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
struct bio_vec *bvecs;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
struct ceph_aio_request *aio_req = NULL;
int num_pages = 0;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
int flags;
ceph: add buffered/direct exclusionary locking for reads and writes xfstest generic/451 intermittently fails. The test does O_DIRECT writes to a file, and then reads back the result using buffered I/O, while running a separate set of tasks that are also doing buffered reads. The client will invalidate the cache prior to a direct write, but it's easy for one of the other readers' replies to race in and reinstantiate the invalidated range with stale data. To fix this, we must to serialize direct I/O writes and buffered reads. We could just sprinkle in some shared locks on the i_rwsem for reads, and increase the exclusive footprint on the write side, but that would cause O_DIRECT writes to end up serialized vs. other direct requests. Instead, borrow the scheme used by nfs.ko. Buffered writes take the i_rwsem exclusively, but buffered reads take a shared lock, allowing them to run in parallel. O_DIRECT requests also take a shared lock, but we need for them to not run in parallel with buffered reads. A flag on the ceph_inode_info is used to indicate whether it's in direct or buffered I/O mode. When a conflicting request is submitted, it will block until the inode can be flipped to the necessary mode. Link: https://tracker.ceph.com/issues/40985 Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-08-03 01:15:39 +08:00
int ret = 0;
libceph: use timespec64 for r_mtime The request mtime field is used all over ceph, and is currently represented as a 'timespec' structure in Linux. This changes it to timespec64 to allow times beyond 2038, modifying all users at the same time. [ Remove now redundant ts variable in writepage_nounlock(). ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-14 04:18:37 +08:00
struct timespec64 mtime = current_time(inode);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
size_t count = iov_iter_count(iter);
loff_t pos = iocb->ki_pos;
bool write = iov_iter_rw(iter) == WRITE;
ceph: only dirty ITER_IOVEC pages for direct read If a page is already locked, attempting to dirty it leads to a deadlock in lock_page(). This is what currently happens to ITER_BVEC pages when a dio-enabled loop device is backed by ceph: $ losetup --direct-io /dev/loop0 /mnt/cephfs/img $ xfs_io -c 'pread 0 4k' /dev/loop0 Follow other file systems and only dirty ITER_IOVEC pages. Cc: stable@kernel.org Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-16 11:22:29 +08:00
bool should_dirty = !write && iter_is_iovec(iter);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (write && ceph_snap(file_inode(file)) != CEPH_NOSNAP)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
return -EROFS;
ceph: include snapc in debug message of write Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-08-16 17:24:58 +08:00
dout("sync_direct_%s on file %p %lld~%u snapc %p seq %lld\n",
(write ? "write" : "read"), file, pos, (unsigned)count,
ceph: fix NULL pointer deref when debugging is enabled Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-04-24 02:18:45 +08:00
snapc, snapc ? snapc->seq : 0);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (write) {
ceph: ignore error from invalidate_inode_pages2_range() in direct write This call can fail if there are dirty pages. The preceding call to filemap_write_and_wait_range() will normally remove dirty pages, but as inode_lock() is not held over calls to ceph_direct_read_write(), it could race with non-direct writes and pages could be dirtied immediately after filemap_write_and_wait_range() returns If there are dirty pages, they will be removed by the subsequent call to truncate_inode_pages_range(), so having them here is not a problem. If the 'ret' value is left holding an error, then in the async IO case (aio_req is not NULL) the loop that would normally call ceph_osdc_start_request() will see the error in 'ret' and abort all requests. This doesn't seem like correct behaviour. So use separate 'ret2' instead of overloading 'ret'. Signed-off-by: NeilBrown <neilb@suse.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Yan, Zheng <zyan@redhat.com>
2016-09-01 22:26:23 +08:00
int ret2 = invalidate_inode_pages2_range(inode->i_mapping,
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
pos >> PAGE_SHIFT,
ceph: pass inclusive lend parameter to filemap_write_and_wait_range() The 'lend' parameter of filemap_write_and_wait_range is required to be inclusive, so follow the rule. Same for invalidate_inode_pages2_range. Signed-off-by: zhengbin <zhengbin13@huawei.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-02-01 19:19:15 +08:00
(pos + count - 1) >> PAGE_SHIFT);
ceph: ignore error from invalidate_inode_pages2_range() in direct write This call can fail if there are dirty pages. The preceding call to filemap_write_and_wait_range() will normally remove dirty pages, but as inode_lock() is not held over calls to ceph_direct_read_write(), it could race with non-direct writes and pages could be dirtied immediately after filemap_write_and_wait_range() returns If there are dirty pages, they will be removed by the subsequent call to truncate_inode_pages_range(), so having them here is not a problem. If the 'ret' value is left holding an error, then in the async IO case (aio_req is not NULL) the loop that would normally call ceph_osdc_start_request() will see the error in 'ret' and abort all requests. This doesn't seem like correct behaviour. So use separate 'ret2' instead of overloading 'ret'. Signed-off-by: NeilBrown <neilb@suse.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Yan, Zheng <zyan@redhat.com>
2016-09-01 22:26:23 +08:00
if (ret2 < 0)
ceph: fix printing wrong return variable in ceph_direct_read_write() Fix printing wrong return variable for invalidate_inode_pages2_range in ceph_direct_read_write(). Signed-off-by: Zhi Zhang <zhang.david2011@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-11-08 21:35:59 +08:00
dout("invalidate_inode_pages2_range returned %d\n", ret2);
ceph: sync read/write considers page cache In the cases where we either do a sync read or a write, we need to make sure that everything in the page cache is flushed. In the case of a sync write we invalidate the relevant pages, so that subsequent read/write reflects the new data written. Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 03:14:41 +08:00
ceph: don't use CEPH_OSD_FLAG_ORDERSNAP Inode can be moved between snap realms. It's possible inode is moved into a snap realm whose seq number is smaller than old snap realm's. So there is no guarantee that seq number inode's snap context always increases. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-08-16 17:27:05 +08:00
flags = /* CEPH_OSD_FLAG_ORDERSNAP | */ CEPH_OSD_FLAG_WRITE;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
} else {
flags = CEPH_OSD_FLAG_READ;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
while (iov_iter_count(iter) > 0) {
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
u64 size = iov_iter_count(iter);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
ssize_t len;
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ceph: fix rsize/wsize capping in ceph_direct_read_write() rsize/wsize cap should be applied before ceph_osdc_new_request() is called. Otherwise, if the size is limited by the cap instead of the stripe unit, ceph_osdc_new_request() would setup an extent op that is bigger than what dio_get_pages_alloc() would pin and add to the page vector, triggering asserts in the messenger. Cc: stable@vger.kernel.org Fixes: 95cca2b44e54 ("ceph: limit osd write size") Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com>
2018-05-03 22:10:09 +08:00
if (write)
size = min_t(u64, size, fsc->mount_options->wsize);
else
size = min_t(u64, size, fsc->mount_options->rsize);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
vino = ceph_vino(inode);
req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
vino, pos, &size, 0,
ceph: nuke startsync op startsync is a no-op, has been for years. Remove it. Link: http://tracker.ceph.com/issues/20604 Signed-off-by: Yanhu Cao <gmayyyha@gmail.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-07-21 17:20:10 +08:00
1,
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
write ? CEPH_OSD_OP_WRITE :
CEPH_OSD_OP_READ,
flags, snapc,
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ci->i_truncate_seq,
ci->i_truncate_size,
false);
if (IS_ERR(req)) {
ret = PTR_ERR(req);
ceph_sync_{,direct_}write: fix an oops on ceph_osdc_new_request() failure ceph_osdc_put_request(ERR_PTR(-error)) oopses. What we want there is break, not goto out. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 10:44:19 +08:00
break;
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
len = iter_get_bvecs_alloc(iter, size, &bvecs, &num_pages);
if (len < 0) {
ceph_sync_direct_write: stop poking into iov_iter guts all needed primitives are there... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 10:58:25 +08:00
ceph_osdc_put_request(req);
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
ret = len;
ceph_sync_direct_write: stop poking into iov_iter guts all needed primitives are there... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 10:58:25 +08:00
break;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
}
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
if (len != size)
osd_req_op_extent_update(req, 0, len);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
/*
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
* To simplify error handling, allow AIO when IO within i_size
* or IO can be satisfied by single OSD request.
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
*/
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (pos == iocb->ki_pos && !is_sync_kiocb(iocb) &&
(len == count || pos + count <= i_size_read(inode))) {
aio_req = kzalloc(sizeof(*aio_req), GFP_KERNEL);
if (aio_req) {
aio_req->iocb = iocb;
aio_req->write = write;
ceph: only dirty ITER_IOVEC pages for direct read If a page is already locked, attempting to dirty it leads to a deadlock in lock_page(). This is what currently happens to ITER_BVEC pages when a dio-enabled loop device is backed by ceph: $ losetup --direct-io /dev/loop0 /mnt/cephfs/img $ xfs_io -c 'pread 0 4k' /dev/loop0 Follow other file systems and only dirty ITER_IOVEC pages. Cc: stable@kernel.org Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-03-16 11:22:29 +08:00
aio_req->should_dirty = should_dirty;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
INIT_LIST_HEAD(&aio_req->osd_reqs);
if (write) {
ceph: re-send AIO write request when getting -EOLDSNAP error When receiving -EOLDSNAP from OSD, we need to re-send corresponding write request. Due to locking issue, we can send new request inside another OSD request's complete callback. So we use worker to re-send request for AIO write. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-24 08:44:20 +08:00
aio_req->mtime = mtime;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
swap(aio_req->prealloc_cf, *pcf);
}
}
/* ignore error */
}
if (write) {
/*
* throw out any page cache pages in this range. this
* may block.
*/
truncate_inode_pages_range(inode->i_mapping, pos,
ceph: fix end offset in truncate_inode_pages_range call Commit e450f4d1a5d6 ("ceph: pass inclusive lend parameter to filemap_write_and_wait_range()") fixed the end offset parameter used to call filemap_write_and_wait_range and invalidate_inode_pages2_range. Unfortunately it missed truncate_inode_pages_range, introducing a regression that is easily detected by xfstest generic/130. The problem is that when doing direct IO it is possible that an extra page is truncated from the page cache when the end offset is page aligned. This can cause data loss if that page hasn't been sync'ed to the OSDs. While there, change code to use PAGE_ALIGN macro instead. Cc: stable@vger.kernel.org Fixes: e450f4d1a5d6 ("ceph: pass inclusive lend parameter to filemap_write_and_wait_range()") Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-02 01:16:34 +08:00
PAGE_ALIGN(pos + len) - 1);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
libceph: switch to calc_target(), part 2 The crux of this is getting rid of ceph_osdc_build_request(), so that MOSDOp can be encoded not before but after calc_target() calculates the actual target. Encoding now happens within ceph_osdc_start_request(). Also nuked is the accompanying bunch of pointers into the encoded buffer that was used to update fields on each send - instead, the entire front is re-encoded. If we want to support target->name_len != base->name_len in the future, there is no other way, because oid is surrounded by other fields in the encoded buffer. Encoding OSD ops and adding data items to the request message were mixed together in osd_req_encode_op(). While we want to re-encode OSD ops, we don't want to add duplicate data items to the message when resending, so all call to ceph_osdc_msg_data_add() are factored out into a new setup_request_data(). Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-05-26 06:29:52 +08:00
req->r_mtime = mtime;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
}
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
osd_req_op_extent_osd_data_bvecs(req, 0, bvecs, num_pages, len);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (aio_req) {
aio_req->total_len += len;
aio_req->num_reqs++;
atomic_inc(&aio_req->pending_reqs);
req->r_callback = ceph_aio_complete_req;
req->r_inode = inode;
req->r_priv = aio_req;
libceph: rename r_unsafe_item to r_private_item This list item remained from when we had safe and unsafe replies (commit vs ack). It has since become a private list item for use by clients. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-08 18:50:09 +08:00
list_add_tail(&req->r_private_item, &aio_req->osd_reqs);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
pos += len;
continue;
}
ret = ceph_osdc_start_request(req->r_osdc, req, false);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
if (!ret)
ret = ceph_osdc_wait_request(&fsc->client->osdc, req);
ceph: add read/write latency metric support Calculate the latency for OSD read requests. Add a new r_end_stamp field to struct ceph_osd_request that will hold the time of that the reply was received. Use that to calculate the RTT for each call, and divide the sum of those by number of calls to get averate RTT. Keep a tally of RTT for OSD writes and number of calls to track average latency of OSD writes. URL: https://tracker.ceph.com/issues/43215 Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-20 11:45:01 +08:00
if (write)
ceph_update_write_latency(metric, req->r_start_latency,
req->r_end_latency, ret);
else
ceph_update_read_latency(metric, req->r_start_latency,
req->r_end_latency, ret);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
size = i_size_read(inode);
if (!write) {
if (ret == -ENOENT)
ret = 0;
if (ret >= 0 && ret < len && pos + ret < size) {
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
struct iov_iter i;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
int zlen = min_t(size_t, len - ret,
size - pos - ret);
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
iov_iter: Separate type from direction and use accessor functions In the iov_iter struct, separate the iterator type from the iterator direction and use accessor functions to access them in most places. Convert a bunch of places to use switch-statements to access them rather then chains of bitwise-AND statements. This makes it easier to add further iterator types. Also, this can be more efficient as to implement a switch of small contiguous integers, the compiler can use ~50% fewer compare instructions than it has to use bitwise-and instructions. Further, cease passing the iterator type into the iterator setup function. The iterator function can set that itself. Only the direction is required. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
iov_iter_bvec(&i, READ, bvecs, num_pages, len);
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
iov_iter_advance(&i, ret);
iov_iter_zero(zlen, &i);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
ret += zlen;
}
if (ret >= 0)
len = ret;
}
ceph: fix iov_iter issues in ceph_direct_read_write() dio_get_pagev_size() and dio_get_pages_alloc() introduced in commit b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") assume that the passed iov_iter is ITER_IOVEC. This isn't the case with splice where it ends up poking into the guts of ITER_BVEC or ITER_PIPE iterators, causing lockups and crashes easily reproduced with generic/095. Rather than trying to figure out gap alignment and stuff pages into a page vector, add a helper for going from iov_iter to a bio_vec array and make use of the new CEPH_OSD_DATA_TYPE_BVECS code. Fixes: b5b98989dc7e ("ceph: combine as many iovec as possile into one OSD request") Link: http://tracker.ceph.com/issues/18130 Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Tested-by: Luis Henriques <lhenriques@suse.com>
2018-05-04 22:57:31 +08:00
put_bvecs(bvecs, num_pages, should_dirty);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ceph_osdc_put_request(req);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (ret < 0)
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
break;
ceph_sync_direct_write: stop poking into iov_iter guts all needed primitives are there... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 10:58:25 +08:00
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
pos += len;
if (!write && pos >= size)
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
break;
ceph_sync_direct_write: stop poking into iov_iter guts all needed primitives are there... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 10:58:25 +08:00
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (write && pos > size) {
if (ceph_inode_set_size(inode, pos))
ceph_sync_direct_write: stop poking into iov_iter guts all needed primitives are there... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 10:58:25 +08:00
ceph_check_caps(ceph_inode(inode),
CHECK_CAPS_AUTHONLY,
NULL);
}
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
}
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (aio_req) {
ceph: fix use-after-free bug in ceph_direct_read_write() ceph_aio_complete() can free the ceph_aio_request struct before the code exits the while loop. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-06-14 11:13:59 +08:00
LIST_HEAD(osd_reqs);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (aio_req->num_reqs == 0) {
kfree(aio_req);
return ret;
}
ceph_get_cap_refs(ci, write ? CEPH_CAP_FILE_WR :
CEPH_CAP_FILE_RD);
ceph: fix use-after-free bug in ceph_direct_read_write() ceph_aio_complete() can free the ceph_aio_request struct before the code exits the while loop. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-06-14 11:13:59 +08:00
list_splice(&aio_req->osd_reqs, &osd_reqs);
ceph: increment/decrement dio counter on async requests Ceph can in some cases issue an async DIO request, in which case we can end up calling ceph_end_io_direct before the I/O is actually complete. That may allow buffered operations to proceed while DIO requests are still in flight. Fix this by incrementing the i_dio_count when issuing an async DIO request, and decrement it when tearing down the aio_req. Fixes: 321fe13c9398 ("ceph: add buffered/direct exclusionary locking for reads and writes") Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-13 22:56:06 +08:00
inode_dio_begin(inode);
ceph: fix use-after-free bug in ceph_direct_read_write() ceph_aio_complete() can free the ceph_aio_request struct before the code exits the while loop. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-06-14 11:13:59 +08:00
while (!list_empty(&osd_reqs)) {
req = list_first_entry(&osd_reqs,
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
struct ceph_osd_request,
libceph: rename r_unsafe_item to r_private_item This list item remained from when we had safe and unsafe replies (commit vs ack). It has since become a private list item for use by clients. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-08 18:50:09 +08:00
r_private_item);
list_del_init(&req->r_private_item);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (ret >= 0)
ret = ceph_osdc_start_request(req->r_osdc,
req, false);
if (ret < 0) {
req->r_result = ret;
libceph: drop msg argument from ceph_osdc_callback_t finish_read(), its only user, uses it to get to hdr.data_len, which is what ->r_result is set to on success. This gains us the ability to safely call callbacks from contexts other than reply, e.g. map check. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-04-28 22:07:24 +08:00
ceph_aio_complete_req(req);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
}
}
return -EIOCBQUEUED;
}
if (ret != -EOLDSNAPC && pos > iocb->ki_pos) {
ret = pos - iocb->ki_pos;
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
iocb->ki_pos = pos;
}
return ret;
}
/*
* Synchronous write, straight from __user pointer or user pages.
*
* If write spans object boundary, just do multiple writes. (For a
* correct atomic write, we should e.g. take write locks on all
* objects, rollback on failure, etc.)
*/
ceph: fix append mode write generic_write_checks() may update 'pos', so we need to pass 'pos' to ceph_sync_write() and ceph_sync_direct_write(); Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
2014-07-28 14:33:46 +08:00
static ssize_t
ceph: set i_head_snapc when getting CEPH_CAP_FILE_WR reference In most cases that snap context is needed, we are holding reference of CEPH_CAP_FILE_WR. So we can set ceph inode's i_head_snapc when getting the CEPH_CAP_FILE_WR reference, and make codes get snap context from i_head_snapc. This makes the code simpler. Another benefit of this change is that we can handle snap notification more elegantly. Especially when snap context is updated while someone else is doing write. The old queue cap_snap code may set cap_snap's context to ether the old context or the new snap context, depending on if i_head_snapc is set. The new queue capp_snap code always set cap_snap's context to the old snap context. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-04-30 14:40:54 +08:00
ceph_sync_write(struct kiocb *iocb, struct iov_iter *from, loff_t pos,
struct ceph_snap_context *snapc)
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
{
struct file *file = iocb->ki_filp;
struct inode *inode = file_inode(file);
struct ceph_inode_info *ci = ceph_inode(inode);
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
struct ceph_vino vino;
struct ceph_osd_request *req;
struct page **pages;
u64 len;
int num_pages;
int written = 0;
int flags;
int ret;
ceph: update the 'approaching max_size' code The old 'approaching max_size' code expects MDS set max_size to '2 * reported_size'. This is no longer true. The new code reports file size when half of previous max_size increment has been used. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-05-22 12:03:32 +08:00
bool check_caps = false;
libceph: use timespec64 for r_mtime The request mtime field is used all over ceph, and is currently represented as a 'timespec' structure in Linux. This changes it to timespec64 to allow times beyond 2038, modifying all users at the same time. [ Remove now redundant ts variable in writepage_nounlock(). ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-14 04:18:37 +08:00
struct timespec64 mtime = current_time(inode);
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
size_t count = iov_iter_count(from);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
if (ceph_snap(file_inode(file)) != CEPH_NOSNAP)
return -EROFS;
ceph: include snapc in debug message of write Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-08-16 17:24:58 +08:00
dout("sync_write on file %p %lld~%u snapc %p seq %lld\n",
file, pos, (unsigned)count, snapc, snapc->seq);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ceph: pass inclusive lend parameter to filemap_write_and_wait_range() The 'lend' parameter of filemap_write_and_wait_range is required to be inclusive, so follow the rule. Same for invalidate_inode_pages2_range. Signed-off-by: zhengbin <zhengbin13@huawei.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-02-01 19:19:15 +08:00
ret = filemap_write_and_wait_range(inode->i_mapping,
pos, pos + count - 1);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
if (ret < 0)
return ret;
ret = invalidate_inode_pages2_range(inode->i_mapping,
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
pos >> PAGE_SHIFT,
ceph: pass inclusive lend parameter to filemap_write_and_wait_range() The 'lend' parameter of filemap_write_and_wait_range is required to be inclusive, so follow the rule. Same for invalidate_inode_pages2_range. Signed-off-by: zhengbin <zhengbin13@huawei.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-02-01 19:19:15 +08:00
(pos + count - 1) >> PAGE_SHIFT);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
if (ret < 0)
dout("invalidate_inode_pages2_range returned %d\n", ret);
ceph: don't use CEPH_OSD_FLAG_ORDERSNAP Inode can be moved between snap realms. It's possible inode is moved into a snap realm whose seq number is smaller than old snap realm's. So there is no guarantee that seq number inode's snap context always increases. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-08-16 17:27:05 +08:00
flags = /* CEPH_OSD_FLAG_ORDERSNAP | */ CEPH_OSD_FLAG_WRITE;
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
while ((len = iov_iter_count(from)) > 0) {
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
size_t left;
int n;
vino = ceph_vino(inode);
req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
libceph: specify position of extent operation allow specifying position of extent operation in multi-operations osd request. This is required for cephfs to convert inline data to normal data (compare xattr, then write object). Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Ilya Dryomov <idryomov@redhat.com>
2014-11-13 14:40:37 +08:00
vino, pos, &len, 0, 1,
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
CEPH_OSD_OP_WRITE, flags, snapc,
ci->i_truncate_seq,
ci->i_truncate_size,
false);
if (IS_ERR(req)) {
ret = PTR_ERR(req);
ceph_sync_{,direct_}write: fix an oops on ceph_osdc_new_request() failure ceph_osdc_put_request(ERR_PTR(-error)) oopses. What we want there is break, not goto out. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 10:44:19 +08:00
break;
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
}
/*
* write from beginning of first page,
* regardless of io alignment
*/
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
num_pages = (len + PAGE_SIZE - 1) >> PAGE_SHIFT;
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ceph: switch some GFP_NOFS memory allocation to GFP_KERNEL GFP_NOFS memory allocation is required for page writeback path. But there is no need to use GFP_NOFS in syscall path and readpage path Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-13 17:27:05 +08:00
pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
if (IS_ERR(pages)) {
ret = PTR_ERR(pages);
goto out;
}
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
left = len;
for (n = 0; n < num_pages; n++) {
ceph: cast PAGE_SIZE to size_t in ceph_sync_write() Use min_t(size_t, ...) instead of plain min(), which does strict type checking, to avoid compile warning on i386. Cc: Jianpeng Ma <majianpeng@gmail.com> Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com> Reviewed-by: Sage Weil <sage@inktank.com>
2014-01-29 01:16:18 +08:00
size_t plen = min_t(size_t, left, PAGE_SIZE);
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
ret = copy_page_from_iter(pages[n], 0, plen, from);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
if (ret != plen) {
ret = -EFAULT;
break;
}
left -= ret;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
if (ret < 0) {
ceph_release_page_vector(pages, num_pages);
goto out;
}
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
req->r_inode = inode;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
osd_req_op_extent_osd_data_pages(req, 0, pages, len, 0,
false, true);
libceph: hold off building osd request Defer building the osd request until just before submitting it in all callers except ceph_writepages_start(). (That caller will be handed in the next patch.) Signed-off-by: Alex Elder <elder@inktank.com> Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
2013-03-15 03:09:06 +08:00
libceph: switch to calc_target(), part 2 The crux of this is getting rid of ceph_osdc_build_request(), so that MOSDOp can be encoded not before but after calc_target() calculates the actual target. Encoding now happens within ceph_osdc_start_request(). Also nuked is the accompanying bunch of pointers into the encoded buffer that was used to update fields on each send - instead, the entire front is re-encoded. If we want to support target->name_len != base->name_len in the future, there is no other way, because oid is surrounded by other fields in the encoded buffer. Encoding OSD ops and adding data items to the request message were mixed together in osd_req_encode_op(). While we want to re-encode OSD ops, we don't want to add duplicate data items to the message when resending, so all call to ceph_osdc_msg_data_add() are factored out into a new setup_request_data(). Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-05-26 06:29:52 +08:00
req->r_mtime = mtime;
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ret = ceph_osdc_start_request(&fsc->client->osdc, req, false);
if (!ret)
ret = ceph_osdc_wait_request(&fsc->client->osdc, req);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: add read/write latency metric support Calculate the latency for OSD read requests. Add a new r_end_stamp field to struct ceph_osd_request that will hold the time of that the reply was received. Use that to calculate the RTT for each call, and divide the sum of those by number of calls to get averate RTT. Keep a tally of RTT for OSD writes and number of calls to track average latency of OSD writes. URL: https://tracker.ceph.com/issues/43215 Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-20 11:45:01 +08:00
ceph_update_write_latency(&fsc->mdsc->metric, req->r_start_latency,
req->r_end_latency, ret);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
out:
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
ceph_osdc_put_request(req);
ceph: when seeing write errors on an inode, switch to sync writes Currently, we don't have a real feedback mechanism in place for when we start seeing buffered writeback errors. If writeback is failing, there is nothing that prevents an application from continuing to dirty pages that aren't being cleaned. In the event that we're seeing write errors of any sort occur on an inode, have the callback set a flag to force further writes to be synchronous. When the next write succeeds, clear the flag to allow buffered writeback to continue. Since this is just a hint to the write submission mechanism, we only take the i_ceph_lock when a lockless check shows that the flag needs to be changed. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng” <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-04-04 20:39:46 +08:00
if (ret != 0) {
ceph_set_error_write(ci);
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
break;
ceph: when seeing write errors on an inode, switch to sync writes Currently, we don't have a real feedback mechanism in place for when we start seeing buffered writeback errors. If writeback is failing, there is nothing that prevents an application from continuing to dirty pages that aren't being cleaned. In the event that we're seeing write errors of any sort occur on an inode, have the callback set a flag to force further writes to be synchronous. When the next write succeeds, clear the flag to allow buffered writeback to continue. Since this is just a hint to the write submission mechanism, we only take the i_ceph_lock when a lockless check shows that the flag needs to be changed. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng” <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-04-04 20:39:46 +08:00
}
ceph_clear_error_write(ci);
pos += len;
written += len;
if (pos > i_size_read(inode)) {
check_caps = ceph_inode_set_size(inode, pos);
if (check_caps)
ceph_check_caps(ceph_inode(inode),
CHECK_CAPS_AUTHONLY,
NULL);
}
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
if (ret != -EOLDSNAPC && written > 0) {
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ret = written;
ceph: Implement writev/pwritev for sync operation. For writev/pwritev sync-operatoin, ceph only do the first iov. I divided the write-sync-operation into two functions. One for direct-write, other for none-direct-sync-write. This is because for none-direct-sync-write we can merge iovs to one. But for direct-write, we can't merge iovs. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-09-12 13:54:26 +08:00
iocb->ki_pos = pos;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
}
return ret;
}
/*
* Wrap generic_file_aio_read with checks for cap bits on the inode.
* Atomically grab references, so that those bits are not released
* back to the MDS mid-read.
*
* Hmm, the sync read case isn't actually async... should it be?
*/
ceph: switch to ->read_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-03 08:28:01 +08:00
static ssize_t ceph_read_iter(struct kiocb *iocb, struct iov_iter *to)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
{
struct file *filp = iocb->ki_filp;
ceph: perform lazy reads when file mode and caps permit If the file mode is marked as "lazy," perform cached/buffered reads when the caps permit it. Adjust the rdcache_gen and invalidation logic accordingly so that we manage our cache based on the FILE_CACHE -or- FILE_LAZYIO cap bits. Signed-off-by: Sage Weil <sage@newdream.net>
2010-05-28 01:40:43 +08:00
struct ceph_file_info *fi = filp->private_data;
fs: remove ki_nbytes There is no need to pass the total request length in the kiocb, as we already get passed in through the iov_iter argument. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2015-02-12 02:56:46 +08:00
size_t len = iov_iter_count(to);
new helper: file_inode(file) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-01-24 06:07:38 +08:00
struct inode *inode = file_inode(filp);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
struct ceph_inode_info *ci = ceph_inode(inode);
ceph: fetch inline data when getting Fcr cap refs we can't use getattr to fetch inline data after getting Fcr caps, because it can cause deadlock. The solution is try bringing inline data to page cache when not holding any cap, and hope the inline data page is still there after getting the Fcr caps. If the page is still there, pin it in page cache for later IO. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:10:07 +08:00
struct page *pinned_page = NULL;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ssize_t ret;
ceph: perform lazy reads when file mode and caps permit If the file mode is marked as "lazy," perform cached/buffered reads when the caps permit it. Adjust the rdcache_gen and invalidation logic accordingly so that we manage our cache based on the FILE_CACHE -or- FILE_LAZYIO cap bits. Signed-off-by: Sage Weil <sage@newdream.net>
2010-05-28 01:40:43 +08:00
int want, got = 0;
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
int retry_op = 0, read = 0;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: fix sync read eof check deadlock If a sync read gets a short result from the OSD, it may need to do a getattr to see if it is short due to reaching end-of-file. The getattr was being done while holding a reference to FILE_RD, which can lead to a deadlock if the MDS is revoking that capability bit and can't process the getattr until it does. We fix this by setting a flag if EOF size validation is needed, and doing the getattr in ceph_aio_read, after the RD cap ref is dropped. If the read needs to be continued, we loop and continue traversing the file. Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 06:04:02 +08:00
again:
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
dout("aio_read %p %llx.%llx %llu~%u trying to get caps on %p\n",
inode, ceph_vinop(inode), iocb->ki_pos, (unsigned)len, inode);
ceph: take the inode lock before acquiring cap refs Most of the time, we (or the vfs layer) takes the inode_lock and then acquires caps, but ceph_read_iter does the opposite, and that can lead to a deadlock. When there are multiple clients treading over the same data, we can end up in a situation where a reader takes caps and then tries to acquire the inode_lock. Another task holds the inode_lock and issues a request to the MDS which needs to revoke the caps, but that can't happen until the inode_lock is unwedged. Fix this by having ceph_read_iter take the inode_lock earlier, before attempting to acquire caps. Fixes: 321fe13c9398 ("ceph: add buffered/direct exclusionary locking for reads and writes") Link: https://tracker.ceph.com/issues/36348 Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-13 22:10:27 +08:00
if (iocb->ki_flags & IOCB_DIRECT)
ceph_start_io_direct(inode);
else
ceph_start_io_read(inode);
ceph: perform lazy reads when file mode and caps permit If the file mode is marked as "lazy," perform cached/buffered reads when the caps permit it. Adjust the rdcache_gen and invalidation logic accordingly so that we manage our cache based on the FILE_CACHE -or- FILE_LAZYIO cap bits. Signed-off-by: Sage Weil <sage@newdream.net>
2010-05-28 01:40:43 +08:00
if (fi->fmode & CEPH_FILE_MODE_LAZY)
want = CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO;
else
want = CEPH_CAP_FILE_CACHE;
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
ret = ceph_get_caps(filp, CEPH_CAP_FILE_RD, want, -1,
&got, &pinned_page);
ceph: take the inode lock before acquiring cap refs Most of the time, we (or the vfs layer) takes the inode_lock and then acquires caps, but ceph_read_iter does the opposite, and that can lead to a deadlock. When there are multiple clients treading over the same data, we can end up in a situation where a reader takes caps and then tries to acquire the inode_lock. Another task holds the inode_lock and issues a request to the MDS which needs to revoke the caps, but that can't happen until the inode_lock is unwedged. Fix this by having ceph_read_iter take the inode_lock earlier, before attempting to acquire caps. Fixes: 321fe13c9398 ("ceph: add buffered/direct exclusionary locking for reads and writes") Link: https://tracker.ceph.com/issues/36348 Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-13 22:10:27 +08:00
if (ret < 0) {
if (iocb->ki_flags & IOCB_DIRECT)
ceph_end_io_direct(inode);
else
ceph_end_io_read(inode);
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
return ret;
ceph: take the inode lock before acquiring cap refs Most of the time, we (or the vfs layer) takes the inode_lock and then acquires caps, but ceph_read_iter does the opposite, and that can lead to a deadlock. When there are multiple clients treading over the same data, we can end up in a situation where a reader takes caps and then tries to acquire the inode_lock. Another task holds the inode_lock and issues a request to the MDS which needs to revoke the caps, but that can't happen until the inode_lock is unwedged. Fix this by having ceph_read_iter take the inode_lock earlier, before attempting to acquire caps. Fixes: 321fe13c9398 ("ceph: add buffered/direct exclusionary locking for reads and writes") Link: https://tracker.ceph.com/issues/36348 Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-13 22:10:27 +08:00
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: perform lazy reads when file mode and caps permit If the file mode is marked as "lazy," perform cached/buffered reads when the caps permit it. Adjust the rdcache_gen and invalidation logic accordingly so that we manage our cache based on the FILE_CACHE -or- FILE_LAZYIO cap bits. Signed-off-by: Sage Weil <sage@newdream.net>
2010-05-28 01:40:43 +08:00
if ((got & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) == 0 ||
mirror O_APPEND and O_DIRECT into iocb->ki_flags ... avoiding write_iter/fcntl races. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2015-04-10 01:52:01 +08:00
(iocb->ki_flags & IOCB_DIRECT) ||
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
(fi->flags & CEPH_F_SYNC)) {
dout("aio_sync_read %p %llx.%llx %llu~%u got cap refs on %s\n",
inode, ceph_vinop(inode), iocb->ki_pos, (unsigned)len,
ceph_cap_string(got));
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
if (ci->i_inline_version == CEPH_INLINE_NONE) {
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (!retry_op && (iocb->ki_flags & IOCB_DIRECT)) {
ret = ceph_direct_read_write(iocb, to,
NULL, NULL);
if (ret >= 0 && ret < len)
retry_op = CHECK_EOF;
} else {
ret = ceph_sync_read(iocb, to, &retry_op);
}
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
} else {
retry_op = READ_INLINE;
}
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
} else {
ceph: track read contexts in ceph_file_info Previously ceph_read_iter() uses current->journal to pass context info to ceph_readpages(), so that ceph_readpages() can distinguish read(2) from readahead(2)/fadvise(2)/madvise(2). The problem is that page fault can happen when copying data to userspace memory. Page fault may call other filesystem's page_mkwrite() if the userspace memory is mapped to a file. The later filesystem may also want to use current->journal. The fix is define a on-stack data structure in ceph_read_iter(), add it to context list in ceph_file_info. ceph_readpages() searches the list, find if there is a context belongs to current thread. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-12-15 11:15:36 +08:00
CEPH_DEFINE_RW_CONTEXT(rw_ctx, got);
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
dout("aio_read %p %llx.%llx %llu~%u got cap refs on %s\n",
ceph: switch to ->read_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-03 08:28:01 +08:00
inode, ceph_vinop(inode), iocb->ki_pos, (unsigned)len,
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
ceph_cap_string(got));
ceph: track read contexts in ceph_file_info Previously ceph_read_iter() uses current->journal to pass context info to ceph_readpages(), so that ceph_readpages() can distinguish read(2) from readahead(2)/fadvise(2)/madvise(2). The problem is that page fault can happen when copying data to userspace memory. Page fault may call other filesystem's page_mkwrite() if the userspace memory is mapped to a file. The later filesystem may also want to use current->journal. The fix is define a on-stack data structure in ceph_read_iter(), add it to context list in ceph_file_info. ceph_readpages() searches the list, find if there is a context belongs to current thread. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-12-15 11:15:36 +08:00
ceph_add_rw_context(fi, &rw_ctx);
ceph: switch to ->read_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-03 08:28:01 +08:00
ret = generic_file_read_iter(iocb, to);
ceph: track read contexts in ceph_file_info Previously ceph_read_iter() uses current->journal to pass context info to ceph_readpages(), so that ceph_readpages() can distinguish read(2) from readahead(2)/fadvise(2)/madvise(2). The problem is that page fault can happen when copying data to userspace memory. Page fault may call other filesystem's page_mkwrite() if the userspace memory is mapped to a file. The later filesystem may also want to use current->journal. The fix is define a on-stack data structure in ceph_read_iter(), add it to context list in ceph_file_info. ceph_readpages() searches the list, find if there is a context belongs to current thread. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-12-15 11:15:36 +08:00
ceph_del_rw_context(fi, &rw_ctx);
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
}
ceph: take the inode lock before acquiring cap refs Most of the time, we (or the vfs layer) takes the inode_lock and then acquires caps, but ceph_read_iter does the opposite, and that can lead to a deadlock. When there are multiple clients treading over the same data, we can end up in a situation where a reader takes caps and then tries to acquire the inode_lock. Another task holds the inode_lock and issues a request to the MDS which needs to revoke the caps, but that can't happen until the inode_lock is unwedged. Fix this by having ceph_read_iter take the inode_lock earlier, before attempting to acquire caps. Fixes: 321fe13c9398 ("ceph: add buffered/direct exclusionary locking for reads and writes") Link: https://tracker.ceph.com/issues/36348 Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-13 22:10:27 +08:00
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
dout("aio_read %p %llx.%llx dropping cap refs on %s = %d\n",
inode, ceph_vinop(inode), ceph_cap_string(got), (int)ret);
ceph: fetch inline data when getting Fcr cap refs we can't use getattr to fetch inline data after getting Fcr caps, because it can cause deadlock. The solution is try bringing inline data to page cache when not holding any cap, and hope the inline data page is still there after getting the Fcr caps. If the page is still there, pin it in page cache for later IO. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:10:07 +08:00
if (pinned_page) {
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
put_page(pinned_page);
ceph: fetch inline data when getting Fcr cap refs we can't use getattr to fetch inline data after getting Fcr caps, because it can cause deadlock. The solution is try bringing inline data to page cache when not holding any cap, and hope the inline data page is still there after getting the Fcr caps. If the page is still there, pin it in page cache for later IO. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:10:07 +08:00
pinned_page = NULL;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph_put_cap_refs(ci, got);
ceph: take the inode lock before acquiring cap refs Most of the time, we (or the vfs layer) takes the inode_lock and then acquires caps, but ceph_read_iter does the opposite, and that can lead to a deadlock. When there are multiple clients treading over the same data, we can end up in a situation where a reader takes caps and then tries to acquire the inode_lock. Another task holds the inode_lock and issues a request to the MDS which needs to revoke the caps, but that can't happen until the inode_lock is unwedged. Fix this by having ceph_read_iter take the inode_lock earlier, before attempting to acquire caps. Fixes: 321fe13c9398 ("ceph: add buffered/direct exclusionary locking for reads and writes") Link: https://tracker.ceph.com/issues/36348 Signed-off-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-11-13 22:10:27 +08:00
if (iocb->ki_flags & IOCB_DIRECT)
ceph_end_io_direct(inode);
else
ceph_end_io_read(inode);
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
if (retry_op > HAVE_RETRIED && ret >= 0) {
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
int statret;
struct page *page = NULL;
loff_t i_size;
if (retry_op == READ_INLINE) {
ceph: switch some GFP_NOFS memory allocation to GFP_KERNEL GFP_NOFS memory allocation is required for page writeback path. But there is no need to use GFP_NOFS in syscall path and readpage path Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-13 17:27:05 +08:00
page = __page_cache_alloc(GFP_KERNEL);
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
if (!page)
return -ENOMEM;
}
ceph: fix sync read eof check deadlock If a sync read gets a short result from the OSD, it may need to do a getattr to see if it is short due to reaching end-of-file. The getattr was being done while holding a reference to FILE_RD, which can lead to a deadlock if the MDS is revoking that capability bit and can't process the getattr until it does. We fix this by setting a flag if EOF size validation is needed, and doing the getattr in ceph_aio_read, after the RD cap ref is dropped. If the read needs to be continued, we loop and continue traversing the file. Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 06:04:02 +08:00
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
statret = __ceph_do_getattr(inode, page,
CEPH_STAT_CAP_INLINE_DATA, !!page);
if (statret < 0) {
ceph: fix error handling in ceph_read_iter In case __ceph_do_getattr returns an error and the retry_op in ceph_read_iter is not READ_INLINE, then it's possible to invoke __free_page on a page which is NULL, this naturally leads to a crash. This can happen when, for example, a process waiting on a MDS reply receives sigterm. Fix this by explicitly checking whether the page is set or not. Cc: stable@vger.kernel.org # 3.19+ Signed-off-by: Nikolay Borisov <kernel@kyup.com> Reviewed-by: Yan, Zheng <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2016-10-10 20:38:18 +08:00
if (page)
__free_page(page);
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
if (statret == -ENODATA) {
BUG_ON(retry_op != READ_INLINE);
goto again;
}
return statret;
}
ceph: fix sync read eof check deadlock If a sync read gets a short result from the OSD, it may need to do a getattr to see if it is short due to reaching end-of-file. The getattr was being done while holding a reference to FILE_RD, which can lead to a deadlock if the MDS is revoking that capability bit and can't process the getattr until it does. We fix this by setting a flag if EOF size validation is needed, and doing the getattr in ceph_aio_read, after the RD cap ref is dropped. If the read needs to be continued, we loop and continue traversing the file. Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 06:04:02 +08:00
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
i_size = i_size_read(inode);
if (retry_op == READ_INLINE) {
ceph: fix reading inline data when i_size > PAGE_SIZE when inode has inline data but its size > PAGE_SIZE (it was truncated to larger size), previous direct read code return -EIO. This patch adds code to return zeros for data whose offset > PAGE_SIZE. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-01-10 11:43:12 +08:00
BUG_ON(ret > 0 || read > 0);
if (iocb->ki_pos < i_size &&
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
iocb->ki_pos < PAGE_SIZE) {
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
loff_t end = min_t(loff_t, i_size,
iocb->ki_pos + len);
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
end = min_t(loff_t, end, PAGE_SIZE);
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
if (statret < end)
zero_user_segment(page, statret, end);
ret = copy_page_to_iter(page,
iocb->ki_pos & ~PAGE_MASK,
end - iocb->ki_pos, to);
iocb->ki_pos += ret;
ceph: fix reading inline data when i_size > PAGE_SIZE when inode has inline data but its size > PAGE_SIZE (it was truncated to larger size), previous direct read code return -EIO. This patch adds code to return zeros for data whose offset > PAGE_SIZE. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-01-10 11:43:12 +08:00
read += ret;
}
if (iocb->ki_pos < i_size && read < len) {
size_t zlen = min_t(size_t, len - read,
i_size - iocb->ki_pos);
ret = iov_iter_zero(zlen, to);
iocb->ki_pos += ret;
read += ret;
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
}
__free_pages(page, 0);
ceph: fix reading inline data when i_size > PAGE_SIZE when inode has inline data but its size > PAGE_SIZE (it was truncated to larger size), previous direct read code return -EIO. This patch adds code to return zeros for data whose offset > PAGE_SIZE. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-01-10 11:43:12 +08:00
return read;
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
}
ceph: fix sync read eof check deadlock If a sync read gets a short result from the OSD, it may need to do a getattr to see if it is short due to reaching end-of-file. The getattr was being done while holding a reference to FILE_RD, which can lead to a deadlock if the MDS is revoking that capability bit and can't process the getattr until it does. We fix this by setting a flag if EOF size validation is needed, and doing the getattr in ceph_aio_read, after the RD cap ref is dropped. If the read needs to be continued, we loop and continue traversing the file. Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 06:04:02 +08:00
/* hit EOF or hole? */
ceph: sync read inline data we can't use getattr to fetch inline data while holding Fr cap, because it can cause deadlock. If we need to sync read inline data, drop cap refs first, then use getattr to fetch inline data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:36:18 +08:00
if (retry_op == CHECK_EOF && iocb->ki_pos < i_size &&
ceph: fix reading inline data when i_size > PAGE_SIZE when inode has inline data but its size > PAGE_SIZE (it was truncated to larger size), previous direct read code return -EIO. This patch adds code to return zeros for data whose offset > PAGE_SIZE. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-01-10 11:43:12 +08:00
ret < len) {
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
dout("sync_read hit hole, ppos %lld < size %lld"
ceph: use i_size_{read,write} to get/set i_size Cap message from MDS can update i_size. In that case, we don't hold i_mutex. So it's unsafe to directly access inode->i_size while holding i_mutex. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-30 11:32:46 +08:00
", reading more\n", iocb->ki_pos, i_size);
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
ceph: fix sync read eof check deadlock If a sync read gets a short result from the OSD, it may need to do a getattr to see if it is short due to reaching end-of-file. The getattr was being done while holding a reference to FILE_RD, which can lead to a deadlock if the MDS is revoking that capability bit and can't process the getattr until it does. We fix this by setting a flag if EOF size validation is needed, and doing the getattr in ceph_aio_read, after the RD cap ref is dropped. If the read needs to be continued, we loop and continue traversing the file. Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 06:04:02 +08:00
read += ret;
len -= ret;
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
retry_op = HAVE_RETRIED;
ceph: fix sync read eof check deadlock If a sync read gets a short result from the OSD, it may need to do a getattr to see if it is short due to reaching end-of-file. The getattr was being done while holding a reference to FILE_RD, which can lead to a deadlock if the MDS is revoking that capability bit and can't process the getattr until it does. We fix this by setting a flag if EOF size validation is needed, and doing the getattr in ceph_aio_read, after the RD cap ref is dropped. If the read needs to be continued, we loop and continue traversing the file. Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 06:04:02 +08:00
goto again;
}
}
ceph: implement readv/preadv for sync operation For readv/preadv sync-operatoin, ceph only do the first iov. Now implement this. Signed-off-by: Jianpeng Ma <majianpeng@gmail.com> Reviewed-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-09-26 14:42:17 +08:00
ceph: fix sync read eof check deadlock If a sync read gets a short result from the OSD, it may need to do a getattr to see if it is short due to reaching end-of-file. The getattr was being done while holding a reference to FILE_RD, which can lead to a deadlock if the MDS is revoking that capability bit and can't process the getattr until it does. We fix this by setting a flag if EOF size validation is needed, and doing the getattr in ceph_aio_read, after the RD cap ref is dropped. If the read needs to be continued, we loop and continue traversing the file. Signed-off-by: Sage Weil <sage@newdream.net>
2010-02-10 06:04:02 +08:00
if (ret >= 0)
ret += read;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
return ret;
}
/*
* Take cap references to avoid releasing caps to MDS mid-write.
*
* If we are synchronous, and write with an old snap context, the OSD
* may return EOLDSNAPC. In that case, retry the write.. _after_
* dropping our cap refs and allowing the pending snap to logically
* complete _before_ this write occurs.
*
* If we are near ENOSPC, write synchronously.
*/
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
static ssize_t ceph_write_iter(struct kiocb *iocb, struct iov_iter *from)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
{
struct file *file = iocb->ki_filp;
ceph: perform lazy writes when file mode and caps permit If we have marked a file as "lazy" (using the ceph ioctl), perform buffered writes when the MDS caps allow it. Signed-off-by: Sage Weil <sage@newdream.net>
2010-05-27 05:31:27 +08:00
struct ceph_file_info *fi = file->private_data;
new helper: file_inode(file) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-01-24 06:07:38 +08:00
struct inode *inode = file_inode(file);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
struct ceph_inode_info *ci = ceph_inode(inode);
ceph: add additional offset check in ceph_write_iter() If the offset is larger or equal to both real file size and max file size, then return -EFBIG. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-19 22:15:26 +08:00
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
ceph: check POOL_FLAG_FULL/NEARFULL in addition to OSDMAP_FULL/NEARFULL CEPH_OSDMAP_FULL/NEARFULL aren't set since mimic, so we need to consult per-pool flags as well. Unfortunately the backwards compatibility here is lacking: - the change that deprecated OSDMAP_FULL/NEARFULL went into mimic, but was guarded by require_osd_release >= RELEASE_LUMINOUS - it was subsequently backported to luminous in v12.2.2, but that makes no difference to clients that only check OSDMAP_FULL/NEARFULL because require_osd_release is not client-facing -- it is for OSDs Since all kernels are affected, the best we can do here is just start checking both map flags and pool flags and send that to stable. These checks are best effort, so take osdc->lock and look up pool flags just once. Remove the FIXME, since filesystem quotas are checked above and RADOS quotas are reflected in POOL_FLAG_FULL: when the pool reaches its quota, both POOL_FLAG_FULL and POOL_FLAG_FULL_QUOTA are set. Cc: stable@vger.kernel.org Reported-by: Yanhu Cao <gmayyyha@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Acked-by: Sage Weil <sage@redhat.com>
2020-03-09 19:03:14 +08:00
struct ceph_osd_client *osdc = &fsc->client->osdc;
ceph: pre-allocate data structure that tracks caps flushing Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-10 17:26:13 +08:00
struct ceph_cap_flush *prealloc_cf;
switch generic_write_checks() to iocb and iter ... returning -E... upon error and amount of data left in iter after (possible) truncation upon success. Note, that normal case gives a non-zero (positive) return value, so any tests for != 0 _must_ be updated. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Conflicts: fs/ext4/file.c
2015-04-10 00:55:47 +08:00
ssize_t count, written = 0;
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
int err, want, got;
ceph: do not execute direct write in parallel if O_APPEND is specified In O_APPEND & O_DIRECT mode, the data from different writers will be possibly overlapping each other since they take the shared lock. For example, both Writer1 and Writer2 are in O_APPEND and O_DIRECT mode: Writer1 Writer2 shared_lock() shared_lock() getattr(CAP_SIZE) getattr(CAP_SIZE) iocb->ki_pos = EOF iocb->ki_pos = EOF write(data1) write(data2) shared_unlock() shared_unlock() The data2 will overlap the data1 from the same file offset, the old EOF. Switch to exclusive lock instead when O_APPEND is specified. Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-04 10:28:25 +08:00
bool direct_lock = false;
ceph: check POOL_FLAG_FULL/NEARFULL in addition to OSDMAP_FULL/NEARFULL CEPH_OSDMAP_FULL/NEARFULL aren't set since mimic, so we need to consult per-pool flags as well. Unfortunately the backwards compatibility here is lacking: - the change that deprecated OSDMAP_FULL/NEARFULL went into mimic, but was guarded by require_osd_release >= RELEASE_LUMINOUS - it was subsequently backported to luminous in v12.2.2, but that makes no difference to clients that only check OSDMAP_FULL/NEARFULL because require_osd_release is not client-facing -- it is for OSDs Since all kernels are affected, the best we can do here is just start checking both map flags and pool flags and send that to stable. These checks are best effort, so take osdc->lock and look up pool flags just once. Remove the FIXME, since filesystem quotas are checked above and RADOS quotas are reflected in POOL_FLAG_FULL: when the pool reaches its quota, both POOL_FLAG_FULL and POOL_FLAG_FULL_QUOTA are set. Cc: stable@vger.kernel.org Reported-by: Yanhu Cao <gmayyyha@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Acked-by: Sage Weil <sage@redhat.com>
2020-03-09 19:03:14 +08:00
u32 map_flags;
u64 pool_flags;
switch generic_write_checks() to iocb and iter ... returning -E... upon error and amount of data left in iter after (possible) truncation upon success. Note, that normal case gives a non-zero (positive) return value, so any tests for != 0 _must_ be updated. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Conflicts: fs/ext4/file.c
2015-04-10 00:55:47 +08:00
loff_t pos;
ceph: add additional offset check in ceph_write_iter() If the offset is larger or equal to both real file size and max file size, then return -EFBIG. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-19 22:15:26 +08:00
loff_t limit = max(i_size_read(inode), fsc->max_file_size);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
if (ceph_snap(inode) != CEPH_NOSNAP)
return -EROFS;
ceph: pre-allocate data structure that tracks caps flushing Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-10 17:26:13 +08:00
prealloc_cf = ceph_alloc_cap_flush();
if (!prealloc_cf)
return -ENOMEM;
ceph: do not execute direct write in parallel if O_APPEND is specified In O_APPEND & O_DIRECT mode, the data from different writers will be possibly overlapping each other since they take the shared lock. For example, both Writer1 and Writer2 are in O_APPEND and O_DIRECT mode: Writer1 Writer2 shared_lock() shared_lock() getattr(CAP_SIZE) getattr(CAP_SIZE) iocb->ki_pos = EOF iocb->ki_pos = EOF write(data1) write(data2) shared_unlock() shared_unlock() The data2 will overlap the data1 from the same file offset, the old EOF. Switch to exclusive lock instead when O_APPEND is specified. Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-04 10:28:25 +08:00
if ((iocb->ki_flags & (IOCB_DIRECT | IOCB_APPEND)) == IOCB_DIRECT)
direct_lock = true;
ceph: fix -EOLDSNAPC handling Need to drop cap reference before retry. Besides, it's better to redo file write checks for each retry because we re-lock inode. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-08-14 10:50:50 +08:00
retry_snap:
ceph: do not execute direct write in parallel if O_APPEND is specified In O_APPEND & O_DIRECT mode, the data from different writers will be possibly overlapping each other since they take the shared lock. For example, both Writer1 and Writer2 are in O_APPEND and O_DIRECT mode: Writer1 Writer2 shared_lock() shared_lock() getattr(CAP_SIZE) getattr(CAP_SIZE) iocb->ki_pos = EOF iocb->ki_pos = EOF write(data1) write(data2) shared_unlock() shared_unlock() The data2 will overlap the data1 from the same file offset, the old EOF. Switch to exclusive lock instead when O_APPEND is specified. Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-04 10:28:25 +08:00
if (direct_lock)
ceph: add buffered/direct exclusionary locking for reads and writes xfstest generic/451 intermittently fails. The test does O_DIRECT writes to a file, and then reads back the result using buffered I/O, while running a separate set of tasks that are also doing buffered reads. The client will invalidate the cache prior to a direct write, but it's easy for one of the other readers' replies to race in and reinstantiate the invalidated range with stale data. To fix this, we must to serialize direct I/O writes and buffered reads. We could just sprinkle in some shared locks on the i_rwsem for reads, and increase the exclusive footprint on the write side, but that would cause O_DIRECT writes to end up serialized vs. other direct requests. Instead, borrow the scheme used by nfs.ko. Buffered writes take the i_rwsem exclusively, but buffered reads take a shared lock, allowing them to run in parallel. O_DIRECT requests also take a shared lock, but we need for them to not run in parallel with buffered reads. A flag on the ceph_inode_info is used to indicate whether it's in direct or buffered I/O mode. When a conflicting request is submitted, it will block until the inode can be flipped to the necessary mode. Link: https://tracker.ceph.com/issues/40985 Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-08-03 01:15:39 +08:00
ceph_start_io_direct(inode);
else
ceph_start_io_write(inode);
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
/* We can write back this queue in page reclaim */
fs: export inode_to_bdi and use it in favor of mapping->backing_dev_info Now that we got rid of the bdi abuse on character devices we can always use sb->s_bdi to get at the backing_dev_info for a file, except for the block device special case. Export inode_to_bdi and replace uses of mapping->backing_dev_info with it to prepare for the removal of mapping->backing_dev_info. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-01-14 17:42:36 +08:00
current->backing_dev_info = inode_to_bdi(inode);
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
ceph: get inode size for each append write Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-09-07 11:35:01 +08:00
if (iocb->ki_flags & IOCB_APPEND) {
err = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE, false);
if (err < 0)
goto out;
}
switch generic_write_checks() to iocb and iter ... returning -E... upon error and amount of data left in iter after (possible) truncation upon success. Note, that normal case gives a non-zero (positive) return value, so any tests for != 0 _must_ be updated. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Conflicts: fs/ext4/file.c
2015-04-10 00:55:47 +08:00
err = generic_write_checks(iocb, from);
if (err <= 0)
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
goto out;
switch generic_write_checks() to iocb and iter ... returning -E... upon error and amount of data left in iter after (possible) truncation upon success. Note, that normal case gives a non-zero (positive) return value, so any tests for != 0 _must_ be updated. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Conflicts: fs/ext4/file.c
2015-04-10 00:55:47 +08:00
pos = iocb->ki_pos;
ceph: add additional offset check in ceph_write_iter() If the offset is larger or equal to both real file size and max file size, then return -EFBIG. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-19 22:15:26 +08:00
if (unlikely(pos >= limit)) {
err = -EFBIG;
goto out;
} else {
iov_iter_truncate(from, limit - pos);
}
switch generic_write_checks() to iocb and iter ... returning -E... upon error and amount of data left in iter after (possible) truncation upon success. Note, that normal case gives a non-zero (positive) return value, so any tests for != 0 _must_ be updated. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Conflicts: fs/ext4/file.c
2015-04-10 00:55:47 +08:00
count = iov_iter_count(from);
ceph: quota: support for ceph.quota.max_bytes Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-01-05 18:47:21 +08:00
if (ceph_quota_is_max_bytes_exceeded(inode, pos + count)) {
err = -EDQUOT;
goto out;
}
fs: Rename file_remove_suid() to file_remove_privs() file_remove_suid() is a misnomer since it removes also file capabilities stored in xattrs and sets S_NOSEC flag. Also should_remove_suid() tells something else than whether file_remove_suid() call is necessary which leads to bugs. Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2015-05-21 22:05:53 +08:00
err = file_remove_privs(file);
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
if (err)
goto out;
err = file_update_time(file);
if (err)
goto out;
ceph: increment change_attribute on local changes We don't set SB_I_VERSION on ceph since we need to manage it ourselves, so we must increment it whenever we update the file times. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-06-06 20:57:27 +08:00
inode_inc_iversion_raw(inode);
ceph: convert inline data to normal data before data write Before any data write, convert inline data to normal data and set i_inline_version to CEPH_INLINE_NONE. The OSD request that saves inline data to object contains 3 operations (CMPXATTR, WRITE and SETXATTR). It compares a xattr named 'inline_version' to prevent old data overwrites newer data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:38:29 +08:00
if (ci->i_inline_version != CEPH_INLINE_NONE) {
err = ceph_uninline_data(file, NULL);
if (err < 0)
goto out;
}
ceph: check POOL_FLAG_FULL/NEARFULL in addition to OSDMAP_FULL/NEARFULL CEPH_OSDMAP_FULL/NEARFULL aren't set since mimic, so we need to consult per-pool flags as well. Unfortunately the backwards compatibility here is lacking: - the change that deprecated OSDMAP_FULL/NEARFULL went into mimic, but was guarded by require_osd_release >= RELEASE_LUMINOUS - it was subsequently backported to luminous in v12.2.2, but that makes no difference to clients that only check OSDMAP_FULL/NEARFULL because require_osd_release is not client-facing -- it is for OSDs Since all kernels are affected, the best we can do here is just start checking both map flags and pool flags and send that to stable. These checks are best effort, so take osdc->lock and look up pool flags just once. Remove the FIXME, since filesystem quotas are checked above and RADOS quotas are reflected in POOL_FLAG_FULL: when the pool reaches its quota, both POOL_FLAG_FULL and POOL_FLAG_FULL_QUOTA are set. Cc: stable@vger.kernel.org Reported-by: Yanhu Cao <gmayyyha@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Acked-by: Sage Weil <sage@redhat.com>
2020-03-09 19:03:14 +08:00
down_read(&osdc->lock);
map_flags = osdc->osdmap->flags;
pool_flags = ceph_pg_pool_flags(osdc->osdmap, ci->i_layout.pool_id);
up_read(&osdc->lock);
if ((map_flags & CEPH_OSDMAP_FULL) ||
(pool_flags & CEPH_POOL_FLAG_FULL)) {
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
err = -ENOSPC;
ceph: don't early drop Fw cap ceph_aio_write() has an optimization that marks CEPH_CAP_FILE_WR cap dirty before data is copied to page cache and inode size is updated. The optimization avoids slow cap revocation caused by balance_dirty_pages(), but introduces inode size update race. If ceph_check_caps() flushes the dirty cap before the inode size is updated, MDS can miss the new inode size. So just remove the optimization. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com>
2013-03-01 10:55:39 +08:00
goto out;
}
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
ceph: fix printk format warnings in file.c Fix printk format warnings by using %zd for 'ssize_t' variables: fs/ceph/file.c:751:2: warning: format '%ld' expects argument of type 'long int', but argument 11 has type 'ssize_t' [-Wformat] fs/ceph/file.c:762:2: warning: format '%ld' expects argument of type 'long int', but argument 11 has type 'ssize_t' [-Wformat] Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Cc: ceph-devel@vger.kernel.org Signed-off-by: Sage Weil <sage@inktank.com>
2013-04-20 05:20:07 +08:00
dout("aio_write %p %llx.%llx %llu~%zd getting caps. i_size %llu\n",
ceph: use i_size_{read,write} to get/set i_size Cap message from MDS can update i_size. In that case, we don't hold i_mutex. So it's unsafe to directly access inode->i_size while holding i_mutex. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-30 11:32:46 +08:00
inode, ceph_vinop(inode), pos, count, i_size_read(inode));
ceph: revert commit 22cddde104 commit 22cddde104 breaks the atomicity of write operation, it also introduces a deadlock between write and truncate. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com> Conflicts: fs/ceph/addr.c
2013-05-02 12:15:58 +08:00
if (fi->fmode & CEPH_FILE_MODE_LAZY)
want = CEPH_CAP_FILE_BUFFER | CEPH_CAP_FILE_LAZYIO;
else
want = CEPH_CAP_FILE_BUFFER;
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
got = 0;
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
err = ceph_get_caps(file, CEPH_CAP_FILE_WR, want, pos + count,
ceph: fetch inline data when getting Fcr cap refs we can't use getattr to fetch inline data after getting Fcr caps, because it can cause deadlock. The solution is try bringing inline data to page cache when not holding any cap, and hope the inline data page is still there after getting the Fcr caps. If the page is still there, pin it in page cache for later IO. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:10:07 +08:00
&got, NULL);
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
if (err < 0)
ceph: take i_mutex before getting Fw cap There is deadlock as illustrated bellow. The fix is taking i_mutex before getting Fw cap reference. write truncate MDS --------------------- -------------------- -------------- get Fw cap lock i_mutex lock i_mutex (blocked) request setattr.size -> <- revoke Fw cap Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com> Reviewed-by: Sage Weil <sage@inktank.com>
2013-04-12 16:11:10 +08:00
goto out;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph: fix printk format warnings in file.c Fix printk format warnings by using %zd for 'ssize_t' variables: fs/ceph/file.c:751:2: warning: format '%ld' expects argument of type 'long int', but argument 11 has type 'ssize_t' [-Wformat] fs/ceph/file.c:762:2: warning: format '%ld' expects argument of type 'long int', but argument 11 has type 'ssize_t' [-Wformat] Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Cc: ceph-devel@vger.kernel.org Signed-off-by: Sage Weil <sage@inktank.com>
2013-04-20 05:20:07 +08:00
dout("aio_write %p %llx.%llx %llu~%zd got cap refs on %s\n",
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
inode, ceph_vinop(inode), pos, count, ceph_cap_string(got));
ceph: revert commit 22cddde104 commit 22cddde104 breaks the atomicity of write operation, it also introduces a deadlock between write and truncate. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com> Conflicts: fs/ceph/addr.c
2013-05-02 12:15:58 +08:00
if ((got & (CEPH_CAP_FILE_BUFFER|CEPH_CAP_FILE_LAZYIO)) == 0 ||
ceph: when seeing write errors on an inode, switch to sync writes Currently, we don't have a real feedback mechanism in place for when we start seeing buffered writeback errors. If writeback is failing, there is nothing that prevents an application from continuing to dirty pages that aren't being cleaned. In the event that we're seeing write errors of any sort occur on an inode, have the callback set a flag to force further writes to be synchronous. When the next write succeeds, clear the flag to allow buffered writeback to continue. Since this is just a hint to the write submission mechanism, we only take the i_ceph_lock when a lockless check shows that the flag needs to be changed. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng” <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-04-04 20:39:46 +08:00
(iocb->ki_flags & IOCB_DIRECT) || (fi->flags & CEPH_F_SYNC) ||
(ci->i_ceph_flags & CEPH_I_ERROR_WRITE)) {
ceph: set i_head_snapc when getting CEPH_CAP_FILE_WR reference In most cases that snap context is needed, we are holding reference of CEPH_CAP_FILE_WR. So we can set ceph inode's i_head_snapc when getting the CEPH_CAP_FILE_WR reference, and make codes get snap context from i_head_snapc. This makes the code simpler. Another benefit of this change is that we can handle snap notification more elegantly. Especially when snap context is updated while someone else is doing write. The old queue cap_snap code may set cap_snap's context to ether the old context or the new snap context, depending on if i_head_snapc is set. The new queue capp_snap code always set cap_snap's context to the old snap context. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-04-30 14:40:54 +08:00
struct ceph_snap_context *snapc;
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
struct iov_iter data;
ceph: set i_head_snapc when getting CEPH_CAP_FILE_WR reference In most cases that snap context is needed, we are holding reference of CEPH_CAP_FILE_WR. So we can set ceph inode's i_head_snapc when getting the CEPH_CAP_FILE_WR reference, and make codes get snap context from i_head_snapc. This makes the code simpler. Another benefit of this change is that we can handle snap notification more elegantly. Especially when snap context is updated while someone else is doing write. The old queue cap_snap code may set cap_snap's context to ether the old context or the new snap context, depending on if i_head_snapc is set. The new queue capp_snap code always set cap_snap's context to the old snap context. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-04-30 14:40:54 +08:00
spin_lock(&ci->i_ceph_lock);
if (__ceph_have_pending_cap_snap(ci)) {
struct ceph_cap_snap *capsnap =
list_last_entry(&ci->i_cap_snaps,
struct ceph_cap_snap,
ci_item);
snapc = ceph_get_snap_context(capsnap->context);
} else {
BUG_ON(!ci->i_head_snapc);
snapc = ceph_get_snap_context(ci->i_head_snapc);
}
spin_unlock(&ci->i_ceph_lock);
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
/* we might need to revert back to that point */
data = *from;
ceph: do not execute direct write in parallel if O_APPEND is specified In O_APPEND & O_DIRECT mode, the data from different writers will be possibly overlapping each other since they take the shared lock. For example, both Writer1 and Writer2 are in O_APPEND and O_DIRECT mode: Writer1 Writer2 shared_lock() shared_lock() getattr(CAP_SIZE) getattr(CAP_SIZE) iocb->ki_pos = EOF iocb->ki_pos = EOF write(data1) write(data2) shared_unlock() shared_unlock() The data2 will overlap the data1 from the same file offset, the old EOF. Switch to exclusive lock instead when O_APPEND is specified. Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-04 10:28:25 +08:00
if (iocb->ki_flags & IOCB_DIRECT)
ceph: Asynchronous IO support The basic idea of AIO support is simple, just call kiocb::ki_complete() in OSD request's complete callback. But there are several special cases. when IO span multiple objects, we need to wait until all OSD requests are complete, then call kiocb::ki_complete(). Error handling in this case is tricky too. For simplify, AIO both span multiple objects and extends i_size are not allowed. Another special case is check EOF for reading (other client can write to the file and extend i_size concurrently). For simplify, the direct-IO/AIO code path does do the check, fallback to normal syn read instead. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-23 21:23:38 +08:00
written = ceph_direct_read_write(iocb, &data, snapc,
&prealloc_cf);
ceph: do not execute direct write in parallel if O_APPEND is specified In O_APPEND & O_DIRECT mode, the data from different writers will be possibly overlapping each other since they take the shared lock. For example, both Writer1 and Writer2 are in O_APPEND and O_DIRECT mode: Writer1 Writer2 shared_lock() shared_lock() getattr(CAP_SIZE) getattr(CAP_SIZE) iocb->ki_pos = EOF iocb->ki_pos = EOF write(data1) write(data2) shared_unlock() shared_unlock() The data2 will overlap the data1 from the same file offset, the old EOF. Switch to exclusive lock instead when O_APPEND is specified. Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-04 10:28:25 +08:00
else
ceph: set i_head_snapc when getting CEPH_CAP_FILE_WR reference In most cases that snap context is needed, we are holding reference of CEPH_CAP_FILE_WR. So we can set ceph inode's i_head_snapc when getting the CEPH_CAP_FILE_WR reference, and make codes get snap context from i_head_snapc. This makes the code simpler. Another benefit of this change is that we can handle snap notification more elegantly. Especially when snap context is updated while someone else is doing write. The old queue cap_snap code may set cap_snap's context to ether the old context or the new snap context, depending on if i_head_snapc is set. The new queue capp_snap code always set cap_snap's context to the old snap context. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-04-30 14:40:54 +08:00
written = ceph_sync_write(iocb, &data, pos, snapc);
ceph: do not execute direct write in parallel if O_APPEND is specified In O_APPEND & O_DIRECT mode, the data from different writers will be possibly overlapping each other since they take the shared lock. For example, both Writer1 and Writer2 are in O_APPEND and O_DIRECT mode: Writer1 Writer2 shared_lock() shared_lock() getattr(CAP_SIZE) getattr(CAP_SIZE) iocb->ki_pos = EOF iocb->ki_pos = EOF write(data1) write(data2) shared_unlock() shared_unlock() The data2 will overlap the data1 from the same file offset, the old EOF. Switch to exclusive lock instead when O_APPEND is specified. Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-04 10:28:25 +08:00
if (direct_lock)
ceph_end_io_direct(inode);
else
ceph: add buffered/direct exclusionary locking for reads and writes xfstest generic/451 intermittently fails. The test does O_DIRECT writes to a file, and then reads back the result using buffered I/O, while running a separate set of tasks that are also doing buffered reads. The client will invalidate the cache prior to a direct write, but it's easy for one of the other readers' replies to race in and reinstantiate the invalidated range with stale data. To fix this, we must to serialize direct I/O writes and buffered reads. We could just sprinkle in some shared locks on the i_rwsem for reads, and increase the exclusive footprint on the write side, but that would cause O_DIRECT writes to end up serialized vs. other direct requests. Instead, borrow the scheme used by nfs.ko. Buffered writes take the i_rwsem exclusively, but buffered reads take a shared lock, allowing them to run in parallel. O_DIRECT requests also take a shared lock, but we need for them to not run in parallel with buffered reads. A flag on the ceph_inode_info is used to indicate whether it's in direct or buffered I/O mode. When a conflicting request is submitted, it will block until the inode can be flipped to the necessary mode. Link: https://tracker.ceph.com/issues/40985 Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-08-03 01:15:39 +08:00
ceph_end_io_write(inode);
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
if (written > 0)
iov_iter_advance(from, written);
ceph: set i_head_snapc when getting CEPH_CAP_FILE_WR reference In most cases that snap context is needed, we are holding reference of CEPH_CAP_FILE_WR. So we can set ceph inode's i_head_snapc when getting the CEPH_CAP_FILE_WR reference, and make codes get snap context from i_head_snapc. This makes the code simpler. Another benefit of this change is that we can handle snap notification more elegantly. Especially when snap context is updated while someone else is doing write. The old queue cap_snap code may set cap_snap's context to ether the old context or the new snap context, depending on if i_head_snapc is set. The new queue capp_snap code always set cap_snap's context to the old snap context. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-04-30 14:40:54 +08:00
ceph_put_snap_context(snapc);
ceph: revert commit 22cddde104 commit 22cddde104 breaks the atomicity of write operation, it also introduces a deadlock between write and truncate. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com> Conflicts: fs/ceph/addr.c
2013-05-02 12:15:58 +08:00
} else {
ceph: introduce i_truncate_mutex I encountered below deadlock when running fsstress wmtruncate work truncate MDS --------------- ------------------ -------------------------- lock i_mutex <- truncate file lock i_mutex (blocked) <- revoking Fcb (filelock to MIX) send request -> handle request (xlock filelock) At the initial time, there are some dirty pages in the page cache. When the kclient receives the truncate message, it reduces inode size and creates some 'out of i_size' dirty pages. wmtruncate work can't truncate these dirty pages because it's blocked by the i_mutex. Later when the kclient receives the cap message that revokes Fcb caps, It can't flush all dirty pages because writepages() only flushes dirty pages within the inode size. When the MDS handles the 'truncate' request from kclient, it waits for the filelock to become stable. But the filelock is stuck in unstable state because it can't finish revoking kclient's Fcb caps. The truncate pagecache locking has already caused lots of trouble for use. I think it's time simplify it by introducing a new mutex. We use the new mutex to prevent concurrent truncate_inode_pages(). There is no need to worry about race between buffered write and truncate_inode_pages(), because our "get caps" mechanism prevents them from concurrent execution. Reviewed-by: Sage Weil <sage@inktank.com> Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
2013-08-13 12:42:15 +08:00
/*
* No need to acquire the i_truncate_mutex. Because
* the MDS revokes Fwb caps before sending truncate
* message to us. We can't get Fwb cap while there
* are pending vmtruncate. So write and vmtruncate
* can not run at the same time
*/
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
written = generic_perform_write(file, from, pos);
ceph_aio_write(): switch to generic_perform_write() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-02-12 11:28:43 +08:00
if (likely(written >= 0))
iocb->ki_pos = pos + written;
ceph: add buffered/direct exclusionary locking for reads and writes xfstest generic/451 intermittently fails. The test does O_DIRECT writes to a file, and then reads back the result using buffered I/O, while running a separate set of tasks that are also doing buffered reads. The client will invalidate the cache prior to a direct write, but it's easy for one of the other readers' replies to race in and reinstantiate the invalidated range with stale data. To fix this, we must to serialize direct I/O writes and buffered reads. We could just sprinkle in some shared locks on the i_rwsem for reads, and increase the exclusive footprint on the write side, but that would cause O_DIRECT writes to end up serialized vs. other direct requests. Instead, borrow the scheme used by nfs.ko. Buffered writes take the i_rwsem exclusively, but buffered reads take a shared lock, allowing them to run in parallel. O_DIRECT requests also take a shared lock, but we need for them to not run in parallel with buffered reads. A flag on the ceph_inode_info is used to indicate whether it's in direct or buffered I/O mode. When a conflicting request is submitted, it will block until the inode can be flipped to the necessary mode. Link: https://tracker.ceph.com/issues/40985 Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-08-03 01:15:39 +08:00
ceph_end_io_write(inode);
ceph: revert commit 22cddde104 commit 22cddde104 breaks the atomicity of write operation, it also introduces a deadlock between write and truncate. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com> Conflicts: fs/ceph/addr.c
2013-05-02 12:15:58 +08:00
}
ceph: avoid carrying Fw cap during write into page cache The generic_file_aio_write call may block on balance_dirty_pages while we flush data to the OSDs. If we hold a reference to the FILE_WR cap during that interval revocation by the MDS (e.g., to do a stat(2)) may be very slow. Reviewed-by: Yehuda Sadeh <yehuda@hq.newdream.net> Signed-off-by: Sage Weil <sage@newdream.net>
2011-07-27 02:27:34 +08:00
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
if (written >= 0) {
ceph: do not call __mark_dirty_inode under i_lock The __mark_dirty_inode helper now takes i_lock as of 250df6ed. Fix the one ceph callers that held i_lock (__ceph_mark_dirty_caps) to return the flags value so that the callers can do it outside of i_lock. Signed-off-by: Sage Weil <sage@newdream.net>
2011-05-05 02:33:47 +08:00
int dirty;
ceph: quota: update MDS when max_bytes is approaching When we're reaching the ceph.quota.max_bytes limit, i.e., when writing more than 1/16th of the space left in a quota realm, update the MDS with the new file size. This mirrors the fuse-client approach with commit 122c50315ed1 ("client: Inform mds file size when approaching quota limit"), in the ceph git tree. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-01-05 18:47:22 +08:00
ceph: use i_ceph_lock instead of i_lock We have been using i_lock to protect all kinds of data structures in the ceph_inode_info struct, including lists of inodes that we need to iterate over while avoiding races with inode destruction. That requires grabbing a reference to the inode with the list lock protected, but igrab() now takes i_lock to check the inode flags. Changing the list lock ordering would be a painful process. However, using a ceph-specific i_ceph_lock in the ceph inode instead of i_lock is a simple mechanical change and avoids the ordering constraints imposed by igrab(). Reported-by: Amon Ott <a.ott@m-privacy.de> Signed-off-by: Sage Weil <sage@newdream.net>
2011-12-01 01:47:09 +08:00
spin_lock(&ci->i_ceph_lock);
ceph: convert inline data to normal data before data write Before any data write, convert inline data to normal data and set i_inline_version to CEPH_INLINE_NONE. The OSD request that saves inline data to object contains 3 operations (CMPXATTR, WRITE and SETXATTR). It compares a xattr named 'inline_version' to prevent old data overwrites newer data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:38:29 +08:00
ci->i_inline_version = CEPH_INLINE_NONE;
ceph: pre-allocate data structure that tracks caps flushing Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-10 17:26:13 +08:00
dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR,
&prealloc_cf);
ceph: use i_ceph_lock instead of i_lock We have been using i_lock to protect all kinds of data structures in the ceph_inode_info struct, including lists of inodes that we need to iterate over while avoiding races with inode destruction. That requires grabbing a reference to the inode with the list lock protected, but igrab() now takes i_lock to check the inode flags. Changing the list lock ordering would be a painful process. However, using a ceph-specific i_ceph_lock in the ceph inode instead of i_lock is a simple mechanical change and avoids the ordering constraints imposed by igrab(). Reported-by: Amon Ott <a.ott@m-privacy.de> Signed-off-by: Sage Weil <sage@newdream.net>
2011-12-01 01:47:09 +08:00
spin_unlock(&ci->i_ceph_lock);
ceph: do not call __mark_dirty_inode under i_lock The __mark_dirty_inode helper now takes i_lock as of 250df6ed. Fix the one ceph callers that held i_lock (__ceph_mark_dirty_caps) to return the flags value so that the callers can do it outside of i_lock. Signed-off-by: Sage Weil <sage@newdream.net>
2011-05-05 02:33:47 +08:00
if (dirty)
__mark_inode_dirty(inode, dirty);
ceph: quota: update MDS when max_bytes is approaching When we're reaching the ceph.quota.max_bytes limit, i.e., when writing more than 1/16th of the space left in a quota realm, update the MDS with the new file size. This mirrors the fuse-client approach with commit 122c50315ed1 ("client: Inform mds file size when approaching quota limit"), in the ceph git tree. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-01-05 18:47:22 +08:00
if (ceph_quota_is_max_bytes_approaching(inode, iocb->ki_pos))
ceph: remove delay check logic from ceph_check_caps() __ceph_caps_file_wanted() already checks 'caps_wanted_delay_min' and 'caps_wanted_delay_max'. There is no need to duplicate the logic in ceph_check_caps() and __send_cap() Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:01 +08:00
ceph_check_caps(ci, 0, NULL);
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
}
ceph: revert commit 22cddde104 commit 22cddde104 breaks the atomicity of write operation, it also introduces a deadlock between write and truncate. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com> Conflicts: fs/ceph/addr.c
2013-05-02 12:15:58 +08:00
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
dout("aio_write %p %llx.%llx %llu~%u dropping cap refs on %s\n",
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
inode, ceph_vinop(inode), pos, (unsigned)count,
ceph: revert commit 22cddde104 commit 22cddde104 breaks the atomicity of write operation, it also introduces a deadlock between write and truncate. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com> Conflicts: fs/ceph/addr.c
2013-05-02 12:15:58 +08:00
ceph_cap_string(got));
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
ceph_put_cap_refs(ci, got);
ceph: revert commit 22cddde104 commit 22cddde104 breaks the atomicity of write operation, it also introduces a deadlock between write and truncate. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com> Conflicts: fs/ceph/addr.c
2013-05-02 12:15:58 +08:00
ceph: fix -EOLDSNAPC handling Need to drop cap reference before retry. Besides, it's better to redo file write checks for each retry because we re-lock inode. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-08-14 10:50:50 +08:00
if (written == -EOLDSNAPC) {
dout("aio_write %p %llx.%llx %llu~%u" "got EOLDSNAPC, retrying\n",
inode, ceph_vinop(inode), pos, (unsigned)count);
goto retry_snap;
}
ceph: use generic_write_sync Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-04-07 23:52:02 +08:00
if (written >= 0) {
ceph: check POOL_FLAG_FULL/NEARFULL in addition to OSDMAP_FULL/NEARFULL CEPH_OSDMAP_FULL/NEARFULL aren't set since mimic, so we need to consult per-pool flags as well. Unfortunately the backwards compatibility here is lacking: - the change that deprecated OSDMAP_FULL/NEARFULL went into mimic, but was guarded by require_osd_release >= RELEASE_LUMINOUS - it was subsequently backported to luminous in v12.2.2, but that makes no difference to clients that only check OSDMAP_FULL/NEARFULL because require_osd_release is not client-facing -- it is for OSDs Since all kernels are affected, the best we can do here is just start checking both map flags and pool flags and send that to stable. These checks are best effort, so take osdc->lock and look up pool flags just once. Remove the FIXME, since filesystem quotas are checked above and RADOS quotas are reflected in POOL_FLAG_FULL: when the pool reaches its quota, both POOL_FLAG_FULL and POOL_FLAG_FULL_QUOTA are set. Cc: stable@vger.kernel.org Reported-by: Yanhu Cao <gmayyyha@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Acked-by: Sage Weil <sage@redhat.com>
2020-03-09 19:03:14 +08:00
if ((map_flags & CEPH_OSDMAP_NEARFULL) ||
(pool_flags & CEPH_POOL_FLAG_NEARFULL))
ceph: use generic_write_sync Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-04-07 23:52:02 +08:00
iocb->ki_flags |= IOCB_DSYNC;
written = generic_write_sync(iocb, written);
ceph: don't early drop Fw cap ceph_aio_write() has an optimization that marks CEPH_CAP_FILE_WR cap dirty before data is copied to page cache and inode size is updated. The optimization avoids slow cap revocation caused by balance_dirty_pages(), but introduces inode size update race. If ceph_check_caps() flushes the dirty cap before the inode size is updated, MDS can miss the new inode size. So just remove the optimization. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Greg Farnum <greg@inktank.com>
2013-03-01 10:55:39 +08:00
}
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
ceph: replace hold_mutex flag with goto All of the early exit paths need to drop the mutex; it is only the normal path through the function that does not. Skip the unlock in that case with a goto out_unlocked. Signed-off-by: Sage Weil <sage@inktank.com> Reviewed-by: Jianpeng Ma <majianpeng@gmail.com>
2013-08-10 00:57:58 +08:00
goto out_unlocked;
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
out:
ceph: do not execute direct write in parallel if O_APPEND is specified In O_APPEND & O_DIRECT mode, the data from different writers will be possibly overlapping each other since they take the shared lock. For example, both Writer1 and Writer2 are in O_APPEND and O_DIRECT mode: Writer1 Writer2 shared_lock() shared_lock() getattr(CAP_SIZE) getattr(CAP_SIZE) iocb->ki_pos = EOF iocb->ki_pos = EOF write(data1) write(data2) shared_unlock() shared_unlock() The data2 will overlap the data1 from the same file offset, the old EOF. Switch to exclusive lock instead when O_APPEND is specified. Signed-off-by: Xiubo Li <xiubli@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-04 10:28:25 +08:00
if (direct_lock)
ceph: add buffered/direct exclusionary locking for reads and writes xfstest generic/451 intermittently fails. The test does O_DIRECT writes to a file, and then reads back the result using buffered I/O, while running a separate set of tasks that are also doing buffered reads. The client will invalidate the cache prior to a direct write, but it's easy for one of the other readers' replies to race in and reinstantiate the invalidated range with stale data. To fix this, we must to serialize direct I/O writes and buffered reads. We could just sprinkle in some shared locks on the i_rwsem for reads, and increase the exclusive footprint on the write side, but that would cause O_DIRECT writes to end up serialized vs. other direct requests. Instead, borrow the scheme used by nfs.ko. Buffered writes take the i_rwsem exclusively, but buffered reads take a shared lock, allowing them to run in parallel. O_DIRECT requests also take a shared lock, but we need for them to not run in parallel with buffered reads. A flag on the ceph_inode_info is used to indicate whether it's in direct or buffered I/O mode. When a conflicting request is submitted, it will block until the inode can be flipped to the necessary mode. Link: https://tracker.ceph.com/issues/40985 Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-08-03 01:15:39 +08:00
ceph_end_io_direct(inode);
else
ceph_end_io_write(inode);
ceph: replace hold_mutex flag with goto All of the early exit paths need to drop the mutex; it is only the normal path through the function that does not. Skip the unlock in that case with a goto out_unlocked. Signed-off-by: Sage Weil <sage@inktank.com> Reviewed-by: Jianpeng Ma <majianpeng@gmail.com>
2013-08-10 00:57:58 +08:00
out_unlocked:
ceph: pre-allocate data structure that tracks caps flushing Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-10 17:26:13 +08:00
ceph_free_cap_flush(prealloc_cf);
ceph: apply write checks in ceph_aio_write copy write checks in __generic_file_aio_write to ceph_aio_write. To make these checks cover sync write path. Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com> Reviewed-by: Alex Elder <elder@inktank.com>
2013-04-12 16:11:13 +08:00
current->backing_dev_info = NULL;
return written ? written : err;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
}
/*
* llseek. be sure to verify file size on SEEK_END.
*/
lseek: the "whence" argument is called "whence" But the kernel decided to call it "origin" instead. Fix most of the sites. Acked-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 07:59:39 +08:00
static loff_t ceph_llseek(struct file *file, loff_t offset, int whence)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
{
struct inode *inode = file->f_mapping->host;
ceph: compare fsc->max_file_size and inode->i_size for max file size limit In ceph_llseek(), we compare fsc->max_file_size and inode->i_size to choose max file size limit. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-19 22:15:27 +08:00
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
ceph: use i_size_{read,write} to get/set i_size Cap message from MDS can update i_size. In that case, we don't hold i_mutex. So it's unsafe to directly access inode->i_size while holding i_mutex. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-30 11:32:46 +08:00
loff_t i_size;
ceph: Correctly return NXIO errors from ceph_llseek ceph_llseek does not correctly return NXIO errors because the 'out' path always returns 'offset'. Fixes: 06222e491e66 ("fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek") Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-07-22 01:43:09 +08:00
loff_t ret;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
wrappers for ->i_mutex access parallel to mutex_{lock,unlock,trylock,is_locked,lock_nested}, inode_foo(inode) being mutex_foo(&inode->i_mutex). Please, use those for access to ->i_mutex; over the coming cycle ->i_mutex will become rwsem, with ->lookup() done with it held only shared. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-01-23 04:40:57 +08:00
inode_lock(inode);
ceph: fix SEEK_CUR, SEEK_SET regression Commit 06222e491e663dac939f04b125c9dc52126a75c4 got the if wrong so that it always evaluates as true. This is semantically harmless, but makes SEEK_CUR and SEEK_SET needlessly query the server. Rewrite the if to explicitly enumerate the cases we DO need a valid i_size to make this code less fragile. Reported-by: Roel Kluin <roel.kluin@gmail.com> Signed-off-by: Sage Weil <sage@newdream.net>
2011-12-14 01:19:26 +08:00
lseek: the "whence" argument is called "whence" But the kernel decided to call it "origin" instead. Fix most of the sites. Acked-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 07:59:39 +08:00
if (whence == SEEK_END || whence == SEEK_DATA || whence == SEEK_HOLE) {
ceph: request xattrs if xattr_version is zero Following sequence of events can happen. - Client releases an inode, queues cap release message. - A 'lookup' reply brings the same inode back, but the reply doesn't contain xattrs because MDS didn't receive the cap release message and thought client already has up-to-data xattrs. The fix is force sending a getattr request to MDS if xattrs_version is 0. The getattr mask is set to CEPH_STAT_CAP_XATTR, so MDS knows client does not have xattr. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-09-16 21:46:17 +08:00
ret = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE, false);
ceph: Correctly return NXIO errors from ceph_llseek ceph_llseek does not correctly return NXIO errors because the 'out' path always returns 'offset'. Fixes: 06222e491e66 ("fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek") Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-07-22 01:43:09 +08:00
if (ret < 0)
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
goto out;
fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek This converts everybody to handle SEEK_HOLE/SEEK_DATA properly. In some cases we just return -EINVAL, in others we do the normal generic thing, and in others we're simply making sure that the properly due-dilligence is done. For example in NFS/CIFS we need to make sure the file size is update properly for the SEEK_HOLE and SEEK_DATA case, but since it calls the generic llseek stuff itself that is all we have to do. Thanks, Signed-off-by: Josef Bacik <josef@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2011-07-19 01:21:38 +08:00
}
ceph: use i_size_{read,write} to get/set i_size Cap message from MDS can update i_size. In that case, we don't hold i_mutex. So it's unsafe to directly access inode->i_size while holding i_mutex. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-30 11:32:46 +08:00
i_size = i_size_read(inode);
lseek: the "whence" argument is called "whence" But the kernel decided to call it "origin" instead. Fix most of the sites. Acked-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 07:59:39 +08:00
switch (whence) {
fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek This converts everybody to handle SEEK_HOLE/SEEK_DATA properly. In some cases we just return -EINVAL, in others we do the normal generic thing, and in others we're simply making sure that the properly due-dilligence is done. For example in NFS/CIFS we need to make sure the file size is update properly for the SEEK_HOLE and SEEK_DATA case, but since it calls the generic llseek stuff itself that is all we have to do. Thanks, Signed-off-by: Josef Bacik <josef@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2011-07-19 01:21:38 +08:00
case SEEK_END:
ceph: use i_size_{read,write} to get/set i_size Cap message from MDS can update i_size. In that case, we don't hold i_mutex. So it's unsafe to directly access inode->i_size while holding i_mutex. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-30 11:32:46 +08:00
offset += i_size;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
break;
case SEEK_CUR:
/*
* Here we special-case the lseek(fd, 0, SEEK_CUR)
* position-querying operation. Avoid rewriting the "same"
* f_pos value back to the file because a concurrent read(),
* write() or lseek() might have altered it
*/
if (offset == 0) {
ceph: Correctly return NXIO errors from ceph_llseek ceph_llseek does not correctly return NXIO errors because the 'out' path always returns 'offset'. Fixes: 06222e491e66 ("fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek") Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-07-22 01:43:09 +08:00
ret = file->f_pos;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
goto out;
}
offset += file->f_pos;
break;
fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek This converts everybody to handle SEEK_HOLE/SEEK_DATA properly. In some cases we just return -EINVAL, in others we do the normal generic thing, and in others we're simply making sure that the properly due-dilligence is done. For example in NFS/CIFS we need to make sure the file size is update properly for the SEEK_HOLE and SEEK_DATA case, but since it calls the generic llseek stuff itself that is all we have to do. Thanks, Signed-off-by: Josef Bacik <josef@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2011-07-19 01:21:38 +08:00
case SEEK_DATA:
ceph: check negative offsets in ceph_llseek() When a user requests SEEK_HOLE or SEEK_DATA with a negative offset ceph_llseek should return -ENXIO. Currently -EINVAL is being returned for SEEK_DATA and 0 for SEEK_HOLE. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-07-28 18:56:40 +08:00
if (offset < 0 || offset >= i_size) {
fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek This converts everybody to handle SEEK_HOLE/SEEK_DATA properly. In some cases we just return -EINVAL, in others we do the normal generic thing, and in others we're simply making sure that the properly due-dilligence is done. For example in NFS/CIFS we need to make sure the file size is update properly for the SEEK_HOLE and SEEK_DATA case, but since it calls the generic llseek stuff itself that is all we have to do. Thanks, Signed-off-by: Josef Bacik <josef@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2011-07-19 01:21:38 +08:00
ret = -ENXIO;
goto out;
}
break;
case SEEK_HOLE:
ceph: check negative offsets in ceph_llseek() When a user requests SEEK_HOLE or SEEK_DATA with a negative offset ceph_llseek should return -ENXIO. Currently -EINVAL is being returned for SEEK_DATA and 0 for SEEK_HOLE. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-07-28 18:56:40 +08:00
if (offset < 0 || offset >= i_size) {
fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek This converts everybody to handle SEEK_HOLE/SEEK_DATA properly. In some cases we just return -EINVAL, in others we do the normal generic thing, and in others we're simply making sure that the properly due-dilligence is done. For example in NFS/CIFS we need to make sure the file size is update properly for the SEEK_HOLE and SEEK_DATA case, but since it calls the generic llseek stuff itself that is all we have to do. Thanks, Signed-off-by: Josef Bacik <josef@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2011-07-19 01:21:38 +08:00
ret = -ENXIO;
goto out;
}
ceph: use i_size_{read,write} to get/set i_size Cap message from MDS can update i_size. In that case, we don't hold i_mutex. So it's unsafe to directly access inode->i_size while holding i_mutex. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-12-30 11:32:46 +08:00
offset = i_size;
fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek This converts everybody to handle SEEK_HOLE/SEEK_DATA properly. In some cases we just return -EINVAL, in others we do the normal generic thing, and in others we're simply making sure that the properly due-dilligence is done. For example in NFS/CIFS we need to make sure the file size is update properly for the SEEK_HOLE and SEEK_DATA case, but since it calls the generic llseek stuff itself that is all we have to do. Thanks, Signed-off-by: Josef Bacik <josef@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2011-07-19 01:21:38 +08:00
break;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
}
ceph: compare fsc->max_file_size and inode->i_size for max file size limit In ceph_llseek(), we compare fsc->max_file_size and inode->i_size to choose max file size limit. Signed-off-by: Chengguang Xu <cgxu519@gmx.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-19 22:15:27 +08:00
ret = vfs_setpos(file, offset, max(i_size, fsc->max_file_size));
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
out:
wrappers for ->i_mutex access parallel to mutex_{lock,unlock,trylock,is_locked,lock_nested}, inode_foo(inode) being mutex_foo(&inode->i_mutex). Please, use those for access to ->i_mutex; over the coming cycle ->i_mutex will become rwsem, with ->lookup() done with it held only shared. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-01-23 04:40:57 +08:00
inode_unlock(inode);
ceph: Correctly return NXIO errors from ceph_llseek ceph_llseek does not correctly return NXIO errors because the 'out' path always returns 'offset'. Fixes: 06222e491e66 ("fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek") Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com> Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-07-22 01:43:09 +08:00
return ret;
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
}
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
static inline void ceph_zero_partial_page(
struct inode *inode, loff_t offset, unsigned size)
{
struct page *page;
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
pgoff_t index = offset >> PAGE_SHIFT;
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
page = find_lock_page(inode->i_mapping, index);
if (page) {
wait_on_page_writeback(page);
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
zero_user(page, offset & (PAGE_SIZE - 1), size);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
unlock_page(page);
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
put_page(page);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
}
}
static void ceph_zero_pagecache_range(struct inode *inode, loff_t offset,
loff_t length)
{
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
loff_t nearly = round_up(offset, PAGE_SIZE);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
if (offset < nearly) {
loff_t size = nearly - offset;
if (length < size)
size = length;
ceph_zero_partial_page(inode, offset, size);
offset += size;
length -= size;
}
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
if (length >= PAGE_SIZE) {
loff_t size = round_down(length, PAGE_SIZE);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
truncate_pagecache_range(inode, offset, offset + size - 1);
offset += size;
length -= size;
}
if (length)
ceph_zero_partial_page(inode, offset, length);
}
static int ceph_zero_partial_object(struct inode *inode,
loff_t offset, loff_t *length)
{
struct ceph_inode_info *ci = ceph_inode(inode);
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
struct ceph_osd_request *req;
int ret = 0;
loff_t zero = 0;
int op;
if (!length) {
op = offset ? CEPH_OSD_OP_DELETE : CEPH_OSD_OP_TRUNCATE;
length = &zero;
} else {
op = CEPH_OSD_OP_ZERO;
}
req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
ceph_vino(inode),
offset, length,
libceph: specify position of extent operation allow specifying position of extent operation in multi-operations osd request. This is required for cephfs to convert inline data to normal data (compare xattr, then write object). Signed-off-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Ilya Dryomov <idryomov@redhat.com>
2014-11-13 14:40:37 +08:00
0, 1, op,
libceph, rbd, ceph: WRITE | ONDISK -> WRITE CEPH_OSD_FLAG_ONDISK is set in account_request(). Signed-off-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Sage Weil <sage@redhat.com>
2017-02-12 01:48:41 +08:00
CEPH_OSD_FLAG_WRITE,
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
NULL, 0, 0, false);
if (IS_ERR(req)) {
ret = PTR_ERR(req);
goto out;
}
libceph: use timespec64 for r_mtime The request mtime field is used all over ceph, and is currently represented as a 'timespec' structure in Linux. This changes it to timespec64 to allow times beyond 2038, modifying all users at the same time. [ Remove now redundant ts variable in writepage_nounlock(). ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-07-14 04:18:37 +08:00
req->r_mtime = inode->i_mtime;
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
ret = ceph_osdc_start_request(&fsc->client->osdc, req, false);
if (!ret) {
ret = ceph_osdc_wait_request(&fsc->client->osdc, req);
if (ret == -ENOENT)
ret = 0;
}
ceph_osdc_put_request(req);
out:
return ret;
}
static int ceph_zero_objects(struct inode *inode, loff_t offset, loff_t length)
{
int ret = 0;
struct ceph_inode_info *ci = ceph_inode(inode);
libceph: define new ceph_file_layout structure Define new ceph_file_layout structure and rename old ceph_file_layout to ceph_file_layout_legacy. This is preparation for adding namespace to ceph_file_layout structure. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-02-03 21:24:49 +08:00
s32 stripe_unit = ci->i_layout.stripe_unit;
s32 stripe_count = ci->i_layout.stripe_count;
s32 object_size = ci->i_layout.object_size;
ceph: fix fallocate division We need to use do_div to divide by a 64-bit value. Signed-off-by: Sage Weil <sage@inktank.com> Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
2013-08-28 03:15:16 +08:00
u64 object_set_size = object_size * stripe_count;
u64 nearly, t;
/* round offset up to next period boundary */
nearly = offset + object_set_size - 1;
t = nearly;
nearly -= do_div(t, object_set_size);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
while (length && offset < nearly) {
loff_t size = length;
ret = ceph_zero_partial_object(inode, offset, &size);
if (ret < 0)
return ret;
offset += size;
length -= size;
}
while (length >= object_set_size) {
int i;
loff_t pos = offset;
for (i = 0; i < stripe_count; ++i) {
ret = ceph_zero_partial_object(inode, pos, NULL);
if (ret < 0)
return ret;
pos += stripe_unit;
}
offset += object_set_size;
length -= object_set_size;
}
while (length) {
loff_t size = length;
ret = ceph_zero_partial_object(inode, offset, &size);
if (ret < 0)
return ret;
offset += size;
length -= size;
}
return ret;
}
static long ceph_fallocate(struct file *file, int mode,
loff_t offset, loff_t length)
{
struct ceph_file_info *fi = file->private_data;
fs: ceph: new helper: file_inode(file) Signed-off-by: Libo Chen <clbchenlibo.chen@huawei.com> Signed-off-by: Sage Weil <sage@inktank.com>
2013-12-11 13:49:11 +08:00
struct inode *inode = file_inode(file);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
struct ceph_inode_info *ci = ceph_inode(inode);
ceph: pre-allocate data structure that tracks caps flushing Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-10 17:26:13 +08:00
struct ceph_cap_flush *prealloc_cf;
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
int want, got = 0;
int dirty;
int ret = 0;
loff_t endoff = 0;
loff_t size;
ceph: only allow punch hole mode in fallocate Current implementation of cephfs fallocate isn't correct as it doesn't really reserve the space in the cluster, which means that a subsequent call to a write may actually fail due to lack of space. In fact, it is currently possible to fallocate an amount space that is larger than the free space in the cluster. It has behaved this way since the initial commit ad7a60de882a ("ceph: punch hole support"). Since there's no easy solution to fix this at the moment, this patch simply removes support for all fallocate operations but FALLOC_FL_PUNCH_HOLE (which implies FALLOC_FL_KEEP_SIZE). Link: https://tracker.ceph.com/issues/36317 Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-10 01:54:28 +08:00
if (mode != (FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
ceph: check unsupported fallocate mode Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
2014-06-26 15:25:17 +08:00
return -EOPNOTSUPP;
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
if (!S_ISREG(inode->i_mode))
return -EOPNOTSUPP;
ceph: pre-allocate data structure that tracks caps flushing Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-10 17:26:13 +08:00
prealloc_cf = ceph_alloc_cap_flush();
if (!prealloc_cf)
return -ENOMEM;
wrappers for ->i_mutex access parallel to mutex_{lock,unlock,trylock,is_locked,lock_nested}, inode_foo(inode) being mutex_foo(&inode->i_mutex). Please, use those for access to ->i_mutex; over the coming cycle ->i_mutex will become rwsem, with ->lookup() done with it held only shared. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-01-23 04:40:57 +08:00
inode_lock(inode);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
if (ceph_snap(inode) != CEPH_NOSNAP) {
ret = -EROFS;
goto unlock;
}
ceph: convert inline data to normal data before data write Before any data write, convert inline data to normal data and set i_inline_version to CEPH_INLINE_NONE. The OSD request that saves inline data to object contains 3 operations (CMPXATTR, WRITE and SETXATTR). It compares a xattr named 'inline_version' to prevent old data overwrites newer data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:38:29 +08:00
if (ci->i_inline_version != CEPH_INLINE_NONE) {
ret = ceph_uninline_data(file, NULL);
if (ret < 0)
goto unlock;
}
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
size = i_size_read(inode);
ceph: only allow punch hole mode in fallocate Current implementation of cephfs fallocate isn't correct as it doesn't really reserve the space in the cluster, which means that a subsequent call to a write may actually fail due to lack of space. In fact, it is currently possible to fallocate an amount space that is larger than the free space in the cluster. It has behaved this way since the initial commit ad7a60de882a ("ceph: punch hole support"). Since there's no easy solution to fix this at the moment, this patch simply removes support for all fallocate operations but FALLOC_FL_PUNCH_HOLE (which implies FALLOC_FL_KEEP_SIZE). Link: https://tracker.ceph.com/issues/36317 Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-10 01:54:28 +08:00
/* Are we punching a hole beyond EOF? */
if (offset >= size)
goto unlock;
if ((offset + length) > size)
length = size - offset;
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
if (fi->fmode & CEPH_FILE_MODE_LAZY)
want = CEPH_CAP_FILE_BUFFER | CEPH_CAP_FILE_LAZYIO;
else
want = CEPH_CAP_FILE_BUFFER;
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
ret = ceph_get_caps(file, CEPH_CAP_FILE_WR, want, endoff, &got, NULL);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
if (ret < 0)
goto unlock;
ceph: only allow punch hole mode in fallocate Current implementation of cephfs fallocate isn't correct as it doesn't really reserve the space in the cluster, which means that a subsequent call to a write may actually fail due to lack of space. In fact, it is currently possible to fallocate an amount space that is larger than the free space in the cluster. It has behaved this way since the initial commit ad7a60de882a ("ceph: punch hole support"). Since there's no easy solution to fix this at the moment, this patch simply removes support for all fallocate operations but FALLOC_FL_PUNCH_HOLE (which implies FALLOC_FL_KEEP_SIZE). Link: https://tracker.ceph.com/issues/36317 Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-10 01:54:28 +08:00
ceph_zero_pagecache_range(inode, offset, length);
ret = ceph_zero_objects(inode, offset, length);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
if (!ret) {
spin_lock(&ci->i_ceph_lock);
ceph: convert inline data to normal data before data write Before any data write, convert inline data to normal data and set i_inline_version to CEPH_INLINE_NONE. The OSD request that saves inline data to object contains 3 operations (CMPXATTR, WRITE and SETXATTR). It compares a xattr named 'inline_version' to prevent old data overwrites newer data. Signed-off-by: Yan, Zheng <zyan@redhat.com>
2014-11-14 22:38:29 +08:00
ci->i_inline_version = CEPH_INLINE_NONE;
ceph: pre-allocate data structure that tracks caps flushing Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-10 17:26:13 +08:00
dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR,
&prealloc_cf);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
spin_unlock(&ci->i_ceph_lock);
if (dirty)
__mark_inode_dirty(inode, dirty);
}
ceph_put_cap_refs(ci, got);
unlock:
wrappers for ->i_mutex access parallel to mutex_{lock,unlock,trylock,is_locked,lock_nested}, inode_foo(inode) being mutex_foo(&inode->i_mutex). Please, use those for access to ->i_mutex; over the coming cycle ->i_mutex will become rwsem, with ->lookup() done with it held only shared. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2016-01-23 04:40:57 +08:00
inode_unlock(inode);
ceph: pre-allocate data structure that tracks caps flushing Signed-off-by: Yan, Zheng <zyan@redhat.com>
2015-06-10 17:26:13 +08:00
ceph_free_cap_flush(prealloc_cf);
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
return ret;
}
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
/*
* This function tries to get FILE_WR capabilities for dst_ci and FILE_RD for
* src_ci. Two attempts are made to obtain both caps, and an error is return if
* this fails; zero is returned on success.
*/
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
static int get_rd_wr_caps(struct file *src_filp, int *src_got,
struct file *dst_filp,
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
loff_t dst_endoff, int *dst_got)
{
int ret = 0;
bool retrying = false;
retry_caps:
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
ret = ceph_get_caps(dst_filp, CEPH_CAP_FILE_WR, CEPH_CAP_FILE_BUFFER,
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
dst_endoff, dst_got, NULL);
if (ret < 0)
return ret;
/*
* Since we're already holding the FILE_WR capability for the dst file,
* we would risk a deadlock by using ceph_get_caps. Thus, we'll do some
* retry dance instead to try to get both capabilities.
*/
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
ret = ceph_try_get_caps(file_inode(src_filp),
CEPH_CAP_FILE_RD, CEPH_CAP_FILE_SHARED,
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
false, src_got);
if (ret <= 0) {
/* Start by dropping dst_ci caps and getting src_ci caps */
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
ceph_put_cap_refs(ceph_inode(file_inode(dst_filp)), *dst_got);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
if (retrying) {
if (!ret)
/* ceph_try_get_caps masks EAGAIN */
ret = -EAGAIN;
return ret;
}
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
ret = ceph_get_caps(src_filp, CEPH_CAP_FILE_RD,
CEPH_CAP_FILE_SHARED, -1, src_got, NULL);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
if (ret < 0)
return ret;
/*... drop src_ci caps too, and retry */
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
ceph_put_cap_refs(ceph_inode(file_inode(src_filp)), *src_got);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
retrying = true;
goto retry_caps;
}
return ret;
}
static void put_rd_wr_caps(struct ceph_inode_info *src_ci, int src_got,
struct ceph_inode_info *dst_ci, int dst_got)
{
ceph_put_cap_refs(src_ci, src_got);
ceph_put_cap_refs(dst_ci, dst_got);
}
/*
* This function does several size-related checks, returning an error if:
* - source file is smaller than off+len
* - destination file size is not OK (inode_newsize_ok())
* - max bytes quotas is exceeded
*/
static int is_file_size_ok(struct inode *src_inode, struct inode *dst_inode,
loff_t src_off, loff_t dst_off, size_t len)
{
loff_t size, endoff;
size = i_size_read(src_inode);
/*
* Don't copy beyond source file EOF. Instead of simply setting length
* to (size - src_off), just drop to VFS default implementation, as the
* local i_size may be stale due to other clients writing to the source
* inode.
*/
if (src_off + len > size) {
dout("Copy beyond EOF (%llu + %zu > %llu)\n",
src_off, len, size);
return -EOPNOTSUPP;
}
size = i_size_read(dst_inode);
endoff = dst_off + len;
if (inode_newsize_ok(dst_inode, endoff))
return -EOPNOTSUPP;
if (ceph_quota_is_max_bytes_exceeded(dst_inode, endoff))
return -EDQUOT;
return 0;
}
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
static ssize_t ceph_do_objects_copy(struct ceph_inode_info *src_ci, u64 *src_off,
struct ceph_inode_info *dst_ci, u64 *dst_off,
struct ceph_fs_client *fsc,
size_t len, unsigned int flags)
{
struct ceph_object_locator src_oloc, dst_oloc;
struct ceph_object_id src_oid, dst_oid;
size_t bytes = 0;
u64 src_objnum, src_objoff, dst_objnum, dst_objoff;
u32 src_objlen, dst_objlen;
u32 object_size = src_ci->i_layout.object_size;
int ret;
src_oloc.pool = src_ci->i_layout.pool_id;
src_oloc.pool_ns = ceph_try_get_string(src_ci->i_layout.pool_ns);
dst_oloc.pool = dst_ci->i_layout.pool_id;
dst_oloc.pool_ns = ceph_try_get_string(dst_ci->i_layout.pool_ns);
while (len >= object_size) {
ceph_calc_file_object_mapping(&src_ci->i_layout, *src_off,
object_size, &src_objnum,
&src_objoff, &src_objlen);
ceph_calc_file_object_mapping(&dst_ci->i_layout, *dst_off,
object_size, &dst_objnum,
&dst_objoff, &dst_objlen);
ceph_oid_init(&src_oid);
ceph_oid_printf(&src_oid, "%llx.%08llx",
src_ci->i_vino.ino, src_objnum);
ceph_oid_init(&dst_oid);
ceph_oid_printf(&dst_oid, "%llx.%08llx",
dst_ci->i_vino.ino, dst_objnum);
/* Do an object remote copy */
ret = ceph_osdc_copy_from(&fsc->client->osdc,
src_ci->i_vino.snap, 0,
&src_oid, &src_oloc,
CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL |
CEPH_OSD_OP_FLAG_FADVISE_NOCACHE,
&dst_oid, &dst_oloc,
CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL |
CEPH_OSD_OP_FLAG_FADVISE_DONTNEED,
dst_ci->i_truncate_seq,
dst_ci->i_truncate_size,
CEPH_OSD_COPY_FROM_FLAG_TRUNCATE_SEQ);
if (ret) {
if (ret == -EOPNOTSUPP) {
fsc->have_copy_from2 = false;
pr_notice("OSDs don't support copy-from2; disabling copy offload\n");
}
dout("ceph_osdc_copy_from returned %d\n", ret);
if (!bytes)
bytes = ret;
goto out;
}
len -= object_size;
bytes += object_size;
*src_off += object_size;
*dst_off += object_size;
}
out:
ceph_oloc_destroy(&src_oloc);
ceph_oloc_destroy(&dst_oloc);
return bytes;
}
vfs: no fallback for ->copy_file_range Now that we have generic_copy_file_range(), remove it as a fallback case when offloads fail. This puts the responsibility for executing fallbacks on the filesystems that implement ->copy_file_range and allows us to add operational validity checks to generic_copy_file_range(). Rework vfs_copy_file_range() to call a new do_copy_file_range() helper to execute the copying callout, and move calls to generic_file_copy_range() into filesystem methods where they currently return failures. [Amir] overlayfs is not responsible of executing the fallback. It is the responsibility of the underlying filesystem. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2019-06-05 23:04:47 +08:00
static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off,
struct file *dst_file, loff_t dst_off,
size_t len, unsigned int flags)
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
{
struct inode *src_inode = file_inode(src_file);
struct inode *dst_inode = file_inode(dst_file);
struct ceph_inode_info *src_ci = ceph_inode(src_inode);
struct ceph_inode_info *dst_ci = ceph_inode(dst_inode);
struct ceph_cap_flush *prealloc_cf;
ceph: allow object copies across different filesystems in the same cluster OSDs are able to perform object copies across different pools. Thus, there's no need to prevent copy_file_range from doing remote copies if the source and destination superblocks are different. Only return -EXDEV if they have different fsid (the cluster ID). Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-09-09 23:48:54 +08:00
struct ceph_fs_client *src_fsc = ceph_inode_to_client(src_inode);
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
loff_t size;
ssize_t ret = -EIO, bytes;
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
u64 src_objnum, dst_objnum, src_objoff, dst_objoff;
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
u32 src_objlen, dst_objlen;
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
int src_got = 0, dst_got = 0, err, dirty;
ceph: allow object copies across different filesystems in the same cluster OSDs are able to perform object copies across different pools. Thus, there's no need to prevent copy_file_range from doing remote copies if the source and destination superblocks are different. Only return -EXDEV if they have different fsid (the cluster ID). Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-09-09 23:48:54 +08:00
if (src_inode->i_sb != dst_inode->i_sb) {
struct ceph_fs_client *dst_fsc = ceph_inode_to_client(dst_inode);
if (ceph_fsid_compare(&src_fsc->client->fsid,
&dst_fsc->client->fsid)) {
dout("Copying files across clusters: src: %pU dst: %pU\n",
&src_fsc->client->fsid, &dst_fsc->client->fsid);
return -EXDEV;
}
}
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
if (ceph_snap(dst_inode) != CEPH_NOSNAP)
return -EROFS;
/*
* Some of the checks below will return -EOPNOTSUPP, which will force a
* fallback to the default VFS copy_file_range implementation. This is
* desirable in several cases (for ex, the 'len' is smaller than the
* size of the objects, or in cases where that would be more
* efficient).
*/
ceph: allow object copies across different filesystems in the same cluster OSDs are able to perform object copies across different pools. Thus, there's no need to prevent copy_file_range from doing remote copies if the source and destination superblocks are different. Only return -EXDEV if they have different fsid (the cluster ID). Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-09-09 23:48:54 +08:00
if (ceph_test_mount_opt(src_fsc, NOCOPYFROM))
ceph: new mount option to disable usage of copy-from op Add a new mount option 'nocopyfrom' that will prevent the usage of the RADOS 'copy-from' operation in cephfs. This could be useful, for example, for an administrator to temporarily mitigate any possible bugs in the 'copy-from' implementation. Currently, only copy_file_range uses this RADOS operation. Setting this mount option will result in this syscall reverting to the default VFS implementation, i.e. to perform the copies locally instead of doing remote object copies. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:46:00 +08:00
return -EOPNOTSUPP;
ceph: use copy-from2 op in copy_file_range Instead of using the copy-from operation, switch copy_file_range to the new copy-from2 operation, which allows to send the truncate_seq and truncate_size parameters. If an OSD does not support the copy-from2 operation it will return -EOPNOTSUPP. In that case, the kernel client will stop trying to do remote object copies for this fs client and will always use the generic VFS copy_file_range. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-01-08 18:03:53 +08:00
if (!src_fsc->have_copy_from2)
return -EOPNOTSUPP;
ceph: don't allow copy_file_range when stripe_count != 1 copy_file_range tries to use the OSD 'copy-from' operation, which simply performs a full object copy. Unfortunately, the implementation of this system call assumes that stripe_count is always set to 1 and doesn't take into account that the data may be striped across an object set. If the file layout has stripe_count different from 1, then the destination file data will be corrupted. For example: Consider a 8 MiB file with 4 MiB object size, stripe_count of 2 and stripe_size of 2 MiB; the first half of the file will be filled with 'A's and the second half will be filled with 'B's: 0 4M 8M Obj1 Obj2 +------+------+ +----+ +----+ file: | AAAA | BBBB | | AA | | AA | +------+------+ |----| |----| | BB | | BB | +----+ +----+ If we copy_file_range this file into a new file (which needs to have the same file layout!), then it will start by copying the object starting at file offset 0 (Obj1). And then it will copy the object starting at file offset 4M -- which is Obj1 again. Unfortunately, the solution for this is to not allow remote object copies to be performed when the file layout stripe_count is not 1 and simply fallback to the default (VFS) copy_file_range implementation. Cc: stable@vger.kernel.org Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-10-31 19:49:39 +08:00
/*
* Striped file layouts require that we copy partial objects, but the
* OSD copy-from operation only supports full-object copies. Limit
* this to non-striped file layouts for now.
*/
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
if ((src_ci->i_layout.stripe_unit != dst_ci->i_layout.stripe_unit) ||
ceph: don't allow copy_file_range when stripe_count != 1 copy_file_range tries to use the OSD 'copy-from' operation, which simply performs a full object copy. Unfortunately, the implementation of this system call assumes that stripe_count is always set to 1 and doesn't take into account that the data may be striped across an object set. If the file layout has stripe_count different from 1, then the destination file data will be corrupted. For example: Consider a 8 MiB file with 4 MiB object size, stripe_count of 2 and stripe_size of 2 MiB; the first half of the file will be filled with 'A's and the second half will be filled with 'B's: 0 4M 8M Obj1 Obj2 +------+------+ +----+ +----+ file: | AAAA | BBBB | | AA | | AA | +------+------+ |----| |----| | BB | | BB | +----+ +----+ If we copy_file_range this file into a new file (which needs to have the same file layout!), then it will start by copying the object starting at file offset 0 (Obj1). And then it will copy the object starting at file offset 4M -- which is Obj1 again. Unfortunately, the solution for this is to not allow remote object copies to be performed when the file layout stripe_count is not 1 and simply fallback to the default (VFS) copy_file_range implementation. Cc: stable@vger.kernel.org Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-10-31 19:49:39 +08:00
(src_ci->i_layout.stripe_count != 1) ||
(dst_ci->i_layout.stripe_count != 1) ||
(src_ci->i_layout.object_size != dst_ci->i_layout.object_size)) {
dout("Invalid src/dst files layout\n");
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
return -EOPNOTSUPP;
ceph: don't allow copy_file_range when stripe_count != 1 copy_file_range tries to use the OSD 'copy-from' operation, which simply performs a full object copy. Unfortunately, the implementation of this system call assumes that stripe_count is always set to 1 and doesn't take into account that the data may be striped across an object set. If the file layout has stripe_count different from 1, then the destination file data will be corrupted. For example: Consider a 8 MiB file with 4 MiB object size, stripe_count of 2 and stripe_size of 2 MiB; the first half of the file will be filled with 'A's and the second half will be filled with 'B's: 0 4M 8M Obj1 Obj2 +------+------+ +----+ +----+ file: | AAAA | BBBB | | AA | | AA | +------+------+ |----| |----| | BB | | BB | +----+ +----+ If we copy_file_range this file into a new file (which needs to have the same file layout!), then it will start by copying the object starting at file offset 0 (Obj1). And then it will copy the object starting at file offset 4M -- which is Obj1 again. Unfortunately, the solution for this is to not allow remote object copies to be performed when the file layout stripe_count is not 1 and simply fallback to the default (VFS) copy_file_range implementation. Cc: stable@vger.kernel.org Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-10-31 19:49:39 +08:00
}
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
if (len < src_ci->i_layout.object_size)
return -EOPNOTSUPP; /* no remote copy will be done */
prealloc_cf = ceph_alloc_cap_flush();
if (!prealloc_cf)
return -ENOMEM;
ceph: add destination file data sync before doing any remote copy If we try to copy into a file that was just written, any data that is remote copied will be overwritten by our buffered writes once they are flushed.  When this happens, the call to invalidate_inode_pages2_range will also return a -EBUSY error. This patch fixes this by also sync'ing the destination file before starting any copy. Fixes: 503f82a9932d ("ceph: support copy_file_range file operation") Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-24 00:53:14 +08:00
/* Start by sync'ing the source and destination files */
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
ret = file_write_and_wait_range(src_file, src_off, (src_off + len));
ceph: add destination file data sync before doing any remote copy If we try to copy into a file that was just written, any data that is remote copied will be overwritten by our buffered writes once they are flushed.  When this happens, the call to invalidate_inode_pages2_range will also return a -EBUSY error. This patch fixes this by also sync'ing the destination file before starting any copy. Fixes: 503f82a9932d ("ceph: support copy_file_range file operation") Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-24 00:53:14 +08:00
if (ret < 0) {
dout("failed to write src file (%zd)\n", ret);
goto out;
}
ret = file_write_and_wait_range(dst_file, dst_off, (dst_off + len));
if (ret < 0) {
dout("failed to write dst file (%zd)\n", ret);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
goto out;
ceph: add destination file data sync before doing any remote copy If we try to copy into a file that was just written, any data that is remote copied will be overwritten by our buffered writes once they are flushed.  When this happens, the call to invalidate_inode_pages2_range will also return a -EBUSY error. This patch fixes this by also sync'ing the destination file before starting any copy. Fixes: 503f82a9932d ("ceph: support copy_file_range file operation") Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-24 00:53:14 +08:00
}
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
/*
* We need FILE_WR caps for dst_ci and FILE_RD for src_ci as other
* clients may have dirty data in their caches. And OSDs know nothing
* about caps, so they can't safely do the remote object copies.
*/
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
err = get_rd_wr_caps(src_file, &src_got,
dst_file, (dst_off + len), &dst_got);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
if (err < 0) {
dout("get_rd_wr_caps returned %d\n", err);
ret = -EOPNOTSUPP;
goto out;
}
ret = is_file_size_ok(src_inode, dst_inode, src_off, dst_off, len);
if (ret < 0)
goto out_caps;
/* Drop dst file cached pages */
ret = invalidate_inode_pages2_range(dst_inode->i_mapping,
dst_off >> PAGE_SHIFT,
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
(dst_off + len) >> PAGE_SHIFT);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
if (ret < 0) {
dout("Failed to invalidate inode pages (%zd)\n", ret);
ret = 0; /* XXX */
}
ceph_calc_file_object_mapping(&src_ci->i_layout, src_off,
src_ci->i_layout.object_size,
&src_objnum, &src_objoff, &src_objlen);
ceph_calc_file_object_mapping(&dst_ci->i_layout, dst_off,
dst_ci->i_layout.object_size,
&dst_objnum, &dst_objoff, &dst_objlen);
/* object-level offsets need to the same */
if (src_objoff != dst_objoff) {
ret = -EOPNOTSUPP;
goto out_caps;
}
/*
* Do a manual copy if the object offset isn't object aligned.
* 'src_objlen' contains the bytes left until the end of the object,
* starting at the src_off
*/
if (src_objoff) {
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
dout("Initial partial copy of %u bytes\n", src_objlen);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
/*
* we need to temporarily drop all caps as we'll be calling
* {read,write}_iter, which will get caps again.
*/
put_rd_wr_caps(src_ci, src_got, dst_ci, dst_got);
ret = do_splice_direct(src_file, &src_off, dst_file,
&dst_off, src_objlen, flags);
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
/* Abort on short copies or on error */
if (ret < src_objlen) {
dout("Failed partial copy (%zd)\n", ret);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
goto out;
}
len -= ret;
ceph: pass filp to ceph_get_caps() Also change several other functions' arguments, no logical changes. This is preparetion for later patch that checks filp error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-07-25 20:16:43 +08:00
err = get_rd_wr_caps(src_file, &src_got,
dst_file, (dst_off + len), &dst_got);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
if (err < 0)
goto out;
err = is_file_size_ok(src_inode, dst_inode,
src_off, dst_off, len);
if (err < 0)
goto out_caps;
}
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
size = i_size_read(dst_inode);
bytes = ceph_do_objects_copy(src_ci, &src_off, dst_ci, &dst_off,
src_fsc, len, flags);
if (bytes <= 0) {
if (!ret)
ret = bytes;
goto out_caps;
}
dout("Copied %zu bytes out of %zu\n", bytes, len);
len -= bytes;
ret += bytes;
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
file_update_time(dst_file);
ceph: increment change_attribute on local changes We don't set SB_I_VERSION on ceph since we need to manage it ourselves, so we must increment it whenever we update the file times. Signed-off-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-06-06 20:57:27 +08:00
inode_inc_iversion_raw(dst_inode);
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
if (dst_off > size) {
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
/* Let the MDS know about dst file size change */
ceph: remove delay check logic from ceph_check_caps() __ceph_caps_file_wanted() already checks 'caps_wanted_delay_min' and 'caps_wanted_delay_max'. There is no need to duplicate the logic in ceph_check_caps() and __send_cap() Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-03-05 20:21:01 +08:00
if (ceph_inode_set_size(dst_inode, dst_off) ||
ceph_quota_is_max_bytes_approaching(dst_inode, dst_off))
ceph_check_caps(dst_ci, CHECK_CAPS_AUTHONLY, NULL);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
}
/* Mark Fw dirty */
spin_lock(&dst_ci->i_ceph_lock);
dst_ci->i_inline_version = CEPH_INLINE_NONE;
dirty = __ceph_mark_dirty_caps(dst_ci, CEPH_CAP_FILE_WR, &prealloc_cf);
spin_unlock(&dst_ci->i_ceph_lock);
if (dirty)
__mark_inode_dirty(dst_inode, dirty);
out_caps:
put_rd_wr_caps(src_ci, src_got, dst_ci, dst_got);
ceph: re-org copy_file_range and fix some error paths This patch re-organizes copy_file_range, trying to fix a few issues in the error handling. Here's the summary: - Abort copy if initial do_splice_direct() returns fewer bytes than requested. - Move the 'size' initialization (with i_size_read()) further down in the code, after the initial call to do_splice_direct(). This avoids issues with a possibly stale value if a manual copy is done. - Move the object copy loop into a separate function. This makes it easier to handle errors (e.g, dirtying caps and updating the MDS metadata if only some objects have been copied before an error has occurred). - Added calls to ceph_oloc_destroy() to avoid leaking memory with src_oloc and dst_oloc - After the object copy loop, the new file size to be reported to the MDS (if there's file size change) is now the actual file size, and not the size after an eventual extra manual copy. - Added a few dout() to show the number of bytes copied in the two manual copies and in the object copy loop. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2020-02-24 21:44:32 +08:00
/*
* Do the final manual copy if we still have some bytes left, unless
* there were errors in remote object copies (len >= object_size).
*/
if (len && (len < src_ci->i_layout.object_size)) {
dout("Final partial copy of %zu bytes\n", len);
bytes = do_splice_direct(src_file, &src_off, dst_file,
&dst_off, len, flags);
if (bytes > 0)
ret += bytes;
else
dout("Failed partial copy (%zd)\n", bytes);
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
}
out:
ceph_free_cap_flush(prealloc_cf);
return ret;
}
vfs: no fallback for ->copy_file_range Now that we have generic_copy_file_range(), remove it as a fallback case when offloads fail. This puts the responsibility for executing fallbacks on the filesystems that implement ->copy_file_range and allows us to add operational validity checks to generic_copy_file_range(). Rework vfs_copy_file_range() to call a new do_copy_file_range() helper to execute the copying callout, and move calls to generic_file_copy_range() into filesystem methods where they currently return failures. [Amir] overlayfs is not responsible of executing the fallback. It is the responsibility of the underlying filesystem. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2019-06-05 23:04:47 +08:00
static ssize_t ceph_copy_file_range(struct file *src_file, loff_t src_off,
struct file *dst_file, loff_t dst_off,
size_t len, unsigned int flags)
{
ssize_t ret;
ret = __ceph_copy_file_range(src_file, src_off, dst_file, dst_off,
len, flags);
vfs: allow copy_file_range to copy across devices We want to enable cross-filesystem copy_file_range functionality where possible, so push the "same superblock only" checks down to the individual filesystem callouts so they can make their own decisions about cross-superblock copy offload and fallack to generic_copy_file_range() for cross-superblock copy. [Amir] We do not call ->remap_file_range() in case the files are not on the same sb and do not call ->copy_file_range() in case the files do not belong to the same filesystem driver. This changes behavior of the copy_file_range(2) syscall, which will now allow cross filesystem in-kernel copy. CIFS already supports cross-superblock copy, between two shares to the same server. This functionality will now be available via the copy_file_range(2) syscall. Cc: Steve French <stfrench@microsoft.com> Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2019-06-05 23:04:50 +08:00
if (ret == -EOPNOTSUPP || ret == -EXDEV)
vfs: no fallback for ->copy_file_range Now that we have generic_copy_file_range(), remove it as a fallback case when offloads fail. This puts the responsibility for executing fallbacks on the filesystems that implement ->copy_file_range and allows us to add operational validity checks to generic_copy_file_range(). Rework vfs_copy_file_range() to call a new do_copy_file_range() helper to execute the copying callout, and move calls to generic_file_copy_range() into filesystem methods where they currently return failures. [Amir] overlayfs is not responsible of executing the fallback. It is the responsibility of the underlying filesystem. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2019-06-05 23:04:47 +08:00
ret = generic_copy_file_range(src_file, src_off, dst_file,
dst_off, len, flags);
return ret;
}
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
const struct file_operations ceph_file_fops = {
.open = ceph_open,
.release = ceph_release,
.llseek = ceph_llseek,
ceph: switch to ->read_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-03 08:28:01 +08:00
.read_iter = ceph_read_iter,
ceph: switch to ->write_iter() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-04 11:09:01 +08:00
.write_iter = ceph_write_iter,
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
.mmap = ceph_mmap,
.fsync = ceph_fsync,
ceph: add flock/fcntl lock support Implement flock inode operation to support advisory file locking. All lock/unlock operations are synchronous with the MDS. Lock state is sent when reconnecting to a recovering MDS to restore the shared lock state. Signed-off-by: Greg Farnum <gregf@hq.newdream.net> Signed-off-by: Sage Weil <sage@newdream.net>
2010-08-03 06:34:23 +08:00
.lock = ceph_lock,
.flock = ceph_flock,
ceph: fix splice read for no Fc capability case When iov_iter type is ITER_PIPE, copy_page_to_iter() increases the page's reference and add the page to a pipe_buffer. It also set the pipe_buffer's ops to page_cache_pipe_buf_ops. The comfirm callback in page_cache_pipe_buf_ops expects the page is from page cache and uptodate, otherwise it return error. For ceph_sync_read() case, pages are not from page cache. So we can't call copy_page_to_iter() when iov_iter type is ITER_PIPE. The fix is using iov_iter_get_pages_alloc() to allocate pages for the pipe. (the code is similar to default_file_splice_read) Signed-off-by: Yan, Zheng <zyan@redhat.com>
2016-11-08 21:54:34 +08:00
.splice_read = generic_file_splice_read,
ceph: switch to iter_file_splice_write() Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2014-04-05 16:40:12 +08:00
.splice_write = iter_file_splice_write,
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
.unlocked_ioctl = ceph_ioctl,
ceph: fix compat_ioctl for ceph_dir_operations The ceph_ioctl function is used both for files and directories, but only the files support doing that in 32-bit compat mode. On the s390 architecture, there is also a problem with invalid 31-bit pointers that need to be passed through compat_ptr(). Use the new compat_ptr_ioctl() to address both issues. Note: When backporting this patch to stable kernels, "compat_ioctl: add compat_ptr_ioctl()" is needed as well. Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Cc: stable@vger.kernel.org Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2018-09-12 02:47:23 +08:00
.compat_ioctl = compat_ptr_ioctl,
ceph: punch hole support This patch implements fallocate and punch hole support for Ceph kernel client. Signed-off-by: Li Wang <liwang@ubuntukylin.com> Signed-off-by: Yunchuan Wen <yunchuanwen@ubuntukylin.com>
2013-08-15 11:51:44 +08:00
.fallocate = ceph_fallocate,
ceph: support copy_file_range file operation This commit implements support for the copy_file_range syscall in cephfs. It is implemented using the RADOS 'copy-from' operation, which allows to do a remote object copy, without the need to download/upload data from/to the OSDs. Some manual copy may however be required if the source/destination file offsets aren't object aligned or if the copy length is smaller than the object size. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2018-10-15 23:45:59 +08:00
.copy_file_range = ceph_copy_file_range,
ceph: file operations File open and close operations, and read and write methods that ensure we have obtained the proper capabilities from the MDS cluster before performing IO on a file. We take references on held capabilities for the duration of the read/write to avoid prematurely releasing them back to the MDS. We implement two main paths for read and write: one that is buffered (and uses generic_aio_{read,write}), and one that is fully synchronous and blocking (operating either on a __user pointer or, if O_DIRECT, directly on user pages). Signed-off-by: Sage Weil <sage@newdream.net>
2009-10-07 02:31:08 +08:00
};
Copy Permalink