linux/fs/afs/write.c

948 lines
22 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/* handling of writes to regular files and writing back to the server
*
* Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*/
#include <linux/backing-dev.h>
#include <linux/slab.h>
#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/writeback.h>
#include <linux/pagevec.h>
#include "internal.h"
/*
* mark a page as having been made dirty and thus needing writeback
*/
int afs_set_page_dirty(struct page *page)
{
_enter("");
return __set_page_dirty_nobuffers(page);
}
/*
* partly or wholly fill a page that's under preparation for writing
*/
static int afs_fill_page(struct afs_vnode *vnode, struct key *key,
loff_t pos, unsigned int len, struct page *page)
{
struct afs_read *req;
size_t p;
void *data;
int ret;
_enter(",,%llu", (unsigned long long)pos);
if (pos >= vnode->vfs_inode.i_size) {
p = pos & ~PAGE_MASK;
ASSERTCMP(p + len, <=, PAGE_SIZE);
data = kmap(page);
memset(data + p, 0, len);
kunmap(page);
return 0;
}
req = kzalloc(struct_size(req, array, 1), GFP_KERNEL);
if (!req)
return -ENOMEM;
afs: Fix directory handling AFS directories are structured blobs that are downloaded just like files and then parsed by the lookup and readdir code and, as such, are currently handled in the pagecache like any other file, with the entire directory content being thrown away each time the directory changes. However, since the blob is a known structure and since the data version counter on a directory increases by exactly one for each change committed to that directory, we can actually edit the directory locally rather than fetching it from the server after each locally-induced change. What we can't do, though, is mix data from the server and data from the client since the server is technically at liberty to rearrange or compress a directory if it sees fit, provided it updates the data version number when it does so and breaks the callback (ie. sends a notification). Further, lookup with lookup-ahead, readdir and, when it arrives, local editing are likely want to scan the whole of a directory. So directory handling needs to be improved to maintain the coherency of the directory blob prior to permitting local directory editing. To this end: (1) If any directory page gets discarded, invalidate and reread the entire directory. (2) If readpage notes that if when it fetches a single page that the version number has changed, the entire directory is flagged for invalidation. (3) Read as much of the directory in one go as we can. Note that this removes local caching of directories in fscache for the moment as we can't pass the pages to fscache_read_or_alloc_pages() since page->lru is in use by the LRU. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
refcount_set(&req->usage, 1);
req->pos = pos;
req->len = len;
req->nr_pages = 1;
afs: Fix directory handling AFS directories are structured blobs that are downloaded just like files and then parsed by the lookup and readdir code and, as such, are currently handled in the pagecache like any other file, with the entire directory content being thrown away each time the directory changes. However, since the blob is a known structure and since the data version counter on a directory increases by exactly one for each change committed to that directory, we can actually edit the directory locally rather than fetching it from the server after each locally-induced change. What we can't do, though, is mix data from the server and data from the client since the server is technically at liberty to rearrange or compress a directory if it sees fit, provided it updates the data version number when it does so and breaks the callback (ie. sends a notification). Further, lookup with lookup-ahead, readdir and, when it arrives, local editing are likely want to scan the whole of a directory. So directory handling needs to be improved to maintain the coherency of the directory blob prior to permitting local directory editing. To this end: (1) If any directory page gets discarded, invalidate and reread the entire directory. (2) If readpage notes that if when it fetches a single page that the version number has changed, the entire directory is flagged for invalidation. (3) Read as much of the directory in one go as we can. Note that this removes local caching of directories in fscache for the moment as we can't pass the pages to fscache_read_or_alloc_pages() since page->lru is in use by the LRU. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
req->pages = req->array;
req->pages[0] = page;
get_page(page);
afs: Overhaul volume and server record caching and fileserver rotation The current code assumes that volumes and servers are per-cell and are never shared, but this is not enforced, and, indeed, public cells do exist that are aliases of each other. Further, an organisation can, say, set up a public cell and a private cell with overlapping, but not identical, sets of servers. The difference is purely in the database attached to the VL servers. The current code will malfunction if it sees a server in two cells as it assumes global address -> server record mappings and that each server is in just one cell. Further, each server may have multiple addresses - and may have addresses of different families (IPv4 and IPv6, say). To this end, the following structural changes are made: (1) Server record management is overhauled: (a) Server records are made independent of cell. The namespace keeps track of them, volume records have lists of them and each vnode has a server on which its callback interest currently resides. (b) The cell record no longer keeps a list of servers known to be in that cell. (c) The server records are now kept in a flat list because there's no single address to sort on. (d) Server records are now keyed by their UUID within the namespace. (e) The addresses for a server are obtained with the VL.GetAddrsU rather than with VL.GetEntryByName, using the server's UUID as a parameter. (f) Cached server records are garbage collected after a period of non-use and are counted out of existence before purging is allowed to complete. This protects the work functions against rmmod. (g) The servers list is now in /proc/fs/afs/servers. (2) Volume record management is overhauled: (a) An RCU-replaceable server list is introduced. This tracks both servers and their coresponding callback interests. (b) The superblock is now keyed on cell record and numeric volume ID. (c) The volume record is now tied to the superblock which mounts it, and is activated when mounted and deactivated when unmounted. This makes it easier to handle the cache cookie without causing a double-use in fscache. (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU to get the server UUID list. (e) The volume name is updated if it is seen to have changed when the volume is updated (the update is keyed on the volume ID). (3) The vlocation record is got rid of and VLDB records are no longer cached. Sufficient information is stored in the volume record, though an update to a volume record is now no longer shared between related volumes (volumes come in bundles of three: R/W, R/O and backup). and the following procedural changes are made: (1) The fileserver cursor introduced previously is now fleshed out and used to iterate over fileservers and their addresses. (2) Volume status is checked during iteration, and the server list is replaced if a change is detected. (3) Server status is checked during iteration, and the address list is replaced if a change is detected. (4) The abort code is saved into the address list cursor and -ECONNABORTED returned in afs_make_call() if a remote abort happened rather than translating the abort into an error message. This allows actions to be taken depending on the abort code more easily. (a) If a VMOVED abort is seen then this is handled by rechecking the volume and restarting the iteration. (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is handled by sleeping for a short period and retrying and/or trying other servers that might serve that volume. A message is also displayed once until the condition has cleared. (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the moment. (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to see if it has been deleted; if not, the fileserver is probably indicating that the volume couldn't be attached and needs salvaging. (e) If statfs() sees one of these aborts, it does not sleep, but rather returns an error, so as not to block the umount program. (5) The fileserver iteration functions in vnode.c are now merged into their callers and more heavily macroised around the cursor. vnode.c is removed. (6) Operations on a particular vnode are serialised on that vnode because the server will lock that vnode whilst it operates on it, so a second op sent will just have to wait. (7) Fileservers are probed with FS.GetCapabilities before being used. This is where service upgrade will be done. (8) A callback interest on a fileserver is set up before an FS operation is performed and passed through to afs_make_call() so that it can be set on the vnode if the operation returns a callback. The callback interest is passed through to afs_iget() also so that it can be set there too. In general, record updating is done on an as-needed basis when we try to access servers, volumes or vnodes rather than offloading it to work items and special threads. Notes: (1) Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998). (2) VBUSY is retried forever for the moment at intervals of 1s. (3) /proc/fs/afs/<cell>/servers no longer exists. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:50 +08:00
ret = afs_fetch_data(vnode, key, req);
afs_put_read(req);
if (ret < 0) {
if (ret == -ENOENT) {
_debug("got NOENT from server"
" - marking file deleted and stale");
set_bit(AFS_VNODE_DELETED, &vnode->flags);
ret = -ESTALE;
}
}
_leave(" = %d", ret);
return ret;
}
/*
* prepare to perform part of a write to a page
*/
int afs_write_begin(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len, unsigned flags,
struct page **_page, void **fsdata)
{
struct afs_vnode *vnode = AFS_FS_I(file_inode(file));
struct page *page;
struct key *key = afs_file_key(file);
unsigned long priv;
unsigned f, from = pos & (PAGE_SIZE - 1);
unsigned t, to = from + 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
pgoff_t index = pos >> PAGE_SHIFT;
int ret;
_enter("{%llx:%llu},{%lx},%u,%u",
vnode->fid.vid, vnode->fid.vnode, index, from, to);
fs: symlink write_begin allocation context fix With the write_begin/write_end aops, page_symlink was broken because it could no longer pass a GFP_NOFS type mask into the point where the allocations happened. They are done in write_begin, which would always assume that the filesystem can be entered from reclaim. This bug could cause filesystem deadlocks. The funny thing with having a gfp_t mask there is that it doesn't really allow the caller to arbitrarily tinker with the context in which it can be called. It couldn't ever be GFP_ATOMIC, for example, because it needs to take the page lock. The only thing any callers care about is __GFP_FS anyway, so turn that into a single flag. Add a new flag for write_begin, AOP_FLAG_NOFS. Filesystems can now act on this flag in their write_begin function. Change __grab_cache_page to accept a nofs argument as well, to honour that flag (while we're there, change the name to grab_cache_page_write_begin which is more instructive and does away with random leading underscores). This is really a more flexible way to go in the end anyway -- if a filesystem happens to want any extra allocations aside from the pagecache ones in ints write_begin function, it may now use GFP_KERNEL (rather than GFP_NOFS) for common case allocations (eg. ocfs2_alloc_write_ctxt, for a random example). [kosaki.motohiro@jp.fujitsu.com: fix ubifs] [kosaki.motohiro@jp.fujitsu.com: fix fuse] Signed-off-by: Nick Piggin <npiggin@suse.de> Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: <stable@kernel.org> [2.6.28.x] Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> [ Cleaned up the calling convention: just pass in the AOP flags untouched to the grab_cache_page_write_begin() function. That just simplifies everybody, and may even allow future expansion of the logic. - Linus ] Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-05 04:00:53 +08:00
page = grab_cache_page_write_begin(mapping, index, flags);
if (!page)
return -ENOMEM;
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 (!PageUptodate(page) && len != PAGE_SIZE) {
ret = afs_fill_page(vnode, key, pos & PAGE_MASK, PAGE_SIZE, page);
if (ret < 0) {
unlock_page(page);
put_page(page);
_leave(" = %d [prep]", ret);
return ret;
}
SetPageUptodate(page);
}
try_again:
/* See if this page is already partially written in a way that we can
* merge the new write with.
*/
t = f = 0;
if (PagePrivate(page)) {
priv = page_private(page);
f = afs_page_dirty_from(priv);
t = afs_page_dirty_to(priv);
ASSERTCMP(f, <=, t);
}
if (f != t) {
afs: Make afs_write_begin() avoid writing to a page that's being stored Make afs_write_begin() wait for a page that's marked PG_writeback because: (1) We need to avoid interference with the data being stored so that the data on the server ends up in a defined state. (2) page->private is used to track the window of dirty data within a page, but it's also used by the storage code to track what's being written, being cleared by the completion notification. Ownership can't be relinquished by the storage code until completion because it a store fails, the data must be remarked dirty. Tracing shows something like the following (edited): x86_64-linux-gn-15940 [1] afs_page_dirty: vn=ffff8800bef33800 9c75 begin 0-125 kworker/u8:3-114 [2] afs_page_dirty: vn=ffff8800bef33800 9c75 store+ 0-125 x86_64-linux-gn-15940 [1] afs_page_dirty: vn=ffff8800bef33800 9c75 begin 0-2052 kworker/u8:3-114 [2] afs_page_dirty: vn=ffff8800bef33800 9c75 clear 0-2052 kworker/u8:3-114 [2] afs_page_dirty: vn=ffff8800bef33800 9c75 store 0-0 kworker/u8:3-114 [2] afs_page_dirty: vn=ffff8800bef33800 9c75 WARN 0-0 The clear (completion) corresponding to the store+ (store continuation from a previous page) happens between the second begin (afs_write_begin) and the store corresponding to that. This results in the second store not seeing any data to write back, leading to the following warning: WARNING: CPU: 2 PID: 114 at ../fs/afs/write.c:403 afs_write_back_from_locked_page+0x19d/0x76c [kafs] Modules linked in: kafs(E) CPU: 2 PID: 114 Comm: kworker/u8:3 Tainted: G E 4.14.0-fscache+ #242 Hardware name: ASUS All Series/H97-PLUS, BIOS 2306 10/09/2014 Workqueue: writeback wb_workfn (flush-afs-2) task: ffff8800cad72600 task.stack: ffff8800cad44000 RIP: 0010:afs_write_back_from_locked_page+0x19d/0x76c [kafs] RSP: 0018:ffff8800cad47aa0 EFLAGS: 00010246 RAX: 0000000000000001 RBX: ffff8800bef33a20 RCX: 0000000000000000 RDX: 000000000000000f RSI: ffffffff81c5d0e0 RDI: ffff8800cad72e78 RBP: ffff8800d31ea1e8 R08: ffff8800c1358000 R09: ffff8800ca00e400 R10: ffff8800cad47a38 R11: ffff8800c5d9e400 R12: 0000000000000000 R13: ffffea0002d9df00 R14: ffffffffa0023c1c R15: 0000000000007fdf FS: 0000000000000000(0000) GS:ffff8800ca700000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f85ac6c4000 CR3: 0000000001c10001 CR4: 00000000001606e0 Call Trace: ? clear_page_dirty_for_io+0x23a/0x267 afs_writepages_region+0x1be/0x286 [kafs] afs_writepages+0x60/0x127 [kafs] do_writepages+0x36/0x70 __writeback_single_inode+0x12f/0x635 writeback_sb_inodes+0x2cc/0x452 __writeback_inodes_wb+0x68/0x9f wb_writeback+0x208/0x470 ? wb_workfn+0x22b/0x565 wb_workfn+0x22b/0x565 ? worker_thread+0x230/0x2ac process_one_work+0x2cc/0x517 ? worker_thread+0x230/0x2ac worker_thread+0x1d4/0x2ac ? rescuer_thread+0x29b/0x29b kthread+0x15d/0x165 ? kthread_create_on_node+0x3f/0x3f ? call_usermodehelper_exec_async+0x118/0x11f ret_from_fork+0x24/0x30 Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-18 08:13:30 +08:00
if (PageWriteback(page)) {
trace_afs_page_dirty(vnode, tracepoint_string("alrdy"),
page->index, priv);
goto flush_conflicting_write;
}
/* If the file is being filled locally, allow inter-write
* spaces to be merged into writes. If it's not, only write
* back what the user gives us.
*/
if (!test_bit(AFS_VNODE_NEW_CONTENT, &vnode->flags) &&
(to < f || from > t))
goto flush_conflicting_write;
}
*_page = page;
_leave(" = 0");
return 0;
/* The previous write and this write aren't adjacent or overlapping, so
* flush the page out.
*/
flush_conflicting_write:
_debug("flush conflict");
ret = write_one_page(page);
if (ret < 0)
goto error;
ret = lock_page_killable(page);
if (ret < 0)
goto error;
goto try_again;
error:
put_page(page);
_leave(" = %d", ret);
return ret;
}
/*
* finalise part of a write to a page
*/
int afs_write_end(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len, unsigned copied,
struct page *page, void *fsdata)
{
struct afs_vnode *vnode = AFS_FS_I(file_inode(file));
struct key *key = afs_file_key(file);
unsigned long priv;
unsigned int f, from = pos & (PAGE_SIZE - 1);
unsigned int t, to = from + copied;
loff_t i_size, maybe_i_size;
int ret = 0;
_enter("{%llx:%llu},{%lx}",
vnode->fid.vid, vnode->fid.vnode, page->index);
if (copied == 0)
goto out;
maybe_i_size = pos + copied;
i_size = i_size_read(&vnode->vfs_inode);
if (maybe_i_size > i_size) {
write_seqlock(&vnode->cb_lock);
i_size = i_size_read(&vnode->vfs_inode);
if (maybe_i_size > i_size)
i_size_write(&vnode->vfs_inode, maybe_i_size);
write_sequnlock(&vnode->cb_lock);
}
if (!PageUptodate(page)) {
if (copied < len) {
/* Try and load any missing data from the server. The
* unmarshalling routine will take care of clearing any
* bits that are beyond the EOF.
*/
ret = afs_fill_page(vnode, key, pos + copied,
len - copied, page);
if (ret < 0)
goto out;
}
SetPageUptodate(page);
}
if (PagePrivate(page)) {
priv = page_private(page);
f = afs_page_dirty_from(priv);
t = afs_page_dirty_to(priv);
if (from < f)
f = from;
if (to > t)
t = to;
priv = afs_page_dirty(f, t);
set_page_private(page, priv);
trace_afs_page_dirty(vnode, tracepoint_string("dirty+"),
page->index, priv);
} else {
priv = afs_page_dirty(from, to);
attach_page_private(page, (void *)priv);
trace_afs_page_dirty(vnode, tracepoint_string("dirty"),
page->index, priv);
}
set_page_dirty(page);
if (PageDirty(page))
_debug("dirtied");
ret = copied;
out:
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);
return ret;
}
/*
* kill all the pages in the given range
*/
static void afs_kill_pages(struct address_space *mapping,
pgoff_t first, pgoff_t last)
{
struct afs_vnode *vnode = AFS_FS_I(mapping->host);
struct pagevec pv;
unsigned count, loop;
_enter("{%llx:%llu},%lx-%lx",
vnode->fid.vid, vnode->fid.vnode, first, last);
pagevec_init(&pv);
do {
_debug("kill %lx-%lx", first, last);
count = last - first + 1;
if (count > PAGEVEC_SIZE)
count = PAGEVEC_SIZE;
pv.nr = find_get_pages_contig(mapping, first, count, pv.pages);
ASSERTCMP(pv.nr, ==, count);
for (loop = 0; loop < count; loop++) {
struct page *page = pv.pages[loop];
ClearPageUptodate(page);
SetPageError(page);
end_page_writeback(page);
if (page->index >= first)
first = page->index + 1;
lock_page(page);
generic_error_remove_page(mapping, page);
unlock_page(page);
}
__pagevec_release(&pv);
} while (first <= last);
_leave("");
}
/*
* Redirty all the pages in a given range.
*/
static void afs_redirty_pages(struct writeback_control *wbc,
struct address_space *mapping,
pgoff_t first, pgoff_t last)
{
struct afs_vnode *vnode = AFS_FS_I(mapping->host);
struct pagevec pv;
unsigned count, loop;
_enter("{%llx:%llu},%lx-%lx",
vnode->fid.vid, vnode->fid.vnode, first, last);
AFS development -----BEGIN PGP SIGNATURE----- iQIVAwUAWgm9V/Sw1s6N8H32AQK5mQ//QGUDZLXsUPCtq0XJq0V+r4MUjNp9tCZR htiuNrEkHSyPpYgCcQ2Aqdl9kndwVXcE7lWT99mp/a0zwNAsp9GOGVhCXUd5R86G XlrBuUYVvBJk18tDsUNWdjRQ0gMHgQSlEnEbsaGiU1bVrpXatI9hL8qoeO78Iy7+ eaJUQLCuCVJq7qMQGhC0hg338vmHVeYhnViXIxq+HFjsMmR9IVanuK+sQr6NSJxS F6RkPxBUPWkRVMHmxTLWj/XSHZwtwu+Mnc/UFYsAPLKEbY0cIohsI8EgfE8U7geU yRVnu3MIOXUXUrZizj9SwVYWdJfneRlINqMbHIO8QXMKR38tnQ0C2/7bgBsXiNPv YdiAyeqL4nM+JthV/rgA3hWgupwBlSb4ubclTphDNxMs5MBIUIK3XUt9GOXDDUZz 2FT/FdrphM2UORaI2AEOi4Q0/nHdin+3rld8fjV0Ree/TPNXwcrOmvy8yGnxFCEp 5b7YLwKrffZGnnS965dhZlnFR6hjndmzFgHdyRrJwc80hXi1Q/+W4F19MoYkkoVK G/gLvD3FbmygmFnjCik9TjUrro6vQxo56H/TuWgHTvYriNGH+D/D7EGUwg4GiXZZ +7vrNw660uXmZiu9i0YacCRyD8lvm7QpmWLb+uHwzfsBE1+C8UetyQ+egSWVdWJO KwPspygWXD4= =3vy0 -----END PGP SIGNATURE----- Merge tag 'afs-next-20171113' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs Pull AFS updates from David Howells: "kAFS filesystem driver overhaul. The major points of the overhaul are: (1) Preliminary groundwork is laid for supporting network-namespacing of kAFS. The remainder of the namespacing work requires some way to pass namespace information to submounts triggered by an automount. This requires something like the mount overhaul that's in progress. (2) sockaddr_rxrpc is used in preference to in_addr for holding addresses internally and add support for talking to the YFS VL server. With this, kAFS can do everything over IPv6 as well as IPv4 if it's talking to servers that support it. (3) Callback handling is overhauled to be generally passive rather than active. 'Callbacks' are promises by the server to tell us about data and metadata changes. Callbacks are now checked when we next touch an inode rather than actively going and looking for it where possible. (4) File access permit caching is overhauled to store the caching information per-inode rather than per-directory, shared over subordinate files. Whilst older AFS servers only allow ACLs on directories (shared to the files in that directory), newer AFS servers break that restriction. To improve memory usage and to make it easier to do mass-key removal, permit combinations are cached and shared. (5) Cell database management is overhauled to allow lighter locks to be used and to make cell records autonomous state machines that look after getting their own DNS records and cleaning themselves up, in particular preventing races in acquiring and relinquishing the fscache token for the cell. (6) Volume caching is overhauled. The afs_vlocation record is got rid of to simplify things and the superblock is now keyed on the cell and the numeric volume ID only. The volume record is tied to a superblock and normal superblock management is used to mediate the lifetime of the volume fscache token. (7) File server record caching is overhauled to make server records independent of cells and volumes. A server can be in multiple cells (in such a case, the administrator must make sure that the VL services for all cells correctly reflect the volumes shared between those cells). Server records are now indexed using the UUID of the server rather than the address since a server can have multiple addresses. (8) File server rotation is overhauled to handle VMOVED, VBUSY (and similar), VOFFLINE and VNOVOL indications and to handle rotation both of servers and addresses of those servers. The rotation will also wait and retry if the server says it is busy. (9) Data writeback is overhauled. Each inode no longer stores a list of modified sections tagged with the key that authorised it in favour of noting the modified region of a page in page->private and storing a list of keys that made modifications in the inode. This simplifies things and allows other keys to be used to actually write to the server if a key that made a modification becomes useless. (10) Writable mmap() is implemented. This allows a kernel to be build entirely on AFS. Note that Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998)" * tag 'afs-next-20171113' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs: (35 commits) afs: Protect call->state changes against signals afs: Trace page dirty/clean afs: Implement shared-writeable mmap afs: Get rid of the afs_writeback record afs: Introduce a file-private data record afs: Use a dynamic port if 7001 is in use afs: Fix directory read/modify race afs: Trace the sending of pages afs: Trace the initiation and completion of client calls afs: Fix documentation on # vs % prefix in mount source specification afs: Fix total-length calculation for multiple-page send afs: Only progress call state at end of Tx phase from rxrpc callback afs: Make use of the YFS service upgrade to fully support IPv6 afs: Overhaul volume and server record caching and fileserver rotation afs: Move server rotation code into its own file afs: Add an address list concept afs: Overhaul cell database management afs: Overhaul permit caching afs: Overhaul the callback handling afs: Rename struct afs_call server member to cm_server ...
2017-11-17 03:41:22 +08:00
pagevec_init(&pv);
do {
_debug("redirty %lx-%lx", first, last);
count = last - first + 1;
if (count > PAGEVEC_SIZE)
count = PAGEVEC_SIZE;
pv.nr = find_get_pages_contig(mapping, first, count, pv.pages);
ASSERTCMP(pv.nr, ==, count);
for (loop = 0; loop < count; loop++) {
struct page *page = pv.pages[loop];
redirty_page_for_writepage(wbc, page);
end_page_writeback(page);
if (page->index >= first)
first = page->index + 1;
}
__pagevec_release(&pv);
} while (first <= last);
_leave("");
}
afs: Fix application of status and callback to be under same lock When applying the status and callback in the response of an operation, apply them in the same critical section so that there's no race between checking the callback state and checking status-dependent state (such as the data version). Fix this by: (1) Allocating a joint {status,callback} record (afs_status_cb) before calling the RPC function for each vnode for which the RPC reply contains a status or a status plus a callback. A flag is set in the record to indicate if a callback was actually received. (2) These records are passed into the RPC functions to be filled in. The afs_decode_status() and yfs_decode_status() functions are removed and the cb_lock is no longer taken. (3) xdr_decode_AFSFetchStatus() and xdr_decode_YFSFetchStatus() no longer update the vnode. (4) xdr_decode_AFSCallBack() and xdr_decode_YFSCallBack() no longer update the vnode. (5) vnodes, expected data-version numbers and callback break counters (cb_break) no longer need to be passed to the reply delivery functions. Note that, for the moment, the file locking functions still need access to both the call and the vnode at the same time. (6) afs_vnode_commit_status() is now given the cb_break value and the expected data_version and the task of applying the status and the callback to the vnode are now done here. This is done under a single taking of vnode->cb_lock. (7) afs_pages_written_back() is now called by afs_store_data() rather than by the reply delivery function. afs_pages_written_back() has been moved to before the call point and is now given the first and last page numbers rather than a pointer to the call. (8) The indicator from YFS.RemoveFile2 as to whether the target file actually got removed (status.abort_code == VNOVNODE) rather than merely dropping a link is now checked in afs_unlink rather than in xdr_decode_YFSFetchStatus(). Supplementary fixes: (*) afs_cache_permit() now gets the caller_access mask from the afs_status_cb object rather than picking it out of the vnode's status record. afs_fetch_status() returns caller_access through its argument list for this purpose also. (*) afs_inode_init_from_status() now uses a write lock on cb_lock rather than a read lock and now sets the callback inside the same critical section. Fixes: c435ee34551e ("afs: Overhaul the callback handling") Signed-off-by: David Howells <dhowells@redhat.com>
2019-05-09 22:16:10 +08:00
/*
* completion of write to server
*/
static void afs_pages_written_back(struct afs_vnode *vnode,
pgoff_t first, pgoff_t last)
{
struct pagevec pv;
unsigned long priv;
unsigned count, loop;
_enter("{%llx:%llu},{%lx-%lx}",
vnode->fid.vid, vnode->fid.vnode, first, last);
pagevec_init(&pv);
do {
_debug("done %lx-%lx", first, last);
count = last - first + 1;
if (count > PAGEVEC_SIZE)
count = PAGEVEC_SIZE;
pv.nr = find_get_pages_contig(vnode->vfs_inode.i_mapping,
first, count, pv.pages);
ASSERTCMP(pv.nr, ==, count);
for (loop = 0; loop < count; loop++) {
priv = (unsigned long)detach_page_private(pv.pages[loop]);
afs: Fix application of status and callback to be under same lock When applying the status and callback in the response of an operation, apply them in the same critical section so that there's no race between checking the callback state and checking status-dependent state (such as the data version). Fix this by: (1) Allocating a joint {status,callback} record (afs_status_cb) before calling the RPC function for each vnode for which the RPC reply contains a status or a status plus a callback. A flag is set in the record to indicate if a callback was actually received. (2) These records are passed into the RPC functions to be filled in. The afs_decode_status() and yfs_decode_status() functions are removed and the cb_lock is no longer taken. (3) xdr_decode_AFSFetchStatus() and xdr_decode_YFSFetchStatus() no longer update the vnode. (4) xdr_decode_AFSCallBack() and xdr_decode_YFSCallBack() no longer update the vnode. (5) vnodes, expected data-version numbers and callback break counters (cb_break) no longer need to be passed to the reply delivery functions. Note that, for the moment, the file locking functions still need access to both the call and the vnode at the same time. (6) afs_vnode_commit_status() is now given the cb_break value and the expected data_version and the task of applying the status and the callback to the vnode are now done here. This is done under a single taking of vnode->cb_lock. (7) afs_pages_written_back() is now called by afs_store_data() rather than by the reply delivery function. afs_pages_written_back() has been moved to before the call point and is now given the first and last page numbers rather than a pointer to the call. (8) The indicator from YFS.RemoveFile2 as to whether the target file actually got removed (status.abort_code == VNOVNODE) rather than merely dropping a link is now checked in afs_unlink rather than in xdr_decode_YFSFetchStatus(). Supplementary fixes: (*) afs_cache_permit() now gets the caller_access mask from the afs_status_cb object rather than picking it out of the vnode's status record. afs_fetch_status() returns caller_access through its argument list for this purpose also. (*) afs_inode_init_from_status() now uses a write lock on cb_lock rather than a read lock and now sets the callback inside the same critical section. Fixes: c435ee34551e ("afs: Overhaul the callback handling") Signed-off-by: David Howells <dhowells@redhat.com>
2019-05-09 22:16:10 +08:00
trace_afs_page_dirty(vnode, tracepoint_string("clear"),
pv.pages[loop]->index, priv);
end_page_writeback(pv.pages[loop]);
}
first += count;
__pagevec_release(&pv);
} while (first <= last);
afs_prune_wb_keys(vnode);
_leave("");
}
afs: Overhaul volume and server record caching and fileserver rotation The current code assumes that volumes and servers are per-cell and are never shared, but this is not enforced, and, indeed, public cells do exist that are aliases of each other. Further, an organisation can, say, set up a public cell and a private cell with overlapping, but not identical, sets of servers. The difference is purely in the database attached to the VL servers. The current code will malfunction if it sees a server in two cells as it assumes global address -> server record mappings and that each server is in just one cell. Further, each server may have multiple addresses - and may have addresses of different families (IPv4 and IPv6, say). To this end, the following structural changes are made: (1) Server record management is overhauled: (a) Server records are made independent of cell. The namespace keeps track of them, volume records have lists of them and each vnode has a server on which its callback interest currently resides. (b) The cell record no longer keeps a list of servers known to be in that cell. (c) The server records are now kept in a flat list because there's no single address to sort on. (d) Server records are now keyed by their UUID within the namespace. (e) The addresses for a server are obtained with the VL.GetAddrsU rather than with VL.GetEntryByName, using the server's UUID as a parameter. (f) Cached server records are garbage collected after a period of non-use and are counted out of existence before purging is allowed to complete. This protects the work functions against rmmod. (g) The servers list is now in /proc/fs/afs/servers. (2) Volume record management is overhauled: (a) An RCU-replaceable server list is introduced. This tracks both servers and their coresponding callback interests. (b) The superblock is now keyed on cell record and numeric volume ID. (c) The volume record is now tied to the superblock which mounts it, and is activated when mounted and deactivated when unmounted. This makes it easier to handle the cache cookie without causing a double-use in fscache. (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU to get the server UUID list. (e) The volume name is updated if it is seen to have changed when the volume is updated (the update is keyed on the volume ID). (3) The vlocation record is got rid of and VLDB records are no longer cached. Sufficient information is stored in the volume record, though an update to a volume record is now no longer shared between related volumes (volumes come in bundles of three: R/W, R/O and backup). and the following procedural changes are made: (1) The fileserver cursor introduced previously is now fleshed out and used to iterate over fileservers and their addresses. (2) Volume status is checked during iteration, and the server list is replaced if a change is detected. (3) Server status is checked during iteration, and the address list is replaced if a change is detected. (4) The abort code is saved into the address list cursor and -ECONNABORTED returned in afs_make_call() if a remote abort happened rather than translating the abort into an error message. This allows actions to be taken depending on the abort code more easily. (a) If a VMOVED abort is seen then this is handled by rechecking the volume and restarting the iteration. (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is handled by sleeping for a short period and retrying and/or trying other servers that might serve that volume. A message is also displayed once until the condition has cleared. (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the moment. (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to see if it has been deleted; if not, the fileserver is probably indicating that the volume couldn't be attached and needs salvaging. (e) If statfs() sees one of these aborts, it does not sleep, but rather returns an error, so as not to block the umount program. (5) The fileserver iteration functions in vnode.c are now merged into their callers and more heavily macroised around the cursor. vnode.c is removed. (6) Operations on a particular vnode are serialised on that vnode because the server will lock that vnode whilst it operates on it, so a second op sent will just have to wait. (7) Fileservers are probed with FS.GetCapabilities before being used. This is where service upgrade will be done. (8) A callback interest on a fileserver is set up before an FS operation is performed and passed through to afs_make_call() so that it can be set on the vnode if the operation returns a callback. The callback interest is passed through to afs_iget() also so that it can be set there too. In general, record updating is done on an as-needed basis when we try to access servers, volumes or vnodes rather than offloading it to work items and special threads. Notes: (1) Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998). (2) VBUSY is retried forever for the moment at intervals of 1s. (3) /proc/fs/afs/<cell>/servers no longer exists. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:50 +08:00
/*
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
* Find a key to use for the writeback. We cached the keys used to author the
* writes on the vnode. *_wbk will contain the last writeback key used or NULL
* and we need to start from there if it's set.
afs: Overhaul volume and server record caching and fileserver rotation The current code assumes that volumes and servers are per-cell and are never shared, but this is not enforced, and, indeed, public cells do exist that are aliases of each other. Further, an organisation can, say, set up a public cell and a private cell with overlapping, but not identical, sets of servers. The difference is purely in the database attached to the VL servers. The current code will malfunction if it sees a server in two cells as it assumes global address -> server record mappings and that each server is in just one cell. Further, each server may have multiple addresses - and may have addresses of different families (IPv4 and IPv6, say). To this end, the following structural changes are made: (1) Server record management is overhauled: (a) Server records are made independent of cell. The namespace keeps track of them, volume records have lists of them and each vnode has a server on which its callback interest currently resides. (b) The cell record no longer keeps a list of servers known to be in that cell. (c) The server records are now kept in a flat list because there's no single address to sort on. (d) Server records are now keyed by their UUID within the namespace. (e) The addresses for a server are obtained with the VL.GetAddrsU rather than with VL.GetEntryByName, using the server's UUID as a parameter. (f) Cached server records are garbage collected after a period of non-use and are counted out of existence before purging is allowed to complete. This protects the work functions against rmmod. (g) The servers list is now in /proc/fs/afs/servers. (2) Volume record management is overhauled: (a) An RCU-replaceable server list is introduced. This tracks both servers and their coresponding callback interests. (b) The superblock is now keyed on cell record and numeric volume ID. (c) The volume record is now tied to the superblock which mounts it, and is activated when mounted and deactivated when unmounted. This makes it easier to handle the cache cookie without causing a double-use in fscache. (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU to get the server UUID list. (e) The volume name is updated if it is seen to have changed when the volume is updated (the update is keyed on the volume ID). (3) The vlocation record is got rid of and VLDB records are no longer cached. Sufficient information is stored in the volume record, though an update to a volume record is now no longer shared between related volumes (volumes come in bundles of three: R/W, R/O and backup). and the following procedural changes are made: (1) The fileserver cursor introduced previously is now fleshed out and used to iterate over fileservers and their addresses. (2) Volume status is checked during iteration, and the server list is replaced if a change is detected. (3) Server status is checked during iteration, and the address list is replaced if a change is detected. (4) The abort code is saved into the address list cursor and -ECONNABORTED returned in afs_make_call() if a remote abort happened rather than translating the abort into an error message. This allows actions to be taken depending on the abort code more easily. (a) If a VMOVED abort is seen then this is handled by rechecking the volume and restarting the iteration. (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is handled by sleeping for a short period and retrying and/or trying other servers that might serve that volume. A message is also displayed once until the condition has cleared. (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the moment. (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to see if it has been deleted; if not, the fileserver is probably indicating that the volume couldn't be attached and needs salvaging. (e) If statfs() sees one of these aborts, it does not sleep, but rather returns an error, so as not to block the umount program. (5) The fileserver iteration functions in vnode.c are now merged into their callers and more heavily macroised around the cursor. vnode.c is removed. (6) Operations on a particular vnode are serialised on that vnode because the server will lock that vnode whilst it operates on it, so a second op sent will just have to wait. (7) Fileservers are probed with FS.GetCapabilities before being used. This is where service upgrade will be done. (8) A callback interest on a fileserver is set up before an FS operation is performed and passed through to afs_make_call() so that it can be set on the vnode if the operation returns a callback. The callback interest is passed through to afs_iget() also so that it can be set there too. In general, record updating is done on an as-needed basis when we try to access servers, volumes or vnodes rather than offloading it to work items and special threads. Notes: (1) Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998). (2) VBUSY is retried forever for the moment at intervals of 1s. (3) /proc/fs/afs/<cell>/servers no longer exists. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:50 +08:00
*/
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
static int afs_get_writeback_key(struct afs_vnode *vnode,
struct afs_wb_key **_wbk)
afs: Overhaul volume and server record caching and fileserver rotation The current code assumes that volumes and servers are per-cell and are never shared, but this is not enforced, and, indeed, public cells do exist that are aliases of each other. Further, an organisation can, say, set up a public cell and a private cell with overlapping, but not identical, sets of servers. The difference is purely in the database attached to the VL servers. The current code will malfunction if it sees a server in two cells as it assumes global address -> server record mappings and that each server is in just one cell. Further, each server may have multiple addresses - and may have addresses of different families (IPv4 and IPv6, say). To this end, the following structural changes are made: (1) Server record management is overhauled: (a) Server records are made independent of cell. The namespace keeps track of them, volume records have lists of them and each vnode has a server on which its callback interest currently resides. (b) The cell record no longer keeps a list of servers known to be in that cell. (c) The server records are now kept in a flat list because there's no single address to sort on. (d) Server records are now keyed by their UUID within the namespace. (e) The addresses for a server are obtained with the VL.GetAddrsU rather than with VL.GetEntryByName, using the server's UUID as a parameter. (f) Cached server records are garbage collected after a period of non-use and are counted out of existence before purging is allowed to complete. This protects the work functions against rmmod. (g) The servers list is now in /proc/fs/afs/servers. (2) Volume record management is overhauled: (a) An RCU-replaceable server list is introduced. This tracks both servers and their coresponding callback interests. (b) The superblock is now keyed on cell record and numeric volume ID. (c) The volume record is now tied to the superblock which mounts it, and is activated when mounted and deactivated when unmounted. This makes it easier to handle the cache cookie without causing a double-use in fscache. (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU to get the server UUID list. (e) The volume name is updated if it is seen to have changed when the volume is updated (the update is keyed on the volume ID). (3) The vlocation record is got rid of and VLDB records are no longer cached. Sufficient information is stored in the volume record, though an update to a volume record is now no longer shared between related volumes (volumes come in bundles of three: R/W, R/O and backup). and the following procedural changes are made: (1) The fileserver cursor introduced previously is now fleshed out and used to iterate over fileservers and their addresses. (2) Volume status is checked during iteration, and the server list is replaced if a change is detected. (3) Server status is checked during iteration, and the address list is replaced if a change is detected. (4) The abort code is saved into the address list cursor and -ECONNABORTED returned in afs_make_call() if a remote abort happened rather than translating the abort into an error message. This allows actions to be taken depending on the abort code more easily. (a) If a VMOVED abort is seen then this is handled by rechecking the volume and restarting the iteration. (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is handled by sleeping for a short period and retrying and/or trying other servers that might serve that volume. A message is also displayed once until the condition has cleared. (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the moment. (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to see if it has been deleted; if not, the fileserver is probably indicating that the volume couldn't be attached and needs salvaging. (e) If statfs() sees one of these aborts, it does not sleep, but rather returns an error, so as not to block the umount program. (5) The fileserver iteration functions in vnode.c are now merged into their callers and more heavily macroised around the cursor. vnode.c is removed. (6) Operations on a particular vnode are serialised on that vnode because the server will lock that vnode whilst it operates on it, so a second op sent will just have to wait. (7) Fileservers are probed with FS.GetCapabilities before being used. This is where service upgrade will be done. (8) A callback interest on a fileserver is set up before an FS operation is performed and passed through to afs_make_call() so that it can be set on the vnode if the operation returns a callback. The callback interest is passed through to afs_iget() also so that it can be set there too. In general, record updating is done on an as-needed basis when we try to access servers, volumes or vnodes rather than offloading it to work items and special threads. Notes: (1) Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998). (2) VBUSY is retried forever for the moment at intervals of 1s. (3) /proc/fs/afs/<cell>/servers no longer exists. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:50 +08:00
{
struct afs_wb_key *wbk = NULL;
struct list_head *p;
int ret = -ENOKEY, ret2;
afs: Overhaul volume and server record caching and fileserver rotation The current code assumes that volumes and servers are per-cell and are never shared, but this is not enforced, and, indeed, public cells do exist that are aliases of each other. Further, an organisation can, say, set up a public cell and a private cell with overlapping, but not identical, sets of servers. The difference is purely in the database attached to the VL servers. The current code will malfunction if it sees a server in two cells as it assumes global address -> server record mappings and that each server is in just one cell. Further, each server may have multiple addresses - and may have addresses of different families (IPv4 and IPv6, say). To this end, the following structural changes are made: (1) Server record management is overhauled: (a) Server records are made independent of cell. The namespace keeps track of them, volume records have lists of them and each vnode has a server on which its callback interest currently resides. (b) The cell record no longer keeps a list of servers known to be in that cell. (c) The server records are now kept in a flat list because there's no single address to sort on. (d) Server records are now keyed by their UUID within the namespace. (e) The addresses for a server are obtained with the VL.GetAddrsU rather than with VL.GetEntryByName, using the server's UUID as a parameter. (f) Cached server records are garbage collected after a period of non-use and are counted out of existence before purging is allowed to complete. This protects the work functions against rmmod. (g) The servers list is now in /proc/fs/afs/servers. (2) Volume record management is overhauled: (a) An RCU-replaceable server list is introduced. This tracks both servers and their coresponding callback interests. (b) The superblock is now keyed on cell record and numeric volume ID. (c) The volume record is now tied to the superblock which mounts it, and is activated when mounted and deactivated when unmounted. This makes it easier to handle the cache cookie without causing a double-use in fscache. (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU to get the server UUID list. (e) The volume name is updated if it is seen to have changed when the volume is updated (the update is keyed on the volume ID). (3) The vlocation record is got rid of and VLDB records are no longer cached. Sufficient information is stored in the volume record, though an update to a volume record is now no longer shared between related volumes (volumes come in bundles of three: R/W, R/O and backup). and the following procedural changes are made: (1) The fileserver cursor introduced previously is now fleshed out and used to iterate over fileservers and their addresses. (2) Volume status is checked during iteration, and the server list is replaced if a change is detected. (3) Server status is checked during iteration, and the address list is replaced if a change is detected. (4) The abort code is saved into the address list cursor and -ECONNABORTED returned in afs_make_call() if a remote abort happened rather than translating the abort into an error message. This allows actions to be taken depending on the abort code more easily. (a) If a VMOVED abort is seen then this is handled by rechecking the volume and restarting the iteration. (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is handled by sleeping for a short period and retrying and/or trying other servers that might serve that volume. A message is also displayed once until the condition has cleared. (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the moment. (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to see if it has been deleted; if not, the fileserver is probably indicating that the volume couldn't be attached and needs salvaging. (e) If statfs() sees one of these aborts, it does not sleep, but rather returns an error, so as not to block the umount program. (5) The fileserver iteration functions in vnode.c are now merged into their callers and more heavily macroised around the cursor. vnode.c is removed. (6) Operations on a particular vnode are serialised on that vnode because the server will lock that vnode whilst it operates on it, so a second op sent will just have to wait. (7) Fileservers are probed with FS.GetCapabilities before being used. This is where service upgrade will be done. (8) A callback interest on a fileserver is set up before an FS operation is performed and passed through to afs_make_call() so that it can be set on the vnode if the operation returns a callback. The callback interest is passed through to afs_iget() also so that it can be set there too. In general, record updating is done on an as-needed basis when we try to access servers, volumes or vnodes rather than offloading it to work items and special threads. Notes: (1) Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998). (2) VBUSY is retried forever for the moment at intervals of 1s. (3) /proc/fs/afs/<cell>/servers no longer exists. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:50 +08:00
spin_lock(&vnode->wb_lock);
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
if (*_wbk)
p = (*_wbk)->vnode_link.next;
else
p = vnode->wb_keys.next;
while (p != &vnode->wb_keys) {
wbk = list_entry(p, struct afs_wb_key, vnode_link);
_debug("wbk %u", key_serial(wbk->key));
ret2 = key_validate(wbk->key);
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
if (ret2 == 0) {
refcount_inc(&wbk->usage);
_debug("USE WB KEY %u", key_serial(wbk->key));
break;
}
wbk = NULL;
if (ret == -ENOKEY)
ret = ret2;
p = p->next;
}
spin_unlock(&vnode->wb_lock);
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
if (*_wbk)
afs_put_wb_key(*_wbk);
*_wbk = wbk;
return 0;
}
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
static void afs_store_data_success(struct afs_operation *op)
{
struct afs_vnode *vnode = op->file[0].vnode;
op->ctime = op->file[0].scb.status.mtime_client;
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
afs_vnode_commit_status(op, &op->file[0]);
if (op->error == 0) {
if (!op->store.laundering)
afs_pages_written_back(vnode, op->store.first, op->store.last);
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
afs_stat_v(vnode, n_stores);
atomic_long_add((op->store.last * PAGE_SIZE + op->store.last_to) -
(op->store.first * PAGE_SIZE + op->store.first_offset),
&afs_v2net(vnode)->n_store_bytes);
}
}
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
static const struct afs_operation_ops afs_store_data_operation = {
.issue_afs_rpc = afs_fs_store_data,
.issue_yfs_rpc = yfs_fs_store_data,
.success = afs_store_data_success,
};
afs: Fix application of status and callback to be under same lock When applying the status and callback in the response of an operation, apply them in the same critical section so that there's no race between checking the callback state and checking status-dependent state (such as the data version). Fix this by: (1) Allocating a joint {status,callback} record (afs_status_cb) before calling the RPC function for each vnode for which the RPC reply contains a status or a status plus a callback. A flag is set in the record to indicate if a callback was actually received. (2) These records are passed into the RPC functions to be filled in. The afs_decode_status() and yfs_decode_status() functions are removed and the cb_lock is no longer taken. (3) xdr_decode_AFSFetchStatus() and xdr_decode_YFSFetchStatus() no longer update the vnode. (4) xdr_decode_AFSCallBack() and xdr_decode_YFSCallBack() no longer update the vnode. (5) vnodes, expected data-version numbers and callback break counters (cb_break) no longer need to be passed to the reply delivery functions. Note that, for the moment, the file locking functions still need access to both the call and the vnode at the same time. (6) afs_vnode_commit_status() is now given the cb_break value and the expected data_version and the task of applying the status and the callback to the vnode are now done here. This is done under a single taking of vnode->cb_lock. (7) afs_pages_written_back() is now called by afs_store_data() rather than by the reply delivery function. afs_pages_written_back() has been moved to before the call point and is now given the first and last page numbers rather than a pointer to the call. (8) The indicator from YFS.RemoveFile2 as to whether the target file actually got removed (status.abort_code == VNOVNODE) rather than merely dropping a link is now checked in afs_unlink rather than in xdr_decode_YFSFetchStatus(). Supplementary fixes: (*) afs_cache_permit() now gets the caller_access mask from the afs_status_cb object rather than picking it out of the vnode's status record. afs_fetch_status() returns caller_access through its argument list for this purpose also. (*) afs_inode_init_from_status() now uses a write lock on cb_lock rather than a read lock and now sets the callback inside the same critical section. Fixes: c435ee34551e ("afs: Overhaul the callback handling") Signed-off-by: David Howells <dhowells@redhat.com>
2019-05-09 22:16:10 +08:00
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
/*
* write to a file
*/
static int afs_store_data(struct address_space *mapping,
pgoff_t first, pgoff_t last,
unsigned offset, unsigned to, bool laundering)
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
{
struct afs_vnode *vnode = AFS_FS_I(mapping->host);
struct afs_operation *op;
struct afs_wb_key *wbk = NULL;
int ret;
_enter("%s{%llx:%llu.%u},%lx,%lx,%x,%x",
vnode->volume->name,
vnode->fid.vid,
vnode->fid.vnode,
vnode->fid.unique,
first, last, offset, to);
afs: Overhaul volume and server record caching and fileserver rotation The current code assumes that volumes and servers are per-cell and are never shared, but this is not enforced, and, indeed, public cells do exist that are aliases of each other. Further, an organisation can, say, set up a public cell and a private cell with overlapping, but not identical, sets of servers. The difference is purely in the database attached to the VL servers. The current code will malfunction if it sees a server in two cells as it assumes global address -> server record mappings and that each server is in just one cell. Further, each server may have multiple addresses - and may have addresses of different families (IPv4 and IPv6, say). To this end, the following structural changes are made: (1) Server record management is overhauled: (a) Server records are made independent of cell. The namespace keeps track of them, volume records have lists of them and each vnode has a server on which its callback interest currently resides. (b) The cell record no longer keeps a list of servers known to be in that cell. (c) The server records are now kept in a flat list because there's no single address to sort on. (d) Server records are now keyed by their UUID within the namespace. (e) The addresses for a server are obtained with the VL.GetAddrsU rather than with VL.GetEntryByName, using the server's UUID as a parameter. (f) Cached server records are garbage collected after a period of non-use and are counted out of existence before purging is allowed to complete. This protects the work functions against rmmod. (g) The servers list is now in /proc/fs/afs/servers. (2) Volume record management is overhauled: (a) An RCU-replaceable server list is introduced. This tracks both servers and their coresponding callback interests. (b) The superblock is now keyed on cell record and numeric volume ID. (c) The volume record is now tied to the superblock which mounts it, and is activated when mounted and deactivated when unmounted. This makes it easier to handle the cache cookie without causing a double-use in fscache. (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU to get the server UUID list. (e) The volume name is updated if it is seen to have changed when the volume is updated (the update is keyed on the volume ID). (3) The vlocation record is got rid of and VLDB records are no longer cached. Sufficient information is stored in the volume record, though an update to a volume record is now no longer shared between related volumes (volumes come in bundles of three: R/W, R/O and backup). and the following procedural changes are made: (1) The fileserver cursor introduced previously is now fleshed out and used to iterate over fileservers and their addresses. (2) Volume status is checked during iteration, and the server list is replaced if a change is detected. (3) Server status is checked during iteration, and the address list is replaced if a change is detected. (4) The abort code is saved into the address list cursor and -ECONNABORTED returned in afs_make_call() if a remote abort happened rather than translating the abort into an error message. This allows actions to be taken depending on the abort code more easily. (a) If a VMOVED abort is seen then this is handled by rechecking the volume and restarting the iteration. (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is handled by sleeping for a short period and retrying and/or trying other servers that might serve that volume. A message is also displayed once until the condition has cleared. (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the moment. (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to see if it has been deleted; if not, the fileserver is probably indicating that the volume couldn't be attached and needs salvaging. (e) If statfs() sees one of these aborts, it does not sleep, but rather returns an error, so as not to block the umount program. (5) The fileserver iteration functions in vnode.c are now merged into their callers and more heavily macroised around the cursor. vnode.c is removed. (6) Operations on a particular vnode are serialised on that vnode because the server will lock that vnode whilst it operates on it, so a second op sent will just have to wait. (7) Fileservers are probed with FS.GetCapabilities before being used. This is where service upgrade will be done. (8) A callback interest on a fileserver is set up before an FS operation is performed and passed through to afs_make_call() so that it can be set on the vnode if the operation returns a callback. The callback interest is passed through to afs_iget() also so that it can be set there too. In general, record updating is done on an as-needed basis when we try to access servers, volumes or vnodes rather than offloading it to work items and special threads. Notes: (1) Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998). (2) VBUSY is retried forever for the moment at intervals of 1s. (3) /proc/fs/afs/<cell>/servers no longer exists. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:50 +08:00
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
ret = afs_get_writeback_key(vnode, &wbk);
if (ret) {
_leave(" = %d [no keys]", ret);
return ret;
afs: Overhaul volume and server record caching and fileserver rotation The current code assumes that volumes and servers are per-cell and are never shared, but this is not enforced, and, indeed, public cells do exist that are aliases of each other. Further, an organisation can, say, set up a public cell and a private cell with overlapping, but not identical, sets of servers. The difference is purely in the database attached to the VL servers. The current code will malfunction if it sees a server in two cells as it assumes global address -> server record mappings and that each server is in just one cell. Further, each server may have multiple addresses - and may have addresses of different families (IPv4 and IPv6, say). To this end, the following structural changes are made: (1) Server record management is overhauled: (a) Server records are made independent of cell. The namespace keeps track of them, volume records have lists of them and each vnode has a server on which its callback interest currently resides. (b) The cell record no longer keeps a list of servers known to be in that cell. (c) The server records are now kept in a flat list because there's no single address to sort on. (d) Server records are now keyed by their UUID within the namespace. (e) The addresses for a server are obtained with the VL.GetAddrsU rather than with VL.GetEntryByName, using the server's UUID as a parameter. (f) Cached server records are garbage collected after a period of non-use and are counted out of existence before purging is allowed to complete. This protects the work functions against rmmod. (g) The servers list is now in /proc/fs/afs/servers. (2) Volume record management is overhauled: (a) An RCU-replaceable server list is introduced. This tracks both servers and their coresponding callback interests. (b) The superblock is now keyed on cell record and numeric volume ID. (c) The volume record is now tied to the superblock which mounts it, and is activated when mounted and deactivated when unmounted. This makes it easier to handle the cache cookie without causing a double-use in fscache. (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU to get the server UUID list. (e) The volume name is updated if it is seen to have changed when the volume is updated (the update is keyed on the volume ID). (3) The vlocation record is got rid of and VLDB records are no longer cached. Sufficient information is stored in the volume record, though an update to a volume record is now no longer shared between related volumes (volumes come in bundles of three: R/W, R/O and backup). and the following procedural changes are made: (1) The fileserver cursor introduced previously is now fleshed out and used to iterate over fileservers and their addresses. (2) Volume status is checked during iteration, and the server list is replaced if a change is detected. (3) Server status is checked during iteration, and the address list is replaced if a change is detected. (4) The abort code is saved into the address list cursor and -ECONNABORTED returned in afs_make_call() if a remote abort happened rather than translating the abort into an error message. This allows actions to be taken depending on the abort code more easily. (a) If a VMOVED abort is seen then this is handled by rechecking the volume and restarting the iteration. (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is handled by sleeping for a short period and retrying and/or trying other servers that might serve that volume. A message is also displayed once until the condition has cleared. (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the moment. (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to see if it has been deleted; if not, the fileserver is probably indicating that the volume couldn't be attached and needs salvaging. (e) If statfs() sees one of these aborts, it does not sleep, but rather returns an error, so as not to block the umount program. (5) The fileserver iteration functions in vnode.c are now merged into their callers and more heavily macroised around the cursor. vnode.c is removed. (6) Operations on a particular vnode are serialised on that vnode because the server will lock that vnode whilst it operates on it, so a second op sent will just have to wait. (7) Fileservers are probed with FS.GetCapabilities before being used. This is where service upgrade will be done. (8) A callback interest on a fileserver is set up before an FS operation is performed and passed through to afs_make_call() so that it can be set on the vnode if the operation returns a callback. The callback interest is passed through to afs_iget() also so that it can be set there too. In general, record updating is done on an as-needed basis when we try to access servers, volumes or vnodes rather than offloading it to work items and special threads. Notes: (1) Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998). (2) VBUSY is retried forever for the moment at intervals of 1s. (3) /proc/fs/afs/<cell>/servers no longer exists. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:50 +08:00
}
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
op = afs_alloc_operation(wbk->key, vnode->volume);
if (IS_ERR(op)) {
afs_put_wb_key(wbk);
return -ENOMEM;
}
afs_op_set_vnode(op, 0, vnode);
op->file[0].dv_delta = 1;
op->store.mapping = mapping;
op->store.first = first;
op->store.last = last;
op->store.first_offset = offset;
op->store.last_to = to;
op->store.laundering = laundering;
op->mtime = vnode->vfs_inode.i_mtime;
afs: Fix interruption of operations The afs filesystem driver allows unstarted operations to be cancelled by signal, but most of these can easily be restarted (mkdir for example). The primary culprits for reproducing this are those applications that use SIGALRM to display a progress counter. File lock-extension operation is marked uninterruptible as we have a limited time in which to do it, and the release op is marked uninterruptible also as if we fail to unlock a file, we'll have to wait 20 mins before anyone can lock it again. The store operation logs a warning if it gets interruption, e.g.: kAFS: Unexpected error from FS.StoreData -4 because it's run from the background - but it can also be run from fdatasync()-type things. However, store options aren't marked interruptible at the moment. Fix this in the following ways: (1) Mark store operations as uninterruptible. It might make sense to relax this for certain situations, but I'm not sure how to make sure that background store ops aren't affected by signals to foreground processes that happen to trigger them. (2) In afs_get_io_locks(), where we're getting the serialisation lock for talking to the fileserver, return ERESTARTSYS rather than EINTR because a lot of the operations (e.g. mkdir) are restartable if we haven't yet started sending the op to the server. Fixes: e49c7b2f6de7 ("afs: Build an abstraction around an "operation" concept") Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-07-08 16:27:07 +08:00
op->flags |= AFS_OPERATION_UNINTR;
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
op->ops = &afs_store_data_operation;
try_next_key:
afs_begin_vnode_operation(op);
afs_wait_for_operation(op);
switch (op->error) {
case -EACCES:
case -EPERM:
case -ENOKEY:
case -EKEYEXPIRED:
case -EKEYREJECTED:
case -EKEYREVOKED:
_debug("next");
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
ret = afs_get_writeback_key(vnode, &wbk);
if (ret == 0) {
key_put(op->key);
op->key = key_get(wbk->key);
goto try_next_key;
}
break;
}
afs_put_wb_key(wbk);
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
_leave(" = %d", op->error);
return afs_put_operation(op);
afs: Overhaul volume and server record caching and fileserver rotation The current code assumes that volumes and servers are per-cell and are never shared, but this is not enforced, and, indeed, public cells do exist that are aliases of each other. Further, an organisation can, say, set up a public cell and a private cell with overlapping, but not identical, sets of servers. The difference is purely in the database attached to the VL servers. The current code will malfunction if it sees a server in two cells as it assumes global address -> server record mappings and that each server is in just one cell. Further, each server may have multiple addresses - and may have addresses of different families (IPv4 and IPv6, say). To this end, the following structural changes are made: (1) Server record management is overhauled: (a) Server records are made independent of cell. The namespace keeps track of them, volume records have lists of them and each vnode has a server on which its callback interest currently resides. (b) The cell record no longer keeps a list of servers known to be in that cell. (c) The server records are now kept in a flat list because there's no single address to sort on. (d) Server records are now keyed by their UUID within the namespace. (e) The addresses for a server are obtained with the VL.GetAddrsU rather than with VL.GetEntryByName, using the server's UUID as a parameter. (f) Cached server records are garbage collected after a period of non-use and are counted out of existence before purging is allowed to complete. This protects the work functions against rmmod. (g) The servers list is now in /proc/fs/afs/servers. (2) Volume record management is overhauled: (a) An RCU-replaceable server list is introduced. This tracks both servers and their coresponding callback interests. (b) The superblock is now keyed on cell record and numeric volume ID. (c) The volume record is now tied to the superblock which mounts it, and is activated when mounted and deactivated when unmounted. This makes it easier to handle the cache cookie without causing a double-use in fscache. (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU to get the server UUID list. (e) The volume name is updated if it is seen to have changed when the volume is updated (the update is keyed on the volume ID). (3) The vlocation record is got rid of and VLDB records are no longer cached. Sufficient information is stored in the volume record, though an update to a volume record is now no longer shared between related volumes (volumes come in bundles of three: R/W, R/O and backup). and the following procedural changes are made: (1) The fileserver cursor introduced previously is now fleshed out and used to iterate over fileservers and their addresses. (2) Volume status is checked during iteration, and the server list is replaced if a change is detected. (3) Server status is checked during iteration, and the address list is replaced if a change is detected. (4) The abort code is saved into the address list cursor and -ECONNABORTED returned in afs_make_call() if a remote abort happened rather than translating the abort into an error message. This allows actions to be taken depending on the abort code more easily. (a) If a VMOVED abort is seen then this is handled by rechecking the volume and restarting the iteration. (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is handled by sleeping for a short period and retrying and/or trying other servers that might serve that volume. A message is also displayed once until the condition has cleared. (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the moment. (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to see if it has been deleted; if not, the fileserver is probably indicating that the volume couldn't be attached and needs salvaging. (e) If statfs() sees one of these aborts, it does not sleep, but rather returns an error, so as not to block the umount program. (5) The fileserver iteration functions in vnode.c are now merged into their callers and more heavily macroised around the cursor. vnode.c is removed. (6) Operations on a particular vnode are serialised on that vnode because the server will lock that vnode whilst it operates on it, so a second op sent will just have to wait. (7) Fileservers are probed with FS.GetCapabilities before being used. This is where service upgrade will be done. (8) A callback interest on a fileserver is set up before an FS operation is performed and passed through to afs_make_call() so that it can be set on the vnode if the operation returns a callback. The callback interest is passed through to afs_iget() also so that it can be set there too. In general, record updating is done on an as-needed basis when we try to access servers, volumes or vnodes rather than offloading it to work items and special threads. Notes: (1) Pre AFS-3.4 servers are no longer supported, though this can be added back if necessary (AFS-3.4 was released in 1998). (2) VBUSY is retried forever for the moment at intervals of 1s. (3) /proc/fs/afs/<cell>/servers no longer exists. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:50 +08:00
}
/*
* Synchronously write back the locked page and any subsequent non-locked dirty
* pages.
*/
static int afs_write_back_from_locked_page(struct address_space *mapping,
struct writeback_control *wbc,
struct page *primary_page,
pgoff_t final_page)
{
struct afs_vnode *vnode = AFS_FS_I(mapping->host);
struct page *pages[8], *page;
unsigned long count, priv;
unsigned n, offset, to, f, t;
pgoff_t start, first, last;
loff_t i_size, end;
int loop, ret;
_enter(",%lx", primary_page->index);
count = 1;
if (test_set_page_writeback(primary_page))
BUG();
/* Find all consecutive lockable dirty pages that have contiguous
* written regions, stopping when we find a page that is not
* immediately lockable, is not dirty or is missing, or we reach the
* end of the range.
*/
start = primary_page->index;
priv = page_private(primary_page);
offset = afs_page_dirty_from(priv);
to = afs_page_dirty_to(priv);
trace_afs_page_dirty(vnode, tracepoint_string("store"),
primary_page->index, priv);
WARN_ON(offset == to);
if (offset == to)
trace_afs_page_dirty(vnode, tracepoint_string("WARN"),
primary_page->index, priv);
if (start >= final_page ||
(to < PAGE_SIZE && !test_bit(AFS_VNODE_NEW_CONTENT, &vnode->flags)))
goto no_more;
start++;
do {
_debug("more %lx [%lx]", start, count);
n = final_page - start + 1;
if (n > ARRAY_SIZE(pages))
n = ARRAY_SIZE(pages);
n = find_get_pages_contig(mapping, start, ARRAY_SIZE(pages), pages);
_debug("fgpc %u", n);
if (n == 0)
goto no_more;
if (pages[0]->index != start) {
do {
put_page(pages[--n]);
} while (n > 0);
goto no_more;
}
for (loop = 0; loop < n; loop++) {
page = pages[loop];
if (to != PAGE_SIZE &&
!test_bit(AFS_VNODE_NEW_CONTENT, &vnode->flags))
break;
if (page->index > final_page)
break;
if (!trylock_page(page))
break;
if (!PageDirty(page) || PageWriteback(page)) {
unlock_page(page);
break;
}
priv = page_private(page);
f = afs_page_dirty_from(priv);
t = afs_page_dirty_to(priv);
if (f != 0 &&
!test_bit(AFS_VNODE_NEW_CONTENT, &vnode->flags)) {
unlock_page(page);
break;
}
to = t;
trace_afs_page_dirty(vnode, tracepoint_string("store+"),
page->index, priv);
if (!clear_page_dirty_for_io(page))
BUG();
if (test_set_page_writeback(page))
BUG();
unlock_page(page);
put_page(page);
}
count += loop;
if (loop < n) {
for (; loop < n; loop++)
put_page(pages[loop]);
goto no_more;
}
start += loop;
} while (start <= final_page && count < 65536);
no_more:
/* We now have a contiguous set of dirty pages, each with writeback
* set; the first page is still locked at this point, but all the rest
* have been unlocked.
*/
unlock_page(primary_page);
first = primary_page->index;
last = first + count - 1;
end = (loff_t)last * PAGE_SIZE + to;
i_size = i_size_read(&vnode->vfs_inode);
_debug("write back %lx[%u..] to %lx[..%u]", first, offset, last, to);
if (end > i_size)
to = i_size & ~PAGE_MASK;
ret = afs_store_data(mapping, first, last, offset, to, false);
switch (ret) {
case 0:
ret = count;
break;
default:
pr_notice("kAFS: Unexpected error from FS.StoreData %d\n", ret);
fallthrough;
case -EACCES:
case -EPERM:
case -ENOKEY:
case -EKEYEXPIRED:
case -EKEYREJECTED:
case -EKEYREVOKED:
afs_redirty_pages(wbc, mapping, first, last);
mapping_set_error(mapping, ret);
break;
case -EDQUOT:
case -ENOSPC:
afs_redirty_pages(wbc, mapping, first, last);
mapping_set_error(mapping, -ENOSPC);
break;
case -EROFS:
case -EIO:
case -EREMOTEIO:
case -EFBIG:
case -ENOENT:
case -ENOMEDIUM:
case -ENXIO:
trace_afs_file_error(vnode, ret, afs_file_error_writeback_fail);
afs_kill_pages(mapping, first, last);
mapping_set_error(mapping, ret);
break;
}
_leave(" = %d", ret);
return ret;
}
/*
* write a page back to the server
* - the caller locked the page for us
*/
int afs_writepage(struct page *page, struct writeback_control *wbc)
{
int ret;
_enter("{%lx},", page->index);
ret = afs_write_back_from_locked_page(page->mapping, wbc, page,
wbc->range_end >> PAGE_SHIFT);
if (ret < 0) {
_leave(" = %d", ret);
return 0;
}
wbc->nr_to_write -= ret;
_leave(" = 0");
return 0;
}
/*
* write a region of pages back to the server
*/
static int afs_writepages_region(struct address_space *mapping,
struct writeback_control *wbc,
pgoff_t index, pgoff_t end, pgoff_t *_next)
{
struct page *page;
int ret, n;
_enter(",,%lx,%lx,", index, end);
do {
n = find_get_pages_range_tag(mapping, &index, end,
PAGECACHE_TAG_DIRTY, 1, &page);
if (!n)
break;
_debug("wback %lx", page->index);
/*
* at this point we hold neither the i_pages lock nor the
* page lock: the page may be truncated or invalidated
* (changing page->mapping to NULL), or even swizzled
* back from swapper_space to tmpfs file mapping
*/
ret = lock_page_killable(page);
if (ret < 0) {
put_page(page);
_leave(" = %d", ret);
return ret;
}
if (page->mapping != mapping || !PageDirty(page)) {
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);
continue;
}
if (PageWriteback(page)) {
unlock_page(page);
if (wbc->sync_mode != WB_SYNC_NONE)
wait_on_page_writeback(page);
put_page(page);
continue;
}
if (!clear_page_dirty_for_io(page))
BUG();
ret = afs_write_back_from_locked_page(mapping, wbc, page, end);
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);
if (ret < 0) {
_leave(" = %d", ret);
return ret;
}
wbc->nr_to_write -= ret;
cond_resched();
} while (index < end && wbc->nr_to_write > 0);
*_next = index;
_leave(" = 0 [%lx]", *_next);
return 0;
}
/*
* write some of the pending data back to the server
*/
int afs_writepages(struct address_space *mapping,
struct writeback_control *wbc)
{
afs: Fix deadlock between writeback and truncate The afs filesystem has a lock[*] that it uses to serialise I/O operations going to the server (vnode->io_lock), as the server will only perform one modification operation at a time on any given file or directory. This prevents the the filesystem from filling up all the call slots to a server with calls that aren't going to be executed in parallel anyway, thereby allowing operations on other files to obtain slots. [*] Note that is probably redundant for directories at least since i_rwsem is used to serialise directory modifications and lookup/reading vs modification. The server does allow parallel non-modification ops, however. When a file truncation op completes, we truncate the in-memory copy of the file to match - but we do it whilst still holding the io_lock, the idea being to prevent races with other operations. However, if writeback starts in a worker thread simultaneously with truncation (whilst notify_change() is called with i_rwsem locked, writeback pays it no heed), it may manage to set PG_writeback bits on the pages that will get truncated before afs_setattr_success() manages to call truncate_pagecache(). Truncate will then wait for those pages - whilst still inside io_lock: # cat /proc/8837/stack [<0>] wait_on_page_bit_common+0x184/0x1e7 [<0>] truncate_inode_pages_range+0x37f/0x3eb [<0>] truncate_pagecache+0x3c/0x53 [<0>] afs_setattr_success+0x4d/0x6e [<0>] afs_wait_for_operation+0xd8/0x169 [<0>] afs_do_sync_operation+0x16/0x1f [<0>] afs_setattr+0x1fb/0x25d [<0>] notify_change+0x2cf/0x3c4 [<0>] do_truncate+0x7f/0xb2 [<0>] do_sys_ftruncate+0xd1/0x104 [<0>] do_syscall_64+0x2d/0x3a [<0>] entry_SYSCALL_64_after_hwframe+0x44/0xa9 The writeback operation, however, stalls indefinitely because it needs to get the io_lock to proceed: # cat /proc/5940/stack [<0>] afs_get_io_locks+0x58/0x1ae [<0>] afs_begin_vnode_operation+0xc7/0xd1 [<0>] afs_store_data+0x1b2/0x2a3 [<0>] afs_write_back_from_locked_page+0x418/0x57c [<0>] afs_writepages_region+0x196/0x224 [<0>] afs_writepages+0x74/0x156 [<0>] do_writepages+0x2d/0x56 [<0>] __writeback_single_inode+0x84/0x207 [<0>] writeback_sb_inodes+0x238/0x3cf [<0>] __writeback_inodes_wb+0x68/0x9f [<0>] wb_writeback+0x145/0x26c [<0>] wb_do_writeback+0x16a/0x194 [<0>] wb_workfn+0x74/0x177 [<0>] process_one_work+0x174/0x264 [<0>] worker_thread+0x117/0x1b9 [<0>] kthread+0xec/0xf1 [<0>] ret_from_fork+0x1f/0x30 and thus deadlock has occurred. Note that whilst afs_setattr() calls filemap_write_and_wait(), the fact that the caller is holding i_rwsem doesn't preclude more pages being dirtied through an mmap'd region. Fix this by: (1) Use the vnode validate_lock to mediate access between afs_setattr() and afs_writepages(): (a) Exclusively lock validate_lock in afs_setattr() around the whole RPC operation. (b) If WB_SYNC_ALL isn't set on entry to afs_writepages(), trying to shared-lock validate_lock and returning immediately if we couldn't get it. (c) If WB_SYNC_ALL is set, wait for the lock. The validate_lock is also used to validate a file and to zap its cache if the file was altered by a third party, so it's probably a good fit for this. (2) Move the truncation outside of the io_lock in setattr, using the same hook as is used for local directory editing. This requires the old i_size to be retained in the operation record as we commit the revised status to the inode members inside the io_lock still, but we still need to know if we reduced the file size. Fixes: d2ddc776a458 ("afs: Overhaul volume and server record caching and fileserver rotation") Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-10-07 21:22:12 +08:00
struct afs_vnode *vnode = AFS_FS_I(mapping->host);
pgoff_t start, end, next;
int ret;
_enter("");
afs: Fix deadlock between writeback and truncate The afs filesystem has a lock[*] that it uses to serialise I/O operations going to the server (vnode->io_lock), as the server will only perform one modification operation at a time on any given file or directory. This prevents the the filesystem from filling up all the call slots to a server with calls that aren't going to be executed in parallel anyway, thereby allowing operations on other files to obtain slots. [*] Note that is probably redundant for directories at least since i_rwsem is used to serialise directory modifications and lookup/reading vs modification. The server does allow parallel non-modification ops, however. When a file truncation op completes, we truncate the in-memory copy of the file to match - but we do it whilst still holding the io_lock, the idea being to prevent races with other operations. However, if writeback starts in a worker thread simultaneously with truncation (whilst notify_change() is called with i_rwsem locked, writeback pays it no heed), it may manage to set PG_writeback bits on the pages that will get truncated before afs_setattr_success() manages to call truncate_pagecache(). Truncate will then wait for those pages - whilst still inside io_lock: # cat /proc/8837/stack [<0>] wait_on_page_bit_common+0x184/0x1e7 [<0>] truncate_inode_pages_range+0x37f/0x3eb [<0>] truncate_pagecache+0x3c/0x53 [<0>] afs_setattr_success+0x4d/0x6e [<0>] afs_wait_for_operation+0xd8/0x169 [<0>] afs_do_sync_operation+0x16/0x1f [<0>] afs_setattr+0x1fb/0x25d [<0>] notify_change+0x2cf/0x3c4 [<0>] do_truncate+0x7f/0xb2 [<0>] do_sys_ftruncate+0xd1/0x104 [<0>] do_syscall_64+0x2d/0x3a [<0>] entry_SYSCALL_64_after_hwframe+0x44/0xa9 The writeback operation, however, stalls indefinitely because it needs to get the io_lock to proceed: # cat /proc/5940/stack [<0>] afs_get_io_locks+0x58/0x1ae [<0>] afs_begin_vnode_operation+0xc7/0xd1 [<0>] afs_store_data+0x1b2/0x2a3 [<0>] afs_write_back_from_locked_page+0x418/0x57c [<0>] afs_writepages_region+0x196/0x224 [<0>] afs_writepages+0x74/0x156 [<0>] do_writepages+0x2d/0x56 [<0>] __writeback_single_inode+0x84/0x207 [<0>] writeback_sb_inodes+0x238/0x3cf [<0>] __writeback_inodes_wb+0x68/0x9f [<0>] wb_writeback+0x145/0x26c [<0>] wb_do_writeback+0x16a/0x194 [<0>] wb_workfn+0x74/0x177 [<0>] process_one_work+0x174/0x264 [<0>] worker_thread+0x117/0x1b9 [<0>] kthread+0xec/0xf1 [<0>] ret_from_fork+0x1f/0x30 and thus deadlock has occurred. Note that whilst afs_setattr() calls filemap_write_and_wait(), the fact that the caller is holding i_rwsem doesn't preclude more pages being dirtied through an mmap'd region. Fix this by: (1) Use the vnode validate_lock to mediate access between afs_setattr() and afs_writepages(): (a) Exclusively lock validate_lock in afs_setattr() around the whole RPC operation. (b) If WB_SYNC_ALL isn't set on entry to afs_writepages(), trying to shared-lock validate_lock and returning immediately if we couldn't get it. (c) If WB_SYNC_ALL is set, wait for the lock. The validate_lock is also used to validate a file and to zap its cache if the file was altered by a third party, so it's probably a good fit for this. (2) Move the truncation outside of the io_lock in setattr, using the same hook as is used for local directory editing. This requires the old i_size to be retained in the operation record as we commit the revised status to the inode members inside the io_lock still, but we still need to know if we reduced the file size. Fixes: d2ddc776a458 ("afs: Overhaul volume and server record caching and fileserver rotation") Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-10-07 21:22:12 +08:00
/* We have to be careful as we can end up racing with setattr()
* truncating the pagecache since the caller doesn't take a lock here
* to prevent it.
*/
if (wbc->sync_mode == WB_SYNC_ALL)
down_read(&vnode->validate_lock);
else if (!down_read_trylock(&vnode->validate_lock))
return 0;
if (wbc->range_cyclic) {
start = mapping->writeback_index;
end = -1;
ret = afs_writepages_region(mapping, wbc, start, end, &next);
if (start > 0 && wbc->nr_to_write > 0 && ret == 0)
ret = afs_writepages_region(mapping, wbc, 0, start,
&next);
mapping->writeback_index = next;
} else if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX) {
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 = (pgoff_t)(LLONG_MAX >> PAGE_SHIFT);
ret = afs_writepages_region(mapping, wbc, 0, end, &next);
if (wbc->nr_to_write > 0)
mapping->writeback_index = next;
} else {
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
start = wbc->range_start >> PAGE_SHIFT;
end = wbc->range_end >> PAGE_SHIFT;
ret = afs_writepages_region(mapping, wbc, start, end, &next);
}
afs: Fix deadlock between writeback and truncate The afs filesystem has a lock[*] that it uses to serialise I/O operations going to the server (vnode->io_lock), as the server will only perform one modification operation at a time on any given file or directory. This prevents the the filesystem from filling up all the call slots to a server with calls that aren't going to be executed in parallel anyway, thereby allowing operations on other files to obtain slots. [*] Note that is probably redundant for directories at least since i_rwsem is used to serialise directory modifications and lookup/reading vs modification. The server does allow parallel non-modification ops, however. When a file truncation op completes, we truncate the in-memory copy of the file to match - but we do it whilst still holding the io_lock, the idea being to prevent races with other operations. However, if writeback starts in a worker thread simultaneously with truncation (whilst notify_change() is called with i_rwsem locked, writeback pays it no heed), it may manage to set PG_writeback bits on the pages that will get truncated before afs_setattr_success() manages to call truncate_pagecache(). Truncate will then wait for those pages - whilst still inside io_lock: # cat /proc/8837/stack [<0>] wait_on_page_bit_common+0x184/0x1e7 [<0>] truncate_inode_pages_range+0x37f/0x3eb [<0>] truncate_pagecache+0x3c/0x53 [<0>] afs_setattr_success+0x4d/0x6e [<0>] afs_wait_for_operation+0xd8/0x169 [<0>] afs_do_sync_operation+0x16/0x1f [<0>] afs_setattr+0x1fb/0x25d [<0>] notify_change+0x2cf/0x3c4 [<0>] do_truncate+0x7f/0xb2 [<0>] do_sys_ftruncate+0xd1/0x104 [<0>] do_syscall_64+0x2d/0x3a [<0>] entry_SYSCALL_64_after_hwframe+0x44/0xa9 The writeback operation, however, stalls indefinitely because it needs to get the io_lock to proceed: # cat /proc/5940/stack [<0>] afs_get_io_locks+0x58/0x1ae [<0>] afs_begin_vnode_operation+0xc7/0xd1 [<0>] afs_store_data+0x1b2/0x2a3 [<0>] afs_write_back_from_locked_page+0x418/0x57c [<0>] afs_writepages_region+0x196/0x224 [<0>] afs_writepages+0x74/0x156 [<0>] do_writepages+0x2d/0x56 [<0>] __writeback_single_inode+0x84/0x207 [<0>] writeback_sb_inodes+0x238/0x3cf [<0>] __writeback_inodes_wb+0x68/0x9f [<0>] wb_writeback+0x145/0x26c [<0>] wb_do_writeback+0x16a/0x194 [<0>] wb_workfn+0x74/0x177 [<0>] process_one_work+0x174/0x264 [<0>] worker_thread+0x117/0x1b9 [<0>] kthread+0xec/0xf1 [<0>] ret_from_fork+0x1f/0x30 and thus deadlock has occurred. Note that whilst afs_setattr() calls filemap_write_and_wait(), the fact that the caller is holding i_rwsem doesn't preclude more pages being dirtied through an mmap'd region. Fix this by: (1) Use the vnode validate_lock to mediate access between afs_setattr() and afs_writepages(): (a) Exclusively lock validate_lock in afs_setattr() around the whole RPC operation. (b) If WB_SYNC_ALL isn't set on entry to afs_writepages(), trying to shared-lock validate_lock and returning immediately if we couldn't get it. (c) If WB_SYNC_ALL is set, wait for the lock. The validate_lock is also used to validate a file and to zap its cache if the file was altered by a third party, so it's probably a good fit for this. (2) Move the truncation outside of the io_lock in setattr, using the same hook as is used for local directory editing. This requires the old i_size to be retained in the operation record as we commit the revised status to the inode members inside the io_lock still, but we still need to know if we reduced the file size. Fixes: d2ddc776a458 ("afs: Overhaul volume and server record caching and fileserver rotation") Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-10-07 21:22:12 +08:00
up_read(&vnode->validate_lock);
_leave(" = %d", ret);
return ret;
}
/*
* write to an AFS file
*/
ssize_t afs_file_write(struct kiocb *iocb, struct iov_iter *from)
{
struct afs_vnode *vnode = AFS_FS_I(file_inode(iocb->ki_filp));
ssize_t result;
size_t count = iov_iter_count(from);
_enter("{%llx:%llu},{%zu},",
vnode->fid.vid, vnode->fid.vnode, count);
if (IS_SWAPFILE(&vnode->vfs_inode)) {
printk(KERN_INFO
"AFS: Attempt to write to active swap file!\n");
return -EBUSY;
}
if (!count)
return 0;
result = generic_file_write_iter(iocb, from);
_leave(" = %zd", result);
return result;
}
/*
* flush any dirty pages for this process, and check for write errors.
* - the return status from this call provides a reliable indication of
* whether any write errors occurred for this process.
*/
int afs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file_inode(file);
struct afs_vnode *vnode = AFS_FS_I(inode);
_enter("{%llx:%llu},{n=%pD},%d",
vnode->fid.vid, vnode->fid.vnode, file,
datasync);
return file_write_and_wait_range(file, start, end);
}
/*
* notification that a previously read-only page is about to become writable
* - if it returns an error, the caller will deliver a bus error signal
*/
vm_fault_t afs_page_mkwrite(struct vm_fault *vmf)
{
struct file *file = vmf->vma->vm_file;
struct inode *inode = file_inode(file);
struct afs_vnode *vnode = AFS_FS_I(inode);
unsigned long priv;
_enter("{{%llx:%llu}},{%lx}",
vnode->fid.vid, vnode->fid.vnode, vmf->page->index);
sb_start_pagefault(inode->i_sb);
/* Wait for the page to be written to the cache before we allow it to
* be modified. We then assume the entire page will need writing back.
*/
#ifdef CONFIG_AFS_FSCACHE
fscache_wait_on_page_write(vnode->cache, vmf->page);
#endif
if (PageWriteback(vmf->page) &&
wait_on_page_bit_killable(vmf->page, PG_writeback) < 0)
return VM_FAULT_RETRY;
if (lock_page_killable(vmf->page) < 0)
return VM_FAULT_RETRY;
/* We mustn't change page->private until writeback is complete as that
* details the portion of the page we need to write back and we might
* need to redirty the page if there's a problem.
*/
wait_on_page_writeback(vmf->page);
priv = afs_page_dirty(0, PAGE_SIZE);
priv = afs_page_dirty_mmapped(priv);
trace_afs_page_dirty(vnode, tracepoint_string("mkwrite"),
vmf->page->index, priv);
if (PagePrivate(vmf->page))
set_page_private(vmf->page, priv);
else
attach_page_private(vmf->page, (void *)priv);
file_update_time(file);
sb_end_pagefault(inode->i_sb);
return VM_FAULT_LOCKED;
}
/*
* Prune the keys cached for writeback. The caller must hold vnode->wb_lock.
*/
void afs_prune_wb_keys(struct afs_vnode *vnode)
{
LIST_HEAD(graveyard);
struct afs_wb_key *wbk, *tmp;
/* Discard unused keys */
spin_lock(&vnode->wb_lock);
if (!mapping_tagged(&vnode->vfs_inode.i_data, PAGECACHE_TAG_WRITEBACK) &&
!mapping_tagged(&vnode->vfs_inode.i_data, PAGECACHE_TAG_DIRTY)) {
list_for_each_entry_safe(wbk, tmp, &vnode->wb_keys, vnode_link) {
if (refcount_read(&wbk->usage) == 1)
list_move(&wbk->vnode_link, &graveyard);
}
}
spin_unlock(&vnode->wb_lock);
while (!list_empty(&graveyard)) {
wbk = list_entry(graveyard.next, struct afs_wb_key, vnode_link);
list_del(&wbk->vnode_link);
afs_put_wb_key(wbk);
}
}
/*
* Clean up a page during invalidation.
*/
int afs_launder_page(struct page *page)
{
struct address_space *mapping = page->mapping;
struct afs_vnode *vnode = AFS_FS_I(mapping->host);
unsigned long priv;
unsigned int f, t;
int ret = 0;
_enter("{%lx}", page->index);
priv = page_private(page);
if (clear_page_dirty_for_io(page)) {
f = 0;
t = PAGE_SIZE;
if (PagePrivate(page)) {
f = afs_page_dirty_from(priv);
t = afs_page_dirty_to(priv);
}
trace_afs_page_dirty(vnode, tracepoint_string("launder"),
page->index, priv);
ret = afs_store_data(mapping, page->index, page->index, t, f, true);
}
priv = (unsigned long)detach_page_private(page);
trace_afs_page_dirty(vnode, tracepoint_string("laundered"),
page->index, priv);
#ifdef CONFIG_AFS_FSCACHE
if (PageFsCache(page)) {
fscache_wait_on_page_write(vnode->cache, page);
fscache_uncache_page(vnode->cache, page);
}
#endif
return ret;
}