linux-stable/fs/xfs/xfs_buf.c

2367 lines
58 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0
/*
* Copyright (c) 2000-2006 Silicon Graphics, Inc.
* All Rights Reserved.
*/
#include "xfs.h"
#include <linux/backing-dev.h>
#include <linux/dax.h>
#include "xfs_shared.h"
#include "xfs_format.h"
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
#include "xfs_mount.h"
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
#include "xfs_trace.h"
#include "xfs_log.h"
#include "xfs_log_recover.h"
xfs: xfs_is_shutdown vs xlog_is_shutdown cage fight I've been chasing a recent resurgence in generic/388 recovery failure and/or corruption events. The events have largely been uninitialised inode chunks being tripped over in log recovery such as: XFS (pmem1): User initiated shutdown received. pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/xfs/xfs_fsops.c:500). Shutting down filesystem. XFS (pmem1): Please unmount the filesystem and rectify the problem(s) XFS (pmem1): Unmounting Filesystem XFS (pmem1): Mounting V5 Filesystem XFS (pmem1): Starting recovery (logdev: internal) XFS (pmem1): bad inode magic/vsn daddr 8723584 #0 (magic=1818) XFS (pmem1): Metadata corruption detected at xfs_inode_buf_verify+0x180/0x190, xfs_inode block 0x851c80 xfs_inode_buf_verify XFS (pmem1): Unmount and run xfs_repair XFS (pmem1): First 128 bytes of corrupted metadata buffer: 00000000: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000010: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000020: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000030: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000040: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000050: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000060: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000070: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ XFS (pmem1): metadata I/O error in "xlog_recover_items_pass2+0x52/0xc0" at daddr 0x851c80 len 32 error 117 XFS (pmem1): log mount/recovery failed: error -117 XFS (pmem1): log mount failed There have been isolated random other issues, too - xfs_repair fails because it finds some corruption in symlink blocks, rmap inconsistencies, etc - but they are nowhere near as common as the uninitialised inode chunk failure. The problem has clearly happened at runtime before recovery has run; I can see the ICREATE log item in the log shortly before the actively recovered range of the log. This means the ICREATE was definitely created and written to the log, but for some reason the tail of the log has been moved past the ordered buffer log item that tracks INODE_ALLOC buffers and, supposedly, prevents the tail of the log moving past the ICREATE log item before the inode chunk buffer is written to disk. Tracing the fsstress processes that are running when the filesystem shut down immediately pin-pointed the problem: user shutdown marks xfs_mount as shutdown godown-213341 [008] 6398.022871: console: [ 6397.915392] XFS (pmem1): User initiated shutdown received. ..... aild tries to push ordered inode cluster buffer xfsaild/pmem1-213314 [001] 6398.022974: xfs_buf_trylock: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 16 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_inode_item_push+0x8e xfsaild/pmem1-213314 [001] 6398.022976: xfs_ilock_nowait: dev 259:1 ino 0x851c80 flags ILOCK_SHARED caller xfs_iflush_cluster+0xae xfs_iflush_cluster() checks xfs_is_shutdown(), returns true, calls xfs_iflush_abort() to kill writeback of the inode. Inode is removed from AIL, drops cluster buffer reference. xfsaild/pmem1-213314 [001] 6398.022977: xfs_ail_delete: dev 259:1 lip 0xffff88880247ed80 old lsn 7/20344 new lsn 7/21000 type XFS_LI_INODE flags IN_AIL xfsaild/pmem1-213314 [001] 6398.022978: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 17 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_iflush_abort+0xd7 ..... All inodes on cluster buffer are aborted, then the cluster buffer itself is aborted and removed from the AIL *without writeback*: xfsaild/pmem1-213314 [001] 6398.023011: xfs_buf_error_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_ioend_fail+0x33 xfsaild/pmem1-213314 [001] 6398.023012: xfs_ail_delete: dev 259:1 lip 0xffff8888053efde8 old lsn 7/20344 new lsn 7/20344 type XFS_LI_BUF flags IN_AIL The inode buffer was at 7/20344 when it was removed from the AIL. xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_item_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_done+0x31 xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_relse+0x39 ..... Userspace is still running, doing stuff. an fsstress process runs syncfs() or sync() and we end up in sync_fs_one_sb() which issues a log force. This pushes on the CIL: fsstress-213322 [001] 6398.024430: xfs_fs_sync_fs: dev 259:1 m_features 0x20000000019ff6e9 opstate (clean|shutdown|inodegc|blockgc) s_flags 0x70810000 caller sync_fs_one_sb+0x26 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x0 caller xfs_fs_sync_fs+0x82 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x5f caller xfs_log_force+0x7c <...>-194402 [001] 6398.024467: kmem_alloc: size 176 flags 0x14 caller xlog_cil_push_work+0x9f And the CIL fills up iclogs with pending changes. This picks up the current tail from the AIL: <...>-194402 [001] 6398.024497: xlog_iclog_get_space: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x0 flags caller xlog_write+0x149 <...>-194402 [001] 6398.024498: xlog_iclog_switch: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x700005408 flags caller xlog_state_get_iclog_space+0x37e <...>-194402 [001] 6398.024521: xlog_iclog_release: dev 259:1 state XLOG_STATE_WANT_SYNC refcnt 1 offset 32256 lsn 0x700005408 flags caller xlog_write+0x5f9 <...>-194402 [001] 6398.024522: xfs_log_assign_tail_lsn: dev 259:1 new tail lsn 7/21000, old lsn 7/20344, last sync 7/21448 And it moves the tail of the log to 7/21000 from 7/20344. This *moves the tail of the log beyond the ICREATE transaction* that was at 7/20344 and pinned by the inode cluster buffer that was cancelled above. .... godown-213341 [008] 6398.027005: xfs_force_shutdown: dev 259:1 tag logerror flags log_io|force_umount file fs/xfs/xfs_fsops.c line_num 500 godown-213341 [008] 6398.027022: console: [ 6397.915406] pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 godown-213341 [008] 6398.030551: console: [ 6397.919546] XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/ And finally the log itself is now shutdown, stopping all further writes to the log. But this is too late to prevent the corruption that moving the tail of the log forwards after we start cancelling writeback causes. The fundamental problem here is that we are using the wrong shutdown checks for log items. We've long conflated mount shutdown with log shutdown state, and I started separating that recently with the atomic shutdown state changes in commit b36d4651e165 ("xfs: make forced shutdown processing atomic"). The changes in that commit series are directly responsible for being able to diagnose this issue because it clearly separated mount shutdown from log shutdown. Essentially, once we start cancelling writeback of log items and removing them from the AIL because the filesystem is shut down, we *cannot* update the journal because we may have cancelled the items that pin the tail of the log. That moves the tail of the log forwards without having written the metadata back, hence we have corrupt in memory state and writing to the journal propagates that to the on-disk state. What commit b36d4651e165 makes clear is that log item state needs to change relative to log shutdown, not mount shutdown. IOWs, anything that aborts metadata writeback needs to check log shutdown state because log items directly affect log consistency. Having them check mount shutdown state introduces the above race condition where we cancel metadata writeback before the log shuts down. To fix this, this patch works through all log items and converts shutdown checks to use xlog_is_shutdown() rather than xfs_is_shutdown(), so that we don't start aborting metadata writeback before we shut off journal writes. AFAICT, this race condition is a zero day IO error handling bug in XFS that dates back to the introduction of XLOG_IO_ERROR, XLOG_STATE_IOERROR and XFS_FORCED_SHUTDOWN back in January 1997. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:13 +00:00
#include "xfs_log_priv.h"
#include "xfs_trans.h"
#include "xfs_buf_item.h"
#include "xfs_errortag.h"
#include "xfs_error.h"
#include "xfs_ag.h"
struct kmem_cache *xfs_buf_cache;
xfs: fix use-after-free race in xfs_buf_rele When looking at a 4.18 based KASAN use after free report, I noticed that racing xfs_buf_rele() may race on dropping the last reference to the buffer and taking the buffer lock. This was the symptom displayed by the KASAN report, but the actual issue that was reported had already been fixed in 4.19-rc1 by commit e339dd8d8b04 ("xfs: use sync buffer I/O for sync delwri queue submission"). Despite this, I think there is still an issue with xfs_buf_rele() in this code: release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock); spin_lock(&bp->b_lock); if (!release) { ..... If two threads race on the b_lock after both dropping a reference and one getting dropping the last reference so release = true, we end up with: CPU 0 CPU 1 atomic_dec_and_lock() atomic_dec_and_lock() spin_lock(&bp->b_lock) spin_lock(&bp->b_lock) <spins> <release = true bp->b_lru_ref = 0> <remove from lists> freebuf = true spin_unlock(&bp->b_lock) xfs_buf_free(bp) <gets lock, reading and writing freed memory> <accesses freed memory> spin_unlock(&bp->b_lock) <reads/writes freed memory> IOWs, we can't safely take bp->b_lock after dropping the hold reference because the buffer may go away at any time after we drop that reference. However, this can be fixed simply by taking the bp->b_lock before we drop the reference. It is safe to nest the pag_buf_lock inside bp->b_lock as the pag_buf_lock is only used to serialise against lookup in xfs_buf_find() and no other locks are held over or under the pag_buf_lock there. Make this clear by documenting the buffer lock orders at the top of the file. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com Signed-off-by: Dave Chinner <david@fromorbit.com>
2018-10-18 06:21:29 +00:00
/*
* Locking orders
*
* xfs_buf_ioacct_inc:
* xfs_buf_ioacct_dec:
* b_sema (caller holds)
* b_lock
*
* xfs_buf_stale:
* b_sema (caller holds)
* b_lock
* lru_lock
*
* xfs_buf_rele:
* b_lock
* pag_buf_lock
* lru_lock
*
* xfs_buftarg_drain_rele
xfs: fix use-after-free race in xfs_buf_rele When looking at a 4.18 based KASAN use after free report, I noticed that racing xfs_buf_rele() may race on dropping the last reference to the buffer and taking the buffer lock. This was the symptom displayed by the KASAN report, but the actual issue that was reported had already been fixed in 4.19-rc1 by commit e339dd8d8b04 ("xfs: use sync buffer I/O for sync delwri queue submission"). Despite this, I think there is still an issue with xfs_buf_rele() in this code: release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock); spin_lock(&bp->b_lock); if (!release) { ..... If two threads race on the b_lock after both dropping a reference and one getting dropping the last reference so release = true, we end up with: CPU 0 CPU 1 atomic_dec_and_lock() atomic_dec_and_lock() spin_lock(&bp->b_lock) spin_lock(&bp->b_lock) <spins> <release = true bp->b_lru_ref = 0> <remove from lists> freebuf = true spin_unlock(&bp->b_lock) xfs_buf_free(bp) <gets lock, reading and writing freed memory> <accesses freed memory> spin_unlock(&bp->b_lock) <reads/writes freed memory> IOWs, we can't safely take bp->b_lock after dropping the hold reference because the buffer may go away at any time after we drop that reference. However, this can be fixed simply by taking the bp->b_lock before we drop the reference. It is safe to nest the pag_buf_lock inside bp->b_lock as the pag_buf_lock is only used to serialise against lookup in xfs_buf_find() and no other locks are held over or under the pag_buf_lock there. Make this clear by documenting the buffer lock orders at the top of the file. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com Signed-off-by: Dave Chinner <david@fromorbit.com>
2018-10-18 06:21:29 +00:00
* lru_lock
* b_lock (trylock due to inversion)
*
* xfs_buftarg_isolate
* lru_lock
* b_lock (trylock due to inversion)
*/
static int __xfs_buf_submit(struct xfs_buf *bp, bool wait);
static inline int
xfs_buf_submit(
struct xfs_buf *bp)
{
return __xfs_buf_submit(bp, !(bp->b_flags & XBF_ASYNC));
}
static inline int
xfs_buf_is_vmapped(
struct xfs_buf *bp)
{
/*
* Return true if the buffer is vmapped.
*
* b_addr is null if the buffer is not mapped, but the code is clever
* enough to know it doesn't have to map a single page, so the check has
* to be both for b_addr and bp->b_page_count > 1.
*/
return bp->b_addr && bp->b_page_count > 1;
}
static inline int
xfs_buf_vmap_len(
struct xfs_buf *bp)
{
return (bp->b_page_count * PAGE_SIZE);
}
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
/*
* Bump the I/O in flight count on the buftarg if we haven't yet done so for
* this buffer. The count is incremented once per buffer (per hold cycle)
* because the corresponding decrement is deferred to buffer release. Buffers
* can undergo I/O multiple times in a hold-release cycle and per buffer I/O
* tracking adds unnecessary overhead. This is used for sychronization purposes
* with unmount (see xfs_buftarg_drain()), so all we really need is a count of
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
* in-flight buffers.
*
* Buffers that are never released (e.g., superblock, iclog buffers) must set
* the XBF_NO_IOACCT flag before I/O submission. Otherwise, the buftarg count
* never reaches zero and unmount hangs indefinitely.
*/
static inline void
xfs_buf_ioacct_inc(
struct xfs_buf *bp)
{
xfs: use ->b_state to fix buffer I/O accounting release race We've had user reports of unmount hangs in xfs_wait_buftarg() that analysis shows is due to btp->bt_io_count == -1. bt_io_count represents the count of in-flight asynchronous buffers and thus should always be >= 0. xfs_wait_buftarg() waits for this value to stabilize to zero in order to ensure that all untracked (with respect to the lru) buffers have completed I/O processing before unmount proceeds to tear down in-core data structures. The value of -1 implies an I/O accounting decrement race. Indeed, the fact that xfs_buf_ioacct_dec() is called from xfs_buf_rele() (where the buffer lock is no longer held) means that bp->b_flags can be updated from an unsafe context. While a user-level reproducer is currently not available, some intrusive hacks to run racing buffer lookups/ioacct/releases from multiple threads was used to successfully manufacture this problem. Existing callers do not expect to acquire the buffer lock from xfs_buf_rele(). Therefore, we can not safely update ->b_flags from this context. It turns out that we already have separate buffer state bits and associated serialization for dealing with buffer LRU state in the form of ->b_state and ->b_lock. Therefore, replace the _XBF_IN_FLIGHT flag with a ->b_state variant, update the I/O accounting wrappers appropriately and make sure they are used with the correct locking. This ensures that buffer in-flight state can be modified at buffer release time without racing with modifications from a buffer lock holder. Fixes: 9c7504aa72b6 ("xfs: track and serialize in-flight async buffers against unmount") Cc: <stable@vger.kernel.org> # v4.8+ Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Nikolay Borisov <nborisov@suse.com> Tested-by: Libor Pechacek <lpechacek@suse.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-05-31 15:22:52 +00:00
if (bp->b_flags & XBF_NO_IOACCT)
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
return;
ASSERT(bp->b_flags & XBF_ASYNC);
xfs: use ->b_state to fix buffer I/O accounting release race We've had user reports of unmount hangs in xfs_wait_buftarg() that analysis shows is due to btp->bt_io_count == -1. bt_io_count represents the count of in-flight asynchronous buffers and thus should always be >= 0. xfs_wait_buftarg() waits for this value to stabilize to zero in order to ensure that all untracked (with respect to the lru) buffers have completed I/O processing before unmount proceeds to tear down in-core data structures. The value of -1 implies an I/O accounting decrement race. Indeed, the fact that xfs_buf_ioacct_dec() is called from xfs_buf_rele() (where the buffer lock is no longer held) means that bp->b_flags can be updated from an unsafe context. While a user-level reproducer is currently not available, some intrusive hacks to run racing buffer lookups/ioacct/releases from multiple threads was used to successfully manufacture this problem. Existing callers do not expect to acquire the buffer lock from xfs_buf_rele(). Therefore, we can not safely update ->b_flags from this context. It turns out that we already have separate buffer state bits and associated serialization for dealing with buffer LRU state in the form of ->b_state and ->b_lock. Therefore, replace the _XBF_IN_FLIGHT flag with a ->b_state variant, update the I/O accounting wrappers appropriately and make sure they are used with the correct locking. This ensures that buffer in-flight state can be modified at buffer release time without racing with modifications from a buffer lock holder. Fixes: 9c7504aa72b6 ("xfs: track and serialize in-flight async buffers against unmount") Cc: <stable@vger.kernel.org> # v4.8+ Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Nikolay Borisov <nborisov@suse.com> Tested-by: Libor Pechacek <lpechacek@suse.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-05-31 15:22:52 +00:00
spin_lock(&bp->b_lock);
if (!(bp->b_state & XFS_BSTATE_IN_FLIGHT)) {
bp->b_state |= XFS_BSTATE_IN_FLIGHT;
percpu_counter_inc(&bp->b_target->bt_io_count);
}
spin_unlock(&bp->b_lock);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
}
/*
* Clear the in-flight state on a buffer about to be released to the LRU or
* freed and unaccount from the buftarg.
*/
static inline void
xfs: use ->b_state to fix buffer I/O accounting release race We've had user reports of unmount hangs in xfs_wait_buftarg() that analysis shows is due to btp->bt_io_count == -1. bt_io_count represents the count of in-flight asynchronous buffers and thus should always be >= 0. xfs_wait_buftarg() waits for this value to stabilize to zero in order to ensure that all untracked (with respect to the lru) buffers have completed I/O processing before unmount proceeds to tear down in-core data structures. The value of -1 implies an I/O accounting decrement race. Indeed, the fact that xfs_buf_ioacct_dec() is called from xfs_buf_rele() (where the buffer lock is no longer held) means that bp->b_flags can be updated from an unsafe context. While a user-level reproducer is currently not available, some intrusive hacks to run racing buffer lookups/ioacct/releases from multiple threads was used to successfully manufacture this problem. Existing callers do not expect to acquire the buffer lock from xfs_buf_rele(). Therefore, we can not safely update ->b_flags from this context. It turns out that we already have separate buffer state bits and associated serialization for dealing with buffer LRU state in the form of ->b_state and ->b_lock. Therefore, replace the _XBF_IN_FLIGHT flag with a ->b_state variant, update the I/O accounting wrappers appropriately and make sure they are used with the correct locking. This ensures that buffer in-flight state can be modified at buffer release time without racing with modifications from a buffer lock holder. Fixes: 9c7504aa72b6 ("xfs: track and serialize in-flight async buffers against unmount") Cc: <stable@vger.kernel.org> # v4.8+ Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Nikolay Borisov <nborisov@suse.com> Tested-by: Libor Pechacek <lpechacek@suse.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-05-31 15:22:52 +00:00
__xfs_buf_ioacct_dec(
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
struct xfs_buf *bp)
{
lockdep_assert_held(&bp->b_lock);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
xfs: use ->b_state to fix buffer I/O accounting release race We've had user reports of unmount hangs in xfs_wait_buftarg() that analysis shows is due to btp->bt_io_count == -1. bt_io_count represents the count of in-flight asynchronous buffers and thus should always be >= 0. xfs_wait_buftarg() waits for this value to stabilize to zero in order to ensure that all untracked (with respect to the lru) buffers have completed I/O processing before unmount proceeds to tear down in-core data structures. The value of -1 implies an I/O accounting decrement race. Indeed, the fact that xfs_buf_ioacct_dec() is called from xfs_buf_rele() (where the buffer lock is no longer held) means that bp->b_flags can be updated from an unsafe context. While a user-level reproducer is currently not available, some intrusive hacks to run racing buffer lookups/ioacct/releases from multiple threads was used to successfully manufacture this problem. Existing callers do not expect to acquire the buffer lock from xfs_buf_rele(). Therefore, we can not safely update ->b_flags from this context. It turns out that we already have separate buffer state bits and associated serialization for dealing with buffer LRU state in the form of ->b_state and ->b_lock. Therefore, replace the _XBF_IN_FLIGHT flag with a ->b_state variant, update the I/O accounting wrappers appropriately and make sure they are used with the correct locking. This ensures that buffer in-flight state can be modified at buffer release time without racing with modifications from a buffer lock holder. Fixes: 9c7504aa72b6 ("xfs: track and serialize in-flight async buffers against unmount") Cc: <stable@vger.kernel.org> # v4.8+ Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Nikolay Borisov <nborisov@suse.com> Tested-by: Libor Pechacek <lpechacek@suse.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-05-31 15:22:52 +00:00
if (bp->b_state & XFS_BSTATE_IN_FLIGHT) {
bp->b_state &= ~XFS_BSTATE_IN_FLIGHT;
percpu_counter_dec(&bp->b_target->bt_io_count);
}
}
static inline void
xfs_buf_ioacct_dec(
struct xfs_buf *bp)
{
spin_lock(&bp->b_lock);
__xfs_buf_ioacct_dec(bp);
spin_unlock(&bp->b_lock);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
}
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
/*
* When we mark a buffer stale, we remove the buffer from the LRU and clear the
* b_lru_ref count so that the buffer is freed immediately when the buffer
* reference count falls to zero. If the buffer is already on the LRU, we need
* to remove the reference that LRU holds on the buffer.
*
* This prevents build-up of stale buffers on the LRU.
*/
void
xfs_buf_stale(
struct xfs_buf *bp)
{
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
ASSERT(xfs_buf_islocked(bp));
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
bp->b_flags |= XBF_STALE;
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
/*
* Clear the delwri status so that a delwri queue walker will not
* flush this buffer to disk now that it is stale. The delwri queue has
* a reference to the buffer, so this is safe to do.
*/
bp->b_flags &= ~_XBF_DELWRI_Q;
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
/*
* Once the buffer is marked stale and unlocked, a subsequent lookup
* could reset b_flags. There is no guarantee that the buffer is
* unaccounted (released to LRU) before that occurs. Drop in-flight
* status now to preserve accounting consistency.
*/
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
spin_lock(&bp->b_lock);
xfs: use ->b_state to fix buffer I/O accounting release race We've had user reports of unmount hangs in xfs_wait_buftarg() that analysis shows is due to btp->bt_io_count == -1. bt_io_count represents the count of in-flight asynchronous buffers and thus should always be >= 0. xfs_wait_buftarg() waits for this value to stabilize to zero in order to ensure that all untracked (with respect to the lru) buffers have completed I/O processing before unmount proceeds to tear down in-core data structures. The value of -1 implies an I/O accounting decrement race. Indeed, the fact that xfs_buf_ioacct_dec() is called from xfs_buf_rele() (where the buffer lock is no longer held) means that bp->b_flags can be updated from an unsafe context. While a user-level reproducer is currently not available, some intrusive hacks to run racing buffer lookups/ioacct/releases from multiple threads was used to successfully manufacture this problem. Existing callers do not expect to acquire the buffer lock from xfs_buf_rele(). Therefore, we can not safely update ->b_flags from this context. It turns out that we already have separate buffer state bits and associated serialization for dealing with buffer LRU state in the form of ->b_state and ->b_lock. Therefore, replace the _XBF_IN_FLIGHT flag with a ->b_state variant, update the I/O accounting wrappers appropriately and make sure they are used with the correct locking. This ensures that buffer in-flight state can be modified at buffer release time without racing with modifications from a buffer lock holder. Fixes: 9c7504aa72b6 ("xfs: track and serialize in-flight async buffers against unmount") Cc: <stable@vger.kernel.org> # v4.8+ Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Nikolay Borisov <nborisov@suse.com> Tested-by: Libor Pechacek <lpechacek@suse.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-05-31 15:22:52 +00:00
__xfs_buf_ioacct_dec(bp);
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
atomic_set(&bp->b_lru_ref, 0);
if (!(bp->b_state & XFS_BSTATE_DISPOSE) &&
(list_lru_del(&bp->b_target->bt_lru, &bp->b_lru)))
atomic_dec(&bp->b_hold);
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
ASSERT(atomic_read(&bp->b_hold) >= 1);
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
spin_unlock(&bp->b_lock);
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
}
static int
xfs_buf_get_maps(
struct xfs_buf *bp,
int map_count)
{
ASSERT(bp->b_maps == NULL);
bp->b_map_count = map_count;
if (map_count == 1) {
bp->b_maps = &bp->__b_map;
return 0;
}
bp->b_maps = kmem_zalloc(map_count * sizeof(struct xfs_buf_map),
KM_NOFS);
if (!bp->b_maps)
return -ENOMEM;
return 0;
}
/*
* Frees b_pages if it was allocated.
*/
static void
xfs_buf_free_maps(
struct xfs_buf *bp)
{
if (bp->b_maps != &bp->__b_map) {
kmem_free(bp->b_maps);
bp->b_maps = NULL;
}
}
static int
_xfs_buf_alloc(
struct xfs_buftarg *target,
struct xfs_buf_map *map,
int nmaps,
xfs_buf_flags_t flags,
struct xfs_buf **bpp)
{
struct xfs_buf *bp;
int error;
int i;
*bpp = NULL;
bp = kmem_cache_zalloc(xfs_buf_cache, GFP_NOFS | __GFP_NOFAIL);
/*
* We don't want certain flags to appear in b_flags unless they are
* specifically set by later operations on the buffer.
*/
flags &= ~(XBF_UNMAPPED | XBF_TRYLOCK | XBF_ASYNC | XBF_READ_AHEAD);
atomic_set(&bp->b_hold, 1);
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
atomic_set(&bp->b_lru_ref, 1);
init_completion(&bp->b_iowait);
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
INIT_LIST_HEAD(&bp->b_lru);
INIT_LIST_HEAD(&bp->b_list);
INIT_LIST_HEAD(&bp->b_li_list);
sema_init(&bp->b_sema, 0); /* held, no waiters */
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
spin_lock_init(&bp->b_lock);
bp->b_target = target;
bp->b_mount = target->bt_mount;
bp->b_flags = flags;
/*
* Set length and io_length to the same value initially.
* I/O routines should use io_length, which will be the same in
* most cases but may be reset (e.g. XFS recovery).
*/
error = xfs_buf_get_maps(bp, nmaps);
if (error) {
kmem_cache_free(xfs_buf_cache, bp);
return error;
}
bp->b_rhash_key = map[0].bm_bn;
bp->b_length = 0;
for (i = 0; i < nmaps; i++) {
bp->b_maps[i].bm_bn = map[i].bm_bn;
bp->b_maps[i].bm_len = map[i].bm_len;
bp->b_length += map[i].bm_len;
}
atomic_set(&bp->b_pin_count, 0);
init_waitqueue_head(&bp->b_waiters);
XFS_STATS_INC(bp->b_mount, xb_create);
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_init(bp, _RET_IP_);
*bpp = bp;
return 0;
}
static void
xfs_buf_free_pages(
struct xfs_buf *bp)
{
uint i;
ASSERT(bp->b_flags & _XBF_PAGES);
if (xfs_buf_is_vmapped(bp))
vm_unmap_ram(bp->b_addr, bp->b_page_count);
for (i = 0; i < bp->b_page_count; i++) {
if (bp->b_pages[i])
__free_page(bp->b_pages[i]);
}
if (current->reclaim_state)
current->reclaim_state->reclaimed_slab += bp->b_page_count;
if (bp->b_pages != bp->b_page_array)
kmem_free(bp->b_pages);
bp->b_pages = NULL;
bp->b_flags &= ~_XBF_PAGES;
}
xfs: lockless buffer lookup Now that we have a standalone fast path for buffer lookup, we can easily convert it to use rcu lookups. When we continually hammer the buffer cache with trylock lookups, we end up with a huge amount of lock contention on the per-ag buffer hash locks: - 92.71% 0.05% [kernel] [k] xfs_inodegc_worker - 92.67% xfs_inodegc_worker - 92.13% xfs_inode_unlink - 91.52% xfs_inactive_ifree - 85.63% xfs_read_agi - 85.61% xfs_trans_read_buf_map - 85.59% xfs_buf_read_map - xfs_buf_get_map - 85.55% xfs_buf_find - 72.87% _raw_spin_lock - do_raw_spin_lock 71.86% __pv_queued_spin_lock_slowpath - 8.74% xfs_buf_rele - 7.88% _raw_spin_lock - 7.88% do_raw_spin_lock 7.63% __pv_queued_spin_lock_slowpath - 1.70% xfs_buf_trylock - 1.68% down_trylock - 1.41% _raw_spin_lock_irqsave - 1.39% do_raw_spin_lock __pv_queued_spin_lock_slowpath - 0.76% _raw_spin_unlock 0.75% do_raw_spin_unlock This is basically hammering the pag->pag_buf_lock from lots of CPUs doing trylocks at the same time. Most of the buffer trylock operations ultimately fail after we've done the lookup, so we're really hammering the buf hash lock whilst making no progress. We can also see significant spinlock traffic on the same lock just under normal operation when lots of tasks are accessing metadata from the same AG, so let's avoid all this by converting the lookup fast path to leverages the rhashtable's ability to do rcu protected lookups. We avoid races with the buffer release path by using atomic_inc_not_zero() on the buffer hold count. Any buffer that is in the LRU will have a non-zero count, thereby allowing the lockless fast path to be taken in most cache hit situations. If the buffer hold count is zero, then it is likely going through the release path so in that case we fall back to the existing lookup miss slow path. The slow path will then do an atomic lookup and insert under the buffer hash lock and hence serialise correctly against buffer release freeing the buffer. The use of rcu protected lookups means that buffer handles now need to be freed by RCU callbacks (same as inodes). We still free the buffer pages before the RCU callback - we won't be trying to access them at all on a buffer that has zero references - but we need the buffer handle itself to be present for the entire rcu protected read side to detect a zero hold count correctly. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2022-07-14 02:05:07 +00:00
static void
xfs_buf_free_callback(
struct callback_head *cb)
{
struct xfs_buf *bp = container_of(cb, struct xfs_buf, b_rcu);
xfs_buf_free_maps(bp);
kmem_cache_free(xfs_buf_cache, bp);
}
static void
xfs_buf_free(
struct xfs_buf *bp)
{
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_free(bp, _RET_IP_);
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
ASSERT(list_empty(&bp->b_lru));
if (bp->b_flags & _XBF_PAGES)
xfs_buf_free_pages(bp);
else if (bp->b_flags & _XBF_KMEM)
2011-03-25 22:16:45 +00:00
kmem_free(bp->b_addr);
xfs: lockless buffer lookup Now that we have a standalone fast path for buffer lookup, we can easily convert it to use rcu lookups. When we continually hammer the buffer cache with trylock lookups, we end up with a huge amount of lock contention on the per-ag buffer hash locks: - 92.71% 0.05% [kernel] [k] xfs_inodegc_worker - 92.67% xfs_inodegc_worker - 92.13% xfs_inode_unlink - 91.52% xfs_inactive_ifree - 85.63% xfs_read_agi - 85.61% xfs_trans_read_buf_map - 85.59% xfs_buf_read_map - xfs_buf_get_map - 85.55% xfs_buf_find - 72.87% _raw_spin_lock - do_raw_spin_lock 71.86% __pv_queued_spin_lock_slowpath - 8.74% xfs_buf_rele - 7.88% _raw_spin_lock - 7.88% do_raw_spin_lock 7.63% __pv_queued_spin_lock_slowpath - 1.70% xfs_buf_trylock - 1.68% down_trylock - 1.41% _raw_spin_lock_irqsave - 1.39% do_raw_spin_lock __pv_queued_spin_lock_slowpath - 0.76% _raw_spin_unlock 0.75% do_raw_spin_unlock This is basically hammering the pag->pag_buf_lock from lots of CPUs doing trylocks at the same time. Most of the buffer trylock operations ultimately fail after we've done the lookup, so we're really hammering the buf hash lock whilst making no progress. We can also see significant spinlock traffic on the same lock just under normal operation when lots of tasks are accessing metadata from the same AG, so let's avoid all this by converting the lookup fast path to leverages the rhashtable's ability to do rcu protected lookups. We avoid races with the buffer release path by using atomic_inc_not_zero() on the buffer hold count. Any buffer that is in the LRU will have a non-zero count, thereby allowing the lockless fast path to be taken in most cache hit situations. If the buffer hold count is zero, then it is likely going through the release path so in that case we fall back to the existing lookup miss slow path. The slow path will then do an atomic lookup and insert under the buffer hash lock and hence serialise correctly against buffer release freeing the buffer. The use of rcu protected lookups means that buffer handles now need to be freed by RCU callbacks (same as inodes). We still free the buffer pages before the RCU callback - we won't be trying to access them at all on a buffer that has zero references - but we need the buffer handle itself to be present for the entire rcu protected read side to detect a zero hold count correctly. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2022-07-14 02:05:07 +00:00
call_rcu(&bp->b_rcu, xfs_buf_free_callback);
}
static int
xfs_buf_alloc_kmem(
struct xfs_buf *bp,
xfs_buf_flags_t flags)
{
xfs_km_flags_t kmflag_mask = KM_NOFS;
size_t size = BBTOB(bp->b_length);
/* Assure zeroed buffer for non-read cases. */
if (!(flags & XBF_READ))
kmflag_mask |= KM_ZERO;
bp->b_addr = kmem_alloc(size, kmflag_mask);
if (!bp->b_addr)
return -ENOMEM;
2011-03-25 22:16:45 +00:00
if (((unsigned long)(bp->b_addr + size - 1) & PAGE_MASK) !=
((unsigned long)bp->b_addr & PAGE_MASK)) {
/* b_addr spans two pages - use alloc_page instead */
kmem_free(bp->b_addr);
bp->b_addr = NULL;
return -ENOMEM;
2011-03-25 22:16:45 +00:00
}
bp->b_offset = offset_in_page(bp->b_addr);
bp->b_pages = bp->b_page_array;
bp->b_pages[0] = kmem_to_page(bp->b_addr);
bp->b_page_count = 1;
bp->b_flags |= _XBF_KMEM;
return 0;
}
static int
xfs_buf_alloc_pages(
struct xfs_buf *bp,
xfs_buf_flags_t flags)
{
gfp_t gfp_mask = __GFP_NOWARN;
long filled = 0;
if (flags & XBF_READ_AHEAD)
gfp_mask |= __GFP_NORETRY;
else
gfp_mask |= GFP_NOFS;
/* Make sure that we have a page list */
bp->b_page_count = DIV_ROUND_UP(BBTOB(bp->b_length), PAGE_SIZE);
if (bp->b_page_count <= XB_PAGES) {
bp->b_pages = bp->b_page_array;
} else {
bp->b_pages = kzalloc(sizeof(struct page *) * bp->b_page_count,
gfp_mask);
if (!bp->b_pages)
return -ENOMEM;
}
bp->b_flags |= _XBF_PAGES;
/* Assure zeroed buffer for non-read cases. */
if (!(flags & XBF_READ))
gfp_mask |= __GFP_ZERO;
2011-03-25 22:16:45 +00:00
/*
* Bulk filling of pages can take multiple calls. Not filling the entire
* array is not an allocation failure, so don't back off if we get at
* least one extra page.
*/
for (;;) {
long last = filled;
filled = alloc_pages_bulk_array(gfp_mask, bp->b_page_count,
bp->b_pages);
if (filled == bp->b_page_count) {
XFS_STATS_INC(bp->b_mount, xb_page_found);
break;
}
if (filled != last)
continue;
if (flags & XBF_READ_AHEAD) {
xfs_buf_free_pages(bp);
return -ENOMEM;
}
XFS_STATS_INC(bp->b_mount, xb_page_retries);
mm: introduce memalloc_retry_wait() Various places in the kernel - largely in filesystems - respond to a memory allocation failure by looping around and re-trying. Some of these cannot conveniently use __GFP_NOFAIL, for reasons such as: - a GFP_ATOMIC allocation, which __GFP_NOFAIL doesn't work on - a need to check for the process being signalled between failures - the possibility that other recovery actions could be performed - the allocation is quite deep in support code, and passing down an extra flag to say if __GFP_NOFAIL is wanted would be clumsy. Many of these currently use congestion_wait() which (in almost all cases) simply waits the given timeout - congestion isn't tracked for most devices. It isn't clear what the best delay is for loops, but it is clear that the various filesystems shouldn't be responsible for choosing a timeout. This patch introduces memalloc_retry_wait() with takes on that responsibility. Code that wants to retry a memory allocation can call this function passing the GFP flags that were used. It will wait however is appropriate. For now, it only considers __GFP_NORETRY and whatever gfpflags_allow_blocking() tests. If blocking is allowed without __GFP_NORETRY, then alloc_page either made some reclaim progress, or waited for a while, before failing. So there is no need for much further waiting. memalloc_retry_wait() will wait until the current jiffie ends. If this condition is not met, then alloc_page() won't have waited much if at all. In that case memalloc_retry_wait() waits about 200ms. This is the delay that most current loops uses. linux/sched/mm.h needs to be included in some files now, but linux/backing-dev.h does not. Link: https://lkml.kernel.org/r/163754371968.13692.1277530886009912421@noble.neil.brown.name Signed-off-by: NeilBrown <neilb@suse.de> Cc: Dave Chinner <david@fromorbit.com> Cc: Michal Hocko <mhocko@suse.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Chao Yu <chao@kernel.org> Cc: Darrick J. Wong <djwong@kernel.org> Cc: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2022-01-14 22:07:14 +00:00
memalloc_retry_wait(gfp_mask);
}
2011-03-25 22:16:45 +00:00
return 0;
}
/*
* Map buffer into kernel address-space if necessary.
*/
STATIC int
_xfs_buf_map_pages(
struct xfs_buf *bp,
2022-04-20 22:44:59 +00:00
xfs_buf_flags_t flags)
{
2011-03-25 22:16:45 +00:00
ASSERT(bp->b_flags & _XBF_PAGES);
if (bp->b_page_count == 1) {
2011-03-25 22:16:45 +00:00
/* A single page buffer is always mappable */
bp->b_addr = page_address(bp->b_pages[0]);
} else if (flags & XBF_UNMAPPED) {
bp->b_addr = NULL;
} else {
int retried = 0;
unsigned nofs_flag;
xfs: use NOIO contexts for vm_map_ram When we map pages in the buffer cache, we can do so in GFP_NOFS contexts. However, the vmap interfaces do not provide any method of communicating this information to memory reclaim, and hence we get lockdep complaining about it regularly and occassionally see hangs that may be vmap related reclaim deadlocks. We can also see these same problems from anywhere where we use vmalloc for a large buffer (e.g. attribute code) inside a transaction context. A typical lockdep report shows up as a reclaim state warning like so: [14046.101458] ================================= [14046.102850] [ INFO: inconsistent lock state ] [14046.102850] 3.14.0-rc4+ #2 Not tainted [14046.102850] --------------------------------- [14046.102850] inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage. [14046.102850] kswapd0/14 [HC0[0]:SC0[0]:HE1:SE1] takes: [14046.102850] (&xfs_dir_ilock_class){++++?+}, at: [<791a04bb>] xfs_ilock+0xff/0x16a [14046.102850] {RECLAIM_FS-ON-W} state was registered at: [14046.102850] [<7904cdb1>] mark_held_locks+0x81/0xe7 [14046.102850] [<7904d390>] lockdep_trace_alloc+0x5c/0xb4 [14046.102850] [<790c2c28>] kmem_cache_alloc_trace+0x2b/0x11e [14046.102850] [<790ba7f4>] vm_map_ram+0x119/0x3e6 [14046.102850] [<7914e124>] _xfs_buf_map_pages+0x5b/0xcf [14046.102850] [<7914ed74>] xfs_buf_get_map+0x67/0x13f [14046.102850] [<7917506f>] xfs_attr_rmtval_set+0x396/0x4d5 [14046.102850] [<7916e8bb>] xfs_attr_leaf_addname+0x18f/0x37d [14046.102850] [<7916ed9e>] xfs_attr_set_int+0x2f5/0x3e8 [14046.102850] [<7916eefc>] xfs_attr_set+0x6b/0x74 [14046.102850] [<79168355>] xfs_xattr_set+0x61/0x81 [14046.102850] [<790e5b10>] generic_setxattr+0x59/0x68 [14046.102850] [<790e4c06>] __vfs_setxattr_noperm+0x58/0xce [14046.102850] [<790e4d0a>] vfs_setxattr+0x8e/0x92 [14046.102850] [<790e4ddd>] setxattr+0xcf/0x159 [14046.102850] [<790e5423>] SyS_lsetxattr+0x88/0xbb [14046.102850] [<79268438>] sysenter_do_call+0x12/0x36 Now, we can't completely remove these traces - mainly because vm_map_ram() will do GFP_KERNEL allocation and that generates the above warning before we get into the reclaim code, but we can turn them all into false positive warnings. To do that, use the method that DM and other IO context code uses to avoid this problem: there is a process flag to tell memory reclaim not to do IO that we can set appropriately. That prevents GFP_KERNEL context reclaim being done from deep inside the vmalloc code in places we can't directly pass a GFP_NOFS context to. That interface has a pair of wrapper functions: memalloc_noio_save() and memalloc_noio_restore(). Adding them around vm_map_ram and the vzalloc call in kmem_alloc_large() will prevent deadlocks and most lockdep reports for this issue. Also, convert the vzalloc() call in kmem_alloc_large() to use __vmalloc() so that we can pass the correct gfp context to the data page allocation routine inside __vmalloc() so that it is clear that GFP_NOFS context is important to this vmalloc call. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-03-07 05:19:14 +00:00
/*
* vm_map_ram() will allocate auxiliary structures (e.g.
xfs: use NOIO contexts for vm_map_ram When we map pages in the buffer cache, we can do so in GFP_NOFS contexts. However, the vmap interfaces do not provide any method of communicating this information to memory reclaim, and hence we get lockdep complaining about it regularly and occassionally see hangs that may be vmap related reclaim deadlocks. We can also see these same problems from anywhere where we use vmalloc for a large buffer (e.g. attribute code) inside a transaction context. A typical lockdep report shows up as a reclaim state warning like so: [14046.101458] ================================= [14046.102850] [ INFO: inconsistent lock state ] [14046.102850] 3.14.0-rc4+ #2 Not tainted [14046.102850] --------------------------------- [14046.102850] inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage. [14046.102850] kswapd0/14 [HC0[0]:SC0[0]:HE1:SE1] takes: [14046.102850] (&xfs_dir_ilock_class){++++?+}, at: [<791a04bb>] xfs_ilock+0xff/0x16a [14046.102850] {RECLAIM_FS-ON-W} state was registered at: [14046.102850] [<7904cdb1>] mark_held_locks+0x81/0xe7 [14046.102850] [<7904d390>] lockdep_trace_alloc+0x5c/0xb4 [14046.102850] [<790c2c28>] kmem_cache_alloc_trace+0x2b/0x11e [14046.102850] [<790ba7f4>] vm_map_ram+0x119/0x3e6 [14046.102850] [<7914e124>] _xfs_buf_map_pages+0x5b/0xcf [14046.102850] [<7914ed74>] xfs_buf_get_map+0x67/0x13f [14046.102850] [<7917506f>] xfs_attr_rmtval_set+0x396/0x4d5 [14046.102850] [<7916e8bb>] xfs_attr_leaf_addname+0x18f/0x37d [14046.102850] [<7916ed9e>] xfs_attr_set_int+0x2f5/0x3e8 [14046.102850] [<7916eefc>] xfs_attr_set+0x6b/0x74 [14046.102850] [<79168355>] xfs_xattr_set+0x61/0x81 [14046.102850] [<790e5b10>] generic_setxattr+0x59/0x68 [14046.102850] [<790e4c06>] __vfs_setxattr_noperm+0x58/0xce [14046.102850] [<790e4d0a>] vfs_setxattr+0x8e/0x92 [14046.102850] [<790e4ddd>] setxattr+0xcf/0x159 [14046.102850] [<790e5423>] SyS_lsetxattr+0x88/0xbb [14046.102850] [<79268438>] sysenter_do_call+0x12/0x36 Now, we can't completely remove these traces - mainly because vm_map_ram() will do GFP_KERNEL allocation and that generates the above warning before we get into the reclaim code, but we can turn them all into false positive warnings. To do that, use the method that DM and other IO context code uses to avoid this problem: there is a process flag to tell memory reclaim not to do IO that we can set appropriately. That prevents GFP_KERNEL context reclaim being done from deep inside the vmalloc code in places we can't directly pass a GFP_NOFS context to. That interface has a pair of wrapper functions: memalloc_noio_save() and memalloc_noio_restore(). Adding them around vm_map_ram and the vzalloc call in kmem_alloc_large() will prevent deadlocks and most lockdep reports for this issue. Also, convert the vzalloc() call in kmem_alloc_large() to use __vmalloc() so that we can pass the correct gfp context to the data page allocation routine inside __vmalloc() so that it is clear that GFP_NOFS context is important to this vmalloc call. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-03-07 05:19:14 +00:00
* pagetables) with GFP_KERNEL, yet we are likely to be under
* GFP_NOFS context here. Hence we need to tell memory reclaim
* that we are in such a context via PF_MEMALLOC_NOFS to prevent
xfs: use NOIO contexts for vm_map_ram When we map pages in the buffer cache, we can do so in GFP_NOFS contexts. However, the vmap interfaces do not provide any method of communicating this information to memory reclaim, and hence we get lockdep complaining about it regularly and occassionally see hangs that may be vmap related reclaim deadlocks. We can also see these same problems from anywhere where we use vmalloc for a large buffer (e.g. attribute code) inside a transaction context. A typical lockdep report shows up as a reclaim state warning like so: [14046.101458] ================================= [14046.102850] [ INFO: inconsistent lock state ] [14046.102850] 3.14.0-rc4+ #2 Not tainted [14046.102850] --------------------------------- [14046.102850] inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage. [14046.102850] kswapd0/14 [HC0[0]:SC0[0]:HE1:SE1] takes: [14046.102850] (&xfs_dir_ilock_class){++++?+}, at: [<791a04bb>] xfs_ilock+0xff/0x16a [14046.102850] {RECLAIM_FS-ON-W} state was registered at: [14046.102850] [<7904cdb1>] mark_held_locks+0x81/0xe7 [14046.102850] [<7904d390>] lockdep_trace_alloc+0x5c/0xb4 [14046.102850] [<790c2c28>] kmem_cache_alloc_trace+0x2b/0x11e [14046.102850] [<790ba7f4>] vm_map_ram+0x119/0x3e6 [14046.102850] [<7914e124>] _xfs_buf_map_pages+0x5b/0xcf [14046.102850] [<7914ed74>] xfs_buf_get_map+0x67/0x13f [14046.102850] [<7917506f>] xfs_attr_rmtval_set+0x396/0x4d5 [14046.102850] [<7916e8bb>] xfs_attr_leaf_addname+0x18f/0x37d [14046.102850] [<7916ed9e>] xfs_attr_set_int+0x2f5/0x3e8 [14046.102850] [<7916eefc>] xfs_attr_set+0x6b/0x74 [14046.102850] [<79168355>] xfs_xattr_set+0x61/0x81 [14046.102850] [<790e5b10>] generic_setxattr+0x59/0x68 [14046.102850] [<790e4c06>] __vfs_setxattr_noperm+0x58/0xce [14046.102850] [<790e4d0a>] vfs_setxattr+0x8e/0x92 [14046.102850] [<790e4ddd>] setxattr+0xcf/0x159 [14046.102850] [<790e5423>] SyS_lsetxattr+0x88/0xbb [14046.102850] [<79268438>] sysenter_do_call+0x12/0x36 Now, we can't completely remove these traces - mainly because vm_map_ram() will do GFP_KERNEL allocation and that generates the above warning before we get into the reclaim code, but we can turn them all into false positive warnings. To do that, use the method that DM and other IO context code uses to avoid this problem: there is a process flag to tell memory reclaim not to do IO that we can set appropriately. That prevents GFP_KERNEL context reclaim being done from deep inside the vmalloc code in places we can't directly pass a GFP_NOFS context to. That interface has a pair of wrapper functions: memalloc_noio_save() and memalloc_noio_restore(). Adding them around vm_map_ram and the vzalloc call in kmem_alloc_large() will prevent deadlocks and most lockdep reports for this issue. Also, convert the vzalloc() call in kmem_alloc_large() to use __vmalloc() so that we can pass the correct gfp context to the data page allocation routine inside __vmalloc() so that it is clear that GFP_NOFS context is important to this vmalloc call. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-03-07 05:19:14 +00:00
* memory reclaim re-entering the filesystem here and
* potentially deadlocking.
*/
nofs_flag = memalloc_nofs_save();
do {
bp->b_addr = vm_map_ram(bp->b_pages, bp->b_page_count,
2020-06-02 04:51:27 +00:00
-1);
if (bp->b_addr)
break;
vm_unmap_aliases();
} while (retried++ <= 1);
memalloc_nofs_restore(nofs_flag);
if (!bp->b_addr)
return -ENOMEM;
}
return 0;
}
/*
* Finding and Reading Buffers
*/
static int
_xfs_buf_obj_cmp(
struct rhashtable_compare_arg *arg,
const void *obj)
{
const struct xfs_buf_map *map = arg->key;
const struct xfs_buf *bp = obj;
/*
* The key hashing in the lookup path depends on the key being the
* first element of the compare_arg, make sure to assert this.
*/
BUILD_BUG_ON(offsetof(struct xfs_buf_map, bm_bn) != 0);
if (bp->b_rhash_key != map->bm_bn)
return 1;
if (unlikely(bp->b_length != map->bm_len)) {
/*
* found a block number match. If the range doesn't
* match, the only way this is allowed is if the buffer
* in the cache is stale and the transaction that made
* it stale has not yet committed. i.e. we are
* reallocating a busy extent. Skip this buffer and
* continue searching for an exact match.
*/
ASSERT(bp->b_flags & XBF_STALE);
return 1;
}
return 0;
}
static const struct rhashtable_params xfs_buf_hash_params = {
.min_size = 32, /* empty AGs have minimal footprint */
.nelem_hint = 16,
.key_len = sizeof(xfs_daddr_t),
.key_offset = offsetof(struct xfs_buf, b_rhash_key),
.head_offset = offsetof(struct xfs_buf, b_rhash_head),
.automatic_shrinking = true,
.obj_cmpfn = _xfs_buf_obj_cmp,
};
int
xfs_buf_hash_init(
struct xfs_perag *pag)
{
spin_lock_init(&pag->pag_buf_lock);
return rhashtable_init(&pag->pag_buf_hash, &xfs_buf_hash_params);
}
void
xfs_buf_hash_destroy(
struct xfs_perag *pag)
{
rhashtable_destroy(&pag->pag_buf_hash);
}
static int
xfs_buf_map_verify(
struct xfs_buftarg *btp,
struct xfs_buf_map *map)
{
xfs_daddr_t eofs;
/* Check for IOs smaller than the sector size / not sector aligned */
ASSERT(!(BBTOB(map->bm_len) < btp->bt_meta_sectorsize));
ASSERT(!(BBTOB(map->bm_bn) & (xfs_off_t)btp->bt_meta_sectormask));
/*
* Corrupted block numbers can get through to here, unfortunately, so we
* have to check that the buffer falls within the filesystem bounds.
*/
eofs = XFS_FSB_TO_BB(btp->bt_mount, btp->bt_mount->m_sb.sb_dblocks);
if (map->bm_bn < 0 || map->bm_bn >= eofs) {
xfs_alert(btp->bt_mount,
"%s: daddr 0x%llx out of range, EOFS 0x%llx",
__func__, map->bm_bn, eofs);
xfs: rework remote attr CRCs Note: this changes the on-disk remote attribute format. I assert that this is OK to do as CRCs are marked experimental and the first kernel it is included in has not yet reached release yet. Further, the userspace utilities are still evolving and so anyone using this stuff right now is a developer or tester using volatile filesystems for testing this feature. Hence changing the format right now to save longer term pain is the right thing to do. The fundamental change is to move from a header per extent in the attribute to a header per filesytem block in the attribute. This means there are more header blocks and the parsing of the attribute data is slightly more complex, but it has the advantage that we always know the size of the attribute on disk based on the length of the data it contains. This is where the header-per-extent method has problems. We don't know the size of the attribute on disk without first knowing how many extents are used to hold it. And we can't tell from a mapping lookup, either, because remote attributes can be allocated contiguously with other attribute blocks and so there is no obvious way of determining the actual size of the atribute on disk short of walking and mapping buffers. The problem with this approach is that if we map a buffer incorrectly (e.g. we make the last buffer for the attribute data too long), we then get buffer cache lookup failure when we map it correctly. i.e. we get a size mismatch on lookup. This is not necessarily fatal, but it's a cache coherency problem that can lead to returning the wrong data to userspace or writing the wrong data to disk. And debug kernels will assert fail if this occurs. I found lots of niggly little problems trying to fix this issue on a 4k block size filesystem, finally getting it to pass with lots of fixes. The thing is, 1024 byte filesystems still failed, and it was getting really complex handling all the corner cases that were showing up. And there were clearly more that I hadn't found yet. It is complex, fragile code, and if we don't fix it now, it will be complex, fragile code forever more. Hence the simple fix is to add a header to each filesystem block. This gives us the same relationship between the attribute data length and the number of blocks on disk as we have without CRCs - it's a linear mapping and doesn't require us to guess anything. It is simple to implement, too - the remote block count calculated at lookup time can be used by the remote attribute set/get/remove code without modification for both CRC and non-CRC filesystems. The world becomes sane again. Because the copy-in and copy-out now need to iterate over each filesystem block, I moved them into helper functions so we separate the block mapping and buffer manupulations from the attribute data and CRC header manipulations. The code becomes much clearer as a result, and it is a lot easier to understand and debug. It also appears to be much more robust - once it worked on 4k block size filesystems, it has worked without failure on 1k block size filesystems, too. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Ben Myers <bpm@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com> (cherry picked from commit ad1858d77771172e08016890f0eb2faedec3ecee)
2013-05-21 08:02:08 +00:00
WARN_ON(1);
return -EFSCORRUPTED;
}
return 0;
}
static int
xfs_buf_find_lock(
struct xfs_buf *bp,
xfs_buf_flags_t flags)
{
if (flags & XBF_TRYLOCK) {
if (!xfs_buf_trylock(bp)) {
XFS_STATS_INC(bp->b_mount, xb_busy_locked);
return -EAGAIN;
}
} else {
xfs_buf_lock(bp);
XFS_STATS_INC(bp->b_mount, xb_get_locked_waited);
}
2011-03-25 22:16:45 +00:00
/*
* if the buffer is stale, clear all the external state associated with
* it. We need to keep flags such as how we allocated the buffer memory
* intact here.
*/
if (bp->b_flags & XBF_STALE) {
ASSERT((bp->b_flags & _XBF_DELWRI_Q) == 0);
bp->b_flags &= _XBF_KMEM | _XBF_PAGES;
bp->b_ops = NULL;
}
return 0;
}
static inline int
xfs_buf_lookup(
struct xfs_perag *pag,
struct xfs_buf_map *map,
xfs_buf_flags_t flags,
struct xfs_buf **bpp)
{
struct xfs_buf *bp;
int error;
xfs: lockless buffer lookup Now that we have a standalone fast path for buffer lookup, we can easily convert it to use rcu lookups. When we continually hammer the buffer cache with trylock lookups, we end up with a huge amount of lock contention on the per-ag buffer hash locks: - 92.71% 0.05% [kernel] [k] xfs_inodegc_worker - 92.67% xfs_inodegc_worker - 92.13% xfs_inode_unlink - 91.52% xfs_inactive_ifree - 85.63% xfs_read_agi - 85.61% xfs_trans_read_buf_map - 85.59% xfs_buf_read_map - xfs_buf_get_map - 85.55% xfs_buf_find - 72.87% _raw_spin_lock - do_raw_spin_lock 71.86% __pv_queued_spin_lock_slowpath - 8.74% xfs_buf_rele - 7.88% _raw_spin_lock - 7.88% do_raw_spin_lock 7.63% __pv_queued_spin_lock_slowpath - 1.70% xfs_buf_trylock - 1.68% down_trylock - 1.41% _raw_spin_lock_irqsave - 1.39% do_raw_spin_lock __pv_queued_spin_lock_slowpath - 0.76% _raw_spin_unlock 0.75% do_raw_spin_unlock This is basically hammering the pag->pag_buf_lock from lots of CPUs doing trylocks at the same time. Most of the buffer trylock operations ultimately fail after we've done the lookup, so we're really hammering the buf hash lock whilst making no progress. We can also see significant spinlock traffic on the same lock just under normal operation when lots of tasks are accessing metadata from the same AG, so let's avoid all this by converting the lookup fast path to leverages the rhashtable's ability to do rcu protected lookups. We avoid races with the buffer release path by using atomic_inc_not_zero() on the buffer hold count. Any buffer that is in the LRU will have a non-zero count, thereby allowing the lockless fast path to be taken in most cache hit situations. If the buffer hold count is zero, then it is likely going through the release path so in that case we fall back to the existing lookup miss slow path. The slow path will then do an atomic lookup and insert under the buffer hash lock and hence serialise correctly against buffer release freeing the buffer. The use of rcu protected lookups means that buffer handles now need to be freed by RCU callbacks (same as inodes). We still free the buffer pages before the RCU callback - we won't be trying to access them at all on a buffer that has zero references - but we need the buffer handle itself to be present for the entire rcu protected read side to detect a zero hold count correctly. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2022-07-14 02:05:07 +00:00
rcu_read_lock();
bp = rhashtable_lookup(&pag->pag_buf_hash, map, xfs_buf_hash_params);
xfs: lockless buffer lookup Now that we have a standalone fast path for buffer lookup, we can easily convert it to use rcu lookups. When we continually hammer the buffer cache with trylock lookups, we end up with a huge amount of lock contention on the per-ag buffer hash locks: - 92.71% 0.05% [kernel] [k] xfs_inodegc_worker - 92.67% xfs_inodegc_worker - 92.13% xfs_inode_unlink - 91.52% xfs_inactive_ifree - 85.63% xfs_read_agi - 85.61% xfs_trans_read_buf_map - 85.59% xfs_buf_read_map - xfs_buf_get_map - 85.55% xfs_buf_find - 72.87% _raw_spin_lock - do_raw_spin_lock 71.86% __pv_queued_spin_lock_slowpath - 8.74% xfs_buf_rele - 7.88% _raw_spin_lock - 7.88% do_raw_spin_lock 7.63% __pv_queued_spin_lock_slowpath - 1.70% xfs_buf_trylock - 1.68% down_trylock - 1.41% _raw_spin_lock_irqsave - 1.39% do_raw_spin_lock __pv_queued_spin_lock_slowpath - 0.76% _raw_spin_unlock 0.75% do_raw_spin_unlock This is basically hammering the pag->pag_buf_lock from lots of CPUs doing trylocks at the same time. Most of the buffer trylock operations ultimately fail after we've done the lookup, so we're really hammering the buf hash lock whilst making no progress. We can also see significant spinlock traffic on the same lock just under normal operation when lots of tasks are accessing metadata from the same AG, so let's avoid all this by converting the lookup fast path to leverages the rhashtable's ability to do rcu protected lookups. We avoid races with the buffer release path by using atomic_inc_not_zero() on the buffer hold count. Any buffer that is in the LRU will have a non-zero count, thereby allowing the lockless fast path to be taken in most cache hit situations. If the buffer hold count is zero, then it is likely going through the release path so in that case we fall back to the existing lookup miss slow path. The slow path will then do an atomic lookup and insert under the buffer hash lock and hence serialise correctly against buffer release freeing the buffer. The use of rcu protected lookups means that buffer handles now need to be freed by RCU callbacks (same as inodes). We still free the buffer pages before the RCU callback - we won't be trying to access them at all on a buffer that has zero references - but we need the buffer handle itself to be present for the entire rcu protected read side to detect a zero hold count correctly. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2022-07-14 02:05:07 +00:00
if (!bp || !atomic_inc_not_zero(&bp->b_hold)) {
rcu_read_unlock();
return -ENOENT;
}
xfs: lockless buffer lookup Now that we have a standalone fast path for buffer lookup, we can easily convert it to use rcu lookups. When we continually hammer the buffer cache with trylock lookups, we end up with a huge amount of lock contention on the per-ag buffer hash locks: - 92.71% 0.05% [kernel] [k] xfs_inodegc_worker - 92.67% xfs_inodegc_worker - 92.13% xfs_inode_unlink - 91.52% xfs_inactive_ifree - 85.63% xfs_read_agi - 85.61% xfs_trans_read_buf_map - 85.59% xfs_buf_read_map - xfs_buf_get_map - 85.55% xfs_buf_find - 72.87% _raw_spin_lock - do_raw_spin_lock 71.86% __pv_queued_spin_lock_slowpath - 8.74% xfs_buf_rele - 7.88% _raw_spin_lock - 7.88% do_raw_spin_lock 7.63% __pv_queued_spin_lock_slowpath - 1.70% xfs_buf_trylock - 1.68% down_trylock - 1.41% _raw_spin_lock_irqsave - 1.39% do_raw_spin_lock __pv_queued_spin_lock_slowpath - 0.76% _raw_spin_unlock 0.75% do_raw_spin_unlock This is basically hammering the pag->pag_buf_lock from lots of CPUs doing trylocks at the same time. Most of the buffer trylock operations ultimately fail after we've done the lookup, so we're really hammering the buf hash lock whilst making no progress. We can also see significant spinlock traffic on the same lock just under normal operation when lots of tasks are accessing metadata from the same AG, so let's avoid all this by converting the lookup fast path to leverages the rhashtable's ability to do rcu protected lookups. We avoid races with the buffer release path by using atomic_inc_not_zero() on the buffer hold count. Any buffer that is in the LRU will have a non-zero count, thereby allowing the lockless fast path to be taken in most cache hit situations. If the buffer hold count is zero, then it is likely going through the release path so in that case we fall back to the existing lookup miss slow path. The slow path will then do an atomic lookup and insert under the buffer hash lock and hence serialise correctly against buffer release freeing the buffer. The use of rcu protected lookups means that buffer handles now need to be freed by RCU callbacks (same as inodes). We still free the buffer pages before the RCU callback - we won't be trying to access them at all on a buffer that has zero references - but we need the buffer handle itself to be present for the entire rcu protected read side to detect a zero hold count correctly. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2022-07-14 02:05:07 +00:00
rcu_read_unlock();
error = xfs_buf_find_lock(bp, flags);
if (error) {
xfs_buf_rele(bp);
return error;
}
trace_xfs_buf_find(bp, flags, _RET_IP_);
*bpp = bp;
return 0;
}
/*
* Insert the new_bp into the hash table. This consumes the perag reference
* taken for the lookup regardless of the result of the insert.
*/
static int
xfs_buf_find_insert(
struct xfs_buftarg *btp,
struct xfs_perag *pag,
struct xfs_buf_map *cmap,
struct xfs_buf_map *map,
int nmaps,
xfs_buf_flags_t flags,
struct xfs_buf **bpp)
{
xfs: Don't allocate new buffers on every call to _xfs_buf_find Stats show that for an 8-way unlink @ ~80,000 unlinks/s we are doing ~1 million cache hit lookups to ~3000 buffer creates. That's almost 3 orders of magnitude more cahce hits than misses, so optimising for cache hits is quite important. In the cache hit case, we do not need to allocate a new buffer in case of a cache miss, so we are effectively hitting the allocator for no good reason for vast the majority of calls to _xfs_buf_find. 8-way create workloads are showing similar cache hit/miss ratios. The result is profiles that look like this: samples pcnt function DSO _______ _____ _______________________________ _________________ 1036.00 10.0% _xfs_buf_find [kernel.kallsyms] 582.00 5.6% kmem_cache_alloc [kernel.kallsyms] 519.00 5.0% __memcpy [kernel.kallsyms] 468.00 4.5% __ticket_spin_lock [kernel.kallsyms] 388.00 3.7% kmem_cache_free [kernel.kallsyms] 331.00 3.2% xfs_log_commit_cil [kernel.kallsyms] Further, there is a fair bit of work involved in initialising a new buffer once a cache miss has occurred and we currently do that under the rbtree spinlock. That increases spinlock hold time on what are heavily used trees. To fix this, remove the initialisation of the buffer from _xfs_buf_find() and only allocate the new buffer once we've had a cache miss. Initialise the buffer immediately after allocating it in xfs_buf_get, too, so that is it ready for insert if we get another cache miss after allocation. This minimises lock hold time and avoids unnecessary allocator churn. The resulting profiles look like: samples pcnt function DSO _______ _____ ___________________________ _________________ 8111.00 9.1% _xfs_buf_find [kernel.kallsyms] 4380.00 4.9% __memcpy [kernel.kallsyms] 4341.00 4.8% __ticket_spin_lock [kernel.kallsyms] 3401.00 3.8% kmem_cache_alloc [kernel.kallsyms] 2856.00 3.2% xfs_log_commit_cil [kernel.kallsyms] 2625.00 2.9% __kmalloc [kernel.kallsyms] 2380.00 2.7% kfree [kernel.kallsyms] 2016.00 2.3% kmem_cache_free [kernel.kallsyms] Showing a significant reduction in time spent doing allocation and freeing from slabs (kmem_cache_alloc and kmem_cache_free). Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2011-09-30 04:45:02 +00:00
struct xfs_buf *new_bp;
struct xfs_buf *bp;
int error;
error = _xfs_buf_alloc(btp, map, nmaps, flags, &new_bp);
if (error)
goto out_drop_pag;
/*
* For buffers that fit entirely within a single page, first attempt to
* allocate the memory from the heap to minimise memory usage. If we
* can't get heap memory for these small buffers, we fall back to using
* the page allocator.
*/
if (BBTOB(new_bp->b_length) >= PAGE_SIZE ||
xfs_buf_alloc_kmem(new_bp, flags) < 0) {
error = xfs_buf_alloc_pages(new_bp, flags);
if (error)
goto out_free_buf;
}
spin_lock(&pag->pag_buf_lock);
bp = rhashtable_lookup_get_insert_fast(&pag->pag_buf_hash,
&new_bp->b_rhash_head, xfs_buf_hash_params);
if (IS_ERR(bp)) {
error = PTR_ERR(bp);
spin_unlock(&pag->pag_buf_lock);
goto out_free_buf;
}
if (bp) {
/* found an existing buffer */
atomic_inc(&bp->b_hold);
spin_unlock(&pag->pag_buf_lock);
error = xfs_buf_find_lock(bp, flags);
if (error)
xfs_buf_rele(bp);
else
*bpp = bp;
goto out_free_buf;
}
/* The new buffer keeps the perag reference until it is freed. */
new_bp->b_pag = pag;
spin_unlock(&pag->pag_buf_lock);
*bpp = new_bp;
return 0;
xfs: Don't allocate new buffers on every call to _xfs_buf_find Stats show that for an 8-way unlink @ ~80,000 unlinks/s we are doing ~1 million cache hit lookups to ~3000 buffer creates. That's almost 3 orders of magnitude more cahce hits than misses, so optimising for cache hits is quite important. In the cache hit case, we do not need to allocate a new buffer in case of a cache miss, so we are effectively hitting the allocator for no good reason for vast the majority of calls to _xfs_buf_find. 8-way create workloads are showing similar cache hit/miss ratios. The result is profiles that look like this: samples pcnt function DSO _______ _____ _______________________________ _________________ 1036.00 10.0% _xfs_buf_find [kernel.kallsyms] 582.00 5.6% kmem_cache_alloc [kernel.kallsyms] 519.00 5.0% __memcpy [kernel.kallsyms] 468.00 4.5% __ticket_spin_lock [kernel.kallsyms] 388.00 3.7% kmem_cache_free [kernel.kallsyms] 331.00 3.2% xfs_log_commit_cil [kernel.kallsyms] Further, there is a fair bit of work involved in initialising a new buffer once a cache miss has occurred and we currently do that under the rbtree spinlock. That increases spinlock hold time on what are heavily used trees. To fix this, remove the initialisation of the buffer from _xfs_buf_find() and only allocate the new buffer once we've had a cache miss. Initialise the buffer immediately after allocating it in xfs_buf_get, too, so that is it ready for insert if we get another cache miss after allocation. This minimises lock hold time and avoids unnecessary allocator churn. The resulting profiles look like: samples pcnt function DSO _______ _____ ___________________________ _________________ 8111.00 9.1% _xfs_buf_find [kernel.kallsyms] 4380.00 4.9% __memcpy [kernel.kallsyms] 4341.00 4.8% __ticket_spin_lock [kernel.kallsyms] 3401.00 3.8% kmem_cache_alloc [kernel.kallsyms] 2856.00 3.2% xfs_log_commit_cil [kernel.kallsyms] 2625.00 2.9% __kmalloc [kernel.kallsyms] 2380.00 2.7% kfree [kernel.kallsyms] 2016.00 2.3% kmem_cache_free [kernel.kallsyms] Showing a significant reduction in time spent doing allocation and freeing from slabs (kmem_cache_alloc and kmem_cache_free). Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2011-09-30 04:45:02 +00:00
out_free_buf:
xfs_buf_free(new_bp);
out_drop_pag:
xfs_perag_put(pag);
return error;
}
/*
xfs: Don't allocate new buffers on every call to _xfs_buf_find Stats show that for an 8-way unlink @ ~80,000 unlinks/s we are doing ~1 million cache hit lookups to ~3000 buffer creates. That's almost 3 orders of magnitude more cahce hits than misses, so optimising for cache hits is quite important. In the cache hit case, we do not need to allocate a new buffer in case of a cache miss, so we are effectively hitting the allocator for no good reason for vast the majority of calls to _xfs_buf_find. 8-way create workloads are showing similar cache hit/miss ratios. The result is profiles that look like this: samples pcnt function DSO _______ _____ _______________________________ _________________ 1036.00 10.0% _xfs_buf_find [kernel.kallsyms] 582.00 5.6% kmem_cache_alloc [kernel.kallsyms] 519.00 5.0% __memcpy [kernel.kallsyms] 468.00 4.5% __ticket_spin_lock [kernel.kallsyms] 388.00 3.7% kmem_cache_free [kernel.kallsyms] 331.00 3.2% xfs_log_commit_cil [kernel.kallsyms] Further, there is a fair bit of work involved in initialising a new buffer once a cache miss has occurred and we currently do that under the rbtree spinlock. That increases spinlock hold time on what are heavily used trees. To fix this, remove the initialisation of the buffer from _xfs_buf_find() and only allocate the new buffer once we've had a cache miss. Initialise the buffer immediately after allocating it in xfs_buf_get, too, so that is it ready for insert if we get another cache miss after allocation. This minimises lock hold time and avoids unnecessary allocator churn. The resulting profiles look like: samples pcnt function DSO _______ _____ ___________________________ _________________ 8111.00 9.1% _xfs_buf_find [kernel.kallsyms] 4380.00 4.9% __memcpy [kernel.kallsyms] 4341.00 4.8% __ticket_spin_lock [kernel.kallsyms] 3401.00 3.8% kmem_cache_alloc [kernel.kallsyms] 2856.00 3.2% xfs_log_commit_cil [kernel.kallsyms] 2625.00 2.9% __kmalloc [kernel.kallsyms] 2380.00 2.7% kfree [kernel.kallsyms] 2016.00 2.3% kmem_cache_free [kernel.kallsyms] Showing a significant reduction in time spent doing allocation and freeing from slabs (kmem_cache_alloc and kmem_cache_free). Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2011-09-30 04:45:02 +00:00
* Assembles a buffer covering the specified range. The code is optimised for
* cache hits, as metadata intensive workloads will see 3 orders of magnitude
* more hits than misses.
*/
int
xfs_buf_get_map(
struct xfs_buftarg *btp,
struct xfs_buf_map *map,
int nmaps,
xfs_buf_flags_t flags,
struct xfs_buf **bpp)
{
struct xfs_perag *pag;
struct xfs_buf *bp = NULL;
struct xfs_buf_map cmap = { .bm_bn = map[0].bm_bn };
int error;
int i;
for (i = 0; i < nmaps; i++)
cmap.bm_len += map[i].bm_len;
xfs: Don't allocate new buffers on every call to _xfs_buf_find Stats show that for an 8-way unlink @ ~80,000 unlinks/s we are doing ~1 million cache hit lookups to ~3000 buffer creates. That's almost 3 orders of magnitude more cahce hits than misses, so optimising for cache hits is quite important. In the cache hit case, we do not need to allocate a new buffer in case of a cache miss, so we are effectively hitting the allocator for no good reason for vast the majority of calls to _xfs_buf_find. 8-way create workloads are showing similar cache hit/miss ratios. The result is profiles that look like this: samples pcnt function DSO _______ _____ _______________________________ _________________ 1036.00 10.0% _xfs_buf_find [kernel.kallsyms] 582.00 5.6% kmem_cache_alloc [kernel.kallsyms] 519.00 5.0% __memcpy [kernel.kallsyms] 468.00 4.5% __ticket_spin_lock [kernel.kallsyms] 388.00 3.7% kmem_cache_free [kernel.kallsyms] 331.00 3.2% xfs_log_commit_cil [kernel.kallsyms] Further, there is a fair bit of work involved in initialising a new buffer once a cache miss has occurred and we currently do that under the rbtree spinlock. That increases spinlock hold time on what are heavily used trees. To fix this, remove the initialisation of the buffer from _xfs_buf_find() and only allocate the new buffer once we've had a cache miss. Initialise the buffer immediately after allocating it in xfs_buf_get, too, so that is it ready for insert if we get another cache miss after allocation. This minimises lock hold time and avoids unnecessary allocator churn. The resulting profiles look like: samples pcnt function DSO _______ _____ ___________________________ _________________ 8111.00 9.1% _xfs_buf_find [kernel.kallsyms] 4380.00 4.9% __memcpy [kernel.kallsyms] 4341.00 4.8% __ticket_spin_lock [kernel.kallsyms] 3401.00 3.8% kmem_cache_alloc [kernel.kallsyms] 2856.00 3.2% xfs_log_commit_cil [kernel.kallsyms] 2625.00 2.9% __kmalloc [kernel.kallsyms] 2380.00 2.7% kfree [kernel.kallsyms] 2016.00 2.3% kmem_cache_free [kernel.kallsyms] Showing a significant reduction in time spent doing allocation and freeing from slabs (kmem_cache_alloc and kmem_cache_free). Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2011-09-30 04:45:02 +00:00
error = xfs_buf_map_verify(btp, &cmap);
if (error)
return error;
pag = xfs_perag_get(btp->bt_mount,
xfs_daddr_to_agno(btp->bt_mount, cmap.bm_bn));
error = xfs_buf_lookup(pag, &cmap, flags, &bp);
if (error && error != -ENOENT)
goto out_put_perag;
/* cache hits always outnumber misses by at least 10:1 */
if (unlikely(!bp)) {
XFS_STATS_INC(btp->bt_mount, xb_miss_locked);
xfs: Don't allocate new buffers on every call to _xfs_buf_find Stats show that for an 8-way unlink @ ~80,000 unlinks/s we are doing ~1 million cache hit lookups to ~3000 buffer creates. That's almost 3 orders of magnitude more cahce hits than misses, so optimising for cache hits is quite important. In the cache hit case, we do not need to allocate a new buffer in case of a cache miss, so we are effectively hitting the allocator for no good reason for vast the majority of calls to _xfs_buf_find. 8-way create workloads are showing similar cache hit/miss ratios. The result is profiles that look like this: samples pcnt function DSO _______ _____ _______________________________ _________________ 1036.00 10.0% _xfs_buf_find [kernel.kallsyms] 582.00 5.6% kmem_cache_alloc [kernel.kallsyms] 519.00 5.0% __memcpy [kernel.kallsyms] 468.00 4.5% __ticket_spin_lock [kernel.kallsyms] 388.00 3.7% kmem_cache_free [kernel.kallsyms] 331.00 3.2% xfs_log_commit_cil [kernel.kallsyms] Further, there is a fair bit of work involved in initialising a new buffer once a cache miss has occurred and we currently do that under the rbtree spinlock. That increases spinlock hold time on what are heavily used trees. To fix this, remove the initialisation of the buffer from _xfs_buf_find() and only allocate the new buffer once we've had a cache miss. Initialise the buffer immediately after allocating it in xfs_buf_get, too, so that is it ready for insert if we get another cache miss after allocation. This minimises lock hold time and avoids unnecessary allocator churn. The resulting profiles look like: samples pcnt function DSO _______ _____ ___________________________ _________________ 8111.00 9.1% _xfs_buf_find [kernel.kallsyms] 4380.00 4.9% __memcpy [kernel.kallsyms] 4341.00 4.8% __ticket_spin_lock [kernel.kallsyms] 3401.00 3.8% kmem_cache_alloc [kernel.kallsyms] 2856.00 3.2% xfs_log_commit_cil [kernel.kallsyms] 2625.00 2.9% __kmalloc [kernel.kallsyms] 2380.00 2.7% kfree [kernel.kallsyms] 2016.00 2.3% kmem_cache_free [kernel.kallsyms] Showing a significant reduction in time spent doing allocation and freeing from slabs (kmem_cache_alloc and kmem_cache_free). Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2011-09-30 04:45:02 +00:00
if (flags & XBF_INCORE)
goto out_put_perag;
/* xfs_buf_find_insert() consumes the perag reference. */
error = xfs_buf_find_insert(btp, pag, &cmap, map, nmaps,
flags, &bp);
if (error)
return error;
} else {
XFS_STATS_INC(btp->bt_mount, xb_get_locked);
xfs_perag_put(pag);
}
/* We do not hold a perag reference anymore. */
if (!bp->b_addr) {
error = _xfs_buf_map_pages(bp, flags);
if (unlikely(error)) {
xfs_warn_ratelimited(btp->bt_mount,
"%s: failed to map %u pages", __func__,
bp->b_page_count);
xfs_buf_relse(bp);
return error;
}
}
xfs: inode recovery readahead can race with inode buffer creation When we do inode readahead in log recovery, we do can do the readahead before we've replayed the icreate transaction that stamps the buffer with inode cores. The inode readahead verifier catches this and marks the buffer as !done to indicate that it doesn't yet contain valid inodes. In adding buffer error notification (i.e. setting b_error = -EIO at the same time as as we clear the done flag) to such a readahead verifier failure, we can then get subsequent inode recovery failing with this error: XFS (dm-0): metadata I/O error: block 0xa00060 ("xlog_recover_do..(read#2)") error 5 numblks 32 This occurs when readahead completion races with icreate item replay such as: inode readahead find buffer lock buffer submit RA io .... icreate recovery xfs_trans_get_buffer find buffer lock buffer <blocks on RA completion> ..... <ra completion> fails verifier clear XBF_DONE set bp->b_error = -EIO release and unlock buffer <icreate gains lock> icreate initialises buffer marks buffer as done adds buffer to delayed write queue releases buffer At this point, we have an initialised inode buffer that is up to date but has an -EIO state registered against it. When we finally get to recovering an inode in that buffer: inode item recovery xfs_trans_read_buffer find buffer lock buffer sees XBF_DONE is set, returns buffer sees bp->b_error is set fail log recovery! Essentially, we need xfs_trans_get_buf_map() to clear the error status of the buffer when doing a lookup. This function returns uninitialised buffers, so the buffer returned can not be in an error state and none of the code that uses this function expects b_error to be set on return. Indeed, there is an ASSERT(!bp->b_error); in the transaction case in xfs_trans_get_buf_map() that would have caught this if log recovery used transactions.... This patch firstly changes the inode readahead failure to set -EIO on the buffer, and secondly changes xfs_buf_get_map() to never return a buffer with an error state set so this first change doesn't cause unexpected log recovery failures. cc: <stable@vger.kernel.org> # 3.12 - current Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-01-11 20:03:44 +00:00
/*
* Clear b_error if this is a lookup from a caller that doesn't expect
* valid data to be found in the buffer.
*/
if (!(flags & XBF_READ))
xfs_buf_ioerror(bp, 0);
XFS_STATS_INC(btp->bt_mount, xb_get);
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_get(bp, flags, _RET_IP_);
*bpp = bp;
return 0;
out_put_perag:
xfs_perag_put(pag);
return error;
}
int
_xfs_buf_read(
struct xfs_buf *bp,
xfs_buf_flags_t flags)
{
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
ASSERT(!(flags & XBF_WRITE));
ASSERT(bp->b_maps[0].bm_bn != XFS_BUF_DADDR_NULL);
bp->b_flags &= ~(XBF_WRITE | XBF_ASYNC | XBF_READ_AHEAD | XBF_DONE);
bp->b_flags |= flags & (XBF_READ | XBF_ASYNC | XBF_READ_AHEAD);
return xfs_buf_submit(bp);
}
/*
* Reverify a buffer found in cache without an attached ->b_ops.
*
* If the caller passed an ops structure and the buffer doesn't have ops
* assigned, set the ops and use it to verify the contents. If verification
* fails, clear XBF_DONE. We assume the buffer has no recorded errors and is
* already in XBF_DONE state on entry.
*
* Under normal operations, every in-core buffer is verified on read I/O
* completion. There are two scenarios that can lead to in-core buffers without
* an assigned ->b_ops. The first is during log recovery of buffers on a V4
* filesystem, though these buffers are purged at the end of recovery. The
* other is online repair, which intentionally reads with a NULL buffer ops to
* run several verifiers across an in-core buffer in order to establish buffer
* type. If repair can't establish that, the buffer will be left in memory
* with NULL buffer ops.
*/
int
xfs_buf_reverify(
struct xfs_buf *bp,
const struct xfs_buf_ops *ops)
{
ASSERT(bp->b_flags & XBF_DONE);
ASSERT(bp->b_error == 0);
if (!ops || bp->b_ops)
return 0;
bp->b_ops = ops;
bp->b_ops->verify_read(bp);
if (bp->b_error)
bp->b_flags &= ~XBF_DONE;
return bp->b_error;
}
int
xfs_buf_read_map(
struct xfs_buftarg *target,
struct xfs_buf_map *map,
int nmaps,
xfs_buf_flags_t flags,
struct xfs_buf **bpp,
const struct xfs_buf_ops *ops,
xfs_failaddr_t fa)
{
struct xfs_buf *bp;
int error;
flags |= XBF_READ;
*bpp = NULL;
error = xfs_buf_get_map(target, map, nmaps, flags, &bp);
if (error)
return error;
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_read(bp, flags, _RET_IP_);
if (!(bp->b_flags & XBF_DONE)) {
/* Initiate the buffer read and wait. */
XFS_STATS_INC(target->bt_mount, xb_get_read);
bp->b_ops = ops;
error = _xfs_buf_read(bp, flags);
/* Readahead iodone already dropped the buffer, so exit. */
if (flags & XBF_ASYNC)
return 0;
} else {
/* Buffer already read; all we need to do is check it. */
error = xfs_buf_reverify(bp, ops);
/* Readahead already finished; drop the buffer and exit. */
if (flags & XBF_ASYNC) {
xfs_buf_relse(bp);
return 0;
}
/* We do not want read in the flags */
bp->b_flags &= ~XBF_READ;
ASSERT(bp->b_ops != NULL || ops == NULL);
}
/*
* If we've had a read error, then the contents of the buffer are
* invalid and should not be used. To ensure that a followup read tries
* to pull the buffer from disk again, we clear the XBF_DONE flag and
* mark the buffer stale. This ensures that anyone who has a current
* reference to the buffer will interpret it's contents correctly and
* future cache lookups will also treat it as an empty, uninitialised
* buffer.
*/
if (error) {
xfs: xfs_is_shutdown vs xlog_is_shutdown cage fight I've been chasing a recent resurgence in generic/388 recovery failure and/or corruption events. The events have largely been uninitialised inode chunks being tripped over in log recovery such as: XFS (pmem1): User initiated shutdown received. pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/xfs/xfs_fsops.c:500). Shutting down filesystem. XFS (pmem1): Please unmount the filesystem and rectify the problem(s) XFS (pmem1): Unmounting Filesystem XFS (pmem1): Mounting V5 Filesystem XFS (pmem1): Starting recovery (logdev: internal) XFS (pmem1): bad inode magic/vsn daddr 8723584 #0 (magic=1818) XFS (pmem1): Metadata corruption detected at xfs_inode_buf_verify+0x180/0x190, xfs_inode block 0x851c80 xfs_inode_buf_verify XFS (pmem1): Unmount and run xfs_repair XFS (pmem1): First 128 bytes of corrupted metadata buffer: 00000000: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000010: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000020: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000030: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000040: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000050: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000060: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000070: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ XFS (pmem1): metadata I/O error in "xlog_recover_items_pass2+0x52/0xc0" at daddr 0x851c80 len 32 error 117 XFS (pmem1): log mount/recovery failed: error -117 XFS (pmem1): log mount failed There have been isolated random other issues, too - xfs_repair fails because it finds some corruption in symlink blocks, rmap inconsistencies, etc - but they are nowhere near as common as the uninitialised inode chunk failure. The problem has clearly happened at runtime before recovery has run; I can see the ICREATE log item in the log shortly before the actively recovered range of the log. This means the ICREATE was definitely created and written to the log, but for some reason the tail of the log has been moved past the ordered buffer log item that tracks INODE_ALLOC buffers and, supposedly, prevents the tail of the log moving past the ICREATE log item before the inode chunk buffer is written to disk. Tracing the fsstress processes that are running when the filesystem shut down immediately pin-pointed the problem: user shutdown marks xfs_mount as shutdown godown-213341 [008] 6398.022871: console: [ 6397.915392] XFS (pmem1): User initiated shutdown received. ..... aild tries to push ordered inode cluster buffer xfsaild/pmem1-213314 [001] 6398.022974: xfs_buf_trylock: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 16 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_inode_item_push+0x8e xfsaild/pmem1-213314 [001] 6398.022976: xfs_ilock_nowait: dev 259:1 ino 0x851c80 flags ILOCK_SHARED caller xfs_iflush_cluster+0xae xfs_iflush_cluster() checks xfs_is_shutdown(), returns true, calls xfs_iflush_abort() to kill writeback of the inode. Inode is removed from AIL, drops cluster buffer reference. xfsaild/pmem1-213314 [001] 6398.022977: xfs_ail_delete: dev 259:1 lip 0xffff88880247ed80 old lsn 7/20344 new lsn 7/21000 type XFS_LI_INODE flags IN_AIL xfsaild/pmem1-213314 [001] 6398.022978: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 17 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_iflush_abort+0xd7 ..... All inodes on cluster buffer are aborted, then the cluster buffer itself is aborted and removed from the AIL *without writeback*: xfsaild/pmem1-213314 [001] 6398.023011: xfs_buf_error_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_ioend_fail+0x33 xfsaild/pmem1-213314 [001] 6398.023012: xfs_ail_delete: dev 259:1 lip 0xffff8888053efde8 old lsn 7/20344 new lsn 7/20344 type XFS_LI_BUF flags IN_AIL The inode buffer was at 7/20344 when it was removed from the AIL. xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_item_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_done+0x31 xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_relse+0x39 ..... Userspace is still running, doing stuff. an fsstress process runs syncfs() or sync() and we end up in sync_fs_one_sb() which issues a log force. This pushes on the CIL: fsstress-213322 [001] 6398.024430: xfs_fs_sync_fs: dev 259:1 m_features 0x20000000019ff6e9 opstate (clean|shutdown|inodegc|blockgc) s_flags 0x70810000 caller sync_fs_one_sb+0x26 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x0 caller xfs_fs_sync_fs+0x82 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x5f caller xfs_log_force+0x7c <...>-194402 [001] 6398.024467: kmem_alloc: size 176 flags 0x14 caller xlog_cil_push_work+0x9f And the CIL fills up iclogs with pending changes. This picks up the current tail from the AIL: <...>-194402 [001] 6398.024497: xlog_iclog_get_space: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x0 flags caller xlog_write+0x149 <...>-194402 [001] 6398.024498: xlog_iclog_switch: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x700005408 flags caller xlog_state_get_iclog_space+0x37e <...>-194402 [001] 6398.024521: xlog_iclog_release: dev 259:1 state XLOG_STATE_WANT_SYNC refcnt 1 offset 32256 lsn 0x700005408 flags caller xlog_write+0x5f9 <...>-194402 [001] 6398.024522: xfs_log_assign_tail_lsn: dev 259:1 new tail lsn 7/21000, old lsn 7/20344, last sync 7/21448 And it moves the tail of the log to 7/21000 from 7/20344. This *moves the tail of the log beyond the ICREATE transaction* that was at 7/20344 and pinned by the inode cluster buffer that was cancelled above. .... godown-213341 [008] 6398.027005: xfs_force_shutdown: dev 259:1 tag logerror flags log_io|force_umount file fs/xfs/xfs_fsops.c line_num 500 godown-213341 [008] 6398.027022: console: [ 6397.915406] pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 godown-213341 [008] 6398.030551: console: [ 6397.919546] XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/ And finally the log itself is now shutdown, stopping all further writes to the log. But this is too late to prevent the corruption that moving the tail of the log forwards after we start cancelling writeback causes. The fundamental problem here is that we are using the wrong shutdown checks for log items. We've long conflated mount shutdown with log shutdown state, and I started separating that recently with the atomic shutdown state changes in commit b36d4651e165 ("xfs: make forced shutdown processing atomic"). The changes in that commit series are directly responsible for being able to diagnose this issue because it clearly separated mount shutdown from log shutdown. Essentially, once we start cancelling writeback of log items and removing them from the AIL because the filesystem is shut down, we *cannot* update the journal because we may have cancelled the items that pin the tail of the log. That moves the tail of the log forwards without having written the metadata back, hence we have corrupt in memory state and writing to the journal propagates that to the on-disk state. What commit b36d4651e165 makes clear is that log item state needs to change relative to log shutdown, not mount shutdown. IOWs, anything that aborts metadata writeback needs to check log shutdown state because log items directly affect log consistency. Having them check mount shutdown state introduces the above race condition where we cancel metadata writeback before the log shuts down. To fix this, this patch works through all log items and converts shutdown checks to use xlog_is_shutdown() rather than xfs_is_shutdown(), so that we don't start aborting metadata writeback before we shut off journal writes. AFAICT, this race condition is a zero day IO error handling bug in XFS that dates back to the introduction of XLOG_IO_ERROR, XLOG_STATE_IOERROR and XFS_FORCED_SHUTDOWN back in January 1997. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:13 +00:00
/*
* Check against log shutdown for error reporting because
* metadata writeback may require a read first and we need to
* report errors in metadata writeback until the log is shut
* down. High level transaction read functions already check
* against mount shutdown, anyway, so we only need to be
* concerned about low level IO interactions here.
*/
if (!xlog_is_shutdown(target->bt_mount->m_log))
xfs_buf_ioerror_alert(bp, fa);
bp->b_flags &= ~XBF_DONE;
xfs_buf_stale(bp);
xfs_buf_relse(bp);
/* bad CRC means corrupted metadata */
if (error == -EFSBADCRC)
error = -EFSCORRUPTED;
return error;
}
*bpp = bp;
return 0;
}
/*
* If we are not low on memory then do the readahead in a deadlock
* safe manner.
*/
void
xfs_buf_readahead_map(
struct xfs_buftarg *target,
struct xfs_buf_map *map,
int nmaps,
const struct xfs_buf_ops *ops)
{
struct xfs_buf *bp;
xfs_buf_read_map(target, map, nmaps,
XBF_TRYLOCK | XBF_ASYNC | XBF_READ_AHEAD, &bp, ops,
__this_address);
}
/*
* Read an uncached buffer from disk. Allocates and returns a locked
* buffer containing the disk contents or nothing. Uncached buffers always have
* a cache index of XFS_BUF_DADDR_NULL so we can easily determine if the buffer
* is cached or uncached during fault diagnosis.
*/
int
xfs_buf_read_uncached(
struct xfs_buftarg *target,
xfs_daddr_t daddr,
size_t numblks,
2022-04-20 22:44:59 +00:00
xfs_buf_flags_t flags,
struct xfs_buf **bpp,
const struct xfs_buf_ops *ops)
{
struct xfs_buf *bp;
int error;
*bpp = NULL;
error = xfs_buf_get_uncached(target, numblks, flags, &bp);
if (error)
return error;
/* set up the buffer for a read IO */
ASSERT(bp->b_map_count == 1);
bp->b_rhash_key = XFS_BUF_DADDR_NULL;
bp->b_maps[0].bm_bn = daddr;
bp->b_flags |= XBF_READ;
bp->b_ops = ops;
xfs_buf_submit(bp);
if (bp->b_error) {
error = bp->b_error;
xfs_buf_relse(bp);
return error;
}
*bpp = bp;
return 0;
}
int
xfs_buf_get_uncached(
struct xfs_buftarg *target,
size_t numblks,
2022-04-20 22:44:59 +00:00
xfs_buf_flags_t flags,
struct xfs_buf **bpp)
{
int error;
struct xfs_buf *bp;
DEFINE_SINGLE_BUF_MAP(map, XFS_BUF_DADDR_NULL, numblks);
*bpp = NULL;
/* flags might contain irrelevant bits, pass only what we care about */
error = _xfs_buf_alloc(target, &map, 1, flags & XBF_NO_IOACCT, &bp);
if (error)
return error;
error = xfs_buf_alloc_pages(bp, flags);
if (error)
goto fail_free_buf;
error = _xfs_buf_map_pages(bp, 0);
if (unlikely(error)) {
xfs_warn(target->bt_mount,
"%s: failed to map pages", __func__);
goto fail_free_buf;
}
trace_xfs_buf_get_uncached(bp, _RET_IP_);
*bpp = bp;
return 0;
fail_free_buf:
xfs_buf_free(bp);
return error;
}
/*
* Increment reference count on buffer, to hold the buffer concurrently
* with another thread which may release (free) the buffer asynchronously.
* Must hold the buffer already to call this function.
*/
void
xfs_buf_hold(
struct xfs_buf *bp)
{
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_hold(bp, _RET_IP_);
atomic_inc(&bp->b_hold);
}
/*
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
* Release a hold on the specified buffer. If the hold count is 1, the buffer is
* placed on LRU or freed (depending on b_lru_ref).
*/
void
xfs_buf_rele(
struct xfs_buf *bp)
{
struct xfs_perag *pag = bp->b_pag;
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
bool release;
bool freebuf = false;
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_rele(bp, _RET_IP_);
if (!pag) {
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
ASSERT(list_empty(&bp->b_lru));
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
if (atomic_dec_and_test(&bp->b_hold)) {
xfs_buf_ioacct_dec(bp);
xfs_buf_free(bp);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
}
return;
}
ASSERT(atomic_read(&bp->b_hold) > 0);
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
xfs: fix use-after-free race in xfs_buf_rele When looking at a 4.18 based KASAN use after free report, I noticed that racing xfs_buf_rele() may race on dropping the last reference to the buffer and taking the buffer lock. This was the symptom displayed by the KASAN report, but the actual issue that was reported had already been fixed in 4.19-rc1 by commit e339dd8d8b04 ("xfs: use sync buffer I/O for sync delwri queue submission"). Despite this, I think there is still an issue with xfs_buf_rele() in this code: release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock); spin_lock(&bp->b_lock); if (!release) { ..... If two threads race on the b_lock after both dropping a reference and one getting dropping the last reference so release = true, we end up with: CPU 0 CPU 1 atomic_dec_and_lock() atomic_dec_and_lock() spin_lock(&bp->b_lock) spin_lock(&bp->b_lock) <spins> <release = true bp->b_lru_ref = 0> <remove from lists> freebuf = true spin_unlock(&bp->b_lock) xfs_buf_free(bp) <gets lock, reading and writing freed memory> <accesses freed memory> spin_unlock(&bp->b_lock) <reads/writes freed memory> IOWs, we can't safely take bp->b_lock after dropping the hold reference because the buffer may go away at any time after we drop that reference. However, this can be fixed simply by taking the bp->b_lock before we drop the reference. It is safe to nest the pag_buf_lock inside bp->b_lock as the pag_buf_lock is only used to serialise against lookup in xfs_buf_find() and no other locks are held over or under the pag_buf_lock there. Make this clear by documenting the buffer lock orders at the top of the file. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com Signed-off-by: Dave Chinner <david@fromorbit.com>
2018-10-18 06:21:29 +00:00
/*
* We grab the b_lock here first to serialise racing xfs_buf_rele()
* calls. The pag_buf_lock being taken on the last reference only
* serialises against racing lookups in xfs_buf_find(). IOWs, the second
* to last reference we drop here is not serialised against the last
* reference until we take bp->b_lock. Hence if we don't grab b_lock
* first, the last "release" reference can win the race to the lock and
* free the buffer before the second-to-last reference is processed,
* leading to a use-after-free scenario.
*/
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
spin_lock(&bp->b_lock);
xfs: fix use-after-free race in xfs_buf_rele When looking at a 4.18 based KASAN use after free report, I noticed that racing xfs_buf_rele() may race on dropping the last reference to the buffer and taking the buffer lock. This was the symptom displayed by the KASAN report, but the actual issue that was reported had already been fixed in 4.19-rc1 by commit e339dd8d8b04 ("xfs: use sync buffer I/O for sync delwri queue submission"). Despite this, I think there is still an issue with xfs_buf_rele() in this code: release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock); spin_lock(&bp->b_lock); if (!release) { ..... If two threads race on the b_lock after both dropping a reference and one getting dropping the last reference so release = true, we end up with: CPU 0 CPU 1 atomic_dec_and_lock() atomic_dec_and_lock() spin_lock(&bp->b_lock) spin_lock(&bp->b_lock) <spins> <release = true bp->b_lru_ref = 0> <remove from lists> freebuf = true spin_unlock(&bp->b_lock) xfs_buf_free(bp) <gets lock, reading and writing freed memory> <accesses freed memory> spin_unlock(&bp->b_lock) <reads/writes freed memory> IOWs, we can't safely take bp->b_lock after dropping the hold reference because the buffer may go away at any time after we drop that reference. However, this can be fixed simply by taking the bp->b_lock before we drop the reference. It is safe to nest the pag_buf_lock inside bp->b_lock as the pag_buf_lock is only used to serialise against lookup in xfs_buf_find() and no other locks are held over or under the pag_buf_lock there. Make this clear by documenting the buffer lock orders at the top of the file. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com Signed-off-by: Dave Chinner <david@fromorbit.com>
2018-10-18 06:21:29 +00:00
release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
if (!release) {
/*
* Drop the in-flight state if the buffer is already on the LRU
* and it holds the only reference. This is racy because we
* haven't acquired the pag lock, but the use of _XBF_IN_FLIGHT
* ensures the decrement occurs only once per-buf.
*/
if ((atomic_read(&bp->b_hold) == 1) && !list_empty(&bp->b_lru))
xfs: use ->b_state to fix buffer I/O accounting release race We've had user reports of unmount hangs in xfs_wait_buftarg() that analysis shows is due to btp->bt_io_count == -1. bt_io_count represents the count of in-flight asynchronous buffers and thus should always be >= 0. xfs_wait_buftarg() waits for this value to stabilize to zero in order to ensure that all untracked (with respect to the lru) buffers have completed I/O processing before unmount proceeds to tear down in-core data structures. The value of -1 implies an I/O accounting decrement race. Indeed, the fact that xfs_buf_ioacct_dec() is called from xfs_buf_rele() (where the buffer lock is no longer held) means that bp->b_flags can be updated from an unsafe context. While a user-level reproducer is currently not available, some intrusive hacks to run racing buffer lookups/ioacct/releases from multiple threads was used to successfully manufacture this problem. Existing callers do not expect to acquire the buffer lock from xfs_buf_rele(). Therefore, we can not safely update ->b_flags from this context. It turns out that we already have separate buffer state bits and associated serialization for dealing with buffer LRU state in the form of ->b_state and ->b_lock. Therefore, replace the _XBF_IN_FLIGHT flag with a ->b_state variant, update the I/O accounting wrappers appropriately and make sure they are used with the correct locking. This ensures that buffer in-flight state can be modified at buffer release time without racing with modifications from a buffer lock holder. Fixes: 9c7504aa72b6 ("xfs: track and serialize in-flight async buffers against unmount") Cc: <stable@vger.kernel.org> # v4.8+ Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Nikolay Borisov <nborisov@suse.com> Tested-by: Libor Pechacek <lpechacek@suse.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-05-31 15:22:52 +00:00
__xfs_buf_ioacct_dec(bp);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
goto out_unlock;
}
/* the last reference has been dropped ... */
xfs: use ->b_state to fix buffer I/O accounting release race We've had user reports of unmount hangs in xfs_wait_buftarg() that analysis shows is due to btp->bt_io_count == -1. bt_io_count represents the count of in-flight asynchronous buffers and thus should always be >= 0. xfs_wait_buftarg() waits for this value to stabilize to zero in order to ensure that all untracked (with respect to the lru) buffers have completed I/O processing before unmount proceeds to tear down in-core data structures. The value of -1 implies an I/O accounting decrement race. Indeed, the fact that xfs_buf_ioacct_dec() is called from xfs_buf_rele() (where the buffer lock is no longer held) means that bp->b_flags can be updated from an unsafe context. While a user-level reproducer is currently not available, some intrusive hacks to run racing buffer lookups/ioacct/releases from multiple threads was used to successfully manufacture this problem. Existing callers do not expect to acquire the buffer lock from xfs_buf_rele(). Therefore, we can not safely update ->b_flags from this context. It turns out that we already have separate buffer state bits and associated serialization for dealing with buffer LRU state in the form of ->b_state and ->b_lock. Therefore, replace the _XBF_IN_FLIGHT flag with a ->b_state variant, update the I/O accounting wrappers appropriately and make sure they are used with the correct locking. This ensures that buffer in-flight state can be modified at buffer release time without racing with modifications from a buffer lock holder. Fixes: 9c7504aa72b6 ("xfs: track and serialize in-flight async buffers against unmount") Cc: <stable@vger.kernel.org> # v4.8+ Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Nikolay Borisov <nborisov@suse.com> Tested-by: Libor Pechacek <lpechacek@suse.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-05-31 15:22:52 +00:00
__xfs_buf_ioacct_dec(bp);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
if (!(bp->b_flags & XBF_STALE) && atomic_read(&bp->b_lru_ref)) {
/*
* If the buffer is added to the LRU take a new reference to the
* buffer for the LRU and clear the (now stale) dispose list
* state flag
*/
if (list_lru_add(&bp->b_target->bt_lru, &bp->b_lru)) {
bp->b_state &= ~XFS_BSTATE_DISPOSE;
atomic_inc(&bp->b_hold);
}
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
spin_unlock(&pag->pag_buf_lock);
} else {
/*
* most of the time buffers will already be removed from the
* LRU, so optimise that case by checking for the
* XFS_BSTATE_DISPOSE flag indicating the last list the buffer
* was on was the disposal list
*/
if (!(bp->b_state & XFS_BSTATE_DISPOSE)) {
list_lru_del(&bp->b_target->bt_lru, &bp->b_lru);
} else {
ASSERT(list_empty(&bp->b_lru));
}
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
ASSERT(!(bp->b_flags & _XBF_DELWRI_Q));
rhashtable_remove_fast(&pag->pag_buf_hash, &bp->b_rhash_head,
xfs_buf_hash_params);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
spin_unlock(&pag->pag_buf_lock);
xfs_perag_put(pag);
freebuf = true;
}
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
out_unlock:
spin_unlock(&bp->b_lock);
if (freebuf)
xfs_buf_free(bp);
}
/*
2011-03-25 22:16:45 +00:00
* Lock a buffer object, if it is not already locked.
*
* If we come across a stale, pinned, locked buffer, we know that we are
* being asked to lock a buffer that has been reallocated. Because it is
* pinned, we know that the log has not been pushed to disk and hence it
* will still be locked. Rather than continuing to have trylock attempts
* fail until someone else pushes the log, push it ourselves before
* returning. This means that the xfsaild will not get stuck trying
* to push on stale inode buffers.
*/
int
xfs_buf_trylock(
struct xfs_buf *bp)
{
int locked;
locked = down_trylock(&bp->b_sema) == 0;
if (locked)
trace_xfs_buf_trylock(bp, _RET_IP_);
else
trace_xfs_buf_trylock_fail(bp, _RET_IP_);
return locked;
}
/*
2011-03-25 22:16:45 +00:00
* Lock a buffer object.
xfs: Improve scalability of busy extent tracking When we free a metadata extent, we record it in the per-AG busy extent array so that it is not re-used before the freeing transaction hits the disk. This array is fixed size, so when it overflows we make further allocation transactions synchronous because we cannot track more freed extents until those transactions hit the disk and are completed. Under heavy mixed allocation and freeing workloads with large log buffers, we can overflow this array quite easily. Further, the array is sparsely populated, which means that inserts need to search for a free slot, and array searches often have to search many more slots that are actually used to check all the busy extents. Quite inefficient, really. To enable this aspect of extent freeing to scale better, we need a structure that can grow dynamically. While in other areas of XFS we have used radix trees, the extents being freed are at random locations on disk so are better suited to being indexed by an rbtree. So, use a per-AG rbtree indexed by block number to track busy extents. This incures a memory allocation when marking an extent busy, but should not occur too often in low memory situations. This should scale to an arbitrary number of extents so should not be a limitation for features such as in-memory aggregation of transactions. However, there are still situations where we can't avoid allocating busy extents (such as allocation from the AGFL). To minimise the overhead of such occurences, we need to avoid doing a synchronous log force while holding the AGF locked to ensure that the previous transactions are safely on disk before we use the extent. We can do this by marking the transaction doing the allocation as synchronous rather issuing a log force. Because of the locking involved and the ordering of transactions, the synchronous transaction provides the same guarantees as a synchronous log force because it ensures that all the prior transactions are already on disk when the synchronous transaction hits the disk. i.e. it preserves the free->allocate order of the extent correctly in recovery. By doing this, we avoid holding the AGF locked while log writes are in progress, hence reducing the length of time the lock is held and therefore we increase the rate at which we can allocate and free from the allocation group, thereby increasing overall throughput. The only problem with this approach is that when a metadata buffer is marked stale (e.g. a directory block is removed), then buffer remains pinned and locked until the log goes to disk. The issue here is that if that stale buffer is reallocated in a subsequent transaction, the attempt to lock that buffer in the transaction will hang waiting the log to go to disk to unlock and unpin the buffer. Hence if someone tries to lock a pinned, stale, locked buffer we need to push on the log to get it unlocked ASAP. Effectively we are trading off a guaranteed log force for a much less common trigger for log force to occur. Ideally we should not reallocate busy extents. That is a much more complex fix to the problem as it involves direct intervention in the allocation btree searches in many places. This is left to a future set of modifications. Finally, now that we track busy extents in allocated memory, we don't need the descriptors in the transaction structure to point to them. We can replace the complex busy chunk infrastructure with a simple linked list of busy extents. This allows us to remove a large chunk of code, making the overall change a net reduction in code size. Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2010-05-21 02:07:08 +00:00
*
* If we come across a stale, pinned, locked buffer, we know that we
* are being asked to lock a buffer that has been reallocated. Because
* it is pinned, we know that the log has not been pushed to disk and
* hence it will still be locked. Rather than sleeping until someone
* else pushes the log, push it ourselves before trying to get the lock.
*/
void
xfs_buf_lock(
struct xfs_buf *bp)
{
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_lock(bp, _RET_IP_);
xfs: Improve scalability of busy extent tracking When we free a metadata extent, we record it in the per-AG busy extent array so that it is not re-used before the freeing transaction hits the disk. This array is fixed size, so when it overflows we make further allocation transactions synchronous because we cannot track more freed extents until those transactions hit the disk and are completed. Under heavy mixed allocation and freeing workloads with large log buffers, we can overflow this array quite easily. Further, the array is sparsely populated, which means that inserts need to search for a free slot, and array searches often have to search many more slots that are actually used to check all the busy extents. Quite inefficient, really. To enable this aspect of extent freeing to scale better, we need a structure that can grow dynamically. While in other areas of XFS we have used radix trees, the extents being freed are at random locations on disk so are better suited to being indexed by an rbtree. So, use a per-AG rbtree indexed by block number to track busy extents. This incures a memory allocation when marking an extent busy, but should not occur too often in low memory situations. This should scale to an arbitrary number of extents so should not be a limitation for features such as in-memory aggregation of transactions. However, there are still situations where we can't avoid allocating busy extents (such as allocation from the AGFL). To minimise the overhead of such occurences, we need to avoid doing a synchronous log force while holding the AGF locked to ensure that the previous transactions are safely on disk before we use the extent. We can do this by marking the transaction doing the allocation as synchronous rather issuing a log force. Because of the locking involved and the ordering of transactions, the synchronous transaction provides the same guarantees as a synchronous log force because it ensures that all the prior transactions are already on disk when the synchronous transaction hits the disk. i.e. it preserves the free->allocate order of the extent correctly in recovery. By doing this, we avoid holding the AGF locked while log writes are in progress, hence reducing the length of time the lock is held and therefore we increase the rate at which we can allocate and free from the allocation group, thereby increasing overall throughput. The only problem with this approach is that when a metadata buffer is marked stale (e.g. a directory block is removed), then buffer remains pinned and locked until the log goes to disk. The issue here is that if that stale buffer is reallocated in a subsequent transaction, the attempt to lock that buffer in the transaction will hang waiting the log to go to disk to unlock and unpin the buffer. Hence if someone tries to lock a pinned, stale, locked buffer we need to push on the log to get it unlocked ASAP. Effectively we are trading off a guaranteed log force for a much less common trigger for log force to occur. Ideally we should not reallocate busy extents. That is a much more complex fix to the problem as it involves direct intervention in the allocation btree searches in many places. This is left to a future set of modifications. Finally, now that we track busy extents in allocated memory, we don't need the descriptors in the transaction structure to point to them. We can replace the complex busy chunk infrastructure with a simple linked list of busy extents. This allows us to remove a large chunk of code, making the overall change a net reduction in code size. Signed-off-by: Dave Chinner <david@fromorbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2010-05-21 02:07:08 +00:00
if (atomic_read(&bp->b_pin_count) && (bp->b_flags & XBF_STALE))
xfs_log_force(bp->b_mount, 0);
down(&bp->b_sema);
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_lock_done(bp, _RET_IP_);
}
void
xfs_buf_unlock(
struct xfs_buf *bp)
{
ASSERT(xfs_buf_islocked(bp));
up(&bp->b_sema);
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
trace_xfs_buf_unlock(bp, _RET_IP_);
}
STATIC void
xfs_buf_wait_unpin(
struct xfs_buf *bp)
{
DECLARE_WAITQUEUE (wait, current);
if (atomic_read(&bp->b_pin_count) == 0)
return;
add_wait_queue(&bp->b_waiters, &wait);
for (;;) {
set_current_state(TASK_UNINTERRUPTIBLE);
if (atomic_read(&bp->b_pin_count) == 0)
break;
io_schedule();
}
remove_wait_queue(&bp->b_waiters, &wait);
set_current_state(TASK_RUNNING);
}
static void
xfs_buf_ioerror_alert_ratelimited(
struct xfs_buf *bp)
{
static unsigned long lasttime;
static struct xfs_buftarg *lasttarg;
if (bp->b_target != lasttarg ||
time_after(jiffies, (lasttime + 5*HZ))) {
lasttime = jiffies;
xfs_buf_ioerror_alert(bp, __this_address);
}
lasttarg = bp->b_target;
}
/*
* Account for this latest trip around the retry handler, and decide if
* we've failed enough times to constitute a permanent failure.
*/
static bool
xfs_buf_ioerror_permanent(
struct xfs_buf *bp,
struct xfs_error_cfg *cfg)
{
struct xfs_mount *mp = bp->b_mount;
if (cfg->max_retries != XFS_ERR_RETRY_FOREVER &&
++bp->b_retries > cfg->max_retries)
return true;
if (cfg->retry_timeout != XFS_ERR_RETRY_FOREVER &&
time_after(jiffies, cfg->retry_timeout + bp->b_first_retry_time))
return true;
/* At unmount we may treat errors differently */
if (xfs_is_unmounting(mp) && mp->m_fail_unmount)
return true;
return false;
}
/*
* On a sync write or shutdown we just want to stale the buffer and let the
* caller handle the error in bp->b_error appropriately.
*
* If the write was asynchronous then no one will be looking for the error. If
* this is the first failure of this type, clear the error state and write the
* buffer out again. This means we always retry an async write failure at least
* once, but we also need to set the buffer up to behave correctly now for
* repeated failures.
*
* If we get repeated async write failures, then we take action according to the
* error configuration we have been set up to use.
*
* Returns true if this function took care of error handling and the caller must
* not touch the buffer again. Return false if the caller should proceed with
* normal I/O completion handling.
*/
static bool
xfs_buf_ioend_handle_error(
struct xfs_buf *bp)
{
struct xfs_mount *mp = bp->b_mount;
struct xfs_error_cfg *cfg;
/*
xfs: xfs_is_shutdown vs xlog_is_shutdown cage fight I've been chasing a recent resurgence in generic/388 recovery failure and/or corruption events. The events have largely been uninitialised inode chunks being tripped over in log recovery such as: XFS (pmem1): User initiated shutdown received. pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/xfs/xfs_fsops.c:500). Shutting down filesystem. XFS (pmem1): Please unmount the filesystem and rectify the problem(s) XFS (pmem1): Unmounting Filesystem XFS (pmem1): Mounting V5 Filesystem XFS (pmem1): Starting recovery (logdev: internal) XFS (pmem1): bad inode magic/vsn daddr 8723584 #0 (magic=1818) XFS (pmem1): Metadata corruption detected at xfs_inode_buf_verify+0x180/0x190, xfs_inode block 0x851c80 xfs_inode_buf_verify XFS (pmem1): Unmount and run xfs_repair XFS (pmem1): First 128 bytes of corrupted metadata buffer: 00000000: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000010: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000020: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000030: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000040: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000050: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000060: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000070: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ XFS (pmem1): metadata I/O error in "xlog_recover_items_pass2+0x52/0xc0" at daddr 0x851c80 len 32 error 117 XFS (pmem1): log mount/recovery failed: error -117 XFS (pmem1): log mount failed There have been isolated random other issues, too - xfs_repair fails because it finds some corruption in symlink blocks, rmap inconsistencies, etc - but they are nowhere near as common as the uninitialised inode chunk failure. The problem has clearly happened at runtime before recovery has run; I can see the ICREATE log item in the log shortly before the actively recovered range of the log. This means the ICREATE was definitely created and written to the log, but for some reason the tail of the log has been moved past the ordered buffer log item that tracks INODE_ALLOC buffers and, supposedly, prevents the tail of the log moving past the ICREATE log item before the inode chunk buffer is written to disk. Tracing the fsstress processes that are running when the filesystem shut down immediately pin-pointed the problem: user shutdown marks xfs_mount as shutdown godown-213341 [008] 6398.022871: console: [ 6397.915392] XFS (pmem1): User initiated shutdown received. ..... aild tries to push ordered inode cluster buffer xfsaild/pmem1-213314 [001] 6398.022974: xfs_buf_trylock: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 16 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_inode_item_push+0x8e xfsaild/pmem1-213314 [001] 6398.022976: xfs_ilock_nowait: dev 259:1 ino 0x851c80 flags ILOCK_SHARED caller xfs_iflush_cluster+0xae xfs_iflush_cluster() checks xfs_is_shutdown(), returns true, calls xfs_iflush_abort() to kill writeback of the inode. Inode is removed from AIL, drops cluster buffer reference. xfsaild/pmem1-213314 [001] 6398.022977: xfs_ail_delete: dev 259:1 lip 0xffff88880247ed80 old lsn 7/20344 new lsn 7/21000 type XFS_LI_INODE flags IN_AIL xfsaild/pmem1-213314 [001] 6398.022978: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 17 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_iflush_abort+0xd7 ..... All inodes on cluster buffer are aborted, then the cluster buffer itself is aborted and removed from the AIL *without writeback*: xfsaild/pmem1-213314 [001] 6398.023011: xfs_buf_error_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_ioend_fail+0x33 xfsaild/pmem1-213314 [001] 6398.023012: xfs_ail_delete: dev 259:1 lip 0xffff8888053efde8 old lsn 7/20344 new lsn 7/20344 type XFS_LI_BUF flags IN_AIL The inode buffer was at 7/20344 when it was removed from the AIL. xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_item_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_done+0x31 xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_relse+0x39 ..... Userspace is still running, doing stuff. an fsstress process runs syncfs() or sync() and we end up in sync_fs_one_sb() which issues a log force. This pushes on the CIL: fsstress-213322 [001] 6398.024430: xfs_fs_sync_fs: dev 259:1 m_features 0x20000000019ff6e9 opstate (clean|shutdown|inodegc|blockgc) s_flags 0x70810000 caller sync_fs_one_sb+0x26 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x0 caller xfs_fs_sync_fs+0x82 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x5f caller xfs_log_force+0x7c <...>-194402 [001] 6398.024467: kmem_alloc: size 176 flags 0x14 caller xlog_cil_push_work+0x9f And the CIL fills up iclogs with pending changes. This picks up the current tail from the AIL: <...>-194402 [001] 6398.024497: xlog_iclog_get_space: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x0 flags caller xlog_write+0x149 <...>-194402 [001] 6398.024498: xlog_iclog_switch: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x700005408 flags caller xlog_state_get_iclog_space+0x37e <...>-194402 [001] 6398.024521: xlog_iclog_release: dev 259:1 state XLOG_STATE_WANT_SYNC refcnt 1 offset 32256 lsn 0x700005408 flags caller xlog_write+0x5f9 <...>-194402 [001] 6398.024522: xfs_log_assign_tail_lsn: dev 259:1 new tail lsn 7/21000, old lsn 7/20344, last sync 7/21448 And it moves the tail of the log to 7/21000 from 7/20344. This *moves the tail of the log beyond the ICREATE transaction* that was at 7/20344 and pinned by the inode cluster buffer that was cancelled above. .... godown-213341 [008] 6398.027005: xfs_force_shutdown: dev 259:1 tag logerror flags log_io|force_umount file fs/xfs/xfs_fsops.c line_num 500 godown-213341 [008] 6398.027022: console: [ 6397.915406] pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 godown-213341 [008] 6398.030551: console: [ 6397.919546] XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/ And finally the log itself is now shutdown, stopping all further writes to the log. But this is too late to prevent the corruption that moving the tail of the log forwards after we start cancelling writeback causes. The fundamental problem here is that we are using the wrong shutdown checks for log items. We've long conflated mount shutdown with log shutdown state, and I started separating that recently with the atomic shutdown state changes in commit b36d4651e165 ("xfs: make forced shutdown processing atomic"). The changes in that commit series are directly responsible for being able to diagnose this issue because it clearly separated mount shutdown from log shutdown. Essentially, once we start cancelling writeback of log items and removing them from the AIL because the filesystem is shut down, we *cannot* update the journal because we may have cancelled the items that pin the tail of the log. That moves the tail of the log forwards without having written the metadata back, hence we have corrupt in memory state and writing to the journal propagates that to the on-disk state. What commit b36d4651e165 makes clear is that log item state needs to change relative to log shutdown, not mount shutdown. IOWs, anything that aborts metadata writeback needs to check log shutdown state because log items directly affect log consistency. Having them check mount shutdown state introduces the above race condition where we cancel metadata writeback before the log shuts down. To fix this, this patch works through all log items and converts shutdown checks to use xlog_is_shutdown() rather than xfs_is_shutdown(), so that we don't start aborting metadata writeback before we shut off journal writes. AFAICT, this race condition is a zero day IO error handling bug in XFS that dates back to the introduction of XLOG_IO_ERROR, XLOG_STATE_IOERROR and XFS_FORCED_SHUTDOWN back in January 1997. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:13 +00:00
* If we've already shutdown the journal because of I/O errors, there's
* no point in giving this a retry.
*/
xfs: xfs_is_shutdown vs xlog_is_shutdown cage fight I've been chasing a recent resurgence in generic/388 recovery failure and/or corruption events. The events have largely been uninitialised inode chunks being tripped over in log recovery such as: XFS (pmem1): User initiated shutdown received. pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/xfs/xfs_fsops.c:500). Shutting down filesystem. XFS (pmem1): Please unmount the filesystem and rectify the problem(s) XFS (pmem1): Unmounting Filesystem XFS (pmem1): Mounting V5 Filesystem XFS (pmem1): Starting recovery (logdev: internal) XFS (pmem1): bad inode magic/vsn daddr 8723584 #0 (magic=1818) XFS (pmem1): Metadata corruption detected at xfs_inode_buf_verify+0x180/0x190, xfs_inode block 0x851c80 xfs_inode_buf_verify XFS (pmem1): Unmount and run xfs_repair XFS (pmem1): First 128 bytes of corrupted metadata buffer: 00000000: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000010: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000020: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000030: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000040: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000050: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000060: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000070: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ XFS (pmem1): metadata I/O error in "xlog_recover_items_pass2+0x52/0xc0" at daddr 0x851c80 len 32 error 117 XFS (pmem1): log mount/recovery failed: error -117 XFS (pmem1): log mount failed There have been isolated random other issues, too - xfs_repair fails because it finds some corruption in symlink blocks, rmap inconsistencies, etc - but they are nowhere near as common as the uninitialised inode chunk failure. The problem has clearly happened at runtime before recovery has run; I can see the ICREATE log item in the log shortly before the actively recovered range of the log. This means the ICREATE was definitely created and written to the log, but for some reason the tail of the log has been moved past the ordered buffer log item that tracks INODE_ALLOC buffers and, supposedly, prevents the tail of the log moving past the ICREATE log item before the inode chunk buffer is written to disk. Tracing the fsstress processes that are running when the filesystem shut down immediately pin-pointed the problem: user shutdown marks xfs_mount as shutdown godown-213341 [008] 6398.022871: console: [ 6397.915392] XFS (pmem1): User initiated shutdown received. ..... aild tries to push ordered inode cluster buffer xfsaild/pmem1-213314 [001] 6398.022974: xfs_buf_trylock: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 16 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_inode_item_push+0x8e xfsaild/pmem1-213314 [001] 6398.022976: xfs_ilock_nowait: dev 259:1 ino 0x851c80 flags ILOCK_SHARED caller xfs_iflush_cluster+0xae xfs_iflush_cluster() checks xfs_is_shutdown(), returns true, calls xfs_iflush_abort() to kill writeback of the inode. Inode is removed from AIL, drops cluster buffer reference. xfsaild/pmem1-213314 [001] 6398.022977: xfs_ail_delete: dev 259:1 lip 0xffff88880247ed80 old lsn 7/20344 new lsn 7/21000 type XFS_LI_INODE flags IN_AIL xfsaild/pmem1-213314 [001] 6398.022978: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 17 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_iflush_abort+0xd7 ..... All inodes on cluster buffer are aborted, then the cluster buffer itself is aborted and removed from the AIL *without writeback*: xfsaild/pmem1-213314 [001] 6398.023011: xfs_buf_error_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_ioend_fail+0x33 xfsaild/pmem1-213314 [001] 6398.023012: xfs_ail_delete: dev 259:1 lip 0xffff8888053efde8 old lsn 7/20344 new lsn 7/20344 type XFS_LI_BUF flags IN_AIL The inode buffer was at 7/20344 when it was removed from the AIL. xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_item_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_done+0x31 xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_relse+0x39 ..... Userspace is still running, doing stuff. an fsstress process runs syncfs() or sync() and we end up in sync_fs_one_sb() which issues a log force. This pushes on the CIL: fsstress-213322 [001] 6398.024430: xfs_fs_sync_fs: dev 259:1 m_features 0x20000000019ff6e9 opstate (clean|shutdown|inodegc|blockgc) s_flags 0x70810000 caller sync_fs_one_sb+0x26 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x0 caller xfs_fs_sync_fs+0x82 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x5f caller xfs_log_force+0x7c <...>-194402 [001] 6398.024467: kmem_alloc: size 176 flags 0x14 caller xlog_cil_push_work+0x9f And the CIL fills up iclogs with pending changes. This picks up the current tail from the AIL: <...>-194402 [001] 6398.024497: xlog_iclog_get_space: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x0 flags caller xlog_write+0x149 <...>-194402 [001] 6398.024498: xlog_iclog_switch: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x700005408 flags caller xlog_state_get_iclog_space+0x37e <...>-194402 [001] 6398.024521: xlog_iclog_release: dev 259:1 state XLOG_STATE_WANT_SYNC refcnt 1 offset 32256 lsn 0x700005408 flags caller xlog_write+0x5f9 <...>-194402 [001] 6398.024522: xfs_log_assign_tail_lsn: dev 259:1 new tail lsn 7/21000, old lsn 7/20344, last sync 7/21448 And it moves the tail of the log to 7/21000 from 7/20344. This *moves the tail of the log beyond the ICREATE transaction* that was at 7/20344 and pinned by the inode cluster buffer that was cancelled above. .... godown-213341 [008] 6398.027005: xfs_force_shutdown: dev 259:1 tag logerror flags log_io|force_umount file fs/xfs/xfs_fsops.c line_num 500 godown-213341 [008] 6398.027022: console: [ 6397.915406] pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 godown-213341 [008] 6398.030551: console: [ 6397.919546] XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/ And finally the log itself is now shutdown, stopping all further writes to the log. But this is too late to prevent the corruption that moving the tail of the log forwards after we start cancelling writeback causes. The fundamental problem here is that we are using the wrong shutdown checks for log items. We've long conflated mount shutdown with log shutdown state, and I started separating that recently with the atomic shutdown state changes in commit b36d4651e165 ("xfs: make forced shutdown processing atomic"). The changes in that commit series are directly responsible for being able to diagnose this issue because it clearly separated mount shutdown from log shutdown. Essentially, once we start cancelling writeback of log items and removing them from the AIL because the filesystem is shut down, we *cannot* update the journal because we may have cancelled the items that pin the tail of the log. That moves the tail of the log forwards without having written the metadata back, hence we have corrupt in memory state and writing to the journal propagates that to the on-disk state. What commit b36d4651e165 makes clear is that log item state needs to change relative to log shutdown, not mount shutdown. IOWs, anything that aborts metadata writeback needs to check log shutdown state because log items directly affect log consistency. Having them check mount shutdown state introduces the above race condition where we cancel metadata writeback before the log shuts down. To fix this, this patch works through all log items and converts shutdown checks to use xlog_is_shutdown() rather than xfs_is_shutdown(), so that we don't start aborting metadata writeback before we shut off journal writes. AFAICT, this race condition is a zero day IO error handling bug in XFS that dates back to the introduction of XLOG_IO_ERROR, XLOG_STATE_IOERROR and XFS_FORCED_SHUTDOWN back in January 1997. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:13 +00:00
if (xlog_is_shutdown(mp->m_log))
goto out_stale;
xfs_buf_ioerror_alert_ratelimited(bp);
/*
* We're not going to bother about retrying this during recovery.
* One strike!
*/
if (bp->b_flags & _XBF_LOGRECOVERY) {
xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
return false;
}
/*
* Synchronous writes will have callers process the error.
*/
if (!(bp->b_flags & XBF_ASYNC))
goto out_stale;
trace_xfs_buf_iodone_async(bp, _RET_IP_);
cfg = xfs_error_get_cfg(mp, XFS_ERR_METADATA, bp->b_error);
if (bp->b_last_error != bp->b_error ||
!(bp->b_flags & (XBF_STALE | XBF_WRITE_FAIL))) {
bp->b_last_error = bp->b_error;
if (cfg->retry_timeout != XFS_ERR_RETRY_FOREVER &&
!bp->b_first_retry_time)
bp->b_first_retry_time = jiffies;
goto resubmit;
}
/*
* Permanent error - we need to trigger a shutdown if we haven't already
* to indicate that inconsistency will result from this action.
*/
if (xfs_buf_ioerror_permanent(bp, cfg)) {
xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
goto out_stale;
}
/* Still considered a transient error. Caller will schedule retries. */
if (bp->b_flags & _XBF_INODES)
xfs_buf_inode_io_fail(bp);
else if (bp->b_flags & _XBF_DQUOTS)
xfs_buf_dquot_io_fail(bp);
else
ASSERT(list_empty(&bp->b_li_list));
xfs_buf_ioerror(bp, 0);
xfs_buf_relse(bp);
return true;
resubmit:
xfs_buf_ioerror(bp, 0);
bp->b_flags |= (XBF_DONE | XBF_WRITE_FAIL);
xfs_buf_submit(bp);
return true;
out_stale:
xfs_buf_stale(bp);
bp->b_flags |= XBF_DONE;
bp->b_flags &= ~XBF_WRITE;
trace_xfs_buf_error_relse(bp, _RET_IP_);
return false;
}
static void
xfs_buf_ioend(
struct xfs_buf *bp)
{
trace_xfs_buf_iodone(bp, _RET_IP_);
/*
* Pull in IO completion errors now. We are guaranteed to be running
* single threaded, so we don't need the lock to read b_io_error.
*/
if (!bp->b_error && bp->b_io_error)
xfs_buf_ioerror(bp, bp->b_io_error);
if (bp->b_flags & XBF_READ) {
if (!bp->b_error && bp->b_ops)
bp->b_ops->verify_read(bp);
if (!bp->b_error)
bp->b_flags |= XBF_DONE;
} else {
if (!bp->b_error) {
bp->b_flags &= ~XBF_WRITE_FAIL;
bp->b_flags |= XBF_DONE;
}
if (unlikely(bp->b_error) && xfs_buf_ioend_handle_error(bp))
return;
/* clear the retry state */
bp->b_last_error = 0;
bp->b_retries = 0;
bp->b_first_retry_time = 0;
/*
* Note that for things like remote attribute buffers, there may
* not be a buffer log item here, so processing the buffer log
* item must remain optional.
*/
if (bp->b_log_item)
xfs_buf_item_done(bp);
if (bp->b_flags & _XBF_INODES)
xfs_buf_inode_iodone(bp);
else if (bp->b_flags & _XBF_DQUOTS)
xfs_buf_dquot_iodone(bp);
}
bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_READ_AHEAD |
_XBF_LOGRECOVERY);
if (bp->b_flags & XBF_ASYNC)
xfs_buf_relse(bp);
else
complete(&bp->b_iowait);
}
static void
xfs_buf_ioend_work(
struct work_struct *work)
{
struct xfs_buf *bp =
container_of(work, struct xfs_buf, b_ioend_work);
xfs: event tracing support Convert the old xfs tracing support that could only be used with the out of tree kdb and xfsidbg patches to use the generic event tracer. To use it make sure CONFIG_EVENT_TRACING is enabled and then enable all xfs trace channels by: echo 1 > /sys/kernel/debug/tracing/events/xfs/enable or alternatively enable single events by just doing the same in one event subdirectory, e.g. echo 1 > /sys/kernel/debug/tracing/events/xfs/xfs_ihold/enable or set more complex filters, etc. In Documentation/trace/events.txt all this is desctribed in more detail. To reads the events do a cat /sys/kernel/debug/tracing/trace Compared to the last posting this patch converts the tracing mostly to the one tracepoint per callsite model that other users of the new tracing facility also employ. This allows a very fine-grained control of the tracing, a cleaner output of the traces and also enables the perf tool to use each tracepoint as a virtual performance counter, allowing us to e.g. count how often certain workloads git various spots in XFS. Take a look at http://lwn.net/Articles/346470/ for some examples. Also the btree tracing isn't included at all yet, as it will require additional core tracing features not in mainline yet, I plan to deliver it later. And the really nice thing about this patch is that it actually removes many lines of code while adding this nice functionality: fs/xfs/Makefile | 8 fs/xfs/linux-2.6/xfs_acl.c | 1 fs/xfs/linux-2.6/xfs_aops.c | 52 - fs/xfs/linux-2.6/xfs_aops.h | 2 fs/xfs/linux-2.6/xfs_buf.c | 117 +-- fs/xfs/linux-2.6/xfs_buf.h | 33 fs/xfs/linux-2.6/xfs_fs_subr.c | 3 fs/xfs/linux-2.6/xfs_ioctl.c | 1 fs/xfs/linux-2.6/xfs_ioctl32.c | 1 fs/xfs/linux-2.6/xfs_iops.c | 1 fs/xfs/linux-2.6/xfs_linux.h | 1 fs/xfs/linux-2.6/xfs_lrw.c | 87 -- fs/xfs/linux-2.6/xfs_lrw.h | 45 - fs/xfs/linux-2.6/xfs_super.c | 104 --- fs/xfs/linux-2.6/xfs_super.h | 7 fs/xfs/linux-2.6/xfs_sync.c | 1 fs/xfs/linux-2.6/xfs_trace.c | 75 ++ fs/xfs/linux-2.6/xfs_trace.h | 1369 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/linux-2.6/xfs_vnode.h | 4 fs/xfs/quota/xfs_dquot.c | 110 --- fs/xfs/quota/xfs_dquot.h | 21 fs/xfs/quota/xfs_qm.c | 40 - fs/xfs/quota/xfs_qm_syscalls.c | 4 fs/xfs/support/ktrace.c | 323 --------- fs/xfs/support/ktrace.h | 85 -- fs/xfs/xfs.h | 16 fs/xfs/xfs_ag.h | 14 fs/xfs/xfs_alloc.c | 230 +----- fs/xfs/xfs_alloc.h | 27 fs/xfs/xfs_alloc_btree.c | 1 fs/xfs/xfs_attr.c | 107 --- fs/xfs/xfs_attr.h | 10 fs/xfs/xfs_attr_leaf.c | 14 fs/xfs/xfs_attr_sf.h | 40 - fs/xfs/xfs_bmap.c | 507 +++------------ fs/xfs/xfs_bmap.h | 49 - fs/xfs/xfs_bmap_btree.c | 6 fs/xfs/xfs_btree.c | 5 fs/xfs/xfs_btree_trace.h | 17 fs/xfs/xfs_buf_item.c | 87 -- fs/xfs/xfs_buf_item.h | 20 fs/xfs/xfs_da_btree.c | 3 fs/xfs/xfs_da_btree.h | 7 fs/xfs/xfs_dfrag.c | 2 fs/xfs/xfs_dir2.c | 8 fs/xfs/xfs_dir2_block.c | 20 fs/xfs/xfs_dir2_leaf.c | 21 fs/xfs/xfs_dir2_node.c | 27 fs/xfs/xfs_dir2_sf.c | 26 fs/xfs/xfs_dir2_trace.c | 216 ------ fs/xfs/xfs_dir2_trace.h | 72 -- fs/xfs/xfs_filestream.c | 8 fs/xfs/xfs_fsops.c | 2 fs/xfs/xfs_iget.c | 111 --- fs/xfs/xfs_inode.c | 67 -- fs/xfs/xfs_inode.h | 76 -- fs/xfs/xfs_inode_item.c | 5 fs/xfs/xfs_iomap.c | 85 -- fs/xfs/xfs_iomap.h | 8 fs/xfs/xfs_log.c | 181 +---- fs/xfs/xfs_log_priv.h | 20 fs/xfs/xfs_log_recover.c | 1 fs/xfs/xfs_mount.c | 2 fs/xfs/xfs_quota.h | 8 fs/xfs/xfs_rename.c | 1 fs/xfs/xfs_rtalloc.c | 1 fs/xfs/xfs_rw.c | 3 fs/xfs/xfs_trans.h | 47 + fs/xfs/xfs_trans_buf.c | 62 - fs/xfs/xfs_vnodeops.c | 8 70 files changed, 2151 insertions(+), 2592 deletions(-) Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Alex Elder <aelder@sgi.com>
2009-12-14 23:14:59 +00:00
xfs_buf_ioend(bp);
}
static void
xfs_buf_ioend_async(
struct xfs_buf *bp)
{
INIT_WORK(&bp->b_ioend_work, xfs_buf_ioend_work);
queue_work(bp->b_mount->m_buf_workqueue, &bp->b_ioend_work);
}
void
__xfs_buf_ioerror(
struct xfs_buf *bp,
int error,
xfs_failaddr_t failaddr)
{
ASSERT(error <= 0 && error >= -1000);
bp->b_error = error;
trace_xfs_buf_ioerror(bp, error, failaddr);
}
void
xfs_buf_ioerror_alert(
struct xfs_buf *bp,
xfs_failaddr_t func)
{
xfs_buf_alert_ratelimited(bp, "XFS: metadata IO error",
"metadata I/O error in \"%pS\" at daddr 0x%llx len %d error %d",
func, (uint64_t)xfs_buf_daddr(bp),
bp->b_length, -bp->b_error);
}
/*
* To simulate an I/O failure, the buffer must be locked and held with at least
* three references. The LRU reference is dropped by the stale call. The buf
* item reference is dropped via ioend processing. The third reference is owned
* by the caller and is dropped on I/O completion if the buffer is XBF_ASYNC.
*/
void
xfs_buf_ioend_fail(
struct xfs_buf *bp)
{
bp->b_flags &= ~XBF_DONE;
xfs_buf_stale(bp);
xfs_buf_ioerror(bp, -EIO);
xfs_buf_ioend(bp);
}
int
xfs_bwrite(
struct xfs_buf *bp)
{
int error;
ASSERT(xfs_buf_islocked(bp));
bp->b_flags |= XBF_WRITE;
bp->b_flags &= ~(XBF_ASYNC | XBF_READ | _XBF_DELWRI_Q |
xfs: reset buffer write failure state on successful completion The buffer write failure flag is intended to control the internal write retry that XFS has historically implemented to help mitigate the severity of transient I/O errors. The flag is set when a buffer is resubmitted from the I/O completion path due to a previous failure. It is checked on subsequent I/O completions to skip the internal retry and fall through to the higher level configurable error handling mechanism. The flag is cleared in the synchronous and delwri submission paths and also checked in various places to log write failure messages. There are a couple minor problems with the current usage of this flag. One is that we issue an internal retry after every submission from xfsaild due to how delwri submission clears the flag. This results in double the expected or configured number of write attempts when under sustained failures. Another more subtle issue is that the flag is never cleared on successful I/O completion. This can cause xfs_wait_buftarg() to suggest that dirty buffers are being thrown away due to the existence of the flag, when the reality is that the flag might still be set because the write succeeded on the retry. Clear the write failure flag on successful I/O completion to address both of these problems. This means that the internal retry attempt occurs once since the last time a buffer write failed and that various other contexts only see the flag set when the immediately previous write attempt has failed. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Allison Collins <allison.henderson@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2020-05-06 20:25:20 +00:00
XBF_DONE);
error = xfs_buf_submit(bp);
if (error)
xfs_force_shutdown(bp->b_mount, SHUTDOWN_META_IO_ERROR);
return error;
}
static void
xfs_buf_bio_end_io(
struct bio *bio)
{
struct xfs_buf *bp = (struct xfs_buf *)bio->bi_private;
if (!bio->bi_status &&
(bp->b_flags & XBF_WRITE) && (bp->b_flags & XBF_ASYNC) &&
XFS_TEST_ERROR(false, bp->b_mount, XFS_ERRTAG_BUF_IOERROR))
bio->bi_status = BLK_STS_IOERR;
/*
* don't overwrite existing errors - otherwise we can lose errors on
* buffers that require multiple bios to complete.
*/
if (bio->bi_status) {
int error = blk_status_to_errno(bio->bi_status);
cmpxchg(&bp->b_io_error, 0, error);
}
if (!bp->b_error && xfs_buf_is_vmapped(bp) && (bp->b_flags & XBF_READ))
invalidate_kernel_vmap_range(bp->b_addr, xfs_buf_vmap_len(bp));
if (atomic_dec_and_test(&bp->b_io_remaining) == 1)
xfs_buf_ioend_async(bp);
bio_put(bio);
}
static void
xfs_buf_ioapply_map(
struct xfs_buf *bp,
int map,
int *buf_offset,
int *count,
blk_opf_t op)
{
int page_index;
unsigned int total_nr_pages = bp->b_page_count;
int nr_pages;
struct bio *bio;
sector_t sector = bp->b_maps[map].bm_bn;
int size;
int offset;
/* skip the pages in the buffer before the start offset */
page_index = 0;
offset = *buf_offset;
while (offset >= PAGE_SIZE) {
page_index++;
offset -= PAGE_SIZE;
}
/*
* Limit the IO size to the length of the current vector, and update the
* remaining IO count for the next time around.
*/
size = min_t(int, BBTOB(bp->b_maps[map].bm_len), *count);
*count -= size;
*buf_offset += size;
next_chunk:
atomic_inc(&bp->b_io_remaining);
nr_pages = bio_max_segs(total_nr_pages);
bio = bio_alloc(bp->b_target->bt_bdev, nr_pages, op, GFP_NOIO);
block: Abstract out bvec iterator Immutable biovecs are going to require an explicit iterator. To implement immutable bvecs, a later patch is going to add a bi_bvec_done member to this struct; for now, this patch effectively just renames things. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Alasdair Kergon <agk@redhat.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: dm-devel@redhat.com Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: Benny Halevy <bhalevy@tonian.com> Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Chris Mason <chris.mason@fusionio.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Cc: Jaegeuk Kim <jaegeuk.kim@samsung.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Joern Engel <joern@logfs.org> Cc: Prasad Joshi <prasadjoshi.linux@gmail.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Ben Myers <bpm@sgi.com> Cc: xfs@oss.sgi.com Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Len Brown <len.brown@intel.com> Cc: Pavel Machek <pavel@ucw.cz> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: "Roger Pau Monné" <roger.pau@citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Jiang Liu <jiang.liu@huawei.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchand@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Peng Tao <tao.peng@emc.com> Cc: Andy Adamson <andros@netapp.com> Cc: fanchaoting <fanchaoting@cn.fujitsu.com> Cc: Jie Liu <jeff.liu@oracle.com> Cc: Sunil Mushran <sunil.mushran@gmail.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Pankaj Kumar <pankaj.km@samsung.com> Cc: Dan Magenheimer <dan.magenheimer@oracle.com> Cc: Mel Gorman <mgorman@suse.de>6
2013-10-11 22:44:27 +00:00
bio->bi_iter.bi_sector = sector;
bio->bi_end_io = xfs_buf_bio_end_io;
bio->bi_private = bp;
2011-03-25 22:16:45 +00:00
for (; size && nr_pages; nr_pages--, page_index++) {
2011-03-25 22:16:45 +00:00
int rbytes, nbytes = PAGE_SIZE - offset;
if (nbytes > size)
nbytes = size;
rbytes = bio_add_page(bio, bp->b_pages[page_index], nbytes,
offset);
if (rbytes < nbytes)
break;
offset = 0;
sector += BTOBB(nbytes);
size -= nbytes;
total_nr_pages--;
}
block: Abstract out bvec iterator Immutable biovecs are going to require an explicit iterator. To implement immutable bvecs, a later patch is going to add a bi_bvec_done member to this struct; for now, this patch effectively just renames things. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Alasdair Kergon <agk@redhat.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: dm-devel@redhat.com Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: Benny Halevy <bhalevy@tonian.com> Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Chris Mason <chris.mason@fusionio.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Cc: Jaegeuk Kim <jaegeuk.kim@samsung.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Joern Engel <joern@logfs.org> Cc: Prasad Joshi <prasadjoshi.linux@gmail.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Ben Myers <bpm@sgi.com> Cc: xfs@oss.sgi.com Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Len Brown <len.brown@intel.com> Cc: Pavel Machek <pavel@ucw.cz> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: "Roger Pau Monné" <roger.pau@citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Jiang Liu <jiang.liu@huawei.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchand@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Peng Tao <tao.peng@emc.com> Cc: Andy Adamson <andros@netapp.com> Cc: fanchaoting <fanchaoting@cn.fujitsu.com> Cc: Jie Liu <jeff.liu@oracle.com> Cc: Sunil Mushran <sunil.mushran@gmail.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Pankaj Kumar <pankaj.km@samsung.com> Cc: Dan Magenheimer <dan.magenheimer@oracle.com> Cc: Mel Gorman <mgorman@suse.de>6
2013-10-11 22:44:27 +00:00
if (likely(bio->bi_iter.bi_size)) {
if (xfs_buf_is_vmapped(bp)) {
flush_kernel_vmap_range(bp->b_addr,
xfs_buf_vmap_len(bp));
}
submit_bio(bio);
if (size)
goto next_chunk;
} else {
/*
* This is guaranteed not to be the last io reference count
* because the caller (xfs_buf_submit) holds a count itself.
*/
atomic_dec(&bp->b_io_remaining);
xfs_buf_ioerror(bp, -EIO);
bio_put(bio);
}
}
STATIC void
_xfs_buf_ioapply(
struct xfs_buf *bp)
{
struct blk_plug plug;
blk_opf_t op;
int offset;
int size;
int i;
/*
* Make sure we capture only current IO errors rather than stale errors
* left over from previous use of the buffer (e.g. failed readahead).
*/
bp->b_error = 0;
if (bp->b_flags & XBF_WRITE) {
op = REQ_OP_WRITE;
/*
* Run the write verifier callback function if it exists. If
* this function fails it will mark the buffer with an error and
* the IO should not be dispatched.
*/
if (bp->b_ops) {
bp->b_ops->verify_write(bp);
if (bp->b_error) {
xfs_force_shutdown(bp->b_mount,
SHUTDOWN_CORRUPT_INCORE);
return;
}
} else if (bp->b_rhash_key != XFS_BUF_DADDR_NULL) {
struct xfs_mount *mp = bp->b_mount;
/*
* non-crc filesystems don't attach verifiers during
* log recovery, so don't warn for such filesystems.
*/
if (xfs_has_crc(mp)) {
xfs_warn(mp,
"%s: no buf ops on daddr 0x%llx len %d",
__func__, xfs_buf_daddr(bp),
bp->b_length);
xfs_hex_dump(bp->b_addr,
XFS_CORRUPTION_DUMP_LEN);
dump_stack();
}
}
} else {
op = REQ_OP_READ;
if (bp->b_flags & XBF_READ_AHEAD)
op |= REQ_RAHEAD;
}
/* we only use the buffer cache for meta-data */
op |= REQ_META;
/*
* Walk all the vectors issuing IO on them. Set up the initial offset
* into the buffer and the desired IO size before we start -
* _xfs_buf_ioapply_vec() will modify them appropriately for each
* subsequent call.
*/
offset = bp->b_offset;
size = BBTOB(bp->b_length);
blk_start_plug(&plug);
for (i = 0; i < bp->b_map_count; i++) {
xfs_buf_ioapply_map(bp, i, &offset, &size, op);
if (bp->b_error)
break;
if (size <= 0)
break; /* all done */
}
blk_finish_plug(&plug);
}
/*
* Wait for I/O completion of a sync buffer and return the I/O error code.
*/
static int
xfs_buf_iowait(
struct xfs_buf *bp)
{
ASSERT(!(bp->b_flags & XBF_ASYNC));
trace_xfs_buf_iowait(bp, _RET_IP_);
wait_for_completion(&bp->b_iowait);
trace_xfs_buf_iowait_done(bp, _RET_IP_);
return bp->b_error;
}
/*
* Buffer I/O submission path, read or write. Asynchronous submission transfers
* the buffer lock ownership and the current reference to the IO. It is not
* safe to reference the buffer after a call to this function unless the caller
* holds an additional reference itself.
*/
static int
__xfs_buf_submit(
struct xfs_buf *bp,
bool wait)
{
int error = 0;
trace_xfs_buf_submit(bp, _RET_IP_);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
ASSERT(!(bp->b_flags & _XBF_DELWRI_Q));
xfs: xfs_is_shutdown vs xlog_is_shutdown cage fight I've been chasing a recent resurgence in generic/388 recovery failure and/or corruption events. The events have largely been uninitialised inode chunks being tripped over in log recovery such as: XFS (pmem1): User initiated shutdown received. pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/xfs/xfs_fsops.c:500). Shutting down filesystem. XFS (pmem1): Please unmount the filesystem and rectify the problem(s) XFS (pmem1): Unmounting Filesystem XFS (pmem1): Mounting V5 Filesystem XFS (pmem1): Starting recovery (logdev: internal) XFS (pmem1): bad inode magic/vsn daddr 8723584 #0 (magic=1818) XFS (pmem1): Metadata corruption detected at xfs_inode_buf_verify+0x180/0x190, xfs_inode block 0x851c80 xfs_inode_buf_verify XFS (pmem1): Unmount and run xfs_repair XFS (pmem1): First 128 bytes of corrupted metadata buffer: 00000000: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000010: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000020: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000030: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000040: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000050: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000060: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000070: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ XFS (pmem1): metadata I/O error in "xlog_recover_items_pass2+0x52/0xc0" at daddr 0x851c80 len 32 error 117 XFS (pmem1): log mount/recovery failed: error -117 XFS (pmem1): log mount failed There have been isolated random other issues, too - xfs_repair fails because it finds some corruption in symlink blocks, rmap inconsistencies, etc - but they are nowhere near as common as the uninitialised inode chunk failure. The problem has clearly happened at runtime before recovery has run; I can see the ICREATE log item in the log shortly before the actively recovered range of the log. This means the ICREATE was definitely created and written to the log, but for some reason the tail of the log has been moved past the ordered buffer log item that tracks INODE_ALLOC buffers and, supposedly, prevents the tail of the log moving past the ICREATE log item before the inode chunk buffer is written to disk. Tracing the fsstress processes that are running when the filesystem shut down immediately pin-pointed the problem: user shutdown marks xfs_mount as shutdown godown-213341 [008] 6398.022871: console: [ 6397.915392] XFS (pmem1): User initiated shutdown received. ..... aild tries to push ordered inode cluster buffer xfsaild/pmem1-213314 [001] 6398.022974: xfs_buf_trylock: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 16 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_inode_item_push+0x8e xfsaild/pmem1-213314 [001] 6398.022976: xfs_ilock_nowait: dev 259:1 ino 0x851c80 flags ILOCK_SHARED caller xfs_iflush_cluster+0xae xfs_iflush_cluster() checks xfs_is_shutdown(), returns true, calls xfs_iflush_abort() to kill writeback of the inode. Inode is removed from AIL, drops cluster buffer reference. xfsaild/pmem1-213314 [001] 6398.022977: xfs_ail_delete: dev 259:1 lip 0xffff88880247ed80 old lsn 7/20344 new lsn 7/21000 type XFS_LI_INODE flags IN_AIL xfsaild/pmem1-213314 [001] 6398.022978: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 17 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_iflush_abort+0xd7 ..... All inodes on cluster buffer are aborted, then the cluster buffer itself is aborted and removed from the AIL *without writeback*: xfsaild/pmem1-213314 [001] 6398.023011: xfs_buf_error_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_ioend_fail+0x33 xfsaild/pmem1-213314 [001] 6398.023012: xfs_ail_delete: dev 259:1 lip 0xffff8888053efde8 old lsn 7/20344 new lsn 7/20344 type XFS_LI_BUF flags IN_AIL The inode buffer was at 7/20344 when it was removed from the AIL. xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_item_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_done+0x31 xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_relse+0x39 ..... Userspace is still running, doing stuff. an fsstress process runs syncfs() or sync() and we end up in sync_fs_one_sb() which issues a log force. This pushes on the CIL: fsstress-213322 [001] 6398.024430: xfs_fs_sync_fs: dev 259:1 m_features 0x20000000019ff6e9 opstate (clean|shutdown|inodegc|blockgc) s_flags 0x70810000 caller sync_fs_one_sb+0x26 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x0 caller xfs_fs_sync_fs+0x82 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x5f caller xfs_log_force+0x7c <...>-194402 [001] 6398.024467: kmem_alloc: size 176 flags 0x14 caller xlog_cil_push_work+0x9f And the CIL fills up iclogs with pending changes. This picks up the current tail from the AIL: <...>-194402 [001] 6398.024497: xlog_iclog_get_space: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x0 flags caller xlog_write+0x149 <...>-194402 [001] 6398.024498: xlog_iclog_switch: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x700005408 flags caller xlog_state_get_iclog_space+0x37e <...>-194402 [001] 6398.024521: xlog_iclog_release: dev 259:1 state XLOG_STATE_WANT_SYNC refcnt 1 offset 32256 lsn 0x700005408 flags caller xlog_write+0x5f9 <...>-194402 [001] 6398.024522: xfs_log_assign_tail_lsn: dev 259:1 new tail lsn 7/21000, old lsn 7/20344, last sync 7/21448 And it moves the tail of the log to 7/21000 from 7/20344. This *moves the tail of the log beyond the ICREATE transaction* that was at 7/20344 and pinned by the inode cluster buffer that was cancelled above. .... godown-213341 [008] 6398.027005: xfs_force_shutdown: dev 259:1 tag logerror flags log_io|force_umount file fs/xfs/xfs_fsops.c line_num 500 godown-213341 [008] 6398.027022: console: [ 6397.915406] pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 godown-213341 [008] 6398.030551: console: [ 6397.919546] XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/ And finally the log itself is now shutdown, stopping all further writes to the log. But this is too late to prevent the corruption that moving the tail of the log forwards after we start cancelling writeback causes. The fundamental problem here is that we are using the wrong shutdown checks for log items. We've long conflated mount shutdown with log shutdown state, and I started separating that recently with the atomic shutdown state changes in commit b36d4651e165 ("xfs: make forced shutdown processing atomic"). The changes in that commit series are directly responsible for being able to diagnose this issue because it clearly separated mount shutdown from log shutdown. Essentially, once we start cancelling writeback of log items and removing them from the AIL because the filesystem is shut down, we *cannot* update the journal because we may have cancelled the items that pin the tail of the log. That moves the tail of the log forwards without having written the metadata back, hence we have corrupt in memory state and writing to the journal propagates that to the on-disk state. What commit b36d4651e165 makes clear is that log item state needs to change relative to log shutdown, not mount shutdown. IOWs, anything that aborts metadata writeback needs to check log shutdown state because log items directly affect log consistency. Having them check mount shutdown state introduces the above race condition where we cancel metadata writeback before the log shuts down. To fix this, this patch works through all log items and converts shutdown checks to use xlog_is_shutdown() rather than xfs_is_shutdown(), so that we don't start aborting metadata writeback before we shut off journal writes. AFAICT, this race condition is a zero day IO error handling bug in XFS that dates back to the introduction of XLOG_IO_ERROR, XLOG_STATE_IOERROR and XFS_FORCED_SHUTDOWN back in January 1997. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:13 +00:00
/*
* On log shutdown we stale and complete the buffer immediately. We can
* be called to read the superblock before the log has been set up, so
* be careful checking the log state.
*
* Checking the mount shutdown state here can result in the log tail
* moving inappropriately on disk as the log may not yet be shut down.
* i.e. failing this buffer on mount shutdown can remove it from the AIL
* and move the tail of the log forwards without having written this
* buffer to disk. This corrupts the log tail state in memory, and
* because the log may not be shut down yet, it can then be propagated
* to disk before the log is shutdown. Hence we check log shutdown
* state here rather than mount state to avoid corrupting the log tail
* on shutdown.
*/
if (bp->b_mount->m_log &&
xlog_is_shutdown(bp->b_mount->m_log)) {
xfs_buf_ioend_fail(bp);
return -EIO;
}
/*
* Grab a reference so the buffer does not go away underneath us. For
* async buffers, I/O completion drops the callers reference, which
* could occur before submission returns.
*/
xfs_buf_hold(bp);
if (bp->b_flags & XBF_WRITE)
xfs_buf_wait_unpin(bp);
/* clear the internal error state to avoid spurious errors */
bp->b_io_error = 0;
/*
* Set the count to 1 initially, this will stop an I/O completion
* callout which happens before we have started all the I/O from calling
* xfs_buf_ioend too early.
*/
atomic_set(&bp->b_io_remaining, 1);
if (bp->b_flags & XBF_ASYNC)
xfs_buf_ioacct_inc(bp);
_xfs_buf_ioapply(bp);
/*
* If _xfs_buf_ioapply failed, we can get back here with only the IO
* reference we took above. If we drop it to zero, run completion so
* that we don't return to the caller with completion still pending.
*/
if (atomic_dec_and_test(&bp->b_io_remaining) == 1) {
if (bp->b_error || !(bp->b_flags & XBF_ASYNC))
xfs_buf_ioend(bp);
else
xfs_buf_ioend_async(bp);
}
if (wait)
error = xfs_buf_iowait(bp);
/*
* Release the hold that keeps the buffer referenced for the entire
* I/O. Note that if the buffer is async, it is not safe to reference
* after this release.
*/
xfs_buf_rele(bp);
return error;
}
void *
xfs_buf_offset(
struct xfs_buf *bp,
size_t offset)
{
struct page *page;
if (bp->b_addr)
return bp->b_addr + offset;
2011-03-25 22:16:45 +00:00
page = bp->b_pages[offset >> PAGE_SHIFT];
return page_address(page) + (offset & (PAGE_SIZE-1));
}
void
xfs_buf_zero(
struct xfs_buf *bp,
size_t boff,
size_t bsize)
{
size_t bend;
bend = boff + bsize;
while (boff < bend) {
struct page *page;
int page_index, page_offset, csize;
page_index = (boff + bp->b_offset) >> PAGE_SHIFT;
page_offset = (boff + bp->b_offset) & ~PAGE_MASK;
page = bp->b_pages[page_index];
csize = min_t(size_t, PAGE_SIZE - page_offset,
BBTOB(bp->b_length) - boff);
ASSERT((csize + page_offset) <= PAGE_SIZE);
memset(page_address(page) + page_offset, 0, csize);
boff += csize;
}
}
/*
* Log a message about and stale a buffer that a caller has decided is corrupt.
*
* This function should be called for the kinds of metadata corruption that
* cannot be detect from a verifier, such as incorrect inter-block relationship
* data. Do /not/ call this function from a verifier function.
*
* The buffer must be XBF_DONE prior to the call. Afterwards, the buffer will
* be marked stale, but b_error will not be set. The caller is responsible for
* releasing the buffer or fixing it.
*/
void
__xfs_buf_mark_corrupt(
struct xfs_buf *bp,
xfs_failaddr_t fa)
{
ASSERT(bp->b_flags & XBF_DONE);
xfs_buf_corruption_error(bp, fa);
xfs_buf_stale(bp);
}
/*
* Handling of buffer targets (buftargs).
*/
/*
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
* Wait for any bufs with callbacks that have been submitted but have not yet
* returned. These buffers will have an elevated hold count, so wait on those
* while freeing all the buffers only held by the LRU.
*/
static enum lru_status
xfs_buftarg_drain_rele(
struct list_head *item,
list_lru: add helpers to isolate items Currently, the isolate callback passed to the list_lru_walk family of functions is supposed to just delete an item from the list upon returning LRU_REMOVED or LRU_REMOVED_RETRY, while nr_items counter is fixed by __list_lru_walk_one after the callback returns. Since the callback is allowed to drop the lock after removing an item (it has to return LRU_REMOVED_RETRY then), the nr_items can be less than the actual number of elements on the list even if we check them under the lock. This makes it difficult to move items from one list_lru_one to another, which is required for per-memcg list_lru reparenting - we can't just splice the lists, we have to move entries one by one. This patch therefore introduces helpers that must be used by callback functions to isolate items instead of raw list_del/list_move. These are list_lru_isolate and list_lru_isolate_move. They not only remove the entry from the list, but also fix the nr_items counter, making sure nr_items always reflects the actual number of elements on the list if checked under the appropriate lock. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Dave Chinner <david@fromorbit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:35 +00:00
struct list_lru_one *lru,
spinlock_t *lru_lock,
void *arg)
{
struct xfs_buf *bp = container_of(item, struct xfs_buf, b_lru);
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
struct list_head *dispose = arg;
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
if (atomic_read(&bp->b_hold) > 1) {
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
/* need to wait, so skip it this pass */
trace_xfs_buf_drain_buftarg(bp, _RET_IP_);
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
return LRU_SKIP;
}
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
if (!spin_trylock(&bp->b_lock))
return LRU_SKIP;
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
/*
* clear the LRU reference count so the buffer doesn't get
* ignored in xfs_buf_rele().
*/
atomic_set(&bp->b_lru_ref, 0);
bp->b_state |= XFS_BSTATE_DISPOSE;
list_lru: add helpers to isolate items Currently, the isolate callback passed to the list_lru_walk family of functions is supposed to just delete an item from the list upon returning LRU_REMOVED or LRU_REMOVED_RETRY, while nr_items counter is fixed by __list_lru_walk_one after the callback returns. Since the callback is allowed to drop the lock after removing an item (it has to return LRU_REMOVED_RETRY then), the nr_items can be less than the actual number of elements on the list even if we check them under the lock. This makes it difficult to move items from one list_lru_one to another, which is required for per-memcg list_lru reparenting - we can't just splice the lists, we have to move entries one by one. This patch therefore introduces helpers that must be used by callback functions to isolate items instead of raw list_del/list_move. These are list_lru_isolate and list_lru_isolate_move. They not only remove the entry from the list, but also fix the nr_items counter, making sure nr_items always reflects the actual number of elements on the list if checked under the appropriate lock. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Dave Chinner <david@fromorbit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:35 +00:00
list_lru_isolate_move(lru, item, dispose);
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
spin_unlock(&bp->b_lock);
return LRU_REMOVED;
}
/*
* Wait for outstanding I/O on the buftarg to complete.
*/
void
xfs_buftarg_wait(
struct xfs_buftarg *btp)
{
xfs: log mount failures don't wait for buffers to be released Recently I've been seeing xfs/051 fail on 1k block size filesystems. Trying to trace the events during the test lead to the problem going away, indicating that it was a race condition that lead to this ASSERT failure: XFS: Assertion failed: atomic_read(&pag->pag_ref) == 0, file: fs/xfs/xfs_mount.c, line: 156 ..... [<ffffffff814e1257>] xfs_free_perag+0x87/0xb0 [<ffffffff814e21b9>] xfs_mountfs+0x4d9/0x900 [<ffffffff814e5dff>] xfs_fs_fill_super+0x3bf/0x4d0 [<ffffffff811d8800>] mount_bdev+0x180/0x1b0 [<ffffffff814e3ff5>] xfs_fs_mount+0x15/0x20 [<ffffffff811d90a8>] mount_fs+0x38/0x170 [<ffffffff811f4347>] vfs_kern_mount+0x67/0x120 [<ffffffff811f7018>] do_mount+0x218/0xd60 [<ffffffff811f7e5b>] SyS_mount+0x8b/0xd0 When I finally caught it with tracing enabled, I saw that AG 2 had an elevated reference count and a buffer was responsible for it. I tracked down the specific buffer, and found that it was missing the final reference count release that would put it back on the LRU and hence be found by xfs_wait_buftarg() calls in the log mount failure handling. The last four traces for the buffer before the assert were (trimmed for relevance) kworker/0:1-5259 xfs_buf_iodone: hold 2 lock 0 flags ASYNC kworker/0:1-5259 xfs_buf_ioerror: hold 2 lock 0 error -5 mount-7163 xfs_buf_lock_done: hold 2 lock 0 flags ASYNC mount-7163 xfs_buf_unlock: hold 2 lock 1 flags ASYNC This is an async write that is completing, so there's nobody waiting for it directly. Hence we call xfs_buf_relse() once all the processing is complete. That does: static inline void xfs_buf_relse(xfs_buf_t *bp) { xfs_buf_unlock(bp); xfs_buf_rele(bp); } Now, it's clear that mount is waiting on the buffer lock, and that it has been released by xfs_buf_relse() and gained by mount. This is expected, because at this point the mount process is in xfs_buf_delwri_submit() waiting for all the IO it submitted to complete. The mount process, however, is waiting on the lock for the buffer because it is in xfs_buf_delwri_submit(). This waits for IO completion, but it doesn't wait for the buffer reference owned by the IO to go away. The mount process collects all the completions, fails the log recovery, and the higher level code then calls xfs_wait_buftarg() to free all the remaining buffers in the filesystem. The issue is that on unlocking the buffer, the scheduler has decided that the mount process has higher priority than the the kworker thread that is running the IO completion, and so immediately switched contexts to the mount process from the semaphore unlock code, hence preventing the kworker thread from finishing the IO completion and releasing the IO reference to the buffer. Hence by the time that xfs_wait_buftarg() is run, the buffer still has an active reference and so isn't on the LRU list that the function walks to free the remaining buffers. Hence we miss that buffer and continue onwards to tear down the mount structures, at which time we get find a stray reference count on the perag structure. On a non-debug kernel, this will be ignored and the structure torn down and freed. Hence when the kworker thread is then rescheduled and the buffer released and freed, it will access a freed perag structure. The problem here is that when the log mount fails, we still need to quiesce the log to ensure that the IO workqueues have returned to idle before we run xfs_wait_buftarg(). By synchronising the workqueues, we ensure that all IO completions are fully processed, not just to the point where buffers have been unlocked. This ensures we don't end up in the situation above. cc: <stable@vger.kernel.org> # 3.18 Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-01-18 21:28:10 +00:00
/*
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
* First wait on the buftarg I/O count for all in-flight buffers to be
* released. This is critical as new buffers do not make the LRU until
* they are released.
*
* Next, flush the buffer workqueue to ensure all completion processing
* has finished. Just waiting on buffer locks is not sufficient for
* async IO as the reference count held over IO is not released until
* after the buffer lock is dropped. Hence we need to ensure here that
* all reference counts have been dropped before we start walking the
* LRU list.
xfs: log mount failures don't wait for buffers to be released Recently I've been seeing xfs/051 fail on 1k block size filesystems. Trying to trace the events during the test lead to the problem going away, indicating that it was a race condition that lead to this ASSERT failure: XFS: Assertion failed: atomic_read(&pag->pag_ref) == 0, file: fs/xfs/xfs_mount.c, line: 156 ..... [<ffffffff814e1257>] xfs_free_perag+0x87/0xb0 [<ffffffff814e21b9>] xfs_mountfs+0x4d9/0x900 [<ffffffff814e5dff>] xfs_fs_fill_super+0x3bf/0x4d0 [<ffffffff811d8800>] mount_bdev+0x180/0x1b0 [<ffffffff814e3ff5>] xfs_fs_mount+0x15/0x20 [<ffffffff811d90a8>] mount_fs+0x38/0x170 [<ffffffff811f4347>] vfs_kern_mount+0x67/0x120 [<ffffffff811f7018>] do_mount+0x218/0xd60 [<ffffffff811f7e5b>] SyS_mount+0x8b/0xd0 When I finally caught it with tracing enabled, I saw that AG 2 had an elevated reference count and a buffer was responsible for it. I tracked down the specific buffer, and found that it was missing the final reference count release that would put it back on the LRU and hence be found by xfs_wait_buftarg() calls in the log mount failure handling. The last four traces for the buffer before the assert were (trimmed for relevance) kworker/0:1-5259 xfs_buf_iodone: hold 2 lock 0 flags ASYNC kworker/0:1-5259 xfs_buf_ioerror: hold 2 lock 0 error -5 mount-7163 xfs_buf_lock_done: hold 2 lock 0 flags ASYNC mount-7163 xfs_buf_unlock: hold 2 lock 1 flags ASYNC This is an async write that is completing, so there's nobody waiting for it directly. Hence we call xfs_buf_relse() once all the processing is complete. That does: static inline void xfs_buf_relse(xfs_buf_t *bp) { xfs_buf_unlock(bp); xfs_buf_rele(bp); } Now, it's clear that mount is waiting on the buffer lock, and that it has been released by xfs_buf_relse() and gained by mount. This is expected, because at this point the mount process is in xfs_buf_delwri_submit() waiting for all the IO it submitted to complete. The mount process, however, is waiting on the lock for the buffer because it is in xfs_buf_delwri_submit(). This waits for IO completion, but it doesn't wait for the buffer reference owned by the IO to go away. The mount process collects all the completions, fails the log recovery, and the higher level code then calls xfs_wait_buftarg() to free all the remaining buffers in the filesystem. The issue is that on unlocking the buffer, the scheduler has decided that the mount process has higher priority than the the kworker thread that is running the IO completion, and so immediately switched contexts to the mount process from the semaphore unlock code, hence preventing the kworker thread from finishing the IO completion and releasing the IO reference to the buffer. Hence by the time that xfs_wait_buftarg() is run, the buffer still has an active reference and so isn't on the LRU list that the function walks to free the remaining buffers. Hence we miss that buffer and continue onwards to tear down the mount structures, at which time we get find a stray reference count on the perag structure. On a non-debug kernel, this will be ignored and the structure torn down and freed. Hence when the kworker thread is then rescheduled and the buffer released and freed, it will access a freed perag structure. The problem here is that when the log mount fails, we still need to quiesce the log to ensure that the IO workqueues have returned to idle before we run xfs_wait_buftarg(). By synchronising the workqueues, we ensure that all IO completions are fully processed, not just to the point where buffers have been unlocked. This ensures we don't end up in the situation above. cc: <stable@vger.kernel.org> # 3.18 Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-01-18 21:28:10 +00:00
*/
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
while (percpu_counter_sum(&btp->bt_io_count))
delay(100);
xfs: prevent dropping ioend completions during buftarg wait xfs_wait_buftarg() waits for all pending I/O, drains the ioend completion workqueue and walks the LRU until all buffers in the cache have been released. This is traditionally an unmount operation` but the mechanism is also reused during filesystem freeze. xfs_wait_buftarg() invokes drain_workqueue() as part of the quiesce, which is intended more for a shutdown sequence in that it indicates to the queue that new operations are not expected once the drain has begun. New work jobs after this point result in a WARN_ON_ONCE() and are otherwise dropped. With filesystem freeze, however, read operations are allowed and can proceed during or after the workqueue drain. If such a read occurs during the drain sequence, the workqueue infrastructure complains about the queued ioend completion work item and drops it on the floor. As a result, the buffer remains on the LRU and the freeze never completes. Despite the fact that the overall buffer cache cleanup is not necessary during freeze, fix up this operation such that it is safe to invoke during non-unmount quiesce operations. Replace the drain_workqueue() call with flush_workqueue(), which runs a similar serialization on pending workqueue jobs without causing new jobs to be dropped. This is safe for unmount as unmount independently locks out new operations by the time xfs_wait_buftarg() is invoked. cc: <stable@vger.kernel.org> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-08-26 06:01:59 +00:00
flush_workqueue(btp->bt_mount->m_buf_workqueue);
}
void
xfs_buftarg_drain(
struct xfs_buftarg *btp)
{
LIST_HEAD(dispose);
int loop = 0;
bool write_fail = false;
xfs_buftarg_wait(btp);
xfs: log mount failures don't wait for buffers to be released Recently I've been seeing xfs/051 fail on 1k block size filesystems. Trying to trace the events during the test lead to the problem going away, indicating that it was a race condition that lead to this ASSERT failure: XFS: Assertion failed: atomic_read(&pag->pag_ref) == 0, file: fs/xfs/xfs_mount.c, line: 156 ..... [<ffffffff814e1257>] xfs_free_perag+0x87/0xb0 [<ffffffff814e21b9>] xfs_mountfs+0x4d9/0x900 [<ffffffff814e5dff>] xfs_fs_fill_super+0x3bf/0x4d0 [<ffffffff811d8800>] mount_bdev+0x180/0x1b0 [<ffffffff814e3ff5>] xfs_fs_mount+0x15/0x20 [<ffffffff811d90a8>] mount_fs+0x38/0x170 [<ffffffff811f4347>] vfs_kern_mount+0x67/0x120 [<ffffffff811f7018>] do_mount+0x218/0xd60 [<ffffffff811f7e5b>] SyS_mount+0x8b/0xd0 When I finally caught it with tracing enabled, I saw that AG 2 had an elevated reference count and a buffer was responsible for it. I tracked down the specific buffer, and found that it was missing the final reference count release that would put it back on the LRU and hence be found by xfs_wait_buftarg() calls in the log mount failure handling. The last four traces for the buffer before the assert were (trimmed for relevance) kworker/0:1-5259 xfs_buf_iodone: hold 2 lock 0 flags ASYNC kworker/0:1-5259 xfs_buf_ioerror: hold 2 lock 0 error -5 mount-7163 xfs_buf_lock_done: hold 2 lock 0 flags ASYNC mount-7163 xfs_buf_unlock: hold 2 lock 1 flags ASYNC This is an async write that is completing, so there's nobody waiting for it directly. Hence we call xfs_buf_relse() once all the processing is complete. That does: static inline void xfs_buf_relse(xfs_buf_t *bp) { xfs_buf_unlock(bp); xfs_buf_rele(bp); } Now, it's clear that mount is waiting on the buffer lock, and that it has been released by xfs_buf_relse() and gained by mount. This is expected, because at this point the mount process is in xfs_buf_delwri_submit() waiting for all the IO it submitted to complete. The mount process, however, is waiting on the lock for the buffer because it is in xfs_buf_delwri_submit(). This waits for IO completion, but it doesn't wait for the buffer reference owned by the IO to go away. The mount process collects all the completions, fails the log recovery, and the higher level code then calls xfs_wait_buftarg() to free all the remaining buffers in the filesystem. The issue is that on unlocking the buffer, the scheduler has decided that the mount process has higher priority than the the kworker thread that is running the IO completion, and so immediately switched contexts to the mount process from the semaphore unlock code, hence preventing the kworker thread from finishing the IO completion and releasing the IO reference to the buffer. Hence by the time that xfs_wait_buftarg() is run, the buffer still has an active reference and so isn't on the LRU list that the function walks to free the remaining buffers. Hence we miss that buffer and continue onwards to tear down the mount structures, at which time we get find a stray reference count on the perag structure. On a non-debug kernel, this will be ignored and the structure torn down and freed. Hence when the kworker thread is then rescheduled and the buffer released and freed, it will access a freed perag structure. The problem here is that when the log mount fails, we still need to quiesce the log to ensure that the IO workqueues have returned to idle before we run xfs_wait_buftarg(). By synchronising the workqueues, we ensure that all IO completions are fully processed, not just to the point where buffers have been unlocked. This ensures we don't end up in the situation above. cc: <stable@vger.kernel.org> # 3.18 Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-01-18 21:28:10 +00:00
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
/* loop until there is nothing left on the lru list. */
while (list_lru_count(&btp->bt_lru)) {
list_lru_walk(&btp->bt_lru, xfs_buftarg_drain_rele,
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
&dispose, LONG_MAX);
while (!list_empty(&dispose)) {
struct xfs_buf *bp;
bp = list_first_entry(&dispose, struct xfs_buf, b_lru);
list_del_init(&bp->b_lru);
xfs: abort metadata writeback on permanent errors If we are doing aysnc writeback of metadata, we can get write errors but have nobody to report them to. At the moment, we simply attempt to reissue the write from io completion in the hope that it's a transient error. When it's not a transient error, the buffer is stuck forever in this loop, and we cannot break out of it. Eventually, unmount will hang because the AIL cannot be emptied and everything goes downhill from them. To solve this problem, only retry the write IO once before aborting it. We don't throw the buffer away because some transient errors can last minutes (e.g. FC path failover) or even hours (thin provisioned devices that have run out of backing space) before they go away. Hence we really want to keep trying until we can't try any more. Because the buffer was not cleaned, however, it does not get removed from the AIL and hence the next pass across the AIL will start IO on it again. As such, we still get the "retry forever" semantics that we currently have, but we allow other access to the buffer in the mean time. Meanwhile the filesystem can continue to modify the buffer and relog it, so the IO errors won't hang the log or the filesystem. Now when we are pushing the AIL, we can see all these "permanent IO error" buffers and we can issue a warning about failures before we retry the IO. We can also catch these buffers when unmounting an issue a corruption warning, too. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
2013-12-12 05:34:38 +00:00
if (bp->b_flags & XBF_WRITE_FAIL) {
write_fail = true;
xfs_buf_alert_ratelimited(bp,
"XFS: Corruption Alert",
"Corruption Alert: Buffer at daddr 0x%llx had permanent write failures!",
(long long)xfs_buf_daddr(bp));
xfs: abort metadata writeback on permanent errors If we are doing aysnc writeback of metadata, we can get write errors but have nobody to report them to. At the moment, we simply attempt to reissue the write from io completion in the hope that it's a transient error. When it's not a transient error, the buffer is stuck forever in this loop, and we cannot break out of it. Eventually, unmount will hang because the AIL cannot be emptied and everything goes downhill from them. To solve this problem, only retry the write IO once before aborting it. We don't throw the buffer away because some transient errors can last minutes (e.g. FC path failover) or even hours (thin provisioned devices that have run out of backing space) before they go away. Hence we really want to keep trying until we can't try any more. Because the buffer was not cleaned, however, it does not get removed from the AIL and hence the next pass across the AIL will start IO on it again. As such, we still get the "retry forever" semantics that we currently have, but we allow other access to the buffer in the mean time. Meanwhile the filesystem can continue to modify the buffer and relog it, so the IO errors won't hang the log or the filesystem. Now when we are pushing the AIL, we can see all these "permanent IO error" buffers and we can issue a warning about failures before we retry the IO. We can also catch these buffers when unmounting an issue a corruption warning, too. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
2013-12-12 05:34:38 +00:00
}
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
xfs_buf_rele(bp);
}
if (loop++ != 0)
delay(100);
}
/*
* If one or more failed buffers were freed, that means dirty metadata
* was thrown away. This should only ever happen after I/O completion
* handling has elevated I/O error(s) to permanent failures and shuts
xfs: xfs_is_shutdown vs xlog_is_shutdown cage fight I've been chasing a recent resurgence in generic/388 recovery failure and/or corruption events. The events have largely been uninitialised inode chunks being tripped over in log recovery such as: XFS (pmem1): User initiated shutdown received. pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/xfs/xfs_fsops.c:500). Shutting down filesystem. XFS (pmem1): Please unmount the filesystem and rectify the problem(s) XFS (pmem1): Unmounting Filesystem XFS (pmem1): Mounting V5 Filesystem XFS (pmem1): Starting recovery (logdev: internal) XFS (pmem1): bad inode magic/vsn daddr 8723584 #0 (magic=1818) XFS (pmem1): Metadata corruption detected at xfs_inode_buf_verify+0x180/0x190, xfs_inode block 0x851c80 xfs_inode_buf_verify XFS (pmem1): Unmount and run xfs_repair XFS (pmem1): First 128 bytes of corrupted metadata buffer: 00000000: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000010: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000020: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000030: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000040: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000050: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000060: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000070: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ XFS (pmem1): metadata I/O error in "xlog_recover_items_pass2+0x52/0xc0" at daddr 0x851c80 len 32 error 117 XFS (pmem1): log mount/recovery failed: error -117 XFS (pmem1): log mount failed There have been isolated random other issues, too - xfs_repair fails because it finds some corruption in symlink blocks, rmap inconsistencies, etc - but they are nowhere near as common as the uninitialised inode chunk failure. The problem has clearly happened at runtime before recovery has run; I can see the ICREATE log item in the log shortly before the actively recovered range of the log. This means the ICREATE was definitely created and written to the log, but for some reason the tail of the log has been moved past the ordered buffer log item that tracks INODE_ALLOC buffers and, supposedly, prevents the tail of the log moving past the ICREATE log item before the inode chunk buffer is written to disk. Tracing the fsstress processes that are running when the filesystem shut down immediately pin-pointed the problem: user shutdown marks xfs_mount as shutdown godown-213341 [008] 6398.022871: console: [ 6397.915392] XFS (pmem1): User initiated shutdown received. ..... aild tries to push ordered inode cluster buffer xfsaild/pmem1-213314 [001] 6398.022974: xfs_buf_trylock: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 16 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_inode_item_push+0x8e xfsaild/pmem1-213314 [001] 6398.022976: xfs_ilock_nowait: dev 259:1 ino 0x851c80 flags ILOCK_SHARED caller xfs_iflush_cluster+0xae xfs_iflush_cluster() checks xfs_is_shutdown(), returns true, calls xfs_iflush_abort() to kill writeback of the inode. Inode is removed from AIL, drops cluster buffer reference. xfsaild/pmem1-213314 [001] 6398.022977: xfs_ail_delete: dev 259:1 lip 0xffff88880247ed80 old lsn 7/20344 new lsn 7/21000 type XFS_LI_INODE flags IN_AIL xfsaild/pmem1-213314 [001] 6398.022978: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 17 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_iflush_abort+0xd7 ..... All inodes on cluster buffer are aborted, then the cluster buffer itself is aborted and removed from the AIL *without writeback*: xfsaild/pmem1-213314 [001] 6398.023011: xfs_buf_error_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_ioend_fail+0x33 xfsaild/pmem1-213314 [001] 6398.023012: xfs_ail_delete: dev 259:1 lip 0xffff8888053efde8 old lsn 7/20344 new lsn 7/20344 type XFS_LI_BUF flags IN_AIL The inode buffer was at 7/20344 when it was removed from the AIL. xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_item_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_done+0x31 xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_relse+0x39 ..... Userspace is still running, doing stuff. an fsstress process runs syncfs() or sync() and we end up in sync_fs_one_sb() which issues a log force. This pushes on the CIL: fsstress-213322 [001] 6398.024430: xfs_fs_sync_fs: dev 259:1 m_features 0x20000000019ff6e9 opstate (clean|shutdown|inodegc|blockgc) s_flags 0x70810000 caller sync_fs_one_sb+0x26 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x0 caller xfs_fs_sync_fs+0x82 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x5f caller xfs_log_force+0x7c <...>-194402 [001] 6398.024467: kmem_alloc: size 176 flags 0x14 caller xlog_cil_push_work+0x9f And the CIL fills up iclogs with pending changes. This picks up the current tail from the AIL: <...>-194402 [001] 6398.024497: xlog_iclog_get_space: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x0 flags caller xlog_write+0x149 <...>-194402 [001] 6398.024498: xlog_iclog_switch: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x700005408 flags caller xlog_state_get_iclog_space+0x37e <...>-194402 [001] 6398.024521: xlog_iclog_release: dev 259:1 state XLOG_STATE_WANT_SYNC refcnt 1 offset 32256 lsn 0x700005408 flags caller xlog_write+0x5f9 <...>-194402 [001] 6398.024522: xfs_log_assign_tail_lsn: dev 259:1 new tail lsn 7/21000, old lsn 7/20344, last sync 7/21448 And it moves the tail of the log to 7/21000 from 7/20344. This *moves the tail of the log beyond the ICREATE transaction* that was at 7/20344 and pinned by the inode cluster buffer that was cancelled above. .... godown-213341 [008] 6398.027005: xfs_force_shutdown: dev 259:1 tag logerror flags log_io|force_umount file fs/xfs/xfs_fsops.c line_num 500 godown-213341 [008] 6398.027022: console: [ 6397.915406] pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 godown-213341 [008] 6398.030551: console: [ 6397.919546] XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/ And finally the log itself is now shutdown, stopping all further writes to the log. But this is too late to prevent the corruption that moving the tail of the log forwards after we start cancelling writeback causes. The fundamental problem here is that we are using the wrong shutdown checks for log items. We've long conflated mount shutdown with log shutdown state, and I started separating that recently with the atomic shutdown state changes in commit b36d4651e165 ("xfs: make forced shutdown processing atomic"). The changes in that commit series are directly responsible for being able to diagnose this issue because it clearly separated mount shutdown from log shutdown. Essentially, once we start cancelling writeback of log items and removing them from the AIL because the filesystem is shut down, we *cannot* update the journal because we may have cancelled the items that pin the tail of the log. That moves the tail of the log forwards without having written the metadata back, hence we have corrupt in memory state and writing to the journal propagates that to the on-disk state. What commit b36d4651e165 makes clear is that log item state needs to change relative to log shutdown, not mount shutdown. IOWs, anything that aborts metadata writeback needs to check log shutdown state because log items directly affect log consistency. Having them check mount shutdown state introduces the above race condition where we cancel metadata writeback before the log shuts down. To fix this, this patch works through all log items and converts shutdown checks to use xlog_is_shutdown() rather than xfs_is_shutdown(), so that we don't start aborting metadata writeback before we shut off journal writes. AFAICT, this race condition is a zero day IO error handling bug in XFS that dates back to the introduction of XLOG_IO_ERROR, XLOG_STATE_IOERROR and XFS_FORCED_SHUTDOWN back in January 1997. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:13 +00:00
* down the journal.
*/
if (write_fail) {
xfs: xfs_is_shutdown vs xlog_is_shutdown cage fight I've been chasing a recent resurgence in generic/388 recovery failure and/or corruption events. The events have largely been uninitialised inode chunks being tripped over in log recovery such as: XFS (pmem1): User initiated shutdown received. pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/xfs/xfs_fsops.c:500). Shutting down filesystem. XFS (pmem1): Please unmount the filesystem and rectify the problem(s) XFS (pmem1): Unmounting Filesystem XFS (pmem1): Mounting V5 Filesystem XFS (pmem1): Starting recovery (logdev: internal) XFS (pmem1): bad inode magic/vsn daddr 8723584 #0 (magic=1818) XFS (pmem1): Metadata corruption detected at xfs_inode_buf_verify+0x180/0x190, xfs_inode block 0x851c80 xfs_inode_buf_verify XFS (pmem1): Unmount and run xfs_repair XFS (pmem1): First 128 bytes of corrupted metadata buffer: 00000000: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000010: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000020: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000030: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000040: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000050: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000060: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ 00000070: 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 ................ XFS (pmem1): metadata I/O error in "xlog_recover_items_pass2+0x52/0xc0" at daddr 0x851c80 len 32 error 117 XFS (pmem1): log mount/recovery failed: error -117 XFS (pmem1): log mount failed There have been isolated random other issues, too - xfs_repair fails because it finds some corruption in symlink blocks, rmap inconsistencies, etc - but they are nowhere near as common as the uninitialised inode chunk failure. The problem has clearly happened at runtime before recovery has run; I can see the ICREATE log item in the log shortly before the actively recovered range of the log. This means the ICREATE was definitely created and written to the log, but for some reason the tail of the log has been moved past the ordered buffer log item that tracks INODE_ALLOC buffers and, supposedly, prevents the tail of the log moving past the ICREATE log item before the inode chunk buffer is written to disk. Tracing the fsstress processes that are running when the filesystem shut down immediately pin-pointed the problem: user shutdown marks xfs_mount as shutdown godown-213341 [008] 6398.022871: console: [ 6397.915392] XFS (pmem1): User initiated shutdown received. ..... aild tries to push ordered inode cluster buffer xfsaild/pmem1-213314 [001] 6398.022974: xfs_buf_trylock: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 16 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_inode_item_push+0x8e xfsaild/pmem1-213314 [001] 6398.022976: xfs_ilock_nowait: dev 259:1 ino 0x851c80 flags ILOCK_SHARED caller xfs_iflush_cluster+0xae xfs_iflush_cluster() checks xfs_is_shutdown(), returns true, calls xfs_iflush_abort() to kill writeback of the inode. Inode is removed from AIL, drops cluster buffer reference. xfsaild/pmem1-213314 [001] 6398.022977: xfs_ail_delete: dev 259:1 lip 0xffff88880247ed80 old lsn 7/20344 new lsn 7/21000 type XFS_LI_INODE flags IN_AIL xfsaild/pmem1-213314 [001] 6398.022978: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 17 pincount 0 lock 0 flags DONE|INODES|PAGES caller xfs_iflush_abort+0xd7 ..... All inodes on cluster buffer are aborted, then the cluster buffer itself is aborted and removed from the AIL *without writeback*: xfsaild/pmem1-213314 [001] 6398.023011: xfs_buf_error_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_ioend_fail+0x33 xfsaild/pmem1-213314 [001] 6398.023012: xfs_ail_delete: dev 259:1 lip 0xffff8888053efde8 old lsn 7/20344 new lsn 7/20344 type XFS_LI_BUF flags IN_AIL The inode buffer was at 7/20344 when it was removed from the AIL. xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_item_relse: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_done+0x31 xfsaild/pmem1-213314 [001] 6398.023012: xfs_buf_rele: dev 259:1 daddr 0x851c80 bbcount 0x20 hold 2 pincount 0 lock 0 flags ASYNC|DONE|STALE|INODES|PAGES caller xfs_buf_item_relse+0x39 ..... Userspace is still running, doing stuff. an fsstress process runs syncfs() or sync() and we end up in sync_fs_one_sb() which issues a log force. This pushes on the CIL: fsstress-213322 [001] 6398.024430: xfs_fs_sync_fs: dev 259:1 m_features 0x20000000019ff6e9 opstate (clean|shutdown|inodegc|blockgc) s_flags 0x70810000 caller sync_fs_one_sb+0x26 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x0 caller xfs_fs_sync_fs+0x82 fsstress-213322 [001] 6398.024430: xfs_log_force: dev 259:1 lsn 0x5f caller xfs_log_force+0x7c <...>-194402 [001] 6398.024467: kmem_alloc: size 176 flags 0x14 caller xlog_cil_push_work+0x9f And the CIL fills up iclogs with pending changes. This picks up the current tail from the AIL: <...>-194402 [001] 6398.024497: xlog_iclog_get_space: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x0 flags caller xlog_write+0x149 <...>-194402 [001] 6398.024498: xlog_iclog_switch: dev 259:1 state XLOG_STATE_ACTIVE refcnt 1 offset 0 lsn 0x700005408 flags caller xlog_state_get_iclog_space+0x37e <...>-194402 [001] 6398.024521: xlog_iclog_release: dev 259:1 state XLOG_STATE_WANT_SYNC refcnt 1 offset 32256 lsn 0x700005408 flags caller xlog_write+0x5f9 <...>-194402 [001] 6398.024522: xfs_log_assign_tail_lsn: dev 259:1 new tail lsn 7/21000, old lsn 7/20344, last sync 7/21448 And it moves the tail of the log to 7/21000 from 7/20344. This *moves the tail of the log beyond the ICREATE transaction* that was at 7/20344 and pinned by the inode cluster buffer that was cancelled above. .... godown-213341 [008] 6398.027005: xfs_force_shutdown: dev 259:1 tag logerror flags log_io|force_umount file fs/xfs/xfs_fsops.c line_num 500 godown-213341 [008] 6398.027022: console: [ 6397.915406] pmem1: writeback error on inode 12621949, offset 1019904, sector 12968096 godown-213341 [008] 6398.030551: console: [ 6397.919546] XFS (pmem1): Log I/O Error (0x6) detected at xfs_fs_goingdown+0xa3/0xf0 (fs/ And finally the log itself is now shutdown, stopping all further writes to the log. But this is too late to prevent the corruption that moving the tail of the log forwards after we start cancelling writeback causes. The fundamental problem here is that we are using the wrong shutdown checks for log items. We've long conflated mount shutdown with log shutdown state, and I started separating that recently with the atomic shutdown state changes in commit b36d4651e165 ("xfs: make forced shutdown processing atomic"). The changes in that commit series are directly responsible for being able to diagnose this issue because it clearly separated mount shutdown from log shutdown. Essentially, once we start cancelling writeback of log items and removing them from the AIL because the filesystem is shut down, we *cannot* update the journal because we may have cancelled the items that pin the tail of the log. That moves the tail of the log forwards without having written the metadata back, hence we have corrupt in memory state and writing to the journal propagates that to the on-disk state. What commit b36d4651e165 makes clear is that log item state needs to change relative to log shutdown, not mount shutdown. IOWs, anything that aborts metadata writeback needs to check log shutdown state because log items directly affect log consistency. Having them check mount shutdown state introduces the above race condition where we cancel metadata writeback before the log shuts down. To fix this, this patch works through all log items and converts shutdown checks to use xlog_is_shutdown() rather than xfs_is_shutdown(), so that we don't start aborting metadata writeback before we shut off journal writes. AFAICT, this race condition is a zero day IO error handling bug in XFS that dates back to the introduction of XLOG_IO_ERROR, XLOG_STATE_IOERROR and XFS_FORCED_SHUTDOWN back in January 1997. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:13 +00:00
ASSERT(xlog_is_shutdown(btp->bt_mount->m_log));
xfs_alert(btp->bt_mount,
"Please run xfs_repair to determine the extent of the problem.");
}
}
static enum lru_status
xfs_buftarg_isolate(
struct list_head *item,
list_lru: add helpers to isolate items Currently, the isolate callback passed to the list_lru_walk family of functions is supposed to just delete an item from the list upon returning LRU_REMOVED or LRU_REMOVED_RETRY, while nr_items counter is fixed by __list_lru_walk_one after the callback returns. Since the callback is allowed to drop the lock after removing an item (it has to return LRU_REMOVED_RETRY then), the nr_items can be less than the actual number of elements on the list even if we check them under the lock. This makes it difficult to move items from one list_lru_one to another, which is required for per-memcg list_lru reparenting - we can't just splice the lists, we have to move entries one by one. This patch therefore introduces helpers that must be used by callback functions to isolate items instead of raw list_del/list_move. These are list_lru_isolate and list_lru_isolate_move. They not only remove the entry from the list, but also fix the nr_items counter, making sure nr_items always reflects the actual number of elements on the list if checked under the appropriate lock. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Dave Chinner <david@fromorbit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:35 +00:00
struct list_lru_one *lru,
spinlock_t *lru_lock,
void *arg)
{
struct xfs_buf *bp = container_of(item, struct xfs_buf, b_lru);
struct list_head *dispose = arg;
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
/*
* we are inverting the lru lock/bp->b_lock here, so use a trylock.
* If we fail to get the lock, just skip it.
*/
if (!spin_trylock(&bp->b_lock))
return LRU_SKIP;
/*
* Decrement the b_lru_ref count unless the value is already
* zero. If the value is already zero, we need to reclaim the
* buffer, otherwise it gets another trip through the LRU.
*/
if (atomic_add_unless(&bp->b_lru_ref, -1, 0)) {
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
spin_unlock(&bp->b_lock);
return LRU_ROTATE;
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
}
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
bp->b_state |= XFS_BSTATE_DISPOSE;
list_lru: add helpers to isolate items Currently, the isolate callback passed to the list_lru_walk family of functions is supposed to just delete an item from the list upon returning LRU_REMOVED or LRU_REMOVED_RETRY, while nr_items counter is fixed by __list_lru_walk_one after the callback returns. Since the callback is allowed to drop the lock after removing an item (it has to return LRU_REMOVED_RETRY then), the nr_items can be less than the actual number of elements on the list even if we check them under the lock. This makes it difficult to move items from one list_lru_one to another, which is required for per-memcg list_lru reparenting - we can't just splice the lists, we have to move entries one by one. This patch therefore introduces helpers that must be used by callback functions to isolate items instead of raw list_del/list_move. These are list_lru_isolate and list_lru_isolate_move. They not only remove the entry from the list, but also fix the nr_items counter, making sure nr_items always reflects the actual number of elements on the list if checked under the appropriate lock. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Dave Chinner <david@fromorbit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:35 +00:00
list_lru_isolate_move(lru, item, dispose);
xfs: rework buffer dispose list tracking In converting the buffer lru lists to use the generic code, the locking for marking the buffers as on the dispose list was lost. This results in confusion in LRU buffer tracking and acocunting, resulting in reference counts being mucked up and filesystem beig unmountable. To fix this, introduce an internal buffer spinlock to protect the state field that holds the dispose list information. Because there is now locking needed around xfs_buf_lru_add/del, and they are used in exactly one place each two lines apart, get rid of the wrappers and code the logic directly in place. Further, the LRU emptying code used on unmount is less than optimal. Convert it to use a dispose list as per a normal shrinker walk, and repeat the walk that fills the dispose list until the LRU is empty. Thi avoids needing to drop and regain the LRU lock for every item being freed, and allows the same logic as the shrinker isolate call to be used. Simpler, easier to understand. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:06 +00:00
spin_unlock(&bp->b_lock);
return LRU_REMOVED;
}
static unsigned long
xfs_buftarg_shrink_scan(
struct shrinker *shrink,
struct shrink_control *sc)
{
struct xfs_buftarg *btp = container_of(shrink,
struct xfs_buftarg, bt_shrinker);
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
LIST_HEAD(dispose);
unsigned long freed;
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
list_lru: introduce list_lru_shrink_{count,walk} Kmem accounting of memcg is unusable now, because it lacks slab shrinker support. That means when we hit the limit we will get ENOMEM w/o any chance to recover. What we should do then is to call shrink_slab, which would reclaim old inode/dentry caches from this cgroup. This is what this patch set is intended to do. Basically, it does two things. First, it introduces the notion of per-memcg slab shrinker. A shrinker that wants to reclaim objects per cgroup should mark itself as SHRINKER_MEMCG_AWARE. Then it will be passed the memory cgroup to scan from in shrink_control->memcg. For such shrinkers shrink_slab iterates over the whole cgroup subtree under the target cgroup and calls the shrinker for each kmem-active memory cgroup. Secondly, this patch set makes the list_lru structure per-memcg. It's done transparently to list_lru users - everything they have to do is to tell list_lru_init that they want memcg-aware list_lru. Then the list_lru will automatically distribute objects among per-memcg lists basing on which cgroup the object is accounted to. This way to make FS shrinkers (icache, dcache) memcg-aware we only need to make them use memcg-aware list_lru, and this is what this patch set does. As before, this patch set only enables per-memcg kmem reclaim when the pressure goes from memory.limit, not from memory.kmem.limit. Handling memory.kmem.limit is going to be tricky due to GFP_NOFS allocations, and it is still unclear whether we will have this knob in the unified hierarchy. This patch (of 9): NUMA aware slab shrinkers use the list_lru structure to distribute objects coming from different NUMA nodes to different lists. Whenever such a shrinker needs to count or scan objects from a particular node, it issues commands like this: count = list_lru_count_node(lru, sc->nid); freed = list_lru_walk_node(lru, sc->nid, isolate_func, isolate_arg, &sc->nr_to_scan); where sc is an instance of the shrink_control structure passed to it from vmscan. To simplify this, let's add special list_lru functions to be used by shrinkers, list_lru_shrink_count() and list_lru_shrink_walk(), which consolidate the nid and nr_to_scan arguments in the shrink_control structure. This will also allow us to avoid patching shrinkers that use list_lru when we make shrink_slab() per-memcg - all we will have to do is extend the shrink_control structure to include the target memcg and make list_lru_shrink_{count,walk} handle this appropriately. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Suggested-by: Dave Chinner <david@fromorbit.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Glauber Costa <glommer@gmail.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:58:47 +00:00
freed = list_lru_shrink_walk(&btp->bt_lru, sc,
xfs_buftarg_isolate, &dispose);
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
while (!list_empty(&dispose)) {
struct xfs_buf *bp;
xfs: add a lru to the XFS buffer cache Introduce a per-buftarg LRU for memory reclaim to operate on. This is the last piece we need to put in place so that we can fully control the buffer lifecycle. This allows XFS to be responsibile for maintaining the working set of buffers under memory pressure instead of relying on the VM reclaim not to take pages we need out from underneath us. The implementation introduces a b_lru_ref counter into the buffer. This is currently set to 1 whenever the buffer is referenced and so is used to determine if the buffer should be added to the LRU or not when freed. Effectively it allows lazy LRU initialisation of the buffer so we do not need to touch the LRU list and locks in xfs_buf_find(). Instead, when the buffer is being released and we drop the last reference to it, we check the b_lru_ref count and if it is none zero we re-add the buffer reference and add the inode to the LRU. The b_lru_ref counter is decremented by the shrinker, and whenever the shrinker comes across a buffer with a zero b_lru_ref counter, if released the LRU reference on the buffer. In the absence of a lookup race, this will result in the buffer being freed. This counting mechanism is used instead of a reference flag so that it is simple to re-introduce buffer-type specific reclaim reference counts to prioritise reclaim more effectively. We still have all those hooks in the XFS code, so this will provide the infrastructure to re-implement that functionality. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-12-02 05:30:55 +00:00
bp = list_first_entry(&dispose, struct xfs_buf, b_lru);
list_del_init(&bp->b_lru);
xfs_buf_rele(bp);
}
return freed;
}
static unsigned long
xfs_buftarg_shrink_count(
struct shrinker *shrink,
struct shrink_control *sc)
{
struct xfs_buftarg *btp = container_of(shrink,
struct xfs_buftarg, bt_shrinker);
list_lru: introduce list_lru_shrink_{count,walk} Kmem accounting of memcg is unusable now, because it lacks slab shrinker support. That means when we hit the limit we will get ENOMEM w/o any chance to recover. What we should do then is to call shrink_slab, which would reclaim old inode/dentry caches from this cgroup. This is what this patch set is intended to do. Basically, it does two things. First, it introduces the notion of per-memcg slab shrinker. A shrinker that wants to reclaim objects per cgroup should mark itself as SHRINKER_MEMCG_AWARE. Then it will be passed the memory cgroup to scan from in shrink_control->memcg. For such shrinkers shrink_slab iterates over the whole cgroup subtree under the target cgroup and calls the shrinker for each kmem-active memory cgroup. Secondly, this patch set makes the list_lru structure per-memcg. It's done transparently to list_lru users - everything they have to do is to tell list_lru_init that they want memcg-aware list_lru. Then the list_lru will automatically distribute objects among per-memcg lists basing on which cgroup the object is accounted to. This way to make FS shrinkers (icache, dcache) memcg-aware we only need to make them use memcg-aware list_lru, and this is what this patch set does. As before, this patch set only enables per-memcg kmem reclaim when the pressure goes from memory.limit, not from memory.kmem.limit. Handling memory.kmem.limit is going to be tricky due to GFP_NOFS allocations, and it is still unclear whether we will have this knob in the unified hierarchy. This patch (of 9): NUMA aware slab shrinkers use the list_lru structure to distribute objects coming from different NUMA nodes to different lists. Whenever such a shrinker needs to count or scan objects from a particular node, it issues commands like this: count = list_lru_count_node(lru, sc->nid); freed = list_lru_walk_node(lru, sc->nid, isolate_func, isolate_arg, &sc->nr_to_scan); where sc is an instance of the shrink_control structure passed to it from vmscan. To simplify this, let's add special list_lru functions to be used by shrinkers, list_lru_shrink_count() and list_lru_shrink_walk(), which consolidate the nid and nr_to_scan arguments in the shrink_control structure. This will also allow us to avoid patching shrinkers that use list_lru when we make shrink_slab() per-memcg - all we will have to do is extend the shrink_control structure to include the target memcg and make list_lru_shrink_{count,walk} handle this appropriately. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Suggested-by: Dave Chinner <david@fromorbit.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Glauber Costa <glommer@gmail.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:58:47 +00:00
return list_lru_shrink_count(&btp->bt_lru, sc);
}
void
xfs_free_buftarg(
struct xfs_buftarg *btp)
{
unregister_shrinker(&btp->bt_shrinker);
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
ASSERT(percpu_counter_sum(&btp->bt_io_count) == 0);
percpu_counter_destroy(&btp->bt_io_count);
list_lru_destroy(&btp->bt_lru);
blkdev_issue_flush(btp->bt_bdev);
fs_put_dax(btp->bt_daxdev, btp->bt_mount);
kmem_free(btp);
}
int
xfs_setsize_buftarg(
xfs_buftarg_t *btp,
unsigned int sectorsize)
{
xfs: allow logical-sector sized O_DIRECT Some time ago, mkfs.xfs started picking the storage physical sector size as the default filesystem "sector size" in order to avoid RMW costs incurred by doing IOs at logical sector size alignments. However, this means that for a filesystem made with i.e. a 4k sector size on an "advanced format" 4k/512 disk, 512-byte direct IOs are no longer allowed. This means that XFS has essentially turned this AF drive into a hard 4K device, from the filesystem on up. XFS's mkfs-specified "sector size" is really just controlling the minimum size & alignment of filesystem metadata. There is no real need to tightly couple XFS's minimal metadata size to the minimum allowed direct IO size; XFS can continue doing metadata in optimal sizes, but still allow smaller DIOs for apps which issue them, for whatever reason. This patch adds a new field to the xfs_buftarg, so that we now track 2 sizes: 1) The metadata sector size, which is the minimum unit and alignment of IO which will be performed by metadata operations. 2) The device logical sector size The first is used internally by the file system for metadata alignment and IOs. The second is used for the minimum allowed direct IO alignment. This has passed xfstests on filesystems made with 4k sectors, including when run under the patch I sent to ignore XFS_IOC_DIOINFO, and issue 512 DIOs anyway. I also directly tested end of block behavior on preallocated, sparse, and existing files when we do a 512 IO into a 4k file on a 4k-sector filesystem, to be sure there were no unexpected behaviors. Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2014-01-21 22:46:23 +00:00
/* Set up metadata sector size info */
btp->bt_meta_sectorsize = sectorsize;
btp->bt_meta_sectormask = sectorsize - 1;
if (set_blocksize(btp->bt_bdev, sectorsize)) {
xfs_warn(btp->bt_mount,
"Cannot set_blocksize to %u on device %pg",
sectorsize, btp->bt_bdev);
return -EINVAL;
}
xfs: allow logical-sector sized O_DIRECT Some time ago, mkfs.xfs started picking the storage physical sector size as the default filesystem "sector size" in order to avoid RMW costs incurred by doing IOs at logical sector size alignments. However, this means that for a filesystem made with i.e. a 4k sector size on an "advanced format" 4k/512 disk, 512-byte direct IOs are no longer allowed. This means that XFS has essentially turned this AF drive into a hard 4K device, from the filesystem on up. XFS's mkfs-specified "sector size" is really just controlling the minimum size & alignment of filesystem metadata. There is no real need to tightly couple XFS's minimal metadata size to the minimum allowed direct IO size; XFS can continue doing metadata in optimal sizes, but still allow smaller DIOs for apps which issue them, for whatever reason. This patch adds a new field to the xfs_buftarg, so that we now track 2 sizes: 1) The metadata sector size, which is the minimum unit and alignment of IO which will be performed by metadata operations. 2) The device logical sector size The first is used internally by the file system for metadata alignment and IOs. The second is used for the minimum allowed direct IO alignment. This has passed xfstests on filesystems made with 4k sectors, including when run under the patch I sent to ignore XFS_IOC_DIOINFO, and issue 512 DIOs anyway. I also directly tested end of block behavior on preallocated, sparse, and existing files when we do a 512 IO into a 4k file on a 4k-sector filesystem, to be sure there were no unexpected behaviors. Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2014-01-21 22:46:23 +00:00
/* Set up device logical sector size mask */
btp->bt_logical_sectorsize = bdev_logical_block_size(btp->bt_bdev);
btp->bt_logical_sectormask = bdev_logical_block_size(btp->bt_bdev) - 1;
return 0;
}
/*
* When allocating the initial buffer target we have not yet
* read in the superblock, so don't know what sized sectors
* are being used at this early stage. Play safe.
*/
STATIC int
xfs_setsize_buftarg_early(
xfs_buftarg_t *btp,
struct block_device *bdev)
{
return xfs_setsize_buftarg(btp, bdev_logical_block_size(bdev));
}
struct xfs_buftarg *
xfs_alloc_buftarg(
struct xfs_mount *mp,
struct block_device *bdev)
{
xfs_buftarg_t *btp;
const struct dax_holder_operations *ops = NULL;
#if defined(CONFIG_FS_DAX) && defined(CONFIG_MEMORY_FAILURE)
ops = &xfs_dax_holder_operations;
#endif
btp = kmem_zalloc(sizeof(*btp), KM_NOFS);
btp->bt_mount = mp;
btp->bt_dev = bdev->bd_dev;
btp->bt_bdev = bdev;
btp->bt_daxdev = fs_dax_get_by_bdev(bdev, &btp->bt_dax_part_off,
mp, ops);
2011-03-25 22:16:45 +00:00
/*
* Buffer IO error rate limiting. Limit it to no more than 10 messages
* per 30 seconds so as to not spam logs too much on repeated errors.
*/
ratelimit_state_init(&btp->bt_ioerror_rl, 30 * HZ,
DEFAULT_RATELIMIT_BURST);
if (xfs_setsize_buftarg_early(btp, bdev))
goto error_free;
list_lru: dynamically adjust node arrays We currently use a compile-time constant to size the node array for the list_lru structure. Due to this, we don't need to allocate any memory at initialization time. But as a consequence, the structures that contain embedded list_lru lists can become way too big (the superblock for instance contains two of them). This patch aims at ameliorating this situation by dynamically allocating the node arrays with the firmware provided nr_node_ids. Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: Dave Chinner <dchinner@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:18 +00:00
if (list_lru_init(&btp->bt_lru))
goto error_free;
list_lru: dynamically adjust node arrays We currently use a compile-time constant to size the node array for the list_lru structure. Due to this, we don't need to allocate any memory at initialization time. But as a consequence, the structures that contain embedded list_lru lists can become way too big (the superblock for instance contains two of them). This patch aims at ameliorating this situation by dynamically allocating the node arrays with the firmware provided nr_node_ids. Signed-off-by: Glauber Costa <glommer@openvz.org> Cc: Dave Chinner <dchinner@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: Arve Hjønnevåg <arve@android.com> Cc: Carlos Maiolino <cmaiolino@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: David Rientjes <rientjes@google.com> Cc: Gleb Natapov <gleb@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: J. Bruce Fields <bfields@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kent Overstreet <koverstreet@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2013-08-28 00:18:18 +00:00
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
if (percpu_counter_init(&btp->bt_io_count, 0, GFP_KERNEL))
goto error_lru;
xfs: track and serialize in-flight async buffers against unmount Newly allocated XFS metadata buffers are added to the LRU once the hold count is released, which typically occurs after I/O completion. There is no other mechanism at current that tracks the existence or I/O state of a new buffer. Further, readahead I/O tends to be submitted asynchronously by nature, which means the I/O can remain in flight and actually complete long after the calling context is gone. This means that file descriptors or any other holds on the filesystem can be released, allowing the filesystem to be unmounted while I/O is still in flight. When I/O completion occurs, core data structures may have been freed, causing completion to run into invalid memory accesses and likely to panic. This problem is reproduced on XFS via directory readahead. A filesystem is mounted, a directory is opened/closed and the filesystem immediately unmounted. The open/close cycle triggers a directory readahead that if delayed long enough, runs buffer I/O completion after the unmount has completed. To address this problem, add a mechanism to track all in-flight, asynchronous buffers using per-cpu counters in the buftarg. The buffer is accounted on the first I/O submission after the current reference is acquired and unaccounted once the buffer is returned to the LRU or freed. Update xfs_wait_buftarg() to wait on all in-flight I/O before walking the LRU list. Once in-flight I/O has completed and the workqueue has drained, all new buffers should have been released onto the LRU. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2016-07-20 01:15:28 +00:00
btp->bt_shrinker.count_objects = xfs_buftarg_shrink_count;
btp->bt_shrinker.scan_objects = xfs_buftarg_shrink_scan;
btp->bt_shrinker.seeks = DEFAULT_SEEKS;
btp->bt_shrinker.flags = SHRINKER_NUMA_AWARE;
mm: shrinkers: provide shrinkers with names Currently shrinkers are anonymous objects. For debugging purposes they can be identified by count/scan function names, but it's not always useful: e.g. for superblock's shrinkers it's nice to have at least an idea of to which superblock the shrinker belongs. This commit adds names to shrinkers. register_shrinker() and prealloc_shrinker() functions are extended to take a format and arguments to master a name. In some cases it's not possible to determine a good name at the time when a shrinker is allocated. For such cases shrinker_debugfs_rename() is provided. The expected format is: <subsystem>-<shrinker_type>[:<instance>]-<id> For some shrinkers an instance can be encoded as (MAJOR:MINOR) pair. After this change the shrinker debugfs directory looks like: $ cd /sys/kernel/debug/shrinker/ $ ls dquota-cache-16 sb-devpts-28 sb-proc-47 sb-tmpfs-42 mm-shadow-18 sb-devtmpfs-5 sb-proc-48 sb-tmpfs-43 mm-zspool:zram0-34 sb-hugetlbfs-17 sb-pstore-31 sb-tmpfs-44 rcu-kfree-0 sb-hugetlbfs-33 sb-rootfs-2 sb-tmpfs-49 sb-aio-20 sb-iomem-12 sb-securityfs-6 sb-tracefs-13 sb-anon_inodefs-15 sb-mqueue-21 sb-selinuxfs-22 sb-xfs:vda1-36 sb-bdev-3 sb-nsfs-4 sb-sockfs-8 sb-zsmalloc-19 sb-bpf-32 sb-pipefs-14 sb-sysfs-26 thp-deferred_split-10 sb-btrfs:vda2-24 sb-proc-25 sb-tmpfs-1 thp-zero-9 sb-cgroup2-30 sb-proc-39 sb-tmpfs-27 xfs-buf:vda1-37 sb-configfs-23 sb-proc-41 sb-tmpfs-29 xfs-inodegc:vda1-38 sb-dax-11 sb-proc-45 sb-tmpfs-35 sb-debugfs-7 sb-proc-46 sb-tmpfs-40 [roman.gushchin@linux.dev: fix build warnings] Link: https://lkml.kernel.org/r/Yr+ZTnLb9lJk6fJO@castle Reported-by: kernel test robot <lkp@intel.com> Link: https://lkml.kernel.org/r/20220601032227.4076670-4-roman.gushchin@linux.dev Signed-off-by: Roman Gushchin <roman.gushchin@linux.dev> Cc: Christophe JAILLET <christophe.jaillet@wanadoo.fr> Cc: Dave Chinner <dchinner@redhat.com> Cc: Hillf Danton <hdanton@sina.com> Cc: Kent Overstreet <kent.overstreet@gmail.com> Cc: Muchun Song <songmuchun@bytedance.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2022-06-01 03:22:24 +00:00
if (register_shrinker(&btp->bt_shrinker, "xfs-buf:%s",
mp->m_super->s_id))
goto error_pcpu;
return btp;
error_pcpu:
percpu_counter_destroy(&btp->bt_io_count);
error_lru:
list_lru_destroy(&btp->bt_lru);
error_free:
kmem_free(btp);
return NULL;
}
/*
* Cancel a delayed write list.
*
* Remove each buffer from the list, clear the delwri queue flag and drop the
* associated buffer reference.
*/
void
xfs_buf_delwri_cancel(
struct list_head *list)
{
struct xfs_buf *bp;
while (!list_empty(list)) {
bp = list_first_entry(list, struct xfs_buf, b_list);
xfs_buf_lock(bp);
bp->b_flags &= ~_XBF_DELWRI_Q;
list_del_init(&bp->b_list);
xfs_buf_relse(bp);
}
}
/*
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
* Add a buffer to the delayed write list.
*
* This queues a buffer for writeout if it hasn't already been. Note that
* neither this routine nor the buffer list submission functions perform
* any internal synchronization. It is expected that the lists are thread-local
* to the callers.
*
* Returns true if we queued up the buffer, or false if it already had
* been on the buffer list.
*/
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
bool
xfs_buf_delwri_queue(
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
struct xfs_buf *bp,
struct list_head *list)
{
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
ASSERT(xfs_buf_islocked(bp));
ASSERT(!(bp->b_flags & XBF_READ));
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
/*
* If the buffer is already marked delwri it already is queued up
* by someone else for imediate writeout. Just ignore it in that
* case.
*/
if (bp->b_flags & _XBF_DELWRI_Q) {
trace_xfs_buf_delwri_queued(bp, _RET_IP_);
return false;
}
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
trace_xfs_buf_delwri_queue(bp, _RET_IP_);
/*
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
* If a buffer gets written out synchronously or marked stale while it
* is on a delwri list we lazily remove it. To do this, the other party
* clears the _XBF_DELWRI_Q flag but otherwise leaves the buffer alone.
* It remains referenced and on the list. In a rare corner case it
* might get readded to a delwri list after the synchronous writeout, in
* which case we need just need to re-add the flag here.
*/
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
bp->b_flags |= _XBF_DELWRI_Q;
if (list_empty(&bp->b_list)) {
atomic_inc(&bp->b_hold);
list_add_tail(&bp->b_list, list);
}
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
return true;
}
/*
* Compare function is more complex than it needs to be because
* the return value is only 32 bits and we are doing comparisons
* on 64 bit values
*/
static int
xfs_buf_cmp(
void *priv,
const struct list_head *a,
const struct list_head *b)
{
struct xfs_buf *ap = container_of(a, struct xfs_buf, b_list);
struct xfs_buf *bp = container_of(b, struct xfs_buf, b_list);
xfs_daddr_t diff;
diff = ap->b_maps[0].bm_bn - bp->b_maps[0].bm_bn;
if (diff < 0)
return -1;
if (diff > 0)
return 1;
return 0;
}
/*
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
* Submit buffers for write. If wait_list is specified, the buffers are
* submitted using sync I/O and placed on the wait list such that the caller can
* iowait each buffer. Otherwise async I/O is used and the buffers are released
* at I/O completion time. In either case, buffers remain locked until I/O
* completes and the buffer is released from the queue.
*/
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
static int
xfs_buf_delwri_submit_buffers(
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
struct list_head *buffer_list,
struct list_head *wait_list)
{
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
struct xfs_buf *bp, *n;
int pinned = 0;
struct blk_plug plug;
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
list_sort(NULL, buffer_list, xfs_buf_cmp);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
blk_start_plug(&plug);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
list_for_each_entry_safe(bp, n, buffer_list, b_list) {
if (!wait_list) {
xfs: check buffer pin state after locking in delwri_submit AIL flushing can get stuck here: [316649.005769] INFO: task xfsaild/pmem1:324525 blocked for more than 123 seconds. [316649.007807] Not tainted 5.17.0-rc6-dgc+ #975 [316649.009186] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. [316649.011720] task:xfsaild/pmem1 state:D stack:14544 pid:324525 ppid: 2 flags:0x00004000 [316649.014112] Call Trace: [316649.014841] <TASK> [316649.015492] __schedule+0x30d/0x9e0 [316649.017745] schedule+0x55/0xd0 [316649.018681] io_schedule+0x4b/0x80 [316649.019683] xfs_buf_wait_unpin+0x9e/0xf0 [316649.021850] __xfs_buf_submit+0x14a/0x230 [316649.023033] xfs_buf_delwri_submit_buffers+0x107/0x280 [316649.024511] xfs_buf_delwri_submit_nowait+0x10/0x20 [316649.025931] xfsaild+0x27e/0x9d0 [316649.028283] kthread+0xf6/0x120 [316649.030602] ret_from_fork+0x1f/0x30 in the situation where flushing gets preempted between the unpin check and the buffer trylock under nowait conditions: blk_start_plug(&plug); list_for_each_entry_safe(bp, n, buffer_list, b_list) { if (!wait_list) { if (xfs_buf_ispinned(bp)) { pinned++; continue; } Here >>>>>> if (!xfs_buf_trylock(bp)) continue; This means submission is stuck until something else triggers a log force to unpin the buffer. To get onto the delwri list to begin with, the buffer pin state has already been checked, and hence it's relatively rare we get a race between flushing and encountering a pinned buffer in delwri submission to begin with. Further, to increase the pin count the buffer has to be locked, so the only way we can hit this race without failing the trylock is to be preempted between the pincount check seeing zero and the trylock being run. Hence to avoid this problem, just invert the order of trylock vs pin check. We shouldn't hit that many pinned buffers here, so optimising away the trylock for pinned buffers should not matter for performance at all. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Chandan Babu R <chandan.babu@oracle.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:10 +00:00
if (!xfs_buf_trylock(bp))
continue;
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
if (xfs_buf_ispinned(bp)) {
xfs: check buffer pin state after locking in delwri_submit AIL flushing can get stuck here: [316649.005769] INFO: task xfsaild/pmem1:324525 blocked for more than 123 seconds. [316649.007807] Not tainted 5.17.0-rc6-dgc+ #975 [316649.009186] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. [316649.011720] task:xfsaild/pmem1 state:D stack:14544 pid:324525 ppid: 2 flags:0x00004000 [316649.014112] Call Trace: [316649.014841] <TASK> [316649.015492] __schedule+0x30d/0x9e0 [316649.017745] schedule+0x55/0xd0 [316649.018681] io_schedule+0x4b/0x80 [316649.019683] xfs_buf_wait_unpin+0x9e/0xf0 [316649.021850] __xfs_buf_submit+0x14a/0x230 [316649.023033] xfs_buf_delwri_submit_buffers+0x107/0x280 [316649.024511] xfs_buf_delwri_submit_nowait+0x10/0x20 [316649.025931] xfsaild+0x27e/0x9d0 [316649.028283] kthread+0xf6/0x120 [316649.030602] ret_from_fork+0x1f/0x30 in the situation where flushing gets preempted between the unpin check and the buffer trylock under nowait conditions: blk_start_plug(&plug); list_for_each_entry_safe(bp, n, buffer_list, b_list) { if (!wait_list) { if (xfs_buf_ispinned(bp)) { pinned++; continue; } Here >>>>>> if (!xfs_buf_trylock(bp)) continue; This means submission is stuck until something else triggers a log force to unpin the buffer. To get onto the delwri list to begin with, the buffer pin state has already been checked, and hence it's relatively rare we get a race between flushing and encountering a pinned buffer in delwri submission to begin with. Further, to increase the pin count the buffer has to be locked, so the only way we can hit this race without failing the trylock is to be preempted between the pincount check seeing zero and the trylock being run. Hence to avoid this problem, just invert the order of trylock vs pin check. We shouldn't hit that many pinned buffers here, so optimising away the trylock for pinned buffers should not matter for performance at all. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Chandan Babu R <chandan.babu@oracle.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2022-03-17 16:09:10 +00:00
xfs_buf_unlock(bp);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
pinned++;
continue;
}
} else {
xfs_buf_lock(bp);
}
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
/*
* Someone else might have written the buffer synchronously or
* marked it stale in the meantime. In that case only the
* _XBF_DELWRI_Q flag got cleared, and we have to drop the
* reference and remove it from the list here.
*/
if (!(bp->b_flags & _XBF_DELWRI_Q)) {
list_del_init(&bp->b_list);
xfs_buf_relse(bp);
continue;
}
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
trace_xfs_buf_delwri_split(bp, _RET_IP_);
/*
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
* If we have a wait list, each buffer (and associated delwri
* queue reference) transfers to it and is submitted
* synchronously. Otherwise, drop the buffer from the delwri
* queue and submit async.
*/
xfs: reset buffer write failure state on successful completion The buffer write failure flag is intended to control the internal write retry that XFS has historically implemented to help mitigate the severity of transient I/O errors. The flag is set when a buffer is resubmitted from the I/O completion path due to a previous failure. It is checked on subsequent I/O completions to skip the internal retry and fall through to the higher level configurable error handling mechanism. The flag is cleared in the synchronous and delwri submission paths and also checked in various places to log write failure messages. There are a couple minor problems with the current usage of this flag. One is that we issue an internal retry after every submission from xfsaild due to how delwri submission clears the flag. This results in double the expected or configured number of write attempts when under sustained failures. Another more subtle issue is that the flag is never cleared on successful I/O completion. This can cause xfs_wait_buftarg() to suggest that dirty buffers are being thrown away due to the existence of the flag, when the reality is that the flag might still be set because the write succeeded on the retry. Clear the write failure flag on successful I/O completion to address both of these problems. This means that the internal retry attempt occurs once since the last time a buffer write failed and that various other contexts only see the flag set when the immediately previous write attempt has failed. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Allison Collins <allison.henderson@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2020-05-06 20:25:20 +00:00
bp->b_flags &= ~_XBF_DELWRI_Q;
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
bp->b_flags |= XBF_WRITE;
if (wait_list) {
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
bp->b_flags &= ~XBF_ASYNC;
list_move_tail(&bp->b_list, wait_list);
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
} else {
bp->b_flags |= XBF_ASYNC;
list_del_init(&bp->b_list);
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
}
__xfs_buf_submit(bp, false);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
}
blk_finish_plug(&plug);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
return pinned;
}
/*
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
* Write out a buffer list asynchronously.
*
* This will take the @buffer_list, write all non-locked and non-pinned buffers
* out and not wait for I/O completion on any of the buffers. This interface
* is only safely useable for callers that can track I/O completion by higher
* level means, e.g. AIL pushing as the @buffer_list is consumed in this
* function.
xfs: clear ail delwri queued bufs on unmount of shutdown fs In the typical unmount case, the AIL is forced out by the unmount sequence before the xfsaild task is stopped. Since AIL items are removed on writeback completion, this means that the AIL ->ail_buf_list delwri queue has been drained. This is not always true in the shutdown case, however. It's possible for buffers to sit on a delwri queue for a period of time across submission attempts if said items are locked or have been relogged and pinned since first added to the queue. If the attempt to log such an item results in a log I/O error, the error processing can shutdown the fs, remove the item from the AIL, stale the buffer (dropping the LRU reference) and clear its delwri queue state. The latter bit means the buffer will be released from a delwri queue on the next submission attempt, but this might never occur if the filesystem has shutdown and the AIL is empty. This means that such buffers are held indefinitely by the AIL delwri queue across destruction of the AIL. Aside from being a memory leak, these buffers can also hold references to in-core perag structures. The latter problem manifests as a generic/475 failure, reproducing the following asserts at unmount time: XFS: Assertion failed: atomic_read(&pag->pag_ref) == 0, file: fs/xfs/xfs_mount.c, line: 151 XFS: Assertion failed: atomic_read(&pag->pag_ref) == 0, file: fs/xfs/xfs_mount.c, line: 132 To prevent this problem, clear the AIL delwri queue as a final step before xfsaild() exit. The !empty state should never occur in the normal case, so add an assert to catch unexpected problems going forward. [dgc: add comment explaining need for xfs_buf_delwri_cancel() after calling xfs_buf_delwri_submit_nowait().] Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
2018-10-18 06:21:49 +00:00
*
* Note: this function will skip buffers it would block on, and in doing so
* leaves them on @buffer_list so they can be retried on a later pass. As such,
* it is up to the caller to ensure that the buffer list is fully submitted or
* cancelled appropriately when they are finished with the list. Failure to
* cancel or resubmit the list until it is empty will result in leaked buffers
* at unmount time.
*/
int
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
xfs_buf_delwri_submit_nowait(
struct list_head *buffer_list)
{
return xfs_buf_delwri_submit_buffers(buffer_list, NULL);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
}
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
/*
* Write out a buffer list synchronously.
*
* This will take the @buffer_list, write all buffers out and wait for I/O
* completion on all of the buffers. @buffer_list is consumed by the function,
* so callers must have some other way of tracking buffers if they require such
* functionality.
*/
int
xfs_buf_delwri_submit(
struct list_head *buffer_list)
{
LIST_HEAD (wait_list);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
int error = 0, error2;
struct xfs_buf *bp;
xfs_buf_delwri_submit_buffers(buffer_list, &wait_list);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
/* Wait for IO to complete. */
while (!list_empty(&wait_list)) {
bp = list_first_entry(&wait_list, struct xfs_buf, b_list);
list_del_init(&bp->b_list);
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
/*
* Wait on the locked buffer, check for errors and unlock and
* release the delwri queue reference.
*/
error2 = xfs_buf_iowait(bp);
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
xfs_buf_relse(bp);
if (!error)
error = error2;
}
xfs: on-stack delayed write buffer lists Queue delwri buffers on a local on-stack list instead of a per-buftarg one, and write back the buffers per-process instead of by waking up xfsbufd. This is now easily doable given that we have very few places left that write delwri buffers: - log recovery: Only done at mount time, and already forcing out the buffers synchronously using xfs_flush_buftarg - quotacheck: Same story. - dquot reclaim: Writes out dirty dquots on the LRU under memory pressure. We might want to look into doing more of this via xfsaild, but it's already more optimal than the synchronous inode reclaim that writes each buffer synchronously. - xfsaild: This is the main beneficiary of the change. By keeping a local list of buffers to write we reduce latency of writing out buffers, and more importably we can remove all the delwri list promotions which were hitting the buffer cache hard under sustained metadata loads. The implementation is very straight forward - xfs_buf_delwri_queue now gets a new list_head pointer that it adds the delwri buffers to, and all callers need to eventually submit the list using xfs_buf_delwi_submit or xfs_buf_delwi_submit_nowait. Buffers that already are on a delwri list are skipped in xfs_buf_delwri_queue, assuming they already are on another delwri list. The biggest change to pass down the buffer list was done to the AIL pushing. Now that we operate on buffers the trylock, push and pushbuf log item methods are merged into a single push routine, which tries to lock the item, and if possible add the buffer that needs writeback to the buffer list. This leads to much simpler code than the previous split but requires the individual IOP_PUSH instances to unlock and reacquire the AIL around calls to blocking routines. Given that xfsailds now also handle writing out buffers, the conditions for log forcing and the sleep times needed some small changes. The most important one is that we consider an AIL busy as long we still have buffers to push, and the other one is that we do increment the pushed LSN for buffers that are under flushing at this moment, but still count them towards the stuck items for restart purposes. Without this we could hammer on stuck items without ever forcing the log and not make progress under heavy random delete workloads on fast flash storage devices. [ Dave Chinner: - rebase on previous patches. - improved comments for XBF_DELWRI_Q handling - fix XBF_ASYNC handling in queue submission (test 106 failure) - rename delwri submit function buffer list parameters for clarity - xfs_efd_item_push() should return XFS_ITEM_PINNED ] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
2012-04-23 05:58:39 +00:00
return error;
}
xfs: push buffer of flush locked dquot to avoid quotacheck deadlock Reclaim during quotacheck can lead to deadlocks on the dquot flush lock: - Quotacheck populates a local delwri queue with the physical dquot buffers. - Quotacheck performs the xfs_qm_dqusage_adjust() bulkstat and dirties all of the dquots. - Reclaim kicks in and attempts to flush a dquot whose buffer is already queud on the quotacheck queue. The flush succeeds but queueing to the reclaim delwri queue fails as the backing buffer is already queued. The flush unlock is now deferred to I/O completion of the buffer from the quotacheck queue. - The dqadjust bulkstat continues and dirties the recently flushed dquot once again. - Quotacheck proceeds to the xfs_qm_flush_one() walk which requires the flush lock to update the backing buffers with the in-core recalculated values. It deadlocks on the redirtied dquot as the flush lock was already acquired by reclaim, but the buffer resides on the local delwri queue which isn't submitted until the end of quotacheck. This is reproduced by running quotacheck on a filesystem with a couple million inodes in low memory (512MB-1GB) situations. This is a regression as of commit 43ff2122e6 ("xfs: on-stack delayed write buffer lists"), which removed a trylock and buffer I/O submission from the quotacheck dquot flush sequence. Quotacheck first resets and collects the physical dquot buffers in a delwri queue. Then, it traverses the filesystem inodes via bulkstat, updates the in-core dquots, flushes the corrected dquots to the backing buffers and finally submits the delwri queue for I/O. Since the backing buffers are queued across the entire quotacheck operation, dquot reclaim cannot possibly complete a dquot flush before quotacheck completes. Therefore, quotacheck must submit the buffer for I/O in order to cycle the flush lock and flush the dirty in-core dquot to the buffer. Add a delwri queue buffer push mechanism to submit an individual buffer for I/O without losing the delwri queue status and use it from quotacheck to avoid the deadlock. This restores quotacheck behavior to as before the regression was introduced. Reported-by: Martin Svec <martin.svec@zoner.cz> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-06-15 04:21:45 +00:00
/*
* Push a single buffer on a delwri queue.
*
* The purpose of this function is to submit a single buffer of a delwri queue
* and return with the buffer still on the original queue. The waiting delwri
* buffer submission infrastructure guarantees transfer of the delwri queue
* buffer reference to a temporary wait list. We reuse this infrastructure to
* transfer the buffer back to the original queue.
*
* Note the buffer transitions from the queued state, to the submitted and wait
* listed state and back to the queued state during this call. The buffer
* locking and queue management logic between _delwri_pushbuf() and
* _delwri_queue() guarantee that the buffer cannot be queued to another list
* before returning.
*/
int
xfs_buf_delwri_pushbuf(
struct xfs_buf *bp,
struct list_head *buffer_list)
{
LIST_HEAD (submit_list);
int error;
ASSERT(bp->b_flags & _XBF_DELWRI_Q);
trace_xfs_buf_delwri_pushbuf(bp, _RET_IP_);
/*
* Isolate the buffer to a new local list so we can submit it for I/O
* independently from the rest of the original list.
*/
xfs_buf_lock(bp);
list_move(&bp->b_list, &submit_list);
xfs_buf_unlock(bp);
/*
* Delwri submission clears the DELWRI_Q buffer flag and returns with
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
* the buffer on the wait list with the original reference. Rather than
xfs: push buffer of flush locked dquot to avoid quotacheck deadlock Reclaim during quotacheck can lead to deadlocks on the dquot flush lock: - Quotacheck populates a local delwri queue with the physical dquot buffers. - Quotacheck performs the xfs_qm_dqusage_adjust() bulkstat and dirties all of the dquots. - Reclaim kicks in and attempts to flush a dquot whose buffer is already queud on the quotacheck queue. The flush succeeds but queueing to the reclaim delwri queue fails as the backing buffer is already queued. The flush unlock is now deferred to I/O completion of the buffer from the quotacheck queue. - The dqadjust bulkstat continues and dirties the recently flushed dquot once again. - Quotacheck proceeds to the xfs_qm_flush_one() walk which requires the flush lock to update the backing buffers with the in-core recalculated values. It deadlocks on the redirtied dquot as the flush lock was already acquired by reclaim, but the buffer resides on the local delwri queue which isn't submitted until the end of quotacheck. This is reproduced by running quotacheck on a filesystem with a couple million inodes in low memory (512MB-1GB) situations. This is a regression as of commit 43ff2122e6 ("xfs: on-stack delayed write buffer lists"), which removed a trylock and buffer I/O submission from the quotacheck dquot flush sequence. Quotacheck first resets and collects the physical dquot buffers in a delwri queue. Then, it traverses the filesystem inodes via bulkstat, updates the in-core dquots, flushes the corrected dquots to the backing buffers and finally submits the delwri queue for I/O. Since the backing buffers are queued across the entire quotacheck operation, dquot reclaim cannot possibly complete a dquot flush before quotacheck completes. Therefore, quotacheck must submit the buffer for I/O in order to cycle the flush lock and flush the dirty in-core dquot to the buffer. Add a delwri queue buffer push mechanism to submit an individual buffer for I/O without losing the delwri queue status and use it from quotacheck to avoid the deadlock. This restores quotacheck behavior to as before the regression was introduced. Reported-by: Martin Svec <martin.svec@zoner.cz> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-06-15 04:21:45 +00:00
* bounce the buffer from a local wait list back to the original list
* after I/O completion, reuse the original list as the wait list.
*/
xfs_buf_delwri_submit_buffers(&submit_list, buffer_list);
/*
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
* The buffer is now locked, under I/O and wait listed on the original
* delwri queue. Wait for I/O completion, restore the DELWRI_Q flag and
* return with the buffer unlocked and on the original queue.
xfs: push buffer of flush locked dquot to avoid quotacheck deadlock Reclaim during quotacheck can lead to deadlocks on the dquot flush lock: - Quotacheck populates a local delwri queue with the physical dquot buffers. - Quotacheck performs the xfs_qm_dqusage_adjust() bulkstat and dirties all of the dquots. - Reclaim kicks in and attempts to flush a dquot whose buffer is already queud on the quotacheck queue. The flush succeeds but queueing to the reclaim delwri queue fails as the backing buffer is already queued. The flush unlock is now deferred to I/O completion of the buffer from the quotacheck queue. - The dqadjust bulkstat continues and dirties the recently flushed dquot once again. - Quotacheck proceeds to the xfs_qm_flush_one() walk which requires the flush lock to update the backing buffers with the in-core recalculated values. It deadlocks on the redirtied dquot as the flush lock was already acquired by reclaim, but the buffer resides on the local delwri queue which isn't submitted until the end of quotacheck. This is reproduced by running quotacheck on a filesystem with a couple million inodes in low memory (512MB-1GB) situations. This is a regression as of commit 43ff2122e6 ("xfs: on-stack delayed write buffer lists"), which removed a trylock and buffer I/O submission from the quotacheck dquot flush sequence. Quotacheck first resets and collects the physical dquot buffers in a delwri queue. Then, it traverses the filesystem inodes via bulkstat, updates the in-core dquots, flushes the corrected dquots to the backing buffers and finally submits the delwri queue for I/O. Since the backing buffers are queued across the entire quotacheck operation, dquot reclaim cannot possibly complete a dquot flush before quotacheck completes. Therefore, quotacheck must submit the buffer for I/O in order to cycle the flush lock and flush the dirty in-core dquot to the buffer. Add a delwri queue buffer push mechanism to submit an individual buffer for I/O without losing the delwri queue status and use it from quotacheck to avoid the deadlock. This restores quotacheck behavior to as before the regression was introduced. Reported-by: Martin Svec <martin.svec@zoner.cz> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-06-15 04:21:45 +00:00
*/
xfs: use sync buffer I/O for sync delwri queue submission If a delwri queue occurs of a buffer that sits on a delwri queue wait list, the queue sets _XBF_DELWRI_Q without changing the state of ->b_list. This occurs, for example, if another thread beats the current delwri waiter thread to the buffer lock after I/O completion. Once the waiter acquires the lock, it removes the buffer from the wait list and leaves a buffer with _XBF_DELWRI_Q set but not populated on a list. This results in a lost buffer submission and in turn can result in assert failures due to _XBF_DELWRI_Q being set on buffer reclaim or filesystem lockups if the buffer happens to cover an item in the AIL. This problem has been reproduced by repeated iterations of xfs/305 on high CPU count (28xcpu) systems with limited memory (~1GB). Dirty dquot reclaim races with an xfsaild push of a separate dquot backed by the same buffer such that the buffer sits on the reclaim wait list at the time xfsaild attempts to queue it. Since the latter dquot has been flush locked but the underlying buffer not submitted for I/O, the dquot pins the AIL and causes the filesystem to livelock. This race is essentially made possible by the buffer lock cycle involved with waiting on a synchronous delwri queue submission. Close the race by using synchronous buffer I/O for respective delwri queue submission. This means the buffer remains locked across the I/O and so is inaccessible from other contexts while in the intermediate wait list state. The sync buffer I/O wait mechanism is factored into a helper such that sync delwri buffer submission and serialization are batched operations. Designed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-07-12 05:26:34 +00:00
error = xfs_buf_iowait(bp);
xfs: push buffer of flush locked dquot to avoid quotacheck deadlock Reclaim during quotacheck can lead to deadlocks on the dquot flush lock: - Quotacheck populates a local delwri queue with the physical dquot buffers. - Quotacheck performs the xfs_qm_dqusage_adjust() bulkstat and dirties all of the dquots. - Reclaim kicks in and attempts to flush a dquot whose buffer is already queud on the quotacheck queue. The flush succeeds but queueing to the reclaim delwri queue fails as the backing buffer is already queued. The flush unlock is now deferred to I/O completion of the buffer from the quotacheck queue. - The dqadjust bulkstat continues and dirties the recently flushed dquot once again. - Quotacheck proceeds to the xfs_qm_flush_one() walk which requires the flush lock to update the backing buffers with the in-core recalculated values. It deadlocks on the redirtied dquot as the flush lock was already acquired by reclaim, but the buffer resides on the local delwri queue which isn't submitted until the end of quotacheck. This is reproduced by running quotacheck on a filesystem with a couple million inodes in low memory (512MB-1GB) situations. This is a regression as of commit 43ff2122e6 ("xfs: on-stack delayed write buffer lists"), which removed a trylock and buffer I/O submission from the quotacheck dquot flush sequence. Quotacheck first resets and collects the physical dquot buffers in a delwri queue. Then, it traverses the filesystem inodes via bulkstat, updates the in-core dquots, flushes the corrected dquots to the backing buffers and finally submits the delwri queue for I/O. Since the backing buffers are queued across the entire quotacheck operation, dquot reclaim cannot possibly complete a dquot flush before quotacheck completes. Therefore, quotacheck must submit the buffer for I/O in order to cycle the flush lock and flush the dirty in-core dquot to the buffer. Add a delwri queue buffer push mechanism to submit an individual buffer for I/O without losing the delwri queue status and use it from quotacheck to avoid the deadlock. This restores quotacheck behavior to as before the regression was introduced. Reported-by: Martin Svec <martin.svec@zoner.cz> Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-06-15 04:21:45 +00:00
bp->b_flags |= _XBF_DELWRI_Q;
xfs_buf_unlock(bp);
return error;
}
void xfs_buf_set_ref(struct xfs_buf *bp, int lru_ref)
{
/*
* Set the lru reference count to 0 based on the error injection tag.
* This allows userspace to disrupt buffer caching for debug/testing
* purposes.
*/
if (XFS_TEST_ERROR(false, bp->b_mount, XFS_ERRTAG_BUF_LRU_REF))
lru_ref = 0;
atomic_set(&bp->b_lru_ref, lru_ref);
}
/*
* Verify an on-disk magic value against the magic value specified in the
* verifier structure. The verifier magic is in disk byte order so the caller is
* expected to pass the value directly from disk.
*/
bool
xfs_verify_magic(
struct xfs_buf *bp,
__be32 dmagic)
{
struct xfs_mount *mp = bp->b_mount;
int idx;
idx = xfs_has_crc(mp);
if (WARN_ON(!bp->b_ops || !bp->b_ops->magic[idx]))
return false;
return dmagic == bp->b_ops->magic[idx];
}
/*
* Verify an on-disk magic value against the magic value specified in the
* verifier structure. The verifier magic is in disk byte order so the caller is
* expected to pass the value directly from disk.
*/
bool
xfs_verify_magic16(
struct xfs_buf *bp,
__be16 dmagic)
{
struct xfs_mount *mp = bp->b_mount;
int idx;
idx = xfs_has_crc(mp);
if (WARN_ON(!bp->b_ops || !bp->b_ops->magic16[idx]))
return false;
return dmagic == bp->b_ops->magic16[idx];
}