2018-06-06 05:42:14 +03:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
2005-04-17 02:20:36 +04:00
|
|
|
/*
|
2005-11-02 06:58:39 +03:00
|
|
|
* Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
|
|
|
|
* All Rights Reserved.
|
2005-04-17 02:20:36 +04:00
|
|
|
*/
|
|
|
|
#include "xfs.h"
|
2005-11-02 06:38:42 +03:00
|
|
|
#include "xfs_fs.h"
|
2019-06-29 05:25:35 +03:00
|
|
|
#include "xfs_shared.h"
|
2013-10-23 03:51:50 +04:00
|
|
|
#include "xfs_format.h"
|
2013-10-23 03:50:10 +04:00
|
|
|
#include "xfs_log_format.h"
|
|
|
|
#include "xfs_trans_resv.h"
|
2005-04-17 02:20:36 +04:00
|
|
|
#include "xfs_mount.h"
|
|
|
|
#include "xfs_inode.h"
|
2013-10-23 03:50:10 +04:00
|
|
|
#include "xfs_trans.h"
|
2005-11-02 06:38:42 +03:00
|
|
|
#include "xfs_inode_item.h"
|
2009-12-15 02:14:59 +03:00
|
|
|
#include "xfs_trace.h"
|
2013-10-23 03:50:10 +04:00
|
|
|
#include "xfs_trans_priv.h"
|
xfs: Properly retry failed inode items in case of error during buffer writeback
When a buffer has been failed during writeback, the inode items into it
are kept flush locked, and are never resubmitted due the flush lock, so,
if any buffer fails to be written, the items in AIL are never written to
disk and never unlocked.
This causes unmount operation to hang due these items flush locked in AIL,
but this also causes the items in AIL to never be written back, even when
the IO device comes back to normal.
I've been testing this patch with a DM-thin device, creating a
filesystem larger than the real device.
When writing enough data to fill the DM-thin device, XFS receives ENOSPC
errors from the device, and keep spinning on xfsaild (when 'retry
forever' configuration is set).
At this point, the filesystem can not be unmounted because of the flush locked
items in AIL, but worse, the items in AIL are never retried at all
(once xfs_inode_item_push() will skip the items that are flush locked),
even if the underlying DM-thin device is expanded to the proper size.
This patch fixes both cases, retrying any item that has been failed
previously, using the infra-structure provided by the previous patch.
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-08-09 04:21:50 +03:00
|
|
|
#include "xfs_buf_item.h"
|
2013-12-13 04:00:43 +04:00
|
|
|
#include "xfs_log.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 19:09:13 +03:00
|
|
|
#include "xfs_log_priv.h"
|
2019-11-02 19:40:53 +03:00
|
|
|
#include "xfs_error.h"
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2017-12-11 14:35:19 +03:00
|
|
|
#include <linux/iversion.h>
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2021-10-12 21:09:23 +03:00
|
|
|
struct kmem_cache *xfs_ili_cache; /* inode log item */
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2010-06-23 12:11:15 +04:00
|
|
|
static inline struct xfs_inode_log_item *INODE_ITEM(struct xfs_log_item *lip)
|
|
|
|
{
|
|
|
|
return container_of(lip, struct xfs_inode_log_item, ili_item);
|
|
|
|
}
|
|
|
|
|
xfs: Fix CIL throttle hang when CIL space used going backwards
A hang with tasks stuck on the CIL hard throttle was reported and
largely diagnosed by Donald Buczek, who discovered that it was a
result of the CIL context space usage decrementing in committed
transactions once the hard throttle limit had been hit and processes
were already blocked. This resulted in the CIL push not waking up
those waiters because the CIL context was no longer over the hard
throttle limit.
The surprising aspect of this was the CIL space usage going
backwards regularly enough to trigger this situation. Assumptions
had been made in design that the relogging process would only
increase the size of the objects in the CIL, and so that space would
only increase.
This change and commit message fixes the issue and documents the
result of an audit of the triggers that can cause the CIL space to
go backwards, how large the backwards steps tend to be, the
frequency in which they occur, and what the impact on the CIL
accounting code is.
Even though the CIL ctx->space_used can go backwards, it will only
do so if the log item is already logged to the CIL and contains a
space reservation for it's entire logged state. This is tracked by
the shadow buffer state on the log item. If the item is not
previously logged in the CIL it has no shadow buffer nor log vector,
and hence the entire size of the logged item copied to the log
vector is accounted to the CIL space usage. i.e. it will always go
up in this case.
If the item has a log vector (i.e. already in the CIL) and the size
decreases, then the existing log vector will be overwritten and the
space usage will go down. This is the only condition where the space
usage reduces, and it can only occur when an item is already tracked
in the CIL. Hence we are safe from CIL space usage underruns as a
result of log items decreasing in size when they are relogged.
Typically this reduction in CIL usage occurs from metadata blocks
being free, such as when a btree block merge occurs or a directory
enter/xattr entry is removed and the da-tree is reduced in size.
This generally results in a reduction in size of around a single
block in the CIL, but also tends to increase the number of log
vectors because the parent and sibling nodes in the tree needs to be
updated when a btree block is removed. If a multi-level merge
occurs, then we see reduction in size of 2+ blocks, but again the
log vector count goes up.
The other vector is inode fork size changes, which only log the
current size of the fork and ignore the previously logged size when
the fork is relogged. Hence if we are removing items from the inode
fork (dir/xattr removal in shortform, extent record removal in
extent form, etc) the relogged size of the inode for can decrease.
No other log items can decrease in size either because they are a
fixed size (e.g. dquots) or they cannot be relogged (e.g. relogging
an intent actually creates a new intent log item and doesn't relog
the old item at all.) Hence the only two vectors for CIL context
size reduction are relogging inode forks and marking buffers active
in the CIL as stale.
Long story short: the majority of the code does the right thing and
handles the reduction in log item size correctly, and only the CIL
hard throttle implementation is problematic and needs fixing. This
patch makes that fix, as well as adds comments in the log item code
that result in items shrinking in size when they are relogged as a
clear reminder that this can and does happen frequently.
The throttle fix is based upon the change Donald proposed, though it
goes further to ensure that once the throttle is activated, it
captures all tasks until the CIL push issues a wakeup, regardless of
whether the CIL space used has gone back under the throttle
threshold.
This ensures that we prevent tasks reducing the CIL slightly under
the throttle threshold and then making more changes that push it
well over the throttle limit. This is acheived by checking if the
throttle wait queue is already active as a condition of throttling.
Hence once we start throttling, we continue to apply the throttle
until the CIL context push wakes everything on the wait queue.
We can use waitqueue_active() for the waitqueue manipulations and
checks as they are all done under the ctx->xc_push_lock. Hence the
waitqueue has external serialisation and we can safely peek inside
the wait queue without holding the internal waitqueue locks.
Many thanks to Donald for his diagnostic and analysis work to
isolate the cause of this hang.
Reported-and-tested-by: Donald Buczek <buczek@molgen.mpg.de>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2021-06-18 18:21:51 +03:00
|
|
|
/*
|
|
|
|
* The logged size of an inode fork is always the current size of the inode
|
|
|
|
* fork. This means that when an inode fork is relogged, the size of the logged
|
|
|
|
* region is determined by the current state, not the combination of the
|
|
|
|
* previously logged state + the current state. This is different relogging
|
|
|
|
* behaviour to most other log items which will retain the size of the
|
|
|
|
* previously logged changes when smaller regions are relogged.
|
|
|
|
*
|
|
|
|
* Hence operations that remove data from the inode fork (e.g. shortform
|
|
|
|
* dir/attr remove, extent form extent removal, etc), the size of the relogged
|
|
|
|
* inode gets -smaller- rather than stays the same size as the previously logged
|
|
|
|
* size and this can result in the committing transaction reducing the amount of
|
|
|
|
* space being consumed by the CIL.
|
|
|
|
*/
|
2013-08-12 14:50:04 +04:00
|
|
|
STATIC void
|
2013-12-13 04:00:43 +04:00
|
|
|
xfs_inode_item_data_fork_size(
|
|
|
|
struct xfs_inode_log_item *iip,
|
2013-08-12 14:50:04 +04:00
|
|
|
int *nvecs,
|
|
|
|
int *nbytes)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
2013-08-12 14:50:04 +04:00
|
|
|
|
2020-05-18 20:28:05 +03:00
|
|
|
switch (ip->i_df.if_format) {
|
2005-04-17 02:20:36 +04:00
|
|
|
case XFS_DINODE_FMT_EXTENTS:
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_DEXT) &&
|
2020-05-18 20:27:22 +03:00
|
|
|
ip->i_df.if_nextents > 0 &&
|
2013-08-12 14:50:04 +04:00
|
|
|
ip->i_df.if_bytes > 0) {
|
|
|
|
/* worst case, doesn't subtract delalloc extents */
|
|
|
|
*nbytes += XFS_IFORK_DSIZE(ip);
|
|
|
|
*nvecs += 1;
|
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_BTREE:
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_DBROOT) &&
|
2013-08-12 14:50:04 +04:00
|
|
|
ip->i_df.if_broot_bytes > 0) {
|
|
|
|
*nbytes += ip->i_df.if_broot_bytes;
|
|
|
|
*nvecs += 1;
|
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_LOCAL:
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_DDATA) &&
|
2013-08-12 14:50:04 +04:00
|
|
|
ip->i_df.if_bytes > 0) {
|
|
|
|
*nbytes += roundup(ip->i_df.if_bytes, 4);
|
|
|
|
*nvecs += 1;
|
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
|
|
|
|
case XFS_DINODE_FMT_DEV:
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
ASSERT(0);
|
|
|
|
break;
|
|
|
|
}
|
2013-12-13 04:00:43 +04:00
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2013-12-13 04:00:43 +04:00
|
|
|
STATIC void
|
|
|
|
xfs_inode_item_attr_fork_size(
|
|
|
|
struct xfs_inode_log_item *iip,
|
|
|
|
int *nvecs,
|
|
|
|
int *nbytes)
|
|
|
|
{
|
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2020-05-18 20:28:05 +03:00
|
|
|
switch (ip->i_afp->if_format) {
|
2005-04-17 02:20:36 +04:00
|
|
|
case XFS_DINODE_FMT_EXTENTS:
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_AEXT) &&
|
2020-05-18 20:27:22 +03:00
|
|
|
ip->i_afp->if_nextents > 0 &&
|
2013-08-12 14:50:04 +04:00
|
|
|
ip->i_afp->if_bytes > 0) {
|
|
|
|
/* worst case, doesn't subtract unused space */
|
|
|
|
*nbytes += XFS_IFORK_ASIZE(ip);
|
|
|
|
*nvecs += 1;
|
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_BTREE:
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_ABROOT) &&
|
2013-08-12 14:50:04 +04:00
|
|
|
ip->i_afp->if_broot_bytes > 0) {
|
|
|
|
*nbytes += ip->i_afp->if_broot_bytes;
|
|
|
|
*nvecs += 1;
|
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_LOCAL:
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_ADATA) &&
|
2013-08-12 14:50:04 +04:00
|
|
|
ip->i_afp->if_bytes > 0) {
|
|
|
|
*nbytes += roundup(ip->i_afp->if_bytes, 4);
|
|
|
|
*nvecs += 1;
|
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
ASSERT(0);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-12-13 04:00:43 +04:00
|
|
|
/*
|
|
|
|
* This returns the number of iovecs needed to log the given inode item.
|
|
|
|
*
|
|
|
|
* We need one iovec for the inode log format structure, one for the
|
|
|
|
* inode core, and possibly one for the inode data/extents/b-tree root
|
|
|
|
* and one for the inode attribute data/extents/b-tree root.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_inode_item_size(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
int *nvecs,
|
|
|
|
int *nbytes)
|
|
|
|
{
|
|
|
|
struct xfs_inode_log_item *iip = INODE_ITEM(lip);
|
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
|
|
|
|
|
|
|
*nvecs += 2;
|
|
|
|
*nbytes += sizeof(struct xfs_inode_log_format) +
|
2020-03-18 18:15:10 +03:00
|
|
|
xfs_log_dinode_size(ip->i_mount);
|
2013-12-13 04:00:43 +04:00
|
|
|
|
|
|
|
xfs_inode_item_data_fork_size(iip, nvecs, nbytes);
|
|
|
|
if (XFS_IFORK_Q(ip))
|
|
|
|
xfs_inode_item_attr_fork_size(iip, nvecs, nbytes);
|
|
|
|
}
|
|
|
|
|
2013-12-13 04:00:43 +04:00
|
|
|
STATIC void
|
2013-12-13 04:00:43 +04:00
|
|
|
xfs_inode_item_format_data_fork(
|
|
|
|
struct xfs_inode_log_item *iip,
|
2013-12-13 04:34:02 +04:00
|
|
|
struct xfs_inode_log_format *ilf,
|
|
|
|
struct xfs_log_vec *lv,
|
|
|
|
struct xfs_log_iovec **vecp)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
2005-04-17 02:20:36 +04:00
|
|
|
size_t data_bytes;
|
|
|
|
|
2020-05-18 20:28:05 +03:00
|
|
|
switch (ip->i_df.if_format) {
|
2005-04-17 02:20:36 +04:00
|
|
|
case XFS_DINODE_FMT_EXTENTS:
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &=
|
2017-10-19 21:07:09 +03:00
|
|
|
~(XFS_ILOG_DDATA | XFS_ILOG_DBROOT | XFS_ILOG_DEV);
|
2012-02-29 13:53:53 +04:00
|
|
|
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_DEXT) &&
|
2020-05-18 20:27:22 +03:00
|
|
|
ip->i_df.if_nextents > 0 &&
|
2012-02-29 13:53:53 +04:00
|
|
|
ip->i_df.if_bytes > 0) {
|
2013-12-13 04:34:04 +04:00
|
|
|
struct xfs_bmbt_rec *p;
|
|
|
|
|
2016-11-08 04:59:42 +03:00
|
|
|
ASSERT(xfs_iext_count(&ip->i_df) > 0);
|
2013-12-13 04:34:04 +04:00
|
|
|
|
|
|
|
p = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_IEXT);
|
|
|
|
data_bytes = xfs_iextents_copy(ip, p, XFS_DATA_FORK);
|
|
|
|
xlog_finish_iovec(lv, *vecp, data_bytes);
|
|
|
|
|
|
|
|
ASSERT(data_bytes <= ip->i_df.if_bytes);
|
|
|
|
|
|
|
|
ilf->ilf_dsize = data_bytes;
|
2013-12-13 04:34:02 +04:00
|
|
|
ilf->ilf_size++;
|
2012-02-29 13:53:53 +04:00
|
|
|
} else {
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &= ~XFS_ILOG_DEXT;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_BTREE:
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &=
|
2017-10-19 21:07:09 +03:00
|
|
|
~(XFS_ILOG_DDATA | XFS_ILOG_DEXT | XFS_ILOG_DEV);
|
2012-02-29 13:53:53 +04:00
|
|
|
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_DBROOT) &&
|
2012-02-29 13:53:53 +04:00
|
|
|
ip->i_df.if_broot_bytes > 0) {
|
2005-04-17 02:20:36 +04:00
|
|
|
ASSERT(ip->i_df.if_broot != NULL);
|
2013-12-13 04:34:02 +04:00
|
|
|
xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_IBROOT,
|
2013-12-13 04:00:43 +04:00
|
|
|
ip->i_df.if_broot,
|
|
|
|
ip->i_df.if_broot_bytes);
|
2013-12-13 04:34:02 +04:00
|
|
|
ilf->ilf_dsize = ip->i_df.if_broot_bytes;
|
|
|
|
ilf->ilf_size++;
|
2012-02-29 13:53:53 +04:00
|
|
|
} else {
|
2012-02-29 13:53:54 +04:00
|
|
|
ASSERT(!(iip->ili_fields &
|
2012-02-29 13:53:53 +04:00
|
|
|
XFS_ILOG_DBROOT));
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &= ~XFS_ILOG_DBROOT;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_LOCAL:
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &=
|
2017-10-19 21:07:09 +03:00
|
|
|
~(XFS_ILOG_DEXT | XFS_ILOG_DBROOT | XFS_ILOG_DEV);
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_DDATA) &&
|
2012-02-29 13:53:53 +04:00
|
|
|
ip->i_df.if_bytes > 0) {
|
2005-04-17 02:20:36 +04:00
|
|
|
/*
|
|
|
|
* Round i_bytes up to a word boundary.
|
2020-08-05 18:49:58 +03:00
|
|
|
* The underlying memory is guaranteed
|
2005-04-17 02:20:36 +04:00
|
|
|
* to be there by xfs_idata_realloc().
|
|
|
|
*/
|
|
|
|
data_bytes = roundup(ip->i_df.if_bytes, 4);
|
2013-12-13 04:00:43 +04:00
|
|
|
ASSERT(ip->i_df.if_u1.if_data != NULL);
|
2021-03-29 21:11:40 +03:00
|
|
|
ASSERT(ip->i_disk_size > 0);
|
2013-12-13 04:34:02 +04:00
|
|
|
xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_ILOCAL,
|
2013-12-13 04:00:43 +04:00
|
|
|
ip->i_df.if_u1.if_data, data_bytes);
|
2013-12-13 04:34:02 +04:00
|
|
|
ilf->ilf_dsize = (unsigned)data_bytes;
|
|
|
|
ilf->ilf_size++;
|
2012-02-29 13:53:53 +04:00
|
|
|
} else {
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &= ~XFS_ILOG_DDATA;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_DEV:
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &=
|
2017-10-19 21:07:09 +03:00
|
|
|
~(XFS_ILOG_DDATA | XFS_ILOG_DBROOT | XFS_ILOG_DEXT);
|
2013-12-13 04:34:02 +04:00
|
|
|
if (iip->ili_fields & XFS_ILOG_DEV)
|
2017-10-19 21:07:09 +03:00
|
|
|
ilf->ilf_u.ilfu_rdev = sysv_encode_dev(VFS_I(ip)->i_rdev);
|
2005-04-17 02:20:36 +04:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
ASSERT(0);
|
|
|
|
break;
|
|
|
|
}
|
2013-12-13 04:00:43 +04:00
|
|
|
}
|
|
|
|
|
2013-12-13 04:00:43 +04:00
|
|
|
STATIC void
|
2013-12-13 04:00:43 +04:00
|
|
|
xfs_inode_item_format_attr_fork(
|
|
|
|
struct xfs_inode_log_item *iip,
|
2013-12-13 04:34:02 +04:00
|
|
|
struct xfs_inode_log_format *ilf,
|
|
|
|
struct xfs_log_vec *lv,
|
|
|
|
struct xfs_log_iovec **vecp)
|
2013-12-13 04:00:43 +04:00
|
|
|
{
|
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
|
|
|
size_t data_bytes;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2020-05-18 20:28:05 +03:00
|
|
|
switch (ip->i_afp->if_format) {
|
2005-04-17 02:20:36 +04:00
|
|
|
case XFS_DINODE_FMT_EXTENTS:
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &=
|
2012-02-29 13:53:53 +04:00
|
|
|
~(XFS_ILOG_ADATA | XFS_ILOG_ABROOT);
|
|
|
|
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_AEXT) &&
|
2020-05-18 20:27:22 +03:00
|
|
|
ip->i_afp->if_nextents > 0 &&
|
2012-02-29 13:53:53 +04:00
|
|
|
ip->i_afp->if_bytes > 0) {
|
2013-12-13 04:34:04 +04:00
|
|
|
struct xfs_bmbt_rec *p;
|
|
|
|
|
2016-11-08 04:59:42 +03:00
|
|
|
ASSERT(xfs_iext_count(ip->i_afp) ==
|
2020-05-18 20:27:22 +03:00
|
|
|
ip->i_afp->if_nextents);
|
2013-12-13 04:34:04 +04:00
|
|
|
|
|
|
|
p = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_IATTR_EXT);
|
|
|
|
data_bytes = xfs_iextents_copy(ip, p, XFS_ATTR_FORK);
|
|
|
|
xlog_finish_iovec(lv, *vecp, data_bytes);
|
|
|
|
|
|
|
|
ilf->ilf_asize = data_bytes;
|
2013-12-13 04:34:02 +04:00
|
|
|
ilf->ilf_size++;
|
2012-02-29 13:53:53 +04:00
|
|
|
} else {
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &= ~XFS_ILOG_AEXT;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_BTREE:
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &=
|
2012-02-29 13:53:53 +04:00
|
|
|
~(XFS_ILOG_ADATA | XFS_ILOG_AEXT);
|
|
|
|
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_ABROOT) &&
|
2012-02-29 13:53:53 +04:00
|
|
|
ip->i_afp->if_broot_bytes > 0) {
|
2005-04-17 02:20:36 +04:00
|
|
|
ASSERT(ip->i_afp->if_broot != NULL);
|
2012-02-29 13:53:53 +04:00
|
|
|
|
2013-12-13 04:34:02 +04:00
|
|
|
xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_IATTR_BROOT,
|
2013-12-13 04:00:43 +04:00
|
|
|
ip->i_afp->if_broot,
|
|
|
|
ip->i_afp->if_broot_bytes);
|
2013-12-13 04:34:02 +04:00
|
|
|
ilf->ilf_asize = ip->i_afp->if_broot_bytes;
|
|
|
|
ilf->ilf_size++;
|
2012-02-29 13:53:53 +04:00
|
|
|
} else {
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &= ~XFS_ILOG_ABROOT;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
case XFS_DINODE_FMT_LOCAL:
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &=
|
2012-02-29 13:53:53 +04:00
|
|
|
~(XFS_ILOG_AEXT | XFS_ILOG_ABROOT);
|
|
|
|
|
2012-02-29 13:53:54 +04:00
|
|
|
if ((iip->ili_fields & XFS_ILOG_ADATA) &&
|
2012-02-29 13:53:53 +04:00
|
|
|
ip->i_afp->if_bytes > 0) {
|
2005-04-17 02:20:36 +04:00
|
|
|
/*
|
|
|
|
* Round i_bytes up to a word boundary.
|
2020-08-05 18:49:58 +03:00
|
|
|
* The underlying memory is guaranteed
|
2005-04-17 02:20:36 +04:00
|
|
|
* to be there by xfs_idata_realloc().
|
|
|
|
*/
|
|
|
|
data_bytes = roundup(ip->i_afp->if_bytes, 4);
|
2013-12-13 04:00:43 +04:00
|
|
|
ASSERT(ip->i_afp->if_u1.if_data != NULL);
|
2013-12-13 04:34:02 +04:00
|
|
|
xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_IATTR_LOCAL,
|
2013-12-13 04:00:43 +04:00
|
|
|
ip->i_afp->if_u1.if_data,
|
|
|
|
data_bytes);
|
2013-12-13 04:34:02 +04:00
|
|
|
ilf->ilf_asize = (unsigned)data_bytes;
|
|
|
|
ilf->ilf_size++;
|
2012-02-29 13:53:53 +04:00
|
|
|
} else {
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields &= ~XFS_ILOG_ADATA;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
ASSERT(0);
|
|
|
|
break;
|
|
|
|
}
|
2013-12-13 04:00:43 +04:00
|
|
|
}
|
|
|
|
|
2020-08-25 02:01:34 +03:00
|
|
|
/*
|
|
|
|
* Convert an incore timestamp to a log timestamp. Note that the log format
|
|
|
|
* specifies host endian format!
|
|
|
|
*/
|
2021-04-21 23:48:27 +03:00
|
|
|
static inline xfs_log_timestamp_t
|
2020-08-25 02:01:34 +03:00
|
|
|
xfs_inode_to_log_dinode_ts(
|
2020-08-17 19:59:07 +03:00
|
|
|
struct xfs_inode *ip,
|
2020-08-25 02:01:34 +03:00
|
|
|
const struct timespec64 tv)
|
|
|
|
{
|
2021-04-21 23:48:27 +03:00
|
|
|
struct xfs_log_legacy_timestamp *lits;
|
2021-04-21 23:48:27 +03:00
|
|
|
xfs_log_timestamp_t its;
|
2020-08-25 02:01:34 +03:00
|
|
|
|
2020-08-17 19:59:07 +03:00
|
|
|
if (xfs_inode_has_bigtime(ip))
|
|
|
|
return xfs_inode_encode_bigtime(tv);
|
|
|
|
|
2021-04-21 23:48:27 +03:00
|
|
|
lits = (struct xfs_log_legacy_timestamp *)&its;
|
2020-08-25 02:01:34 +03:00
|
|
|
lits->t_sec = tv.tv_sec;
|
|
|
|
lits->t_nsec = tv.tv_nsec;
|
|
|
|
|
|
|
|
return its;
|
|
|
|
}
|
|
|
|
|
2021-03-29 21:11:38 +03:00
|
|
|
/*
|
|
|
|
* The legacy DMAPI fields are only present in the on-disk and in-log inodes,
|
|
|
|
* but not in the in-memory one. But we are guaranteed to have an inode buffer
|
|
|
|
* in memory when logging an inode, so we can just copy it from the on-disk
|
|
|
|
* inode to the in-log inode here so that recovery of file system with these
|
|
|
|
* fields set to non-zero values doesn't lose them. For all other cases we zero
|
|
|
|
* the fields.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
xfs_copy_dm_fields_to_log_dinode(
|
|
|
|
struct xfs_inode *ip,
|
|
|
|
struct xfs_log_dinode *to)
|
|
|
|
{
|
|
|
|
struct xfs_dinode *dip;
|
|
|
|
|
|
|
|
dip = xfs_buf_offset(ip->i_itemp->ili_item.li_buf,
|
|
|
|
ip->i_imap.im_boffset);
|
|
|
|
|
|
|
|
if (xfs_iflags_test(ip, XFS_IPRESERVE_DM_FIELDS)) {
|
|
|
|
to->di_dmevmask = be32_to_cpu(dip->di_dmevmask);
|
|
|
|
to->di_dmstate = be16_to_cpu(dip->di_dmstate);
|
|
|
|
} else {
|
|
|
|
to->di_dmevmask = 0;
|
|
|
|
to->di_dmstate = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-02-09 08:54:58 +03:00
|
|
|
static void
|
2016-02-09 08:54:58 +03:00
|
|
|
xfs_inode_to_log_dinode(
|
|
|
|
struct xfs_inode *ip,
|
2016-02-09 08:54:58 +03:00
|
|
|
struct xfs_log_dinode *to,
|
|
|
|
xfs_lsn_t lsn)
|
2016-02-09 08:54:58 +03:00
|
|
|
{
|
2016-02-09 08:54:58 +03:00
|
|
|
struct inode *inode = VFS_I(ip);
|
|
|
|
|
2016-02-09 08:54:58 +03:00
|
|
|
to->di_magic = XFS_DINODE_MAGIC;
|
2020-05-18 20:28:05 +03:00
|
|
|
to->di_format = xfs_ifork_format(&ip->i_df);
|
2020-02-21 19:31:27 +03:00
|
|
|
to->di_uid = i_uid_read(inode);
|
|
|
|
to->di_gid = i_gid_read(inode);
|
2021-03-29 21:11:39 +03:00
|
|
|
to->di_projid_lo = ip->i_projid & 0xffff;
|
|
|
|
to->di_projid_hi = ip->i_projid >> 16;
|
2016-02-09 08:54:58 +03:00
|
|
|
|
2016-02-09 08:54:58 +03:00
|
|
|
memset(to->di_pad, 0, sizeof(to->di_pad));
|
2016-02-09 08:54:58 +03:00
|
|
|
memset(to->di_pad3, 0, sizeof(to->di_pad3));
|
2020-08-17 19:59:07 +03:00
|
|
|
to->di_atime = xfs_inode_to_log_dinode_ts(ip, inode->i_atime);
|
|
|
|
to->di_mtime = xfs_inode_to_log_dinode_ts(ip, inode->i_mtime);
|
|
|
|
to->di_ctime = xfs_inode_to_log_dinode_ts(ip, inode->i_ctime);
|
2016-02-09 08:54:58 +03:00
|
|
|
to->di_nlink = inode->i_nlink;
|
2016-02-09 08:54:58 +03:00
|
|
|
to->di_gen = inode->i_generation;
|
2016-02-09 08:54:58 +03:00
|
|
|
to->di_mode = inode->i_mode;
|
2016-02-09 08:54:58 +03:00
|
|
|
|
2021-03-29 21:11:40 +03:00
|
|
|
to->di_size = ip->i_disk_size;
|
2021-03-29 21:11:40 +03:00
|
|
|
to->di_nblocks = ip->i_nblocks;
|
2021-03-29 21:11:41 +03:00
|
|
|
to->di_extsize = ip->i_extsize;
|
2020-05-18 20:27:22 +03:00
|
|
|
to->di_nextents = xfs_ifork_nextents(&ip->i_df);
|
|
|
|
to->di_anextents = xfs_ifork_nextents(ip->i_afp);
|
2021-03-29 21:11:44 +03:00
|
|
|
to->di_forkoff = ip->i_forkoff;
|
2020-05-18 20:28:05 +03:00
|
|
|
to->di_aformat = xfs_ifork_format(ip->i_afp);
|
2021-03-29 21:11:44 +03:00
|
|
|
to->di_flags = ip->i_diflags;
|
2016-02-09 08:54:58 +03:00
|
|
|
|
2021-03-29 21:11:38 +03:00
|
|
|
xfs_copy_dm_fields_to_log_dinode(ip, to);
|
|
|
|
|
2017-10-09 21:37:22 +03:00
|
|
|
/* log a dummy value to ensure log structure is fully initialised */
|
|
|
|
to->di_next_unlinked = NULLAGINO;
|
|
|
|
|
2021-08-19 04:46:37 +03:00
|
|
|
if (xfs_has_v3inodes(ip->i_mount)) {
|
2020-03-18 18:15:11 +03:00
|
|
|
to->di_version = 3;
|
2017-12-11 14:35:19 +03:00
|
|
|
to->di_changecount = inode_peek_iversion(inode);
|
2021-03-29 21:11:45 +03:00
|
|
|
to->di_crtime = xfs_inode_to_log_dinode_ts(ip, ip->i_crtime);
|
2021-03-29 21:11:45 +03:00
|
|
|
to->di_flags2 = ip->i_diflags2;
|
2021-03-29 21:11:42 +03:00
|
|
|
to->di_cowextsize = ip->i_cowextsize;
|
2016-02-09 08:54:58 +03:00
|
|
|
to->di_ino = ip->i_ino;
|
|
|
|
to->di_lsn = lsn;
|
|
|
|
memset(to->di_pad2, 0, sizeof(to->di_pad2));
|
|
|
|
uuid_copy(&to->di_uuid, &ip->i_mount->m_sb.sb_meta_uuid);
|
2016-02-09 08:54:58 +03:00
|
|
|
to->di_flushiter = 0;
|
|
|
|
} else {
|
2020-03-18 18:15:11 +03:00
|
|
|
to->di_version = 2;
|
2021-03-29 21:11:42 +03:00
|
|
|
to->di_flushiter = ip->i_flushiter;
|
2016-02-09 08:54:58 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Format the inode core. Current timestamp data is only in the VFS inode
|
|
|
|
* fields, so we need to grab them from there. Hence rather than just copying
|
|
|
|
* the XFS inode core structure, format the fields directly into the iovec.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
xfs_inode_item_format_core(
|
|
|
|
struct xfs_inode *ip,
|
|
|
|
struct xfs_log_vec *lv,
|
|
|
|
struct xfs_log_iovec **vecp)
|
|
|
|
{
|
|
|
|
struct xfs_log_dinode *dic;
|
|
|
|
|
|
|
|
dic = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_ICORE);
|
2016-02-09 08:54:58 +03:00
|
|
|
xfs_inode_to_log_dinode(ip, dic, ip->i_itemp->ili_item.li_lsn);
|
2020-03-18 18:15:10 +03:00
|
|
|
xlog_finish_iovec(lv, *vecp, xfs_log_dinode_size(ip->i_mount));
|
2016-02-09 08:54:58 +03:00
|
|
|
}
|
|
|
|
|
2013-12-13 04:00:43 +04:00
|
|
|
/*
|
|
|
|
* This is called to fill in the vector of log iovecs for the given inode
|
|
|
|
* log item. It fills the first item with an inode log format structure,
|
|
|
|
* the second with the on-disk inode structure, and a possible third and/or
|
|
|
|
* fourth with the inode data/extents/b-tree root and inode attributes
|
|
|
|
* data/extents/b-tree root.
|
2017-10-09 21:37:22 +03:00
|
|
|
*
|
|
|
|
* Note: Always use the 64 bit inode log format structure so we don't
|
|
|
|
* leave an uninitialised hole in the format item on 64 bit systems. Log
|
|
|
|
* recovery on 32 bit systems handles this just fine, so there's no reason
|
|
|
|
* for not using an initialising the properly padded structure all the time.
|
2013-12-13 04:00:43 +04:00
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_inode_item_format(
|
|
|
|
struct xfs_log_item *lip,
|
2013-12-13 04:34:02 +04:00
|
|
|
struct xfs_log_vec *lv)
|
2013-12-13 04:00:43 +04:00
|
|
|
{
|
|
|
|
struct xfs_inode_log_item *iip = INODE_ITEM(lip);
|
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
2013-12-13 04:34:02 +04:00
|
|
|
struct xfs_log_iovec *vecp = NULL;
|
2017-10-09 21:37:22 +03:00
|
|
|
struct xfs_inode_log_format *ilf;
|
2013-12-13 04:00:43 +04:00
|
|
|
|
2013-12-13 04:34:05 +04:00
|
|
|
ilf = xlog_prepare_iovec(lv, &vecp, XLOG_REG_TYPE_IFORMAT);
|
|
|
|
ilf->ilf_type = XFS_LI_INODE;
|
|
|
|
ilf->ilf_ino = ip->i_ino;
|
|
|
|
ilf->ilf_blkno = ip->i_imap.im_blkno;
|
|
|
|
ilf->ilf_len = ip->i_imap.im_len;
|
|
|
|
ilf->ilf_boffset = ip->i_imap.im_boffset;
|
|
|
|
ilf->ilf_fields = XFS_ILOG_CORE;
|
|
|
|
ilf->ilf_size = 2; /* format + core */
|
2017-10-09 21:37:22 +03:00
|
|
|
|
|
|
|
/*
|
|
|
|
* make sure we don't leak uninitialised data into the log in the case
|
|
|
|
* when we don't log every field in the inode.
|
|
|
|
*/
|
|
|
|
ilf->ilf_dsize = 0;
|
|
|
|
ilf->ilf_asize = 0;
|
|
|
|
ilf->ilf_pad = 0;
|
2017-10-19 21:07:09 +03:00
|
|
|
memset(&ilf->ilf_u, 0, sizeof(ilf->ilf_u));
|
2017-10-09 21:37:22 +03:00
|
|
|
|
|
|
|
xlog_finish_iovec(lv, vecp, sizeof(*ilf));
|
2013-12-13 04:00:43 +04:00
|
|
|
|
2016-02-09 08:54:58 +03:00
|
|
|
xfs_inode_item_format_core(ip, lv, &vecp);
|
2013-12-13 04:34:02 +04:00
|
|
|
xfs_inode_item_format_data_fork(iip, ilf, lv, &vecp);
|
2013-12-13 04:00:43 +04:00
|
|
|
if (XFS_IFORK_Q(ip)) {
|
2013-12-13 04:34:02 +04:00
|
|
|
xfs_inode_item_format_attr_fork(iip, ilf, lv, &vecp);
|
2013-12-13 04:00:43 +04:00
|
|
|
} else {
|
|
|
|
iip->ili_fields &=
|
|
|
|
~(XFS_ILOG_ADATA | XFS_ILOG_ABROOT | XFS_ILOG_AEXT);
|
|
|
|
}
|
|
|
|
|
2013-12-13 04:34:05 +04:00
|
|
|
/* update the format with the exact fields we actually logged */
|
|
|
|
ilf->ilf_fields |= (iip->ili_fields & ~XFS_ILOG_TIMESTAMP);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is called to pin the inode associated with the inode log
|
2010-02-18 15:43:22 +03:00
|
|
|
* item in memory so it cannot be written out.
|
2005-04-17 02:20:36 +04:00
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_inode_item_pin(
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_log_item *lip)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_inode *ip = INODE_ITEM(lip)->ili_inode;
|
2010-02-18 15:43:22 +03:00
|
|
|
|
2010-06-23 12:11:15 +04:00
|
|
|
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
ASSERT(lip->li_buf);
|
2010-06-23 12:11:15 +04:00
|
|
|
|
|
|
|
trace_xfs_inode_pin(ip, _RET_IP_);
|
|
|
|
atomic_inc(&ip->i_pincount);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is called to unpin the inode associated with the inode log
|
|
|
|
* item which was previously pinned with a call to xfs_inode_item_pin().
|
2010-02-18 15:43:22 +03:00
|
|
|
*
|
|
|
|
* Also wake up anyone in xfs_iunpin_wait() if the count goes to 0.
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
*
|
|
|
|
* Note that unpin can race with inode cluster buffer freeing marking the buffer
|
|
|
|
* stale. In that case, flush completions are run from the buffer unpin call,
|
|
|
|
* which may happen before the inode is unpinned. If we lose the race, there
|
|
|
|
* will be no buffer attached to the log item, but the inode will be marked
|
|
|
|
* XFS_ISTALE.
|
2005-04-17 02:20:36 +04:00
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_inode_item_unpin(
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_log_item *lip,
|
2010-06-23 12:11:15 +04:00
|
|
|
int remove)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_inode *ip = INODE_ITEM(lip)->ili_inode;
|
2010-02-18 15:43:22 +03:00
|
|
|
|
2010-03-08 03:24:07 +03:00
|
|
|
trace_xfs_inode_unpin(ip, _RET_IP_);
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
ASSERT(lip->li_buf || xfs_iflags_test(ip, XFS_ISTALE));
|
2010-02-18 15:43:22 +03:00
|
|
|
ASSERT(atomic_read(&ip->i_pincount) > 0);
|
|
|
|
if (atomic_dec_and_test(&ip->i_pincount))
|
2011-12-19 00:00:10 +04:00
|
|
|
wake_up_bit(&ip->i_flags, __XFS_IPINNED_BIT);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
STATIC uint
|
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 09:58:39 +04:00
|
|
|
xfs_inode_item_push(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
struct list_head *buffer_list)
|
2018-03-08 01:59:39 +03:00
|
|
|
__releases(&lip->li_ailp->ail_lock)
|
|
|
|
__acquires(&lip->li_ailp->ail_lock)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_inode_log_item *iip = INODE_ITEM(lip);
|
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
xfs: Properly retry failed inode items in case of error during buffer writeback
When a buffer has been failed during writeback, the inode items into it
are kept flush locked, and are never resubmitted due the flush lock, so,
if any buffer fails to be written, the items in AIL are never written to
disk and never unlocked.
This causes unmount operation to hang due these items flush locked in AIL,
but this also causes the items in AIL to never be written back, even when
the IO device comes back to normal.
I've been testing this patch with a DM-thin device, creating a
filesystem larger than the real device.
When writing enough data to fill the DM-thin device, XFS receives ENOSPC
errors from the device, and keep spinning on xfsaild (when 'retry
forever' configuration is set).
At this point, the filesystem can not be unmounted because of the flush locked
items in AIL, but worse, the items in AIL are never retried at all
(once xfs_inode_item_push() will skip the items that are flush locked),
even if the underlying DM-thin device is expanded to the proper size.
This patch fixes both cases, retrying any item that has been failed
previously, using the infra-structure provided by the previous patch.
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-08-09 04:21:50 +03:00
|
|
|
struct xfs_buf *bp = lip->li_buf;
|
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 09:58:39 +04:00
|
|
|
uint rval = XFS_ITEM_SUCCESS;
|
|
|
|
int error;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2020-06-30 00:49:19 +03:00
|
|
|
ASSERT(iip->ili_item.li_buf);
|
|
|
|
|
|
|
|
if (xfs_ipincount(ip) > 0 || xfs_buf_ispinned(bp) ||
|
|
|
|
(ip->i_flags & XFS_ISTALE))
|
2005-04-17 02:20:36 +04:00
|
|
|
return XFS_ITEM_PINNED;
|
|
|
|
|
2020-08-18 02:41:01 +03:00
|
|
|
if (xfs_iflags_test(ip, XFS_IFLUSHING))
|
2020-06-30 00:49:19 +03:00
|
|
|
return XFS_ITEM_FLUSHING;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2020-06-30 00:49:19 +03:00
|
|
|
if (!xfs_buf_trylock(bp))
|
|
|
|
return XFS_ITEM_LOCKED;
|
2012-04-23 09:58:36 +04:00
|
|
|
|
2020-06-30 00:49:19 +03:00
|
|
|
spin_unlock(&lip->li_ailp->ail_lock);
|
2012-06-11 18:39:43 +04:00
|
|
|
|
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 09:58:39 +04:00
|
|
|
/*
|
2020-06-30 00:49:19 +03:00
|
|
|
* We need to hold a reference for flushing the cluster buffer as it may
|
|
|
|
* fail the buffer without IO submission. In which case, we better get a
|
|
|
|
* reference for that completion because otherwise we don't get a
|
|
|
|
* reference for IO until we queue the buffer for delwri submission.
|
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 09:58:39 +04:00
|
|
|
*/
|
2020-06-30 00:49:19 +03:00
|
|
|
xfs_buf_hold(bp);
|
2020-06-30 00:49:20 +03:00
|
|
|
error = xfs_iflush_cluster(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 09:58:39 +04:00
|
|
|
if (!error) {
|
|
|
|
if (!xfs_buf_delwri_queue(bp, buffer_list))
|
|
|
|
rval = XFS_ITEM_FLUSHING;
|
|
|
|
xfs_buf_relse(bp);
|
2020-06-30 00:49:19 +03:00
|
|
|
} else {
|
2020-06-30 00:49:20 +03:00
|
|
|
/*
|
|
|
|
* Release the buffer if we were unable to flush anything. On
|
|
|
|
* any other error, the buffer has already been released.
|
|
|
|
*/
|
|
|
|
if (error == -EAGAIN)
|
|
|
|
xfs_buf_relse(bp);
|
2020-03-27 18:29:55 +03:00
|
|
|
rval = XFS_ITEM_LOCKED;
|
2020-06-30 00:49:19 +03:00
|
|
|
}
|
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 09:58:39 +04:00
|
|
|
|
2018-03-08 01:59:39 +03:00
|
|
|
spin_lock(&lip->li_ailp->ail_lock);
|
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 09:58:39 +04:00
|
|
|
return rval;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Unlock the inode associated with the inode log item.
|
|
|
|
*/
|
|
|
|
STATIC void
|
2019-06-29 05:27:32 +03:00
|
|
|
xfs_inode_item_release(
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_log_item *lip)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_inode_log_item *iip = INODE_ITEM(lip);
|
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
2010-06-24 05:36:58 +04:00
|
|
|
unsigned short lock_flags;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2011-07-08 16:34:47 +04:00
|
|
|
ASSERT(ip->i_itemp != NULL);
|
|
|
|
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2010-06-24 05:36:58 +04:00
|
|
|
lock_flags = iip->ili_lock_flags;
|
|
|
|
iip->ili_lock_flags = 0;
|
2011-09-19 19:00:54 +04:00
|
|
|
if (lock_flags)
|
2011-07-08 16:34:47 +04:00
|
|
|
xfs_iunlock(ip, lock_flags);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2010-11-30 07:15:46 +03:00
|
|
|
* This is called to find out where the oldest active copy of the inode log
|
|
|
|
* item in the on disk log resides now that the last log write of it completed
|
|
|
|
* at the given lsn. Since we always re-log all dirty data in an inode, the
|
|
|
|
* latest copy in the on disk log is the only one that matters. Therefore,
|
|
|
|
* simply return the given lsn.
|
|
|
|
*
|
|
|
|
* If the inode has been marked stale because the cluster is being freed, we
|
|
|
|
* don't want to (re-)insert this inode into the AIL. There is a race condition
|
|
|
|
* where the cluster buffer may be unpinned before the inode is inserted into
|
|
|
|
* the AIL during transaction committed processing. If the buffer is unpinned
|
|
|
|
* before the inode item has been committed and inserted, then it is possible
|
xfs: unpin stale inodes directly in IOP_COMMITTED
When inodes are marked stale in a transaction, they are treated
specially when the inode log item is being inserted into the AIL.
It tries to avoid moving the log item forward in the AIL due to a
race condition with the writing the underlying buffer back to disk.
The was "fixed" in commit de25c18 ("xfs: avoid moving stale inodes
in the AIL").
To avoid moving the item forward, we return a LSN smaller than the
commit_lsn of the completing transaction, thereby trying to trick
the commit code into not moving the inode forward at all. I'm not
sure this ever worked as intended - it assumes the inode is already
in the AIL, but I don't think the returned LSN would have been small
enough to prevent moving the inode. It appears that the reason it
worked is that the lower LSN of the inodes meant they were inserted
into the AIL and flushed before the inode buffer (which was moved to
the commit_lsn of the transaction).
The big problem is that with delayed logging, the returning of the
different LSN means insertion takes the slow, non-bulk path. Worse
yet is that insertion is to a position -before- the commit_lsn so it
is doing a AIL traversal on every insertion, and has to walk over
all the items that have already been inserted into the AIL. It's
expensive.
To compound the matter further, with delayed logging inodes are
likely to go from clean to stale in a single checkpoint, which means
they aren't even in the AIL at all when we come across them at AIL
insertion time. Hence these were all getting inserted into the AIL
when they simply do not need to be as inodes marked XFS_ISTALE are
never written back.
Transactional/recovery integrity is maintained in this case by the
other items in the unlink transaction that were modified (e.g. the
AGI btree blocks) and committed in the same checkpoint.
So to fix this, simply unpin the stale inodes directly in
xfs_inode_item_committed() and return -1 to indicate that the AIL
insertion code does not need to do any further processing of these
inodes.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
2011-07-04 09:27:36 +04:00
|
|
|
* for the buffer to be written and IO completes before the inode is inserted
|
2010-11-30 07:15:46 +03:00
|
|
|
* into the AIL. In that case, we'd be inserting a clean, stale inode into the
|
|
|
|
* AIL which will never get removed. It will, however, get reclaimed which
|
|
|
|
* triggers an assert in xfs_inode_free() complaining about freein an inode
|
|
|
|
* still in the AIL.
|
|
|
|
*
|
xfs: unpin stale inodes directly in IOP_COMMITTED
When inodes are marked stale in a transaction, they are treated
specially when the inode log item is being inserted into the AIL.
It tries to avoid moving the log item forward in the AIL due to a
race condition with the writing the underlying buffer back to disk.
The was "fixed" in commit de25c18 ("xfs: avoid moving stale inodes
in the AIL").
To avoid moving the item forward, we return a LSN smaller than the
commit_lsn of the completing transaction, thereby trying to trick
the commit code into not moving the inode forward at all. I'm not
sure this ever worked as intended - it assumes the inode is already
in the AIL, but I don't think the returned LSN would have been small
enough to prevent moving the inode. It appears that the reason it
worked is that the lower LSN of the inodes meant they were inserted
into the AIL and flushed before the inode buffer (which was moved to
the commit_lsn of the transaction).
The big problem is that with delayed logging, the returning of the
different LSN means insertion takes the slow, non-bulk path. Worse
yet is that insertion is to a position -before- the commit_lsn so it
is doing a AIL traversal on every insertion, and has to walk over
all the items that have already been inserted into the AIL. It's
expensive.
To compound the matter further, with delayed logging inodes are
likely to go from clean to stale in a single checkpoint, which means
they aren't even in the AIL at all when we come across them at AIL
insertion time. Hence these were all getting inserted into the AIL
when they simply do not need to be as inodes marked XFS_ISTALE are
never written back.
Transactional/recovery integrity is maintained in this case by the
other items in the unlink transaction that were modified (e.g. the
AGI btree blocks) and committed in the same checkpoint.
So to fix this, simply unpin the stale inodes directly in
xfs_inode_item_committed() and return -1 to indicate that the AIL
insertion code does not need to do any further processing of these
inodes.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
2011-07-04 09:27:36 +04:00
|
|
|
* To avoid this, just unpin the inode directly and return a LSN of -1 so the
|
|
|
|
* transaction committed code knows that it does not need to do any further
|
|
|
|
* processing on the item.
|
2005-04-17 02:20:36 +04:00
|
|
|
*/
|
|
|
|
STATIC xfs_lsn_t
|
|
|
|
xfs_inode_item_committed(
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_log_item *lip,
|
2005-04-17 02:20:36 +04:00
|
|
|
xfs_lsn_t lsn)
|
|
|
|
{
|
2010-11-30 07:15:46 +03:00
|
|
|
struct xfs_inode_log_item *iip = INODE_ITEM(lip);
|
|
|
|
struct xfs_inode *ip = iip->ili_inode;
|
|
|
|
|
xfs: unpin stale inodes directly in IOP_COMMITTED
When inodes are marked stale in a transaction, they are treated
specially when the inode log item is being inserted into the AIL.
It tries to avoid moving the log item forward in the AIL due to a
race condition with the writing the underlying buffer back to disk.
The was "fixed" in commit de25c18 ("xfs: avoid moving stale inodes
in the AIL").
To avoid moving the item forward, we return a LSN smaller than the
commit_lsn of the completing transaction, thereby trying to trick
the commit code into not moving the inode forward at all. I'm not
sure this ever worked as intended - it assumes the inode is already
in the AIL, but I don't think the returned LSN would have been small
enough to prevent moving the inode. It appears that the reason it
worked is that the lower LSN of the inodes meant they were inserted
into the AIL and flushed before the inode buffer (which was moved to
the commit_lsn of the transaction).
The big problem is that with delayed logging, the returning of the
different LSN means insertion takes the slow, non-bulk path. Worse
yet is that insertion is to a position -before- the commit_lsn so it
is doing a AIL traversal on every insertion, and has to walk over
all the items that have already been inserted into the AIL. It's
expensive.
To compound the matter further, with delayed logging inodes are
likely to go from clean to stale in a single checkpoint, which means
they aren't even in the AIL at all when we come across them at AIL
insertion time. Hence these were all getting inserted into the AIL
when they simply do not need to be as inodes marked XFS_ISTALE are
never written back.
Transactional/recovery integrity is maintained in this case by the
other items in the unlink transaction that were modified (e.g. the
AGI btree blocks) and committed in the same checkpoint.
So to fix this, simply unpin the stale inodes directly in
xfs_inode_item_committed() and return -1 to indicate that the AIL
insertion code does not need to do any further processing of these
inodes.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
2011-07-04 09:27:36 +04:00
|
|
|
if (xfs_iflags_test(ip, XFS_ISTALE)) {
|
|
|
|
xfs_inode_item_unpin(lip, 0);
|
|
|
|
return -1;
|
|
|
|
}
|
2010-06-23 12:11:15 +04:00
|
|
|
return lsn;
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void
|
|
|
|
xfs_inode_item_committing(
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_log_item *lip,
|
2021-06-18 18:21:52 +03:00
|
|
|
xfs_csn_t seq)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2021-06-18 18:21:52 +03:00
|
|
|
INODE_ITEM(lip)->ili_commit_seq = seq;
|
2019-06-29 05:27:32 +03:00
|
|
|
return xfs_inode_item_release(lip);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
2011-10-28 13:54:24 +04:00
|
|
|
static const struct xfs_item_ops xfs_inode_item_ops = {
|
2010-06-23 12:11:15 +04:00
|
|
|
.iop_size = xfs_inode_item_size,
|
|
|
|
.iop_format = xfs_inode_item_format,
|
|
|
|
.iop_pin = xfs_inode_item_pin,
|
|
|
|
.iop_unpin = xfs_inode_item_unpin,
|
2019-06-29 05:27:32 +03:00
|
|
|
.iop_release = xfs_inode_item_release,
|
2010-06-23 12:11:15 +04:00
|
|
|
.iop_committed = xfs_inode_item_committed,
|
|
|
|
.iop_push = xfs_inode_item_push,
|
2019-06-29 05:27:32 +03:00
|
|
|
.iop_committing = xfs_inode_item_committing,
|
2005-04-17 02:20:36 +04:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Initialize the inode log item for a newly allocated (in-core) inode.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
xfs_inode_item_init(
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_inode *ip,
|
|
|
|
struct xfs_mount *mp)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2010-06-23 12:11:15 +04:00
|
|
|
struct xfs_inode_log_item *iip;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
|
|
|
ASSERT(ip->i_itemp == NULL);
|
2021-10-12 21:09:23 +03:00
|
|
|
iip = ip->i_itemp = kmem_cache_zalloc(xfs_ili_cache,
|
2020-07-22 19:23:10 +03:00
|
|
|
GFP_KERNEL | __GFP_NOFAIL);
|
2005-04-17 02:20:36 +04:00
|
|
|
|
|
|
|
iip->ili_inode = ip;
|
2020-06-30 00:48:46 +03:00
|
|
|
spin_lock_init(&iip->ili_lock);
|
2010-03-23 02:10:00 +03:00
|
|
|
xfs_log_item_init(mp, &iip->ili_item, XFS_LI_INODE,
|
|
|
|
&xfs_inode_item_ops);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Free the inode log item and any memory hanging off of it.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
xfs_inode_item_destroy(
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
struct xfs_inode *ip)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
struct xfs_inode_log_item *iip = ip->i_itemp;
|
|
|
|
|
|
|
|
ASSERT(iip->ili_item.li_buf == NULL);
|
|
|
|
|
|
|
|
ip->i_itemp = NULL;
|
|
|
|
kmem_free(iip->ili_item.li_lv_shadow);
|
2021-10-12 21:09:23 +03:00
|
|
|
kmem_cache_free(xfs_ili_cache, iip);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
2020-06-30 00:49:20 +03:00
|
|
|
* We only want to pull the item from the AIL if it is actually there
|
|
|
|
* and its location in the log has not changed since we started the
|
|
|
|
* flush. Thus, we only bother if the inode's lsn has not changed.
|
2005-04-17 02:20:36 +04:00
|
|
|
*/
|
2020-06-30 00:49:20 +03:00
|
|
|
static void
|
|
|
|
xfs_iflush_ail_updates(
|
|
|
|
struct xfs_ail *ailp,
|
|
|
|
struct list_head *list)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
2020-06-30 00:49:20 +03:00
|
|
|
struct xfs_log_item *lip;
|
|
|
|
xfs_lsn_t tail_lsn = 0;
|
2010-12-20 04:03:17 +03:00
|
|
|
|
2020-06-30 00:49:20 +03:00
|
|
|
/* this is an opencoded batch version of xfs_trans_ail_delete */
|
|
|
|
spin_lock(&ailp->ail_lock);
|
|
|
|
list_for_each_entry(lip, list, li_bio_list) {
|
|
|
|
xfs_lsn_t lsn;
|
2010-12-20 04:03:17 +03:00
|
|
|
|
2020-06-30 00:49:20 +03:00
|
|
|
clear_bit(XFS_LI_FAILED, &lip->li_flags);
|
|
|
|
if (INODE_ITEM(lip)->ili_flush_lsn != lip->li_lsn)
|
2020-06-30 00:49:18 +03:00
|
|
|
continue;
|
|
|
|
|
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 19:09:13 +03:00
|
|
|
/*
|
|
|
|
* dgc: Not sure how this happens, but it happens very
|
|
|
|
* occassionaly via generic/388. xfs_iflush_abort() also
|
|
|
|
* silently handles this same "under writeback but not in AIL at
|
|
|
|
* shutdown" condition via xfs_trans_ail_delete().
|
|
|
|
*/
|
|
|
|
if (!test_bit(XFS_LI_IN_AIL, &lip->li_flags)) {
|
|
|
|
ASSERT(xlog_is_shutdown(lip->li_log));
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
2020-06-30 00:49:20 +03:00
|
|
|
lsn = xfs_ail_delete_one(ailp, lip);
|
|
|
|
if (!tail_lsn && lsn)
|
|
|
|
tail_lsn = lsn;
|
2010-12-20 04:03:17 +03:00
|
|
|
}
|
2020-06-30 00:49:20 +03:00
|
|
|
xfs_ail_update_finish(ailp, tail_lsn);
|
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2020-06-30 00:49:20 +03:00
|
|
|
/*
|
|
|
|
* Walk the list of inodes that have completed their IOs. If they are clean
|
|
|
|
* remove them from the list and dissociate them from the buffer. Buffers that
|
|
|
|
* are still dirty remain linked to the buffer and on the list. Caller must
|
|
|
|
* handle them appropriately.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
xfs_iflush_finish(
|
|
|
|
struct xfs_buf *bp,
|
|
|
|
struct list_head *list)
|
|
|
|
{
|
|
|
|
struct xfs_log_item *lip, *n;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
2020-06-30 00:49:20 +03:00
|
|
|
list_for_each_entry_safe(lip, n, list, li_bio_list) {
|
|
|
|
struct xfs_inode_log_item *iip = INODE_ITEM(lip);
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
bool drop_buffer = false;
|
|
|
|
|
2020-06-30 00:48:46 +03:00
|
|
|
spin_lock(&iip->ili_lock);
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Remove the reference to the cluster buffer if the inode is
|
2020-06-30 00:49:20 +03:00
|
|
|
* clean in memory and drop the buffer reference once we've
|
|
|
|
* dropped the locks we hold.
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
*/
|
|
|
|
ASSERT(iip->ili_item.li_buf == bp);
|
|
|
|
if (!iip->ili_fields) {
|
|
|
|
iip->ili_item.li_buf = NULL;
|
2020-06-30 00:49:20 +03:00
|
|
|
list_del_init(&lip->li_bio_list);
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
drop_buffer = true;
|
|
|
|
}
|
2010-12-20 04:03:17 +03:00
|
|
|
iip->ili_last_fields = 0;
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
iip->ili_flush_lsn = 0;
|
2020-06-30 00:48:46 +03:00
|
|
|
spin_unlock(&iip->ili_lock);
|
2020-08-18 02:41:01 +03:00
|
|
|
xfs_iflags_clear(iip->ili_inode, XFS_IFLUSHING);
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
if (drop_buffer)
|
|
|
|
xfs_buf_rele(bp);
|
2010-12-20 04:03:17 +03:00
|
|
|
}
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
2020-06-30 00:49:20 +03:00
|
|
|
/*
|
|
|
|
* Inode buffer IO completion routine. It is responsible for removing inodes
|
2020-08-18 02:41:01 +03:00
|
|
|
* attached to the buffer from the AIL if they have not been re-logged and
|
|
|
|
* completing the inode flush.
|
2020-06-30 00:49:20 +03:00
|
|
|
*/
|
|
|
|
void
|
2020-09-01 20:55:29 +03:00
|
|
|
xfs_buf_inode_iodone(
|
2020-06-30 00:49:20 +03:00
|
|
|
struct xfs_buf *bp)
|
|
|
|
{
|
|
|
|
struct xfs_log_item *lip, *n;
|
|
|
|
LIST_HEAD(flushed_inodes);
|
|
|
|
LIST_HEAD(ail_updates);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Pull the attached inodes from the buffer one at a time and take the
|
|
|
|
* appropriate action on them.
|
|
|
|
*/
|
|
|
|
list_for_each_entry_safe(lip, n, &bp->b_li_list, li_bio_list) {
|
|
|
|
struct xfs_inode_log_item *iip = INODE_ITEM(lip);
|
|
|
|
|
|
|
|
if (xfs_iflags_test(iip->ili_inode, XFS_ISTALE)) {
|
|
|
|
xfs_iflush_abort(iip->ili_inode);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!iip->ili_last_fields)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
/* Do an unlocked check for needing the AIL lock. */
|
|
|
|
if (iip->ili_flush_lsn == lip->li_lsn ||
|
|
|
|
test_bit(XFS_LI_FAILED, &lip->li_flags))
|
|
|
|
list_move_tail(&lip->li_bio_list, &ail_updates);
|
|
|
|
else
|
|
|
|
list_move_tail(&lip->li_bio_list, &flushed_inodes);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!list_empty(&ail_updates)) {
|
|
|
|
xfs_iflush_ail_updates(bp->b_mount->m_ail, &ail_updates);
|
|
|
|
list_splice_tail(&ail_updates, &flushed_inodes);
|
|
|
|
}
|
|
|
|
|
|
|
|
xfs_iflush_finish(bp, &flushed_inodes);
|
|
|
|
if (!list_empty(&flushed_inodes))
|
|
|
|
list_splice_tail(&flushed_inodes, &bp->b_li_list);
|
|
|
|
}
|
|
|
|
|
2020-09-01 20:55:29 +03:00
|
|
|
void
|
|
|
|
xfs_buf_inode_io_fail(
|
|
|
|
struct xfs_buf *bp)
|
|
|
|
{
|
|
|
|
struct xfs_log_item *lip;
|
|
|
|
|
|
|
|
list_for_each_entry(lip, &bp->b_li_list, li_bio_list)
|
|
|
|
set_bit(XFS_LI_FAILED, &lip->li_flags);
|
|
|
|
}
|
|
|
|
|
2005-04-17 02:20:36 +04:00
|
|
|
/*
|
2020-08-18 02:41:01 +03:00
|
|
|
* This is the inode flushing abort routine. It is called when
|
2012-04-23 09:58:41 +04:00
|
|
|
* the filesystem is shutting down to clean up the inode state. It is
|
|
|
|
* responsible for removing the inode item from the AIL if it has not been
|
2020-08-18 02:41:01 +03:00
|
|
|
* re-logged and clearing the inode's flush state.
|
2005-04-17 02:20:36 +04:00
|
|
|
*/
|
|
|
|
void
|
|
|
|
xfs_iflush_abort(
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
struct xfs_inode *ip)
|
2005-04-17 02:20:36 +04:00
|
|
|
{
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
struct xfs_inode_log_item *iip = ip->i_itemp;
|
|
|
|
struct xfs_buf *bp = NULL;
|
2005-04-17 02:20:36 +04:00
|
|
|
|
|
|
|
if (iip) {
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
/*
|
|
|
|
* Clear the failed bit before removing the item from the AIL so
|
|
|
|
* xfs_trans_ail_delete() doesn't try to clear and release the
|
|
|
|
* buffer attached to the log item before we are done with it.
|
|
|
|
*/
|
|
|
|
clear_bit(XFS_LI_FAILED, &iip->ili_item.li_flags);
|
2020-05-06 23:27:04 +03:00
|
|
|
xfs_trans_ail_delete(&iip->ili_item, 0);
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
|
2005-04-17 02:20:36 +04:00
|
|
|
/*
|
|
|
|
* Clear the inode logging fields so no more flushes are
|
|
|
|
* attempted.
|
|
|
|
*/
|
2020-06-30 00:48:46 +03:00
|
|
|
spin_lock(&iip->ili_lock);
|
2020-06-30 00:48:45 +03:00
|
|
|
iip->ili_last_fields = 0;
|
2012-02-29 13:53:54 +04:00
|
|
|
iip->ili_fields = 0;
|
xfs: optimise away log forces on timestamp updates for fdatasync
xfs: timestamp updates cause excessive fdatasync log traffic
Sage Weil reported that a ceph test workload was writing to the
log on every fdatasync during an overwrite workload. Event tracing
showed that the only metadata modification being made was the
timestamp updates during the write(2) syscall, but fdatasync(2)
is supposed to ignore them. The key observation was that the
transactions in the log all looked like this:
INODE: #regs: 4 ino: 0x8b flags: 0x45 dsize: 32
And contained a flags field of 0x45 or 0x85, and had data and
attribute forks following the inode core. This means that the
timestamp updates were triggering dirty relogging of previously
logged parts of the inode that hadn't yet been flushed back to
disk.
There are two parts to this problem. The first is that XFS relogs
dirty regions in subsequent transactions, so it carries around the
fields that have been dirtied since the last time the inode was
written back to disk, not since the last time the inode was forced
into the log.
The second part is that on v5 filesystems, the inode change count
update during inode dirtying also sets the XFS_ILOG_CORE flag, so
on v5 filesystems this makes a timestamp update dirty the entire
inode.
As a result when fdatasync is run, it looks at the dirty fields in
the inode, and sees more than just the timestamp flag, even though
the only metadata change since the last fdatasync was just the
timestamps. Hence we force the log on every subsequent fdatasync
even though it is not needed.
To fix this, add a new field to the inode log item that tracks
changes since the last time fsync/fdatasync forced the log to flush
the changes to the journal. This flag is updated when we dirty the
inode, but we do it before updating the change count so it does not
carry the "core dirty" flag from timestamp updates. The fields are
zeroed when the inode is marked clean (due to writeback/freeing) or
when an fsync/datasync forces the log. Hence if we only dirty the
timestamps on the inode between fsync/fdatasync calls, the fdatasync
will not trigger another log force.
Over 100 runs of the test program:
Ext4 baseline:
runtime: 1.63s +/- 0.24s
avg lat: 1.59ms +/- 0.24ms
iops: ~2000
XFS, vanilla kernel:
runtime: 2.45s +/- 0.18s
avg lat: 2.39ms +/- 0.18ms
log forces: ~400/s
iops: ~1000
XFS, patched kernel:
runtime: 1.49s +/- 0.26s
avg lat: 1.46ms +/- 0.25ms
log forces: ~30/s
iops: ~1500
Reported-by: Sage Weil <sage@redhat.com>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
2015-11-03 05:14:59 +03:00
|
|
|
iip->ili_fsync_fields = 0;
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
iip->ili_flush_lsn = 0;
|
|
|
|
bp = iip->ili_item.li_buf;
|
|
|
|
iip->ili_item.li_buf = NULL;
|
2020-06-30 00:49:18 +03:00
|
|
|
list_del_init(&iip->ili_item.li_bio_list);
|
2020-06-30 00:48:46 +03:00
|
|
|
spin_unlock(&iip->ili_lock);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
2020-08-18 02:41:01 +03:00
|
|
|
xfs_iflags_clear(ip, XFS_IFLUSHING);
|
xfs: pin inode backing buffer to the inode log item
When we dirty an inode, we are going to have to write it disk at
some point in the near future. This requires the inode cluster
backing buffer to be present in memory. Unfortunately, under severe
memory pressure we can reclaim the inode backing buffer while the
inode is dirty in memory, resulting in stalling the AIL pushing
because it has to do a read-modify-write cycle on the cluster
buffer.
When we have no memory available, the read of the cluster buffer
blocks the AIL pushing process, and this causes all sorts of issues
for memory reclaim as it requires inode writeback to make forwards
progress. Allocating a cluster buffer causes more memory pressure,
and results in more cluster buffers to be reclaimed, resulting in
more RMW cycles to be done in the AIL context and everything then
backs up on AIL progress. Only the synchronous inode cluster
writeback in the the inode reclaim code provides some level of
forwards progress guarantees that prevent OOM-killer rampages in
this situation.
Fix this by pinning the inode backing buffer to the inode log item
when the inode is first dirtied (i.e. in xfs_trans_log_inode()).
This may mean the first modification of an inode that has been held
in cache for a long time may block on a cluster buffer read, but
we can do that in transaction context and block safely until the
buffer has been allocated and read.
Once we have the cluster buffer, the inode log item takes a
reference to it, pinning it in memory, and attaches it to the log
item for future reference. This means we can always grab the cluster
buffer from the inode log item when we need it.
When the inode is finally cleaned and removed from the AIL, we can
drop the reference the inode log item holds on the cluster buffer.
Once all inodes on the cluster buffer are clean, the cluster buffer
will be unpinned and it will be available for memory reclaim to
reclaim again.
This avoids the issues with needing to do RMW cycles in the AIL
pushing context, and hence allows complete non-blocking inode
flushing to be performed by the AIL pushing context.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-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>
2020-06-30 00:49:15 +03:00
|
|
|
if (bp)
|
|
|
|
xfs_buf_rele(bp);
|
2005-04-17 02:20:36 +04:00
|
|
|
}
|
|
|
|
|
2006-06-09 08:55:38 +04:00
|
|
|
/*
|
2017-10-09 21:37:22 +03:00
|
|
|
* convert an xfs_inode_log_format struct from the old 32 bit version
|
|
|
|
* (which can have different field alignments) to the native 64 bit version
|
2006-06-09 08:55:38 +04:00
|
|
|
*/
|
|
|
|
int
|
|
|
|
xfs_inode_item_format_convert(
|
2017-10-09 21:37:22 +03:00
|
|
|
struct xfs_log_iovec *buf,
|
|
|
|
struct xfs_inode_log_format *in_f)
|
2006-06-09 08:55:38 +04:00
|
|
|
{
|
2017-10-09 21:37:22 +03:00
|
|
|
struct xfs_inode_log_format_32 *in_f32 = buf->i_addr;
|
|
|
|
|
2019-11-02 19:40:53 +03:00
|
|
|
if (buf->i_len != sizeof(*in_f32)) {
|
|
|
|
XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, NULL);
|
2017-10-09 21:37:22 +03:00
|
|
|
return -EFSCORRUPTED;
|
2019-11-02 19:40:53 +03:00
|
|
|
}
|
2017-10-09 21:37:22 +03:00
|
|
|
|
|
|
|
in_f->ilf_type = in_f32->ilf_type;
|
|
|
|
in_f->ilf_size = in_f32->ilf_size;
|
|
|
|
in_f->ilf_fields = in_f32->ilf_fields;
|
|
|
|
in_f->ilf_asize = in_f32->ilf_asize;
|
|
|
|
in_f->ilf_dsize = in_f32->ilf_dsize;
|
|
|
|
in_f->ilf_ino = in_f32->ilf_ino;
|
2017-10-19 21:07:09 +03:00
|
|
|
memcpy(&in_f->ilf_u, &in_f32->ilf_u, sizeof(in_f->ilf_u));
|
2017-10-09 21:37:22 +03:00
|
|
|
in_f->ilf_blkno = in_f32->ilf_blkno;
|
|
|
|
in_f->ilf_len = in_f32->ilf_len;
|
|
|
|
in_f->ilf_boffset = in_f32->ilf_boffset;
|
|
|
|
return 0;
|
2006-06-09 08:55:38 +04:00
|
|
|
}
|