Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-11-23 04:31:50 +00:00
Code Issues Packages Projects Releases Wiki Activity
a47bd1e0e6
linux/fs/xfs/scrub/health.c
Darrick J. Wong d5c88131db xfs: allow queued AG intents to drain before scrubbing 
When a writer thread executes a chain of log intent items, the AG header
buffer locks will cycle during a transaction roll to get from one intent
item to the next in a chain.  Although scrub takes all AG header buffer
locks, this isn't sufficient to guard against scrub checking an AG while
that writer thread is in the middle of finishing a chain because there's
no higher level locking primitive guarding allocation groups.

When there's a collision, cross-referencing between data structures
(e.g. rmapbt and refcountbt) yields false corruption events; if repair
is running, this results in incorrect repairs, which is catastrophic.

Fix this by adding to the perag structure the count of active intents
and make scrub wait until it has both AG header buffer locks and the
intent counter reaches zero.

One quirk of the drain code is that deferred bmap updates also bump and
drop the intent counter.  A fundamental decision made during the design
phase of the reverse mapping feature is that updates to the rmapbt
records are always made by the same code that updates the primary
metadata.  In other words, callers of bmapi functions expect that the
bmapi functions will queue deferred rmap updates.

Some parts of the reflink code queue deferred refcount (CUI) and bmap
(BUI) updates in the same head transaction, but the deferred work
manager completely finishes the CUI before the BUI work is started.  As
a result, the CUI drops the intent count long before the deferred rmap
(RUI) update even has a chance to bump the intent count.  The only way
to keep the intent count elevated between the CUI and RUI is for the BUI
to bump the counter until the RUI has been created.

A second quirk of the intent drain code is that deferred work items must
increment the intent counter as soon as the work item is added to the
transaction.  When a BUI completes and queues an RUI, the RUI must
increment the counter before the BUI decrements it.  The only way to
accomplish this is to require that the counter be bumped as soon as the
deferred work item is created in memory.

In the next patches we'll improve on this facility, but this patch
provides the basic functionality.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
2023-04-11 18:59:58 -07:00

236 lines
7.3 KiB
C
Raw Blame History

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2019-2023 Oracle. All Rights Reserved.
* Author: Darrick J. Wong <djwong@kernel.org>
*/
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_shared.h"
#include "xfs_format.h"
#include "xfs_trans_resv.h"
#include "xfs_mount.h"
#include "xfs_btree.h"
#include "xfs_trans_resv.h"
#include "xfs_mount.h"
#include "xfs_ag.h"
#include "xfs_health.h"
#include "scrub/scrub.h"
#include "scrub/health.h"
/*
* Scrub and In-Core Filesystem Health Assessments
* ===============================================
*
* Online scrub and repair have the time and the ability to perform stronger
* checks than we can do from the metadata verifiers, because they can
* cross-reference records between data structures. Therefore, scrub is in a
* good position to update the online filesystem health assessments to reflect
* the good/bad state of the data structure.
*
* We therefore extend scrub in the following ways to achieve this:
*
* 1. Create a "sick_mask" field in the scrub context. When we're setting up a
* scrub call, set this to the default XFS_SICK_* flag(s) for the selected
* scrub type (call it A). Scrub and repair functions can override the default
* sick_mask value if they choose.
*
* 2. If the scrubber returns a runtime error code, we exit making no changes
* to the incore sick state.
*
* 3. If the scrubber finds that A is clean, use sick_mask to clear the incore
* sick flags before exiting.
*
* 4. If the scrubber finds that A is corrupt, use sick_mask to set the incore
* sick flags. If the user didn't want to repair then we exit, leaving the
* metadata structure unfixed and the sick flag set.
*
* 5. Now we know that A is corrupt and the user wants to repair, so run the
* repairer. If the repairer returns an error code, we exit with that error
* code, having made no further changes to the incore sick state.
*
* 6. If repair rebuilds A correctly and the subsequent re-scrub of A is clean,
* use sick_mask to clear the incore sick flags. This should have the effect
* that A is no longer marked sick.
*
* 7. If repair rebuilds A incorrectly, the re-scrub will find it corrupt and
* use sick_mask to set the incore sick flags. This should have no externally
* visible effect since we already set them in step (4).
*
* There are some complications to this story, however. For certain types of
* complementary metadata indices (e.g. inobt/finobt), it is easier to rebuild
* both structures at the same time. The following principles apply to this
* type of repair strategy:
*
* 8. Any repair function that rebuilds multiple structures should update
* sick_mask_visible to reflect whatever other structures are rebuilt, and
* verify that all the rebuilt structures can pass a scrub check. The outcomes
* of 5-7 still apply, but with a sick_mask that covers everything being
* rebuilt.
*/
/* Map our scrub type to a sick mask and a set of health update functions. */
enum xchk_health_group {
XHG_FS = 1,
XHG_RT,
XHG_AG,
XHG_INO,
};
struct xchk_health_map {
enum xchk_health_group group;
unsigned int sick_mask;
};
static const struct xchk_health_map type_to_health_flag[XFS_SCRUB_TYPE_NR] = {
[XFS_SCRUB_TYPE_SB] = { XHG_AG, XFS_SICK_AG_SB },
[XFS_SCRUB_TYPE_AGF] = { XHG_AG, XFS_SICK_AG_AGF },
[XFS_SCRUB_TYPE_AGFL] = { XHG_AG, XFS_SICK_AG_AGFL },
[XFS_SCRUB_TYPE_AGI] = { XHG_AG, XFS_SICK_AG_AGI },
[XFS_SCRUB_TYPE_BNOBT] = { XHG_AG, XFS_SICK_AG_BNOBT },
[XFS_SCRUB_TYPE_CNTBT] = { XHG_AG, XFS_SICK_AG_CNTBT },
[XFS_SCRUB_TYPE_INOBT] = { XHG_AG, XFS_SICK_AG_INOBT },
[XFS_SCRUB_TYPE_FINOBT] = { XHG_AG, XFS_SICK_AG_FINOBT },
[XFS_SCRUB_TYPE_RMAPBT] = { XHG_AG, XFS_SICK_AG_RMAPBT },
[XFS_SCRUB_TYPE_REFCNTBT] = { XHG_AG, XFS_SICK_AG_REFCNTBT },
[XFS_SCRUB_TYPE_INODE] = { XHG_INO, XFS_SICK_INO_CORE },
[XFS_SCRUB_TYPE_BMBTD] = { XHG_INO, XFS_SICK_INO_BMBTD },
[XFS_SCRUB_TYPE_BMBTA] = { XHG_INO, XFS_SICK_INO_BMBTA },
[XFS_SCRUB_TYPE_BMBTC] = { XHG_INO, XFS_SICK_INO_BMBTC },
[XFS_SCRUB_TYPE_DIR] = { XHG_INO, XFS_SICK_INO_DIR },
[XFS_SCRUB_TYPE_XATTR] = { XHG_INO, XFS_SICK_INO_XATTR },
[XFS_SCRUB_TYPE_SYMLINK] = { XHG_INO, XFS_SICK_INO_SYMLINK },
[XFS_SCRUB_TYPE_PARENT] = { XHG_INO, XFS_SICK_INO_PARENT },
[XFS_SCRUB_TYPE_RTBITMAP] = { XHG_RT, XFS_SICK_RT_BITMAP },
[XFS_SCRUB_TYPE_RTSUM] = { XHG_RT, XFS_SICK_RT_SUMMARY },
[XFS_SCRUB_TYPE_UQUOTA] = { XHG_FS, XFS_SICK_FS_UQUOTA },
[XFS_SCRUB_TYPE_GQUOTA] = { XHG_FS, XFS_SICK_FS_GQUOTA },
[XFS_SCRUB_TYPE_PQUOTA] = { XHG_FS, XFS_SICK_FS_PQUOTA },
[XFS_SCRUB_TYPE_FSCOUNTERS] = { XHG_FS, XFS_SICK_FS_COUNTERS },
};
/* Return the health status mask for this scrub type. */
unsigned int
xchk_health_mask_for_scrub_type(
__u32 scrub_type)
{
return type_to_health_flag[scrub_type].sick_mask;
}
/*
* Update filesystem health assessments based on what we found and did.
*
* If the scrubber finds errors, we mark sick whatever's mentioned in
* sick_mask, no matter whether this is a first scan or an
* evaluation of repair effectiveness.
*
* Otherwise, no direct corruption was found, so mark whatever's in
* sick_mask as healthy.
*/
void
xchk_update_health(
struct xfs_scrub *sc)
{
struct xfs_perag *pag;
bool bad;
if (!sc->sick_mask)
return;
bad = (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT |
XFS_SCRUB_OFLAG_XCORRUPT));
switch (type_to_health_flag[sc->sm->sm_type].group) {
case XHG_AG:
pag = xfs_perag_get(sc->mp, sc->sm->sm_agno);
if (bad)
xfs_ag_mark_sick(pag, sc->sick_mask);
else
xfs_ag_mark_healthy(pag, sc->sick_mask);
xfs_perag_put(pag);
break;
case XHG_INO:
if (!sc->ip)
return;
if (bad)
xfs_inode_mark_sick(sc->ip, sc->sick_mask);
else
xfs_inode_mark_healthy(sc->ip, sc->sick_mask);
break;
case XHG_FS:
if (bad)
xfs_fs_mark_sick(sc->mp, sc->sick_mask);
else
xfs_fs_mark_healthy(sc->mp, sc->sick_mask);
break;
case XHG_RT:
if (bad)
xfs_rt_mark_sick(sc->mp, sc->sick_mask);
else
xfs_rt_mark_healthy(sc->mp, sc->sick_mask);
break;
default:
ASSERT(0);
break;
}
}
/* Is the given per-AG btree healthy enough for scanning? */
bool
xchk_ag_btree_healthy_enough(
struct xfs_scrub *sc,
struct xfs_perag *pag,
xfs_btnum_t btnum)
{
unsigned int mask = 0;
/*
* We always want the cursor if it's the same type as whatever we're
* scrubbing, even if we already know the structure is corrupt.
*
* Otherwise, we're only interested in the btree for cross-referencing.
* If we know the btree is bad then don't bother, just set XFAIL.
*/
switch (btnum) {
case XFS_BTNUM_BNO:
if (sc->sm->sm_type == XFS_SCRUB_TYPE_BNOBT)
return true;
mask = XFS_SICK_AG_BNOBT;
break;
case XFS_BTNUM_CNT:
if (sc->sm->sm_type == XFS_SCRUB_TYPE_CNTBT)
return true;
mask = XFS_SICK_AG_CNTBT;
break;
case XFS_BTNUM_INO:
if (sc->sm->sm_type == XFS_SCRUB_TYPE_INOBT)
return true;
mask = XFS_SICK_AG_INOBT;
break;
case XFS_BTNUM_FINO:
if (sc->sm->sm_type == XFS_SCRUB_TYPE_FINOBT)
return true;
mask = XFS_SICK_AG_FINOBT;
break;
case XFS_BTNUM_RMAP:
if (sc->sm->sm_type == XFS_SCRUB_TYPE_RMAPBT)
return true;
mask = XFS_SICK_AG_RMAPBT;
break;
case XFS_BTNUM_REFC:
if (sc->sm->sm_type == XFS_SCRUB_TYPE_REFCNTBT)
return true;
mask = XFS_SICK_AG_REFCNTBT;
break;
default:
ASSERT(0);
return true;
}
if (xfs_ag_has_sickness(pag, mask)) {
sc->sm->sm_flags |= XFS_SCRUB_OFLAG_XFAIL;
return false;
}
return true;
}
Reference in New Issue View Git Blame Copy Permalink