2018-06-06 09:42:14 +07:00
|
|
|
// SPDX-License-Identifier: GPL-2.0+
|
2016-10-03 23:11:20 +07:00
|
|
|
/*
|
|
|
|
* Copyright (C) 2016 Oracle. All Rights Reserved.
|
|
|
|
* Author: Darrick J. Wong <darrick.wong@oracle.com>
|
|
|
|
*/
|
|
|
|
#include "xfs.h"
|
|
|
|
#include "xfs_fs.h"
|
|
|
|
#include "xfs_format.h"
|
|
|
|
#include "xfs_log_format.h"
|
|
|
|
#include "xfs_trans_resv.h"
|
2016-10-03 23:11:21 +07:00
|
|
|
#include "xfs_bit.h"
|
2018-02-23 05:41:25 +07:00
|
|
|
#include "xfs_shared.h"
|
2016-10-03 23:11:20 +07:00
|
|
|
#include "xfs_mount.h"
|
2016-10-03 23:11:21 +07:00
|
|
|
#include "xfs_defer.h"
|
2016-10-03 23:11:20 +07:00
|
|
|
#include "xfs_trans.h"
|
|
|
|
#include "xfs_trans_priv.h"
|
|
|
|
#include "xfs_buf_item.h"
|
|
|
|
#include "xfs_refcount_item.h"
|
|
|
|
#include "xfs_log.h"
|
2016-10-03 23:11:21 +07:00
|
|
|
#include "xfs_refcount.h"
|
2016-10-03 23:11:20 +07:00
|
|
|
|
|
|
|
|
|
|
|
kmem_zone_t *xfs_cui_zone;
|
|
|
|
kmem_zone_t *xfs_cud_zone;
|
|
|
|
|
|
|
|
static inline struct xfs_cui_log_item *CUI_ITEM(struct xfs_log_item *lip)
|
|
|
|
{
|
|
|
|
return container_of(lip, struct xfs_cui_log_item, cui_item);
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
xfs_cui_item_free(
|
|
|
|
struct xfs_cui_log_item *cuip)
|
|
|
|
{
|
|
|
|
if (cuip->cui_format.cui_nextents > XFS_CUI_MAX_FAST_EXTENTS)
|
|
|
|
kmem_free(cuip);
|
|
|
|
else
|
|
|
|
kmem_zone_free(xfs_cui_zone, cuip);
|
|
|
|
}
|
|
|
|
|
2018-04-03 10:08:27 +07:00
|
|
|
/*
|
|
|
|
* Freeing the CUI requires that we remove it from the AIL if it has already
|
|
|
|
* been placed there. However, the CUI may not yet have been placed in the AIL
|
|
|
|
* when called by xfs_cui_release() from CUD processing due to the ordering of
|
|
|
|
* committed vs unpin operations in bulk insert operations. Hence the reference
|
|
|
|
* count to ensure only the last caller frees the CUI.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
xfs_cui_release(
|
|
|
|
struct xfs_cui_log_item *cuip)
|
|
|
|
{
|
|
|
|
ASSERT(atomic_read(&cuip->cui_refcount) > 0);
|
|
|
|
if (atomic_dec_and_test(&cuip->cui_refcount)) {
|
|
|
|
xfs_trans_ail_remove(&cuip->cui_item, SHUTDOWN_LOG_IO_ERROR);
|
|
|
|
xfs_cui_item_free(cuip);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2016-10-03 23:11:20 +07:00
|
|
|
STATIC void
|
|
|
|
xfs_cui_item_size(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
int *nvecs,
|
|
|
|
int *nbytes)
|
|
|
|
{
|
|
|
|
struct xfs_cui_log_item *cuip = CUI_ITEM(lip);
|
|
|
|
|
|
|
|
*nvecs += 1;
|
|
|
|
*nbytes += xfs_cui_log_format_sizeof(cuip->cui_format.cui_nextents);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is called to fill in the vector of log iovecs for the
|
|
|
|
* given cui log item. We use only 1 iovec, and we point that
|
|
|
|
* at the cui_log_format structure embedded in the cui item.
|
|
|
|
* It is at this point that we assert that all of the extent
|
|
|
|
* slots in the cui item have been filled.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cui_item_format(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
struct xfs_log_vec *lv)
|
|
|
|
{
|
|
|
|
struct xfs_cui_log_item *cuip = CUI_ITEM(lip);
|
|
|
|
struct xfs_log_iovec *vecp = NULL;
|
|
|
|
|
|
|
|
ASSERT(atomic_read(&cuip->cui_next_extent) ==
|
|
|
|
cuip->cui_format.cui_nextents);
|
|
|
|
|
|
|
|
cuip->cui_format.cui_type = XFS_LI_CUI;
|
|
|
|
cuip->cui_format.cui_size = 1;
|
|
|
|
|
|
|
|
xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_CUI_FORMAT, &cuip->cui_format,
|
|
|
|
xfs_cui_log_format_sizeof(cuip->cui_format.cui_nextents));
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Pinning has no meaning for an cui item, so just return.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cui_item_pin(
|
|
|
|
struct xfs_log_item *lip)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The unpin operation is the last place an CUI is manipulated in the log. It is
|
|
|
|
* either inserted in the AIL or aborted in the event of a log I/O error. In
|
|
|
|
* either case, the CUI transaction has been successfully committed to make it
|
|
|
|
* this far. Therefore, we expect whoever committed the CUI to either construct
|
|
|
|
* and commit the CUD or drop the CUD's reference in the event of error. Simply
|
|
|
|
* drop the log's CUI reference now that the log is done with it.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cui_item_unpin(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
int remove)
|
|
|
|
{
|
|
|
|
struct xfs_cui_log_item *cuip = CUI_ITEM(lip);
|
|
|
|
|
|
|
|
xfs_cui_release(cuip);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* CUI items have no locking or pushing. However, since CUIs are pulled from
|
|
|
|
* the AIL when their corresponding CUDs are committed to disk, their situation
|
|
|
|
* is very similar to being pinned. Return XFS_ITEM_PINNED so that the caller
|
|
|
|
* will eventually flush the log. This should help in getting the CUI out of
|
|
|
|
* the AIL.
|
|
|
|
*/
|
|
|
|
STATIC uint
|
|
|
|
xfs_cui_item_push(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
struct list_head *buffer_list)
|
|
|
|
{
|
|
|
|
return XFS_ITEM_PINNED;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The CUI has been either committed or aborted if the transaction has been
|
|
|
|
* cancelled. If the transaction was cancelled, an CUD isn't going to be
|
|
|
|
* constructed and thus we free the CUI here directly.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cui_item_unlock(
|
|
|
|
struct xfs_log_item *lip)
|
|
|
|
{
|
2018-05-09 21:47:34 +07:00
|
|
|
if (test_bit(XFS_LI_ABORTED, &lip->li_flags))
|
2018-04-03 10:08:27 +07:00
|
|
|
xfs_cui_release(CUI_ITEM(lip));
|
2016-10-03 23:11:20 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The CUI is logged only once and cannot be moved in the log, so simply return
|
|
|
|
* the lsn at which it's been logged.
|
|
|
|
*/
|
|
|
|
STATIC xfs_lsn_t
|
|
|
|
xfs_cui_item_committed(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
xfs_lsn_t lsn)
|
|
|
|
{
|
|
|
|
return lsn;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The CUI dependency tracking op doesn't do squat. It can't because
|
|
|
|
* it doesn't know where the free extent is coming from. The dependency
|
|
|
|
* tracking has to be handled by the "enclosing" metadata object. For
|
|
|
|
* example, for inodes, the inode is locked throughout the extent freeing
|
|
|
|
* so the dependency should be recorded there.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cui_item_committing(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
xfs_lsn_t lsn)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is the ops vector shared by all cui log items.
|
|
|
|
*/
|
|
|
|
static const struct xfs_item_ops xfs_cui_item_ops = {
|
|
|
|
.iop_size = xfs_cui_item_size,
|
|
|
|
.iop_format = xfs_cui_item_format,
|
|
|
|
.iop_pin = xfs_cui_item_pin,
|
|
|
|
.iop_unpin = xfs_cui_item_unpin,
|
|
|
|
.iop_unlock = xfs_cui_item_unlock,
|
|
|
|
.iop_committed = xfs_cui_item_committed,
|
|
|
|
.iop_push = xfs_cui_item_push,
|
|
|
|
.iop_committing = xfs_cui_item_committing,
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Allocate and initialize an cui item with the given number of extents.
|
|
|
|
*/
|
|
|
|
struct xfs_cui_log_item *
|
|
|
|
xfs_cui_init(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
uint nextents)
|
|
|
|
|
|
|
|
{
|
|
|
|
struct xfs_cui_log_item *cuip;
|
|
|
|
|
|
|
|
ASSERT(nextents > 0);
|
|
|
|
if (nextents > XFS_CUI_MAX_FAST_EXTENTS)
|
|
|
|
cuip = kmem_zalloc(xfs_cui_log_item_sizeof(nextents),
|
|
|
|
KM_SLEEP);
|
|
|
|
else
|
|
|
|
cuip = kmem_zone_zalloc(xfs_cui_zone, KM_SLEEP);
|
|
|
|
|
|
|
|
xfs_log_item_init(mp, &cuip->cui_item, XFS_LI_CUI, &xfs_cui_item_ops);
|
|
|
|
cuip->cui_format.cui_nextents = nextents;
|
|
|
|
cuip->cui_format.cui_id = (uintptr_t)(void *)cuip;
|
|
|
|
atomic_set(&cuip->cui_next_extent, 0);
|
|
|
|
atomic_set(&cuip->cui_refcount, 2);
|
|
|
|
|
|
|
|
return cuip;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct xfs_cud_log_item *CUD_ITEM(struct xfs_log_item *lip)
|
|
|
|
{
|
|
|
|
return container_of(lip, struct xfs_cud_log_item, cud_item);
|
|
|
|
}
|
|
|
|
|
|
|
|
STATIC void
|
|
|
|
xfs_cud_item_size(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
int *nvecs,
|
|
|
|
int *nbytes)
|
|
|
|
{
|
|
|
|
*nvecs += 1;
|
|
|
|
*nbytes += sizeof(struct xfs_cud_log_format);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is called to fill in the vector of log iovecs for the
|
|
|
|
* given cud log item. We use only 1 iovec, and we point that
|
|
|
|
* at the cud_log_format structure embedded in the cud item.
|
|
|
|
* It is at this point that we assert that all of the extent
|
|
|
|
* slots in the cud item have been filled.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cud_item_format(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
struct xfs_log_vec *lv)
|
|
|
|
{
|
|
|
|
struct xfs_cud_log_item *cudp = CUD_ITEM(lip);
|
|
|
|
struct xfs_log_iovec *vecp = NULL;
|
|
|
|
|
|
|
|
cudp->cud_format.cud_type = XFS_LI_CUD;
|
|
|
|
cudp->cud_format.cud_size = 1;
|
|
|
|
|
|
|
|
xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_CUD_FORMAT, &cudp->cud_format,
|
|
|
|
sizeof(struct xfs_cud_log_format));
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Pinning has no meaning for an cud item, so just return.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cud_item_pin(
|
|
|
|
struct xfs_log_item *lip)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Since pinning has no meaning for an cud item, unpinning does
|
|
|
|
* not either.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cud_item_unpin(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
int remove)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* There isn't much you can do to push on an cud item. It is simply stuck
|
|
|
|
* waiting for the log to be flushed to disk.
|
|
|
|
*/
|
|
|
|
STATIC uint
|
|
|
|
xfs_cud_item_push(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
struct list_head *buffer_list)
|
|
|
|
{
|
|
|
|
return XFS_ITEM_PINNED;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The CUD is either committed or aborted if the transaction is cancelled. If
|
|
|
|
* the transaction is cancelled, drop our reference to the CUI and free the
|
|
|
|
* CUD.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cud_item_unlock(
|
|
|
|
struct xfs_log_item *lip)
|
|
|
|
{
|
|
|
|
struct xfs_cud_log_item *cudp = CUD_ITEM(lip);
|
|
|
|
|
2018-05-09 21:47:34 +07:00
|
|
|
if (test_bit(XFS_LI_ABORTED, &lip->li_flags)) {
|
2016-10-03 23:11:20 +07:00
|
|
|
xfs_cui_release(cudp->cud_cuip);
|
|
|
|
kmem_zone_free(xfs_cud_zone, cudp);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When the cud item is committed to disk, all we need to do is delete our
|
|
|
|
* reference to our partner cui item and then free ourselves. Since we're
|
|
|
|
* freeing ourselves we must return -1 to keep the transaction code from
|
|
|
|
* further referencing this item.
|
|
|
|
*/
|
|
|
|
STATIC xfs_lsn_t
|
|
|
|
xfs_cud_item_committed(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
xfs_lsn_t lsn)
|
|
|
|
{
|
|
|
|
struct xfs_cud_log_item *cudp = CUD_ITEM(lip);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Drop the CUI reference regardless of whether the CUD has been
|
|
|
|
* aborted. Once the CUD transaction is constructed, it is the sole
|
|
|
|
* responsibility of the CUD to release the CUI (even if the CUI is
|
|
|
|
* aborted due to log I/O error).
|
|
|
|
*/
|
|
|
|
xfs_cui_release(cudp->cud_cuip);
|
|
|
|
kmem_zone_free(xfs_cud_zone, cudp);
|
|
|
|
|
|
|
|
return (xfs_lsn_t)-1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The CUD dependency tracking op doesn't do squat. It can't because
|
|
|
|
* it doesn't know where the free extent is coming from. The dependency
|
|
|
|
* tracking has to be handled by the "enclosing" metadata object. For
|
|
|
|
* example, for inodes, the inode is locked throughout the extent freeing
|
|
|
|
* so the dependency should be recorded there.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_cud_item_committing(
|
|
|
|
struct xfs_log_item *lip,
|
|
|
|
xfs_lsn_t lsn)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is the ops vector shared by all cud log items.
|
|
|
|
*/
|
|
|
|
static const struct xfs_item_ops xfs_cud_item_ops = {
|
|
|
|
.iop_size = xfs_cud_item_size,
|
|
|
|
.iop_format = xfs_cud_item_format,
|
|
|
|
.iop_pin = xfs_cud_item_pin,
|
|
|
|
.iop_unpin = xfs_cud_item_unpin,
|
|
|
|
.iop_unlock = xfs_cud_item_unlock,
|
|
|
|
.iop_committed = xfs_cud_item_committed,
|
|
|
|
.iop_push = xfs_cud_item_push,
|
|
|
|
.iop_committing = xfs_cud_item_committing,
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Allocate and initialize an cud item with the given number of extents.
|
|
|
|
*/
|
|
|
|
struct xfs_cud_log_item *
|
|
|
|
xfs_cud_init(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
struct xfs_cui_log_item *cuip)
|
|
|
|
|
|
|
|
{
|
|
|
|
struct xfs_cud_log_item *cudp;
|
|
|
|
|
|
|
|
cudp = kmem_zone_zalloc(xfs_cud_zone, KM_SLEEP);
|
|
|
|
xfs_log_item_init(mp, &cudp->cud_item, XFS_LI_CUD, &xfs_cud_item_ops);
|
|
|
|
cudp->cud_cuip = cuip;
|
|
|
|
cudp->cud_format.cud_cui_id = cuip->cui_format.cui_id;
|
|
|
|
|
|
|
|
return cudp;
|
|
|
|
}
|
2016-10-03 23:11:21 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Process a refcount update intent item that was recovered from the log.
|
|
|
|
* We need to update the refcountbt.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
xfs_cui_recover(
|
|
|
|
struct xfs_mount *mp,
|
xfs: log recovery should replay deferred ops in order
As part of testing log recovery with dm_log_writes, Amir Goldstein
discovered an error in the deferred ops recovery that lead to corruption
of the filesystem metadata if a reflink+rmap filesystem happened to shut
down midway through a CoW remap:
"This is what happens [after failed log recovery]:
"Phase 1 - find and verify superblock...
"Phase 2 - using internal log
" - zero log...
" - scan filesystem freespace and inode maps...
" - found root inode chunk
"Phase 3 - for each AG...
" - scan (but don't clear) agi unlinked lists...
" - process known inodes and perform inode discovery...
" - agno = 0
"data fork in regular inode 134 claims CoW block 376
"correcting nextents for inode 134
"bad data fork in inode 134
"would have cleared inode 134"
Hou Tao dissected the log contents of exactly such a crash:
"According to the implementation of xfs_defer_finish(), these ops should
be completed in the following sequence:
"Have been done:
"(1) CUI: Oper (160)
"(2) BUI: Oper (161)
"(3) CUD: Oper (194), for CUI Oper (160)
"(4) RUI A: Oper (197), free rmap [0x155, 2, -9]
"Should be done:
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI A
"(8) RUD: for RUI B
"Actually be done by xlog_recover_process_intents()
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI B
"(8) RUD: for RUI A
"So the rmap entry [0x155, 2, -9] for COW should be freed firstly,
then a new rmap entry [0x155, 2, 137] will be added. However, as we can see
from the log record in post_mount.log (generated after umount) and the trace
print, the new rmap entry [0x155, 2, 137] are added firstly, then the rmap
entry [0x155, 2, -9] are freed."
When reconstructing the internal log state from the log items found on
disk, it's required that deferred ops replay in exactly the same order
that they would have had the filesystem not gone down. However,
replaying unfinished deferred ops can create /more/ deferred ops. These
new deferred ops are finished in the wrong order. This causes fs
corruption and replay crashes, so let's create a single defer_ops to
handle the subsequent ops created during replay, then use one single
transaction at the end of log recovery to ensure that everything is
replayed in the same order as they're supposed to be.
Reported-by: Amir Goldstein <amir73il@gmail.com>
Analyzed-by: Hou Tao <houtao1@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-22 11:53:02 +07:00
|
|
|
struct xfs_cui_log_item *cuip,
|
|
|
|
struct xfs_defer_ops *dfops)
|
2016-10-03 23:11:21 +07:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
int error = 0;
|
2016-10-03 23:11:22 +07:00
|
|
|
unsigned int refc_type;
|
2016-10-03 23:11:21 +07:00
|
|
|
struct xfs_phys_extent *refc;
|
|
|
|
xfs_fsblock_t startblock_fsb;
|
|
|
|
bool op_ok;
|
2016-10-03 23:11:22 +07:00
|
|
|
struct xfs_cud_log_item *cudp;
|
|
|
|
struct xfs_trans *tp;
|
|
|
|
struct xfs_btree_cur *rcur = NULL;
|
|
|
|
enum xfs_refcount_intent_type type;
|
|
|
|
xfs_fsblock_t new_fsb;
|
|
|
|
xfs_extlen_t new_len;
|
|
|
|
struct xfs_bmbt_irec irec;
|
|
|
|
bool requeue_only = false;
|
2016-10-03 23:11:21 +07:00
|
|
|
|
|
|
|
ASSERT(!test_bit(XFS_CUI_RECOVERED, &cuip->cui_flags));
|
|
|
|
|
|
|
|
/*
|
|
|
|
* First check the validity of the extents described by the
|
|
|
|
* CUI. If any are bad, then assume that all are bad and
|
|
|
|
* just toss the CUI.
|
|
|
|
*/
|
|
|
|
for (i = 0; i < cuip->cui_format.cui_nextents; i++) {
|
|
|
|
refc = &cuip->cui_format.cui_extents[i];
|
|
|
|
startblock_fsb = XFS_BB_TO_FSB(mp,
|
|
|
|
XFS_FSB_TO_DADDR(mp, refc->pe_startblock));
|
|
|
|
switch (refc->pe_flags & XFS_REFCOUNT_EXTENT_TYPE_MASK) {
|
|
|
|
case XFS_REFCOUNT_INCREASE:
|
|
|
|
case XFS_REFCOUNT_DECREASE:
|
|
|
|
case XFS_REFCOUNT_ALLOC_COW:
|
|
|
|
case XFS_REFCOUNT_FREE_COW:
|
|
|
|
op_ok = true;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
op_ok = false;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (!op_ok || startblock_fsb == 0 ||
|
|
|
|
refc->pe_len == 0 ||
|
|
|
|
startblock_fsb >= mp->m_sb.sb_dblocks ||
|
|
|
|
refc->pe_len >= mp->m_sb.sb_agblocks ||
|
|
|
|
(refc->pe_flags & ~XFS_REFCOUNT_EXTENT_FLAGS)) {
|
|
|
|
/*
|
|
|
|
* This will pull the CUI from the AIL and
|
|
|
|
* free the memory associated with it.
|
|
|
|
*/
|
|
|
|
set_bit(XFS_CUI_RECOVERED, &cuip->cui_flags);
|
|
|
|
xfs_cui_release(cuip);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-10-03 23:11:22 +07:00
|
|
|
/*
|
|
|
|
* Under normal operation, refcount updates are deferred, so we
|
|
|
|
* wouldn't be adding them directly to a transaction. All
|
|
|
|
* refcount updates manage reservation usage internally and
|
|
|
|
* dynamically by deferring work that won't fit in the
|
|
|
|
* transaction. Normally, any work that needs to be deferred
|
|
|
|
* gets attached to the same defer_ops that scheduled the
|
|
|
|
* refcount update. However, we're in log recovery here, so we
|
2018-02-23 05:41:25 +07:00
|
|
|
* we use the passed in defer_ops and to finish up any work that
|
|
|
|
* doesn't fit. We need to reserve enough blocks to handle a
|
|
|
|
* full btree split on either end of the refcount range.
|
2016-10-03 23:11:22 +07:00
|
|
|
*/
|
2018-02-23 05:41:25 +07:00
|
|
|
error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate,
|
|
|
|
mp->m_refc_maxlevels * 2, 0, XFS_TRANS_RESERVE, &tp);
|
2016-10-03 23:11:22 +07:00
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
cudp = xfs_trans_get_cud(tp, cuip);
|
|
|
|
|
|
|
|
for (i = 0; i < cuip->cui_format.cui_nextents; i++) {
|
|
|
|
refc = &cuip->cui_format.cui_extents[i];
|
|
|
|
refc_type = refc->pe_flags & XFS_REFCOUNT_EXTENT_TYPE_MASK;
|
|
|
|
switch (refc_type) {
|
|
|
|
case XFS_REFCOUNT_INCREASE:
|
|
|
|
case XFS_REFCOUNT_DECREASE:
|
|
|
|
case XFS_REFCOUNT_ALLOC_COW:
|
|
|
|
case XFS_REFCOUNT_FREE_COW:
|
|
|
|
type = refc_type;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
error = -EFSCORRUPTED;
|
|
|
|
goto abort_error;
|
|
|
|
}
|
|
|
|
if (requeue_only) {
|
|
|
|
new_fsb = refc->pe_startblock;
|
|
|
|
new_len = refc->pe_len;
|
|
|
|
} else
|
|
|
|
error = xfs_trans_log_finish_refcount_update(tp, cudp,
|
xfs: log recovery should replay deferred ops in order
As part of testing log recovery with dm_log_writes, Amir Goldstein
discovered an error in the deferred ops recovery that lead to corruption
of the filesystem metadata if a reflink+rmap filesystem happened to shut
down midway through a CoW remap:
"This is what happens [after failed log recovery]:
"Phase 1 - find and verify superblock...
"Phase 2 - using internal log
" - zero log...
" - scan filesystem freespace and inode maps...
" - found root inode chunk
"Phase 3 - for each AG...
" - scan (but don't clear) agi unlinked lists...
" - process known inodes and perform inode discovery...
" - agno = 0
"data fork in regular inode 134 claims CoW block 376
"correcting nextents for inode 134
"bad data fork in inode 134
"would have cleared inode 134"
Hou Tao dissected the log contents of exactly such a crash:
"According to the implementation of xfs_defer_finish(), these ops should
be completed in the following sequence:
"Have been done:
"(1) CUI: Oper (160)
"(2) BUI: Oper (161)
"(3) CUD: Oper (194), for CUI Oper (160)
"(4) RUI A: Oper (197), free rmap [0x155, 2, -9]
"Should be done:
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI A
"(8) RUD: for RUI B
"Actually be done by xlog_recover_process_intents()
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI B
"(8) RUD: for RUI A
"So the rmap entry [0x155, 2, -9] for COW should be freed firstly,
then a new rmap entry [0x155, 2, 137] will be added. However, as we can see
from the log record in post_mount.log (generated after umount) and the trace
print, the new rmap entry [0x155, 2, 137] are added firstly, then the rmap
entry [0x155, 2, -9] are freed."
When reconstructing the internal log state from the log items found on
disk, it's required that deferred ops replay in exactly the same order
that they would have had the filesystem not gone down. However,
replaying unfinished deferred ops can create /more/ deferred ops. These
new deferred ops are finished in the wrong order. This causes fs
corruption and replay crashes, so let's create a single defer_ops to
handle the subsequent ops created during replay, then use one single
transaction at the end of log recovery to ensure that everything is
replayed in the same order as they're supposed to be.
Reported-by: Amir Goldstein <amir73il@gmail.com>
Analyzed-by: Hou Tao <houtao1@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-22 11:53:02 +07:00
|
|
|
dfops, type, refc->pe_startblock, refc->pe_len,
|
2016-10-03 23:11:22 +07:00
|
|
|
&new_fsb, &new_len, &rcur);
|
|
|
|
if (error)
|
|
|
|
goto abort_error;
|
|
|
|
|
|
|
|
/* Requeue what we didn't finish. */
|
|
|
|
if (new_len > 0) {
|
|
|
|
irec.br_startblock = new_fsb;
|
|
|
|
irec.br_blockcount = new_len;
|
|
|
|
switch (type) {
|
|
|
|
case XFS_REFCOUNT_INCREASE:
|
|
|
|
error = xfs_refcount_increase_extent(
|
xfs: log recovery should replay deferred ops in order
As part of testing log recovery with dm_log_writes, Amir Goldstein
discovered an error in the deferred ops recovery that lead to corruption
of the filesystem metadata if a reflink+rmap filesystem happened to shut
down midway through a CoW remap:
"This is what happens [after failed log recovery]:
"Phase 1 - find and verify superblock...
"Phase 2 - using internal log
" - zero log...
" - scan filesystem freespace and inode maps...
" - found root inode chunk
"Phase 3 - for each AG...
" - scan (but don't clear) agi unlinked lists...
" - process known inodes and perform inode discovery...
" - agno = 0
"data fork in regular inode 134 claims CoW block 376
"correcting nextents for inode 134
"bad data fork in inode 134
"would have cleared inode 134"
Hou Tao dissected the log contents of exactly such a crash:
"According to the implementation of xfs_defer_finish(), these ops should
be completed in the following sequence:
"Have been done:
"(1) CUI: Oper (160)
"(2) BUI: Oper (161)
"(3) CUD: Oper (194), for CUI Oper (160)
"(4) RUI A: Oper (197), free rmap [0x155, 2, -9]
"Should be done:
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI A
"(8) RUD: for RUI B
"Actually be done by xlog_recover_process_intents()
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI B
"(8) RUD: for RUI A
"So the rmap entry [0x155, 2, -9] for COW should be freed firstly,
then a new rmap entry [0x155, 2, 137] will be added. However, as we can see
from the log record in post_mount.log (generated after umount) and the trace
print, the new rmap entry [0x155, 2, 137] are added firstly, then the rmap
entry [0x155, 2, -9] are freed."
When reconstructing the internal log state from the log items found on
disk, it's required that deferred ops replay in exactly the same order
that they would have had the filesystem not gone down. However,
replaying unfinished deferred ops can create /more/ deferred ops. These
new deferred ops are finished in the wrong order. This causes fs
corruption and replay crashes, so let's create a single defer_ops to
handle the subsequent ops created during replay, then use one single
transaction at the end of log recovery to ensure that everything is
replayed in the same order as they're supposed to be.
Reported-by: Amir Goldstein <amir73il@gmail.com>
Analyzed-by: Hou Tao <houtao1@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-22 11:53:02 +07:00
|
|
|
tp->t_mountp, dfops, &irec);
|
2016-10-03 23:11:22 +07:00
|
|
|
break;
|
|
|
|
case XFS_REFCOUNT_DECREASE:
|
|
|
|
error = xfs_refcount_decrease_extent(
|
xfs: log recovery should replay deferred ops in order
As part of testing log recovery with dm_log_writes, Amir Goldstein
discovered an error in the deferred ops recovery that lead to corruption
of the filesystem metadata if a reflink+rmap filesystem happened to shut
down midway through a CoW remap:
"This is what happens [after failed log recovery]:
"Phase 1 - find and verify superblock...
"Phase 2 - using internal log
" - zero log...
" - scan filesystem freespace and inode maps...
" - found root inode chunk
"Phase 3 - for each AG...
" - scan (but don't clear) agi unlinked lists...
" - process known inodes and perform inode discovery...
" - agno = 0
"data fork in regular inode 134 claims CoW block 376
"correcting nextents for inode 134
"bad data fork in inode 134
"would have cleared inode 134"
Hou Tao dissected the log contents of exactly such a crash:
"According to the implementation of xfs_defer_finish(), these ops should
be completed in the following sequence:
"Have been done:
"(1) CUI: Oper (160)
"(2) BUI: Oper (161)
"(3) CUD: Oper (194), for CUI Oper (160)
"(4) RUI A: Oper (197), free rmap [0x155, 2, -9]
"Should be done:
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI A
"(8) RUD: for RUI B
"Actually be done by xlog_recover_process_intents()
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI B
"(8) RUD: for RUI A
"So the rmap entry [0x155, 2, -9] for COW should be freed firstly,
then a new rmap entry [0x155, 2, 137] will be added. However, as we can see
from the log record in post_mount.log (generated after umount) and the trace
print, the new rmap entry [0x155, 2, 137] are added firstly, then the rmap
entry [0x155, 2, -9] are freed."
When reconstructing the internal log state from the log items found on
disk, it's required that deferred ops replay in exactly the same order
that they would have had the filesystem not gone down. However,
replaying unfinished deferred ops can create /more/ deferred ops. These
new deferred ops are finished in the wrong order. This causes fs
corruption and replay crashes, so let's create a single defer_ops to
handle the subsequent ops created during replay, then use one single
transaction at the end of log recovery to ensure that everything is
replayed in the same order as they're supposed to be.
Reported-by: Amir Goldstein <amir73il@gmail.com>
Analyzed-by: Hou Tao <houtao1@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-22 11:53:02 +07:00
|
|
|
tp->t_mountp, dfops, &irec);
|
2016-10-03 23:11:22 +07:00
|
|
|
break;
|
2016-10-03 23:11:39 +07:00
|
|
|
case XFS_REFCOUNT_ALLOC_COW:
|
|
|
|
error = xfs_refcount_alloc_cow_extent(
|
xfs: log recovery should replay deferred ops in order
As part of testing log recovery with dm_log_writes, Amir Goldstein
discovered an error in the deferred ops recovery that lead to corruption
of the filesystem metadata if a reflink+rmap filesystem happened to shut
down midway through a CoW remap:
"This is what happens [after failed log recovery]:
"Phase 1 - find and verify superblock...
"Phase 2 - using internal log
" - zero log...
" - scan filesystem freespace and inode maps...
" - found root inode chunk
"Phase 3 - for each AG...
" - scan (but don't clear) agi unlinked lists...
" - process known inodes and perform inode discovery...
" - agno = 0
"data fork in regular inode 134 claims CoW block 376
"correcting nextents for inode 134
"bad data fork in inode 134
"would have cleared inode 134"
Hou Tao dissected the log contents of exactly such a crash:
"According to the implementation of xfs_defer_finish(), these ops should
be completed in the following sequence:
"Have been done:
"(1) CUI: Oper (160)
"(2) BUI: Oper (161)
"(3) CUD: Oper (194), for CUI Oper (160)
"(4) RUI A: Oper (197), free rmap [0x155, 2, -9]
"Should be done:
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI A
"(8) RUD: for RUI B
"Actually be done by xlog_recover_process_intents()
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI B
"(8) RUD: for RUI A
"So the rmap entry [0x155, 2, -9] for COW should be freed firstly,
then a new rmap entry [0x155, 2, 137] will be added. However, as we can see
from the log record in post_mount.log (generated after umount) and the trace
print, the new rmap entry [0x155, 2, 137] are added firstly, then the rmap
entry [0x155, 2, -9] are freed."
When reconstructing the internal log state from the log items found on
disk, it's required that deferred ops replay in exactly the same order
that they would have had the filesystem not gone down. However,
replaying unfinished deferred ops can create /more/ deferred ops. These
new deferred ops are finished in the wrong order. This causes fs
corruption and replay crashes, so let's create a single defer_ops to
handle the subsequent ops created during replay, then use one single
transaction at the end of log recovery to ensure that everything is
replayed in the same order as they're supposed to be.
Reported-by: Amir Goldstein <amir73il@gmail.com>
Analyzed-by: Hou Tao <houtao1@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-22 11:53:02 +07:00
|
|
|
tp->t_mountp, dfops,
|
2016-10-03 23:11:39 +07:00
|
|
|
irec.br_startblock,
|
|
|
|
irec.br_blockcount);
|
|
|
|
break;
|
|
|
|
case XFS_REFCOUNT_FREE_COW:
|
|
|
|
error = xfs_refcount_free_cow_extent(
|
xfs: log recovery should replay deferred ops in order
As part of testing log recovery with dm_log_writes, Amir Goldstein
discovered an error in the deferred ops recovery that lead to corruption
of the filesystem metadata if a reflink+rmap filesystem happened to shut
down midway through a CoW remap:
"This is what happens [after failed log recovery]:
"Phase 1 - find and verify superblock...
"Phase 2 - using internal log
" - zero log...
" - scan filesystem freespace and inode maps...
" - found root inode chunk
"Phase 3 - for each AG...
" - scan (but don't clear) agi unlinked lists...
" - process known inodes and perform inode discovery...
" - agno = 0
"data fork in regular inode 134 claims CoW block 376
"correcting nextents for inode 134
"bad data fork in inode 134
"would have cleared inode 134"
Hou Tao dissected the log contents of exactly such a crash:
"According to the implementation of xfs_defer_finish(), these ops should
be completed in the following sequence:
"Have been done:
"(1) CUI: Oper (160)
"(2) BUI: Oper (161)
"(3) CUD: Oper (194), for CUI Oper (160)
"(4) RUI A: Oper (197), free rmap [0x155, 2, -9]
"Should be done:
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI A
"(8) RUD: for RUI B
"Actually be done by xlog_recover_process_intents()
"(5) BUD: for BUI Oper (161)
"(6) RUI B: add rmap [0x155, 2, 137]
"(7) RUD: for RUI B
"(8) RUD: for RUI A
"So the rmap entry [0x155, 2, -9] for COW should be freed firstly,
then a new rmap entry [0x155, 2, 137] will be added. However, as we can see
from the log record in post_mount.log (generated after umount) and the trace
print, the new rmap entry [0x155, 2, 137] are added firstly, then the rmap
entry [0x155, 2, -9] are freed."
When reconstructing the internal log state from the log items found on
disk, it's required that deferred ops replay in exactly the same order
that they would have had the filesystem not gone down. However,
replaying unfinished deferred ops can create /more/ deferred ops. These
new deferred ops are finished in the wrong order. This causes fs
corruption and replay crashes, so let's create a single defer_ops to
handle the subsequent ops created during replay, then use one single
transaction at the end of log recovery to ensure that everything is
replayed in the same order as they're supposed to be.
Reported-by: Amir Goldstein <amir73il@gmail.com>
Analyzed-by: Hou Tao <houtao1@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-22 11:53:02 +07:00
|
|
|
tp->t_mountp, dfops,
|
2016-10-03 23:11:39 +07:00
|
|
|
irec.br_startblock,
|
|
|
|
irec.br_blockcount);
|
|
|
|
break;
|
2016-10-03 23:11:22 +07:00
|
|
|
default:
|
|
|
|
ASSERT(0);
|
|
|
|
}
|
|
|
|
if (error)
|
|
|
|
goto abort_error;
|
|
|
|
requeue_only = true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
xfs_refcount_finish_one_cleanup(tp, rcur, error);
|
2016-10-03 23:11:21 +07:00
|
|
|
set_bit(XFS_CUI_RECOVERED, &cuip->cui_flags);
|
2016-10-03 23:11:22 +07:00
|
|
|
error = xfs_trans_commit(tp);
|
|
|
|
return error;
|
|
|
|
|
|
|
|
abort_error:
|
|
|
|
xfs_refcount_finish_one_cleanup(tp, rcur, error);
|
|
|
|
xfs_trans_cancel(tp);
|
2016-10-03 23:11:21 +07:00
|
|
|
return error;
|
|
|
|
}
|