linux-stable.git/fs/ext4, branch v5.8.8

ext4: limit the length of per-inode prealloc list

2020-09-03T09:29:37+00:00

[ Upstream commit 27bc446e2def38db3244a6eb4bb1d6312936610a ]

In the scenario of writing sparse files, the per-inode prealloc list may
be very long, resulting in high overhead for ext4_mb_use_preallocated().
To circumvent this problem, we limit the maximum length of per-inode
prealloc list to 512 and allow users to modify it.

After patching, we observed that the sys ratio of cpu has dropped, and
the system throughput has increased significantly. We created a process
to write the sparse file, and the running time of the process on the
fixed kernel was significantly reduced, as follows:

Running time on unfixed kernel：
[root@TENCENT64 ~]# time taskset 0x01 ./sparse /data1/sparce.dat
real    0m2.051s
user    0m0.008s
sys     0m2.026s

Running time on fixed kernel：
[root@TENCENT64 ~]# time taskset 0x01 ./sparse /data1/sparce.dat
real    0m0.471s
user    0m0.004s
sys     0m0.395s

Signed-off-by: Chunguang Xu 
Link: https://lore.kernel.org/r/d7a98178-056b-6db5-6bce-4ead23f4a257@gmail.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

fs: prevent BUG_ON in submit_bh_wbc()

2020-09-03T09:29:31+00:00

[ Upstream commit 377254b2cd2252c7c3151b113cbdf93a7736c2e9 ]

If a device is hot-removed --- for example, when a physical device is
unplugged from pcie slot or a nbd device's network is shutdown ---
this can result in a BUG_ON() crash in submit_bh_wbc().  This is
because the when the block device dies, the buffer heads will have
their Buffer_Mapped flag get cleared, leading to the crash in
submit_bh_wbc.

We had attempted to work around this problem in commit a17712c8
("ext4: check superblock mapped prior to committing").  Unfortunately,
it's still possible to hit the BUG_ON(!buffer_mapped(bh)) if the
device dies between when the work-around check in ext4_commit_super()
and when submit_bh_wbh() is finally called:

Code path:
ext4_commit_super
    judge if 'buffer_mapped(sbh)' is false, return <== commit a17712c8
          lock_buffer(sbh)
          ...
          unlock_buffer(sbh)
               __sync_dirty_buffer(sbh,...
                    lock_buffer(sbh)
                        judge if 'buffer_mapped(sbh))' is false, return <== added by this patch
                            submit_bh(...,sbh)
                                submit_bh_wbc(...,sbh,...)

[100722.966497] kernel BUG at fs/buffer.c:3095! <== BUG_ON(!buffer_mapped(bh))' in submit_bh_wbc()
[100722.966503] invalid opcode: 0000 [#1] SMP
[100722.966566] task: ffff8817e15a9e40 task.stack: ffffc90024744000
[100722.966574] RIP: 0010:submit_bh_wbc+0x180/0x190
[100722.966575] RSP: 0018:ffffc90024747a90 EFLAGS: 00010246
[100722.966576] RAX: 0000000000620005 RBX: ffff8818a80603a8 RCX: 0000000000000000
[100722.966576] RDX: ffff8818a80603a8 RSI: 0000000000020800 RDI: 0000000000000001
[100722.966577] RBP: ffffc90024747ac0 R08: 0000000000000000 R09: ffff88207f94170d
[100722.966578] R10: 00000000000437c8 R11: 0000000000000001 R12: 0000000000020800
[100722.966578] R13: 0000000000000001 R14: 000000000bf9a438 R15: ffff88195f333000
[100722.966580] FS:  00007fa2eee27700(0000) GS:ffff88203d840000(0000) knlGS:0000000000000000
[100722.966580] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[100722.966581] CR2: 0000000000f0b008 CR3: 000000201a622003 CR4: 00000000007606e0
[100722.966582] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[100722.966583] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[100722.966583] PKRU: 55555554
[100722.966583] Call Trace:
[100722.966588]  __sync_dirty_buffer+0x6e/0xd0
[100722.966614]  ext4_commit_super+0x1d8/0x290 [ext4]
[100722.966626]  __ext4_std_error+0x78/0x100 [ext4]
[100722.966635]  ? __ext4_journal_get_write_access+0xca/0x120 [ext4]
[100722.966646]  ext4_reserve_inode_write+0x58/0xb0 [ext4]
[100722.966655]  ? ext4_dirty_inode+0x48/0x70 [ext4]
[100722.966663]  ext4_mark_inode_dirty+0x53/0x1e0 [ext4]
[100722.966671]  ? __ext4_journal_start_sb+0x6d/0xf0 [ext4]
[100722.966679]  ext4_dirty_inode+0x48/0x70 [ext4]
[100722.966682]  __mark_inode_dirty+0x17f/0x350
[100722.966686]  generic_update_time+0x87/0xd0
[100722.966687]  touch_atime+0xa9/0xd0
[100722.966690]  generic_file_read_iter+0xa09/0xcd0
[100722.966694]  ? page_cache_tree_insert+0xb0/0xb0
[100722.966704]  ext4_file_read_iter+0x4a/0x100 [ext4]
[100722.966707]  ? __inode_security_revalidate+0x4f/0x60
[100722.966709]  __vfs_read+0xec/0x160
[100722.966711]  vfs_read+0x8c/0x130
[100722.966712]  SyS_pread64+0x87/0xb0
[100722.966716]  do_syscall_64+0x67/0x1b0
[100722.966719]  entry_SYSCALL64_slow_path+0x25/0x25

To address this, add the check of 'buffer_mapped(bh)' to
__sync_dirty_buffer().  This also has the benefit of fixing this for
other file systems.

With this addition, we can drop the workaround in ext4_commit_supper().

[ Commit description rewritten by tytso. ]

Signed-off-by: Xianting Tian 
Link: https://lore.kernel.org/r/1596211825-8750-1-git-send-email-xianting_tian@126.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

ext4: correctly restore system zone info when remount fails

2020-09-03T09:29:31+00:00

[ Upstream commit 0f5bde1db174f6c471f0bd27198575719dabe3e5 ]

When remounting filesystem fails late during remount handling and
block_validity mount option is also changed during the remount, we fail
to restore system zone information to a state matching the mount option.
This is mostly harmless, just the block validity checking will not match
the situation described by the mount option. Make sure these two are always
consistent.

Reported-by: Lukas Czerner 
Reviewed-by: Lukas Czerner 
Signed-off-by: Jan Kara 
Link: https://lore.kernel.org/r/20200728130437.7804-7-jack@suse.cz
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

ext4: handle error of ext4_setup_system_zone() on remount

2020-09-03T09:29:31+00:00

[ Upstream commit d176b1f62f242ab259ff665a26fbac69db1aecba ]

ext4_setup_system_zone() can fail. Handle the failure in ext4_remount().

Reviewed-by: Lukas Czerner 
Signed-off-by: Jan Kara 
Link: https://lore.kernel.org/r/20200728130437.7804-2-jack@suse.cz
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

ext4: handle option set by mount flags correctly

2020-09-03T09:29:30+00:00

[ Upstream commit f25391ebb475d3ffb3aa61bb90e3594c841749ef ]

Currently there is a problem with mount options that can be both set by
vfs using mount flags or by a string parsing in ext4.

i_version/iversion options gets lost after remount, for example

$ mount -o i_version /dev/pmem0 /mnt
$ grep pmem0 /proc/self/mountinfo | grep i_version
310 95 259:0 / /mnt rw,relatime shared:163 - ext4 /dev/pmem0 rw,seclabel,i_version
$ mount -o remount,ro /mnt
$ grep pmem0 /proc/self/mountinfo | grep i_version

nolazytime gets ignored by ext4 on remount, for example

$ mount -o lazytime /dev/pmem0 /mnt
$ grep pmem0 /proc/self/mountinfo | grep lazytime
310 95 259:0 / /mnt rw,relatime shared:163 - ext4 /dev/pmem0 rw,lazytime,seclabel
$ mount -o remount,nolazytime /mnt
$ grep pmem0 /proc/self/mountinfo | grep lazytime
310 95 259:0 / /mnt rw,relatime shared:163 - ext4 /dev/pmem0 rw,lazytime,seclabel

Fix it by applying the SB_LAZYTIME and SB_I_VERSION flags from *flags to
s_flags before we parse the option and use the resulting state of the
same flags in *flags at the end of successful remount.

Signed-off-by: Lukas Czerner 
Reviewed-by: Ritesh Harjani 
Link: https://lore.kernel.org/r/20200723150526.19931-1-lczerner@redhat.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

ext4: abort the filesystem if failed to async write metadata buffer

2020-09-03T09:29:30+00:00

[ Upstream commit bc71726c725767205757821df364acff87f92ac5 ]

There is a risk of filesystem inconsistency if we failed to async write
back metadata buffer in the background. Because of current buffer's end
io procedure is handled by end_buffer_async_write() in the block layer,
and it only clear the buffer's uptodate flag and mark the write_io_error
flag, so ext4 cannot detect such failure immediately. In most cases of
getting metadata buffer (e.g. ext4_read_inode_bitmap()), although the
buffer's data is actually uptodate, it may still read data from disk
because the buffer's uptodate flag has been cleared. Finally, it may
lead to on-disk filesystem inconsistency if reading old data from the
disk successfully and write them out again.

This patch detect bdev mapping->wb_err when getting journal's write
access and mark the filesystem error if bdev's mapping->wb_err was
increased, this could prevent further writing and potential
inconsistency.

Signed-off-by: zhangyi (F) 
Suggested-by: Jan Kara 
Link: https://lore.kernel.org/r/20200620025427.1756360-2-yi.zhang@huawei.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

ext4: skip non-loaded groups at cr=0/1 when scanning for good groups

2020-09-03T09:29:30+00:00

[ Upstream commit c1d2c7d47e15482bb23cda83a5021e60f624a09c ]

cr=0 is supposed to be an optimization to save CPU cycles, but if
buddy data (in memory) is not initialized then all this makes no sense
as we have to do sync IO taking a lot of cycles.  Also, at cr=0
mballoc doesn't choose any available chunk.  cr=1 also skips groups
using heuristic based on avg. fragment size.  It's more useful to skip
such groups and switch to cr=2 where groups will be scanned for
available chunks.  However, we always read the first block group in a
flex_bg so metadata blocks will get read into the first flex_bg if
possible.

Using sparse image and dm-slow virtual device of 120TB was
simulated, then the image was formatted and filled using debugfs to
mark ~85% of available space as busy.  mount process w/o the patch
couldn't complete in half an hour (according to vmstat it would take
~10-11 hours).  With the patch applied mount took ~20 seconds.

Lustre-bug-id: https://jira.whamcloud.com/browse/LU-12988
Signed-off-by: Alex Zhuravlev 
Reviewed-by: Andreas Dilger 
Reviewed-by: Artem Blagodarenko 
Signed-off-by: Sasha Levin

ext4: handle read only external journal device

2020-09-03T09:29:30+00:00

[ Upstream commit 273108fa5015eeffc4bacfa5ce272af3434b96e4 ]

Ext4 uses blkdev_get_by_dev() to get the block_device for journal device
which does check to see if the read-only block device was opened
read-only.

As a result ext4 will hapily proceed mounting the file system with
external journal on read-only device. This is bad as we would not be
able to use the journal leading to errors later on.

Instead of simply failing to mount file system in this case, treat it in
a similar way we treat internal journal on read-only device. Allow to
mount with -o noload in read-only mode.

This can be reproduced easily like this:

mke2fs -F -O journal_dev $JOURNAL_DEV 100M
mkfs.$FSTYPE -F -J device=$JOURNAL_DEV $FS_DEV
blockdev --setro $JOURNAL_DEV
mount $FS_DEV $MNT
touch $MNT/file
umount $MNT

leading to error like this

[ 1307.318713] ------------[ cut here ]------------
[ 1307.323362] generic_make_request: Trying to write to read-only block-device dm-2 (partno 0)
[ 1307.331741] WARNING: CPU: 36 PID: 3224 at block/blk-core.c:855 generic_make_request_checks+0x2c3/0x580
[ 1307.341041] Modules linked in: ext4 mbcache jbd2 rfkill intel_rapl_msr intel_rapl_common isst_if_commd
[ 1307.419445] CPU: 36 PID: 3224 Comm: jbd2/dm-2 Tainted: G        W I       5.8.0-rc5 #2
[ 1307.427359] Hardware name: Dell Inc. PowerEdge R740/01KPX8, BIOS 2.3.10 08/15/2019
[ 1307.434932] RIP: 0010:generic_make_request_checks+0x2c3/0x580
[ 1307.440676] Code: 94 03 00 00 48 89 df 48 8d 74 24 08 c6 05 cf 2b 18 01 01 e8 7f a4 ff ff 48 c7 c7 50e
[ 1307.459420] RSP: 0018:ffffc0d70eb5fb48 EFLAGS: 00010286
[ 1307.464646] RAX: 0000000000000000 RBX: ffff9b33b2978300 RCX: 0000000000000000
[ 1307.471780] RDX: ffff9b33e12a81e0 RSI: ffff9b33e1298000 RDI: ffff9b33e1298000
[ 1307.478913] RBP: ffff9b7b9679e0c0 R08: 0000000000000837 R09: 0000000000000024
[ 1307.486044] R10: 0000000000000000 R11: ffffc0d70eb5f9f0 R12: 0000000000000400
[ 1307.493177] R13: 0000000000000000 R14: 0000000000000001 R15: 0000000000000000
[ 1307.500308] FS:  0000000000000000(0000) GS:ffff9b33e1280000(0000) knlGS:0000000000000000
[ 1307.508396] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1307.514142] CR2: 000055eaf4109000 CR3: 0000003dee40a006 CR4: 00000000007606e0
[ 1307.521273] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 1307.528407] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 1307.535538] PKRU: 55555554
[ 1307.538250] Call Trace:
[ 1307.540708]  generic_make_request+0x30/0x340
[ 1307.544985]  submit_bio+0x43/0x190
[ 1307.548393]  ? bio_add_page+0x62/0x90
[ 1307.552068]  submit_bh_wbc+0x16a/0x190
[ 1307.555833]  jbd2_write_superblock+0xec/0x200 [jbd2]
[ 1307.560803]  jbd2_journal_update_sb_log_tail+0x65/0xc0 [jbd2]
[ 1307.566557]  jbd2_journal_commit_transaction+0x2ae/0x1860 [jbd2]
[ 1307.572566]  ? check_preempt_curr+0x7a/0x90
[ 1307.576756]  ? update_curr+0xe1/0x1d0
[ 1307.580421]  ? account_entity_dequeue+0x7b/0xb0
[ 1307.584955]  ? newidle_balance+0x231/0x3d0
[ 1307.589056]  ? __switch_to_asm+0x42/0x70
[ 1307.592986]  ? __switch_to_asm+0x36/0x70
[ 1307.596918]  ? lock_timer_base+0x67/0x80
[ 1307.600851]  kjournald2+0xbd/0x270 [jbd2]
[ 1307.604873]  ? finish_wait+0x80/0x80
[ 1307.608460]  ? commit_timeout+0x10/0x10 [jbd2]
[ 1307.612915]  kthread+0x114/0x130
[ 1307.616152]  ? kthread_park+0x80/0x80
[ 1307.619816]  ret_from_fork+0x22/0x30
[ 1307.623400] ---[ end trace 27490236265b1630 ]---

Signed-off-by: Lukas Czerner 
Reviewed-by: Andreas Dilger 
Link: https://lore.kernel.org/r/20200717090605.2612-1-lczerner@redhat.com
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

ext4: don't BUG on inconsistent journal feature

2020-09-03T09:29:29+00:00

[ Upstream commit 11215630aada28307ba555a43138db6ac54fa825 ]

A customer has reported a BUG_ON in ext4_clear_journal_err() hitting
during an LTP testing. Either this has been caused by a test setup
issue where the filesystem was being overwritten while LTP was mounting
it or the journal replay has overwritten the superblock with invalid
data. In either case it is preferable we don't take the machine down
with a BUG_ON. So handle the situation of unexpectedly missing
has_journal feature more gracefully. We issue warning and fail the mount
in the cases where the race window is narrow and the failed check is
most likely a programming error. In cases where fs corruption is more
likely, we do full ext4_error() handling before failing mount / remount.

Reviewed-by: Lukas Czerner 
Signed-off-by: Jan Kara 
Link: https://lore.kernel.org/r/20200710140759.18031-1-jack@suse.cz
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin

ext4: check journal inode extents more carefully

2020-08-26T09:49:08+00:00

[ Upstream commit ce9f24cccdc019229b70a5c15e2b09ad9c0ab5d1 ]

Currently, system zones just track ranges of block, that are "important"
fs metadata (bitmaps, group descriptors, journal blocks, etc.). This
however complicates how extent tree (or indirect blocks) can be checked
for inodes that actually track such metadata - currently the journal
inode but arguably we should be treating quota files or resize inode
similarly. We cannot run __ext4_ext_check() on such metadata inodes when
loading their extents as that would immediately trigger the validity
checks and so we just hack around that and special-case the journal
inode. This however leads to a situation that a journal inode which has
extent tree of depth at least one can have invalid extent tree that gets
unnoticed until ext4_cache_extents() crashes.

To overcome this limitation, track inode number each system zone belongs
to (0 is used for zones not belonging to any inode). We can then verify
inode number matches the expected one when verifying extent tree and
thus avoid the false errors. With this there's no need to to
special-case journal inode during extent tree checking anymore so remove
it.

Fixes: 0a944e8a6c66 ("ext4: don't perform block validity checks on the journal inode")
Reported-by: Wolfgang Frisch 
Reviewed-by: Lukas Czerner 
Signed-off-by: Jan Kara 
Link: https://lore.kernel.org/r/20200728130437.7804-4-jack@suse.cz
Signed-off-by: Theodore Ts'o 
Signed-off-by: Sasha Levin