* [Bug 204049] [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646
2019-07-02 8:04 [Bug 204049] New: [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646 bugzilla-daemon
@ 2019-07-18 20:00 ` bugzilla-daemon
2019-07-18 20:19 ` bugzilla-daemon
` (4 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: bugzilla-daemon @ 2019-07-18 20:00 UTC (permalink / raw)
To: linux-xfs
https://bugzilla.kernel.org/show_bug.cgi?id=204049
--- Comment #1 from Luis Chamberlain (mcgrof@kernel.org) ---
I reported an immediate v4.19.58 vanilla crash with generic/388 but with the
"xfs_nocrc" and "xfs_reflink" configuration as per oscheck's testing:
The "xfs_nocrc":
# xfs_info /dev/loop5
meta-data=/dev/loop5 isize=256 agcount=4, agsize=1310720 blks
= sectsz=512 attr=2, projid32bit=1
= crc=0 finobt=0, sparse=0, rmapbt=0
= reflink=0
data = bsize=4096 blocks=5242880, imaxpct=25
= sunit=0 swidth=0 blks
naming =version 2 bsize=4096 ascii-ci=0, ftype=1
log =internal log bsize=4096 blocks=2560, version=2
= sectsz=512 sunit=0 blks, lazy-count=1
realtime =none extsz=4096 blocks=0, rtextents=0
The "xfs_reflink" configuration:
# xfs_info /dev/loop5
meta-data=/dev/loop5 isize=512 agcount=4, agsize=1310720 blks
= sectsz=512 attr=2, projid32bit=1
= crc=1 finobt=1, sparse=1, rmapbt=1
= reflink=1
data = bsize=4096 blocks=5242880, imaxpct=25
= sunit=0 swidth=0 blks
naming =version 2 bsize=4096 ascii-ci=0, ftype=1
log =internal log bsize=4096 blocks=3693, version=2
= sectsz=512 sunit=0 blks, lazy-count=1
realtime =none extsz=4096 blocks=0, rtextents=0
This is being tracked on this bug report:
https://bugzilla.kernel.org/show_bug.cgi?id=204223
The configuration above has rmapbt=1, you have rmapbt=0, at least in
discussions over which types of configurations to test for stable long ago on
the mailing list using rmapbt=0 with reflink was not one which we set out to
cover, so curious what the motivation was for tracking problems with it were
now. I'll just refer to this configuration then as "xfs_reflink_normapbt" and
I'll consider tracking it for stable depending on why you set out to cover it
as well.
I cannot reproduce your crash on v4.19.58 with your same configuration,
""xfs_reflink_normapbt", at least so far I've ran the test 15 times in a loop
and I see no failure. The other crashes occur within 1-3 times of running the
test. How many times did you run the test for it to crash on the system?
I'll leave the test running a bit longer just in case.
Given what I am seeing though, it seems likely there may be a regression here.
Could you bisect? We at least now have an idea of what to expect around the
v4.19 for different configurations including yours.
--
You are receiving this mail because:
You are watching the assignee of the bug.
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug 204049] [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646
2019-07-02 8:04 [Bug 204049] New: [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646 bugzilla-daemon
2019-07-18 20:00 ` [Bug 204049] " bugzilla-daemon
@ 2019-07-18 20:19 ` bugzilla-daemon
2019-07-18 20:29 ` bugzilla-daemon
` (3 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: bugzilla-daemon @ 2019-07-18 20:19 UTC (permalink / raw)
To: linux-xfs
https://bugzilla.kernel.org/show_bug.cgi?id=204049
--- Comment #2 from Luis Chamberlain (mcgrof@kernel.org) ---
I spoke too soon, after 30 runs I managed to hit the same crash that I reported
on kz#204223 [0] with the same configuration you used to report here this bug,
"xfs_reflink_normapbt".
So, if you bisect, it will still fail, but you should hit at lest a different
crash, the one being tracked on kz#204223.
[0] https://bugzilla.kernel.org/show_bug.cgi?id=204223
--
You are receiving this mail because:
You are watching the assignee of the bug.
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug 204049] [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646
2019-07-02 8:04 [Bug 204049] New: [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646 bugzilla-daemon
2019-07-18 20:00 ` [Bug 204049] " bugzilla-daemon
2019-07-18 20:19 ` bugzilla-daemon
@ 2019-07-18 20:29 ` bugzilla-daemon
2019-07-19 19:37 ` bugzilla-daemon
` (2 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: bugzilla-daemon @ 2019-07-18 20:29 UTC (permalink / raw)
To: linux-xfs
https://bugzilla.kernel.org/show_bug.cgi?id=204049
--- Comment #3 from Luis Chamberlain (mcgrof@kernel.org) ---
And, do you not get a crash with the "xfs_nocrc" or "xfs_reflink"
configurations? What if you run the test in a loop. To test in a loop I use
oscheck's naggy-check.sh [0]. Provided you have configured your sections named
in the fstests configuration as I have on oscheck [1], you should be able to
use it as follows until a failure is found:
./naggy-check.sh -s xfs_reflink_normapbt -f generic/388
I just added xfs_reflink_normapbt to the upstream oscheck demo config.
[0] https://gitlab.com/mcgrof/oscheck/blob/master/naggy-check.sh
[1] https://gitlab.com/mcgrof/oscheck/blob/master/fstests-configs/xfs.config
--
You are receiving this mail because:
You are watching the assignee of the bug.
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug 204049] [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646
2019-07-02 8:04 [Bug 204049] New: [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646 bugzilla-daemon
` (2 preceding siblings ...)
2019-07-18 20:29 ` bugzilla-daemon
@ 2019-07-19 19:37 ` bugzilla-daemon
2019-07-19 21:29 ` bugzilla-daemon
2019-07-19 21:45 ` bugzilla-daemon
5 siblings, 0 replies; 8+ messages in thread
From: bugzilla-daemon @ 2019-07-19 19:37 UTC (permalink / raw)
To: linux-xfs
https://bugzilla.kernel.org/show_bug.cgi?id=204049
--- Comment #4 from Luis Chamberlain (mcgrof@kernel.org) ---
The crash observed on stable kernels can be fixed with commit 6958d11f77 ("xfs:
don't trip over uninitialized buffer on extent read of corrupted inode") merged
on v5.1.
I can't reproduce the immediate panic on v5.1 with the "xfs_reflink_normapbt"
anymore, as such I believe this seems like a regression, and you should be able
to bisect to v5.1 as the good kernel.
--
You are receiving this mail because:
You are watching the assignee of the bug.
^ permalink raw reply [flat|nested] 8+ messages in thread* [Bug 204049] [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646
2019-07-02 8:04 [Bug 204049] New: [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646 bugzilla-daemon
` (3 preceding siblings ...)
2019-07-19 19:37 ` bugzilla-daemon
@ 2019-07-19 21:29 ` bugzilla-daemon
2019-07-19 21:37 ` Bill O'Donnell
2019-07-19 21:45 ` bugzilla-daemon
5 siblings, 1 reply; 8+ messages in thread
From: bugzilla-daemon @ 2019-07-19 21:29 UTC (permalink / raw)
To: linux-xfs
https://bugzilla.kernel.org/show_bug.cgi?id=204049
--- Comment #5 from Luis Chamberlain (mcgrof@kernel.org) ---
(In reply to Luis Chamberlain from comment #4)
> The crash observed on stable kernels can be fixed with commit 6958d11f77
> ("xfs: don't trip over uninitialized buffer on extent read of corrupted
> inode") merged on v5.1.
>
> I can't reproduce the immediate panic on v5.1 with the
> "xfs_reflink_normapbt" anymore, as such I believe this seems like a
> regression, and you should be able to bisect to v5.1 as the good kernel.
Correction, it just took longer to crash. Same crash as in kz#204223 [0] on
v5.1. That issue is still not fixed.
[0] https://bugzilla.kernel.org/show_bug.cgi?id=204223
--
You are receiving this mail because:
You are watching the assignee of the bug.
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: [Bug 204049] [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646
2019-07-19 21:29 ` bugzilla-daemon
@ 2019-07-19 21:37 ` Bill O'Donnell
0 siblings, 0 replies; 8+ messages in thread
From: Bill O'Donnell @ 2019-07-19 21:37 UTC (permalink / raw)
To: bugzilla-daemon; +Cc: linux-xfs
On Fri, Jul 19, 2019 at 09:29:04PM +0000, bugzilla-daemon@bugzilla.kernel.org wrote:
> https://bugzilla.kernel.org/show_bug.cgi?id=204049
>
> --- Comment #5 from Luis Chamberlain (mcgrof@kernel.org) ---
> (In reply to Luis Chamberlain from comment #4)
> > The crash observed on stable kernels can be fixed with commit 6958d11f77
> > ("xfs: don't trip over uninitialized buffer on extent read of corrupted
> > inode") merged on v5.1.
> >
> > I can't reproduce the immediate panic on v5.1 with the
> > "xfs_reflink_normapbt" anymore, as such I believe this seems like a
> > regression, and you should be able to bisect to v5.1 as the good kernel.
>
> Correction, it just took longer to crash. Same crash as in kz#204223 [0] on
> v5.1. That issue is still not fixed.
In my experience,
many iterations of g/388 will definitely ferret out elusive failures.
It's sometimes tough to distinguish between real problems and abheritions.
-Bill
>
> [0] https://bugzilla.kernel.org/show_bug.cgi?id=204223
>
> --
> You are receiving this mail because:
> You are watching the assignee of the bug.
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Bug 204049] [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646
2019-07-02 8:04 [Bug 204049] New: [xfstests generic/388]: XFS: Assertion failed: ip->i_d.di_format != XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK), file: fs/xfs/xfs_inode.c, line: 3646 bugzilla-daemon
` (4 preceding siblings ...)
2019-07-19 21:29 ` bugzilla-daemon
@ 2019-07-19 21:45 ` bugzilla-daemon
5 siblings, 0 replies; 8+ messages in thread
From: bugzilla-daemon @ 2019-07-19 21:45 UTC (permalink / raw)
To: linux-xfs
https://bugzilla.kernel.org/show_bug.cgi?id=204049
--- Comment #6 from billodo@redhat.com ---
On Fri, Jul 19, 2019 at 09:29:04PM +0000, bugzilla-daemon@bugzilla.kernel.org
wrote:
> https://bugzilla.kernel.org/show_bug.cgi?id=204049
>
> --- Comment #5 from Luis Chamberlain (mcgrof@kernel.org) ---
> (In reply to Luis Chamberlain from comment #4)
> > The crash observed on stable kernels can be fixed with commit 6958d11f77
> > ("xfs: don't trip over uninitialized buffer on extent read of corrupted
> > inode") merged on v5.1.
> >
> > I can't reproduce the immediate panic on v5.1 with the
> > "xfs_reflink_normapbt" anymore, as such I believe this seems like a
> > regression, and you should be able to bisect to v5.1 as the good kernel.
>
> Correction, it just took longer to crash. Same crash as in kz#204223 [0] on
> v5.1. That issue is still not fixed.
In my experience,
many iterations of g/388 will definitely ferret out elusive failures.
It's sometimes tough to distinguish between real problems and abheritions.
-Bill
>
> [0] https://bugzilla.kernel.org/show_bug.cgi?id=204223
>
> --
> You are receiving this mail because:
> You are watching the assignee of the bug.
--
You are receiving this mail because:
You are watching the assignee of the bug.
^ permalink raw reply [flat|nested] 8+ messages in thread