* Corruption of root fs during git bisect of drm system hang
@ 2013-07-10 9:06 Markus Trippelsdorf
2013-07-11 0:31 ` Dave Chinner
2013-07-11 0:37 ` Stan Hoeppner
0 siblings, 2 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-10 9:06 UTC (permalink / raw)
To: xfs
While bisecting a system hang, caused by the drm gpu subsystem, my root fs got
corrupted:
# xfs_repair /dev/sdc2
Phase 1 - find and verify superblock...
Phase 2 - using internal log
- zero log...
- scan filesystem freespace and inode maps...
agi unlinked bucket 6 is 682886 in ag 3 (inode=101346182)
agi unlinked bucket 7 is 11335 in ag 3 (inode=100674631)
agi unlinked bucket 10 is 682890 in ag 3 (inode=101346186)
agi unlinked bucket 21 is 981 in ag 3 (inode=100664277)
agi unlinked bucket 23 is 5704343 in ag 3 (inode=106367639)
agi unlinked bucket 29 is 211421 in ag 3 (inode=100874717)
agi unlinked bucket 31 is 7681375 in ag 3 (inode=108344671)
agi unlinked bucket 34 is 3480162 in ag 3 (inode=104143458)
agi unlinked bucket 40 is 211432 in ag 3 (inode=100874728)
agi unlinked bucket 41 is 2704937 in ag 3 (inode=103368233)
agi unlinked bucket 45 is 594669 in ag 3 (inode=101257965)
agi unlinked bucket 62 is 11902 in ag 3 (inode=100675198)
- found root inode chunk
Phase 3 - for each AG...
- scan and clear agi unlinked lists...
- process known inodes and perform inode discovery...
- agno = 0
- agno = 1
- agno = 2
7f560b7fe700: Badness in key lookup (length)
bp=(bno 46883808, len 16384 bytes) key=(bno 46883808, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 46888976, len 16384 bytes) key=(bno 46888976, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 46889264, len 16384 bytes) key=(bno 46889264, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 46989024, len 16384 bytes) key=(bno 46989024, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 47180640, len 16384 bytes) key=(bno 47180640, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 47224768, len 16384 bytes) key=(bno 47224768, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 48235776, len 16384 bytes) key=(bno 48235776, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 48623392, len 16384 bytes) key=(bno 48623392, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 49735472, len 16384 bytes) key=(bno 49735472, len 8192 bytes)
7f560b7fe700: Badness in key lookup (length)
bp=(bno 50723984, len 16384 bytes) key=(bno 50723984, len 8192 bytes)
- agno = 3
- process newly discovered inodes...
Phase 4 - check for duplicate blocks...
- setting up duplicate extent list...
- check for inodes claiming duplicate blocks...
- agno = 0
- agno = 1
- agno = 2
- agno = 3
Phase 5 - rebuild AG headers and trees...
- reset superblock...
Phase 6 - check inode connectivity...
- resetting contents of realtime bitmap and summary inodes
- traversing filesystem ...
- traversal finished ...
- moving disconnected inodes to lost+found ...
disconnected inode 100664277, moving to lost+found
disconnected inode 100674631, moving to lost+found
disconnected inode 100675198, moving to lost+found
disconnected inode 100874717, moving to lost+found
disconnected inode 100874728, moving to lost+found
disconnected inode 101257965, moving to lost+found
disconnected inode 101346182, moving to lost+found
disconnected inode 101346186, moving to lost+found
disconnected inode 103368233, moving to lost+found
disconnected inode 104143458, moving to lost+found
disconnected inode 106367639, moving to lost+found
disconnected inode 108344671, moving to lost+found
Phase 7 - verify and correct link counts...
cache_purge: shake on cache 0x12a6030 left 1 nodes!?
done
This happend twice in the last few days and thus appears to be reproducible.
My root fs lives on a small SSD:
/dev/root on / type xfs (rw,relatime,attr2,inode64,logbsize=256k,noquota)
/dev/root xfs 30G 15G 16G 50% /
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-10 9:06 Corruption of root fs during git bisect of drm system hang Markus Trippelsdorf
@ 2013-07-11 0:31 ` Dave Chinner
2013-07-11 3:36 ` Markus Trippelsdorf
2013-07-11 0:37 ` Stan Hoeppner
1 sibling, 1 reply; 37+ messages in thread
From: Dave Chinner @ 2013-07-11 0:31 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: xfs
On Wed, Jul 10, 2013 at 11:06:34AM +0200, Markus Trippelsdorf wrote:
> While bisecting a system hang, caused by the drm gpu subsystem, my root fs got
> corrupted:
I don't see any corruption being repaired....
>
> # xfs_repair /dev/sdc2
> Phase 1 - find and verify superblock...
> Phase 2 - using internal log
> - zero log...
> - scan filesystem freespace and inode maps...
> agi unlinked bucket 6 is 682886 in ag 3 (inode=101346182)
> agi unlinked bucket 7 is 11335 in ag 3 (inode=100674631)
> agi unlinked bucket 10 is 682890 in ag 3 (inode=101346186)
> agi unlinked bucket 21 is 981 in ag 3 (inode=100664277)
> agi unlinked bucket 23 is 5704343 in ag 3 (inode=106367639)
> agi unlinked bucket 29 is 211421 in ag 3 (inode=100874717)
> agi unlinked bucket 31 is 7681375 in ag 3 (inode=108344671)
> agi unlinked bucket 34 is 3480162 in ag 3 (inode=104143458)
> agi unlinked bucket 40 is 211432 in ag 3 (inode=100874728)
> agi unlinked bucket 41 is 2704937 in ag 3 (inode=103368233)
> agi unlinked bucket 45 is 594669 in ag 3 (inode=101257965)
> agi unlinked bucket 62 is 11902 in ag 3 (inode=100675198)
That's a filesystem that has unlinked inodes on the unlinked list.
They get cleaned up during log replay. All the other "errors" are
related to cleaning these up....
So what is making you think there is a corruption? What's the error
being reported when you are using the filesystem? i.e. what's the
entire process you go through before you get to finding this
problem?
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-10 9:06 Corruption of root fs during git bisect of drm system hang Markus Trippelsdorf
2013-07-11 0:31 ` Dave Chinner
@ 2013-07-11 0:37 ` Stan Hoeppner
2013-07-11 3:47 ` Markus Trippelsdorf
1 sibling, 1 reply; 37+ messages in thread
From: Stan Hoeppner @ 2013-07-11 0:37 UTC (permalink / raw)
To: xfs
On 7/10/2013 4:06 AM, Markus Trippelsdorf wrote:
> While bisecting a system hang, caused by the drm gpu subsystem, my root fs got
> corrupted:
...
> My root fs lives on a small SSD:
>
> /dev/root on / type xfs (rw,relatime,attr2,inode64,logbsize=256k,noquota)
> /dev/root xfs 30G 15G 16G 50% /
Does your SSD honor barriers and cache flushes?
--
Stan
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 0:31 ` Dave Chinner
@ 2013-07-11 3:36 ` Markus Trippelsdorf
2013-07-11 3:58 ` Dave Chinner
0 siblings, 1 reply; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-11 3:36 UTC (permalink / raw)
To: Dave Chinner; +Cc: xfs
On 2013.07.11 at 10:31 +1000, Dave Chinner wrote:
> On Wed, Jul 10, 2013 at 11:06:34AM +0200, Markus Trippelsdorf wrote:
> > While bisecting a system hang, caused by the drm gpu subsystem, my root fs got
> > corrupted:
>
> I don't see any corruption being repaired....
>
> >
> > # xfs_repair /dev/sdc2
> > Phase 1 - find and verify superblock...
> > Phase 2 - using internal log
> > - zero log...
> > - scan filesystem freespace and inode maps...
> > agi unlinked bucket 6 is 682886 in ag 3 (inode=101346182)
> > agi unlinked bucket 7 is 11335 in ag 3 (inode=100674631)
> > agi unlinked bucket 10 is 682890 in ag 3 (inode=101346186)
> > agi unlinked bucket 21 is 981 in ag 3 (inode=100664277)
> > agi unlinked bucket 23 is 5704343 in ag 3 (inode=106367639)
> > agi unlinked bucket 29 is 211421 in ag 3 (inode=100874717)
> > agi unlinked bucket 31 is 7681375 in ag 3 (inode=108344671)
> > agi unlinked bucket 34 is 3480162 in ag 3 (inode=104143458)
> > agi unlinked bucket 40 is 211432 in ag 3 (inode=100874728)
> > agi unlinked bucket 41 is 2704937 in ag 3 (inode=103368233)
> > agi unlinked bucket 45 is 594669 in ag 3 (inode=101257965)
> > agi unlinked bucket 62 is 11902 in ag 3 (inode=100675198)
>
> That's a filesystem that has unlinked inodes on the unlinked list.
> They get cleaned up during log replay. All the other "errors" are
> related to cleaning these up....
>
> So what is making you think there is a corruption? What's the error
> being reported when you are using the filesystem? i.e. what's the
> entire process you go through before you get to finding this
> problem?
I was loosing my KDE settings bit by bit with every reboot during the
bisection. First my window-rules disappeared, then my desktop background
changed to default, then my taskbar moved from top to the bottom, etc.
In the end I had to restore all my .files from backup.
And please note that xfs_repair unlinked the inodes _after_ the
filesystem has been mounted and unmounted normally.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 0:37 ` Stan Hoeppner
@ 2013-07-11 3:47 ` Markus Trippelsdorf
0 siblings, 0 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-11 3:47 UTC (permalink / raw)
To: Stan Hoeppner; +Cc: xfs
On 2013.07.10 at 19:37 -0500, Stan Hoeppner wrote:
> On 7/10/2013 4:06 AM, Markus Trippelsdorf wrote:
> > While bisecting a system hang, caused by the drm gpu subsystem, my root fs got
> > corrupted:
> ...
> > My root fs lives on a small SSD:
> >
> > /dev/root on / type xfs (rw,relatime,attr2,inode64,logbsize=256k,noquota)
> > /dev/root xfs 30G 15G 16G 50% /
>
> Does your SSD honor barriers and cache flushes?
Yes, I hope so:
Enabled Supported:
...
* Mandatory FLUSH_CACHE
* FLUSH_CACHE_EXT
But please note that the system hang mentioned above turned out to be a
locking problem. So the SSD had plenty of time to process any out-
standing commands before reboot.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 3:36 ` Markus Trippelsdorf
@ 2013-07-11 3:58 ` Dave Chinner
2013-07-11 4:12 ` Stan Hoeppner
2013-07-11 4:15 ` Markus Trippelsdorf
0 siblings, 2 replies; 37+ messages in thread
From: Dave Chinner @ 2013-07-11 3:58 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: xfs
On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> On 2013.07.11 at 10:31 +1000, Dave Chinner wrote:
> > On Wed, Jul 10, 2013 at 11:06:34AM +0200, Markus Trippelsdorf wrote:
> > > While bisecting a system hang, caused by the drm gpu subsystem, my root fs got
> > > corrupted:
> >
> > I don't see any corruption being repaired....
> >
> > >
> > > # xfs_repair /dev/sdc2
> > > Phase 1 - find and verify superblock...
> > > Phase 2 - using internal log
> > > - zero log...
> > > - scan filesystem freespace and inode maps...
> > > agi unlinked bucket 6 is 682886 in ag 3 (inode=101346182)
> > > agi unlinked bucket 7 is 11335 in ag 3 (inode=100674631)
> > > agi unlinked bucket 10 is 682890 in ag 3 (inode=101346186)
> > > agi unlinked bucket 21 is 981 in ag 3 (inode=100664277)
> > > agi unlinked bucket 23 is 5704343 in ag 3 (inode=106367639)
> > > agi unlinked bucket 29 is 211421 in ag 3 (inode=100874717)
> > > agi unlinked bucket 31 is 7681375 in ag 3 (inode=108344671)
> > > agi unlinked bucket 34 is 3480162 in ag 3 (inode=104143458)
> > > agi unlinked bucket 40 is 211432 in ag 3 (inode=100874728)
> > > agi unlinked bucket 41 is 2704937 in ag 3 (inode=103368233)
> > > agi unlinked bucket 45 is 594669 in ag 3 (inode=101257965)
> > > agi unlinked bucket 62 is 11902 in ag 3 (inode=100675198)
> >
> > That's a filesystem that has unlinked inodes on the unlinked list.
> > They get cleaned up during log replay. All the other "errors" are
> > related to cleaning these up....
> >
> > So what is making you think there is a corruption? What's the error
> > being reported when you are using the filesystem? i.e. what's the
> > entire process you go through before you get to finding this
> > problem?
>
> I was loosing my KDE settings bit by bit with every reboot during the
> bisection. First my window-rules disappeared, then my desktop background
> changed to default, then my taskbar moved from top to the bottom, etc.
> In the end I had to restore all my .files from backup.
That's not filesystem corruption. That sounds more like someone not
using fsync in the apropriate place when overwriting a file....
> And please note that xfs_repair unlinked the inodes _after_ the
> filesystem has been mounted and unmounted normally.
Which means we might not be processing the unlinked lists correctly
and leaking them. If repair is finding the inodes in the AGI
unlinked lists, then recovery should be finding them, too. Not
processing them and not clearing the AGI bucket tends to imply that
recovery failed to read the AGI buffer.
What error messages are in dmesg, if any? And what kernel are you
running?
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 3:58 ` Dave Chinner
@ 2013-07-11 4:12 ` Stan Hoeppner
2013-07-11 9:07 ` Markus Trippelsdorf
2013-07-11 4:15 ` Markus Trippelsdorf
1 sibling, 1 reply; 37+ messages in thread
From: Stan Hoeppner @ 2013-07-11 4:12 UTC (permalink / raw)
To: Dave Chinner; +Cc: Markus Trippelsdorf, xfs
On 7/10/2013 10:58 PM, Dave Chinner wrote:
> On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
>> I was loosing my KDE settings bit by bit with every reboot during the
>> bisection. First my window-rules disappeared, then my desktop background
>> changed to default, then my taskbar moved from top to the bottom, etc.
>> In the end I had to restore all my .files from backup.
>
> That's not filesystem corruption. That sounds more like someone not
> using fsync in the apropriate place when overwriting a file....
>From Sandeen's blog, March 2009:
"I dunno how to resolve this right now. I talked to some nice KDE folks
on irc; they basically want atomic writes, either you get your old file
or your new file post-crash; and tempfile/sync/rename does this – but
the fsync hurts on 78% of the Linux filesystems out there. So their
KSaveFile class doesn’t fsync. So what to do, what to do.."
That's 4 years ago. Is it possible the KDE devs are still not using
fsync? Sure seems likely given Markus' problem.
--
Stan
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 3:58 ` Dave Chinner
2013-07-11 4:12 ` Stan Hoeppner
@ 2013-07-11 4:15 ` Markus Trippelsdorf
1 sibling, 0 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-11 4:15 UTC (permalink / raw)
To: Dave Chinner; +Cc: xfs
On 2013.07.11 at 13:58 +1000, Dave Chinner wrote:
> On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> > On 2013.07.11 at 10:31 +1000, Dave Chinner wrote:
> > > On Wed, Jul 10, 2013 at 11:06:34AM +0200, Markus Trippelsdorf wrote:
> > > > While bisecting a system hang, caused by the drm gpu subsystem, my root fs got
> > > > corrupted:
> > >
> > > That's a filesystem that has unlinked inodes on the unlinked list.
> > > They get cleaned up during log replay. All the other "errors" are
> > > related to cleaning these up....
> > >
> > > So what is making you think there is a corruption? What's the error
> > > being reported when you are using the filesystem? i.e. what's the
> > > entire process you go through before you get to finding this
> > > problem?
> >
> > I was loosing my KDE settings bit by bit with every reboot during the
> > bisection. First my window-rules disappeared, then my desktop background
> > changed to default, then my taskbar moved from top to the bottom, etc.
> > In the end I had to restore all my .files from backup.
>
> That's not filesystem corruption. That sounds more like someone not
> using fsync in the apropriate place when overwriting a file....
Ok. Sorry for using the wrong term.
> > And please note that xfs_repair unlinked the inodes _after_ the
> > filesystem has been mounted and unmounted normally.
>
> Which means we might not be processing the unlinked lists correctly
> and leaking them. If repair is finding the inodes in the AGI
> unlinked lists, then recovery should be finding them, too. Not
> processing them and not clearing the AGI bucket tends to imply that
> recovery failed to read the AGI buffer.
>
> What error messages are in dmesg, if any? And what kernel are you
> running?
There are no error messages in dmesg. I'm running the latest Linus tree.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 4:12 ` Stan Hoeppner
@ 2013-07-11 9:07 ` Markus Trippelsdorf
2013-07-11 11:28 ` Markus Trippelsdorf
2013-07-12 2:17 ` Dave Chinner
0 siblings, 2 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-11 9:07 UTC (permalink / raw)
To: Stan Hoeppner; +Cc: xfs
On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
>
> >> I was loosing my KDE settings bit by bit with every reboot during the
> >> bisection. First my window-rules disappeared, then my desktop background
> >> changed to default, then my taskbar moved from top to the bottom, etc.
> >> In the end I had to restore all my .files from backup.
> >
> > That's not filesystem corruption. That sounds more like someone not
> > using fsync in the apropriate place when overwriting a file....
>
> From Sandeen's blog, March 2009:
>
> "I dunno how to resolve this right now. I talked to some nice KDE folks
> on irc; they basically want atomic writes, either you get your old file
> or your new file post-crash; and tempfile/sync/rename does this – but
> the fsync hurts on 78% of the Linux filesystems out there. So their
> KSaveFile class doesn’t fsync. So what to do, what to do.."
>
> That's 4 years ago. Is it possible the KDE devs are still not using
> fsync? Sure seems likely given Markus' problem.
Looking at the source:
http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
it appears that one can set an environment variable KDE_EXTRA_FSYNC to
address this issue.
However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
loose my KDE settings in case of a crash. So the whole fsync thing might
be a red herring.
What's more this time I endend up with undeletable files in /tmp (for
example .X0-lock) after the crash:
(/dev/sdb was mounted and unmounted normally before I ran xfs_repair)
t@ubunt:~# xfs_repair /dev/sdb
Phase 1 - find and verify superblock...
Phase 2 - using internal log
- zero log...
- scan filesystem freespace and inode maps...
agi unlinked bucket 0 is 683435008 in ag 2 (inode=4978402304)
agi unlinked bucket 1 is 683435009 in ag 2 (inode=4978402305)
- found root inode chunk
Phase 3 - for each AG...
- scan and clear agi unlinked lists...
- process known inodes and perform inode discovery...
- agno = 0
imap claims a free inode 1897177 is in use, correcting imap and clearing inode
cleared inode 1897177
imap claims a free inode 1897179 is in use, correcting imap and clearing inode
cleared inode 1897179
imap claims a free inode 1897180 is in use, correcting imap and clearing inode
cleared inode 1897180
imap claims a free inode 1897181 is in use, correcting imap and clearing inode
cleared inode 1897181
- agno = 1
imap claims a free inode 2152321787 is in use, correcting imap and clearing inode
cleared inode 2152321787
imap claims a free inode 2152321789 is in use, correcting imap and clearing inode
cleared inode 2152321789
imap claims a free inode 2152321790 is in use, correcting imap and clearing inode
cleared inode 2152321790
imap claims a free inode 2152321792 is in use, correcting imap and clearing inode
cleared inode 2152321792
7f8fdbde0700: Badness in key lookup (length)
bp=(bno 1806856096, len 16384 bytes) key=(bno 1806856096, len 8192 bytes)
imap claims a free inode 2547803922 is in use, correcting imap and clearing inode
cleared inode 2547803922
- agno = 2
imap claims a free inode 4297828992 is in use, correcting imap and clearing inode
cleared inode 4297828992
imap claims a free inode 4297828996 is in use, correcting imap and clearing inode
cleared inode 4297828996
imap claims a free inode 4298016921 is in use, correcting imap and clearing inode
cleared inode 4298016921
imap claims a free inode 4299215923 is in use, correcting imap and clearing inode
cleared inode 4299215923
imap claims a free inode 4299249355 is in use, correcting imap and clearing inode
cleared inode 4299249355
imap claims a free inode 4299249356 is in use, correcting imap and clearing inode
cleared inode 4299249356
imap claims a free inode 4425478382 is in use, correcting imap and clearing inode
cleared inode 4425478382
imap claims a free inode 4425843325 is in use, correcting imap and clearing inode
cleared inode 4425843325
imap claims a free inode 4425843327 is in use, correcting imap and clearing inode
cleared inode 4425843327
- agno = 3
imap claims a free inode 6442719595 is in use, correcting imap and clearing inode
cleared inode 6442719595
imap claims a free inode 6443102082 is in use, correcting imap and clearing inode
cleared inode 6443102082
imap claims a free inode 6443102083 is in use, correcting imap and clearing inode
cleared inode 6443102083
imap claims a free inode 6443102093 is in use, correcting imap and clearing inode
cleared inode 6443102093
imap claims a free inode 6443102105 is in use, correcting imap and clearing inode
cleared inode 6443102105
imap claims a free inode 6443102106 is in use, correcting imap and clearing inode
cleared inode 6443102106
- process newly discovered inodes...
Phase 4 - check for duplicate blocks...
- setting up duplicate extent list...
- check for inodes claiming duplicate blocks...
- agno = 0
- agno = 3
- agno = 2
- agno = 1
entry "fcron.pid" at block 0 offset 1448 in directory inode 4297829059 references free inode 4297828992
clearing inode number in entry at offset 1448...
entry "fcron.fifo" at block 0 offset 1472 in directory inode 4297829059 references free inode 4297828996
clearing inode number in entry at offset 1472...
entry "syslog-ng.pid" at block 0 offset 1496 in directory inode 4297829059 references free inode 4425843325
clearing inode number in entry at offset 1496...
entry "squid.pid" at block 0 offset 1544 in directory inode 4297829059 references free inode 4425843327
clearing inode number in entry at offset 1544...
entry "KSMserver__0" in shortform directory 1897175 references free inode 4393990
junking entry "KSMserver__0" in directory inode 1897175
entry "syslog-ng.persist" in shortform directory 6455045890 references free inode 6442719595
junking entry "syslog-ng.persist" in directory inode 6455045890
corrected i8 count in directory 6455045890, was 5, now 4
entry "blkid.tab" in shortform directory 2526452945 references free inode 2547803923
junking entry "blkid.tab" in directory inode 2526452945
entry "watch" in shortform directory 2526452947 references free inode 1897177
junking entry "watch" in directory inode 2526452947
entry "+pci:0000:00:00.0" at block 0 offset 2056 in directory inode 6814850679 references free inode 6443102082
clearing inode number in entry at offset 2056...
entry "+pci:0000:00:06.0" at block 0 offset 2200 in directory inode 6814850679 references free inode 6443102105
clearing inode number in entry at offset 2200...
entry "+pci:0000:00:11.0" at block 0 offset 2272 in directory inode 6814850679 references free inode 6443102093
clearing inode number in entry at offset 2272...
entry "+pci:0000:00:12.1" at block 0 offset 2416 in directory inode 6814850679 references free inode 6443102083
clearing inode number in entry at offset 2416...
entry "+pci:0000:00:13.2" at block 0 offset 2704 in directory inode 6814850679 references free inode 6443102106
clearing inode number in entry at offset 2704...
entry "303" in shortform directory 2684075842 references free inode 2152321790
junking entry "303" in directory inode 2684075842
entry ".X0-lock" at block 0 offset 120 in directory inode 2684075878 references free inode 2152321787
clearing inode number in entry at offset 120...
entry "orbit-markus" at block 0 offset 224 in directory inode 2684075878 references free inode 4299249355
clearing inode number in entry at offset 224...
entry "OSL_PIPE_1000_SingleOfficeIPC_b9c45eb0c44ccbd12ea27aa1b043919f" at block 0 offset 248 in directory inode 2684075878 references free inode 2152321792
clearing inode number in entry at offset 248...
Phase 5 - rebuild AG headers and trees...
- reset superblock...
Phase 6 - check inode connectivity...
- resetting contents of realtime bitmap and summary inodes
- traversing filesystem ...
entry "queue.bin" in directory inode 2526452947 references already connected inode 2152321794.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 9:07 ` Markus Trippelsdorf
@ 2013-07-11 11:28 ` Markus Trippelsdorf
2013-07-11 20:24 ` Stan Hoeppner
2013-07-12 2:17 ` Dave Chinner
1 sibling, 1 reply; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-11 11:28 UTC (permalink / raw)
To: Stan Hoeppner; +Cc: xfs
On 2013.07.11 at 11:07 +0200, Markus Trippelsdorf wrote:
> On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> > On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> >
> > >> I was loosing my KDE settings bit by bit with every reboot during the
> > >> bisection. First my window-rules disappeared, then my desktop background
> > >> changed to default, then my taskbar moved from top to the bottom, etc.
> > >> In the end I had to restore all my .files from backup.
> > >
> > > That's not filesystem corruption. That sounds more like someone not
> > > using fsync in the apropriate place when overwriting a file....
> >
> > From Sandeen's blog, March 2009:
> >
> > "I dunno how to resolve this right now. I talked to some nice KDE folks
> > on irc; they basically want atomic writes, either you get your old file
> > or your new file post-crash; and tempfile/sync/rename does this – but
> > the fsync hurts on 78% of the Linux filesystems out there. So their
> > KSaveFile class doesn’t fsync. So what to do, what to do.."
> >
> > That's 4 years ago. Is it possible the KDE devs are still not using
> > fsync? Sure seems likely given Markus' problem.
>
> Looking at the source:
> http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
> it appears that one can set an environment variable KDE_EXTRA_FSYNC to
> address this issue.
>
> However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
> loose my KDE settings in case of a crash. So the whole fsync thing might
> be a red herring.
It turned out that the KDE_EXTRA_FSYNC variable doesn't affect KDE
config file handling at all.
So I've added an fsync in kconfigini.cpp (KConfigIniBackend::writeConfig)
and now I don't loose my settings anymore during kernel crash testing.
That is until xfs eats my KDE config files (kwinrulesr in this case):
root@ubunt:/home/markus# xfs_repair /dev/sdc2
Phase 1 - find and verify superblock...
Phase 2 - using internal log
- zero log...
- scan filesystem freespace and inode maps...
agi unlinked bucket 55 is 406711 in ag 3 (inode=101070007)
- found root inode chunk
Phase 3 - for each AG...
- scan and clear agi unlinked lists...
- process known inodes and perform inode discovery...
- agno = 0
imap claims a free inode 858183 is in use, correcting imap and clearing inode
cleared inode 858183
- agno = 1
imap claims a free inode 40112137 is in use, correcting imap and clearing inode
cleared inode 40112137
imap claims a free inode 40112354 is in use, correcting imap and clearing inode
cleared inode 40112354
- agno = 2
imap claims a free inode 68162927 is in use, correcting imap and clearing inode
cleared inode 68162927
7f336f1b6700: Badness in key lookup (length)
bp=(bno 47086672, len 16384 bytes) key=(bno 47086672, len 8192 bytes)
- agno = 3
imap claims a free inode 100865109 is in use, correcting imap and clearing inode
cleared inode 100865109
imap claims a free inode 101069993 is in use, correcting imap and clearing inode
cleared inode 101069993
imap claims a free inode 101070010 is in use, correcting imap and clearing inode
cleared inode 101070010
imap claims a free inode 101070015 is in use, correcting imap and clearing inode
cleared inode 101070015
- process newly discovered inodes...
Phase 4 - check for duplicate blocks...
- setting up duplicate extent list...
- check for inodes claiming duplicate blocks...
- agno = 0
- agno = 2
- agno = 1
- agno = 3
entry "mytexts.bau" in shortform directory 67333623 references free inode 68162927
junking entry "mytexts.bau" in directory inode 67333623
entry "dialog.xlc" in shortform directory 252098 references free inode 858183
junking entry "dialog.xlc" in directory inode 252098
entry "evolocal.odb" in shortform directory 100870253 references free inode 100865109
junking entry "evolocal.odb" in directory inode 100870253
entry "kwinrulesrc" at block 0 offset 2552 in directory inode 103698564 references free inode 101070010
clearing inode number in entry at offset 2552...
entry "kwinrulesrcbhc578.new" at block 0 offset 3224 in directory inode 103698564 references free inode 101070015
clearing inode number in entry at offset 3224...
entry "Module1.xba" in shortform directory 40112359 references free inode 40112354
junking entry "Module1.xba" in directory inode 40112359
entry "script.xlb" in shortform directory 40112359 references free inode 40112137
junking entry "script.xlb" in directory inode 40112359
Phase 5 - rebuild AG headers and trees...
- reset superblock...
Phase 6 - check inode connectivity...
- resetting contents of realtime bitmap and summary inodes
- traversing filesystem ...
bad hash table for directory inode 103698564 (no data entry): rebuilding
rebuilding directory inode 103698564
- traversal finished ...
- moving disconnected inodes to lost+found ...
disconnected inode 101070007, moving to lost+found
Phase 7 - verify and correct link counts...
cache_purge: shake on cache 0x1bc6030 left 1 nodes!?
done
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 11:28 ` Markus Trippelsdorf
@ 2013-07-11 20:24 ` Stan Hoeppner
2013-07-11 20:40 ` Markus Trippelsdorf
0 siblings, 1 reply; 37+ messages in thread
From: Stan Hoeppner @ 2013-07-11 20:24 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: xfs
On 7/11/2013 6:28 AM, Markus Trippelsdorf wrote:
...
>> Looking at the source:
>> http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
>> it appears that one can set an environment variable KDE_EXTRA_FSYNC to
>> address this issue.
>>
>> However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
>> loose my KDE settings in case of a crash. So the whole fsync thing might
>> be a red herring.
>
> It turned out that the KDE_EXTRA_FSYNC variable doesn't affect KDE
> config file handling at all.
> So I've added an fsync in kconfigini.cpp (KConfigIniBackend::writeConfig)
> and now I don't loose my settings anymore during kernel crash testing.
>
> That is until xfs eats my KDE config files (kwinrulesr in this case):
Adding fsync in kconfigini.cpp apparently doesn't force fsync for all
KDE file operations. You also have some Open Office files getting hosed
due to lack of fsync. XFS is not the cause of these problems.
The problem is that all of this desktop code was developed atop EXT3
which flushed to disk every 5 seconds. This made programmers sloppy as
they didn't have to fsync to make sure data hit disk. This problem has
been covered extensively by many, including Eric in other posts on his
blog. There's a really simple way to test this: mount with sync.
Report results after the next crash. If no files are corrupted then
you've verified the problem lay squarely on the shoulders of these
desktop developers who have abdicated their responsibility to make sure
their file changes hit the disk, instead of relying on a broken
filesystem do it for them.
Worth noting, using EXT4 without the EXT3 flush emulation enabled will
yield similar file corruption upon a crash.
--
Stan
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 20:24 ` Stan Hoeppner
@ 2013-07-11 20:40 ` Markus Trippelsdorf
2013-07-11 23:01 ` Stan Hoeppner
2013-07-12 2:38 ` Dave Chinner
0 siblings, 2 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-11 20:40 UTC (permalink / raw)
To: Stan Hoeppner; +Cc: xfs
On 2013.07.11 at 15:24 -0500, Stan Hoeppner wrote:
> On 7/11/2013 6:28 AM, Markus Trippelsdorf wrote:
> ...
> >> Looking at the source:
> >> http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
> >> it appears that one can set an environment variable KDE_EXTRA_FSYNC to
> >> address this issue.
> >>
> >> However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
> >> loose my KDE settings in case of a crash. So the whole fsync thing might
> >> be a red herring.
> >
> > It turned out that the KDE_EXTRA_FSYNC variable doesn't affect KDE
> > config file handling at all.
> > So I've added an fsync in kconfigini.cpp (KConfigIniBackend::writeConfig)
> > and now I don't loose my settings anymore during kernel crash testing.
> >
> > That is until xfs eats my KDE config files (kwinrulesr in this case):
>
> Adding fsync in kconfigini.cpp apparently doesn't force fsync for all
> KDE file operations. You also have some Open Office files getting hosed
> due to lack of fsync. XFS is not the cause of these problems.
>
> The problem is that all of this desktop code was developed atop EXT3
> which flushed to disk every 5 seconds. This made programmers sloppy as
> they didn't have to fsync to make sure data hit disk. This problem has
> been covered extensively by many, including Eric in other posts on his
> blog. There's a really simple way to test this: mount with sync.
> Report results after the next crash. If no files are corrupted then
> you've verified the problem lay squarely on the shoulders of these
> desktop developers who have abdicated their responsibility to make sure
> their file changes hit the disk, instead of relying on a broken
> filesystem do it for them.
>
> Worth noting, using EXT4 without the EXT3 flush emulation enabled will
> yield similar file corruption upon a crash.
I'm not so sure. Of course a journaled filesystem is not a database
replacement, but wouldn't it be easier to address this issue in xfs
directly instead of hoping in vain that application developers will
fix their code someday?
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 20:40 ` Markus Trippelsdorf
@ 2013-07-11 23:01 ` Stan Hoeppner
2013-07-12 2:38 ` Dave Chinner
1 sibling, 0 replies; 37+ messages in thread
From: Stan Hoeppner @ 2013-07-11 23:01 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: xfs
On 7/11/2013 3:40 PM, Markus Trippelsdorf wrote:
> On 2013.07.11 at 15:24 -0500, Stan Hoeppner wrote:
>> On 7/11/2013 6:28 AM, Markus Trippelsdorf wrote:
>> ...
>>>> Looking at the source:
>>>> http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
>>>> it appears that one can set an environment variable KDE_EXTRA_FSYNC to
>>>> address this issue.
>>>>
>>>> However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
>>>> loose my KDE settings in case of a crash. So the whole fsync thing might
>>>> be a red herring.
>>>
>>> It turned out that the KDE_EXTRA_FSYNC variable doesn't affect KDE
>>> config file handling at all.
>>> So I've added an fsync in kconfigini.cpp (KConfigIniBackend::writeConfig)
>>> and now I don't loose my settings anymore during kernel crash testing.
>>>
>>> That is until xfs eats my KDE config files (kwinrulesr in this case):
>>
>> Adding fsync in kconfigini.cpp apparently doesn't force fsync for all
>> KDE file operations. You also have some Open Office files getting hosed
>> due to lack of fsync. XFS is not the cause of these problems.
>>
>> The problem is that all of this desktop code was developed atop EXT3
>> which flushed to disk every 5 seconds. This made programmers sloppy as
>> they didn't have to fsync to make sure data hit disk. This problem has
>> been covered extensively by many, including Eric in other posts on his
>> blog. There's a really simple way to test this: mount with sync.
>> Report results after the next crash. If no files are corrupted then
>> you've verified the problem lay squarely on the shoulders of these
>> desktop developers who have abdicated their responsibility to make sure
>> their file changes hit the disk, instead of relying on a broken
>> filesystem do it for them.
>>
>> Worth noting, using EXT4 without the EXT3 flush emulation enabled will
>> yield similar file corruption upon a crash.
>
> I'm not so sure. Of course a journaled filesystem is not a database
> replacement, but wouldn't it be easier to address this issue in xfs
> directly instead of hoping in vain that application developers will
> fix their code someday?
Apparently you missed the O_PONIES debate some 4 years ago. Take a
guess as to what happens to XFS performance if it is modified to "fix"
this app dev broken file on crash problem. Hint: a bunch of previously
asynchronous operations must now to be forced to be synchronous.
--
Stan
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 9:07 ` Markus Trippelsdorf
2013-07-11 11:28 ` Markus Trippelsdorf
@ 2013-07-12 2:17 ` Dave Chinner
2013-07-12 7:07 ` Markus Trippelsdorf
1 sibling, 1 reply; 37+ messages in thread
From: Dave Chinner @ 2013-07-12 2:17 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: Stan Hoeppner, xfs
On Thu, Jul 11, 2013 at 11:07:55AM +0200, Markus Trippelsdorf wrote:
> On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> > On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> >
> > >> I was loosing my KDE settings bit by bit with every reboot during the
> > >> bisection. First my window-rules disappeared, then my desktop background
> > >> changed to default, then my taskbar moved from top to the bottom, etc.
> > >> In the end I had to restore all my .files from backup.
> > >
> > > That's not filesystem corruption. That sounds more like someone not
> > > using fsync in the apropriate place when overwriting a file....
> >
> > From Sandeen's blog, March 2009:
> >
> > "I dunno how to resolve this right now. I talked to some nice KDE folks
> > on irc; they basically want atomic writes, either you get your old file
> > or your new file post-crash; and tempfile/sync/rename does this – but
> > the fsync hurts on 78% of the Linux filesystems out there. So their
> > KSaveFile class doesn’t fsync. So what to do, what to do.."
> >
> > That's 4 years ago. Is it possible the KDE devs are still not using
> > fsync? Sure seems likely given Markus' problem.
>
> Looking at the source:
> http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
> it appears that one can set an environment variable KDE_EXTRA_FSYNC to
> address this issue.
>
> However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
> loose my KDE settings in case of a crash. So the whole fsync thing might
> be a red herring.
>
> What's more this time I endend up with undeletable files in /tmp (for
> example .X0-lock) after the crash:
>
> (/dev/sdb was mounted and unmounted normally before I ran xfs_repair)
>
> t@ubunt:~# xfs_repair /dev/sdb
> Phase 1 - find and verify superblock...
> Phase 2 - using internal log
> - zero log...
> - scan filesystem freespace and inode maps...
> agi unlinked bucket 0 is 683435008 in ag 2 (inode=4978402304)
> agi unlinked bucket 1 is 683435009 in ag 2 (inode=4978402305)
> - found root inode chunk
Again, these are signs that log recovery has not completed
successfully or that for some reason it thought the log was clean.
Can you please post the dmesg output after the crash when you go
through the mount/unmount process before you run xfs_repair?
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-11 20:40 ` Markus Trippelsdorf
2013-07-11 23:01 ` Stan Hoeppner
@ 2013-07-12 2:38 ` Dave Chinner
1 sibling, 0 replies; 37+ messages in thread
From: Dave Chinner @ 2013-07-12 2:38 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: Stan Hoeppner, xfs
On Thu, Jul 11, 2013 at 10:40:33PM +0200, Markus Trippelsdorf wrote:
> On 2013.07.11 at 15:24 -0500, Stan Hoeppner wrote:
> > On 7/11/2013 6:28 AM, Markus Trippelsdorf wrote:
> > ...
> > >> Looking at the source:
> > >> http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
> > >> it appears that one can set an environment variable KDE_EXTRA_FSYNC to
> > >> address this issue.
> > >>
> > >> However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
> > >> loose my KDE settings in case of a crash. So the whole fsync thing might
> > >> be a red herring.
> > >
> > > It turned out that the KDE_EXTRA_FSYNC variable doesn't affect KDE
> > > config file handling at all.
> > > So I've added an fsync in kconfigini.cpp (KConfigIniBackend::writeConfig)
> > > and now I don't loose my settings anymore during kernel crash testing.
> > >
> > > That is until xfs eats my KDE config files (kwinrulesr in this case):
> >
> > Adding fsync in kconfigini.cpp apparently doesn't force fsync for all
> > KDE file operations. You also have some Open Office files getting hosed
> > due to lack of fsync. XFS is not the cause of these problems.
> >
> > The problem is that all of this desktop code was developed atop EXT3
> > which flushed to disk every 5 seconds. This made programmers sloppy as
> > they didn't have to fsync to make sure data hit disk. This problem has
> > been covered extensively by many, including Eric in other posts on his
> > blog. There's a really simple way to test this: mount with sync.
> > Report results after the next crash. If no files are corrupted then
> > you've verified the problem lay squarely on the shoulders of these
> > desktop developers who have abdicated their responsibility to make sure
> > their file changes hit the disk, instead of relying on a broken
> > filesystem do it for them.
> >
> > Worth noting, using EXT4 without the EXT3 flush emulation enabled will
> > yield similar file corruption upon a crash.
>
> I'm not so sure. Of course a journaled filesystem is not a database
> replacement, but wouldn't it be easier to address this issue in xfs
> directly instead of hoping in vain that application developers will
> fix their code someday?
The problem is that there is a small minority of vocal users who
complain loudly and vigorously that something is slow when
application developers use proper caution and ensure files are
safely written using fsync. Those users yell and scream that they
care more about speed than they do about losing their config
settings on a crash, and demand the problem be fixed. Hence we end
up with special environment variables that nobody knows about that
try to provide some measure of data integrity. As you've found out,
it's not sufficient.
It's not up to the filesystem to enforce a "you must do everything
safely" policy. The filesystem provides mechanisms for users and
developers to decide if they want to be fast or safe. Unfortunately
for us, while XFS is pretty fast even when running in "safe" mode,
other filesystems aren't, and that's where the problem lies.
If you want everything to be safe, mount the filesystem with -o
sync. But it will be slow. The only way to be fast and safe is for
applications to Do The Right Thing - no hacks in the filesystem can
provide both fast and safe with compromising either fast or safe in
some manner for someone.
It's unfortunate that after several years of educating people to use
fsync when data integrity is important that we are seeing a
significant back-slide to trying to avoid fsync again. it appeared
recently on the ext4 list, when a gnome developer said they turned
off fsync because users were complaining, trying to rely on a side
effect of ext4 data=ordered mode for integrity and they failed and
users started reporting that they were losing files on crashes....
This is an application layer problem, not a filesystem layer problem.
The filesystems can provide mechanisms to try to help minimise the
impact of requiring data integrity operations, but we haven't been
able to get any significant set of userspace developers to agree on
a sane set of functionality that filesystems can provide over and
above what POSIX already gives them.
And besides, a filesystem can't fix the problems of applications
that use fsync to write inconsequential data that doesn't need
persistence across crashs. Thats clearly an application problem....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-12 2:17 ` Dave Chinner
@ 2013-07-12 7:07 ` Markus Trippelsdorf
2013-07-13 9:05 ` Markus Trippelsdorf
2013-07-15 2:28 ` Dave Chinner
0 siblings, 2 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-12 7:07 UTC (permalink / raw)
To: Dave Chinner; +Cc: Stan Hoeppner, xfs
On 2013.07.12 at 12:17 +1000, Dave Chinner wrote:
> On Thu, Jul 11, 2013 at 11:07:55AM +0200, Markus Trippelsdorf wrote:
> > On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> > > On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > > > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> > >
> > > >> I was loosing my KDE settings bit by bit with every reboot during the
> > > >> bisection. First my window-rules disappeared, then my desktop background
> > > >> changed to default, then my taskbar moved from top to the bottom, etc.
> > > >> In the end I had to restore all my .files from backup.
> > > >
> > > > That's not filesystem corruption. That sounds more like someone not
> > > > using fsync in the apropriate place when overwriting a file....
> > >
> > > From Sandeen's blog, March 2009:
> > >
> > > "I dunno how to resolve this right now. I talked to some nice KDE folks
> > > on irc; they basically want atomic writes, either you get your old file
> > > or your new file post-crash; and tempfile/sync/rename does this – but
> > > the fsync hurts on 78% of the Linux filesystems out there. So their
> > > KSaveFile class doesn’t fsync. So what to do, what to do.."
> > >
> > > That's 4 years ago. Is it possible the KDE devs are still not using
> > > fsync? Sure seems likely given Markus' problem.
> >
> > Looking at the source:
> > http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
> > it appears that one can set an environment variable KDE_EXTRA_FSYNC to
> > address this issue.
> >
> > However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
> > loose my KDE settings in case of a crash. So the whole fsync thing might
> > be a red herring.
> >
> > What's more this time I endend up with undeletable files in /tmp (for
> > example .X0-lock) after the crash:
> >
> > (/dev/sdb was mounted and unmounted normally before I ran xfs_repair)
> >
> > t@ubunt:~# xfs_repair /dev/sdb
> > Phase 1 - find and verify superblock...
> > Phase 2 - using internal log
> > - zero log...
> > - scan filesystem freespace and inode maps...
> > agi unlinked bucket 0 is 683435008 in ag 2 (inode=4978402304)
> > agi unlinked bucket 1 is 683435009 in ag 2 (inode=4978402305)
> > - found root inode chunk
>
> Again, these are signs that log recovery has not completed
> successfully or that for some reason it thought the log was clean.
> Can you please post the dmesg output after the crash when you go
> through the mount/unmount process before you run xfs_repair?
Sure.
First boot after crash:
XFS (sdb2): Mounting Filesystem
XFS (sdb2): Starting recovery (logdev: internal)
XFS (sdb2): Ending recovery (logdev: internal)
Second boot after crash:
XFS (sdb2): Mounting Filesystem
XFS (sdb2): Ending clean mount
I then boot Ubuntu from another disc to run xfs_repair.
And looking through my logs I see this WARNING:
------------[ cut here ]------------
WARNING: CPU: 0 PID: 439 at fs/inode.c:280 drop_nlink+0x33/0x40()
CPU: 0 PID: 439 Comm: gconfd-2 Not tainted 3.10.0-08982-g6d128e1-dirty #42
Hardware name: System manufacturer System Product Name/M4A78T-E, BIOS 3503 04/13/2011
0000000000000009 ffffffff8157d030 0000000000000000 ffffffff81060788
ffff8801f8608cc8 ffff880205998230 ffff8801f7bede58 0000000000000000
ffff8801f86083c0 ffffffff8110ce93 ffff8801f8608b40 ffffffff811b7104
Call Trace:
[<ffffffff8157d030>] ? dump_stack+0x41/0x51
[<ffffffff81060788>] ? warn_slowpath_common+0x68/0x80
[<ffffffff8110ce93>] ? drop_nlink+0x33/0x40
[<ffffffff811b7104>] ? xfs_droplink+0x24/0x60
[<ffffffff811b84ed>] ? xfs_remove+0x24d/0x380
[<ffffffff811b1657>] ? xfs_vn_unlink+0x37/0x80
[<ffffffff8110414e>] ? vfs_unlink+0x6e/0xe0
[<ffffffff8110432a>] ? do_unlinkat+0x16a/0x220
[<ffffffff810f4fa9>] ? SyS_faccessat+0x149/0x200
[<ffffffff81583292>] ? system_call_fastpath+0x16/0x1b
---[ end trace de5865b7c20ab8e4 ]---
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-12 7:07 ` Markus Trippelsdorf
@ 2013-07-13 9:05 ` Markus Trippelsdorf
2013-07-15 2:28 ` Dave Chinner
1 sibling, 0 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-13 9:05 UTC (permalink / raw)
To: Dave Chinner; +Cc: Stan Hoeppner, xfs
On 2013.07.12 at 09:07 +0200, Markus Trippelsdorf wrote:
> On 2013.07.12 at 12:17 +1000, Dave Chinner wrote:
> > On Thu, Jul 11, 2013 at 11:07:55AM +0200, Markus Trippelsdorf wrote:
> > > On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> > > > On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > > > > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> > > >
> > > > >> I was loosing my KDE settings bit by bit with every reboot during the
> > > > >> bisection. First my window-rules disappeared, then my desktop background
> > > > >> changed to default, then my taskbar moved from top to the bottom, etc.
> > > > >> In the end I had to restore all my .files from backup.
> > > > >
> > > > > That's not filesystem corruption. That sounds more like someone not
> > > > > using fsync in the apropriate place when overwriting a file....
> > > >
> > > > From Sandeen's blog, March 2009:
> > > >
> > > > "I dunno how to resolve this right now. I talked to some nice KDE folks
> > > > on irc; they basically want atomic writes, either you get your old file
> > > > or your new file post-crash; and tempfile/sync/rename does this – but
> > > > the fsync hurts on 78% of the Linux filesystems out there. So their
> > > > KSaveFile class doesn’t fsync. So what to do, what to do.."
> > > >
> > > > That's 4 years ago. Is it possible the KDE devs are still not using
> > > > fsync? Sure seems likely given Markus' problem.
> > >
> > > Looking at the source:
> > > http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
> > > it appears that one can set an environment variable KDE_EXTRA_FSYNC to
> > > address this issue.
> > >
> > > However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
> > > loose my KDE settings in case of a crash. So the whole fsync thing might
> > > be a red herring.
> > >
> > > What's more this time I endend up with undeletable files in /tmp (for
> > > example .X0-lock) after the crash:
> > >
> > > (/dev/sdb was mounted and unmounted normally before I ran xfs_repair)
> > >
> > > t@ubunt:~# xfs_repair /dev/sdb
> > > Phase 1 - find and verify superblock...
> > > Phase 2 - using internal log
> > > - zero log...
> > > - scan filesystem freespace and inode maps...
> > > agi unlinked bucket 0 is 683435008 in ag 2 (inode=4978402304)
> > > agi unlinked bucket 1 is 683435009 in ag 2 (inode=4978402305)
> > > - found root inode chunk
> >
> > Again, these are signs that log recovery has not completed
> > successfully or that for some reason it thought the log was clean.
> > Can you please post the dmesg output after the crash when you go
> > through the mount/unmount process before you run xfs_repair?
>
> Sure.
> First boot after crash:
> XFS (sdb2): Mounting Filesystem
> XFS (sdb2): Starting recovery (logdev: internal)
> XFS (sdb2): Ending recovery (logdev: internal)
Some further observations:
When I boot 3.2.0 after the crash log recovery works fine.
When I boot 3.9.0 after the crash I get the following:
[ 2.332989] XFS (sdc2): Mounting Filesystem
[ 2.406206] XFS (sdc2): Starting recovery (logdev: internal)
[ 2.418147] XFS (sdc2): log record CRC mismatch: found 0xdbcaef48, expected 0x69e7934e.
[ 2.432718] ffffc9000063e000: 00 00 00 01 00 00 00 00 69 01 00 00 32 d6 93 e5 ........i...2...
[ 2.440218] ffffc9000063e010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.448367] XFS (sdc2): log record CRC mismatch: found 0xaf1a53d, expected 0x38ec3424.
[ 2.463336] ffffc9000063e000: 00 00 00 01 00 00 00 00 69 01 00 00 9a d5 a8 e7 ........i.......
[ 2.470979] ffffc9000063e010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.479128] XFS (sdc2): log record CRC mismatch: found 0x8e2572f5, expected 0x7a48b5fb.
[ 2.482963] ffffc9000063e000: 00 00 00 01 00 00 00 00 69 01 00 00 be 27 a7 7a ........i....'.z
[ 2.484917] ffffc9000063e010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.487348] XFS (sdc2): log record CRC mismatch: found 0x96c174ce, expected 0x2e164f6f.
[ 2.491305] ffffc9000063e000: 00 00 00 01 00 00 00 00 69 01 00 00 fc 4a 96 e7 ........i....J..
[ 2.493334] ffffc9000063e010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.495923] XFS (sdc2): log record CRC mismatch: found 0x7faa3171, expected 0xff793468.
[ 2.499998] ffffc9000063e000: 00 00 00 01 00 00 00 00 69 01 00 00 6e 87 7d 90 ........i...n.}.
[ 2.502069] ffffc9000063e010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.504629] XFS (sdc2): log record CRC mismatch: found 0x52b46483, expected 0xc34c4ddd.
[ 2.508760] ffffc9000063e000: 00 00 00 01 00 00 00 00 69 01 00 00 7e 36 3f 2b ........i...~6?+
[ 2.510865] ffffc9000063e010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.513712] XFS (sdc2): log record CRC mismatch: found 0xdbcaef48, expected 0x69e7934e.
[ 2.517892] ffffc90000edf000: 00 00 00 01 00 00 00 00 69 01 00 00 32 d6 93 e5 ........i...2...
[ 2.520026] ffffc90000edf010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.526166] XFS (sdc2): log record CRC mismatch: found 0xaf1a53d, expected 0x38ec3424.
[ 2.530421] ffffc90000edf000: 00 00 00 01 00 00 00 00 69 01 00 00 9a d5 a8 e7 ........i.......
[ 2.532584] ffffc90000edf010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.539422] XFS (sdc2): log record CRC mismatch: found 0x8e2572f5, expected 0x7a48b5fb.
[ 2.544853] ffffc90000edf000: 00 00 00 01 00 00 00 00 69 01 00 00 be 27 a7 7a ........i....'.z
[ 2.547606] ffffc90000edf010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.560042] XFS (sdc2): log record CRC mismatch: found 0x96c174ce, expected 0x2e164f6f.
[ 2.577113] ffffc90000edf000: 00 00 00 01 00 00 00 00 69 01 00 00 fc 4a 96 e7 ........i....J..
[ 2.585729] ffffc90000edf010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.589138] usb 4-2: new full-speed USB device number 2 using ohci_hcd
[ 2.614466] XFS (sdc2): log record CRC mismatch: found 0x7faa3171, expected 0xff793468.
[ 2.625827] tsc: Refined TSC clocksource calibration: 3210.828 MHz
[ 2.625838] Switching to clocksource tsc
[ 2.648762] ffffc90000edf000: 00 00 00 01 00 00 00 00 69 01 00 00 6e 87 7d 90 ........i...n.}.
[ 2.657431] ffffc90000edf010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.673246] XFS (sdc2): log record CRC mismatch: found 0x52b46483, expected 0xc34c4ddd.
[ 2.691869] ffffc90000edf000: 00 00 00 01 00 00 00 00 69 01 00 00 7e 36 3f 2b ........i...~6?+
[ 2.701352] ffffc90000edf010: 00 00 00 10 69 00 00 00 4e 41 52 54 2a 00 00 00 ....i...NART*...
[ 2.714524] XFS (sdc2): Ending recovery (logdev: internal)
[ 2.723389] VFS: Mounted root (xfs filesystem) readonly on device 8:34.
[ 2.732808] devtmpfs: mounted
When I boot the current Linus tree after the crash log recovery fails silently.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-12 7:07 ` Markus Trippelsdorf
2013-07-13 9:05 ` Markus Trippelsdorf
@ 2013-07-15 2:28 ` Dave Chinner
2013-07-15 6:47 ` Markus Trippelsdorf
1 sibling, 1 reply; 37+ messages in thread
From: Dave Chinner @ 2013-07-15 2:28 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: Stan Hoeppner, xfs
On Fri, Jul 12, 2013 at 09:07:21AM +0200, Markus Trippelsdorf wrote:
> On 2013.07.12 at 12:17 +1000, Dave Chinner wrote:
> > On Thu, Jul 11, 2013 at 11:07:55AM +0200, Markus Trippelsdorf wrote:
> > > On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> > > > On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > > > > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> > > >
> > > > >> I was loosing my KDE settings bit by bit with every reboot during the
> > > > >> bisection. First my window-rules disappeared, then my desktop background
> > > > >> changed to default, then my taskbar moved from top to the bottom, etc.
> > > > >> In the end I had to restore all my .files from backup.
> > > > >
> > > > > That's not filesystem corruption. That sounds more like someone not
> > > > > using fsync in the apropriate place when overwriting a file....
> > > >
> > > > From Sandeen's blog, March 2009:
> > > >
> > > > "I dunno how to resolve this right now. I talked to some nice KDE folks
> > > > on irc; they basically want atomic writes, either you get your old file
> > > > or your new file post-crash; and tempfile/sync/rename does this – but
> > > > the fsync hurts on 78% of the Linux filesystems out there. So their
> > > > KSaveFile class doesn’t fsync. So what to do, what to do.."
> > > >
> > > > That's 4 years ago. Is it possible the KDE devs are still not using
> > > > fsync? Sure seems likely given Markus' problem.
> > >
> > > Looking at the source:
> > > http://api.kde.org/4.10-api/kdelibs-apidocs/kdecore/html/ksavefile_8cpp_source.html#l00219
> > > it appears that one can set an environment variable KDE_EXTRA_FSYNC to
> > > address this issue.
> > >
> > > However in my case it doesn't help. Even with KDE_EXTRA_FSYNC=1 I still
> > > loose my KDE settings in case of a crash. So the whole fsync thing might
> > > be a red herring.
> > >
> > > What's more this time I endend up with undeletable files in /tmp (for
> > > example .X0-lock) after the crash:
> > >
> > > (/dev/sdb was mounted and unmounted normally before I ran xfs_repair)
> > >
> > > t@ubunt:~# xfs_repair /dev/sdb
> > > Phase 1 - find and verify superblock...
> > > Phase 2 - using internal log
> > > - zero log...
> > > - scan filesystem freespace and inode maps...
> > > agi unlinked bucket 0 is 683435008 in ag 2 (inode=4978402304)
> > > agi unlinked bucket 1 is 683435009 in ag 2 (inode=4978402305)
> > > - found root inode chunk
> >
> > Again, these are signs that log recovery has not completed
> > successfully or that for some reason it thought the log was clean.
> > Can you please post the dmesg output after the crash when you go
> > through the mount/unmount process before you run xfs_repair?
>
> Sure.
> First boot after crash:
> XFS (sdb2): Mounting Filesystem
> XFS (sdb2): Starting recovery (logdev: internal)
> XFS (sdb2): Ending recovery (logdev: internal)
>
> Second boot after crash:
> XFS (sdb2): Mounting Filesystem
> XFS (sdb2): Ending clean mount
>
> I then boot Ubuntu from another disc to run xfs_repair.
That's what shoul dhave been in the initial description of your
problem.
> And looking through my logs I see this WARNING:
>
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 439 at fs/inode.c:280 drop_nlink+0x33/0x40()
> CPU: 0 PID: 439 Comm: gconfd-2 Not tainted 3.10.0-08982-g6d128e1-dirty #42
> Hardware name: System manufacturer System Product Name/M4A78T-E, BIOS 3503 04/13/2011
> 0000000000000009 ffffffff8157d030 0000000000000000 ffffffff81060788
> ffff8801f8608cc8 ffff880205998230 ffff8801f7bede58 0000000000000000
> ffff8801f86083c0 ffffffff8110ce93 ffff8801f8608b40 ffffffff811b7104
> Call Trace:
> [<ffffffff8157d030>] ? dump_stack+0x41/0x51
> [<ffffffff81060788>] ? warn_slowpath_common+0x68/0x80
> [<ffffffff8110ce93>] ? drop_nlink+0x33/0x40
> [<ffffffff811b7104>] ? xfs_droplink+0x24/0x60
> [<ffffffff811b84ed>] ? xfs_remove+0x24d/0x380
> [<ffffffff811b1657>] ? xfs_vn_unlink+0x37/0x80
> [<ffffffff8110414e>] ? vfs_unlink+0x6e/0xe0
> [<ffffffff8110432a>] ? do_unlinkat+0x16a/0x220
> [<ffffffff810f4fa9>] ? SyS_faccessat+0x149/0x200
> [<ffffffff81583292>] ? system_call_fastpath+0x16/0x1b
When did that occur? Before the crash, after the first/second mount?
after you ran repair?
> Some further observations:
>
> When I boot 3.2.0 after the crash log recovery works fine.
>
> When I boot 3.9.0 after the crash I get the following:
>
> [ 2.332989] XFS (sdc2): Mounting Filesystem
> [ 2.406206] XFS (sdc2): Starting recovery (logdev: internal)
> [ 2.418147] XFS (sdc2): log record CRC mismatch: found 0xdbcaef48, expected 0x69e7934e.
Just informational - indicating that the log records don't have
valid CRCs in them because 3.2 didn't calculate them. If you are
getting them when after a crash on a 3.9+ kernel, then there's a
problem writing to the log....
> When I boot the current Linus tree after the crash log recovery fails silently.
dmesg output, please. Indeed, what does "fails silently" mean? the
filesystem doesn't mount but no error is given?
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: Corruption of root fs during git bisect of drm system hang
2013-07-15 2:28 ` Dave Chinner
@ 2013-07-15 6:47 ` Markus Trippelsdorf
2013-07-19 12:22 ` [Bisected] " Markus Trippelsdorf
0 siblings, 1 reply; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-15 6:47 UTC (permalink / raw)
To: Dave Chinner; +Cc: Stan Hoeppner, xfs
On 2013.07.15 at 12:28 +1000, Dave Chinner wrote:
> On Fri, Jul 12, 2013 at 09:07:21AM +0200, Markus Trippelsdorf wrote:
> > On 2013.07.12 at 12:17 +1000, Dave Chinner wrote:
> > > On Thu, Jul 11, 2013 at 11:07:55AM +0200, Markus Trippelsdorf wrote:
> > > > On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> > > > > On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > > > > > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> > > > >
> > > > > >> I was loosing my KDE settings bit by bit with every reboot during the
> > > > > >> bisection. First my window-rules disappeared, then my desktop background
> > > > > >> changed to default, then my taskbar moved from top to the bottom, etc.
> > > > > >> In the end I had to restore all my .files from backup.
> > > > > >
> > > > > > That's not filesystem corruption. That sounds more like someone not
> > > > > > using fsync in the apropriate place when overwriting a file....
> > > > >
> > > > t@ubunt:~# xfs_repair /dev/sdb
> > > > Phase 1 - find and verify superblock...
> > > > Phase 2 - using internal log
> > > > - zero log...
> > > > - scan filesystem freespace and inode maps...
> > > > agi unlinked bucket 0 is 683435008 in ag 2 (inode=4978402304)
> > > > agi unlinked bucket 1 is 683435009 in ag 2 (inode=4978402305)
> > > > - found root inode chunk
> > >
> > > Again, these are signs that log recovery has not completed
> > > successfully or that for some reason it thought the log was clean.
> > > Can you please post the dmesg output after the crash when you go
> > > through the mount/unmount process before you run xfs_repair?
> >
> > Sure.
> > First boot after crash:
> > XFS (sdb2): Mounting Filesystem
> > XFS (sdb2): Starting recovery (logdev: internal)
> > XFS (sdb2): Ending recovery (logdev: internal)
> >
> > Second boot after crash:
> > XFS (sdb2): Mounting Filesystem
> > XFS (sdb2): Ending clean mount
> >
> > I then boot Ubuntu from another disc to run xfs_repair.
>
> That's what shoul dhave been in the initial description of your
> problem.
>
> > And looking through my logs I see this WARNING:
> >
> > ------------[ cut here ]------------
> > WARNING: CPU: 0 PID: 439 at fs/inode.c:280 drop_nlink+0x33/0x40()
> > CPU: 0 PID: 439 Comm: gconfd-2 Not tainted 3.10.0-08982-g6d128e1-dirty #42
> > Hardware name: System manufacturer System Product Name/M4A78T-E, BIOS 3503 04/13/2011
> > 0000000000000009 ffffffff8157d030 0000000000000000 ffffffff81060788
> > ffff8801f8608cc8 ffff880205998230 ffff8801f7bede58 0000000000000000
> > ffff8801f86083c0 ffffffff8110ce93 ffff8801f8608b40 ffffffff811b7104
> > Call Trace:
> > [<ffffffff8157d030>] ? dump_stack+0x41/0x51
> > [<ffffffff81060788>] ? warn_slowpath_common+0x68/0x80
> > [<ffffffff8110ce93>] ? drop_nlink+0x33/0x40
> > [<ffffffff811b7104>] ? xfs_droplink+0x24/0x60
> > [<ffffffff811b84ed>] ? xfs_remove+0x24d/0x380
> > [<ffffffff811b1657>] ? xfs_vn_unlink+0x37/0x80
> > [<ffffffff8110414e>] ? vfs_unlink+0x6e/0xe0
> > [<ffffffff8110432a>] ? do_unlinkat+0x16a/0x220
> > [<ffffffff810f4fa9>] ? SyS_faccessat+0x149/0x200
> > [<ffffffff81583292>] ? system_call_fastpath+0x16/0x1b
>
> When did that occur? Before the crash, after the first/second mount?
> after you ran repair?
After the first mount.
> > Some further observations:
> >
> > When I boot 3.2.0 after the crash log recovery works fine.
> >
> > When I boot 3.9.0 after the crash I get the following:
> >
> > [ 2.332989] XFS (sdc2): Mounting Filesystem
> > [ 2.406206] XFS (sdc2): Starting recovery (logdev: internal)
> > [ 2.418147] XFS (sdc2): log record CRC mismatch: found 0xdbcaef48, expected 0x69e7934e.
>
> Just informational - indicating that the log records don't have
> valid CRCs in them because 3.2 didn't calculate them. If you are
> getting them when after a crash on a 3.9+ kernel, then there's a
> problem writing to the log....
The crash always occurred on the current Linus tree kernel...
> > When I boot the current Linus tree after the crash log recovery fails silently.
>
> dmesg output, please. Indeed, what does "fails silently" mean? the
> filesystem doesn't mount but no error is given?
Again, there is no dmesg output. XFS tells me that it's "Ending recovery
(logdev: internal)" without any errors, when indeed it didn't recover
the log at all. It then mounts the filesystem normally (rw) in this
unclean state. That's when the WARNING I postend above happend.
The fact that when I boot 3.2.0 after the crash (that occurred running
the current Linus tree) log recovery works just fine, point to the new
CRC implementation as the reason for this bug.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-15 6:47 ` Markus Trippelsdorf
@ 2013-07-19 12:22 ` Markus Trippelsdorf
2013-07-19 12:41 ` Stefan Ring
` (2 more replies)
0 siblings, 3 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-19 12:22 UTC (permalink / raw)
To: Dave Chinner; +Cc: Ben Myers, Mark Tinguely, Stan Hoeppner, xfs
On 2013.07.15 at 08:47 +0200, Markus Trippelsdorf wrote:
> On 2013.07.15 at 12:28 +1000, Dave Chinner wrote:
> > On Fri, Jul 12, 2013 at 09:07:21AM +0200, Markus Trippelsdorf wrote:
> > > On 2013.07.12 at 12:17 +1000, Dave Chinner wrote:
> > > > On Thu, Jul 11, 2013 at 11:07:55AM +0200, Markus Trippelsdorf wrote:
> > > > > On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> > > > > > On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > > > > > > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> > > > > >
> > > > > > >> I was loosing my KDE settings bit by bit with every reboot during the
> > > > > > >> bisection. First my window-rules disappeared, then my desktop background
> > > > > > >> changed to default, then my taskbar moved from top to the bottom, etc.
> > > > > > >> In the end I had to restore all my .files from backup.
> > > > > > >
> > > > > > > That's not filesystem corruption. That sounds more like someone not
> > > > > > > using fsync in the apropriate place when overwriting a file....
> > > > > >
> > > > > t@ubunt:~# xfs_repair /dev/sdb
> > > > > Phase 1 - find and verify superblock...
> > > > > Phase 2 - using internal log
> > > > > - zero log...
> > > > > - scan filesystem freespace and inode maps...
> > > > > agi unlinked bucket 0 is 683435008 in ag 2 (inode=4978402304)
> > > > > agi unlinked bucket 1 is 683435009 in ag 2 (inode=4978402305)
> > > > > - found root inode chunk
> > > >
> > > > Again, these are signs that log recovery has not completed
> > > > successfully or that for some reason it thought the log was clean.
> > > > Can you please post the dmesg output after the crash when you go
> > > > through the mount/unmount process before you run xfs_repair?
> > >
> > > Sure.
> > > First boot after crash:
> > > XFS (sdb2): Mounting Filesystem
> > > XFS (sdb2): Starting recovery (logdev: internal)
> > > XFS (sdb2): Ending recovery (logdev: internal)
> > >
> > > Second boot after crash:
> > > XFS (sdb2): Mounting Filesystem
> > > XFS (sdb2): Ending clean mount
> > >
> > > I then boot Ubuntu from another disc to run xfs_repair.
> >
> > That's what shoul dhave been in the initial description of your
> > problem.
> >
> > > And looking through my logs I see this WARNING:
> > >
> > > ------------[ cut here ]------------
> > > WARNING: CPU: 0 PID: 439 at fs/inode.c:280 drop_nlink+0x33/0x40()
> > > CPU: 0 PID: 439 Comm: gconfd-2 Not tainted 3.10.0-08982-g6d128e1-dirty #42
> > > Hardware name: System manufacturer System Product Name/M4A78T-E, BIOS 3503 04/13/2011
> > > 0000000000000009 ffffffff8157d030 0000000000000000 ffffffff81060788
> > > ffff8801f8608cc8 ffff880205998230 ffff8801f7bede58 0000000000000000
> > > ffff8801f86083c0 ffffffff8110ce93 ffff8801f8608b40 ffffffff811b7104
> > > Call Trace:
> > > [<ffffffff8157d030>] ? dump_stack+0x41/0x51
> > > [<ffffffff81060788>] ? warn_slowpath_common+0x68/0x80
> > > [<ffffffff8110ce93>] ? drop_nlink+0x33/0x40
> > > [<ffffffff811b7104>] ? xfs_droplink+0x24/0x60
> > > [<ffffffff811b84ed>] ? xfs_remove+0x24d/0x380
> > > [<ffffffff811b1657>] ? xfs_vn_unlink+0x37/0x80
> > > [<ffffffff8110414e>] ? vfs_unlink+0x6e/0xe0
> > > [<ffffffff8110432a>] ? do_unlinkat+0x16a/0x220
> > > [<ffffffff810f4fa9>] ? SyS_faccessat+0x149/0x200
> > > [<ffffffff81583292>] ? system_call_fastpath+0x16/0x1b
> >
> > When did that occur? Before the crash, after the first/second mount?
> > after you ran repair?
>
> After the first mount.
>
> > > Some further observations:
> > >
> > > When I boot 3.2.0 after the crash log recovery works fine.
> > >
> > > When I boot 3.9.0 after the crash I get the following:
> > >
> > > [ 2.332989] XFS (sdc2): Mounting Filesystem
> > > [ 2.406206] XFS (sdc2): Starting recovery (logdev: internal)
> > > [ 2.418147] XFS (sdc2): log record CRC mismatch: found 0xdbcaef48, expected 0x69e7934e.
> >
> > Just informational - indicating that the log records don't have
> > valid CRCs in them because 3.2 didn't calculate them. If you are
> > getting them when after a crash on a 3.9+ kernel, then there's a
> > problem writing to the log....
>
> The crash always occurred on the current Linus tree kernel...
>
> > > When I boot the current Linus tree after the crash log recovery fails silently.
> >
> > dmesg output, please. Indeed, what does "fails silently" mean? the
> > filesystem doesn't mount but no error is given?
>
> Again, there is no dmesg output. XFS tells me that it's "Ending recovery
> (logdev: internal)" without any errors, when indeed it didn't recover
> the log at all. It then mounts the filesystem normally (rw) in this
> unclean state. That's when the WARNING I postend above happend.
I've bisected this issue to the following commit:
commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
Author: Dave Chinner <dchinner@redhat.com>
Date: Thu Jun 27 16:04:49 2013 +1000
xfs: don't do IO when creating an new inode
Reverting this commit on top of the Linus tree "solves" all problems for
me. IOW I no longer loose my KDE and LibreOffice config files during a
crash. Log recovery now works fine and xfs_repair shows no issues.
So users of 3.11.0-rc1 beware. Only run this version if you have
up-to-date backups handy.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 12:22 ` [Bisected] " Markus Trippelsdorf
@ 2013-07-19 12:41 ` Stefan Ring
2013-07-19 12:51 ` Markus Trippelsdorf
2013-07-19 21:11 ` Mark Tinguely
2013-07-20 1:48 ` Dave Chinner
2 siblings, 1 reply; 37+ messages in thread
From: Stefan Ring @ 2013-07-19 12:41 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: Ben Myers, Mark Tinguely, Stan Hoeppner, Linux fs XFS
> I've bisected this issue to the following commit:
>
> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> Author: Dave Chinner <dchinner@redhat.com>
> Date: Thu Jun 27 16:04:49 2013 +1000
>
> xfs: don't do IO when creating an new inode
>
> Reverting this commit on top of the Linus tree "solves" all problems for
> me. IOW I no longer loose my KDE and LibreOffice config files during a
> crash. Log recovery now works fine and xfs_repair shows no issues.
>
> So users of 3.11.0-rc1 beware. Only run this version if you have
> up-to-date backups handy.
What I miss in this thread is a distinction between filesystem
corruption on the one hand and a few zeroed files on the other. The
latter may be a nuisance, but it is expected behavior, while the
former should never happen, period, if I'm not mistaken.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 12:41 ` Stefan Ring
@ 2013-07-19 12:51 ` Markus Trippelsdorf
2013-07-19 16:02 ` Eric Sandeen
0 siblings, 1 reply; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-19 12:51 UTC (permalink / raw)
To: Stefan Ring; +Cc: Ben Myers, Mark Tinguely, Stan Hoeppner, Linux fs XFS
On 2013.07.19 at 14:41 +0200, Stefan Ring wrote:
> > I've bisected this issue to the following commit:
> >
> > commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> > Author: Dave Chinner <dchinner@redhat.com>
> > Date: Thu Jun 27 16:04:49 2013 +1000
> >
> > xfs: don't do IO when creating an new inode
> >
> > Reverting this commit on top of the Linus tree "solves" all problems for
> > me. IOW I no longer loose my KDE and LibreOffice config files during a
> > crash. Log recovery now works fine and xfs_repair shows no issues.
> >
> > So users of 3.11.0-rc1 beware. Only run this version if you have
> > up-to-date backups handy.
>
> What I miss in this thread is a distinction between filesystem
> corruption on the one hand and a few zeroed files on the other. The
> latter may be a nuisance, but it is expected behavior, while the
> former should never happen, period, if I'm not mistaken.
Well, it is natural that fs developers at first try to blame userspace.
Unfortunately it turned out that in this case there is filesystem
corruption. (Fortunately this normally happens only very rarely on rc1
kernels).
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 12:51 ` Markus Trippelsdorf
@ 2013-07-19 16:02 ` Eric Sandeen
2013-07-19 16:32 ` Markus Trippelsdorf
0 siblings, 1 reply; 37+ messages in thread
From: Eric Sandeen @ 2013-07-19 16:02 UTC (permalink / raw)
To: Markus Trippelsdorf
Cc: Stefan Ring, Ben Myers, Mark Tinguely, Stan Hoeppner, Linux fs XFS
On 7/19/13 7:51 AM, Markus Trippelsdorf wrote:
> On 2013.07.19 at 14:41 +0200, Stefan Ring wrote:
>>> I've bisected this issue to the following commit:
>>>
>>> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
>>> Author: Dave Chinner <dchinner@redhat.com>
>>> Date: Thu Jun 27 16:04:49 2013 +1000
>>>
>>> xfs: don't do IO when creating an new inode
>>>
>>> Reverting this commit on top of the Linus tree "solves" all problems for
>>> me. IOW I no longer loose my KDE and LibreOffice config files during a
>>> crash. Log recovery now works fine and xfs_repair shows no issues.
>>>
>>> So users of 3.11.0-rc1 beware. Only run this version if you have
>>> up-to-date backups handy.
Are you certain about that bisection point? All that does is
say: When we allocate a new inode, assign it a random generation
number, rather than reading it from disk & incrementing the
older generation number, AFAICS. So it simply avoids a read IO.
I wonder if simply changing IO patterns on the SSD changes how
it's doing caching & destaging <handwave>.
>> What I miss in this thread is a distinction between filesystem
>> corruption on the one hand and a few zeroed files on the other. The
>> latter may be a nuisance, but it is expected behavior, while the
>> former should never happen, period, if I'm not mistaken.
>
> Well, it is natural that fs developers at first try to blame userspace.
I disagree with that, we just need to be clear about your scenarios,
and what integrity guarantees should apply.
> Unfortunately it turned out that in this case there is filesystem
> corruption. (Fortunately this normally happens only very rarely on rc1
> kernels).
Corruption is when you get back data that you did not write,
or metadata which is inconsistent or unreadable even after a proper
log replay.
Corruption is _not_ unsynced, buffered data that was lost on a
crash or poweroff.
But I might not have followed the thread properly, and I might
misunderstand your situation.
When you experience this lost file [data] scenario, was it after an
orderly reboot, or after a crash and/or system reset?
-Eric
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 16:02 ` Eric Sandeen
@ 2013-07-19 16:32 ` Markus Trippelsdorf
2013-07-19 19:13 ` Ben Myers
2013-07-19 19:23 ` Eric Sandeen
0 siblings, 2 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-19 16:32 UTC (permalink / raw)
To: Eric Sandeen
Cc: Stefan Ring, Ben Myers, Mark Tinguely, Stan Hoeppner, Linux fs XFS
On 2013.07.19 at 11:02 -0500, Eric Sandeen wrote:
> On 7/19/13 7:51 AM, Markus Trippelsdorf wrote:
> > On 2013.07.19 at 14:41 +0200, Stefan Ring wrote:
> >>> I've bisected this issue to the following commit:
> >>>
> >>> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> >>> Author: Dave Chinner <dchinner@redhat.com>
> >>> Date: Thu Jun 27 16:04:49 2013 +1000
> >>>
> >>> xfs: don't do IO when creating an new inode
> >>>
> >>> Reverting this commit on top of the Linus tree "solves" all problems for
> >>> me. IOW I no longer loose my KDE and LibreOffice config files during a
> >>> crash. Log recovery now works fine and xfs_repair shows no issues.
> >>>
> >>> So users of 3.11.0-rc1 beware. Only run this version if you have
> >>> up-to-date backups handy.
>
> Are you certain about that bisection point? All that does is
> say: When we allocate a new inode, assign it a random generation
> number, rather than reading it from disk & incrementing the
> older generation number, AFAICS. So it simply avoids a read IO.
Yes, I'm sure.
As I wrote above I also double-checked by reverting the commit on top of
the current Linus tree.
> I wonder if simply changing IO patterns on the SSD changes how
> it's doing caching & destaging <handwave>.
No. The corruption also happens on my conventional (spinning) drives.
> >> What I miss in this thread is a distinction between filesystem
> >> corruption on the one hand and a few zeroed files on the other. The
> >> latter may be a nuisance, but it is expected behavior, while the
> >> former should never happen, period, if I'm not mistaken.
> >
> > Well, it is natural that fs developers at first try to blame userspace.
>
> I disagree with that, we just need to be clear about your scenarios,
> and what integrity guarantees should apply.
>
> > Unfortunately it turned out that in this case there is filesystem
> > corruption. (Fortunately this normally happens only very rarely on rc1
> > kernels).
>
> Corruption is when you get back data that you did not write,
> or metadata which is inconsistent or unreadable even after a proper
> log replay.
>
> Corruption is _not_ unsynced, buffered data that was lost on a
> crash or poweroff.
>
> But I might not have followed the thread properly, and I might
> misunderstand your situation.
>
> When you experience this lost file [data] scenario, was it after an
> orderly reboot, or after a crash and/or system reset?
To reproduce this issue simply boot into your desktop and then hit
sysrq-c and reboot. After log replay without error messages, the
filesystem is in an inconsistent state and many small config files are
lost. There are also undeletable files. You need to run xfs_repair
manually to bring the filesystem back to normal.
When cca9f93a52d is reverted, you don't loose your config files and the
filesystem is OK after log replay. xfs_repair reports no issues at all.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 16:32 ` Markus Trippelsdorf
@ 2013-07-19 19:13 ` Ben Myers
2013-07-19 19:56 ` Markus Trippelsdorf
2013-07-19 19:23 ` Eric Sandeen
1 sibling, 1 reply; 37+ messages in thread
From: Ben Myers @ 2013-07-19 19:13 UTC (permalink / raw)
To: Markus Trippelsdorf
Cc: Stefan Ring, Eric Sandeen, Mark Tinguely, Stan Hoeppner, Linux fs XFS
Hey Markus,
On Fri, Jul 19, 2013 at 06:32:20PM +0200, Markus Trippelsdorf wrote:
> On 2013.07.19 at 11:02 -0500, Eric Sandeen wrote:
> > > Unfortunately it turned out that in this case there is filesystem
> > > corruption. (Fortunately this normally happens only very rarely on rc1
> > > kernels).
> >
> > Corruption is when you get back data that you did not write,
> > or metadata which is inconsistent or unreadable even after a proper
> > log replay.
> >
> > Corruption is _not_ unsynced, buffered data that was lost on a
> > crash or poweroff.
> >
> > But I might not have followed the thread properly, and I might
> > misunderstand your situation.
> >
> > When you experience this lost file [data] scenario, was it after an
> > orderly reboot, or after a crash and/or system reset?
>
> To reproduce this issue simply boot into your desktop and then hit
> sysrq-c and reboot. After log replay without error messages, the
> filesystem is in an inconsistent state and many small config files are
> lost. There are also undeletable files. You need to run xfs_repair
> manually to bring the filesystem back to normal.
>
> When cca9f93a52d is reverted, you don't loose your config files and the
> filesystem is OK after log replay. xfs_repair reports no issues at all.
I'm a bit late to the party, but I wanted to give this a try.
On the machine I tried, I was not able to reproduce any corruption with a
echo b > /proc/sysrq-trigger
xfs_repair -n found no problems at all. I'll try it on a few more.
Could you post some of your latest xfs_repair output? And, have you been able
to reproduce this on more than one machine? I may have missed that detail
earlier in the thread.
Thanks much,
Ben
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 16:32 ` Markus Trippelsdorf
2013-07-19 19:13 ` Ben Myers
@ 2013-07-19 19:23 ` Eric Sandeen
2013-07-19 19:53 ` Markus Trippelsdorf
1 sibling, 1 reply; 37+ messages in thread
From: Eric Sandeen @ 2013-07-19 19:23 UTC (permalink / raw)
To: Markus Trippelsdorf
Cc: Stefan Ring, Ben Myers, Mark Tinguely, Stan Hoeppner, Linux fs XFS
On 7/19/13 11:32 AM, Markus Trippelsdorf wrote:
> On 2013.07.19 at 11:02 -0500, Eric Sandeen wrote:
>> On 7/19/13 7:51 AM, Markus Trippelsdorf wrote:
>>> On 2013.07.19 at 14:41 +0200, Stefan Ring wrote:
>>>>> I've bisected this issue to the following commit:
>>>>>
>>>>> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
>>>>> Author: Dave Chinner <dchinner@redhat.com>
>>>>> Date: Thu Jun 27 16:04:49 2013 +1000
>>>>>
>>>>> xfs: don't do IO when creating an new inode
>>>>>
>>>>> Reverting this commit on top of the Linus tree "solves" all problems for
>>>>> me. IOW I no longer loose my KDE and LibreOffice config files during a
>>>>> crash. Log recovery now works fine and xfs_repair shows no issues.
>>>>>
>>>>> So users of 3.11.0-rc1 beware. Only run this version if you have
>>>>> up-to-date backups handy.
>>
>> Are you certain about that bisection point? All that does is
>> say: When we allocate a new inode, assign it a random generation
>> number, rather than reading it from disk & incrementing the
>> older generation number, AFAICS. So it simply avoids a read IO.
>
> Yes, I'm sure.
> As I wrote above I also double-checked by reverting the commit on top of
> the current Linus tree.
>
>> I wonder if simply changing IO patterns on the SSD changes how
>> it's doing caching & destaging <handwave>.
>
> No. The corruption also happens on my conventional (spinning) drives.
>
>>>> What I miss in this thread is a distinction between filesystem
>>>> corruption on the one hand and a few zeroed files on the other. The
>>>> latter may be a nuisance, but it is expected behavior, while the
>>>> former should never happen, period, if I'm not mistaken.
>>>
>>> Well, it is natural that fs developers at first try to blame userspace.
>>
>> I disagree with that, we just need to be clear about your scenarios,
>> and what integrity guarantees should apply.
>>
>>> Unfortunately it turned out that in this case there is filesystem
>>> corruption. (Fortunately this normally happens only very rarely on rc1
>>> kernels).
>>
>> Corruption is when you get back data that you did not write,
>> or metadata which is inconsistent or unreadable even after a proper
>> log replay.
>>
>> Corruption is _not_ unsynced, buffered data that was lost on a
>> crash or poweroff.
>>
>> But I might not have followed the thread properly, and I might
>> misunderstand your situation.
>>
>> When you experience this lost file [data] scenario, was it after an
>> orderly reboot, or after a crash and/or system reset?
>
> To reproduce this issue simply boot into your desktop and then hit
> sysrq-c and reboot.
Ok, a crash, so at a minimum, some buffered data loss is 100% expected.
> After log replay without error messages, the
> filesystem is in an inconsistent state
What exactly do you mean by inconsistent state? Sorry to be pedantic here.
> and many small config files are
> lost.
Written how long ago? Were they fsynced?
I suppose you are unsure about that, if they're app-written.
> There are also undeletable files.
What happens when you try to delete them?
> You need to run xfs_repair
> manually to bring the filesystem back to normal.
And what is the repair output?
Can you show an exact sequence of events, capturing all relevant output from repair and/or dmesg, etc, just so we see exactly what you see?
Thanks,
-Eric
> When cca9f93a52d is reverted, you don't loose your config files and the
> filesystem is OK after log replay. xfs_repair reports no issues at all.
>
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 19:23 ` Eric Sandeen
@ 2013-07-19 19:53 ` Markus Trippelsdorf
0 siblings, 0 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-19 19:53 UTC (permalink / raw)
To: Eric Sandeen
Cc: Stefan Ring, Ben Myers, Mark Tinguely, Stan Hoeppner, Linux fs XFS
On 2013.07.19 at 14:23 -0500, Eric Sandeen wrote:
> On 7/19/13 11:32 AM, Markus Trippelsdorf wrote:
> > On 2013.07.19 at 11:02 -0500, Eric Sandeen wrote:
> >> On 7/19/13 7:51 AM, Markus Trippelsdorf wrote:
> >>> On 2013.07.19 at 14:41 +0200, Stefan Ring wrote:
> >>>>> I've bisected this issue to the following commit:
> >>>>>
> >>>>> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> >>>>> Author: Dave Chinner <dchinner@redhat.com>
> >>>>> Date: Thu Jun 27 16:04:49 2013 +1000
> >>>>>
> >>>>> xfs: don't do IO when creating an new inode
> >>>>>
> >>>>> Reverting this commit on top of the Linus tree "solves" all problems for
> >>>>> me. IOW I no longer loose my KDE and LibreOffice config files during a
> >>>>> crash. Log recovery now works fine and xfs_repair shows no issues.
> >>>>>
> >>>>> So users of 3.11.0-rc1 beware. Only run this version if you have
> >>>>> up-to-date backups handy.
> >>
> >> Are you certain about that bisection point? All that does is
> >> say: When we allocate a new inode, assign it a random generation
> >> number, rather than reading it from disk & incrementing the
> >> older generation number, AFAICS. So it simply avoids a read IO.
> >
> > Yes, I'm sure.
> > As I wrote above I also double-checked by reverting the commit on top of
> > the current Linus tree.
> >
> >> I wonder if simply changing IO patterns on the SSD changes how
> >> it's doing caching & destaging <handwave>.
> >
> > No. The corruption also happens on my conventional (spinning) drives.
> >
> >>>> What I miss in this thread is a distinction between filesystem
> >>>> corruption on the one hand and a few zeroed files on the other. The
> >>>> latter may be a nuisance, but it is expected behavior, while the
> >>>> former should never happen, period, if I'm not mistaken.
> >>>
> >>> Well, it is natural that fs developers at first try to blame userspace.
> >>
> >> I disagree with that, we just need to be clear about your scenarios,
> >> and what integrity guarantees should apply.
> >>
> >>> Unfortunately it turned out that in this case there is filesystem
> >>> corruption. (Fortunately this normally happens only very rarely on rc1
> >>> kernels).
> >>
> >> Corruption is when you get back data that you did not write,
> >> or metadata which is inconsistent or unreadable even after a proper
> >> log replay.
> >>
> >> Corruption is _not_ unsynced, buffered data that was lost on a
> >> crash or poweroff.
> >>
> >> But I might not have followed the thread properly, and I might
> >> misunderstand your situation.
> >>
> >> When you experience this lost file [data] scenario, was it after an
> >> orderly reboot, or after a crash and/or system reset?
> >
> > To reproduce this issue simply boot into your desktop and then hit
> > sysrq-c and reboot.
>
> Ok, a crash, so at a minimum, some buffered data loss is 100% expected.
Sure.
> > After log replay without error messages, the
> > filesystem is in an inconsistent state
>
> What exactly do you mean by inconsistent state? Sorry to be pedantic here.
By inconsistent state I mean a filesystem state that forces you to run
xfs_repair to get back to normal.
> > and many small config files are
> > lost.
>
> Written how long ago? Were they fsynced?
> I suppose you are unsure about that, if they're app-written.
I hit sysrq-c ~10 seconds after the KDE session is fully functional.
As I've wrote above I added an fsync to the KDE config file handler. So
the files should be fsynced.
> > There are also undeletable files.
>
> What happens when you try to delete them?
They show up as "?????? ??????" in "ls -l" and I get an error when I
try to delete them. (I don't recall the exact error message)
See for example the /tmp/.X0-lock file that I mentioned earlier in this
thread.
> > You need to run xfs_repair
> > manually to bring the filesystem back to normal.
>
> And what is the repair output?
See the outputs I've posted in this thread before. It's always a
variation thereof.
> Can you show an exact sequence of events, capturing all relevant output from repair and/or dmesg, etc, just so we see exactly what you see?
I already did that.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 19:13 ` Ben Myers
@ 2013-07-19 19:56 ` Markus Trippelsdorf
2013-07-19 20:28 ` Markus Trippelsdorf
0 siblings, 1 reply; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-19 19:56 UTC (permalink / raw)
To: Ben Myers
Cc: Stefan Ring, Eric Sandeen, Mark Tinguely, Stan Hoeppner, Linux fs XFS
On 2013.07.19 at 14:13 -0500, Ben Myers wrote:
> Hey Markus,
>
> On Fri, Jul 19, 2013 at 06:32:20PM +0200, Markus Trippelsdorf wrote:
> > On 2013.07.19 at 11:02 -0500, Eric Sandeen wrote:
> > > > Unfortunately it turned out that in this case there is filesystem
> > > > corruption. (Fortunately this normally happens only very rarely on rc1
> > > > kernels).
> > >
> > > Corruption is when you get back data that you did not write,
> > > or metadata which is inconsistent or unreadable even after a proper
> > > log replay.
> > >
> > > Corruption is _not_ unsynced, buffered data that was lost on a
> > > crash or poweroff.
> > >
> > > But I might not have followed the thread properly, and I might
> > > misunderstand your situation.
> > >
> > > When you experience this lost file [data] scenario, was it after an
> > > orderly reboot, or after a crash and/or system reset?
> >
> > To reproduce this issue simply boot into your desktop and then hit
> > sysrq-c and reboot. After log replay without error messages, the
> > filesystem is in an inconsistent state and many small config files are
> > lost. There are also undeletable files. You need to run xfs_repair
> > manually to bring the filesystem back to normal.
> >
> > When cca9f93a52d is reverted, you don't loose your config files and the
> > filesystem is OK after log replay. xfs_repair reports no issues at all.
>
> I'm a bit late to the party, but I wanted to give this a try.
>
> On the machine I tried, I was not able to reproduce any corruption with a
>
> echo b > /proc/sysrq-trigger
>
> xfs_repair -n found no problems at all. I'll try it on a few more.
>
> Could you post some of your latest xfs_repair output? And, have you been able
> to reproduce this on more than one machine? I may have missed that detail
> earlier in the thread.
I didn't save the xfs_repair output on every run. See the examples that
I've posted in this thread before.
And the issue always happens on the same machine.
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 19:56 ` Markus Trippelsdorf
@ 2013-07-19 20:28 ` Markus Trippelsdorf
0 siblings, 0 replies; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-19 20:28 UTC (permalink / raw)
To: Ben Myers
Cc: Stefan Ring, Eric Sandeen, Mark Tinguely, Stan Hoeppner, Linux fs XFS
On 2013.07.19 at 21:56 +0200, Markus Trippelsdorf wrote:
> On 2013.07.19 at 14:13 -0500, Ben Myers wrote:
> > Hey Markus,
> >
> > On Fri, Jul 19, 2013 at 06:32:20PM +0200, Markus Trippelsdorf wrote:
> > > On 2013.07.19 at 11:02 -0500, Eric Sandeen wrote:
> > > > > Unfortunately it turned out that in this case there is filesystem
> > > > > corruption. (Fortunately this normally happens only very rarely on rc1
> > > > > kernels).
> > > >
> > > > Corruption is when you get back data that you did not write,
> > > > or metadata which is inconsistent or unreadable even after a proper
> > > > log replay.
> > > >
> > > > Corruption is _not_ unsynced, buffered data that was lost on a
> > > > crash or poweroff.
> > > >
> > > > But I might not have followed the thread properly, and I might
> > > > misunderstand your situation.
> > > >
> > > > When you experience this lost file [data] scenario, was it after an
> > > > orderly reboot, or after a crash and/or system reset?
> > >
> > > To reproduce this issue simply boot into your desktop and then hit
> > > sysrq-c and reboot. After log replay without error messages, the
> > > filesystem is in an inconsistent state and many small config files are
> > > lost. There are also undeletable files. You need to run xfs_repair
> > > manually to bring the filesystem back to normal.
> > >
> > > When cca9f93a52d is reverted, you don't loose your config files and the
> > > filesystem is OK after log replay. xfs_repair reports no issues at all.
> >
> > I'm a bit late to the party, but I wanted to give this a try.
> >
> > On the machine I tried, I was not able to reproduce any corruption with a
> >
> > echo b > /proc/sysrq-trigger
> >
> > xfs_repair -n found no problems at all. I'll try it on a few more.
> >
> > Could you post some of your latest xfs_repair output? And, have you been able
> > to reproduce this on more than one machine? I may have missed that detail
> > earlier in the thread.
>
> I didn't save the xfs_repair output on every run. See the examples that
> I've posted in this thread before.
> And the issue always happens on the same machine.
To add some more info I'll attach the full dmseg of my machine.
I/O scheduler is deadline on both disks.
[ 0.000000] Linux version 3.11.0-rc1-00181-gb8a33fc-dirty (markus@x4) (gcc version 4.8.1 (GCC) ) #87 SMP Fri Jul 19 19:42:54 CEST 2013
[ 0.000000] Command line: root=PARTUUID=F61ADF02-9A53-485C-9BD4-3DD2F964C27C init=/sbin/minit rootflags=logbsize=262144 fbcon=rotate:3 radeon.dpm=1 drm_kms_helper.poll=0 quiet
[ 0.000000] KERNEL supported cpus:
[ 0.000000] AMD AuthenticAMD
[ 0.000000] e820: BIOS-provided physical RAM map:
[ 0.000000] BIOS-e820: [mem 0x0000000000000100-0x000000000009fbff] usable
[ 0.000000] BIOS-e820: [mem 0x000000000009fc00-0x000000000009ffff] reserved
[ 0.000000] BIOS-e820: [mem 0x00000000000e6000-0x00000000000fffff] reserved
[ 0.000000] BIOS-e820: [mem 0x0000000000100000-0x00000000dfe8ffff] usable
[ 0.000000] BIOS-e820: [mem 0x00000000dfe90000-0x00000000dfea7fff] ACPI data
[ 0.000000] BIOS-e820: [mem 0x00000000dfea8000-0x00000000dfecffff] ACPI NVS
[ 0.000000] BIOS-e820: [mem 0x00000000dfed0000-0x00000000dfefffff] reserved
[ 0.000000] BIOS-e820: [mem 0x00000000fff00000-0x00000000ffffffff] reserved
[ 0.000000] BIOS-e820: [mem 0x0000000100000000-0x000000021fffffff] usable
[ 0.000000] NX (Execute Disable) protection: active
[ 0.000000] SMBIOS 2.5 present.
[ 0.000000] DMI: System manufacturer System Product Name/M4A78T-E, BIOS 3503 04/13/2011
[ 0.000000] e820: update [mem 0x00000000-0x00000fff] usable ==> reserved
[ 0.000000] e820: remove [mem 0x000a0000-0x000fffff] usable
[ 0.000000] e820: last_pfn = 0x220000 max_arch_pfn = 0x400000000
[ 0.000000] MTRR default type: uncachable
[ 0.000000] MTRR fixed ranges enabled:
[ 0.000000] 00000-9FFFF write-back
[ 0.000000] A0000-EFFFF uncachable
[ 0.000000] F0000-FFFFF write-protect
[ 0.000000] MTRR variable ranges enabled:
[ 0.000000] 0 base 000000000000 mask FFFF80000000 write-back
[ 0.000000] 1 base 000080000000 mask FFFFC0000000 write-back
[ 0.000000] 2 base 0000C0000000 mask FFFFE0000000 write-back
[ 0.000000] 3 disabled
[ 0.000000] 4 disabled
[ 0.000000] 5 disabled
[ 0.000000] 6 disabled
[ 0.000000] 7 disabled
[ 0.000000] TOM2: 0000000220000000 aka 8704M
[ 0.000000] x86 PAT enabled: cpu 0, old 0x7010600070106, new 0x7010600070106
[ 0.000000] e820: update [mem 0xe0000000-0xffffffff] usable ==> reserved
[ 0.000000] e820: last_pfn = 0xdfe90 max_arch_pfn = 0x400000000
[ 0.000000] Base memory trampoline at [ffff880000099000] 99000 size 24576
[ 0.000000] Using GB pages for direct mapping
[ 0.000000] init_memory_mapping: [mem 0x00000000-0x000fffff]
[ 0.000000] [mem 0x00000000-0x000fffff] page 4k
[ 0.000000] BRK [0x01bb6000, 0x01bb6fff] PGTABLE
[ 0.000000] BRK [0x01bb7000, 0x01bb7fff] PGTABLE
[ 0.000000] BRK [0x01bb8000, 0x01bb8fff] PGTABLE
[ 0.000000] init_memory_mapping: [mem 0x21fe00000-0x21fffffff]
[ 0.000000] [mem 0x21fe00000-0x21fffffff] page 2M
[ 0.000000] BRK [0x01bb9000, 0x01bb9fff] PGTABLE
[ 0.000000] init_memory_mapping: [mem 0x21c000000-0x21fdfffff]
[ 0.000000] [mem 0x21c000000-0x21fdfffff] page 2M
[ 0.000000] init_memory_mapping: [mem 0x200000000-0x21bffffff]
[ 0.000000] [mem 0x200000000-0x21bffffff] page 2M
[ 0.000000] init_memory_mapping: [mem 0x00100000-0xdfe8ffff]
[ 0.000000] [mem 0x00100000-0x001fffff] page 4k
[ 0.000000] [mem 0x00200000-0x3fffffff] page 2M
[ 0.000000] [mem 0x40000000-0xbfffffff] page 1G
[ 0.000000] [mem 0xc0000000-0xdfdfffff] page 2M
[ 0.000000] [mem 0xdfe00000-0xdfe8ffff] page 4k
[ 0.000000] init_memory_mapping: [mem 0x100000000-0x1ffffffff]
[ 0.000000] [mem 0x100000000-0x1ffffffff] page 1G
[ 0.000000] ACPI: RSDP 00000000000fb880 00024 (v02 ACPIAM)
[ 0.000000] ACPI: XSDT 00000000dfe90100 0005C (v01 041311 XSDT1656 20110413 MSFT 00000097)
[ 0.000000] ACPI: FACP 00000000dfe90290 000F4 (v03 041311 FACP1656 20110413 MSFT 00000097)
[ 0.000000] ACPI BIOS Warning (bug): Optional FADT field Pm2ControlBlock has zero address or length: 0x0000000000000000/0x1 (20130517/tbfadt-603)
[ 0.000000] ACPI: DSDT 00000000dfe90450 0E6FE (v01 A1152 A1152000 00000000 INTL 20060113)
[ 0.000000] ACPI: FACS 00000000dfea8000 00040
[ 0.000000] ACPI: APIC 00000000dfe90390 0007C (v01 041311 APIC1656 20110413 MSFT 00000097)
[ 0.000000] ACPI: MCFG 00000000dfe90410 0003C (v01 041311 OEMMCFG 20110413 MSFT 00000097)
[ 0.000000] ACPI: OEMB 00000000dfea8040 00072 (v01 041311 OEMB1656 20110413 MSFT 00000097)
[ 0.000000] ACPI: SRAT 00000000dfe9f450 000E8 (v01 AMD FAM_F_10 00000002 AMD 00000001)
[ 0.000000] ACPI: HPET 00000000dfe9f540 00038 (v01 041311 OEMHPET 20110413 MSFT 00000097)
[ 0.000000] ACPI: SSDT 00000000dfe9f580 0088C (v01 A M I POWERNOW 00000001 AMD 00000001)
[ 0.000000] ACPI: Local APIC address 0xfee00000
[ 0.000000] [ffffea0000000000-ffffea00087fffff] PMD -> [ffff880217600000-ffff88021f5fffff] on node 0
[ 0.000000] Zone ranges:
[ 0.000000] DMA [mem 0x00001000-0x00ffffff]
[ 0.000000] DMA32 [mem 0x01000000-0xffffffff]
[ 0.000000] Normal [mem 0x100000000-0x21fffffff]
[ 0.000000] Movable zone start for each node
[ 0.000000] Early memory node ranges
[ 0.000000] node 0: [mem 0x00001000-0x0009efff]
[ 0.000000] node 0: [mem 0x00100000-0xdfe8ffff]
[ 0.000000] node 0: [mem 0x100000000-0x21fffffff]
[ 0.000000] On node 0 totalpages: 2096686
[ 0.000000] DMA zone: 64 pages used for memmap
[ 0.000000] DMA zone: 21 pages reserved
[ 0.000000] DMA zone: 3998 pages, LIFO batch:0
[ 0.000000] DMA32 zone: 14267 pages used for memmap
[ 0.000000] DMA32 zone: 913040 pages, LIFO batch:31
[ 0.000000] Normal zone: 18432 pages used for memmap
[ 0.000000] Normal zone: 1179648 pages, LIFO batch:31
[ 0.000000] ACPI: PM-Timer IO Port: 0x808
[ 0.000000] ACPI: Local APIC address 0xfee00000
[ 0.000000] ACPI: LAPIC (acpi_id[0x01] lapic_id[0x00] enabled)
[ 0.000000] ACPI: LAPIC (acpi_id[0x02] lapic_id[0x01] enabled)
[ 0.000000] ACPI: LAPIC (acpi_id[0x03] lapic_id[0x02] enabled)
[ 0.000000] ACPI: LAPIC (acpi_id[0x04] lapic_id[0x03] enabled)
[ 0.000000] ACPI: LAPIC (acpi_id[0x05] lapic_id[0x84] disabled)
[ 0.000000] ACPI: LAPIC (acpi_id[0x06] lapic_id[0x85] disabled)
[ 0.000000] ACPI: IOAPIC (id[0x04] address[0xfec00000] gsi_base[0])
[ 0.000000] IOAPIC[0]: apic_id 4, version 33, address 0xfec00000, GSI 0-23
[ 0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
[ 0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 low level)
[ 0.000000] ACPI: IRQ0 used by override.
[ 0.000000] ACPI: IRQ2 used by override.
[ 0.000000] ACPI: IRQ9 used by override.
[ 0.000000] Using ACPI (MADT) for SMP configuration information
[ 0.000000] ACPI: HPET id: 0x8300 base: 0xfed00000
[ 0.000000] smpboot: Allowing 4 CPUs, 0 hotplug CPUs
[ 0.000000] nr_irqs_gsi: 40
[ 0.000000] e820: [mem 0xdff00000-0xffefffff] available for PCI devices
[ 0.000000] setup_percpu: NR_CPUS:4 nr_cpumask_bits:4 nr_cpu_ids:4 nr_node_ids:1
[ 0.000000] PERCPU: Embedded 24 pages/cpu @ffff88021fc00000 s75328 r0 d22976 u524288
[ 0.000000] pcpu-alloc: s75328 r0 d22976 u524288 alloc=1*2097152
[ 0.000000] pcpu-alloc: [0] 0 1 2 3
[ 0.000000] Built 1 zonelists in Zone order, mobility grouping on. Total pages: 2063902
[ 0.000000] Kernel command line: root=PARTUUID=F61ADF02-9A53-485C-9BD4-3DD2F964C27C init=/sbin/minit rootflags=logbsize=262144 fbcon=rotate:3 radeon.dpm=1 drm_kms_helper.poll=0 quiet
[ 0.000000] PID hash table entries: 4096 (order: 3, 32768 bytes)
[ 0.000000] Dentry cache hash table entries: 1048576 (order: 11, 8388608 bytes)
[ 0.000000] Inode-cache hash table entries: 524288 (order: 10, 4194304 bytes)
[ 0.000000] Memory: 8164696K/8386744K available (5654K kernel code, 434K rwdata, 2068K rodata, 688K init, 600K bss, 222048K reserved)
[ 0.000000] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=4, Nodes=1
[ 0.000000] Hierarchical RCU implementation.
[ 0.000000] NR_IRQS:4352 nr_irqs:712 16
[ 0.000000] Console: colour VGA+ 80x25
[ 0.000000] console [tty0] enabled
[ 0.000000] hpet clockevent registered
[ 0.000000] tsc: Fast TSC calibration using PIT
[ 0.003333] tsc: Detected 3210.683 MHz processor
[ 0.000002] Calibrating delay loop (skipped), value calculated using timer frequency.. 6423.92 BogoMIPS (lpj=10702276)
[ 0.000003] pid_max: default: 32768 minimum: 301
[ 0.000026] Mount-cache hash table entries: 256
[ 0.000148] tseg: 0000000000
[ 0.000152] CPU: Physical Processor ID: 0
[ 0.000153] CPU: Processor Core ID: 3
[ 0.000154] mce: CPU supports 6 MCE banks
[ 0.000157] LVT offset 0 assigned for vector 0xf9
[ 0.000160] process: using AMD E400 aware idle routine
[ 0.000161] Last level iTLB entries: 4KB 512, 2MB 16, 4MB 8
Last level dTLB entries: 4KB 512, 2MB 128, 4MB 64
tlb_flushall_shift: 4
[ 0.000217] Freeing SMP alternatives memory: 20K (ffffffff81b1a000 - ffffffff81b1f000)
[ 0.000218] ACPI: Core revision 20130517
[ 0.002779] ACPI: All ACPI Tables successfully acquired
[ 0.003164] ..TIMER: vector=0x30 apic1=0 pin1=2 apic2=-1 pin2=-1
[ 0.036174] smpboot: CPU0: AMD Phenom(tm) II X4 955 Processor (fam: 10, model: 04, stepping: 02)
[ 0.142539] Performance Events: AMD PMU driver.
[ 0.142542] ... version: 0
[ 0.142542] ... bit width: 48
[ 0.142543] ... generic registers: 4
[ 0.142543] ... value mask: 0000ffffffffffff
[ 0.142544] ... max period: 00007fffffffffff
[ 0.142544] ... fixed-purpose events: 0
[ 0.142545] ... event mask: 000000000000000f
[ 0.142621] MCE: In-kernel MCE decoding enabled.
[ 0.155796] process: System has AMD C1E enabled
[ 0.155806] process: Switch to broadcast mode on CPU1
[ 0.169182] process: Switch to broadcast mode on CPU2
[ 0.142659] smpboot: Booting Node 0, Processors #1 #2 #3 OK
[ 0.182310] Brought up 4 CPUs
[ 0.182311] smpboot: Total of 4 processors activated (25695.69 BogoMIPS)
[ 0.182317] process: Switch to broadcast mode on CPU3
[ 0.188006] process: Switch to broadcast mode on CPU0
[ 0.188189] devtmpfs: initialized
[ 0.188372] NET: Registered protocol family 16
[ 0.188438] node 0 link 0: io port [1000, ffffff]
[ 0.188439] TOM: 00000000e0000000 aka 3584M
[ 0.188441] Fam 10h mmconf [mem 0xe0000000-0xefffffff]
[ 0.188442] node 0 link 0: mmio [a0000, bffff]
[ 0.188444] node 0 link 0: mmio [e0000000, efffffff] ==> none
[ 0.188445] node 0 link 0: mmio [f0000000, fbcfffff]
[ 0.188446] node 0 link 0: mmio [fbd00000, fbefffff]
[ 0.188447] node 0 link 0: mmio [fbf00000, ffefffff]
[ 0.188448] TOM2: 0000000220000000 aka 8704M
[ 0.188449] bus: [bus 00-07] on node 0 link 0
[ 0.188450] bus: 00 [io 0x0000-0xffff]
[ 0.188450] bus: 00 [mem 0x000a0000-0x000bffff]
[ 0.188451] bus: 00 [mem 0xf0000000-0xffffffff]
[ 0.188452] bus: 00 [mem 0x220000000-0xfcffffffff]
[ 0.188487] ACPI: bus type PCI registered
[ 0.188491] PCI: Using configuration type 1 for base access
[ 0.188491] PCI: Using configuration type 1 for extended access
[ 0.189301] bio: create slab <bio-0> at 0
[ 0.189417] ACPI: Added _OSI(Module Device)
[ 0.189418] ACPI: Added _OSI(Processor Device)
[ 0.189419] ACPI: Added _OSI(3.0 _SCP Extensions)
[ 0.189419] ACPI: Added _OSI(Processor Aggregator Device)
[ 0.190052] ACPI: EC: Look up EC in DSDT
[ 0.191049] ACPI: Executed 3 blocks of module-level executable AML code
[ 0.272841] ACPI: Interpreter enabled
[ 0.272846] ACPI: (supports S0 S5)
[ 0.272846] ACPI: Using IOAPIC for interrupt routing
[ 0.272867] PCI: Using host bridge windows from ACPI; if necessary, use "pci=nocrs" and report a bug
[ 0.276246] ACPI: PCI Root Bridge [PCI0] (domain 0000 [bus 00-ff])
[ 0.276355] PCI host bridge to bus 0000:00
[ 0.276358] pci_bus 0000:00: root bus resource [bus 00-ff]
[ 0.276359] pci_bus 0000:00: root bus resource [io 0x0000-0x0cf7]
[ 0.276360] pci_bus 0000:00: root bus resource [io 0x0d00-0xffff]
[ 0.276362] pci_bus 0000:00: root bus resource [mem 0x000a0000-0x000bffff]
[ 0.276363] pci_bus 0000:00: root bus resource [mem 0x000d0000-0x000dffff]
[ 0.276364] pci_bus 0000:00: root bus resource [mem 0xdff00000-0xdfffffff]
[ 0.276365] pci_bus 0000:00: root bus resource [mem 0xf0000000-0xfebfffff]
[ 0.276374] pci 0000:00:00.0: [1022:9600] type 00 class 0x060000
[ 0.276429] pci 0000:00:01.0: [1022:9602] type 01 class 0x060400
[ 0.276475] pci 0000:00:06.0: [1022:9606] type 01 class 0x060400
[ 0.276501] pci 0000:00:06.0: PME# supported from D0 D3hot D3cold
[ 0.276546] pci 0000:00:11.0: [1002:4391] type 00 class 0x010601
[ 0.276561] pci 0000:00:11.0: reg 0x10: [io 0xc000-0xc007]
[ 0.276570] pci 0000:00:11.0: reg 0x14: [io 0xb000-0xb003]
[ 0.276578] pci 0000:00:11.0: reg 0x18: [io 0xa000-0xa007]
[ 0.276586] pci 0000:00:11.0: reg 0x1c: [io 0x9000-0x9003]
[ 0.276595] pci 0000:00:11.0: reg 0x20: [io 0x8000-0x800f]
[ 0.276603] pci 0000:00:11.0: reg 0x24: [mem 0xfbcffc00-0xfbcfffff]
[ 0.276676] pci 0000:00:12.0: [1002:4397] type 00 class 0x0c0310
[ 0.276687] pci 0000:00:12.0: reg 0x10: [mem 0xfbcfd000-0xfbcfdfff]
[ 0.276773] pci 0000:00:12.1: [1002:4398] type 00 class 0x0c0310
[ 0.276785] pci 0000:00:12.1: reg 0x10: [mem 0xfbcfe000-0xfbcfefff]
[ 0.276876] pci 0000:00:12.2: [1002:4396] type 00 class 0x0c0320
[ 0.276893] pci 0000:00:12.2: reg 0x10: [mem 0xfbcff800-0xfbcff8ff]
[ 0.276971] pci 0000:00:12.2: supports D1 D2
[ 0.276972] pci 0000:00:12.2: PME# supported from D0 D1 D2 D3hot
[ 0.277019] pci 0000:00:13.0: [1002:4397] type 00 class 0x0c0310
[ 0.277031] pci 0000:00:13.0: reg 0x10: [mem 0xfbcfb000-0xfbcfbfff]
[ 0.277119] pci 0000:00:13.1: [1002:4398] type 00 class 0x0c0310
[ 0.277130] pci 0000:00:13.1: reg 0x10: [mem 0xfbcfc000-0xfbcfcfff]
[ 0.277224] pci 0000:00:13.2: [1002:4396] type 00 class 0x0c0320
[ 0.277241] pci 0000:00:13.2: reg 0x10: [mem 0xfbcff400-0xfbcff4ff]
[ 0.277318] pci 0000:00:13.2: supports D1 D2
[ 0.277319] pci 0000:00:13.2: PME# supported from D0 D1 D2 D3hot
[ 0.277371] pci 0000:00:14.0: [1002:4385] type 00 class 0x0c0500
[ 0.277484] pci 0000:00:14.1: [1002:439c] type 00 class 0x01018a
[ 0.277499] pci 0000:00:14.1: reg 0x10: [io 0x0000-0x0007]
[ 0.277508] pci 0000:00:14.1: reg 0x14: [io 0x0000-0x0003]
[ 0.277516] pci 0000:00:14.1: reg 0x18: [io 0x0000-0x0007]
[ 0.277524] pci 0000:00:14.1: reg 0x1c: [io 0x0000-0x0003]
[ 0.277532] pci 0000:00:14.1: reg 0x20: [io 0xff00-0xff0f]
[ 0.277608] pci 0000:00:14.2: [1002:4383] type 00 class 0x040300
[ 0.277627] pci 0000:00:14.2: reg 0x10: [mem 0xfbcf4000-0xfbcf7fff 64bit]
[ 0.277689] pci 0000:00:14.2: PME# supported from D0 D3hot D3cold
[ 0.277726] pci 0000:00:14.3: [1002:439d] type 00 class 0x060100
[ 0.277823] pci 0000:00:14.4: [1002:4384] type 01 class 0x060401
[ 0.277886] pci 0000:00:14.5: [1002:4399] type 00 class 0x0c0310
[ 0.277898] pci 0000:00:14.5: reg 0x10: [mem 0xfbcfa000-0xfbcfafff]
[ 0.277989] pci 0000:00:18.0: [1022:1200] type 00 class 0x060000
[ 0.278031] pci 0000:00:18.1: [1022:1201] type 00 class 0x060000
[ 0.278069] pci 0000:00:18.2: [1022:1202] type 00 class 0x060000
[ 0.278107] pci 0000:00:18.3: [1022:1203] type 00 class 0x060000
[ 0.278147] pci 0000:00:18.4: [1022:1204] type 00 class 0x060000
[ 0.278217] pci 0000:01:05.0: [1002:9614] type 00 class 0x030000
[ 0.278223] pci 0000:01:05.0: reg 0x10: [mem 0xf0000000-0xf7ffffff pref]
[ 0.278227] pci 0000:01:05.0: reg 0x14: [io 0xd000-0xd0ff]
[ 0.278230] pci 0000:01:05.0: reg 0x18: [mem 0xfbee0000-0xfbeeffff]
[ 0.278237] pci 0000:01:05.0: reg 0x24: [mem 0xfbd00000-0xfbdfffff]
[ 0.278251] pci 0000:01:05.0: supports D1 D2
[ 0.278275] pci 0000:01:05.1: [1002:960f] type 00 class 0x040300
[ 0.278281] pci 0000:01:05.1: reg 0x10: [mem 0xfbefc000-0xfbefffff]
[ 0.278305] pci 0000:01:05.1: supports D1 D2
[ 0.278349] pci 0000:00:01.0: PCI bridge to [bus 01]
[ 0.278351] pci 0000:00:01.0: bridge window [io 0xd000-0xdfff]
[ 0.278353] pci 0000:00:01.0: bridge window [mem 0xfbd00000-0xfbefffff]
[ 0.278355] pci 0000:00:01.0: bridge window [mem 0xf0000000-0xf7ffffff 64bit pref]
[ 0.278409] pci 0000:02:00.0: [1969:1026] type 00 class 0x020000
[ 0.278444] pci 0000:02:00.0: reg 0x10: [mem 0xfbfc0000-0xfbffffff 64bit]
[ 0.278464] pci 0000:02:00.0: reg 0x18: [io 0xec00-0xec7f]
[ 0.278613] pci 0000:02:00.0: PME# supported from D3hot D3cold
[ 0.278648] pci 0000:00:06.0: PCI bridge to [bus 02]
[ 0.278650] pci 0000:00:06.0: bridge window [io 0xe000-0xefff]
[ 0.278652] pci 0000:00:06.0: bridge window [mem 0xfbf00000-0xfbffffff]
[ 0.278715] pci 0000:00:14.4: PCI bridge to [bus 03] (subtractive decode)
[ 0.278722] pci 0000:00:14.4: bridge window [io 0x0000-0x0cf7] (subtractive decode)
[ 0.278724] pci 0000:00:14.4: bridge window [io 0x0d00-0xffff] (subtractive decode)
[ 0.278725] pci 0000:00:14.4: bridge window [mem 0x000a0000-0x000bffff] (subtractive decode)
[ 0.278726] pci 0000:00:14.4: bridge window [mem 0x000d0000-0x000dffff] (subtractive decode)
[ 0.278727] pci 0000:00:14.4: bridge window [mem 0xdff00000-0xdfffffff] (subtractive decode)
[ 0.278728] pci 0000:00:14.4: bridge window [mem 0xf0000000-0xfebfffff] (subtractive decode)
[ 0.278737] pci_bus 0000:00: on NUMA node 0
[ 0.278739] acpi PNP0A03:00: ACPI _OSC support notification failed, disabling PCIe ASPM
[ 0.278740] acpi PNP0A03:00: Unable to request _OSC control (_OSC support mask: 0x08)
[ 0.279178] ACPI: PCI Interrupt Link [LNKA] (IRQs 4 7 10 11 12 14 15) *0, disabled.
[ 0.279217] ACPI: PCI Interrupt Link [LNKB] (IRQs 4 7 10 11 12 14 15) *0, disabled.
[ 0.279246] ACPI: PCI Interrupt Link [LNKC] (IRQs 4 7 10 11 12 14 15) *0, disabled.
[ 0.279275] ACPI: PCI Interrupt Link [LNKD] (IRQs 4 7 10 11 12 14 15) *0, disabled.
[ 0.279305] ACPI: PCI Interrupt Link [LNKE] (IRQs 4 7 10 11 12 14 15) *0, disabled.
[ 0.279334] ACPI: PCI Interrupt Link [LNKF] (IRQs 4 7 10 11 12 14 15) *0, disabled.
[ 0.279363] ACPI: PCI Interrupt Link [LNKG] (IRQs 4 7 10 11 12 14 15) *0, disabled.
[ 0.279392] ACPI: PCI Interrupt Link [LNKH] (IRQs 4 7 10 11 12 14 15) *0, disabled.
[ 0.279498] ACPI: Enabled 1 GPEs in block 00 to 1F
[ 0.279502] ACPI: \_SB_.PCI0: notify handler is installed
[ 0.279525] Found 1 acpi root devices
[ 0.279582] SCSI subsystem initialized
[ 0.279631] libata version 3.00 loaded.
[ 0.279633] ACPI: bus type USB registered
[ 0.279642] usbcore: registered new interface driver usbfs
[ 0.279647] usbcore: registered new interface driver hub
[ 0.279667] usbcore: registered new device driver usb
[ 0.279685] EDAC MC: Ver: 3.0.0
[ 0.279787] Advanced Linux Sound Architecture Driver Initialized.
[ 0.279788] PCI: Using ACPI for IRQ routing
[ 0.279789] PCI: pci_cache_line_size set to 64 bytes
[ 0.279838] e820: reserve RAM buffer [mem 0x0009fc00-0x0009ffff]
[ 0.279839] e820: reserve RAM buffer [mem 0xdfe90000-0xdfffffff]
[ 0.279914] hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0, 0
[ 0.279917] hpet0: 4 comparators, 32-bit 14.318180 MHz counter
[ 0.281965] Switched to clocksource hpet
[ 0.281992] pnp: PnP ACPI init
[ 0.281997] ACPI: bus type PNP registered
[ 0.282080] system 00:00: Plug and Play ACPI device, IDs PNP0c02 (active)
[ 0.282106] pnp 00:01: [dma 4]
[ 0.282117] pnp 00:01: Plug and Play ACPI device, IDs PNP0200 (active)
[ 0.282135] pnp 00:02: Plug and Play ACPI device, IDs PNP0b00 (active)
[ 0.282148] pnp 00:03: Plug and Play ACPI device, IDs PNP0800 (active)
[ 0.282164] pnp 00:04: Plug and Play ACPI device, IDs PNP0c04 (active)
[ 0.282193] pnp 00:05: Plug and Play ACPI device, IDs PNP0103 (active)
[ 0.282256] system 00:06: [mem 0xfec00000-0xfec00fff] could not be reserved
[ 0.282257] system 00:06: [mem 0xfee00000-0xfee00fff] has been reserved
[ 0.282258] system 00:06: Plug and Play ACPI device, IDs PNP0c02 (active)
[ 0.282411] system 00:07: [io 0x04d0-0x04d1] has been reserved
[ 0.282412] system 00:07: [io 0x040b] has been reserved
[ 0.282413] system 00:07: [io 0x04d6] has been reserved
[ 0.282415] system 00:07: [io 0x0c00-0x0c01] has been reserved
[ 0.282416] system 00:07: [io 0x0c14] has been reserved
[ 0.282417] system 00:07: [io 0x0c50-0x0c51] has been reserved
[ 0.282418] system 00:07: [io 0x0c52] has been reserved
[ 0.282419] system 00:07: [io 0x0c6c] has been reserved
[ 0.282421] system 00:07: [io 0x0c6f] has been reserved
[ 0.282422] system 00:07: [io 0x0cd0-0x0cd1] has been reserved
[ 0.282423] system 00:07: [io 0x0cd2-0x0cd3] has been reserved
[ 0.282424] system 00:07: [io 0x0cd4-0x0cd5] has been reserved
[ 0.282426] system 00:07: [io 0x0cd6-0x0cd7] has been reserved
[ 0.282427] system 00:07: [io 0x0cd8-0x0cdf] has been reserved
[ 0.282428] system 00:07: [io 0x0b00-0x0b3f] has been reserved
[ 0.282429] system 00:07: [io 0x0800-0x089f] could not be reserved
[ 0.282431] system 00:07: [io 0x0b00-0x0b0f] has been reserved
[ 0.282432] system 00:07: [io 0x0b20-0x0b3f] has been reserved
[ 0.282433] system 00:07: [io 0x0900-0x090f] has been reserved
[ 0.282435] system 00:07: [io 0x0910-0x091f] has been reserved
[ 0.282436] system 00:07: [io 0xfe00-0xfefe] has been reserved
[ 0.282438] system 00:07: [mem 0xdff00000-0xdfffffff] has been reserved
[ 0.282439] system 00:07: [mem 0xffb80000-0xffbfffff] has been reserved
[ 0.282440] system 00:07: [mem 0xfec10000-0xfec1001f] has been reserved
[ 0.282442] system 00:07: Plug and Play ACPI device, IDs PNP0c02 (active)
[ 0.282860] system 00:08: [io 0x0230-0x023f] has been reserved
[ 0.282862] system 00:08: [io 0x0290-0x029f] has been reserved
[ 0.282863] system 00:08: [io 0x0f40-0x0f4f] has been reserved
[ 0.282864] system 00:08: [io 0x0a30-0x0a3f] has been reserved
[ 0.282866] system 00:08: Plug and Play ACPI device, IDs PNP0c02 (active)
[ 0.282901] system 00:09: [mem 0xe0000000-0xefffffff] has been reserved
[ 0.282902] system 00:09: Plug and Play ACPI device, IDs PNP0c02 (active)
[ 0.283037] system 00:0a: [mem 0x00000000-0x0009ffff] could not be reserved
[ 0.283038] system 00:0a: [mem 0x000c0000-0x000cffff] could not be reserved
[ 0.283040] system 00:0a: [mem 0x000e0000-0x000fffff] could not be reserved
[ 0.283041] system 00:0a: [mem 0x00100000-0xdfefffff] could not be reserved
[ 0.283042] system 00:0a: [mem 0xfec00000-0xffffffff] could not be reserved
[ 0.283044] system 00:0a: Plug and Play ACPI device, IDs PNP0c01 (active)
[ 0.283115] pnp: PnP ACPI: found 11 devices
[ 0.283115] ACPI: bus type PNP unregistered
[ 0.289521] pci 0000:00:01.0: PCI bridge to [bus 01]
[ 0.289523] pci 0000:00:01.0: bridge window [io 0xd000-0xdfff]
[ 0.289525] pci 0000:00:01.0: bridge window [mem 0xfbd00000-0xfbefffff]
[ 0.289527] pci 0000:00:01.0: bridge window [mem 0xf0000000-0xf7ffffff 64bit pref]
[ 0.289530] pci 0000:00:06.0: PCI bridge to [bus 02]
[ 0.289531] pci 0000:00:06.0: bridge window [io 0xe000-0xefff]
[ 0.289533] pci 0000:00:06.0: bridge window [mem 0xfbf00000-0xfbffffff]
[ 0.289536] pci 0000:00:14.4: PCI bridge to [bus 03]
[ 0.289626] pci_bus 0000:00: resource 4 [io 0x0000-0x0cf7]
[ 0.289627] pci_bus 0000:00: resource 5 [io 0x0d00-0xffff]
[ 0.289628] pci_bus 0000:00: resource 6 [mem 0x000a0000-0x000bffff]
[ 0.289629] pci_bus 0000:00: resource 7 [mem 0x000d0000-0x000dffff]
[ 0.289630] pci_bus 0000:00: resource 8 [mem 0xdff00000-0xdfffffff]
[ 0.289631] pci_bus 0000:00: resource 9 [mem 0xf0000000-0xfebfffff]
[ 0.289633] pci_bus 0000:01: resource 0 [io 0xd000-0xdfff]
[ 0.289634] pci_bus 0000:01: resource 1 [mem 0xfbd00000-0xfbefffff]
[ 0.289635] pci_bus 0000:01: resource 2 [mem 0xf0000000-0xf7ffffff 64bit pref]
[ 0.289636] pci_bus 0000:02: resource 0 [io 0xe000-0xefff]
[ 0.289637] pci_bus 0000:02: resource 1 [mem 0xfbf00000-0xfbffffff]
[ 0.289638] pci_bus 0000:03: resource 4 [io 0x0000-0x0cf7]
[ 0.289639] pci_bus 0000:03: resource 5 [io 0x0d00-0xffff]
[ 0.289641] pci_bus 0000:03: resource 6 [mem 0x000a0000-0x000bffff]
[ 0.289642] pci_bus 0000:03: resource 7 [mem 0x000d0000-0x000dffff]
[ 0.289643] pci_bus 0000:03: resource 8 [mem 0xdff00000-0xdfffffff]
[ 0.289644] pci_bus 0000:03: resource 9 [mem 0xf0000000-0xfebfffff]
[ 0.289664] NET: Registered protocol family 2
[ 0.289745] TCP established hash table entries: 65536 (order: 8, 1048576 bytes)
[ 0.289966] TCP bind hash table entries: 65536 (order: 8, 1048576 bytes)
[ 0.290173] TCP: Hash tables configured (established 65536 bind 65536)
[ 0.290207] TCP: reno registered
[ 0.290209] UDP hash table entries: 4096 (order: 5, 131072 bytes)
[ 0.290247] UDP-Lite hash table entries: 4096 (order: 5, 131072 bytes)
[ 0.290318] NET: Registered protocol family 1
[ 0.290323] pci 0000:00:01.0: MSI quirk detected; subordinate MSI disabled
[ 0.555536] pci 0000:01:05.0: Boot video device
[ 0.555543] PCI: CLS 64 bytes, default 64
[ 0.555554] PCI-DMA: Using software bounce buffering for IO (SWIOTLB)
[ 0.555556] software IO TLB [mem 0xdbe90000-0xdfe90000] (64MB) mapped at [ffff8800dbe90000-ffff8800dfe8ffff]
[ 0.555666] kvm: Nested Virtualization enabled
[ 0.555667] kvm: Nested Paging enabled
[ 0.555821] LVT offset 1 assigned for vector 0x400
[ 0.555823] IBS: LVT offset 1 assigned
[ 0.555842] perf: AMD IBS detected (0x0000001f)
[ 0.556060] microcode: CPU0: patch_level=0x010000db
[ 0.556066] microcode: CPU1: patch_level=0x010000db
[ 0.556077] microcode: CPU2: patch_level=0x010000db
[ 0.556087] microcode: CPU3: patch_level=0x010000db
[ 0.556114] microcode: Microcode Update Driver: v2.00 <tigran@aivazian.fsnet.co.uk>, Peter Oruba
[ 0.557221] SGI XFS with security attributes, large block/inode numbers, no debug enabled
[ 0.557469] 9p: Installing v9fs 9p2000 file system support
[ 0.557481] msgmni has been set to 15946
[ 0.557728] Block layer SCSI generic (bsg) driver version 0.4 loaded (major 253)
[ 0.557729] io scheduler noop registered
[ 0.557730] io scheduler deadline registered (default)
[ 0.557773] ACPI: processor limited to max C-state 1
[ 0.558662] Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
[ 0.558992] [drm] Initialized drm 1.1.0 20060810
[ 0.558999] [drm] radeon kernel modesetting enabled.
[ 0.559160] [drm] initializing kernel modesetting (RS780 0x1002:0x9614 0x1043:0x834D).
[ 0.559166] [drm] register mmio base: 0xFBEE0000
[ 0.559166] [drm] register mmio size: 65536
[ 0.559695] ATOM BIOS: 113
[ 0.559709] radeon 0000:01:05.0: VRAM: 128M 0x00000000C0000000 - 0x00000000C7FFFFFF (128M used)
[ 0.559710] radeon 0000:01:05.0: GTT: 512M 0x00000000A0000000 - 0x00000000BFFFFFFF
[ 0.559713] [drm] Detected VRAM RAM=128M, BAR=128M
[ 0.559714] [drm] RAM width 32bits DDR
[ 0.559784] [TTM] Zone kernel: Available graphics memory: 4082358 kiB
[ 0.559785] [TTM] Zone dma32: Available graphics memory: 2097152 kiB
[ 0.559785] [TTM] Initializing pool allocator
[ 0.559789] [TTM] Initializing DMA pool allocator
[ 0.559806] [drm] radeon: 128M of VRAM memory ready
[ 0.559807] [drm] radeon: 512M of GTT memory ready.
[ 0.559814] [drm] GART: num cpu pages 131072, num gpu pages 131072
[ 0.560470] [drm] Loading RS780 Microcode
[ 0.566021] [drm] PCIE GART of 512M enabled (table at 0x00000000C0040000).
[ 0.566075] radeon 0000:01:05.0: WB enabled
[ 0.566077] radeon 0000:01:05.0: fence driver on ring 0 use gpu addr 0x00000000a0000c00 and cpu addr 0xffff880215c46c00
[ 0.566078] radeon 0000:01:05.0: fence driver on ring 3 use gpu addr 0x00000000a0000c0c and cpu addr 0xffff880215c46c0c
[ 0.566080] [drm] Supports vblank timestamp caching Rev 1 (10.10.2010).
[ 0.566080] [drm] Driver supports precise vblank timestamp query.
[ 0.566097] [drm] radeon: irq initialized.
[ 0.566331] radeon 0000:01:05.0: setting latency timer to 64
[ 0.598447] [drm] ring test on 0 succeeded in 1 usecs
[ 0.598503] [drm] ring test on 3 succeeded in 1 usecs
[ 0.598789] [drm] ib test on ring 0 succeeded in 0 usecs
[ 0.598801] [drm] ib test on ring 3 succeeded in 0 usecs
[ 0.598884] [drm] Radeon Display Connectors
[ 0.598885] [drm] Connector 0:
[ 0.598886] [drm] VGA-1
[ 0.598887] [drm] DDC: 0x7e40 0x7e40 0x7e44 0x7e44 0x7e48 0x7e48 0x7e4c 0x7e4c
[ 0.598887] [drm] Encoders:
[ 0.598888] [drm] CRT1: INTERNAL_KLDSCP_DAC1
[ 0.598888] [drm] Connector 1:
[ 0.598889] [drm] DVI-D-1
[ 0.598889] [drm] HPD3
[ 0.598890] [drm] DDC: 0x7e50 0x7e50 0x7e54 0x7e54 0x7e58 0x7e58 0x7e5c 0x7e5c
[ 0.598890] [drm] Encoders:
[ 0.598891] [drm] DFP3: INTERNAL_KLDSCP_LVTMA
[ 0.598904] == power state 0 ==
[ 0.598905] ui class: none
[ 0.598906] internal class: boot
[ 0.598907] caps: video
[ 0.598908] uvd vclk: 0 dclk: 0
[ 0.598909] power level 0 sclk: 50000 vddc_index: 2
[ 0.598909] power level 1 sclk: 50000 vddc_index: 2
[ 0.598909] status: c r b
[ 0.598911] == power state 1 ==
[ 0.598911] ui class: performance
[ 0.598911] internal class: none
[ 0.598912] caps: video
[ 0.598913] uvd vclk: 0 dclk: 0
[ 0.598913] power level 0 sclk: 50000 vddc_index: 1
[ 0.598914] power level 1 sclk: 70000 vddc_index: 2
[ 0.598914] status:
[ 0.598914] == power state 2 ==
[ 0.598915] ui class: none
[ 0.598915] internal class: uvd
[ 0.598916] caps: video
[ 0.598916] uvd vclk: 53300 dclk: 40000
[ 0.598917] power level 0 sclk: 50000 vddc_index: 1
[ 0.598917] power level 1 sclk: 50000 vddc_index: 1
[ 0.598918] status:
[ 0.598971] switching from power state:
[ 0.598971] ui class: none
[ 0.598971] internal class: boot
[ 0.598972] caps: video
[ 0.598973] uvd vclk: 0 dclk: 0
[ 0.598973] power level 0 sclk: 50000 vddc_index: 2
[ 0.598974] power level 1 sclk: 50000 vddc_index: 2
[ 0.598974] status: c b
[ 0.598975] switching to power state:
[ 0.598975] ui class: performance
[ 0.598975] internal class: none
[ 0.598976] caps: video
[ 0.598977] uvd vclk: 0 dclk: 0
[ 0.598977] power level 0 sclk: 50000 vddc_index: 1
[ 0.598978] power level 1 sclk: 70000 vddc_index: 2
[ 0.598978] status: r
[ 0.599096] [drm] radeon: dpm initialized
[ 0.641400] [drm] fb mappable at 0xF0142000
[ 0.641401] [drm] vram apper at 0xF0000000
[ 0.641402] [drm] size 7299072
[ 0.641402] [drm] fb depth is 24
[ 0.641402] [drm] pitch is 6912
[ 0.641448] fbcon: radeondrmfb (fb0) is primary device
[ 0.658594] Console: switching to colour frame buffer device 131x105
[ 0.665515] radeon 0000:01:05.0: fb0: radeondrmfb frame buffer device
[ 0.665516] radeon 0000:01:05.0: registered panic notifier
[ 0.665564] [drm] Initialized radeon 2.34.0 20080528 for 0000:01:05.0 on minor 0
[ 0.665702] loop: module loaded
[ 0.665790] ahci 0000:00:11.0: version 3.0
[ 0.665966] ahci 0000:00:11.0: AHCI 0001.0100 32 slots 6 ports 3 Gbps 0x3f impl SATA mode
[ 0.665968] ahci 0000:00:11.0: flags: 64bit ncq sntf ilck pm led clo pmp pio slum part ccc
[ 0.666475] scsi0 : ahci
[ 0.666532] scsi1 : ahci
[ 0.666579] scsi2 : ahci
[ 0.666625] scsi3 : ahci
[ 0.666673] scsi4 : ahci
[ 0.666719] scsi5 : ahci
[ 0.666738] ata1: SATA max UDMA/133 abar m1024@0xfbcffc00 port 0xfbcffd00 irq 22
[ 0.666740] ata2: SATA max UDMA/133 abar m1024@0xfbcffc00 port 0xfbcffd80 irq 22
[ 0.666742] ata3: SATA max UDMA/133 abar m1024@0xfbcffc00 port 0xfbcffe00 irq 22
[ 0.666744] ata4: SATA max UDMA/133 abar m1024@0xfbcffc00 port 0xfbcffe80 irq 22
[ 0.666746] ata5: SATA max UDMA/133 abar m1024@0xfbcffc00 port 0xfbcfff00 irq 22
[ 0.666748] ata6: SATA max UDMA/133 abar m1024@0xfbcffc00 port 0xfbcfff80 irq 22
[ 0.667060] scsi6 : pata_atiixp
[ 0.667106] scsi7 : pata_atiixp
[ 0.667141] ata7: PATA max UDMA/100 cmd 0x1f0 ctl 0x3f6 bmdma 0xff00 irq 14
[ 0.667142] ata8: PATA max UDMA/100 cmd 0x170 ctl 0x376 bmdma 0xff08 irq 15
[ 0.667227] tun: Universal TUN/TAP device driver, 1.6
[ 0.667228] tun: (C) 1999-2004 Max Krasnyansky <maxk@qualcomm.com>
[ 0.712271] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
[ 0.712272] ehci-pci: EHCI PCI platform driver
[ 0.712363] ehci-pci 0000:00:12.2: EHCI Host Controller
[ 0.712390] ehci-pci 0000:00:12.2: new USB bus registered, assigned bus number 1
[ 0.712395] QUIRK: Enable AMD PLL fix
[ 0.712396] ehci-pci 0000:00:12.2: applying AMD SB700/SB800/Hudson-2/3 EHCI dummy qh workaround
[ 0.712398] ehci-pci 0000:00:12.2: applying AMD SB600/SB700 USB freeze workaround
[ 0.712410] ehci-pci 0000:00:12.2: debug port 1
[ 0.712440] ehci-pci 0000:00:12.2: irq 17, io mem 0xfbcff800
[ 0.722179] ehci-pci 0000:00:12.2: USB 2.0 started, EHCI 1.00
[ 0.722256] hub 1-0:1.0: USB hub found
[ 0.722258] hub 1-0:1.0: 6 ports detected
[ 0.722389] ehci-pci 0000:00:13.2: EHCI Host Controller
[ 0.722417] ehci-pci 0000:00:13.2: new USB bus registered, assigned bus number 2
[ 0.722419] ehci-pci 0000:00:13.2: applying AMD SB700/SB800/Hudson-2/3 EHCI dummy qh workaround
[ 0.722420] ehci-pci 0000:00:13.2: applying AMD SB600/SB700 USB freeze workaround
[ 0.722432] ehci-pci 0000:00:13.2: debug port 1
[ 0.722457] ehci-pci 0000:00:13.2: irq 19, io mem 0xfbcff400
[ 0.732190] ehci-pci 0000:00:13.2: USB 2.0 started, EHCI 1.00
[ 0.732249] hub 2-0:1.0: USB hub found
[ 0.732250] hub 2-0:1.0: 6 ports detected
[ 0.732313] ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver
[ 0.732313] ohci-pci: OHCI PCI platform driver
[ 0.732378] ohci-pci 0000:00:12.0: OHCI PCI host controller
[ 0.732407] ohci-pci 0000:00:12.0: new USB bus registered, assigned bus number 3
[ 0.732423] ohci-pci 0000:00:12.0: irq 16, io mem 0xfbcfd000
[ 0.789613] hub 3-0:1.0: USB hub found
[ 0.789618] hub 3-0:1.0: 3 ports detected
[ 0.789722] ohci-pci 0000:00:12.1: OHCI PCI host controller
[ 0.789751] ohci-pci 0000:00:12.1: new USB bus registered, assigned bus number 4
[ 0.789761] ohci-pci 0000:00:12.1: irq 16, io mem 0xfbcfe000
[ 0.826093] ata7.00: ATAPI: HL-DT-STDVD-RAM GH22NP20, 1.03, max UDMA/66
[ 0.832656] ata7.00: configured for UDMA/66
[ 0.846358] hub 4-0:1.0: USB hub found
[ 0.846364] hub 4-0:1.0: 3 ports detected
[ 0.846483] ohci-pci 0000:00:13.0: OHCI PCI host controller
[ 0.846514] ohci-pci 0000:00:13.0: new USB bus registered, assigned bus number 5
[ 0.846525] ohci-pci 0000:00:13.0: irq 18, io mem 0xfbcfb000
[ 0.902980] hub 5-0:1.0: USB hub found
[ 0.902984] hub 5-0:1.0: 3 ports detected
[ 0.903087] ohci-pci 0000:00:13.1: OHCI PCI host controller
[ 0.903116] ohci-pci 0000:00:13.1: new USB bus registered, assigned bus number 6
[ 0.903127] ohci-pci 0000:00:13.1: irq 18, io mem 0xfbcfc000
[ 0.959690] hub 6-0:1.0: USB hub found
[ 0.959694] hub 6-0:1.0: 3 ports detected
[ 0.959798] ohci-pci 0000:00:14.5: OHCI PCI host controller
[ 0.959827] ohci-pci 0000:00:14.5: new USB bus registered, assigned bus number 7
[ 0.959837] ohci-pci 0000:00:14.5: irq 18, io mem 0xfbcfa000
[ 0.978929] ata6: SATA link down (SStatus 0 SControl 300)
[ 0.978971] ata2: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 0.979843] ata2.00: ATA-8: ST1500DL003-9VT16L, CC32, max UDMA/133
[ 0.979846] ata2.00: 2930277168 sectors, multi 16: LBA48 NCQ (depth 31/32)
[ 0.980890] ata2.00: configured for UDMA/133
[ 0.985598] ata4: SATA link down (SStatus 0 SControl 300)
[ 0.985628] ata1: SATA link down (SStatus 0 SControl 300)
[ 0.985659] ata5: SATA link down (SStatus 0 SControl 300)
[ 0.985699] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 0.985747] scsi 1:0:0:0: Direct-Access ATA ST1500DL003-9VT1 CC32 PQ: 0 ANSI: 5
[ 0.985840] sd 1:0:0:0: Attached scsi generic sg0 type 0
[ 0.985888] ata3.00: ATA-8: OCZ VERTEX-TURBO, 1.7, max UDMA/133
[ 0.985890] ata3.00: 62533296 sectors, multi 1: LBA48 NCQ (depth 31/32), AA
[ 0.985905] sd 1:0:0:0: [sda] 2930277168 512-byte logical blocks: (1.50 TB/1.36 TiB)
[ 0.986022] sd 1:0:0:0: [sda] Write Protect is off
[ 0.986024] sd 1:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 0.986069] sd 1:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ 0.987918] ata3.00: configured for UDMA/133
[ 0.988012] scsi 2:0:0:0: Direct-Access ATA OCZ VERTEX-TURBO 1.7 PQ: 0 ANSI: 5
[ 0.988091] sd 2:0:0:0: Attached scsi generic sg1 type 0
[ 0.988108] sd 2:0:0:0: [sdb] 62533296 512-byte logical blocks: (32.0 GB/29.8 GiB)
[ 0.988269] sd 2:0:0:0: [sdb] Write Protect is off
[ 0.988274] sd 2:0:0:0: [sdb] Mode Sense: 00 3a 00 00
[ 0.988319] sd 2:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ 0.989292] scsi 6:0:0:0: CD-ROM HL-DT-ST DVD-RAM GH22NP20 1.03 PQ: 0 ANSI: 5
[ 0.990975] sdb: sdb1 sdb2
[ 0.991322] sd 2:0:0:0: [sdb] Attached SCSI disk
[ 0.991617] sr0: scsi3-mmc drive: 48x/48x writer dvd-ram cd/rw xa/form2 cdda tray
[ 0.991619] cdrom: Uniform CD-ROM driver Revision: 3.20
[ 0.991671] sr 6:0:0:0: Attached scsi CD-ROM sr0
[ 0.991739] sr 6:0:0:0: Attached scsi generic sg2 type 5
[ 1.004419] sda: unknown partition table
[ 1.004657] sd 1:0:0:0: [sda] Attached SCSI disk
[ 1.016348] hub 7-0:1.0: USB hub found
[ 1.016352] hub 7-0:1.0: 2 ports detected
[ 1.016404] usbcore: registered new interface driver usblp
[ 1.016421] usbcore: registered new interface driver usb-storage
[ 1.016443] i8042: PNP: No PS/2 controller found. Probing ports directly.
[ 1.016966] serio: i8042 KBD port at 0x60,0x64 irq 1
[ 1.016970] serio: i8042 AUX port at 0x60,0x64 irq 12
[ 1.017031] mousedev: PS/2 mouse device common for all mice
[ 1.017063] rtc_cmos 00:02: RTC can wake from S4
[ 1.017156] rtc_cmos 00:02: rtc core: registered rtc_cmos as rtc0
[ 1.017178] rtc_cmos 00:02: alarms up to one month, y3k, 114 bytes nvram, hpet irqs
[ 1.017183] i2c /dev entries driver
[ 1.017413] AMD64 EDAC driver v3.4.0
[ 1.017422] EDAC amd64: DRAM ECC enabled.
[ 1.017443] EDAC amd64: F10h detected (node 0).
[ 1.017453] EDAC MC: DCT0 chip selects:
[ 1.017454] EDAC amd64: MC: 0: 1024MB 1: 1024MB
[ 1.017455] EDAC amd64: MC: 2: 1024MB 3: 1024MB
[ 1.017456] EDAC amd64: MC: 4: 0MB 5: 0MB
[ 1.017456] EDAC amd64: MC: 6: 0MB 7: 0MB
[ 1.017457] EDAC MC: DCT1 chip selects:
[ 1.017458] EDAC amd64: MC: 0: 1024MB 1: 1024MB
[ 1.017458] EDAC amd64: MC: 2: 1024MB 3: 1024MB
[ 1.017459] EDAC amd64: MC: 4: 0MB 5: 0MB
[ 1.017459] EDAC amd64: MC: 6: 0MB 7: 0MB
[ 1.017460] EDAC amd64: using x4 syndromes.
[ 1.017461] EDAC amd64: MCT channel count: 2
[ 1.017474] EDAC amd64: CS0: Unbuffered DDR3 RAM
[ 1.017475] EDAC amd64: CS1: Unbuffered DDR3 RAM
[ 1.017475] EDAC amd64: CS2: Unbuffered DDR3 RAM
[ 1.017476] EDAC amd64: CS3: Unbuffered DDR3 RAM
[ 1.017565] EDAC MC0: Giving out device to 'amd64_edac' 'F10h': DEV 0000:00:18.2
[ 1.017576] EDAC PCI0: Giving out device to module 'amd64_edac' controller 'EDAC PCI controller': DEV '0000:00:18.2' (POLLED)
[ 1.017579] cpuidle: using governor ladder
[ 1.017579] cpuidle: using governor menu
[ 1.017586] hidraw: raw HID events driver (C) Jiri Kosina
[ 1.017604] usbcore: registered new interface driver usbhid
[ 1.017605] usbhid: USB HID core driver
[ 1.033112] snd_hda_intel 0000:01:05.1: setting latency timer to 64
[ 1.036076] usbcore: registered new interface driver snd-usb-audio
[ 1.036087] Netfilter messages via NETLINK v0.30.
[ 1.036103] nf_conntrack version 0.5.0 (16384 buckets, 65536 max)
[ 1.036170] ctnetlink v0.93: registering with nfnetlink.
[ 1.036455] ip_tables: (C) 2000-2006 Netfilter Core Team
[ 1.036505] TCP: cubic registered
[ 1.036507] NET: Registered protocol family 17
[ 1.036517] 9pnet: Installing 9P2000 support
[ 1.036715] registered taskstats version 1
[ 1.037059] rtc_cmos 00:02: setting system clock to 2013-07-19 17:51:55 UTC (1374256315)
[ 1.037092] acpi-cpufreq: overriding BIOS provided _PSD data
[ 1.037531] ALSA device list:
[ 1.037532] #0: HDA ATI SB at 0xfbcf4000 irq 16
[ 1.037532] #1: HDA ATI HDMI at 0xfbefc000 irq 19
[ 1.095857] XFS (sdb2): Mounting Filesystem
[ 1.138144] XFS (sdb2): Ending clean mount
[ 1.138172] VFS: Mounted root (xfs filesystem) readonly on device 8:18.
[ 1.142877] devtmpfs: mounted
[ 1.144968] Freeing unused kernel memory: 688K (ffffffff81a6e000 - ffffffff81b1a000)
[ 1.144971] Write protecting the kernel read-only data: 10240k
[ 1.148692] Freeing unused kernel memory: 484K (ffff880001587000 - ffff880001600000)
[ 1.161654] Freeing unused kernel memory: 2028K (ffff880001805000 - ffff880001a00000)
[ 1.559092] tsc: Refined TSC clocksource calibration: 3210.827 MHz
[ 1.621229] XFS (sda): Mounting Filesystem
[ 1.722500] usb 4-2: new full-speed USB device number 2 using ohci-pci
[ 1.908899] logitech-djreceiver 0003:046D:C52B.0003: hiddev0,hidraw0: USB HID v1.11 Device [Logitech USB Receiver] on usb-0000:00:12.1-2/input2
[ 1.913493] input: Logitech Unifying Device. Wireless PID:4026 as /devices/pci0000:00/0000:00:12.1/usb4/4-2/4-2:1.2/0003:046D:C52B.0003/input/input0
[ 1.913717] logitech-djdevice 0003:046D:C52B.0004: input,hidraw1: USB HID v1.11 Keyboard [Logitech Unifying Device. Wireless PID:4026] on usb-0000:00:12.1-2:1
[ 2.039242] usb 3-1: new low-speed USB device number 2 using ohci-pci
[ 2.063203] XFS (sda): Ending clean mount
[ 2.200891] input: HID 046a:0011 as /devices/pci0000:00/0000:00:12.0/usb3/3-1/3-1:1.0/input/input1
[ 2.201061] hid-generic 0003:046A:0011.0005: input,hidraw2: USB HID v1.10 Keyboard [HID 046a:0011] on usb-0000:00:12.0-1/input0
[ 2.373959] udevd[97]: starting eudev version 1.0
[ 2.559376] Switched to clocksource tsc
[ 2.790907] ATL1E 0000:02:00.0 eth0: NIC Link is Up <1000 Mbps Full Duplex>
[ 4.379267] Adding 3071996k swap on /var/cache/swapfile.img. Priority:-1 extents:2 across:8430928k
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 12:22 ` [Bisected] " Markus Trippelsdorf
2013-07-19 12:41 ` Stefan Ring
@ 2013-07-19 21:11 ` Mark Tinguely
2013-07-20 3:18 ` Dave Chinner
2013-07-20 1:48 ` Dave Chinner
2 siblings, 1 reply; 37+ messages in thread
From: Mark Tinguely @ 2013-07-19 21:11 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: Ben Myers, Stan Hoeppner, xfs
On 07/19/13 07:22, Markus Trippelsdorf wrote:
>
> I've bisected this issue to the following commit:
>
> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> Author: Dave Chinner<dchinner@redhat.com>
> Date: Thu Jun 27 16:04:49 2013 +1000
>
> xfs: don't do IO when creating an new inode
>
> Reverting this commit on top of the Linus tree "solves" all problems for
> me. IOW I no longer loose my KDE and LibreOffice config files during a
> crash. Log recovery now works fine and xfs_repair shows no issues.
>
> So users of 3.11.0-rc1 beware. Only run this version if you have
> up-to-date backups handy.
>
I reviewed the above patch and liked it but, I think I recreated the
above mentioned problem with a simple script:
cp /root/.bash_history /root/.lesshst /root/.pwclientrc /root/.viminfo
/root/.bash_profile /root/.lesshst.YCJCDz /root/.quiltrc /somexfsdir
sync
echo 'c' > /proc/sysrq-trigger
.... reboot, remount ...
cd /somexfsdir
# ls -la
ls: cannot access .bash_history: No such file or directory
ls: cannot access .lesshst: No such file or directory
ls: cannot access .pwclientrc: No such file or directory
ls: cannot access .viminfo: No such file or directory
ls: cannot access .bash_profile: No such file or directory
ls: cannot access .lesshst.YCJCDz: No such file or directory
ls: cannot access .quiltrc: No such file or directory
total 4
drwxr-xr-x 2 root root 131 Jul 19 15:32 .
drwxr-xr-x 28 root root 4096 Jul 19 15:35 ..
?????????? ? ? ? ? ? .bash_history
?????????? ? ? ? ? ? .bash_profile
?????????? ? ? ? ? ? .lesshst
?????????? ? ? ? ? ? .lesshst.YCJCDz
?????????? ? ? ? ? ? .pwclientrc
?????????? ? ? ? ? ? .quiltrc
?????????? ? ? ? ? ? .viminfo
# cat .bash_history
cat: .bash_history: No such file or directory
xfs_db> inode 131
xfs_db> p
core.magic = 0x494e
core.mode = 0
core.version = 2
core.format = 2 (extents)
core.nlinkv2 = 0
core.onlink = 0
core.projid_lo = 0
core.projid_hi = 0
core.uid = 0
core.gid = 0
core.flushiter = 1
core.atime.sec = Fri Jul 19 15:26:13 2013
core.atime.nsec = 990813003
core.mtime.sec = Fri Jul 19 15:26:13 2013
core.mtime.nsec = 990813003
core.ctime.sec = Fri Jul 19 15:30:34 2013
core.ctime.nsec = 822788719
core.size = 0
core.nblocks = 0
core.extsize = 0
core.nextents = 0
core.naextents = 0
core.forkoff = 0
core.aformat = 2 (extents)
core.dmevmask = 0
core.dmstate = 0
core.newrtbm = 0
core.prealloc = 0
core.realtime = 0
core.immutable = 0
core.append = 0
core.sync = 0
core.noatime = 0
core.nodump = 0
core.rtinherit = 0
core.projinherit = 0
core.nosymlinks = 0
core.extsz = 0
core.extszinherit = 0
core.nodefrag = 0
core.filestream = 0
core.gen = 3707503345
next_unlinked = null
u = (empty)
revert the above commit and the problem goes away.
Output:
The files are small files that I could find on the test box:
-rw------- 1 root root 28158 Jul 19 15:47 .bash_history
-rw-r--r-- 1 root root 43 Jul 19 15:47 .bash_profile
-rw------- 1 root root 1046 Jul 19 15:47 .lesshst
-rw------- 1 root root 919 Jul 19 15:47 .lesshst.YCJCDz
-rw-r--r-- 1 root root 344 Jul 19 15:47 .pwclientrc
-rw-r--r-- 1 root root 2502 Jul 19 15:47 .quiltrc
-rw------- 1 root root 21895 Jul 19 15:47 .viminfo
And they diff the same as the originals.
core.magic = 0x494e
core.mode = 0100600
core.version = 2
core.format = 2 (extents)
core.nlinkv2 = 1
core.onlink = 0
core.projid_lo = 0
core.projid_hi = 0
core.uid = 0
core.gid = 0
core.flushiter = 1
core.atime.sec = Fri Jul 19 15:56:04 2013
core.atime.nsec = 954825196
core.mtime.sec = Fri Jul 19 15:47:18 2013
core.mtime.nsec = 366686434
core.ctime.sec = Fri Jul 19 15:47:18 2013
core.ctime.nsec = 366686434
core.size = 28158
core.nblocks = 7
core.extsize = 0
core.nextents = 1
core.naextents = 0
core.forkoff = 0
core.aformat = 2 (extents)
core.dmevmask = 0
core.dmstate = 0
core.newrtbm = 0
core.prealloc = 0
core.realtime = 0
core.immutable = 0
core.append = 0
core.sync = 0
core.noatime = 0
core.nodump = 0
core.rtinherit = 0
core.projinherit = 0
core.nosymlinks = 0
core.extsz = 0
core.extszinherit = 0
core.nodefrag = 0
core.filestream = 0
core.gen = 0
next_unlinked = null
u.bmx[0] = [startoff,startblock,blockcount,extentflag] 0:[0,12,7,0]
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 12:22 ` [Bisected] " Markus Trippelsdorf
2013-07-19 12:41 ` Stefan Ring
2013-07-19 21:11 ` Mark Tinguely
@ 2013-07-20 1:48 ` Dave Chinner
2013-07-22 10:22 ` Dave Chinner
2 siblings, 1 reply; 37+ messages in thread
From: Dave Chinner @ 2013-07-20 1:48 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: Ben Myers, Mark Tinguely, Stan Hoeppner, xfs
On Fri, Jul 19, 2013 at 02:22:35PM +0200, Markus Trippelsdorf wrote:
> On 2013.07.15 at 08:47 +0200, Markus Trippelsdorf wrote:
> > On 2013.07.15 at 12:28 +1000, Dave Chinner wrote:
> > > On Fri, Jul 12, 2013 at 09:07:21AM +0200, Markus Trippelsdorf wrote:
> > > > On 2013.07.12 at 12:17 +1000, Dave Chinner wrote:
> > > > > On Thu, Jul 11, 2013 at 11:07:55AM +0200, Markus Trippelsdorf wrote:
> > > > > > On 2013.07.10 at 23:12 -0500, Stan Hoeppner wrote:
> > > > > > > On 7/10/2013 10:58 PM, Dave Chinner wrote:
> > > > > > > > On Thu, Jul 11, 2013 at 05:36:21AM +0200, Markus Trippelsdorf wrote:
> > > > > > >
> > > > > > > >> I was loosing my KDE settings bit by bit with every reboot during the
> > > > > > > >> bisection. First my window-rules disappeared, then my desktop background
> > > > > > > >> changed to default, then my taskbar moved from top to the bottom, etc.
> > > > > > > >> In the end I had to restore all my .files from backup.
> > > > > > > >
> > > > > > > > That's not filesystem corruption. That sounds more like someone not
> > > > > > > > using fsync in the apropriate place when overwriting a file....
> > > > > > >
> > > > > > t@ubunt:~# xfs_repair /dev/sdb
> > > > > > Phase 1 - find and verify superblock...
> > > > > > Phase 2 - using internal log
> > > > > > - zero log...
> > > > > > - scan filesystem freespace and inode maps...
> > > > > > agi unlinked bucket 0 is 683435008 in ag 2 (inode=4978402304)
> > > > > > agi unlinked bucket 1 is 683435009 in ag 2 (inode=4978402305)
> > > > > > - found root inode chunk
> > > > >
> > > > > Again, these are signs that log recovery has not completed
> > > > > successfully or that for some reason it thought the log was clean.
> > > > > Can you please post the dmesg output after the crash when you go
> > > > > through the mount/unmount process before you run xfs_repair?
> > > >
> > > > Sure.
> > > > First boot after crash:
> > > > XFS (sdb2): Mounting Filesystem
> > > > XFS (sdb2): Starting recovery (logdev: internal)
> > > > XFS (sdb2): Ending recovery (logdev: internal)
> > > >
> > > > Second boot after crash:
> > > > XFS (sdb2): Mounting Filesystem
> > > > XFS (sdb2): Ending clean mount
> > > >
> > > > I then boot Ubuntu from another disc to run xfs_repair.
> > >
> > > That's what shoul dhave been in the initial description of your
> > > problem.
> > >
> > > > And looking through my logs I see this WARNING:
> > > >
> > > > ------------[ cut here ]------------
> > > > WARNING: CPU: 0 PID: 439 at fs/inode.c:280 drop_nlink+0x33/0x40()
> > > > CPU: 0 PID: 439 Comm: gconfd-2 Not tainted 3.10.0-08982-g6d128e1-dirty #42
> > > > Hardware name: System manufacturer System Product Name/M4A78T-E, BIOS 3503 04/13/2011
> > > > 0000000000000009 ffffffff8157d030 0000000000000000 ffffffff81060788
> > > > ffff8801f8608cc8 ffff880205998230 ffff8801f7bede58 0000000000000000
> > > > ffff8801f86083c0 ffffffff8110ce93 ffff8801f8608b40 ffffffff811b7104
> > > > Call Trace:
> > > > [<ffffffff8157d030>] ? dump_stack+0x41/0x51
> > > > [<ffffffff81060788>] ? warn_slowpath_common+0x68/0x80
> > > > [<ffffffff8110ce93>] ? drop_nlink+0x33/0x40
> > > > [<ffffffff811b7104>] ? xfs_droplink+0x24/0x60
> > > > [<ffffffff811b84ed>] ? xfs_remove+0x24d/0x380
> > > > [<ffffffff811b1657>] ? xfs_vn_unlink+0x37/0x80
> > > > [<ffffffff8110414e>] ? vfs_unlink+0x6e/0xe0
> > > > [<ffffffff8110432a>] ? do_unlinkat+0x16a/0x220
> > > > [<ffffffff810f4fa9>] ? SyS_faccessat+0x149/0x200
> > > > [<ffffffff81583292>] ? system_call_fastpath+0x16/0x1b
> > >
> > > When did that occur? Before the crash, after the first/second mount?
> > > after you ran repair?
> >
> > After the first mount.
> >
> > > > Some further observations:
> > > >
> > > > When I boot 3.2.0 after the crash log recovery works fine.
> > > >
> > > > When I boot 3.9.0 after the crash I get the following:
> > > >
> > > > [ 2.332989] XFS (sdc2): Mounting Filesystem
> > > > [ 2.406206] XFS (sdc2): Starting recovery (logdev: internal)
> > > > [ 2.418147] XFS (sdc2): log record CRC mismatch: found 0xdbcaef48, expected 0x69e7934e.
> > >
> > > Just informational - indicating that the log records don't have
> > > valid CRCs in them because 3.2 didn't calculate them. If you are
> > > getting them when after a crash on a 3.9+ kernel, then there's a
> > > problem writing to the log....
> >
> > The crash always occurred on the current Linus tree kernel...
> >
> > > > When I boot the current Linus tree after the crash log recovery fails silently.
> > >
> > > dmesg output, please. Indeed, what does "fails silently" mean? the
> > > filesystem doesn't mount but no error is given?
> >
> > Again, there is no dmesg output. XFS tells me that it's "Ending recovery
> > (logdev: internal)" without any errors, when indeed it didn't recover
> > the log at all. It then mounts the filesystem normally (rw) in this
> > unclean state. That's when the WARNING I postend above happend.
>
> I've bisected this issue to the following commit:
>
> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> Author: Dave Chinner <dchinner@redhat.com>
> Date: Thu Jun 27 16:04:49 2013 +1000
>
> xfs: don't do IO when creating an new inode
>
> Reverting this commit on top of the Linus tree "solves" all problems for
> me. IOW I no longer loose my KDE and LibreOffice config files during a
> crash. Log recovery now works fine and xfs_repair shows no issues.
Thanks for bisecting this, Marcus.
I'll admit, right now it doesn't make a lot of sense to me - I don't
immediately see a connection between not reading an inode during the
create phase and unlinked list and directory corruption after a
crash. But now you've identified a change that might be the cause,
I have an avenue of investigation I can follow.
Indeed, in the time I've taken to write this mail I've thought of
2-3 possible causes that I need to investigate....
> So users of 3.11.0-rc1 beware. Only run this version if you have
> up-to-date backups handy.
Don't be so dramatic - very few people are doing what you are doing,
so let's try to understand the root cause of problem before jumping
to rash conclusions....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-19 21:11 ` Mark Tinguely
@ 2013-07-20 3:18 ` Dave Chinner
2013-07-20 17:21 ` Mark Tinguely
0 siblings, 1 reply; 37+ messages in thread
From: Dave Chinner @ 2013-07-20 3:18 UTC (permalink / raw)
To: Mark Tinguely; +Cc: Ben Myers, Stan Hoeppner, Markus Trippelsdorf, xfs
On Fri, Jul 19, 2013 at 04:11:28PM -0500, Mark Tinguely wrote:
> On 07/19/13 07:22, Markus Trippelsdorf wrote:
> >
> >I've bisected this issue to the following commit:
> >
> > commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> > Author: Dave Chinner<dchinner@redhat.com>
> > Date: Thu Jun 27 16:04:49 2013 +1000
> >
> > xfs: don't do IO when creating an new inode
> >
> >Reverting this commit on top of the Linus tree "solves" all problems for
> >me. IOW I no longer loose my KDE and LibreOffice config files during a
> >crash. Log recovery now works fine and xfs_repair shows no issues.
> >
> >So users of 3.11.0-rc1 beware. Only run this version if you have
> >up-to-date backups handy.
> >
>
> I reviewed the above patch and liked it but, I think I recreated the
> above mentioned problem with a simple script:
>
> cp /root/.bash_history /root/.lesshst /root/.pwclientrc
> /root/.viminfo /root/.bash_profile /root/.lesshst.YCJCDz
> /root/.quiltrc /somexfsdir
> sync
> echo 'c' > /proc/sysrq-trigger
> .... reboot, remount ...
> cd /somexfsdir
I've only reproduced the problem *once* with this method - the first
time I tried. Then I mkfs'd the filesystem rather than repairing it
and I haven't been able to reproduce it since. So the problem is
far more subtle that just copying some files, running sync and
crashing the machine - there's some kind of initial or timing
condition that we are missing that triggers it...
The one interesting thing I noticed was that the generation number
in the crash case was non-zero. That's an important piece of
information, and:
> # cat .bash_history
> cat: .bash_history: No such file or directory
>
> xfs_db> inode 131
> xfs_db> p
> core.magic = 0x494e
> core.mode = 0
That's a "free" inode, and why XFS considers it invalid when the
lookup sees it.
> core.gen = 3707503345
You saw it as well, Mark.
That means it has actually been allocated and written to disk at
some point in time. That is, inodes allocated by mkfs in the root
inode chunk have a generation number of zero. For this to have a
non-zero generation number, it means that had to be written after
allocation - either before the sync or during log recovery.
Unfortunately, without the 'xfs_logprint -t -i <dev>' output from
prior to mounting the filesystem which demonstrates te problem, I
can't tell if the issue is a recovery problem or something that
happened before the crash....
> revert the above commit and the problem goes away.
....
> core.mode = 0100600
Not an free inode...
> core.gen = 0
And, importantly, the generation number is zero, as would be
expected for an inode in the root chunk.
FWIW, if you can reproduce this on demand, Mark, is to see if
mounting "-o ikeep" makes the problem go away as this optimisation
is only used on filesystems that are configured to free inode
chunks...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-20 3:18 ` Dave Chinner
@ 2013-07-20 17:21 ` Mark Tinguely
2013-07-21 7:37 ` Dave Chinner
0 siblings, 1 reply; 37+ messages in thread
From: Mark Tinguely @ 2013-07-20 17:21 UTC (permalink / raw)
To: Dave Chinner; +Cc: Ben Myers, Stan Hoeppner, Markus Trippelsdorf, xfs
On 07/19/13 22:18, Dave Chinner wrote:
> On Fri, Jul 19, 2013 at 04:11:28PM -0500, Mark Tinguely wrote:
>> On 07/19/13 07:22, Markus Trippelsdorf wrote:
>>>
>>> I've bisected this issue to the following commit:
>>>
>>> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
>>> Author: Dave Chinner<dchinner@redhat.com>
>>> Date: Thu Jun 27 16:04:49 2013 +1000
>>>
>>> xfs: don't do IO when creating an new inode
>>>
>>> Reverting this commit on top of the Linus tree "solves" all problems for
>>> me. IOW I no longer loose my KDE and LibreOffice config files during a
>>> crash. Log recovery now works fine and xfs_repair shows no issues.
>>>
>>> So users of 3.11.0-rc1 beware. Only run this version if you have
>>> up-to-date backups handy.
>>>
>>
>> I reviewed the above patch and liked it but, I think I recreated the
>> above mentioned problem with a simple script:
>>
>> cp /root/.bash_history /root/.lesshst /root/.pwclientrc
>> /root/.viminfo /root/.bash_profile /root/.lesshst.YCJCDz
>> /root/.quiltrc /somexfsdir
>> sync
>> echo 'c'> /proc/sysrq-trigger
>> .... reboot, remount ...
>> cd /somexfsdir
>
> I've only reproduced the problem *once* with this method - the first
> time I tried. Then I mkfs'd the filesystem rather than repairing it
> and I haven't been able to reproduce it since. So the problem is
> far more subtle that just copying some files, running sync and
> crashing the machine - there's some kind of initial or timing
> condition that we are missing that triggers it...
>
> The one interesting thing I noticed was that the generation number
> in the crash case was non-zero. That's an important piece of
> information, and:
>
>> # cat .bash_history
>> cat: .bash_history: No such file or directory
>>
>> xfs_db> inode 131
>> xfs_db> p
>> core.magic = 0x494e
>> core.mode = 0
>
> That's a "free" inode, and why XFS considers it invalid when the
> lookup sees it.
>
>> core.gen = 3707503345
>
> You saw it as well, Mark.
>
> That means it has actually been allocated and written to disk at
> some point in time. That is, inodes allocated by mkfs in the root
> inode chunk have a generation number of zero. For this to have a
> non-zero generation number, it means that had to be written after
> allocation - either before the sync or during log recovery.
>
> Unfortunately, without the 'xfs_logprint -t -i<dev>' output from
> prior to mounting the filesystem which demonstrates te problem, I
> can't tell if the issue is a recovery problem or something that
> happened before the crash....
>
>> revert the above commit and the problem goes away.
> ....
>> core.mode = 0100600
>
> Not an free inode...
>
>> core.gen = 0
>
> And, importantly, the generation number is zero, as would be
> expected for an inode in the root chunk.
>
> FWIW, if you can reproduce this on demand, Mark, is to see if
> mounting "-o ikeep" makes the problem go away as this optimisation
> is only used on filesystems that are configured to free inode
> chunks...
>
> Cheers,
>
> Dave.
Yeah, I thought of the logprint and the ikeep afterwards.
I tried the script today and it did not reproduce the problem. The
logprint and the mounted filesystem was empty. I will rebuild the
sources to eliminate some patched kernel versions on that box and
experiment with the sync and the shooting of the kernel.
--Mark.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-20 17:21 ` Mark Tinguely
@ 2013-07-21 7:37 ` Dave Chinner
0 siblings, 0 replies; 37+ messages in thread
From: Dave Chinner @ 2013-07-21 7:37 UTC (permalink / raw)
To: Mark Tinguely; +Cc: Ben Myers, Stan Hoeppner, Markus Trippelsdorf, xfs
On Sat, Jul 20, 2013 at 12:21:47PM -0500, Mark Tinguely wrote:
> On 07/19/13 22:18, Dave Chinner wrote:
> >On Fri, Jul 19, 2013 at 04:11:28PM -0500, Mark Tinguely wrote:
> >>On 07/19/13 07:22, Markus Trippelsdorf wrote:
> >>>
> >>>I've bisected this issue to the following commit:
> >>>
> >>> commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> >>> Author: Dave Chinner<dchinner@redhat.com>
> >>> Date: Thu Jun 27 16:04:49 2013 +1000
> >>>
> >>> xfs: don't do IO when creating an new inode
> >>>
> >>>Reverting this commit on top of the Linus tree "solves" all problems for
> >>>me. IOW I no longer loose my KDE and LibreOffice config files during a
> >>>crash. Log recovery now works fine and xfs_repair shows no issues.
....
> >I've only reproduced the problem *once* with this method - the first
> >time I tried. Then I mkfs'd the filesystem rather than repairing it
> >and I haven't been able to reproduce it since. So the problem is
> >far more subtle that just copying some files, running sync and
> >crashing the machine - there's some kind of initial or timing
> >condition that we are missing that triggers it...
> >
> >The one interesting thing I noticed was that the generation number
> >in the crash case was non-zero. That's an important piece of
> >information, and:
....
> >That means it has actually been allocated and written to disk at
> >some point in time. That is, inodes allocated by mkfs in the root
> >inode chunk have a generation number of zero. For this to have a
> >non-zero generation number, it means that had to be written after
> >allocation - either before the sync or during log recovery.
> >
> >Unfortunately, without the 'xfs_logprint -t -i<dev>' output from
> >prior to mounting the filesystem which demonstrates te problem, I
> >can't tell if the issue is a recovery problem or something that
> >happened before the crash....
....
> I tried the script today and it did not reproduce the problem. The
> logprint and the mounted filesystem was empty. I will rebuild the
> sources to eliminate some patched kernel versions on that box and
> experiment with the sync and the shooting of the kernel.
No need - I've worked out yesterday how to reproduce it reliably and
what the root cause of the problem is. My 'net connection was down
yesterday, so I wasn't even sure if my emails would get out after
queued them and left for an overnight trip....
Basically, the problem takes two iterations to trigger. Do this:
mkfs.xfs
mount
copy files
umount
mount
remove files
umount.
This gives files in the inode chunk mode = 0 and flushiter = 2. now
run:
mount
copy files
sync
godown (*)
umount
(*) you don't need to crash the box to trip this problem.
And when you run:
mount
ls -l
The output of ls will have missing files.
The problem is log recovery sees the flushiter of the inodes being
allocated as 0 (because that's what the patch that avoids reading
the inodes during create sets it too), but the flushiter of the
inode on disk is 2, and then log recovery says "inode on disk is
more recent than the inode core being recovered, don't do recovery".
And that's all there is to it. di_flushiter is no longer necessary
as we log all inode modifications now, but we left it there because
we thought it was harmless:
/*
* bump the flush iteration count, used to detect flushes which
* postdate a log record during recovery. This is redundant as we now
* log every change and hence this can't happen. Still, it doesn't hurt.
*/
ip->i_d.di_flushiter++;
In this case, clearly it does hurt. :/
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-20 1:48 ` Dave Chinner
@ 2013-07-22 10:22 ` Dave Chinner
2013-07-22 10:47 ` Markus Trippelsdorf
0 siblings, 1 reply; 37+ messages in thread
From: Dave Chinner @ 2013-07-22 10:22 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: Ben Myers, Mark Tinguely, Stan Hoeppner, xfs
On Sat, Jul 20, 2013 at 11:48:36AM +1000, Dave Chinner wrote:
> On Fri, Jul 19, 2013 at 02:22:35PM +0200, Markus Trippelsdorf wrote:
> > On 2013.07.15 at 08:47 +0200, Markus Trippelsdorf wrote:
> > I've bisected this issue to the following commit:
> >
> > commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> > Author: Dave Chinner <dchinner@redhat.com>
> > Date: Thu Jun 27 16:04:49 2013 +1000
> >
> > xfs: don't do IO when creating an new inode
> >
> > Reverting this commit on top of the Linus tree "solves" all problems for
> > me. IOW I no longer loose my KDE and LibreOffice config files during a
> > crash. Log recovery now works fine and xfs_repair shows no issues.
>
> Thanks for bisecting this, Marcus.
>
> I'll admit, right now it doesn't make a lot of sense to me - I don't
> immediately see a connection between not reading an inode during the
> create phase and unlinked list and directory corruption after a
> crash. But now you've identified a change that might be the cause,
> I have an avenue of investigation I can follow.
>
> Indeed, in the time I've taken to write this mail I've thought of
> 2-3 possible causes that I need to investigate....
Hi Markus, can you test the patch I just posted to the list titled
"xfs: di_flushiter considered harmful" and see if it fixes your
problem? Archive link here:
http://oss.sgi.com/pipermail/xfs/2013-July/028331.html
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-22 10:22 ` Dave Chinner
@ 2013-07-22 10:47 ` Markus Trippelsdorf
2013-07-22 22:54 ` Dave Chinner
0 siblings, 1 reply; 37+ messages in thread
From: Markus Trippelsdorf @ 2013-07-22 10:47 UTC (permalink / raw)
To: Dave Chinner; +Cc: Ben Myers, Mark Tinguely, Stan Hoeppner, xfs
On 2013.07.22 at 20:22 +1000, Dave Chinner wrote:
> On Sat, Jul 20, 2013 at 11:48:36AM +1000, Dave Chinner wrote:
> > On Fri, Jul 19, 2013 at 02:22:35PM +0200, Markus Trippelsdorf wrote:
> > > On 2013.07.15 at 08:47 +0200, Markus Trippelsdorf wrote:
> > > I've bisected this issue to the following commit:
> > >
> > > commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> > > Author: Dave Chinner <dchinner@redhat.com>
> > > Date: Thu Jun 27 16:04:49 2013 +1000
> > >
> > > xfs: don't do IO when creating an new inode
> > >
> > > Reverting this commit on top of the Linus tree "solves" all problems for
> > > me. IOW I no longer loose my KDE and LibreOffice config files during a
> > > crash. Log recovery now works fine and xfs_repair shows no issues.
> >
> > Thanks for bisecting this, Marcus.
> >
> > I'll admit, right now it doesn't make a lot of sense to me - I don't
> > immediately see a connection between not reading an inode during the
> > create phase and unlinked list and directory corruption after a
> > crash. But now you've identified a change that might be the cause,
> > I have an avenue of investigation I can follow.
> >
> > Indeed, in the time I've taken to write this mail I've thought of
> > 2-3 possible causes that I need to investigate....
>
> Hi Markus, can you test the patch I just posted to the list titled
> "xfs: di_flushiter considered harmful" and see if it fixes your
> problem? Archive link here:
>
> http://oss.sgi.com/pipermail/xfs/2013-July/028331.html
Unfortunately no. I still get the same corruption with this patch
applied.
(It's embarrassing to mention, but please add:
Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
to the next iteration of this patch.
Thanks.)
--
Markus
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [Bisected] Corruption of root fs during git bisect of drm system hang
2013-07-22 10:47 ` Markus Trippelsdorf
@ 2013-07-22 22:54 ` Dave Chinner
0 siblings, 0 replies; 37+ messages in thread
From: Dave Chinner @ 2013-07-22 22:54 UTC (permalink / raw)
To: Markus Trippelsdorf; +Cc: Ben Myers, Mark Tinguely, Stan Hoeppner, xfs
On Mon, Jul 22, 2013 at 12:47:43PM +0200, Markus Trippelsdorf wrote:
> On 2013.07.22 at 20:22 +1000, Dave Chinner wrote:
> > On Sat, Jul 20, 2013 at 11:48:36AM +1000, Dave Chinner wrote:
> > > On Fri, Jul 19, 2013 at 02:22:35PM +0200, Markus Trippelsdorf wrote:
> > > > On 2013.07.15 at 08:47 +0200, Markus Trippelsdorf wrote:
> > > > I've bisected this issue to the following commit:
> > > >
> > > > commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> > > > Author: Dave Chinner <dchinner@redhat.com>
> > > > Date: Thu Jun 27 16:04:49 2013 +1000
> > > >
> > > > xfs: don't do IO when creating an new inode
> > > >
> > > > Reverting this commit on top of the Linus tree "solves" all problems for
> > > > me. IOW I no longer loose my KDE and LibreOffice config files during a
> > > > crash. Log recovery now works fine and xfs_repair shows no issues.
> > >
> > > Thanks for bisecting this, Marcus.
> > >
> > > I'll admit, right now it doesn't make a lot of sense to me - I don't
> > > immediately see a connection between not reading an inode during the
> > > create phase and unlinked list and directory corruption after a
> > > crash. But now you've identified a change that might be the cause,
> > > I have an avenue of investigation I can follow.
> > >
> > > Indeed, in the time I've taken to write this mail I've thought of
> > > 2-3 possible causes that I need to investigate....
> >
> > Hi Markus, can you test the patch I just posted to the list titled
> > "xfs: di_flushiter considered harmful" and see if it fixes your
> > problem? Archive link here:
> >
> > http://oss.sgi.com/pipermail/xfs/2013-July/028331.html
>
> Unfortunately no. I still get the same corruption with this patch
> applied.
Umm, really? can you please put together a simple reproducer then?
Because it definitely fixes the problem that Mark reproduced...
> (It's embarrassing to mention, but please add:
> Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
> to the next iteration of this patch.
> Thanks.)
The reason I asked you to test it was so I could confirm that it was
fixing the problem you've reported and so I could added reported-by
and tested-by tags to it.
Indeed, if it doesn't fix your problem, then it's not fixing the bug
you reported, and so adding such tags is wrong.... ;)
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 37+ messages in thread
end of thread, other threads:[~2013-07-22 22:54 UTC | newest]
Thread overview: 37+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-07-10 9:06 Corruption of root fs during git bisect of drm system hang Markus Trippelsdorf
2013-07-11 0:31 ` Dave Chinner
2013-07-11 3:36 ` Markus Trippelsdorf
2013-07-11 3:58 ` Dave Chinner
2013-07-11 4:12 ` Stan Hoeppner
2013-07-11 9:07 ` Markus Trippelsdorf
2013-07-11 11:28 ` Markus Trippelsdorf
2013-07-11 20:24 ` Stan Hoeppner
2013-07-11 20:40 ` Markus Trippelsdorf
2013-07-11 23:01 ` Stan Hoeppner
2013-07-12 2:38 ` Dave Chinner
2013-07-12 2:17 ` Dave Chinner
2013-07-12 7:07 ` Markus Trippelsdorf
2013-07-13 9:05 ` Markus Trippelsdorf
2013-07-15 2:28 ` Dave Chinner
2013-07-15 6:47 ` Markus Trippelsdorf
2013-07-19 12:22 ` [Bisected] " Markus Trippelsdorf
2013-07-19 12:41 ` Stefan Ring
2013-07-19 12:51 ` Markus Trippelsdorf
2013-07-19 16:02 ` Eric Sandeen
2013-07-19 16:32 ` Markus Trippelsdorf
2013-07-19 19:13 ` Ben Myers
2013-07-19 19:56 ` Markus Trippelsdorf
2013-07-19 20:28 ` Markus Trippelsdorf
2013-07-19 19:23 ` Eric Sandeen
2013-07-19 19:53 ` Markus Trippelsdorf
2013-07-19 21:11 ` Mark Tinguely
2013-07-20 3:18 ` Dave Chinner
2013-07-20 17:21 ` Mark Tinguely
2013-07-21 7:37 ` Dave Chinner
2013-07-20 1:48 ` Dave Chinner
2013-07-22 10:22 ` Dave Chinner
2013-07-22 10:47 ` Markus Trippelsdorf
2013-07-22 22:54 ` Dave Chinner
2013-07-11 4:15 ` Markus Trippelsdorf
2013-07-11 0:37 ` Stan Hoeppner
2013-07-11 3:47 ` Markus Trippelsdorf
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.