Search String: Display: Description: Sort:

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*deadlocked\s+xfs\s*$/: 51 ]

Total 51 documents matching your query.

1. deadlocked xfs (score: 1)
Author:
Date: Thu, 10 Jul 2008 21:33:11 -0500
while running fs_mark: fs_mark -d /mnt/test -D 256 -n 100000 -t 4 -s 20480 -F -S 0 which runs 4 threads making 20k files,things got stuffed up in the log, see sysrq-w output at: http://sandeen.fedora
/archives/xfs/2008-07/msg00177.html (7,946 bytes)

2. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 12:48:11 +1000
Thanks for the report Eric. This looks very similar to a deadlock Lachlan recently hit in the patch for "Use atomics for iclog reference counting" http://oss.sgi.com/archives/xfs/2008-02/msg00130.htm
/archives/xfs/2008-07/msg00178.html (10,331 bytes)

3. Re: deadlocked xfs (score: 1)
Author:
Date: Thu, 10 Jul 2008 21:56:48 -0500
thanks I'll look into that. Forgot to mention this was on 2.6.26-rc2 -Eric
/archives/xfs/2008-07/msg00179.html (9,995 bytes)

4. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 13:21:39 +1000
Okay, what Mark is talking about is pv#983925. Details below from the bug - from Lachlan and me. I'm sorry that this info didn't go out sooner. I'm not sure if Lachlan got any further with this but h
/archives/xfs/2008-07/msg00180.html (16,751 bytes)

5. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 13:22:58 +1000
Nice to know - why didn't anyone email me or report this to the list when the bug was first found? I mean, I wrote that code, I know what it is supposed to be doing and as a result should be able hel
/archives/xfs/2008-07/msg00181.html (9,969 bytes)

6. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 13:50:10 +1000
Dave Chinner wrote: On Fri, Jul 11, 2008 at 12:48:11PM +1000, Mark Goodwin wrote: Thanks for the report Eric. This looks very similar to a deadlock Lachlan recently hit in the patch for "Use atomics
/archives/xfs/2008-07/msg00182.html (11,034 bytes)

7. Re: deadlocked xfs (score: 1)
Author:
Date: Thu, 10 Jul 2008 23:02:12 -0500
how 'bout I test my testcase w/ the patch removed to at least be sure that's it. :) Then maybe we can see if Dave, having been alerted to the problem, has a notion of how to fix it? -Eric
/archives/xfs/2008-07/msg00183.html (8,724 bytes)

8. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 14:04:47 +1000
Just FYI, this is not a 'decrement and test' operation - it's an 'decrement and return locked if zero' operation. Best to explain it is this comment in lib/dec_and_lock.c: /* * This is an implementat
/archives/xfs/2008-07/msg00184.html (14,778 bytes)

9. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 14:05:36 +1000
No, do not back it out. I just posted the fix. Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx
/archives/xfs/2008-07/msg00185.html (10,687 bytes)

10. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 14:08:12 +1000
Remaining info from bug....
/archives/xfs/2008-07/msg00186.html (24,959 bytes)

11. Re: deadlocked xfs (score: 1)
Author:
Date: Thu, 10 Jul 2008 23:10:21 -0500
Dave, you frighten me. Testing now, will let it run a while. -Eric
/archives/xfs/2008-07/msg00187.html (9,051 bytes)

12. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 14:13:23 +1000
Excellent. And it happened pretty reliably before, right? Thanks, Tim.
/archives/xfs/2008-07/msg00188.html (9,439 bytes)

13. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 00:17:36 -0400
Independent ofwether it actually fixes the bug (which I think it will) this looks good. Doing anything with the return value from atomic_read except for printing it is most likely bogus, and this one
/archives/xfs/2008-07/msg00189.html (10,420 bytes)

14. Re: deadlocked xfs (score: 1)
Author:
Date: Thu, 10 Jul 2008 23:18:31 -0500
It did, yep, though sometimes took a while. But, w/ the patch it actually deadlocked really quickly (< 2000 inodes written) -Eric
/archives/xfs/2008-07/msg00190.html (9,812 bytes)

15. Re: deadlocked xfs (score: 1)
Author:
Date: Thu, 10 Jul 2008 23:26:58 -0500
ok testing a fixed up version ... will let you know in my AM if it survived. -Eric
/archives/xfs/2008-07/msg00191.html (9,288 bytes)

16. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 14:27:32 +1000
Yeah, inverted the if/else logic, I think. Like I said, untested. I'll send out an updated patch as soon as I get uml running again. Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx
/archives/xfs/2008-07/msg00192.html (10,831 bytes)

17. Re: deadlocked xfs (score: 1)
Author:
Date: Fri, 11 Jul 2008 14:44:58 +1000
Fine, although I didn't write the above :) Gee, they have an atomic_ almost everything :-) When I was talking to Lachlan he was saying it would need to make the atomic_read and the dec somehow atomic
/archives/xfs/2008-07/msg00193.html (13,314 bytes)

18. deadlocked xfs (score: 1)
Author: xxxxxx>
Date: Thu, 10 Jul 2008 21:33:11 -0500
while running fs_mark: fs_mark -d /mnt/test -D 256 -n 100000 -t 4 -s 20480 -F -S 0 which runs 4 threads making 20k files,things got stuffed up in the log, see sysrq-w output at: http://sandeen.fedora
/archives/xfs/2008-07/msg00800.html (7,946 bytes)

19. Re: deadlocked xfs (score: 1)
Author: xxxxxx>
Date: Fri, 11 Jul 2008 12:48:11 +1000
Thanks for the report Eric. This looks very similar to a deadlock Lachlan recently hit in the patch for "Use atomics for iclog reference counting" http://oss.sgi.com/archives/xfs/2008-02/msg00130.htm
/archives/xfs/2008-07/msg00801.html (10,331 bytes)

20. Re: deadlocked xfs (score: 1)
Author: xxxxxx>
Date: Thu, 10 Jul 2008 21:56:48 -0500
thanks I'll look into that. Forgot to mention this was on 2.6.26-rc2 -Eric
/archives/xfs/2008-07/msg00802.html (9,995 bytes)


This search system is powered by Namazu