Search String: Display: Description: Sort:

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*\[2\.6\.27\-rc4\]\s+XFS\s+i_lock\s+vs\s+i_iolock\.\.\.\s*$/: 42 ]

Total 42 documents matching your query.

1. [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author:
Date: Fri, 22 Aug 2008 22:12:59 +0100
On 2.6.27-rc4 with various debug options enabled, lockdep claims lock ordering issues with XFS [1] - easiest reproducer is just running xfs_fsr. Mount options I was using were 'nobarrier,noatime,nodi
/archives/xfs/2008-08/msg00354.html (13,054 bytes)

2. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author:
Date: Mon, 25 Aug 2008 11:02:13 +1000
False positive. We do: xfs_lock_two_inodes(ip, tip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); ..... xfs_iunlock(ip, XFS_ILOCK_EXCL); xfs_iunlock(tip, XFS_ILOCK_EXCL); ..... xfs_lock_two_inodes(ip, tip, XFS_
/archives/xfs/2008-08/msg00368.html (11,250 bytes)

3. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author:
Date: Mon, 25 Aug 2008 12:12:23 +1000
Dave Chinner wrote: On Fri, Aug 22, 2008 at 10:12:59PM +0100, Daniel J Blueman wrote: On 2.6.27-rc4 with various debug options enabled, lockdep claims lock ordering issues with XFS [1] - easiest repr
/archives/xfs/2008-08/msg00372.html (11,859 bytes)

4. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author:
Date: Mon, 25 Aug 2008 13:55:42 +1000
Yeah, that'd work, but it implllies that we no longer allow xfs_lock_two_inodes() to take both inode locks at once. It would need a comment blaming^Wexplaining why lockdep requires us to do this, and
/archives/xfs/2008-08/msg00374.html (11,704 bytes)

5. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author:
Date: Mon, 25 Aug 2008 08:57:44 +0200
Does the annotation I used for double_lock_balance()/double_unlock_balance() work? Basically, it assumes the held lock (this_rq) has subclass 0, but because double_lock_balance() can unlock and reloc
/archives/xfs/2008-08/msg00375.html (11,463 bytes)

6. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author:
Date: Mon, 25 Aug 2008 08:59:33 +0200
How can you take two locks in one go? It seems to me you always need to take them one after another, and as soon as you do that, you have ordering constraints. Of course it could be that doesn't matt
/archives/xfs/2008-08/msg00376.html (11,861 bytes)

7. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: >
Date: Mon, 25 Aug 2008 23:55:32 +0200
Yes, you would. Except that in all other places we only have a single iolock involved, so the ordering of the second iolock and second ilock don't matter. Because of that I think declaring that xfs_l
/archives/xfs/2008-08/msg00385.html (10,114 bytes)

8. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: >
Date: Tue, 26 Aug 2008 11:55:49 +1000
It doesn't take them both inode locks in one go - it does them separately in a given order via xfs_ilock(). Basically there are two layers of constraints here - xfs_ilock() handles the order withing
/archives/xfs/2008-08/msg00389.html (12,421 bytes)

9. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: >
Date: Tue, 26 Aug 2008 12:45:47 +1000
Agreed. Patch below. Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx XFS: prevent lockdep false positives when locking two inodes If we call xfs_lock_two_inodes() to grab both the iolock and the il
/archives/xfs/2008-08/msg00396.html (13,572 bytes)

10. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: x>
Date: Tue, 26 Aug 2008 15:35:08 -0400
Looks good. We probably don't need the #ifdef DEBUG as ASSERT is debug-only anyway.
/archives/xfs/2008-08/msg00405.html (10,609 bytes)

11. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: x>
Date: Tue, 26 Aug 2008 21:13:33 +0100
Hi Dave, Good to get your patch and HCH's ack...thanks! I'll pursue testing and touchdown in < 24 hrs. Daniel -- Daniel J Blueman
/archives/xfs/2008-08/msg00406.html (15,514 bytes)

12. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: x>
Date: Tue, 26 Aug 2008 22:34:40 +0100
Excellent - confirmed it addresses the lockdep report I was seeing before and doesn't introduce any regressions. Thanks, Daniel -- Daniel J Blueman
/archives/xfs/2008-08/msg00407.html (16,220 bytes)

13. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: xx>
Date: Fri, 29 Aug 2008 11:20:27 +1000
/archives/xfs/2008-08/msg00428.html (10,305 bytes)

14. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: xx>
Date: Fri, 29 Aug 2008 11:37:08 +1000
Below. -- Dave Chinner david@xxxxxxxxxxxxx XFS: prevent lockdep false positives when locking two inodes If we call xfs_lock_two_inodes() to grab both the iolock and the ilock, then drop the ilocks on
/archives/xfs/2008-08/msg00429.html (13,484 bytes)

15. [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: xxx>
Date: Fri, 22 Aug 2008 22:12:59 +0100
On 2.6.27-rc4 with various debug options enabled, lockdep claims lock ordering issues with XFS [1] - easiest reproducer is just running xfs_fsr. Mount options I was using were 'nobarrier,noatime,nodi
/archives/xfs/2008-08/msg00802.html (13,054 bytes)

16. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: xxx>
Date: Mon, 25 Aug 2008 11:02:13 +1000
False positive. We do: xfs_lock_two_inodes(ip, tip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); ..... xfs_iunlock(ip, XFS_ILOCK_EXCL); xfs_iunlock(tip, XFS_ILOCK_EXCL); ..... xfs_lock_two_inodes(ip, tip, XFS_
/archives/xfs/2008-08/msg00816.html (11,250 bytes)

17. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: xxx>
Date: Mon, 25 Aug 2008 12:12:23 +1000
Dave Chinner wrote: On Fri, Aug 22, 2008 at 10:12:59PM +0100, Daniel J Blueman wrote: On 2.6.27-rc4 with various debug options enabled, lockdep claims lock ordering issues with XFS [1] - easiest repr
/archives/xfs/2008-08/msg00820.html (11,859 bytes)

18. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: xxx>
Date: Mon, 25 Aug 2008 13:55:42 +1000
Yeah, that'd work, but it implllies that we no longer allow xfs_lock_two_inodes() to take both inode locks at once. It would need a comment blaming^Wexplaining why lockdep requires us to do this, and
/archives/xfs/2008-08/msg00822.html (11,704 bytes)

19. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: xxx>
Date: Mon, 25 Aug 2008 08:57:44 +0200
Does the annotation I used for double_lock_balance()/double_unlock_balance() work? Basically, it assumes the held lock (this_rq) has subclass 0, but because double_lock_balance() can unlock and reloc
/archives/xfs/2008-08/msg00823.html (11,463 bytes)

20. Re: [2.6.27-rc4] XFS i_lock vs i_iolock... (score: 1)
Author: xxx>
Date: Mon, 25 Aug 2008 08:59:33 +0200
How can you take two locks in one go? It seems to me you always need to take them one after another, and as soon as you do that, you have ordering constraints. Of course it could be that doesn't matt
/archives/xfs/2008-08/msg00824.html (11,861 bytes)


This search system is powered by Namazu