* [2.5/2.6] buffer layer error at fs/buffer.c:2800 when unlinking
@ 2003-08-03 14:51 Karol Kozimor
2003-08-06 23:59 ` Nathan Scott
0 siblings, 1 reply; 4+ messages in thread
From: Karol Kozimor @ 2003-08-03 14:51 UTC (permalink / raw)
To: linux-kernel
Hi,
I've recently managed to nail down the problems that have been occuring on
my machine at least since 2.5.59. I use XFS as my rootfs, no other
filesystems are compiled in, but I doubt the filesystem is to blame, since
the problem does not appear under 2.4, despite the similar codebase.
Here's the guts: upon unlinking files in /var/{run,lock/subsys} (typically
during shutdown, but not only), and when doing "dd if=/dev/urandom
of=/etc/random-seed count=1 bs=512" the following traces appear:
#v+
buffer layer error at fs/buffer.c:2800
Call Trace:
[<c014f1c6>] drop_buffers+0xb3/0xb9
[<c014f208>] try_to_free_buffers+0x3c/0x96
[<c01db1a1>] linvfs_release_page+0x74/0x78
[<c014d2e6>] try_to_release_page+0x5c/0x6c
[<c014d3d9>] block_invalidatepage+0xe3/0xf6
[<c013a9db>] do_invalidatepage+0x27/0x2b
[<c013aa66>] truncate_complete_page+0x87/0x89
[<c013abc3>] truncate_inode_pages+0xcd/0x2be
[<c01df4f1>] linvfs_unlink+0x5c/0x5e
[<c0160e4a>] generic_delete_inode+0xc0/0xc5
[<c0160fce>] iput+0x55/0x6f
[<c01584bb>] sys_unlink+0x86/0x135
[<c010905f>] syscall_call+0x7/0xb
buffer layer error at fs/buffer.c:2800
Call Trace:
[<c014f1c6>] drop_buffers+0xb3/0xb9
[<c014f208>] try_to_free_buffers+0x3c/0x96
[<c01db1a1>] linvfs_release_page+0x74/0x78
[<c014d2e6>] try_to_release_page+0x5c/0x6c
[<c014d3d9>] block_invalidatepage+0xe3/0xf6
[<c013a9db>] do_invalidatepage+0x27/0x2b
[<c013aa66>] truncate_complete_page+0x87/0x89
[<c013abc3>] truncate_inode_pages+0xcd/0x2be
[<c01df4f1>] linvfs_unlink+0x5c/0x5e
[<c0160e4a>] generic_delete_inode+0xc0/0xc5
[<c0160fce>] iput+0x55/0x6f
[<c01584bb>] sys_unlink+0x86/0x135
[<c010905f>] syscall_call+0x7/0xb
// that's pcmcia: one trace for /var/run/cardmgr.pid, one for
// /var/lock/subsys/pcmcia
buffer layer error at fs/buffer.c:2800
Call Trace:
[<c014f1c6>] drop_buffers+0xb3/0xb9
[<c014f208>] try_to_free_buffers+0x3c/0x96
[<c01db1a1>] linvfs_release_page+0x74/0x78
[<c014d2e6>] try_to_release_page+0x5c/0x6c
[<c014d3d9>] block_invalidatepage+0xe3/0xf6
[<c013a9db>] do_invalidatepage+0x27/0x2b
[<c013aa66>] truncate_complete_page+0x87/0x89
[<c013abc3>] truncate_inode_pages+0xcd/0x2be
[<c016ba43>] padzero+0x28/0x2a
[<c018e6ad>] xfs_bmap_last_offset+0xc2/0x119
[<c018d929>] xfs_bmap_search_extents+0x70/0x85
[<c01b55c1>] xfs_file_last_byte+0xea/0xf7
[<c01b566c>] xfs_itruncate_start+0x9e/0xe2
[<c01d0bfd>] xfs_setattr+0xdd1/0xfc2
[<c01df9d8>] linvfs_setattr+0x11d/0x1c6
[<c0161931>] notify_change+0x150/0x183
[<c0148fd9>] do_truncate+0x4b/0x62
[<c01d0e3b>] xfs_access+0x4d/0x5b
[<c01562e8>] permission+0x46/0x48
[<c01575ec>] may_open+0x169/0x1ba
[<c01576e9>] open_namei+0xac/0x3f3
[<c0149ffd>] filp_open+0x43/0x69
[<c014a3e4>] sys_open+0x5b/0x8b
[<c010905f>] syscall_call+0x7/0xb
// that's dd
buffer layer error at fs/buffer.c:2800
Call Trace:
[<c014f1c6>] drop_buffers+0xb3/0xb9
[<c014f208>] try_to_free_buffers+0x3c/0x96
[<c01db1a1>] linvfs_release_page+0x74/0x78
[<c014d2e6>] try_to_release_page+0x5c/0x6c
[<c014d3d9>] block_invalidatepage+0xe3/0xf6
[<c013a9db>] do_invalidatepage+0x27/0x2b
[<c013aa66>] truncate_complete_page+0x87/0x89
[<c013abc3>] truncate_inode_pages+0xcd/0x2be
[<c01df4f1>] linvfs_unlink+0x5c/0x5e
[<c0160e4a>] generic_delete_inode+0xc0/0xc5
[<c0160fce>] iput+0x55/0x6f
[<c01584bb>] sys_unlink+0x86/0x135
[<c010905f>] syscall_call+0x7/0xb
// that's another one -- probably /var/run/sshd.pid
#v-
Oddly enough, those errors do not appear when I touch and remove the files
manually. I can observe this behaviour on any kernel version between 2.5.59
and 2.6.0-test2. I didn't test earlier versions. The traces come from
2.6.0-test2.
FYI: I never had any fs corruption whatsoever, just those annoying
messages.
I'll be happy to provide any additional information.
Please Cc: me, as I can't handle lkml traffic.
Best regards,
--
Karol 'sziwan' Kozimor
sziwan@hell.org.pl
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [2.5/2.6] buffer layer error at fs/buffer.c:2800 when unlinking
2003-08-03 14:51 [2.5/2.6] buffer layer error at fs/buffer.c:2800 when unlinking Karol Kozimor
@ 2003-08-06 23:59 ` Nathan Scott
2003-08-10 22:16 ` Karol Kozimor
0 siblings, 1 reply; 4+ messages in thread
From: Nathan Scott @ 2003-08-06 23:59 UTC (permalink / raw)
To: Karol Kozimor; +Cc: linux-kernel
On Sun, Aug 03, 2003 at 04:51:13PM +0200, Karol Kozimor wrote:
> Hi,
>
> I've recently managed to nail down the problems that have been occuring on
> my machine at least since 2.5.59. I use XFS as my rootfs, no other
> filesystems are compiled in, but I doubt the filesystem is to blame, since
> the problem does not appear under 2.4, despite the similar codebase.
> ...
> Here's the guts: upon unlinking files in /var/{run,lock/subsys} (typically
> during shutdown, but not only), and when doing "dd if=/dev/urandom
> of=/etc/random-seed count=1 bs=512" the following traces appear:
> #v+
> buffer layer error at fs/buffer.c:2800
> Call Trace:
> [<c014f1c6>] drop_buffers+0xb3/0xb9
> [<c014f208>] try_to_free_buffers+0x3c/0x96
> [<c01db1a1>] linvfs_release_page+0x74/0x78
> [<c014d2e6>] try_to_release_page+0x5c/0x6c
Hi there,
This is indeed an XFS issue (thanks for reporting it), the
patch below fixes it.
cheers.
--
Nathan
--- /usr/tmp/TmpDir.2013-0/linux/fs/xfs/linux/xfs_aops.c_1.45 2003-08-07 09:55:53.000000000 +1000
+++ linux/fs/xfs/linux/xfs_aops.c 2003-08-07 09:22:10.808194848 +1000
@@ -803,7 +803,7 @@
bh = bh->b_this_page;
} while (offset < end_offset);
- if (uptodate)
+ if (uptodate && bh == head)
SetPageUptodate(page);
if (startio)
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [2.5/2.6] buffer layer error at fs/buffer.c:2800 when unlinking
2003-08-06 23:59 ` Nathan Scott
@ 2003-08-10 22:16 ` Karol Kozimor
2003-08-11 0:34 ` Nathan Scott
0 siblings, 1 reply; 4+ messages in thread
From: Karol Kozimor @ 2003-08-10 22:16 UTC (permalink / raw)
To: Nathan Scott; +Cc: linux-kernel, linux-xfs
Thus wrote Nathan Scott:
> This is indeed an XFS issue (thanks for reporting it), the
> patch below fixes it.
Thanks, it works fine now. I've still got one issue with XFS (this: [1] may
be helpful) left, though I didn't manage yet to reproduce it under
2.6.0-test3 (though I've seen it with 2.6.0-test2, even with the above
patch applied), so I'll start bothering you when (if, hopefully) this
happens. Thanks again,
[1] http://marc.theaimsgroup.com/?l=linux-xfs&m=105240964125502&w=2
(I reported this once or twice on linux-xfs, however unsuccessfully)
--
Karol 'sziwan' Kozimor
sziwan@hell.org.pl
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [2.5/2.6] buffer layer error at fs/buffer.c:2800 when unlinking
2003-08-10 22:16 ` Karol Kozimor
@ 2003-08-11 0:34 ` Nathan Scott
0 siblings, 0 replies; 4+ messages in thread
From: Nathan Scott @ 2003-08-11 0:34 UTC (permalink / raw)
To: linux-kernel, linux-xfs
On Mon, Aug 11, 2003 at 12:16:40AM +0200, Karol Kozimor wrote:
> Thus wrote Nathan Scott:
> > This is indeed an XFS issue (thanks for reporting it), the
> > patch below fixes it.
>
> Thanks, it works fine now. I've still got one issue with XFS (this: [1] may
> be helpful) left, though I didn't manage yet to reproduce it under
> 2.6.0-test3 (though I've seen it with 2.6.0-test2, even with the above
> patch applied), so I'll start bothering you when (if, hopefully) this
> happens. Thanks again,
>
> [1] http://marc.theaimsgroup.com/?l=linux-xfs&m=105240964125502&w=2
> (I reported this once or twice on linux-xfs, however unsuccessfully)
A random corruption problem was fixed in test3 also, that may
well be what this issue is. Let us know if you see it again.
thanks.
--
Nathan
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2003-08-11 0:37 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-08-03 14:51 [2.5/2.6] buffer layer error at fs/buffer.c:2800 when unlinking Karol Kozimor
2003-08-06 23:59 ` Nathan Scott
2003-08-10 22:16 ` Karol Kozimor
2003-08-11 0:34 ` Nathan Scott
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).