mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [Bug 1549] New: /bin/sync hangs running reaim on 8-way x86 using ext2 file system
@ 2003-11-18  1:06 Martin J. Bligh
  0 siblings, 0 replies; only message in thread
From: Martin J. Bligh @ 2003-11-18  1:06 UTC (permalink / raw)
  To: linux-kernel

http://bugme.osdl.org/show_bug.cgi?id=1549

           Summary: /bin/sync hangs running reaim on 8-way x86 using ext2
                    file system
    Kernel Version: 2.6.0-test9-bk19
            Status: NEW
          Severity: normal
             Owner: akpm@digeo.com
         Submitter: dmo@osdl.org


Distribution:rh9
Hardware Environment:8-processor, 700 mgz pentium III, 8 gibabytes memory
Software Environment:reaim test:http://sourceforge.net/projects/re-aim-7
Problem Description: test hangs because /bin/sync never completes.  The sync
binaries are usually hung either trying to lock a page in the page cache (this
is the most common case):
sync          D DA1D1E3C  4048 13674      1         13807 13806 (NOTLB)
da1d1e48 00000082 da1d0000 da1d1e3c c0231e48 f70f6c00 f69c6a00 c0226e16
       f69c6a00 c633cbc0 f69c6a00 f69c6a00 dd7d0040 c633cbc0 000038ea ef356c53
       00000092 f700e040 c633cbc0 c11a2908 c6281bf4 da1d1e54 c011e7f8 00000000
Call Trace:
 [<c0231e48>] deadline_next_request+0x28/0x40
 [<c0226e16>] elv_next_request+0x16/0x110
 [<c011e7f8>] io_schedule+0x28/0x40
 [<c013e8ef>] __lock_page+0xbf/0xf0
 [<c011fb10>] autoremove_wake_function+0x0/0x50
 [<c011fb10>] autoremove_wake_function+0x0/0x50
 [<c0185748>] mpage_writepages+0x318/0x350
 [<c0167ca0>] blkdev_writepage+0x0/0x30
 [<c01691ff>] generic_writepages+0x1f/0x23
 [<c014447e>] do_writepages+0x1e/0x40
 [<c013e209>] __filemap_fdatawrite+0xe9/0x110
 [<c013e247>] filemap_fdatawrite+0x17/0x20
 [<c0160ac6>] sync_blockdev+0x26/0x50
 [<c01840d9>] sync_inodes+0x89/0xa0
 [<c0160c54>] do_sync+0x44/0x70
 [<c0160c8f>] sys_sync+0xf/0x20
 [<c0109729>] sysenter_past_esp+0x52/0x71

This happens with both deadline and anticipatory IO schedulers.
It happens about once every 3 or 4 runs of the "database" form of the
test.

Steps to reproduce:

submit the reaim test to an stp8 machine.  specify no arguments to the reaim
will run the database test form.  It may take 3 or 4 runs to observe this
test failure.  Watch for tests that run for more than 3 hours. This also happens
less frequently on a four-way machine.


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2003-11-18  1:06 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-11-18  1:06 [Bug 1549] New: /bin/sync hangs running reaim on 8-way x86 using ext2 file system Martin J. Bligh

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®