[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <18399.36935.640758.796880@frecb006361.adech.frec.bull.fr>
Date: Tue, 18 Mar 2008 10:49:59 +0100
From: Solofo.Ramangalahy@...l.net
To: linux-ext4@...r.kernel.org
Subject: [2.6.25-rc5-ext4-36c86] attempt to access beyond end of device
Hello,
During stress testing (workload: racer from ltp + fio/iometer), here
is an error I am encountering:
8<------------------------------------------------------------------------------
kernel: WARNING: at fs/buffer.c:1680 __block_write_full_page+0xd4/0x2af()
kernel: Modules linked in:
kernel: Pid: 13113, comm: dd Not tainted 2.6.25-rc5-ext4-36c86 #2
kernel:
kernel: Call Trace:
kernel: [<ffffffff8022dfb9>] warn_on_slowpath+0x51/0x63
kernel: [<ffffffff80317b48>] jbd2_journal_stop+0x24b/0x257
kernel: [<ffffffff802fe73d>] __ext4_journal_stop+0x1f/0x3d
kernel: [<ffffffff802f8cf6>] ext4_da_get_block_write+0x136/0x147
kernel: [<ffffffff803182d5>] start_this_handle+0x32a/0x45b
kernel: [<ffffffff8029cf48>] __block_write_full_page+0xd4/0x2af
kernel: [<ffffffff802f74f5>] ext4_get_block+0x0/0xe2
kernel: [<ffffffff802f8ede>] ext4_writeback_writepage+0xa8/0xf5
kernel: [<ffffffff802a2558>] __mpage_writepage+0x4c3/0x4ff
kernel: [<ffffffff802fe73d>] __ext4_journal_stop+0x1f/0x3d
kernel: [<ffffffff802f8cf6>] ext4_da_get_block_write+0x136/0x147
kernel: [<ffffffff80259375>] find_get_pages+0x2f/0x6d
kernel: [<ffffffff802a2667>] mpage_da_submit_io+0xd3/0x12d
kernel: [<ffffffff802f8bc0>] ext4_da_get_block_write+0x0/0x147
kernel: [<ffffffff802a2749>] mpage_da_writepages+0x88/0x94
kernel: [<ffffffff8025ad35>] __rmqueue+0x1b/0x1bd
kernel: [<ffffffff8031c531>] jbd2_journal_cancel_revoke+0x125/0x166
kernel: [<ffffffff802f8bc0>] ext4_da_get_block_write+0x0/0x147
kernel: [<ffffffff8025d5e3>] do_writepages+0x20/0x2d
kernel: [<ffffffff80297b79>] __writeback_single_inode+0x140/0x294
kernel: [<ffffffff80298058>] sync_sb_inodes+0x1c2/0x2ad
kernel: [<ffffffff8029848d>] writeback_inodes+0x67/0xb8
kernel: [<ffffffff8025da55>] balance_dirty_pages_ratelimited_nr+0x186/0x2f8
kernel: [<ffffffff80258119>] generic_file_buffered_write+0x1f5/0x622
kernel: [<ffffffff802f6091>] ext4_mark_iloc_dirty+0x459/0x4cc
kernel: [<ffffffff80317b48>] jbd2_journal_stop+0x24b/0x257
kernel: [<ffffffff80258889>] __generic_file_aio_write_nolock+0x343/0x377
kernel: [<ffffffff80258921>] generic_file_aio_write+0x64/0xc4
kernel: [<ffffffff802f42c3>] ext4_file_write+0xa0/0x125
kernel: [<ffffffff8027e4e0>] do_sync_write+0xd1/0x118
kernel: [<ffffffff8023f2c5>] autoremove_wake_function+0x0/0x2e
kernel: [<ffffffff8027ec31>] vfs_write+0xad/0x136
kernel: [<ffffffff8027f166>] sys_write+0x45/0x6e
kernel: [<ffffffff8020b04b>] system_call_after_swapgs+0x7b/0x80
kernel:
kernel: ---[ end trace 05e1c77d40267d33 ]---
kernel: attempt to access beyond end of device
kernel: sdc: rw=1, want=355304755152, limit=16064512
kernel: lost page write due to I/O error on sdc
8<------------------------------------------------------------------------------
at this point, there is a flurry of such messages, flooding the
console, eventually needing hard reboot.
This was also observed with linux-2.6.25-rc4-ext4-1.
So far, I have not observed it with "-o mballoc,extents,nodelalloc".
Discussing it on irc:
. the filesystem was not full (hence this is probably not due to
"delalloc not having block reservation").
. using CONFIG_DEBUG_PAGEALLOC and CONFIG_SLAB_DEBUG did not help
diagnostic.
. "attempt to access beyond end of device" may have multiple causes.
I will continue to diagnose, but appreciate any further insight.
Attached .config and /var/log/messages,
--
solofo
View attachment "config-2.6.25-rc5-ext4-36c86" of type "text/plain" (36623 bytes)
View attachment "messages" of type "text/plain" (28033 bytes)
Powered by blists - more mailing lists