2.5.70-mm1 panic with as-scheduler ?

Andrew,

Here is the BUG() we got on 2.5.70-mm1.
Is this a known problem with AS ?

Thanks,
Badari

kernel BUG at drivers/block/as-iosched.c:363!

invalid operand: 0000 [#1]
CPU:    5
EIP:    0060:[<c02b4810>]    Not tainted VLI
EFLAGS: 00010046

EIP is at as_find_arq_hash+0xb0/0xc0
eax: eb950a40 ebx: eb950a40 ecx: eb950a40 edx: eb950a20
esi: 00000000 edi: ce633de8 ebp: f7b5a6d8 esp: eb569cec
ds: 007b es: 007b ss: 0068
Process ftest03 (pid: 29908, threadinfo=eb568000 task=f78006d0)

Stack: 0026ae00 00000000 00000000 00000000 0026ae08 00000000 c02b5bf1 f7b5dda0
       0026ae00 00000000 00000001 00000001 f7b5dda0 00000000 f7b5a800 00000008
       00000000 c02ac769 f7b5a800 eb569d64 d1b309a0 c02af330 f7b5a800 eb569d64
Call Trace:
 [<c02b5bf1>] as_merge+0xf1/0x1d0
 [<c02ac769>] elv_merge+0x29/0x30
 [<c02af330>] __make_request+0x2c0/0x4e0
 [<c02af717>] generic_make_request+0x1c7/0x290
 [<c0159ff0>] submit_bh+0x90/0x1e0
 [<c02af834>] submit_bio+0x54/0xa0
 [<c01584e0>] __block_write_full_page+0x250/0x460
 [<c0159eb8>] block_write_full_page+0xd8/0xe0
 [<c0191df0>] ext3_get_block+0x0/0xb0
 [<c0192a46>] ext3_writepage+0x166/0x280
 [<c0191df0>] ext3_get_block+0x0/0xb0
 [<c01928c0>] bget_one+0x0/0x10
 [<c01788b8>] mpage_writepages+0x238/0x305
 [<c0176b76>] __mark_inode_dirty+0xf6/0x100
 [<c01928e0>] ext3_writepage+0x0/0x280
 [<c0194840>] ext3_setattr+0xf0/0x1a0
 [<c013a88c>] generic_file_writev+0x5c/0x80
 [<c0155656>] do_readv_writev+0x1c6/0x2d0
 [<c013d5c6>] do_writepages+0x36/0x40
 [<c0137be8>] __filemap_fdatawrite+0xd8/0xe0
 [<c0137c07>] filemap_fdatawrite+0x17/0x20
 [<c0156ad3>] sys_fsync+0xb3/0x110
 [<c010943f>] syscall_call+0x7/0xb
 
Code: 1c 3b 70 3c 75 cf 8b 47 14 31 d2 8b 4c 24 04 03 47 0c 8b 34 24 13 57 10 

31
f0 31 d1 89 fa 09 c1 75 c9 eb cf 8d b4 26 00 00 00 00 <0f> 0b 6b 01 cb 4c 44 c0 eb 95 8d b6 00 00 00 00 8b 54 24 08 31

Yes, -mm2 has the fix.

Jens Axboe