Distribution: SuSE 8.0 with 2.6.0-test3
Hardware Environment: Described in testplan located at
http://ltp.sf.net/nfs
Software Environment: glibc 2.2.5 and nfs-utils 1.0.5 on server.
Problem Description: One NFS client hit the 1701 bug, while the other hit
the
1702. Both hit after a few hours of performing a loop of random file I/O
system calls over NFS. Afterwards, even though the test was still running,
all
NFS activity had stopped on both clients. Also, a cat /proc/slabinfo
produced
a Segmentation Fault. Here's an example of both bugs:
===========================
kernel BUG at mm/slab.c:1701!
invalid operand: 0000 [#1]
CPU: 0
EIP: 0060:[cache_alloc_refill+258/720] Not tainted
EFLAGS: 00010002
EIP is at cache_alloc_refill+0x102/0x2d0
eax: 00000000 ebx: 00000006 ecx: c74a7000 edx: c74a7018
esi: 00000007 edi: 00000004 ebp: c43dfcd8 esp: c43dfca4
ds: 007b es: 007b ss: 0068
Process fsstress (pid: 26441, threadinfo=c43de000 task=cf139940)
Stack: 00000000 cfdaee98 00000286 cfda6958 00000006 c74a7000 00000006
cfdaeea4
cfdaeea4 cfdaeeac 00001000 c74a7018 cfda6948 c43dfcfc c0130ea4
cfdaee98
000000d0 00000000 cefd8b80 cfe82fec c0426b98 c43dfd1c c43dfd0c
c019719b
Call Trace:
[kmem_cache_alloc+76/280] kmem_cache_alloc+0x4c/0x118
[nfs_alloc_inode+19/64] nfs_alloc_inode+0x13/0x40
[alloc_inode+22/332] alloc_inode+0x16/0x14c
[get_new_inode+18/196] get_new_inode+0x12/0xc4
[iget5_locked+125/136] iget5_locked+0x7d/0x88
[nfs_find_actor+0/176] nfs_find_actor+0x0/0xb0
[nfs_init_locked+0/60] nfs_init_locked+0x0/0x3c
[__nfs_fhget+95/808] __nfs_fhget+0x5f/0x328
[nfs_find_actor+0/176] nfs_find_actor+0x0/0xb0
[nfs_init_locked+0/60] nfs_init_locked+0x0/0x3c
[nfs_fhget+67/76] nfs_fhget+0x43/0x4c
[nfs_instantiate+87/172] nfs_instantiate+0x57/0xac
[nfs_symlink+325/432] nfs_symlink+0x145/0x1b0
[call_timeout+32/264] call_timeout+0x20/0x108
[nfs_permission+284/392] nfs_permission+0x11c/0x188
[permission+38/60] permission+0x26/0x3c
[vfs_symlink+95/136] vfs_symlink+0x5f/0x88
[sys_symlink+121/204] sys_symlink+0x79/0xcc
[syscall_call+7/11] syscall_call+0x7/0xb
Code: 0f 0b a5 06 87 8a 38 c0 89 f6 85 c0 7c 04 39 d8 72 08 0f 0b
===========================
kernel BUG at mm/slab.c:1702!
invalid operand: 0000 [#1]
CPU: 0
EIP: 0060:[cache_alloc_refill+276/720] Not tainted
EFLAGS: 00010086
EIP is at cache_alloc_refill+0x114/0x2d0
eax: cfdb875c ebx: 00000006 ecx: cfdb1eac edx: cfdb1ec4
esi: 00000002 edi: 00000007 ebp: c345fcd8 esp: c345fca4
ds: 007b es: 007b ss: 0068
Process fsstress (pid: 15862, threadinfo=c345e000 task=c7ce59c0)
Stack: 00000000 cfdb1e98 00000286 cfdb0958 00000006 cfdb1eac 00000006
cfdb1ea4
cfdb1ea4 cfdb1eac 00000000 cfdb1ec4 cfdb0948 c345fcfc c0130ea4
cfdb1e98
000000d0 00000000 cf034b80 cfe8c080 c0426bb0 c345fd1c c345fd0c
c019719b
Call Trace:
[kmem_cache_alloc+76/280] kmem_cache_alloc+0x4c/0x118
[nfs_alloc_inode+19/64] nfs_alloc_inode+0x13/0x40
[alloc_inode+22/332] alloc_inode+0x16/0x14c
[get_new_inode+18/196] get_new_inode+0x12/0xc4
[iget5_locked+125/136] iget5_locked+0x7d/0x88
[nfs_find_actor+0/176] nfs_find_actor+0x0/0xb0
[nfs_init_locked+0/60] nfs_init_locked+0x0/0x3c
[__nfs_fhget+95/808] __nfs_fhget+0x5f/0x328
[nfs_find_actor+0/176] nfs_find_actor+0x0/0xb0
[nfs_init_locked+0/60] nfs_init_locked+0x0/0x3c
[nfs_fhget+67/76] nfs_fhget+0x43/0x4c
[nfs_instantiate+87/172] nfs_instantiate+0x57/0xac
[nfs_mknod+172/232] nfs_mknod+0xac/0xe8
[parse_rock_ridge_inode_internal+696/1504]
parse_rock_ridge_inode_internal+0x2b8/0x5e0
[nfs_permission+354/392] nfs_permission+0x162/0x188
[permission+38/60] permission+0x26/0x3c
[vfs_mknod+166/208] vfs_mknod+0xa6/0xd0
[sys_mknod+283/372] sys_mknod+0x11b/0x174
[syscall_call+7/11] syscall_call+0x7/0xb
Code: 0f 0b a6 06 87 8a 38 c0 8b 55 f8 8b 04 82 83 f8 ff 75 d5 8b
===========================
Steps to reproduce: See the http://ltp.sf.net/nfs for the test plan and
location of the test, fsstress, that produces the random file i/o related
system calls. I'm thinking this may be an NFS issue, rather than slab, so
I
decided to start here first.
-Robbie
Robert V. Williamson <[email protected]>
Linux Test Project
IBM Linux Technology Center
Web: http://ltp.sourceforge.net
IRC: #ltp on freenode.irc.net
====================
"Only two things are infinite, the universe and human stupidity, and I'm
not sure about the former." -Albert Einstein
-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs