i enabled page alloc debugging in 2.6.9-rc2.... and hit this BUG. i was
running fsx-odirect, but the callback here shows we are in the normal NFS
cached async write path. a second panic shows we hit this in the NFS
readdir path too, so i suspect this is independent of the NFS client.
dual CPU 1.26GHz P-III system with 1GB RAM.
kernel BUG at include/asm/spinlock.h:90!
invalid operand: 0000 [#1]
SMP DEBUG_PAGEALLOC
Modules linked in:
CPU: 1
EIP: 0060:[<c0396f06>] Not tainted VLI
EFLAGS: 00010202 (2.6.9-rc2)
EIP is at _spin_unlock+0x16/0x30
eax: 00000001 ebx: 00000000 ecx: 00000002 edx: f7eb693c
esi: f7eb6800 edi: f7eb693c ebp: d89d1060 esp: f70a7bb4
ds: 007b es: 007b ss: 0068
Process fsx-odirect (pid: 924, threadinfo=f70a7000 task=f75cf6d0)
Stack: c03305bb f7eb6800 c1adfc80 c1aa0174 c1aa018c f7f2b210 c034478d
d89d1060
f70a7d10 00000090 c1aa0160 f70a7c1c f6baf040 d89d1060 f6baf1bc
c0346922
d89d1060 f70a7d10 f6aeaac4 c037fe1e f6aeabb8 00000000 40000206
00000040
Call Trace:
[<c03305bb>] dev_queue_xmit+0x1db/0x280
[<c034478d>] ip_finish_output+0xbd/0x1c0
[<c0346922>] ip_push_pending_frames+0x2e2/0x430
[<c037fe1e>] rpc_sleep_on+0x3e/0x60
[<c036512e>] udp_push_pending_frames+0x14e/0x280
[<c0365b0e>] udp_sendpage+0xde/0x150
[<c01264e5>] update_process_times+0x45/0x60
[<c011ad50>] autoremove_wake_function+0x0/0x60
[<c036d7f7>] inet_sendpage+0x67/0xd0
[<c036d790>] inet_sendpage+0x0/0xd0
[<c03941f9>] xprt_sock_udp_send_request+0x2b9/0x360
[<c01d0300>] nfs3_xdr_writeargs+0x120/0x130
[<c01d01e0>] nfs3_xdr_writeargs+0x0/0x130
[<c037ebdf>] xprt_transmit+0x4f/0x200
[<c037d4d2>] call_transmit+0x52/0xc0
[<c03803f1>] __rpc_execute+0x101/0x4c0
[<c037ce12>] rpc_call_setup+0x52/0x80
[<c037cb91>] rpc_clnt_sigmask+0x91/0xb0
[<c01cc418>] nfs_execute_write+0x58/0xa0
[<c01cc60c>] nfs_flush_multi+0x1ac/0x270
[<c01cb250>] nfs_writepage+0x0/0x240
[<c01cc95b>] nfs_flush_list+0x6b/0xf0
[<c01cd5ae>] nfs_flush_inode+0x9e/0xb0
[<c01cb52d>] nfs_writepages+0x9d/0x150
[<c013f205>] do_writepages+0x25/0x50
[<c0138930>] __filemap_fdatawrite_range+0xc0/0xd0
[<c0138977>] filemap_fdatawrite+0x37/0x40
[<c014d9f5>] msync_interval+0xb5/0x110
[<c014db98>] sys_msync+0x148/0x152
[<c01044af>] syscall_call+0x7/0xb
Code: 28 01 79 05 e8 98 ef ff ff c3 0f 0b c7 00 4e bb 3a c0 eb ea 90 81 78
04 ad 4e ad de 89 c2 75 15 0f b6 02 84 c0 7f 04 c6 02 01 c3 <0f> 0b 5a 00
4e bb 3a c0 eb f2 0f 0b 59 00 4e bb 3a c0 eb e1 8d
<0>Kernel panic - not syncing: Fatal exception in interrupt
- Chuck Lever
--
corporate: <cel at netapp dot com>
personal: <chucklever at bigfoot dot com>