Skip to contentSkip to conversation

oops report: bk-head, reiser, sklin98

DISCUSSION Ray Lehtiniemi 1 participant 1 message Feb 16, 2005

lore
Share & tools
lore
#0 lore raw

hi all

i found an oops on my screen this morning. in all likelihood, it's just a delayed crash due to some network experimentation i was doing yesterday, but it seemed severe enough to post to the list, just in case it's real.

kernel BUG at fs/buffer.c:2890!
invalid operand: 0000 [#1]
SMP

   Modules linked in: pktgen
   CPU:    0
   EIP:    0060:[<c015a54c>]    Not tainted VLI
   EFLAGS: 00010246   (2.6.11-rc3-rayl-smp)
   EIP is at try_to_free_buffers+0x8c/0xb0
   eax: 20000000   ebx: c12719a0   ecx: ddf7a768   edx: 00000001
   esi: 00000001   edi: 00000001   ebp: dea2fefc   esp: dea2fe2c
   ds: 007b   es: 007b   ss: 0068
   Process ntpd (pid: 1939, threadinfo=dea2e000 task=ddf46a60)
   Stack: f0000000 c12719a0 00000001 00000001 c019aa9c 00000060 00000001 fffffffb
          ddf7a6bc c019c1d1 000029e6 00000000 00000001 00000060 dea2fefc fffffffb
          00000000 ded941b4 00000000 01d94180 00000086 fffffffb 00000000 b7e3e000
   Call Trace:
    [<c019aa9c>] reiserfs_unprepare_pages+0x3c/0x70
    [<c019c1d1>] reiserfs_file_write+0x621/0x630
    [<c016d00c>] d_alloc+0x12c/0x180
    [<c01f8d1a>] _atomic_dec_and_lock+0x2a/0x60
    [<c016c67a>] dput+0x18a/0x1c0
    [<c01629fa>] path_release+0xa/0x30
    [<c0164549>] open_namei+0xc9/0x600
    [<c0154d17>] filp_open+0x27/0x50
    [<c0155d13>] vfs_write+0xa3/0x120
    [<c0155e41>] sys_write+0x41/0x70
    [<c01024ad>] sysenter_past_esp+0x52/0x75
   Code: 58 04 e8 38 02 00 00 3b 1c 24 89 d8 75 f1 89 f2 8b 5c 24 04 89 d0 8b 74 24 08 8b 7c 24 0c 83 c4 10 c3 89 d8 e8 a6 3a fe ff eb bb <0f> 0b 4a 0b 56 40 37 c0 eb 84 89 e2 89 d8 e8 c1 fe ff ff 89 c6

upon typing 'reboot', i was rewarded with a stream of:

linux:~ # reboot

ide: failed opcode was: unknown
hda: drive not ready for command
ide0: reset: success
hda: status error: status=0x10 { SeekComplete }

ide: failed opcode was: unknown
end_request: I/O error, dev hda, sector 3076147
hda: drive not ready for command
hda: status error: status=0x10 { SeekComplete }

the box is currently having problems rebooting.... i'm just looking into whether the drive got corrupted, or whether this new chassis has stability issues, since i've only had it running for a week or so.

this is on a nexcom nsa1086 chassis with a single P4 hyperthreaded processor. i'm running bk-head from yesterday, together with an updated sklin98 patch which i received from my vendor, nexcom.com. it is version 8.12 beta 01, dated dec 08, 2004. the install.sh script produced a patch which seemed to apply cleanly to bk-head.

it seems very likely this oops is caused by this driver... yesterday, i noticed that four of the gig ethernet ports managed by this driver were producing an infinite stream of interrupts (several hundred thousand per second, according to /proc/interrupts...). also, the pktgen.ko module produced some BUG()-type warnings on several occasions, which unfortunately i didn't record.

please cc me if it warrants a reply.

thanks

Signature · 2 lines
----------------------------------------------------------------------
     Ray L   <[email protected]>