Re: 2.2.13pre13 chronic oops

Jarno Paananen (jpaana@s2.org)
09 Oct 1999 17:47:44 +0300


Scott Lampert <fortunato@altair.heavymetal.org> writes:

| We're running 2.2.13pre13 here and keep seeing this oops many times
| throughout a few days:

Again this very same oops that I have got under 2.3.18ac10 and
other people have got under 2.2.13pre15 and 2.2.10 compiled with gcc
2.7.2.3 and egcs 1.1.2. I really does look like a genuine bug in
the dcache or somewhere else.

| Unable to handle kernel paging request at virtual address 74617453
| current->tss.cr3 = 06de9000, %cr3 = 06de9000
| *pde = 00000000
| Oops: 0000
| CPU: 0
| EIP: 0010:[<c012e6a4>]
| EFLAGS: 00010a97
| eax: 000017f8 ebx: 7461743b ecx: 00007240 edx: 00000c39
| esi: 00000000 edi: c55b700d ebp: 74617453 esp: c5d37f0c

Ebp value of 0 has been popping up, I also had value 0x20 once...

| ds: 0018 es: 0018 ss: 0018
| Process in.identd (pid: 30957, process nr: 157, stackpage=c5d37000)
| Stack: c55b700d 00000001 c0213378 c55b700a 00007240 00000003 c0129834 c3962920
| c5d37f54 c5d37f54 c0129a93 c3962920 c5d37f54 00000001 cb2a7320 ffffffe9
| 00000001 bffff580 c55b700a 00000003 00007240 c0129c26 c55b7000 c3962920
| Call Trace: [<c0129834>] [<c0129a93>] [<c0129c26>] [<c01227a4>] [<c01229e6>] [<c
| 01079b8>]
| Code: 8b 6d 00 8b 4c 24 18 39 4b 48 75 54 8b 4c 24 24 39 4b 0c 75
|
|
| Running the latest ksymoops on it shows:
|
| >>EIP; c012e6a4 <d_lookup+5c/d0> <=====
| Trace; c0129834 <cached_lookup+10/54>
| Trace; c0129a93 <lookup_dentry+113/1e8>
| Trace; c0129c26 <open_namei+66/2ec>
| Trace; c01227a4 <filp_open+44/f0>
| Trace; c01229e6 <sys_open+36/94>
| Trace; c01079b8 <system_call+34/38>
| Code; c012e6a4 <d_lookup+5c/d0>
| 00000000 <_EIP>:
| Code; c012e6a4 <d_lookup+5c/d0> <=====
| 0: 8b 6d 00 movl 0x0(%ebp),%ebp <=====

The very same instruction again. I have had different call
chains, but they are the same from open_namei onward.

| This doesn't mean anything to me, but perhaps it can help someone debug
| this. The problematic system's configuration is:
|
| Pentium III 450 - 440BX Chipset
| 256 megs RAM
| Adaptec AHA-2940 Ultra2 SCSI host adapter
| Intel EtherExpress Pro 10/100
| Seagate ST39173LW
| HP C1557A Cartridge Tape Drive

I'm seeing a bit of a pattern here. I also have 2940U2W SCSI
adapter and Peter Hanecak also had SCSI, didn't mention the
adapter type though. If it's also Adaptec, we might have a
common denominator here, if not, it could be the SCSI-layer itself.
I also have eepro, but it should matter, should it? :) I Cc:ed
the Adaptec driver maintainer if he could be of help here.

I have SMP-system and you don't so it shouldn't be a race condition
either.

| I'm going to try bumping the kernel up to the latest 2.2.13preXX
| tonight and see if it helps any.

Most probably doesn't, but you can always try :) I'm trying
2.3.20pre2 myself now.

// Jarno

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/