[2.2.11] Multiple oopses in d_alloc

Florian Weimer (fw@s.netic.de)
18 Aug 1999 09:07:59 +0200


Configuration: vanilla 2.2.11, compiled with gcc-2.95, Intel Pentium
266 MHz, 256 MB RAM, Flashpoint SCSI controller, one IDE drive (only
used for irrelevant data, a VFAT filesystem was mounted using autofs),
custom sound driver module was loaded but inactive for hours (I don't
think it's the culprit), NFS compiled into the kernel, but nothing has
been mounted since the last boot. At the time of the oops, the system
wasn't loaded at all.

There weren't any problems before (except reproducible crashes in the
buffer management code of Emacs 20.4, which, due to their reproducible
nature, I wouldn't attribute to flaky hardware.)

Here's the oops, fed through ksymoops with some problems:

Aug 16 18:31:10 deneb kernel: Unable to handle kernel paging request at virtual address 2ed14464
Aug 16 18:31:10 deneb kernel: current->tss.cr3 = 0e12c000, %cr3 = 0e12c000
Aug 16 18:31:10 deneb kernel: *pde = 00000000
Aug 16 18:31:10 deneb kernel: Oops: 0002
Aug 16 18:31:10 deneb kernel: CPU: 0
Aug 16 18:31:10 deneb kernel: EIP: 0010:[d_alloc+134/340]
Aug 16 18:31:10 deneb kernel: EFLAGS: 00010207
Aug 16 18:31:10 deneb kernel: eax: 0000000f ebx: 2ed14404 ecx: 00000003 edx: c2319ef4
Aug 16 18:31:10 deneb kernel: esi: c1a90040 edi: 2ed14464 ebp: c4ff5260 esp: c2319e7c
Aug 16 18:31:10 deneb kernel: ds: 0018 es: 0018 ss: 0018
Aug 16 18:31:10 deneb kernel: Process nnrpd (pid: 7796, process nr: 51, stackpage=c2319000)
Aug 16 18:31:10 deneb kernel: Stack: c4ff9760 c2319ef4 c2319ef4 c1a90040 0d0fa727 2ed14464 ce4094c0 c2319f64
Aug 16 18:31:10 deneb kernel: 00001fd8 c012bb6b c4ff5260 c2319ef4 00000001 00000206 c2319ef4 00000000
Aug 16 18:31:10 deneb kernel: c1a9004f 00000001 c2319eec c2319f64 c2319ef4 c012bd7c c4ff5260 c2319ef4
Aug 16 18:31:10 deneb kernel: Call Trace: [real_lookup+63/140] [lookup_dentry+304/504] [open_namei+105/776] [filp_open+74/264] [filp_open+17/264] [sys_open+63/144] [sys_open+38/144]

Despite the warning messages, the disassembled listing is correct
(I verified it by hand):

Aug 16 18:31:10 deneb kernel: Code: f3 a5 a8 02 74 02 66 a5 a8 01 74 01 a4 8b 54 24 34 8b 7c 24
Warning: trailing garbage ignored on Code: line
Text: 'Code: f3 a5 a8 02 74 02 66 a5 a8 01 74 01 a4 8b 54 24 34 8b 7c 24 '
Garbage: ' '

Code: 00000000 Before first symbol 0000000000000000 <_IP>: <===
Code: 00000000 Before first symbol 0: f3 a5 repz movsl %ds:(%esi),%es:(%edi) <===
Code: 00000002 Before first symbol 2: a8 02 testb $0x2,%al
Code: 00000004 Before first symbol 4: 74 02 je 00000008 Before first symbol
Code: 00000006 Before first symbol 6: 66 a5 movsw %ds:(%esi),%es:(%edi)
Code: 00000008 Before first symbol 8: a8 01 testb $0x1,%al
Code: 0000000a Before first symbol a: 74 01 je 0000000d Before first symbol
Code: 0000000c Before first symbol c: a4 movsb %ds:(%esi),%es:(%edi)
Code: 0000000d Before first symbol d: 8b 54 24 34 movl 0x34(%esp,1),%edx
Code: 00000011 Before first symbol 11: 8b 7c 24 00 movl 0x0(%esp,1),%edi

More oopses followed, for example at the movsb instruction above, killing
almost every process which tried lookup something in the file system.

After a reboot, I couldn't reproduce the problem.

-
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/