Re: Steve's crashing 1.3 machine, cured?
Linus Torvalds (torvalds@cs.helsinki.fi)
Sat, 13 Apr 1996 14:07:42 +0300 (EET DST)
On 12 Apr 1996, Steven L Baur wrote:
>=20
> The machine I have that was having so much trouble with Linux 1.3, bu=
t
> not Linux 1.2 is cured, I think. I've been running 1.3.85 most of th=
e
> week (3 days, 7 hours & counting) with modules enabled and the latest
> version of __generic_memcpy_tofs. For the most part it has run
> smoothly.
>=20
> I just got an oops from insmod (called from kerneld):
>=20
> Apr 12 18:51:00 deanna kernel: Unable to handle kernel NULL pointer d=
ereference at virtual address c0000000
> Apr 12 18:51:00 deanna kernel: current->tss.cr3 =3D 01053000, xr3 =3D=
01053000
> Apr 12 18:51:00 deanna kernel: *pde =3D 00102067
> Apr 12 18:51:00 deanna kernel: *pte =3D 00000027
> Apr 12 18:51:00 deanna kernel: Oops: 0000
> Apr 12 18:51:00 deanna kernel: CPU: 0
> Apr 12 18:51:00 deanna kernel: EIP: 0010:[<02836554>]
> Apr 12 18:51:00 deanna kernel: EFLAGS: 00010212
> Apr 12 18:51:00 deanna kernel: eax: 02836534 ebx: 01496018 ecx: 0=
0000000 edx: 01496614
> Apr 12 18:51:00 deanna kernel: esi: 000000be edi: 00dee118 ebp: 0=
000111d esp: 00ddef2c
> Apr 12 18:51:00 deanna kernel: ds: 0018 es: 0018 fs: 002b gs: 0=
02b ss: 0018
> Apr 12 18:51:00 deanna kernel: Process insmod (pid: 2109, process nr:=
82, stackpage=3D00dde000)
> Apr 12 18:51:00 deanna kernel: Stack: 00113b11 00564c0c bffff4b0 0804=
7918 bffff424 00101c2c 0186b063 0011d57d=20
> Apr 12 18:51:00 deanna kernel: c2c00000 00017000 000005fc 0000=
0000 01496018 01496614 000008a4 00000010=20
> Apr 12 18:51:00 deanna kernel: 00dee118 02836534 028365d0 6e75=
6f73 00950064 0095f31d 0095f71c 00000001=20
> Apr 12 18:51:01 deanna kernel: Call Trace: [sys_init_module+1013/1072=
] [vmalloc+69/100] [<02836534>] [<028365d0>] [do_bottom_half+59/96] [sy=
stem_call+89/160]=20
> Apr 12 18:51:01 deanna kernel: Code: 00 01 00 01 02 01 00 01 00 01 00=
01 00 01 00 01 00 01 00 01=20
> Apr 12 18:51:01 deanna kernel: Detected scsi CD-ROM sr0 at scsi0, cha=
nnel 0, id 5, lun 0
>=20
> Is this a rational oops this time?
Looks rational. It also looks like some module did something bad (you=20
can't call vmalloc() from a bottom half handler, and that is what it=20
looks like it's doing).
> This is from starting xmixer (prior to accessing an audio cd). Sound
> support is compiled as a module, and I got an earlier message:
>=20
> Apr 11 10:18:29 deanna insmod: A module named sound already exists
> Apr 11 10:18:30 deanna kernel: Detected scsi CD-ROM sr0 at scsi0, cha=
nnel 0, id 5, lun 0
>=20
> Which is abnormal -- I don't think the first message should have occu=
rred.
Sounds like a bad kerneld/module interaction. Maybe the sound module wa=
s=20
busy being loaded when kerneld started loading it again, or something=20
like this. Bj=F6rn, is there anything that protects against that kind o=
f=20
"re-entrancy" problem?
Linus