First I must say I have not sent anythink like that to LKML yet,
but gonna do my best to give as much and as good information as
I can.
Background:
I am using Debian unstable/SID. Because of some problems it
sometimes shutsdown uncleanly - after logiing out from GNOME
computer just freezes. It is unstable, normally I am using
NVidia binary driver and VMware binary kernel modules so I
wasn't surprised by this.
But I also use ReiserFS on LVM (only root is outside LVM),
so I _was_ surprised when filesystem related problems came up.
It was so bad I had to use fsck.reiserfs --rebuild-tree on
unmounted partitions in single user mode.
Problem:
Today I was doing dist-upgrade - it froze my computer.
I *disabled loading all binary modules*, restarted my machine
and run the process again. Below You can see what it caused:
Configuring bison (1.31-2)...
vs-7050: new entry is found, new inode == 0
invalid operand: 0000
CPU: 0
EIP: 0010:[<d081c239>] Not tainted
EFLAGS: 00010286
eax: 0000002f ebx: d082bea0 ecx: ffffe061 edx: c4aea000
esi: sffe8000 edi: ffffffef ebp: c4aebf0c esp: c4aebcec
ds: 0018 es: 0018 ss: 0018
Process update-alternat (pid: 542, stacpage = c4aeb000)
Stack: d082d79a d0831b20 d082bea0 c4aebd10 c4a129a0 caebdd8 d0812299
cffe8000
d082bea0 c4a12a1c c4a12a1c c4a129a0 c49ac060 00000000 00000000 00000000
00000000 00000000 00000000 00000000 d08301cf 00000000 00000039 00013850
Calltrace: [<d082d79a>] [<d0831b20>] [<d082bea0>] [<d0812299>]
[<d082bea0>]
[<d08301cf>] [<d0828baf>] [<c013cd99>] [<c013cf96>] [<c0106f43>]
Code: 0f 0b 68 20 1b 83 d0 85 f6 74 0f 0f b7 46 08 50 e8 ba 4d 91
/var/lib/dpkg/info/bison.postinst: line 10: 542 Segmentation fault
the 10th line of the script is just "fi"
Other stuff:
I am also getting strange errors like denying access to some files
while doing ls on a directory content - as root of course.
System data:
kernel 2.4.18-pre3 with latest LVM patch and latest kernel preemtion
patch (as of 15.I.2002)
Own thoughts:
I think that's all. Just wanna say I a bit dissapointed with reiserfs.
When installing it (took me 2 days to do the transition from lilo, ext2
to grub, LVM and ReiserFS ;-) I thought it is meant to be immune against
system crashes and power failures.
Now gotta go into single user mode now and --rebuild-tree again on
every each partition... Maybe anyone has a better idea ?
I will be glad to see any responses, please Cc: them to me if You
please ( mailto:greg@sente.pl ) - or give me an address of any _live_
LKML news gateway.
Best regards
Grzegorz Prokopski
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/