Re: ext2fs corruption on 2.3.7

Jeremy Katz (katzj@linuxpower.org)
Mon, 21 Jun 1999 14:24:58 -0400 (EDT)

On Mon, 21 Jun 1999, Ingo Molnar wrote:
> On Mon, 21 Jun 1999, Jeremy Katz wrote:
>
> > I think that some file system corruption still exists in 2.3.7. I copied
> > over the quake2 pak0.pak from my p200 running 2.3.6 to my dual pIII 450
> > running 2.3.7 and am getting different md5sums between the file on the two
> > machines. I have tested the copy via both ftp and nfs with the same end
> > result (though the md5sum on the 2.3.7 system was different each time).
> >
> > Unfortunately, I haven't the _slightest_ idea of where to start looking
> > for the problem, but I'm willing to test patches or even take a look at it
> > myself if someone can point me in the right direction.
>
> I've attached my latest pagecache fixes, it fixedsa (rare) data corruption
> bug, and also contains fixes for the swap oopses reported earlier today.
> Does this fix your problem?

Well, swap wasn't oops'ing on me before, but it does now. So, I disabled
my swap to test the actual copying stuff. On the positive side, it appears
that the file corruption isn't occuring anymore... I was able to get it
consistently transferring this file before, but I've copied it
successfully over ftp now. A copy across NFS still appears to corrupt the
file.

Now trying to reproduce the oops so I can run it through ksymsoops causes
no problems, both swaps are added fine. It appears that the oops only
occurs when swap is activated on boot. ksymsoops follows:

EFLAGS: 00010006
eax: c02897c0 ebx: 000007e1 ecx: ffffffff edx: 00000001
esi: 0000fc30 edi: c02897f0 ebp: 00000202 esp: cfe2ff70
ds: 0018 es: 0018 ss: 0018
Process initlog (pid: 6, process nr: 7, stackpage=cfe2f000)
Stack: 00000000 c008fce0 fffffff0 c013715b 00000001 0804edb0 cfc3e000 c00837c8
00000287 c0137dc4 cfc3e000 cfc2e000 000001f4 0804eda8 bffffa4c c06f2198
cfe2e000 cfc3e000 c00837c0 00000001 c0109fb8 0804eda8 00000001 000001f4
Call Trace: [<c013715b>] [<c0137dc4>] [<c0109fb8>]
Code: f0 0f bb 1a 19 c0 85 c0 74 3b 8b 54 24 10 f7 da 31 f2 89 d0

Trace: c013715b <free_wait+cb/dc>
Trace: c0137dc4 <sys_poll+15c/1a8>
Trace: c0109fb8 <system_call+34/38>
Code: 00000000 Before first symbol 00000000 <_IP>: <===
Code: 00000000 Before first symbol 0: f0 0f bb 1a lock btcl %ebx,(%edx) <===
Code: 00000004 Before first symbol 4: 19 c0 sbbl %eax,%eax
Code: 00000006 Before first symbol 6: 85 c0 testl %eax,%eax
Code: 00000008 Before first symbol 8: 74 3b je 00000045 Before first symbol
Code: 0000000a Before first symbol a: 8b 54 24 10 movl 0x10(%esp,1),%edx
Code: 0000000e Before first symbol e: f7 da negl %edx
Code: 00000010 Before first symbol 10: 31 f2 xorl %esi,%edx
Code: 00000012 Before first symbol 12: 89 d0 movl %edx,%eax

Jeremy

-- 
Jeremy Katz
http://linuxpower.org




-
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/