Kalin KOZHUHAROV wrote:
> Hi, there.
> Long time no posting - didn't have kernel problems for long time :-)
> 
> That is why I am still running 2.6.11.11 (2.6.12 elsewhere). Will move
> to 2.6.13 soon.
> 
> Yesterday just bought a new SATAII drive (Seagate Barracuda 7200.8
> ST3300831AS) and while trying to rsync some data from the old drives the
> rsync process died with segfault. My SiI3112 controller is not SATAII,
> but it should work in SATA mode, have another drive for year+. Looking
> at the dmesg I saw 3 oopses (see the shortened .dmesg file). Run the
> ksymoops and got some output (see .ksymoops.bz2).
> 
> Although it does not seem very related to the drive, that is the only
> recent change in hardware, in software: udev . The machine (MB: A7V8X
> Deluxe) was working stable for 6 months with a few restarts.
> 
> As far as reproducibility goes, apart from those 3 oopses everything is
> OK, didn't even have to restart and am now continuing to rsync some
> 200GB more.
> 
> Any ideas as to what caused this?

OK, I upgraded to the latest 2.6.13 kernel and still got (similar?) oops.

Looking again at it it might be NFS (using v4 recently) related.

Will provide config if needed.

Attaching the new oops.


Any ideas what are all the warnings from ksymoops like that:

...
Warning (compare_maps): vmlinux symbol __crc_IO_APIC_get_PCI_irq_vector 
not found in System.map.  Ignoring System.map entry
Warning (compare_maps): vmlinux symbol __crc_I_BDEV not found in 
System.map.  Ignoring System.map entry
Warning (compare_maps): vmlinux symbol __crc_SELECT_DRIVE not found in 
System.map.  Ignoring System.map entry
...

Running ksymoops without "-v" option clears them, but... I shall I omit 
the "-m" option when I use "-v"?

BTW, runnig with CONFIG_PRINTK_TIME=y cannot pipe directly into ksymoops 
:-( Disabling.

Kalin.

-- 
|[ ~~~~~~~~~~~~~~~~~~~~~~ ]|
+-> http://ThinRope.net/ <-+
|[ ______________________ ]|