Re: slab corruption and stuck CPU under 2.6.8rc2

Re: slab corruption and stuck CPU under 2.6.8rc2

[Ben Collins]

> Jul 27 09:18:44 scratchmonkey kernel: read_lock(00000000006fde40) CPU#0 stuck at 0045ae18, writer PC(00451808):CPU(2)
> Jul 27 09:18:44 scratchmonkey kernel: read_lock(00000000006fde40) CPU#0 stuck at 0045ae18, writer PC(00451808):CPU(2)

Try looking up the 0x006fde40, 0x0045ae18 and 0x00451808 addresses in the
System.map for your kernel. See if we can find out where this is
occuring.

-- 
Debian     - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
WatchGuard - http://www.watchguard.com/

slab corruption and stuck CPU under 2.6.8rc2

[Chris Ricker]

I'm getting the following:

Jul 26 06:25:34 scratchmonkey kernel: Slab corruption: startÿfff80038fd4a80, len\x168
Jul 26 06:25:34 scratchmonkey kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 26 06:25:34 scratchmonkey kernel: Prev obj: startÿfff80038fd49d8, len\x168
Jul 26 06:25:34 scratchmonkey kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 26 06:25:34 scratchmonkey kernel: 010: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 26 06:25:34 scratchmonkey kernel: Next obj: startÿfff80038fd4b28, len\x168
Jul 26 06:25:34 scratchmonkey kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 26 06:25:34 scratchmonkey kernel: 010: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 27 09:18:44 scratchmonkey kernel: Slab corruption: startÿfff8002fda0a80, len\x168
Jul 27 09:18:44 scratchmonkey kernel: read_lock(00000000006fde40) CPU#0 stuck at 0045ae18, writer PC(00451808):CPU(2)
Jul 27 09:18:44 scratchmonkey kernel: read_lock(00000000006fde40) CPU#0 stuck at 0045ae18, writer PC(00451808):CPU(2)
Jul 27 09:18:44 scratchmonkey kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 27 09:18:44 scratchmonkey kernel: Prev obj: startÿfff8002fda09d8, len\x168
Jul 27 09:18:44 scratchmonkey kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 27 09:18:44 scratchmonkey kernel: 010: 6b 6b 6b 6b 6b 6b 6bread_lock(00000000006fde40) CPU#0 stuck at 0045ae18, writer PC(00451808):CPU(2)
Jul 27 09:18:44 scratchmonkey kernel:  6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 27 09:18:44 scratchmonkey kernel: Next obj: startÿfff8002fda0b28, len\x168
Jul 27 09:18:44 scratchmonkey kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b
Jul 27 09:18:44 scratchmonkey kernel: 010: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b

in my logs

Hardware problem or software?

This is on an Ultra 60, 2 300 MHz CPUs, Symbios 53c875 SCSI card, qfe 
ethernet, low-end Sun video (ATI Rage XL-based)

thanks,
chris

Re: slab corruption and stuck CPU under 2.6.8rc2

[Chris Ricker]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1015 bytes --]

On Tue, 27 Jul 2004, Ben Collins wrote:

> > Jul 27 09:18:44 scratchmonkey kernel: read_lock(00000000006fde40) CPU#0 stuck at 0045ae18, writer PC(00451808):CPU(2)
> > Jul 27 09:18:44 scratchmonkey kernel: read_lock(00000000006fde40) CPU#0 stuck at 0045ae18, writer PC(00451808):CPU(2)
> 
> Try looking up the 0x006fde40, 0x0045ae18 and 0x00451808 addresses in the
> System.map for your kernel. See if we can find out where this is
> occuring.

0x006fde40 is tasklist_lock:

00000000006fde00 D kernel_flag
00000000006fde40 D tasklist_lock
00000000006fde80 D mmlist_lock

0x0045ae18 doesnt match exactly:

000000000045ad40 T send_sig_info
000000000045ade0 T send_sig
000000000045ae00 T send_group_sig_info
000000000045ae60 T force_sig
000000000045ae80 T kill_pg

0x00451808 doesn't match exactly:

0000000000451560 T exit_fs
0000000000451680 T exit_mm
00000000004517a0 t exit_notify
0000000000452100 T do_exit
0000000000452580 T complete_and_exit
00000000004525a0 T sys_exit

.config also attached....


thanks,
chris

[-- Attachment #2: Type: APPLICATION/octet-stream, Size: 6425 bytes --]

Re: slab corruption and stuck CPU under 2.6.8rc2

[William Lee Irwin III]

On Tue, 27 Jul 2004, Ben Collins wrote:
>> Try looking up the 0x006fde40, 0x0045ae18 and 0x00451808 addresses in the
>> System.map for your kernel. See if we can find out where this is
>> occuring.

On Tue, Jul 27, 2004 at 10:43:29AM -0400, Chris Ricker wrote:
> 0x006fde40 is tasklist_lock:
[...]
> 0x0045ae18 doesnt match exactly:
[...]
> 0x00451808 doesn't match exactly:
[...]

Just for future reference, they don't need to match exactly. If you can
find the two symbols it goes between, it's generally the lower of those
two.


-- wli