kernel panic in slab.c

kernel panic in slab.c

[Hubert Figuiere]

Hi,

I'm getting kernel panics (during boot up or 5 minutes after, it depends
of the wind, the external temperature, or how old is the captain ;-) )
when I activate CONFIG_SLAB_DEBUG.

Here is the output:
kernel BUG at slab.c:1263!
Oops: Exception in kernel mode, sig: 4

Anybody have had the same problem ?
Disabling DEBUG is fine, but I was really wondering why I was getting
this.

kernel is linux-2.4-devel BK tree in version 2.4.23-pre5


Hub
--
Hubert Figuière - Freebox SA. - +33 1 73 50 2 563
AIM/Yahoo!: hfiguiere, ICQ: 307453487

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: kernel panic in slab.c

[Benjamin Herrenschmidt]

On Thu, 2003-09-25 at 14:31, Hubert Figuiere wrote:
> Hi,
>
> I'm getting kernel panics (during boot up or 5 minutes after, it depends
> of the wind, the external temperature, or how old is the captain ;-) )
> when I activate CONFIG_SLAB_DEBUG.
>
> Here is the output:
> kernel BUG at slab.c:1263!
> Oops: Exception in kernel mode, sig: 4
>
> Anybody have had the same problem ?
> Disabling DEBUG is fine, but I was really wondering why I was getting
> this.
>
> kernel is linux-2.4-devel BK tree in version 2.4.23-pre5

Hi Hubert !

Nice to see you using Linux again :)

What machine is this ?

The bug shows that the poison checker found that a freed block got
overriden. It didn't display more infos about the kind of pattern
it found and/or a backtrace ? That would be helpful.

I've been trying to track down such kind of memory corruption for
some time now, so if you have a good handle on it, I'm interested,
as I suspect something wrong is going on in recent 2.4s

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: kernel panic in slab.c

[Hubert Figuiere]

Hi,

On Thu, 2003-09-25 at 15:44, Benjamin Herrenschmidt wrote:

> Nice to see you using Linux again :)

I never stopped :-)

>
> What machine is this ?

Custom network hardware with a PrPMC750 Motorola board as MPU. We have a
few tiny patches for our custom devices, but nothing important here
IMHO.


> The bug shows that the poison checker found that a freed block got
> overriden. It didn't display more infos about the kind of pattern
> it found and/or a backtrace ? That would be helpful.

I might investigate.

> I've been trying to track down such kind of memory corruption for
> some time now, so if you have a good handle on it, I'm interested,
> as I suspect something wrong is going on in recent 2.4s

Fine, then I feel like investigating it since memory corruption are
probably the worst bugs. I don't have a good handle on the problem. I
often crash while running either "bird" or "swapper" (uptime don't
exceed 5 or 10 minutes). Even worse, sometime it crashes during boot up.
The only thing I'm sure is that it is fully reproducible as it plagged
me for quite sometime. I first thaught that hardware was at fault (like
defective MPU), but my co-workers don't have the problem, and in fact
disabled debugging.


Thanks

Hub
--
Hubert Figuière - Freebox SA. - +33 1 73 50 2 563
AIM/Yahoo!: hfiguiere, ICQ: 307453487

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: kernel panic in slab.c

[Benjamin Herrenschmidt]

> Fine, then I feel like investigating it since memory corruption are
> probably the worst bugs. I don't have a good handle on the problem. I
> often crash while running either "bird" or "swapper" (uptime don't
> exceed 5 or 10 minutes). Even worse, sometime it crashes during boot up.
> The only thing I'm sure is that it is fully reproducible as it plagged
> me for quite sometime. I first thaught that hardware was at fault (like
> defective MPU), but my co-workers don't have the problem, and in fact
> disabled debugging.

Great, let me know what you find out.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

RE: kernel panic in slab.c

[Bill Rogers]

Hi,

I am seeing the same error.  Did you determine the cause?  Thanks.

Bill


kernel panic in slab.c
Hubert Figuiere hub-linux-ppc at trantor.staff.proxad.net  =
<mailto:linuxppc-dev%40ozlabs.org?Subject=3Dkernel%20panic%20in%20slab.c&=
In-Reply-To=3D1064497450.702.12.camel%40gaston>
Fri Sep 26 00:14:24 EST 2003=20
*	Previous message: kernel panic in slab.c  <015712.html>
*	Next message: kernel panic in slab.c  <015714.html>
	*	Messages sorted by: [ date ] <date.html> [ thread ] <thread.html> [ =
subject ] <subject.html> [ author ] <author.html>=20

Hi,

On Thu, 2003-09-25 at 15:44, Benjamin Herrenschmidt wrote:

> Nice to see you using Linux again :)

I never stopped :-)

>
> What machine is this ?

Custom network hardware with a PrPMC750 Motorola board as MPU. We have a
few tiny patches for our custom devices, but nothing important here
IMHO.


> The bug shows that the poison checker found that a freed block got
> overriden. It didn't display more infos about the kind of pattern
> it found and/or a backtrace ? That would be helpful.

I might investigate.

> I've been trying to track down such kind of memory corruption for
> some time now, so if you have a good handle on it, I'm interested,
> as I suspect something wrong is going on in recent 2.4s

Fine, then I feel like investigating it since memory corruption are
probably the worst bugs. I don't have a good handle on the problem. I
often crash while running either "bird" or "swapper" (uptime don't
exceed 5 or 10 minutes). Even worse, sometime it crashes during boot up.
The only thing I'm sure is that it is fully reproducible as it plagged
me for quite sometime. I first thaught that hardware was at fault (like
defective MPU), but my co-workers don't have the problem, and in fact
disabled debugging.


Thanks

Hub
--
Hubert Figui=E8re - Freebox SA. - +33 1 73 50 2 563
AIM/Yahoo!: hfiguiere, ICQ: 307453487

** Sent via the linuxppc-dev mail list. See <http://lists.linuxppc.org/>