Re: Kernel Oops with HIMEM+VM in 2.4.19,20

From: Denis Vlasenko (vda@port.imtp.ilyichevsk.odessa.ua)
Date: Fri Jan 10 2003 - 04:08:39 EST


On 10 January 2003 10:37, Anthony Lau wrote:
> Hello,
>
> I am getting reproducible kernel oops and random segmentation
> faults whenever the kernel starts using VM. Without any VM pages
> being used, the system is stable.
>
> I have tested kernels compiled with and without HIMEM support
> (all other kernel config options identical). Without HIMEM 4GB
> support, the system is stable for weeks. With HIMEM 4GB support,
> the system starts oops'ing and seg. faulting when VM starts
> being used.

You mean when your system starts to swap? Details?
(How much/how heavy it swaps before oops? vmstat output?)

> My system info:
>
> 1.5GB physical RAM (MemTest86 run for 2 times, no errors)
> 2.0GB VM on a partition
> Aopen AX34u with Via Apollo Pro 133T chipset
>
> Sample Oops from logs:
[snipped]
>
> Because of the symptoms, I think that there could be some
> incompatibility between Himem and the VM subsystem. Of course
> I may have just configured my kernel incorrectly.
>
> Any help is appreciated and I will gladly supply more logs
> if I knew which ones would be useful.

Kernel version and .config?
Arrange klogd to be started with -x. Process oopses with ksymoops.
Then contact VM people (listed in no particular order):

William Irwin <wli@holomorphy.com> [02 jul 2002]
        Send bug reports and/or feature requests related to many tasks,
        rmap, space consumption, or allocators to me. I'm involved in
        * rmap
        * memory allocators
        * reducing space consumed by data structures (e.g. struct page)
        * issues arising in workloads with many tasks
        * kernel janitoring
        See also:
        Rik van Riel <riel@surriel.com>
        Andrea Arcangeli <andrea@suse.de>
        Martin Bligh <Martin.Bligh@us.ibm.com>
        Andrew Morton <akpm@digeo.com>

Andrea Arcangeli <andrea@suse.de> [28 mar 2002]
        Send VM related bug reports and patches to me.
        I'm especially interested in VM issues with:
        * lots of RAM and CPUs
        * NUMA
        * heavy swap scenarios
        * performance of I/O intensive workloads (in particular
          with lots of async buffer flushing involved)
        See also Martin J. Bligh <Martin.Bligh@us.ibm.com> entry
        Mail also:
        Arjan van de Ven <arjanv@redhat.com>

Martin J. Bligh <Martin.Bligh@us.ibm.com> [28 mar 2002]
        I'm interested in VM issues with lots (>4G for i386)
        of RAM, lots of CPUs, NUMA

Rik van Riel <riel@conectiva.com.br> [07 feb 2002]
        Send me VM related stuff, please CC to linux-mm@kvack.org

--
vda
-
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/



This archive was generated by hypermail 2b29 : Wed Jan 15 2003 - 22:00:32 EST