Hi
I can reproduce a crash with 2.6.19-rc6-mm2 triggered when debugging a
program with gdb that uses pthreads. No oops or anything strange seems
to be printed by the kernel, but the box appears to stop doing disk IO.
--
Pedro Larroy Tovar, pedro at larroy dot com | http://pedro.larroy.com/
* Las patentes de programaci?n son nocivas para la innovaci?n *
http://proinnova.hispalinux.es/
Pedro Larroy Tovar wrote:
> Hi
>
> I can reproduce a crash with 2.6.19-rc6-mm2 triggered when debugging a
> program with gdb that uses pthreads. No oops or anything strange seems
> to be printed by the kernel, but the box appears to stop doing disk IO.
>
Hm, I wonder if this is related walt's problem running things under gdb?
Pedro, do you ever see oops or BUG messages?
Thanks,
J
Jeremy Fitzhardinge wrote:
> Pedro Larroy Tovar wrote:
>> Hi
>>
>> I can reproduce a crash with 2.6.19-rc6-mm2 triggered when debugging a
>> program with gdb that uses pthreads. No oops or anything strange seems
>> to be printed by the kernel, but the box appears to stop doing disk IO.
> Hm, I wonder if this is related walt's problem running things under gdb?
Jeremy, I redid my git-bisect from scratch and came up with a different
commit for the gdb breakage:
commit f95d47caae5302a63d92be9a0292abc90e2a14e1
Author: Jeremy Fitzhardinge <[email protected]>
Date: Thu Dec 7 02:14:02 2006 +0100
[PATCH] i386: Use %gs as the PDA base-segment in the kernel
This commit is the one which causes gdb to halt with this error
no matter what executable I try to run:
Warning:
Cannot insert breakpoint -2.
Error accessing memory address 0xd74b: Input/output error.
I tried to git-revert just this one commit, but I get merge conflicts
I don't how to resolve.
BTW, I just discovered tons of kernel debugging config options which
were turned off -- I just turned several of them on. Maybe I can give
you better info now.
walt wrote:
> Jeremy Fitzhardinge wrote:
>
>> Pedro Larroy Tovar wrote:
>>
>>> Hi
>>>
>>> I can reproduce a crash with 2.6.19-rc6-mm2 triggered when debugging a
>>> program with gdb that uses pthreads. No oops or anything strange seems
>>> to be printed by the kernel, but the box appears to stop doing disk IO.
>>>
>
>
>> Hm, I wonder if this is related walt's problem running things under gdb?
>>
>
> Jeremy, I redid my git-bisect from scratch and came up with a different
> commit for the gdb breakage:
>
> commit f95d47caae5302a63d92be9a0292abc90e2a14e1
> Author: Jeremy Fitzhardinge <[email protected]>
> Date: Thu Dec 7 02:14:02 2006 +0100
> [PATCH] i386: Use %gs as the PDA base-segment in the kernel
>
> This commit is the one which causes gdb to halt with this error
> no matter what executable I try to run:
> Warning:
> Cannot insert breakpoint -2.
> Error accessing memory address 0xd74b: Input/output error.
>
> I tried to git-revert just this one commit, but I get merge conflicts
> I don't how to resolve.
>
> BTW, I just discovered tons of kernel debugging config options which
> were turned off -- I just turned several of them on. Maybe I can give
> you better info now.
>
OK, this is the changeset I'd expect to cause problems. I'll try to
repo it here.
J