Re: using mce_inject I get: RIP 10:<ffffffffa012c909>{ttm_bo_unref+0xf/0x45 [ttm]}

[huang ying]

On Sat, Aug 27, 2011 at 11:03 PM, Justin P. Mattock
<justinmattock@xxxxxxxxx> wrote:
> On 08/23/2011 01:15 PM, Luck, Tony wrote:
>>>
>>> its easily fixable, but not sure its a good idea due to bisect going
>>> through commits(afraid I might go astray with the bisect if I add any
>>> patches).
>>
>> Rather than fixing a bad build - you can try moving to a nearby commit
>> (use "gitk" to get a view of the structure around the commit that git
>> bisect suggested). ÂIn the early stages of a bisection, it doesn't really
>> matter much if you build the mid-point that bisect provided, or some
>> nearby on - just be sure to mark good/bad the commit you actually built.
>>
>> -Tony
>>
>>
>
> well.. after bisecting(with no results), I found that something in my
> .config was causing this, so after looking through, I found that having
> X86_MCE_INJECT = y causes the pauses when the timeouts occur
>
> let me know if I need to supply any info.

Which test case cause the pause?  Some test case with "timeout" in
name may cause timeout between CPUs.  Or you can try boot system with
kernel parameter "mce=3,0", which will disable timeout.

Best Regards,
Huang Ying
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/