Le 02/04/11 19:31, Nix a écrit :
> On 2 Apr 2011, [email protected] spake thusly:
>
>> On 29 Mar 2011, [email protected] outgrape:
>>
>>> On 29 Mar 2011, Kenneth Crudup verbalised:
>>>
>>>> Has anyone been able to get the 2.6.38 series of kernels to resume 
>>>> properly?
>>>>
>>>> No matter what I've tried (incl. removing the NVidia driver), it panic()s 
>>>> on
>>>> resume even the first time.
>>> Works perfectly for me. Dozens of resumes.
>> What an idiot I am, I jinxed it. Lockup at atomic copy/restore time
>> (with video signal off due to Radeon KMS being copy/restored).
> This is a regression in 2.6.38.2, not present in .38.1. Bisected to:
>
> commit ff518ea26654e05d325d996f6e3a7f5f569cc2d5
> Author: Yinghai Lu <[email protected]>
> Date:   Fri Feb 18 11:30:30 2011 +0000
>
>     x86: Cleanup highmap after brk is concluded
>
>     commit e5f15b45ddf3afa2bbbb10c7ea34fb32b6de0a0e upstream.
>
>     Now cleanup_highmap actually is in two steps: one is early in head64.c
>     and only clears above _end; a second one is in init_memory_mapping() and
>     tries to clean from _brk_end to _end.
>     It should check if those boundaries are PMD_SIZE aligned but currently
>     does not.
>     Also init_memory_mapping() is called several times for numa or memory
>     hotplug, so we really should not handle initial kernel mappings there.
>
>     This patch moves cleanup_highmap() down after _brk_end is settled so
>     we can do everything in one step.
>     Also we honor max_pfn_mapped in the implementation of cleanup_highmap.
>
>     Signed-off-by: Yinghai Lu <[email protected]>
>     Signed-off-by: Stefano Stabellini <[email protected]>
>     LKML-Reference: <alpine.DEB.2.00.1103171739050.3382@kaball-desktop>
>     Signed-off-by: H. Peter Anvin <[email protected]>
>     Signed-off-by: Greg Kroah-Hartman <[email protected]>
>
> Reverting this commit atop 2.6.38.2 allows ToI to resume again. Cc:ing
> Yinghai and Rafael, since this reportedly affects in-kernel resumption
> as well (though I haven't tried it, others have).
>
> I suspect this commit needs to be reverted from -stable.
>
I can confirm this. Without reverting this commit, resume always hangs
at "doing atomic restore".

François Valenduc

_______________________________________________
stable mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/stable

Reply via email to