Re: 2.6.38.2 breaks suspend to disk

[Yinghai Lu <yinghai@kernel.org>]

On 03/31/2011 02:48 PM, Rafael J. Wysocki wrote:
> On Thursday, March 31, 2011, Stefano Stabellini wrote:
>> On Thu, 31 Mar 2011, Michael Leun wrote:
>>> Nope, unfortunately this does not change anything (neither when
>>> applying ontop of the tree resulting from bisecting, nor ontop
>>> 2.6.38.2).
>>>
>>> Just to make shure there was no mistake when bisecting I did
>>>
>>> patch -R -p1<~/linux-2.6.38.y.git-ff518ea26654e05d325d996f6e3a7f5f569cc2d5.txt
>>>
>>> and then indeed it worked again.
>>
>> This patch fixes it for me.
>>
>>
>> ---
>>
>> save cr4 to mmu_cr4_features at boot time
>>
>> Signed-off-by: Stefano Stabellini<stefano.stabellini@eu.citrix.com>
>
> Acked-by: Rafael J. Wysocki<rjw@sisk.pl>
>
> Peter, Ingo, please regard this as urgent, it breaks -stable.
>
> Thanks,
> Rafael
>
>
>> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
>> index 5a0484a..0943eb2 100644
>> --- a/arch/x86/kernel/setup.c
>> +++ b/arch/x86/kernel/setup.c
>> @@ -891,6 +891,7 @@ void __init setup_arch(char **cmdline_p)
>>   		max_low_pfn = max_pfn;
>>
>>   	high_memory = (void *)__va(max_pfn * PAGE_SIZE - 1) + 1;
>> +	mmu_cr4_features = read_cr4();
>>   #endif
>>
>>   	/*

Stefano, can you explain why it fixes the problem?

or original patch uncover the bug?

Thanks

Yinghai