Re: [34-longterm 078/260] x86, tsc: Fix a preemption leak inrestore_sched_clock_state()

[Nico Schottelius]

Good morning,

Paul Gortmaker [Sun, Jan 02, 2011 at 02:16:14AM -0500]:
> From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> 
> commit 55496c896b8a695140045099d4e0175cf09d4eae upstream.
> 
> diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c
> index c65f741..4094ae0 100644
> --- a/arch/x86/kernel/tsc.c
> +++ b/arch/x86/kernel/tsc.c
> @@ -655,7 +655,7 @@ void restore_sched_clock_state(void)
>  
>  	local_irq_save(flags);
>  
> -	get_cpu_var(cyc2ns_offset) = 0;
> +	__get_cpu_var(cyc2ns_offset) = 0;
>  	offset = cyc2ns_suspend - sched_clock();

If I see that correctly, this has already been merged before
2.6.37-rc7. As 2.6.37-rc7 again has the dark screen issue [0]
I assume this does not fix this issue.

I am compiling v2.6.37-rc8-33-gb518a64 right now, to give it
a try.

Cheers,

Nico

[0]: Subject: 2.6.37-rc7: screen black after several suspends


-- 
PGP key: 7ED9 F7D3 6B10 81D7 0EC5  5C09 D7DC C8E4 3187 7DF0

Attachment: pgp00000.pgp
Description: PGP signature