Re: [RFC PATCH 2/3] x86, MCE: Avoid potential deadlock in MCE

[Borislav Petkov <bp@alien8.de>]

On Mon, Jul 21, 2014 at 03:44:06PM -0700, Tony Luck wrote:
> 
> This is how much cleaner things could be with a couple of task_struct
> fields instead of the mce_info silliness ... untested.

...

> diff --git a/include/linux/sched.h b/include/linux/sched.h
> index 0376b054a0d0..91db69a4acd7 100644
> --- a/include/linux/sched.h
> +++ b/include/linux/sched.h
> @@ -1655,6 +1655,10 @@ struct task_struct {
>  	unsigned int	sequential_io;
>  	unsigned int	sequential_io_avg;
>  #endif
> +#ifdef CONFIG_MEMORY_FAILURE
> +	__u64	paddr;
> +	int	restartable;
> +#endif

Right, I don't see anything wrong with this approach especially as
task_struct is full of CONFIG_* ifdeffery for members used with
different features.

Adding 12 more bytes for CONFIG_MEMORY_FAILURE shouldn't hurt anyone. If
we really want to save space, we can use the highest significant byte of
paddr for a bit to say "restartable" or not.

So I think we should make it into a patch and push it upstream.

-- 
Regards/Gruss,
    Boris.

Sent from a fat crate under my desk. Formatting is fine.
--