Re: [PATCH] x86/xsave: Robustify and merge macros

From: Quentin Casasnovas <quentin.casasnovas@oracle.com>
To: Borislav Petkov <bp@alien8.de>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>,
	X86 ML <x86@kernel.org>, LKML <linux-kernel@vger.kernel.org>,
	"H. Peter Anvin" <hpa@zytor.com>, Ingo Molnar <mingo@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Oleg Nesterov <oleg@redhat.com>,
	Andy Lutomirski <luto@amacapital.net>
Subject: Re: [PATCH] x86/xsave: Robustify and merge macros
Date: Sat, 4 Apr 2015 12:11:55 +0200	[thread overview]
Message-ID: <20150404101155.GB5922@chrystal.uk.oracle.com> (raw)
In-Reply-To: <20150404092536.GB21152@pd.tnic>

On Sat, Apr 04, 2015 at 11:25:36AM +0200, Borislav Petkov wrote:
> On Sat, Apr 04, 2015 at 10:36:11AM +0200, Quentin Casasnovas wrote:
> > Since all of these are compile time constants, could we not use the safe
> > variant on that same page?  Not that I'm too worried about the signed right
> > shift but heh that would be portable and should not impact performance
> > anyway, so no added value in using the optimized version is there?
> 
> Seems to work with the experimental diff below. I need to do
> 
> 	-(-(x < y))
> 
> with the last term though as we're working with s32s.
> 
> ---
> diff --git a/arch/x86/include/asm/alternative-asm.h b/arch/x86/include/asm/alternative-asm.h
> index 44a1fc5439d3..2cb6da2716bf 100644
> --- a/arch/x86/include/asm/alternative-asm.h
> +++ b/arch/x86/include/asm/alternative-asm.h
> @@ -53,14 +53,14 @@
>   * Shamelessly stolen and adapted from:
>   * http://graphics.stanford.edu/~seander/bithacks.html#IntegerMinOrMax
>   */
> -#define alt_max_short(a,b)	(((a) - (((a) - (b)) & (((a) - (b)) >> 15))) & 0xffff)
> +#define alt_max_short(a, b)	((a) ^ (((a) ^ (b)) & -(-((a) < (b)))))
>

So I'm not claiming I've spent the time to fully understand this macro but
it looks like it's doing the right thing on my dummy tests:

  http://pastebin.com/DDhtZQgX

Did you also change it in the alternative.h file BTW?

>  .macro ALTERNATIVE_2 oldinstr, newinstr1, feature1, newinstr2, feature2
>  140:
>  	\oldinstr
>  141:
> -	.skip -((alt_max_short(new_len1, new_len2) - old_len) > 0) * \
> -		(alt_max_short(new_len1, new_len2) - old_len),0x90
> +	.skip -((alt_max_short(new_len1, new_len2) - (old_len)) > 0) * \
> +		(alt_max_short(new_len1, new_len2) - (old_len)),0x90
>  142:

Good catch for the missing parenthesis!

> > 
> > Looks good to me and I find it much easier to understand here :)
> > 
> 
> Cool. Please give it more critical staring as we're under time pressure
> here.
> 

So I _think_ it's OK but it would be re-assuring if somebody else could
have a look as well just in case.. :)

Do you have a cleaned up version of the patch you're planning to apply on
top of tip/master instead of just snippets?  This way we can hammer it with
different calls to ALTERNATIVE_2 and alternative_2 to check it's good?
I'll have to leave soonish though..

Quentin