Re: [PATCH 2 of 3] memcpy32 for x86_64

[Denis Vlasenko <vda@ilport.com.ua>]

On Friday 13 January 2006 11:56, Chris Wedgwood wrote:
> On Thu, Jan 12, 2006 at 08:04:41AM -0800, Bryan O'Sullivan wrote:
> 
> > This is true for 64-bit writes over Hypertransport
> 
> is this something that will always be or just something current
> hardware does?

Yes, why risking that things will go wrong?
Also you'll get shorter code. Instead of

> +     .globl memcpy32
> +memcpy32:
> +     movl %edx,%ecx
> +     shrl $1,%ecx
> +     andl $1,%edx
> +     rep movsq
> +     movl %edx,%ecx
> +     rep movsd
> +     ret

you need just

	.globl memcpy32
memcpy32:
	movl %edx,%ecx
	rep movsd
	ret

With properly written inlined asms code will be
reduced to just "rep movsd".
--
vda
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/