Re: [PATCH] Re: SSE related security hole

Linus Torvalds (torvalds@transmeta.com)
Sat, 20 Apr 2002 16:23:50 -0700 (PDT)


On Sat, 20 Apr 2002, Andrea Arcangeli wrote:
>
> pxor+xorps is definitely faster than fxrestor on athlon-mp.

Andrea, that's not the _comparison_.

The "fxrestor" replaces the "fninit" too, so you have to take that into
account.

> fxrestor on athlon-mp 1600, on cold cache (the "default fpu state" will
> be cold most of the time, it's only ever used at the first math fault of
> a task):

Except it's _never_ cold-cache the way it's coded now. In fact it's always
hot-cache - which are exactly the numbers I posted.

> Go ahead and try yourself setting the #if to 0 or 1. The benchmark is
> very accurate.

It may have high precision, but since it's testing something that has
nothing to do with the problem at hand, it's basically 100% useless.

Linus

-
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/