This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [PATCH] more x86-64 corrections
- From: "Jan Beulich" <JBeulich at novell dot com>
- To: <hjl at lucon dot org>
- Cc: <nickc at redhat dot com>,<binutils at sources dot redhat dot com>
- Date: Fri, 23 Jul 2004 16:35:09 +0200
- Subject: Re: [PATCH] more x86-64 corrections
I'm not happy with that, but then let's leave this hunk out. However,
your change to revert to allow the segment overrides is then incomplete:
You either allow them all (cs,ds,es,fs,gs,ss), or you disallow all that
are pointless (cs,ds,es,ss). Or, as I indicated, you suppress their use
as segment override elsewhere, thus *only* allowing them as branch
prediction hints (but for a standalone prefix I can't imagine how one
would do that correctly). Jan
>>> "H. J. Lu" <hjl@lucon.org> 23.07.04 16:04:47 >>>
On Fri, Jul 23, 2004 at 02:54:38PM +0200, Jan Beulich wrote:
> additions that should be done. Finally, intermediately cs and ds
> segment
> overrides got wrongly re-enabled in 64-bit mode. This patch tries to
> address
> all of this.
>
Nick, please don't apply this patch. It is wrong for Intel EM64T.
Jan, please bootstrap gcc 3.4.1 with -mtune=nocona on x86-64 and let
us know how it goes.
H.J.