This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [PATCH, i386]: Fix PR 41900, call *%esp shouldn't be generated because of CPU errata

From: Uros Bizjak <ubizjak at gmail dot com>
To: Richard Henderson <rth at redhat dot com>
Cc: gcc-patches at gcc dot gnu dot org
Date: Fri, 13 Nov 2009 08:28:25 +0100
Subject: Re: [PATCH, i386]: Fix PR 41900, call *%esp shouldn't be generated because of CPU errata
References: <5787cf470911022348l14af710ei9118d9777147d6f3@mail.gmail.com> <4AFCAAD4.9000700@redhat.com>

On Fri, Nov 13, 2009 at 1:39 AM, Richard Henderson <rth@redhat.com> wrote:
> On 11/02/2009 11:48 PM, Uros Bizjak wrote:
>>
>> ? ? ? ?(TARGET_CALL_ESP): New define.
>
> Please don't bother with this and all the pattern duplication.
> Just put a comment in front of the call pattern and use "lsm".

But all new targets (newer than P3) can use %esp here without problems...

IMO, since we already went through the pain of duplicating (in
mainline and backports to 4.3 and 4.4), I think that removing
specialized patterns and limiting existing ones would be a step
backward. If you are concerned with maintenance costs, I can perhaps
compensate extra growth by macroizing 32 and 64bit call patterns.

Uros.

Follow-Ups:
- Re: [PATCH, i386]: Fix PR 41900, call *%esp shouldn't be generated because of CPU errata
  - From: Richard Henderson

References:
- [PATCH, i386]: Fix PR 41900, call *%esp shouldn't be generated because of CPU errata
  - From: Uros Bizjak
- Re: [PATCH, i386]: Fix PR 41900, call *%esp shouldn't be generated because of CPU errata
  - From: Richard Henderson

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]