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: revised patch committed.

From: Paolo Bonzini <bonzini at gnu dot org>
To: Amir Gonnen <amirgonnen at gmail dot com>
Cc: Kenneth dot Zadeck at naturalbridge dot com, gcc-patches at gcc dot gnu dot org
Date: Thu, 5 Nov 2009 16:50:28 +0100
Subject: Re: revised patch committed.
References: <df56f4570911040438p65601a3cu5e650e0b54889cf4@mail.gmail.com> <87skctn9ss.fsf@moria.site> <df56f4570911050728s114f1044ybe9ac94149d423fd@mail.gmail.com>

On Thu, Nov 5, 2009 at 16:28, Amir Gonnen <amirgonnen@gmail.com> wrote:
> Thanks Kenny.
> My target does not support zero_extract with non-constant values in
> the 2nd and 3rd field, so I can't really check this on my existing
> code-base and regressions.
> Anyway I suspect there is some problem with dse pass, after applying
> that patch to df_scan.
> It seems that zero_extract is not handled well and as a result dse
> ignores the effect of stores with zero_extract on the lhs, and creates
> wrong rtl.

So if I understand correctly the patch changes a missed optimization
bug to wrong code?

Paolo

Follow-Ups:
- Re: revised patch committed.
  - From: Amir Gonnen

References:
- fwprop and zero_extract
  - From: Amir Gonnen
- revised patch committed.
  - From: Kenneth Zadeck
- Re: revised patch committed.
  - From: Amir Gonnen

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