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] combine.c: Fix rtl-optimization/27477.

From: Kazu Hirata <kazu at codesourcery dot com>
To: Roger Sayle <roger at eyesopen dot com>
Cc: gcc-patches at gcc dot gnu dot org
Date: Wed, 17 May 2006 17:32:38 -0400
Subject: Re: [patch] combine.c: Fix rtl-optimization/27477.
References: <Pine.LNX.4.44.0605171437110.30997-100000@www.eyesopen.com>

Hi Roger,

This is OK for mainline.  Thanks.  This is a reasonable constraint
on combine's insn splitting.

Thanks.

However there does appear to be something else very odd going on
in the PR27477 failure.  Notice in comment #6, which final
instruction is the "jump_insn" and which is the regular "insn"!?

My guess would be that when the combiner splits an unrecognizable insn consisting of two sets, it is a bit sloppy about which insn should go first. If one of the two sets is a jump_insn, that would be a big problem, but I don't have a testcase that exposes the latent bug.

Kazu Hirata

References:
- Re: [patch] combine.c: Fix rtl-optimization/27477.
  - From: Roger Sayle

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