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: Fix sh-elf ICE for builtins-18.c

From: Jim Wilson <wilson at tuliptree dot org>
To: Joern Rennecke <joern dot rennecke at superh dot com>
Cc: gcc-patches at gcc dot gnu dot org
Date: Tue, 10 Jun 2003 22:17:55 -0700
Subject: Re: Fix sh-elf ICE for builtins-18.c
References: <3EE4AE65.51C62DC9@superh.com>

Joern Rennecke wrote:

	* rtlanal.c (reg_overlap_mentioned_p): Handle CONCAT.
	* expr.c (convert_move): Handle moves between two CONCATs.

The rtlanal.c patch seems to be completely unnecessary. With an unpatched tree, we get to convert_move, and then try to use the multi-word emit-no-conflict block code which uses reg_overlap_mentioned_p. But fixing this isn't very helpful, since the following code won't work anyways.

So you added code earlier in convert_move to handle CONCATs specially, thus obviating the need for the reg_overlap_mentioned_p patch. I built a sh-elf cross, and the testcase works fine with the expr.c patch only.

The expr.c patch is approved.

The rtlanal.c patch is not. I don't think backend code like rtlanal should have to know anything about concat.

If you have a testcase that shows the rtlanal.c patch is useful, I'd like to see it.

Jim

References:
- Fix sh-elf ICE for builtins-18.c
  - From: Joern Rennecke

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