This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Committed: fix build error for cris-*-* due to rtx canonicalization change
- From: Peter Bergner <bergner at vnet dot ibm dot com>
- To: Paolo Bonzini <bonzini at gnu dot org>
- Cc: Hans-Peter Nilsson <hans-peter dot nilsson at axis dot com>, gcc-patches at gcc dot gnu dot org
- Date: Tue, 08 Apr 2008 10:12:44 -0500
- Subject: Re: Committed: fix build error for cris-*-* due to rtx canonicalization change
- References: <200804080931.m389VqMi030634@ignucius.se.axis.com> <47FB3F45.5060904@gnu.org>
On Tue, 2008-04-08 at 11:47 +0200, Paolo Bonzini wrote:
> I think there are going to be many more of these assumptions; I don't
> like at all that hunk of rtlanal.c and I think it should be reverted for
> now. It also caused a code generation regression for i386.
Due to the issues we had with PR28690 and the patch to fix it, I
was cautious with the rtlanal.c change and asked HJ to SPEC test it
for me to make sure this didn't cause any problems on x86/x86_64.
He reported that he didn't see any performance changes with the patch:
http://gcc.gnu.org/ml/gcc-patches/2008-03/msg01013.html
Are you sure that hunk caused the i386 code gen regression?
I'll note that in another post, I did mention that this hunk
has already been reverted.
Peter