This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch 3.0, 3.1] Fix a regression (execute/920612-1.c) on h8300
- To: Kazu Hirata <kazu at hxi dot com>
- Subject: Re: [patch 3.0, 3.1] Fix a regression (execute/920612-1.c) on h8300
- From: law at redhat dot com
- Date: Tue, 12 Jun 2001 18:47:54 -0700
- cc: gcc-patches at gcc dot gnu dot org, mark at codesourcery dot com
- Reply-To: law at redhat dot com
In message <200105181739.KAA22268@cygnus.com>you write:
> Hi,
>
> Attached is a patch for both the branch and the mainline to fix the
> regression on execute/920612-1.c caused by
>
> http://gcc.gnu.org/ml/gcc-cvs/2000-07/msg00318.html
>
> On H8, conditional branch instructions do not work well if the
> overflow flag is set. Thus, to obtain correct cc0, we need to insert
> a test instruction that clears the overflow flag. Originally, the
> change was meant to optimize increments/decrements followed by
> conditional branches, but the optimization does not work due to the
> subtle reason described above.
>
> I removed OK_FOR_T because the only use was in this optimization.
>
> OK to apply?
You need a ChangeLog entry for this patch. It looks fine to me for both
the release branch and the mainline sources. However Mark has final
say for anything going onto the branch.
jeff