This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Fix stack corruption in regrename
- To: Alexandre Oliva <aoliva at redhat dot com>
- Subject: Re: Fix stack corruption in regrename
- From: Bernd Schmidt <bernds at redhat dot com>
- Date: Tue, 28 Aug 2001 11:17:44 +0100 (BST)
- cc: <mark at codesourcery dot com>, Richard Henderson <rth at redhat dot com>, <gcc-patches at gcc dot gnu dot org>
On 28 Aug 2001, Alexandre Oliva wrote:
> On Aug 27, 2001, Bernd Schmidt <bernds@redhat.com> wrote:
>
> > On 27 Aug 2001, Alexandre Oliva wrote:
> >> This is still unreviewed. It's too late for 3.0.1, but it would be
> >> nice to fix this stack-corruption problem for 3.0.2.
> >>
> >> http://gcc.gnu.org/ml/gcc-patches/2001-08/msg00024.html
>
> > Oh, sorry. The patch is ok; 3.1 and 3.0.
>
> Thanks. My pre-check-in test passed in mainline, but bootstrap failed
> in the 3.0 branch. regmove marks insns with QImode and HImode in
> certain cases, and it happened to mark a USE insn as such, which
> caused the following sanity check to fail while compiling c-lex.c for
> stage2 on i686-pc-linux-gnu. I don't see where the difference between
> 3.0 and 3.1 stems from; the code in regmove appears identical, and it
> doesn't appear to care at all about cleaning up modes of insns.
>
> What should I do now, just hold off from installing this in 3.0? Is
> this still ok for 3.1?
Just add a piece of code that clears all insn modes at the start of
reload.
Bernd