This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: PR2876 vs. reload
>>>>> "Bernd" == Bernd Schmidt <bernds@redhat.com> writes:
Bernd> On Tue, 5 Jun 2001, Mark Mitchell wrote:
>> The failure is a regression from GCC 2.95. Bernd, would you
>> be willing to look at this?
Bernd> I can't reproduce the problem, but does this patch fix it?
I'll try it. It would be great if you could explain what's going on
here, and why you think this might fix it.
Also, reproducing it should be easy. On an i686-pc-linux-gnu box,
download 2876.ii, and compile with `cc1plus -O3 -quiet 2876.ii'. That
causes the crash for me.
There is a little variability possibly because the crash comes from us
encountering a pseudo at a point when we are updating a HARD_REG_SET;
we walk off the end of the HARD_REG_SET and clobber the stack. This
happens in clear_dead_regs in regrename.c; we find a note for register
173. You can set a breakpoint there if you want to see what's going
on.
Thanks for the patch,
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com