This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
gcse crash fix (Was: Re: gcc-3_0 march=athlon && (-Os || -ggdb))
- To: Jan Hubicka <jh at suse dot cz>
- Subject: gcse crash fix (Was: Re: gcc-3_0 march=athlon && (-Os || -ggdb))
- From: Alasdair Baird <alasdair at wildcat dot demon dot co dot uk>
- Date: Mon, 4 Jun 2001 18:49:18 +0100
- Cc: gcc-patches at gcc dot gnu dot org
- References: <20010602105816.E4267@atrey.karlin.mff.cuni.cz><20010602110933.A27492@debian-home.lcisp.com><20010602184349.A17823@atrey.karlin.mff.cuni.cz><20010602191714.D17823@atrey.karlin.mff.cuni.cz><15129.13878.887589.875989@wildcat.demon.co.uk><20010603005442.A2935@atrey.karlin.mff.cuni.cz>
> > BOOT_CFLAGS="-g -O -Os -march=athlon"
> >
> > it all goes horribly wrong:
> >
...
> This is unrelated bug related to Richard Kenner's gcse changes. GCSE pass
> now do CSE over expressions that are valid as (insn (set (reg) (expr))).
...
> I've updated it to use signle_set call. In longer term we should GCSE all
...
> Sun Jun 3 00:52:04 CEST 2001 Jan Hubicka <jh@suse.cz>
> * gcse.c (handle_avail_expr): Be prepared to handle single_set
> parallels.
With this patch, booting with BOOT_CFLAGS="-g -O2 -Os -march=athlon" and
checking enabled (--enable-checking=misc,tree,rtl,gc) definitely proceeds
beyond the point where it failed previously. I won't say that the bootstrap
definitely succeeds, because it is still in the middle of stage2, having been
started sometime yesterday. This combination of checking and flags takes
many, many hours (days?) and much memory to run on my slightly ancient box.
In particular, compiling insn-extract.c and insn-recog.c seem to swill
memory and cpu like there will be no tomorrow.
Appologies that my post slightly diverted this thread away from the original
"-Os on 3.0" topic. And thanks for the patch.
Alasdair.