This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: mem sharing fix [Re: Your flow.c change on 4/8]
- To: Richard Henderson <rth at cygnus dot com>,Rod Stewart <stewart at nexus dot carleton dot ca>
- Subject: Re: mem sharing fix [Re: Your flow.c change on 4/8]
- From: Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>
- Date: Thu, 13 Apr 2000 03:44:45 +0200
- Cc: Richard Earnshaw <rearnsha at arm dot com>,gcc-patches at gcc dot gnu dot org
- References: <20000412153922.A13400@cygnus.com> <Pine.LNX.4.10.10004121843180.14205-101000@nexus.carleton.ca> <20000412164508.B13475@cygnus.com>
Am Thu, 13 Apr 2000 schrieb Richard Henderson:
>> /usr/src/redhat/BUILD/gcc/gcc/except.c: In function `init_insn_eh_region':
>> /usr/src/redhat/BUILD/gcc/gcc/except.c:3171: Internal compiler error in
>> `verify_wide_reg_1', at flow.c:2626
>
>Arg. I had Kaz' autoinc patch in my tree, which has the bug that
>it doesn't check for a hard reg before it tests for an eliminable
>register. Which means that autoinc never does anything at all.
>
>Reverting that, I reproduce the bug in the original small test
>and the except.i test.
>
>Here's a fix. Jeff apparently had a similar problem while doing
>PA64 development. I have no idea how this is suddenly triggered.
>
>
>r~
>
>
>Mon Sep 13 00:32:43 1999 Jeffrey A Law (law@cygnus.com)
>
> * function.c (purge_addressof): Unshare any shared rtl created by
> purge_addressof and its children.
Sorry, this doesn't solve the bootstrap fail on powerpc-linux-gnu. The testcase
I just posted still segfaults too. I'll continue to look into it tomorrow,
unless somebody else beats me :-).
Franz.