This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC] Fix PR rtl-optimization/33732
- From: Kenneth Zadeck <zadeck at naturalbridge dot com>
- To: John David Anglin <dave at hiauly1 dot hia dot nrc dot ca>
- Cc: rask at sygehus dot dk, ebotcazou at libertysurf dot fr, bonzini at gnu dot org, seongbae dot park at gmail dot com, dave dot anglin at nrc dot ca, gcc-patches at gcc dot gnu dot org
- Date: Wed, 07 Nov 2007 21:34:41 -0500
- Subject: Re: [RFC] Fix PR rtl-optimization/33732
- References: <200711080221.lA82Ld14009220@hiauly1.hia.nrc.ca>
John David Anglin wrote:
>> John David Anglin wrote:
>>
>>>> FWIW, I'm going to test this patch, but I can't any PA-RISC testing.
>>>>
>>>>
>>> I'll try it tonight.
>>>
>>> Thanks,
>>> Dave
>>>
>>>
>> i have done things like this before and it breaks odd ports in odd
>> ways. i will not accept this patch until it is tested on many of the ports.
>>
>
> Breaks ada build in stage2. Turning off ada, I see
>
> /test/gnu/gcc/objdir/./prev-gcc/xgcc -B/test/gnu/gcc/objdir/./prev-gcc/ -B/opt/g
> nu/gcc/gcc-4.3.0/hppa2.0w-hp-hpux11.11/bin/ -c -g -O2 -DIN_GCC -W -Wall -Wwr
> ite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -Wmi
> ssing-format-attribute -pedantic -Wno-long-long -Wno-variadic-macros
> -Wno-overlength-strings -Werror -fno-common -DHAVE_CONFIG_H -I. -I. -I../../gcc/gcc -I../../gcc/gcc/. -I../../gcc/gcc/../include -I../../gcc/gcc/../libcpp/include -I/opt/gnu/gcc/gcc-4.3.0/include -I/opt/gnu/gcc/gcc-4.3.0/include -I../../gcc/gcc/../libdecnumber -I../../gcc/gcc/../libdecnumber/dpd -I../libdecnumber ../../gcc/gcc/local-alloc.c -o local-alloc.o
> ../../gcc/gcc/local-alloc.c: In function 'find_free_reg':
> ../../gcc/gcc/local-alloc.c:2350: internal compiler error: in df_reg_chain_unlink, at df-scan.c:795
>
> Dave
>
The kinds of errors that i saw were not being able to get a register
very late in reload as well as inconsistencies between the number of
words being reserved for the stack and the number of slots actually
being used by reload.
This bug is new, but not at all surprising.
kenny