This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [testsuite] i386 reload issue
- To: Jakub Jelinek <jakub at redhat dot com>
- Subject: Re: [testsuite] i386 reload issue
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Thu, 24 Aug 2000 10:34:54 -0600
- cc: gcc-patches at gcc dot gnu dot org
- Reply-To: law at cygnus dot com
In message <20000824122429.E21753@devserv.devel.redhat.com>you write:
> I know call_insn_operand rejects this, but current development sources
> (well, have tried 0731 patched snapshot and vanilla CVS from 0717) exhibit
> it still.
> call_insn_operand is in fact never called in this testcase
> if (reload_in_progress || reload_completed), reload only cares about the
> "m" constraint which is ok, and as constant_call_address_operand called fro
> m
> call_value_1 define_insn body rejects it, we end up with call *$0.
We shouldn't be using 'm'. That says accept anything that is a valid
memory operand. [ Though one could argue that reload should have verified
that the argument was a valid call_address_operand. ]
jeff