This is the mail archive of the
mailing list for the GCC project.
Re: PR rtl-optimization/15248 -- semi-latent reload bug
- From: Bernd Schmidt <bernds_cb1 at t-online dot de>
- To: Joern RENNECKE <joern dot rennecke at st dot com>
- Cc: Mark Mitchell <mark at codesourcery dot com>, law at redhat dot com, gcc-patches at gcc dot gnu dot org
- Date: Tue, 07 Jun 2005 15:19:49 +0200
- Subject: Re: PR rtl-optimization/15248 -- semi-latent reload bug
- References: <42A065CB.email@example.com>
Joern RENNECKE wrote:
The comment needs updating.
Yes, I'll fix that in the next iteration.
> ! /* If this register is being made equivalent to a MEM
> ! and the MEM is not SET_SRC, the equivalencing insn
> ! is one with the MEM as a SET_DEST and it occurs
> ! So don't mark this insn now. */
The original comment was probably true when considering execution time,
necessarily when looking at the static instructions.
It should be, since local-alloc only considers registers that occur in a
single basic block for this kind of equivalence.
Anyhow, with the help of Eric Botcazou I've found out what's wrong - it
doesn't correctly handle equivalences made for parameter registers
correctly in all cases. I'm currently testing a new patch, which is
unfortunately a bit larger than the previous one.