This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
RE: Issues with GCSE pre step and double hard registers
- From: Claudiu Zissulescu <Claudiu dot Zissulescu at synopsys dot com>
- To: Steven Bosscher <stevenb dot gcc at gmail dot com>
- Cc: "gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>, "claziss at gmail dot com" <claziss at gmail dot com>, Francois Bedard <Francois dot Bedard at synopsys dot com>, "joern dot rennecke at embecosm dot com" <joern dot rennecke at embecosm dot com>
- Date: Wed, 11 Dec 2013 10:41:31 +0000
- Subject: RE: Issues with GCSE pre step and double hard registers
- Authentication-results: sourceware.org; auth=none
- References: <098ECE41A0A6114BB2A07F1EC238DE8965251999 at DE02WEMBXB dot internal dot synopsys dot com> <CABu31nOaeh=s9C-Z_nVQ5F42UG=UcoKTzwkJhsPDPudq0eocDA at mail dot gmail dot com>
> I don't think this is not the right fix for the problem. GCSE doesn't handle
> expressions containing hard registers, oprs_unchanged_p should never even
> see expressions involving hard registers.
I was afraid of this one.
>
> What is the expression that is recorded as anticipated in insn 38? Is it "mho:SI
> 0>>0x3" or "udiv(r159:SI,0xa)" from the REG_EQUAL note?
>
Instruction 38.
As far as I can see, a solution is to use a split pattern after register allocation to describe the multiplication (now it is done in the expand). Another one, is to force the gcse to check the regnotes (REG_EQUAL) rather than the instruction itself.
Best,
Claudiu