This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC] Emit 'clr' instead of 'mov 0' on SPARC
- From: "David S. Miller" <davem at davemloft dot net>
- To: Eric Botcazou <ebotcazou at libertysurf dot fr>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Fri, 22 Apr 2005 15:23:08 -0700
- Subject: Re: [RFC] Emit 'clr' instead of 'mov 0' on SPARC
- References: <200504222208.30729.ebotcazou@libertysurf.fr>
On Fri, 22 Apr 2005 22:08:30 +0200
Eric Botcazou <ebotcazou@libertysurf.fr> wrote:
> 'clr' is not emitted in the usual cases. Any specific reasons? Any
> objections to the following change?
'clr' is just a synthetic instruction which expands in the
assembler to "mov 0, %reg". I think we could safely remove
it, and having less constraints in such patterns at least
theoretically makes the compiler faster.
I'm actually ambivalent, you can decide :)
About sethi... I'm wondering if I did that for a reason long
ago. Every time I used to move these constraints around, some
new reload failure would pop up during the testsuite :-)