[PATCH] D13142: [PowerPC] Replace cntlz[.] with cntlzw[.]

Hal Finkel via llvm-commits llvm-commits at lists.llvm.org
Fri Sep 25 16:57:30 PDT 2015


----- Original Message -----
> From: "Bill Wendling" <isanbard at gmail.com>
> To: "Hal Finkel" <hfinkel at anl.gov>, reviews+D13142+public+dfe703d8c1b56443 at reviews.llvm.org
> Cc: llvm-commits at lists.llvm.org, tom at bumblecow.com, "David Fang" <fang at csl.cornell.edu>
> Sent: Friday, September 25, 2015 6:55:10 PM
> Subject: Re: [PATCH] D13142: [PowerPC] Replace cntlz[.] with cntlzw[.]
> 
> Unfortunately, I only vaguely recall this patch. Is there a reason
> not to have support for both? I don't think we support POWER in LLVM
> though (I could be wrong there) ... You may need to ask one of the
> ppc experts here.
> 

The only question is: Did the Darwin assembler require the POWER spelling for some reason? If not, then we don't need to generate it.

 -Hal

> 
> On Fri, Sep 25, 2015, 5:46 AM Hal Finkel < hfinkel at anl.gov > wrote:
> 
> 
> [+David Fang; maybe you can comment on Darwin]
> 
> ----- Original Message -----
> > From: hfinkel at anl.gov
> > To: tom at bumblecow.com , isanbard at gmail.com , hfinkel at anl.gov
> > Cc: llvm-commits at lists.llvm.org
> > Sent: Friday, September 25, 2015 7:44:12 AM
> > Subject: Re: [PATCH] D13142: [PowerPC] Replace cntlz[.] with
> > cntlzw[.]
> > 
> > hfinkel added a comment.
> > 
> > I think this is fine, although given that we've accepted the older
> > form for a long time (and it was apparently in use, at least on
> > some
> > older Darwin systems), we should be kind to our users and continue
> > to accept it as an alias. Please add the InstAlias defs and
> > associated parsing tests (set InstAlias's Emit parameter to 0).
> > 
> > Also, as best I can tell from
> > https://llvm.org/bugs/show_bug.cgi?id=1095 and
> > http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20070115/042738.html
> > ,
> > Darwin also wanted the newer mnemonic spelling. This is confusing,
> > given that we have old tests specifically for the old form. Bill,
> > if
> > you can remember from that long ago, can you comment on what we
> > should do here for Darwin?
> > 
> > 
> > http://reviews.llvm.org/D13142
> > 
> > 
> > 
> > 
> 
> --
> Hal Finkel
> Assistant Computational Scientist
> Leadership Computing Facility
> Argonne National Laboratory
> 

-- 
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory


More information about the llvm-commits mailing list