This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Bit twiddling builtins
- From: Richard Henderson <rth at redhat dot com>
- To: Falk Hueffner <falk dot hueffner at student dot uni-tuebingen dot de>
- Cc: H?kan Hjort <d95hjort at dtek dot chalmers dot se>, gcc-patches at gcc dot gnu dot org
- Date: Thu, 23 Jan 2003 18:57:18 -0800
- Subject: Re: Bit twiddling builtins
- References: <10BCDB75-0648-11D7-AA5B-00039372607E@physics.uc.edu> <200212022339.SAA29028@makai.watson.ibm.com> <20030120114128.GA8612@safelogic.se> <874r82l0bo.fsf@student.uni-tuebingen.de> <20030121232416.GA21849@redhat.com> <87lm1d4f47.fsf@student.uni-tuebingen.de> <20030123015624.GA23039@redhat.com> <874r7zlezl.fsf_-_@student.uni-tuebingen.de>
On Thu, Jan 23, 2003 at 10:42:54PM +0100, Falk Hueffner wrote:
> I still can't manage to generate ffsdi. Should I perhaps add ffsl and
> ffsll while I'm at it?
Yes. I thought those already existed. Turns out they only
do in glibc, not gcc.
> That might confuse some backends, though,
> because they assume they'll only see ffs in SImode.
I don't see that. We've already got the mode attached to the
pattern. Should be ok as-is.
> ARM already has a __builtin_clz, that would need to be moved out of
> the way, too.
Right.
> When I add a pattern for clzdi to the Alpha backend, gcc will reduce
> clzsi to it via zero extension.
It will? I can't think of why it would do this... I guess I'll
try out your patch and see what happens.
r~