This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [PATCH, take 2] implement drem() and fmod() as built-in x87intrinsic

From: Uros Bizjak <uros at kss-loka dot si>
To: Martin Reinecke <martin at MPA-Garching dot MPG dot DE>
Cc: gcc-patches at gcc dot gnu dot org
Date: Fri, 07 May 2004 12:19:13 +0200
Subject: Re: [PATCH, take 2] implement drem() and fmod() as built-in x87intrinsic
References: <409B5070.5030804@mpa-garching.mpg.de>

Martin Reinecke wrote:


bugtest.c:6: error: unrecognizable insn:
(insn 17 16 18 1 (set (reg:CC 17 flags)
        (unspec:CC [
                (reg:HI 65)
            ] 25)) -1 (insn_list 16 (nil))
    (expr_list:REG_DEAD (reg:HI 65)
        (nil)))
bugtest.c:6: internal compiler error: in extract_insn, at recog.c:2074
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

Failing instruction, x86_sahf_1, is defined in config/i386/i386.md as:

(define_insn "x86_sahf_1"
 [(set (reg:CC 17)
   (unspec:CC [(match_operand:HI 0 "register_operand" "a")] UNSPEC_SAHF))]
 "!TARGET_64BIT"
 "sahf"
 [(set_attr "length" "1")
  (set_attr "athlon_decode" "vector")
  (set_attr "mode" "SI")])

I don't know, why it has "!TARGET_64BIT" constraint, could you just remove !TARGET_64BIT?

Should I file a PR for this?

No, we have just figure out, why x86_sahf_1 has !TARGET_64BIT constraint.

P.S.: Thanks a lot for implementing all these new math builtins in gcc; they give a great performance boost for our numerical simulations!

Thanks, nice to hear this!

Uros.

References:
- Re: [PATCH, take 2] implement drem() and fmod() as built-in x87intrinsic
  - From: Martin Reinecke

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]