[Bug target/46209] New: pmovmskb, useless sign extension

[tbptbp at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46209

           Summary: pmovmskb, useless sign extension
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassig...@gcc.gnu.org
        ReportedBy: tbp...@gmail.com
              Host: x86_64-pc-linux-gnu
            Target: x86_64-pc-linux-gnu


Hello,

$ cat movemsk.c
#include <xmmintrin.h>
typedef long unsigned int uint64_t;
uint64_t foo128(__m128i x) { return _mm_movemask_epi8(x); }
uint64_t foo64(__m64 x) { return _mm_movemask_pi8(x); }
$ /usr/local/gcc-4.6-20101026/bin/gcc  -O3 -march=native movemsk.c -S -o -
foo128:
.LFB516:
    .cfi_startproc
    pmovmskb    %xmm0, %eax
    cltq
    ret
foo64:
.LFB517:
    .cfi_startproc
    movdq2q    %xmm0, %mm0
    movq    %xmm0, -8(%rsp)
    pmovmskb    %mm0, %eax
    cltq
    ret

I won't discuss the interesting mmx code generation but to point that in both
cases, as per Intel doc, there's no need to extend the result; a sign extension
is even slightly more wrong.

$ /usr/local/gcc-4.6-20101026/bin/gcc  -v
Using built-in specs.
COLLECT_GCC=/usr/local/gcc-4.6-20101026/bin/gcc
COLLECT_LTO_WRAPPER=/usr/local/gcc-4.6-20101026/bin/../libexec/gcc/x86_64-unknown-linux-gnu/4.6.0/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../configure --prefix=/usr/local/gcc-4.6.0
--enable-languages=c,c++ --enable-threads=posix --disable-nls
--with-system-zlib --disable-bootstrap --enable-mpfr --enable-gold --enable-lto
--with-ppl --with-cloog --with-arch=native --enable-checking=release
Thread model: posix
gcc version 4.6.0 20101026 (experimental) (GCC)