[Bug target/50155] New: [4.7 Regression] AVX2 support broke -mavx

[Bug target/50155] New: [4.7 Regression] AVX2 support broke -mavx

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50155

             Bug #: 50155
           Summary: [4.7 Regression] AVX2 support broke -mavx
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: jakub@gcc.gnu.org
                CC: hjl.tools@gmail.com, uros@gcc.gnu.org


void
foo (int x, double *a, double *b, double c)
{
  int i;

  for (i = 0; i < x; i++)
    *a++ = *b++ * i / c;
}

fails to assemble with -O3 -mavx -m64 on x86_64-linux (at least when using
binutils that don't support AVX2).
vpaddd  %ymm6, %ymm1, %ymm5
is AFAIK an AVX2 insn, not AVX, yet it is emitted by the addv8si3 pattern
(*<plusminus_insn><code>3 insn etc. was previously using VI_128 mode iterator,
but now is using VI, guess it wants to use a mode iterator similar to VI, but
which uses TARGET_AVX2 instead of TARGET_AVX guards for the 32 byte integer
modes (VI_AVX2 or VI_1248_AVX2?)).

[Bug target/50155] [4.7 Regression] AVX2 support broke -mavx

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50155

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot       |kirill.yukhin at intel dot
                   |gnu.org                     |com
   Target Milestone|---                         |4.7.0

[Bug target/50155] [4.7 Regression] AVX2 support broke -mavx

[kirill.yukhin at intel dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50155

--- Comment #1 from Yukhin Kirill <kirill.yukhin at intel dot com> 2011-08-22 18:52:40 UTC ---
Hi,
thanks, for investigation.
Here is a patch:
http://gcc.gnu.org/ml/gcc-patches/2011-08/msg01808.html

K

[Bug target/50155] [4.7 Regression] AVX2 support broke -mavx

[uros at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50155

--- Comment #2 from uros at gcc dot gnu.org 2011-08-22 21:01:51 UTC ---
Author: uros
Date: Mon Aug 22 21:01:46 2011
New Revision: 177974

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=177974
Log:
    PR target/50155
    * config/i386/sse.md (VI_AVX2): New.
    (<plusminus_insn><mode>3): Use VI_AVX2 mode iterator.
    (*<plusminus_insn><mode>3): Ditto.
    (<sse2_avx2>_andnot<mode>3): Ditto.
    (*andnot<mode>3): Fix order of cond operands.
    Add asserts for correct TARGET_xxx.
    (*<any_logic:code><mode>3): Ditto.

testsuite/ChangeLog:

    PR target/50155
    * gcc.target/i386/pr50155.c: New test.


Added:
    trunk/gcc/testsuite/gcc.target/i386/pr50155.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/i386/sse.md
    trunk/gcc/testsuite/ChangeLog

[Bug target/50155] [4.7 Regression] AVX2 support broke -mavx

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50155

Uros Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |x86-avx
             Status|UNCONFIRMED                 |RESOLVED
                URL|                            |http://gcc.gnu.org/ml/gcc-p
                   |                            |atches/2011-08/msg01834.htm
                   |                            |l
         Resolution|                            |FIXED

--- Comment #3 from Uros Bizjak <ubizjak at gmail dot com> 2011-08-22 21:14:59 UTC ---
Fixed.