[Bug target/47759] New: _mm_empty() intrinsic fails to serve as a boundary between MMX and x87 code due to optimizations

[Bug target/47759] New: _mm_empty() intrinsic fails to serve as a boundary between MMX and x87 code due to optimizations

[siarhei.siamashka at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47759

           Summary: _mm_empty() intrinsic fails to serve as a boundary
                    between MMX and x87 code due to optimizations
           Product: gcc
           Version: 4.5.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: siarhei.siamashka@gmail.com
              Host: i686-pc-linux-gnu


Created attachment 23355
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23355
mm_empty_testcase.c

The attached testcase fails when compiled with -O2 or -O3 optimizations, but
works with -O1.

I'm actually not precisely sure how this code is expected to behave because
intrinsics are x86 architecture specific and C standard can't be used as a
reference. But my guess is that if the optimizer would not be allowed to
arbitrarily move code across _mm_empty() boundary, then the problem would
disappear.

[Bug target/47759] _mm_empty() intrinsic fails to serve as a boundary between MMX and x87 code due to optimizations

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47759

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.02.15 21:57:56
     Ever Confirmed|0                           |1

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> 2011-02-15 21:57:56 UTC ---
Confirmed, the scheduler does not know that the mmx instructions touch the x87
registers.

[Bug target/47759] _mm_empty() intrinsic fails to serve as a boundary between MMX and x87 code due to optimizations

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47759

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> 2011-02-15 21:59:12 UTC ---
Note this is a target specific issue and not a generic scheduler issue as the
target does not communicate to the scheduler that the x87 registers and the MMX
register overlap.