[Bug c/36399] New: FSF GCC ABI bug on darwin/x86-32

[Bug c/36399] New: FSF GCC ABI bug on darwin/x86-32

[sabre at nondot dot org]

This testcase:

#include <xmmintrin.h>
__m128i r(__m128 d1, __m128 d2, __m128 d3, __m128i r, int t, __m128i s) {return
r+s;}

compiles to:

_r:
        subl    $12, %esp
        movdqa  %xmm3, %xmm0
        paddq   32(%esp), %xmm0
        addl    $12, %esp
        ret

with apple gcc 4.0 and 4.2.

It compiles to this with FSF 4.2:

_r:
        subl    $12, %esp
        movdqa  48(%esp), %xmm0
        paddq   16(%esp), %xmm0
        addl    $12, %esp
        ret

Since apple GCC defines the ABI on darwin, this is an FSF GCC ABI bug for that
platform.


-- 
           Summary: FSF GCC ABI bug on darwin/x86-32
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: sabre at nondot dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] FSF GCC ABI bug on darwin/x86-32

[rguenth at gcc dot gnu dot org]

------- Comment #1 from rguenth at gcc dot gnu dot org  2008-05-31 10:58 -------
As apple implements the darwin ABI in the FSF gcc it's apples turn to fix it.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c                           |target
 GCC target triplet|                            |i?86-apple-darwin
           Keywords|                            |ABI, wrong-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] FSF GCC ABI bug on darwin/x86-32

[echristo at apple dot com]

------- Comment #2 from echristo at apple dot com  2008-06-03 00:37 -------
Confirmed.


-- 

echristo at apple dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2008-06-03 00:37:51
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] FSF GCC ABI bug on darwin/x86-32

[fxcoudert at gcc dot gnu dot org]

------- Comment #3 from fxcoudert at gcc dot gnu dot org  2009-06-03 21:18 -------
Here's what it gives with Apple's gcc:

_r:
        pushl   %ebp
        movl    %esp, %ebp
        subl    $72, %esp
        movaps  %xmm0, -24(%ebp)
        movaps  %xmm1, -40(%ebp)
        movaps  %xmm2, -56(%ebp)
        movdqa  %xmm3, -72(%ebp)
        movdqa  24(%ebp), %xmm0
        paddq   -72(%ebp), %xmm0
        leave
        ret

and with FSF's gcc, both 4.3.2 and and 4.4.0 give:

_r:
        pushl   %ebp
        movl    %esp, %ebp
        subl    $56, %esp
        movaps  %xmm0, -24(%ebp)
        movaps  %xmm1, -40(%ebp)
        movaps  %xmm2, -56(%ebp)
        movdqa  40(%ebp), %xmm1
        movdqa  8(%ebp), %xmm0
        paddq   %xmm1, %xmm0
        leave
        ret


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |4.2.0 4.3.2 4.4.0 4.5.0
   Last reconfirmed|2008-06-03 00:37:51         |2009-06-03 21:18:11
               date|                            |
            Version|unknown                     |4.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] FSF GCC ABI bug on darwin/x86-32

[dalej at gcc dot gnu dot org]

------- Comment #4 from dalej at gcc dot gnu dot org  2009-06-04 00:06 -------
Yeah.  Looking through our local changes, I think you just need this line in
i386.h:

#define SSE_REGPARM_MAX (TARGET_64BIT ? 8 : (TARGET_MACHO ? 4 : (TARGET_SSE ? 3
: 0)))

However Apple employees can't currently touch GPLv3 code, so we can't do the
testing.  Stan, can you do this?


-- 

dalej at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |stanshebs at earthlink dot
                   |                            |net


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] FSF GCC ABI bug on darwin/x86-32

[hjl dot tools at gmail dot com]

------- Comment #5 from hjl dot tools at gmail dot com  2009-06-07 14:27 -------
(In reply to comment #4)
> Yeah.  Looking through our local changes, I think you just need this line in
> i386.h:
> 
> #define SSE_REGPARM_MAX (TARGET_64BIT ? 8 : (TARGET_MACHO ? 4 : (TARGET_SSE ? 3
> : 0)))
> 
> However Apple employees can't currently touch GPLv3 code, so we can't do the
> testing.  Stan, can you do this?
> 

Shouldn't it be

#define SSE_REGPARM_MAX (TARGET_64BIT ? 8 : (TARGET_SSE ? (TARGET_MACHO ? 4 : 3
: 0)))


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] FSF GCC ABI bug on darwin/x86-32

[dalej at gcc dot gnu dot org]

------- Comment #6 from dalej at gcc dot gnu dot org  2009-06-08 19:17 -------
That needs to be parenthesized differently; if that's fixed, I have no
objection to doing it that way, but since all MachO targets have SSE, that
works too.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[fxcoudert at gcc dot gnu dot org]

------- Comment #7 from fxcoudert at gcc dot gnu dot org  2009-12-29 20:21 -------
Patch for current trunk:

Index: gcc/config/i386/i386.h
===================================================================
--- gcc/config/i386/i386.h      (revision 155505)
+++ gcc/config/i386/i386.h      (working copy)
@@ -1810,7 +1810,7 @@
 #define X86_64_SSE_REGPARM_MAX 8
 #define X86_64_MS_SSE_REGPARM_MAX 4

-#define X86_32_SSE_REGPARM_MAX (TARGET_SSE ? 3 : 0)
+#define X86_32_SSE_REGPARM_MAX (TARGET_SSE ? (TARGET_MACHO ? 4 : 3) : 0)

 #define SSE_REGPARM_MAX                                                       
\
   (TARGET_64BIT ? (TARGET_64BIT_MS_ABI ? X86_64_MS_SSE_REGPARM_MAX     \


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fxcoudert at gcc dot gnu dot
                   |                            |org
           Keywords|                            |patch
   Last reconfirmed|2009-06-03 21:18:11         |2009-12-29 20:21:12
               date|                            |
            Summary|FSF GCC ABI bug on          |ABI bug on darwin/x86-32
                   |darwin/x86-32               |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[echristo at apple dot com]

------- Comment #8 from echristo at apple dot com  2009-12-29 22:21 -------
That patch looks correct for at least x86-32


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[howarth at nitro dot med dot uc dot edu]

------- Comment #9 from howarth at nitro dot med dot uc dot edu  2010-03-11 14:13 -------
Should we apply the patch from comment 7 before gcc 4.5 branches?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[rguenth at gcc dot gnu dot org]

------- Comment #10 from rguenth at gcc dot gnu dot org  2010-03-11 14:24 -------
(In reply to comment #9)
> Should we apply the patch from comment 7 before gcc 4.5 branches?

Well, you certainly need to follow patch submission rules.  Thus, post
to gcc-patches w/ a changelog entry and stating how you tested the patch.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[howarth at nitro dot med dot uc dot edu]

------- Comment #11 from howarth at nitro dot med dot uc dot edu  2010-03-12 04:06 -------
Patch posted at http://gcc.gnu.org/ml/gcc-patches/2010-03/msg00472.html.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[mrs at gcc dot gnu dot org]

------- Comment #12 from mrs at gcc dot gnu dot org  2010-03-18 22:57 -------
Subject: Bug 36399

Author: mrs
Date: Thu Mar 18 22:56:38 2010
New Revision: 157553

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=157553
Log:
2010-03-11  Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
            Jack Howarth <howarth@bromo.med.uc.edu>

        PR target/36399
        * gcc/config/i386/i386.h: Fix ABI on darwin x86-32.
        * gcc/testsuite/gcc.target/i386/push-1.c: Don't xfail
          scan-assembler-not "movups" on darwin.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/i386/i386.h
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gcc.target/i386/push-1.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[mikestump at comcast dot net]

------- Comment #13 from mikestump at comcast dot net  2010-03-18 22:57 -------
Fix checked in as r157553.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[mrs at gcc dot gnu dot org]

------- Comment #14 from mrs at gcc dot gnu dot org  2010-03-23 00:10 -------
This has now been fixed.


-- 

mrs at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399

[Bug target/36399] ABI bug on darwin/x86-32

[mrs at gcc dot gnu dot org]

-- 

mrs at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.5.0
            Version|4.5.0                       |4.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36399