[Bug target/36710] New: FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[Bug target/36710] New: FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[dominiq at lps dot ens dot fr]

On i686-apple-darwin9 at revision 137407,
gcc.dg/torture/fp-int-convert-float128.c -m32 -Os fails with:

[ibook-dhum] f90/bug% gcc44 -m32 -Os
/opt/gcc/_gcc_clean/gcc/testsuite/gcc.dg/torture/fp-int-convert-float128.c
/opt/gcc/_gcc_clean/gcc/testsuite/gcc.dg/torture/fp-int-convert-float128.c: In
function 'main':
/opt/gcc/_gcc_clean/gcc/testsuite/gcc.dg/torture/fp-int-convert-float128.c:19:
error: unrecognizable insn:
(insn 25 24 26 3
/opt/gcc/_gcc_clean/gcc/testsuite/gcc.dg/torture/fp-int-convert-float128.c:13
(set (mem:TF (pre_dec:SI (reg/f:SI 7 sp)) [0 S16 A8])
        (reg:TF 1735 [ fv2.2 ])) -1 (nil))
/opt/gcc/_gcc_clean/gcc/testsuite/gcc.dg/torture/fp-int-convert-float128.c:19:
internal compiler error: in extract_insn, at recog.c:1989


-- 
           Summary: FAIL: gcc.dg/torture/fp-int-convert-float128.c  -Os
                    (internal compiler error)
           Product: gcc
           Version: 4.4.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: dominiq at lps dot ens dot fr
 GCC build triplet: i686-apple-darwin9
  GCC host triplet: i686-apple-darwin9
GCC target triplet: i686-apple-darwin9


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[dominiq at lps dot ens dot fr]

------- Comment #1 from dominiq at lps dot ens dot fr  2008-07-03 08:50 -------
Forgot to say that the test passes with -m64.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[ubizjak at gmail dot com]

------- Comment #2 from ubizjak at gmail dot com  2008-07-03 10:06 -------
A couple of splitters for TFmode push_operand should be defined.

I will look into it.


-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |ubizjak at gmail dot com
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2008-07-03 10:06:01
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[ubizjak at gmail dot com]

------- Comment #3 from ubizjak at gmail dot com  2008-07-03 12:22 -------
Created an attachment (id=15846)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15846&action=view)
patch

Dominique, can you regtest attached patch on i686-darwin?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[dominiq at lps dot ens dot fr]

------- Comment #4 from dominiq at lps dot ens dot fr  2008-07-03 12:33 -------
Subject: Re:  FAIL: gcc.dg/torture/fp-int-convert-float128.c  -Os  (internal
compiler error)

> Dominique, can you regtest attached patch on i686-darwin?

I'll do it, but I have to leave now and won't be around for the next hours.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[hjl dot tools at gmail dot com]

------- Comment #5 from hjl dot tools at gmail dot com  2008-07-03 12:34 -------
We should add this testcase as gcc.target/i386/float128-2.c.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[ubizjak at gmail dot com]

------- Comment #6 from ubizjak at gmail dot com  2008-07-03 13:02 -------
Index: gcc.target/i386/float128-2.c
===================================================================
--- gcc.target/i386/float128-2.c        (revision 0)
+++ gcc.target/i386/float128-2.c        (revision 0)
@@ -0,0 +1,17 @@
+/* PR target/36710 */
+
+/* { dg-do run { target *-*-linux* *-*-darwin* } } */
+/* { dg-options "-Os -msse2" } */
+
+#include "sse2-check.h"
+
+extern void abort (void);
+
+static void
+sse2_test (void)
+{
+  static volatile __float128 a = 123.0q;
+
+  if ((int) a != 123)
+    abort ();
+}


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[hjl dot tools at gmail dot com]

------- Comment #7 from hjl dot tools at gmail dot com  2008-07-03 13:03 -------
Created an attachment (id=15848)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15848&action=view)
A patch + a testase

Here is Uros' patch + a testcase which is also reproducible on Linux.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[ubizjak at gmail dot com]

------- Comment #8 from ubizjak at gmail dot com  2008-07-03 13:16 -------
> Created an attachment (id=15848)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15848&action=view) [edit]
> A patch + a testase

This failure is specific to -mno-accumulate-outgoing-args (-Os is just a
shortcut for this looong-worded compile flag). With -mn-a-o-a, compiler tries
to push everything to the stack. Since there is no "push %xmm", compilation
breaks.

I can put "-O2 -mno-accumulate-outgoing-args" into dg-options if this is more
informative, but there is no need for torture to cycle all optimization flags.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[hjl dot tools at gmail dot com]

------- Comment #9 from hjl dot tools at gmail dot com  2008-07-03 14:59 -------
(In reply to comment #8)
> > Created an attachment (id=15848)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15848&action=view) [edit]
> > A patch + a testase
> 
> This failure is specific to -mno-accumulate-outgoing-args (-Os is just a
> shortcut for this looong-worded compile flag). With -mn-a-o-a, compiler tries
> to push everything to the stack. Since there is no "push %xmm", compilation
> breaks.
> 
> I can put "-O2 -mno-accumulate-outgoing-args" into dg-options if this is more
> informative, but there is no need for torture to cycle all optimization flags.

-Os is OK with me.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[dominiq at lps dot ens dot fr]

------- Comment #10 from dominiq at lps dot ens dot fr  2008-07-03 17:35 -------
With the patch in comment #3 the error is gone, now regtesting.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[dominiq at lps dot ens dot fr]

------- Comment #11 from dominiq at lps dot ens dot fr  2008-07-03 17:36 -------
I also see a lot of "bus error" in the obj-c++ regtests. Is this related or
should I fill another PR?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[hjl dot tools at gmail dot com]

------- Comment #12 from hjl dot tools at gmail dot com  2008-07-03 17:47 -------
(In reply to comment #11)
> I also see a lot of "bus error" in the obj-c++ regtests. Is this related or
> should I fill another PR?
> 

Are they related to __float128? If not, open a new PR.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[uros at gcc dot gnu dot org]

------- Comment #13 from uros at gcc dot gnu dot org  2008-07-03 18:57 -------
Subject: Bug 36710

Author: uros
Date: Thu Jul  3 18:57:11 2008
New Revision: 137427

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=137427
Log:
        PR target/36710
        * config/i386/i386.md (mode): Add TF to mode attribute.
        (*pushtf_sse): New insn pattern.
        (pushtf splitters): New splitters.

testsuite/ChangeLog:

        PR target/36710
        * gcc.target/i386/float128-2.c: New test.

Added:
    trunk/gcc/testsuite/gcc.target/i386/float128-2.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/i386/i386.md
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710

[Bug target/36710] FAIL: gcc.dg/torture/fp-int-convert-float128.c -Os (internal compiler error)

[ubizjak at gmail dot com]

------- Comment #14 from ubizjak at gmail dot com  2008-07-03 19:04 -------
Fixed.


-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |http://gcc.gnu.org/ml/gcc-
                   |                            |patches/2008-
                   |                            |07/msg00232.html
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36710