[Bug tree-optimization/45844] New: FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[Bug tree-optimization/45844] New: FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

           Summary: FAIL: gfortran.dg/vect/pr45714-b.f  -O  (internal
                    compiler error)
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: dominiq@lps.ens.fr
                CC: irar@il.ibm.com
              Host: powerpc-apple-darwin9
            Target: powerpc-apple-darwin9
             Build: powerpc-apple-darwin9


When compiled with -O3 -m64 -mcpu=power7 -ffast-math -mveclibabi=mass the test
gfortran.dg/vect/pr45714-b.f gives the following ICE:

[karma] f90/bug% gfc -c -O3 -m64 -mcpu=power7 -ffast-math -mveclibabi=mass
/opt/gcc/_gcc_clean/gcc/testsuite/gfortran.dg/vect/pr45714-b.f
/opt/gcc/_gcc_clean/gcc/testsuite/gfortran.dg/vect/pr45714-b.f: In function
'MAIN__':
/opt/gcc/_gcc_clean/gcc/testsuite/gfortran.dg/vect/pr45714-b.f:25:0: error:
insn does not satisfy its constraints:
(insn 39 141 33 2 (set (reg:V2DF 108 v31 [orig:162 vect_cst_.14 ] [162])
        (vec_duplicate:V2DF (mem/u/c/i:DF (lo_sum:DI (reg:DI 2 r2)
                    (const:DI (unspec:DI [
                                (symbol_ref/u:DI ("*LC1") [flags 0x2])
                            ] 50))) [2 S8 A64])))
/opt/gcc/_gcc_clean/gcc/testsuite/gfortran.dg/vect/pr45714-b.f:14 873
{vsx_splat_v2df}
     (nil))
/opt/gcc/_gcc_clean/gcc/testsuite/gfortran.dg/vect/pr45714-b.f:25:0: internal
compiler error: in reload_cse_simplify_operands, at postreload.c:402

The test has been introduced at revision 164420.

[Bug tree-optimization/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[irar at il dot ibm.com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

Ira Rosen <irar at il dot ibm.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2010.10.03 11:43:15
     Ever Confirmed|0                           |1

--- Comment #1 from Ira Rosen <irar at il dot ibm.com> 2010-10-03 11:43:15 UTC ---
I guess, it's a DI/DF problem...
If so, the code the vectorizer produces seems to be ok:

vector(2) real(kind=8) vect_cst_.14;
real(kind=8) reciptmp.5;
real(kind=4) dtime;
real(kind=8) D.1357;
...

D.1357_12 = (real(kind=8)) dtime_11(D);
reciptmp.5_60 = 1.0e+0 / D.1357_12;
...
vect_cst_.14_51 = {reciptmp.5_60, reciptmp.5_60};


VDI first shows up in ira.

pr45714-b.f.194r.sched1:
(insn 35 116 118 2 (set (reg:V2DF 158 [ vect_cst_.14 ])
        (vec_duplicate:V2DF (reg:DF 162 [ reciptmp.5 ]))) pr45714-b.f:14 993
{vsx_splat_v2df}
     (expr_list:REG_DEAD (reg:DF 162 [ reciptmp.5 ])
        (nil)))

pr45714-b.f.195r.ira:
(insn 35 136 118 2 (set (reg:V2DF 108 31 [orig:158 vect_cst_.14 ] [158])
        (vec_duplicate:V2DF (mem/u/c/i:DF (lo_sum:DI (reg:DI 11 11)
                    (const:DI (unspec:DI [
                                (symbol_ref/u:DI ("*.LC3") [flags 0x2])
                            ] 49))) [2 S8 A64]))) pr45714-b.f:14 993
{vsx_splat_v2df}
     (nil))

[Bug tree-optimization/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

--- Comment #2 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2010-10-03 12:01:16 UTC ---
Could this pr be related to pr45585?

[Bug tree-optimization/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2010-11-07 13:53:58 UTC ---
> Could this pr be related to pr45585?

The fix for pr45585 does not works for this pr.

[Bug tree-optimization/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[amodra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

Alan Modra <amodra at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|powerpc-apple-darwin9       |powerpc-apple-darwin9,
                   |                            |powerpc64-linux
                 CC|                            |amodra at gmail dot com

--- Comment #4 from Alan Modra <amodra at gmail dot com> 2011-03-15 09:18:31 UTC ---
Seen also on powerpc64-linux.  The problem here is that the vsx_splat_V2DF/V2DI
load and store instructions only support mem[reg] or mem[reg+reg] addressing,
but the MEM mode is not a vsx vector mode.  So reg_offset_addressing_ok_p used
in rs6000_legitimate_address_p and other address predicates returns true,
wrongly allowing lo_sum addresses.

[Bug target/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[amodra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

Alan Modra <amodra at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot       |amodra at gmail dot com
                   |gnu.org                     |

--- Comment #5 from Alan Modra <amodra at gmail dot com> 2011-03-15 09:40:35 UTC ---
Testing a fix

[Bug target/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[amodra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

Alan Modra <amodra at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code, patch
             Status|NEW                         |ASSIGNED
                URL|                            |http://gcc.gnu.org/ml/gcc-p
                   |                            |atches/2011-03/msg00829.htm
                   |                            |l

[Bug target/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[amodra at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

--- Comment #6 from Alan Modra <amodra at gcc dot gnu.org> 2011-03-16 06:26:33 UTC ---
Author: amodra
Date: Wed Mar 16 06:26:29 2011
New Revision: 171031

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=171031
Log:
    PR target/45844
    * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
    create invalid offset address for vsx splat insn.
    * config/rs6000/predicates.md (splat_input_operand): New.
    * config/rs6000/vsx.md (vsx_splat_*): Use it.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/rs6000/predicates.md
    trunk/gcc/config/rs6000/rs6000.c
    trunk/gcc/config/rs6000/vsx.md

[Bug target/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

--- Comment #7 from William J. Schmidt <wschmidt at gcc dot gnu.org> 2011-03-16 18:00:32 UTC ---
Author: wschmidt
Date: Wed Mar 16 18:00:23 2011
New Revision: 171057

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=171057
Log:
gcc:

    Backport from mainline:
    2010-09-19  Ira Rosen  <irar@il.ibm.com>

    PR tree-optimization/45714
    * tree-vect-stmts.c (vect_transform_stmt): Use a dummy statement
    created in vectorizable_call instead of the original statement in
    def stmt updates.

    Backport from mainline:
    2011-03-16  Alan Modra  <amodra@gmail.com>

    PR target/45844
    * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
    create invalid offset address for vsx splat insn.
    * config/rs6000/predicates.md (splat_input_operand): New.
    * config/rs6000/vsx.md (vsx_splat_*): Use it.


gcc/testsuite:

    Backport from mainline:
    2010-09-19  Ira Rosen  <irar@il.ibm.com>

    PR tree-optimization/45714
    * gfortran.dg/vect/pr45714-a.f: New test.
    * gfortran.dg/vect/pr45714-b.f: New test.

Added:
    branches/ibm/gcc-4_5-branch/gcc/testsuite/gfortran.dg/vect/pr45714-a.f
    branches/ibm/gcc-4_5-branch/gcc/testsuite/gfortran.dg/vect/pr45714-b.f
Modified:
    branches/ibm/gcc-4_5-branch/gcc/ChangeLog.ibm
    branches/ibm/gcc-4_5-branch/gcc/config/rs6000/predicates.md
    branches/ibm/gcc-4_5-branch/gcc/config/rs6000/rs6000.c
    branches/ibm/gcc-4_5-branch/gcc/config/rs6000/vsx.md
    branches/ibm/gcc-4_5-branch/gcc/testsuite/ChangeLog.ibm
    branches/ibm/gcc-4_5-branch/gcc/tree-vect-stmts.c

[Bug target/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

--- Comment #8 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-03-22 22:03:41 UTC ---
Any plan to backport the fix or should I close this pr?

[Bug target/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[amodra at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

--- Comment #9 from Alan Modra <amodra at gmail dot com> 2011-03-28 01:44:36 UTC ---
I an planning to backport to 4.6.1

[Bug target/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[dougkwan at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

--- Comment #10 from dougkwan at gcc dot gnu.org 2012-06-01 20:43:37 UTC ---
Author: dougkwan
Date: Fri Jun  1 20:43:31 2012
New Revision: 188126

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=188126
Log:
2012-06-01   Doug Kwan  <dougkwan@google.com>

    Backport r171031 from upstream trunk.
    2011-03-16  Alan Modra  <amodra@gmail.com>

        PR target/45844
        * config/rs6000/rs6000.c (rs6000_legitimize_reload_address):
        Don't create invalid offset address for vsx splat insn.
        * config/rs6000/predicates.md (splat_input_operand): New.
        * config/rs6000/vsx.md (vsx_splat_*): Use it.

    * contrib/testsuite-management/powerpc64-grtev2-linux-gnu.xfail:
    Remove expected failures because now the bug is fixed.

Modified:
    branches/google/gcc-4_6/ChangeLog.google-4_6
   
branches/google/gcc-4_6/contrib/testsuite-management/powerpc64-grtev2-linux-gnu.xfail
    branches/google/gcc-4_6/gcc/ChangeLog.google-4_6
    branches/google/gcc-4_6/gcc/config/rs6000/predicates.md
    branches/google/gcc-4_6/gcc/config/rs6000/rs6000.c
    branches/google/gcc-4_6/gcc/config/rs6000/vsx.md

[Bug target/45844] FAIL: gfortran.dg/vect/pr45714-b.f -O (internal compiler error)

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45844

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

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

--- Comment #11 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2012-08-31 14:40:57 UTC ---
Backported to 4.6 at revision 188126, closing as fixed.