[Bug target/60839] New: PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[Bug target/60839] New: PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[sebastian.huber@embedded-brains.de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

            Bug ID: 60839
           Summary: PowerPC: internal compiler error: in extract_insn, at
                    recog.c:2154
           Product: gcc
           Version: 4.8.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: sebastian.huber@embedded-brains.de

With a GCC 4.8 and 4.9 snapshot from 2014-04-14 I get an ICE during GCC build:

make[4]: Entering directory
`/scratch/git-build/b-gcc-git-powerpc-rtems4.11/powerpc-rtems4.11/m8540/libgcc'
# If this is the top-level multilib, build all the other
# multilibs.
/scratch/git-build/b-gcc-git-powerpc-rtems4.11/./gcc/xgcc
-B/scratch/git-build/b-gcc-git-powerpc-rtems4.11/./gcc/ -nostdinc
-B/scratch/git-build/b-gcc-git-powerpc-rtems4.11/powerpc-rtems4.11/newlib/
-isystem
/scratch/git-build/b-gcc-git-powerpc-rtems4.11/powerpc-rtems4.11/newlib/targ-include
-isystem /home/sh/archive/gcc-git/newlib/libc/include
-B/opt/rtems-4.11/powerpc-rtems4.11/bin/
-B/opt/rtems-4.11/powerpc-rtems4.11/lib/ -isystem
/opt/rtems-4.11/powerpc-rtems4.11/include -isystem
/opt/rtems-4.11/powerpc-rtems4.11/sys-include    -g -O2 -mcpu=8540 -O2
-I/home/sh/archive/gcc-git/libgcc/../newlib/libc/sys/rtems/include -g -O2
-DIN_GCC -DCROSS_DIRECTORY_STRUCTURE  -W -Wall -Wwrite-strings -Wcast-qual
-Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem
./include   -g -DIN_LIBGCC2 -fbuilding-libgcc -fno-stack-protector
-Dinhibit_libc  -I. -I. -I../../.././gcc -I/home/sh/archive/gcc-git/libgcc
-I/home/sh/archive/gcc-git/libgcc/. -I/home/sh/archive/gcc-git/libgcc/../gcc
-I/home/sh/archive/gcc-git/libgcc/../include  -DHAVE_CC_TLS  -o _powidf2.o -MT
_powidf2.o -MD -MP -MF _powidf2.dep -DL_powidf2 -c
/home/sh/archive/gcc-git/libgcc/libgcc2.c -fvisibility=hidden -DHIDE_EXPORTS
/home/sh/archive/gcc-git/libgcc/libgcc2.c: In function '__powidf2':
/home/sh/archive/gcc-git/libgcc/libgcc2.c:1777:1: error: unrecognizable insn:
 }
 ^
(insn 2 9 3 2 (set (reg/v:DF 131 [ x ])
        (reg:DF 3 3 [ x ])) /home/sh/archive/gcc-git/libgcc/libgcc2.c:1767 -1
     (nil))
/home/sh/archive/gcc-git/libgcc/libgcc2.c:1777:1: internal compiler error: in
extract_insn, at recog.c:2154
0x6cf2f5 _fatal_insn(char const*, rtx_def const*, char const*, int, char
const*)
        /home/sh/archive/gcc-git/gcc/rtl-error.c:109
0x6cf329 _fatal_insn_not_found(rtx_def const*, char const*, int, char const*)
        /home/sh/archive/gcc-git/gcc/rtl-error.c:117
0x6a30b3 extract_insn(rtx_def*)
        /home/sh/archive/gcc-git/gcc/recog.c:2154
0x5a8b33 instantiate_virtual_regs_in_insn
        /home/sh/archive/gcc-git/gcc/function.c:1561
0x5a8b33 instantiate_virtual_regs
        /home/sh/archive/gcc-git/gcc/function.c:1928
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.
make[4]: *** [_powidf2.o] Error 1

[Bug target/60839] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[sebastian.huber@embedded-brains.de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

Sebastian Huber <sebastian.huber@embedded-brains.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |powerpc-rtems4.11
                 CC|                            |meissner at linux dot vnet.ibm.com
                   |                            |,
                   |                            |wschmidt at linux dot vnet.ibm.com
      Known to work|                            |4.8.2
      Known to fail|                            |4.8.3, 4.9.0

--- Comment #1 from Sebastian Huber <sebastian.huber@embedded-brains.de> ---
A git bisect showed that this is the commit that breaks the GCC build:

66c7a00630e9445ac478dfa1bd6ce900bf6cd55e is the first bad commit
commit 66c7a00630e9445ac478dfa1bd6ce900bf6cd55e
Author: wschmidt <wschmidt@138bc75d-0d04-0410-961f-82ee72b054a4>
Date:   Fri Apr 4 15:14:01 2014 +0000

    [gcc]

    2014-04-04  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

        Back port mainline subversion id 209025.
        2014-04-02  Michael Meissner  <meissner@linux.vnet.ibm.com>

        PR target/60735
        * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
        software floating point or no floating point registers, do not
        allow any type in the FPRs.  Eliminate a test for SPE SIMD types
        in GPRs that occurs after we tested for GPRs that would never be
        true.

        * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
        Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
        since the FMOVE64 type is DFmode/DDmode.  If TARGET_E500_DOUBLE,
        specifically allow DDmode, since that does not use the SPE SIMD
        instructions.



    git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gcc-4_8-branch@209116
138bc75d-0d04-0410-961f-82ee72b054a4

:040000 040000 3d009a9aa1e94d5aaf9c697401c810e32bd06138
7f587451ff81b62bc286b06faff08b80383ab166 M      gcc

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
   Target Milestone|---                         |4.8.3
            Summary|PowerPC: internal compiler  |[4.8/4.9/4.10 Regression]
                   |error: in extract_insn, at  |PowerPC: internal compiler
                   |recog.c:2154                |error: in extract_insn, at
                   |                            |recog.c:2154

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
P1 regression on the 4.8 branch.  I suppose it would help if you provide
preprocessed source.

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[sebastian.huber@embedded-brains.de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

--- Comment #3 from Sebastian Huber <sebastian.huber@embedded-brains.de> ---
Created attachment 32599
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=32599&action=edit
Pre-processed source of libgcc2.c

Command line without pre-processor relevant options:

xgcc -mcpu=8540 -g -O2 -W -Wall -Wwrite-strings -Wcast-qual -Wstrict-prototypes
-Wmissing-prototypes -Wold-style-definition -g -fbuilding-libgcc
-fno-stack-protector -c libgcc2.i -fvisibility=hidden

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

Bill Schmidt <wschmidt at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |wschmidt at gcc dot gnu.org

--- Comment #4 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Mike is out today.  I'll see if I can sort out what's going on, but this will
probably need his input.

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

--- Comment #5 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Per discussion on IRC, we are going to revert this patch on 4.8, 4.9, and 4.10.
 This will mean that PR60735 will have to be reopened for a better fix.  The
patch seems to leave things in a worse state than previously (affects more
ports and more code), so this seems best for now.  This still must be fixed
properly prior to 4.8.3 release.

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

--- Comment #6 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Author: wschmidt
Date: Tue Apr 15 18:20:01 2014
New Revision: 209425

URL: http://gcc.gnu.org/viewcvs?rev=209425&root=gcc&view=rev
Log:
2014-04-15  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

    PR target/60839
    Revert following patch

    2014-04-02  Michael Meissner  <meissner@linux.vnet.ibm.com>

    PR target/60735
    * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
    software floating point or no floating point registers, do not
    allow any type in the FPRs.  Eliminate a test for SPE SIMD types
    in GPRs that occurs after we tested for GPRs that would never be
    true.

    * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
    Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
    since the FMOVE64 type is DFmode/DDmode.  If TARGET_E500_DOUBLE,
    specifically allow DDmode, since that does not use the SPE SIMD
    instructions.


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

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

--- Comment #7 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Author: wschmidt
Date: Tue Apr 15 18:25:09 2014
New Revision: 209426

URL: http://gcc.gnu.org/viewcvs?rev=209426&root=gcc&view=rev
Log:
2014-04-15  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

    PR target/60839
    Revert the following patch

    2014-04-02  Michael Meissner  <meissner@linux.vnet.ibm.com>

    PR target/60735
    * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
    software floating point or no floating point registers, do not
    allow any type in the FPRs.  Eliminate a test for SPE SIMD types
    in GPRs that occurs after we tested for GPRs that would never be
    true.

    * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
    Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
    since the FMOVE64 type is DFmode/DDmode.  If TARGET_E500_DOUBLE,
    specifically allow DDmode, since that does not use the SPE SIMD
    instructions.


Modified:
    branches/gcc-4_9-branch/gcc/ChangeLog
    branches/gcc-4_9-branch/gcc/config/rs6000/rs6000.c
    branches/gcc-4_9-branch/gcc/config/rs6000/rs6000.md

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

--- Comment #8 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Author: wschmidt
Date: Tue Apr 15 18:30:21 2014
New Revision: 209430

URL: http://gcc.gnu.org/viewcvs?rev=209430&root=gcc&view=rev
Log:
2014-04-15  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

    PR target/60839
    Revert the following patch

    2014-04-04  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

    Back port mainline subversion id 209025.
    2014-04-02  Michael Meissner  <meissner@linux.vnet.ibm.com>

    PR target/60735
    * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
    software floating point or no floating point registers, do not
    allow any type in the FPRs.  Eliminate a test for SPE SIMD types
    in GPRs that occurs after we tested for GPRs that would never be
    true.

    * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
    Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
    since the FMOVE64 type is DFmode/DDmode.  If TARGET_E500_DOUBLE,
    specifically allow DDmode, since that does not use the SPE SIMD
    instructions.


Modified:
    branches/gcc-4_8-branch/gcc/ChangeLog
    branches/gcc-4_8-branch/gcc/config/rs6000/rs6000.c
    branches/gcc-4_8-branch/gcc/config/rs6000/rs6000.md

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

--- Comment #9 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Sebastian, sorry for the problems.  Please double check that reverting this
patch has fixed your bootstrap issue and mark the bug resolved if so.  Thanks!

[Bug target/60839] [4.8/4.9/4.10 Regression] PowerPC: internal compiler error: in extract_insn, at recog.c:2154

[sebastian.huber@embedded-brains.de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60839

Sebastian Huber <sebastian.huber@embedded-brains.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED
      Known to fail|4.8.3, 4.9.0                |

--- Comment #10 from Sebastian Huber <sebastian.huber@embedded-brains.de> ---
Thanks for the quick fix.  It run the GCC test suite and it works again.