[Bug bootstrap/48371] New: [4.7 Regression] bootstrap failure

[Bug bootstrap/48371] New: [4.7 Regression] bootstrap failure

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

           Summary: [4.7 Regression] bootstrap failure
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: hjl.tools@gmail.com


On Linux/x86, revision171734 failed to bootstrap:

http://gcc.gnu.org/ml/gcc-regression/2011-03/msg00594.html
http://gcc.gnu.org/ml/gcc-regression/2011-03/msg00595.html

../../src-trunk/libiberty/floatformat.c: In function 'floatformat_from_double':
../../src-trunk/libiberty/floatformat.c:699:1: error: insn does not satisfy its
constraints:
(insn 159 157 278 17 (parallel [
            (set (reg:SI 21 xmm0)
                (unsigned_fix:SI (reg/v:DF 22 xmm1 [orig:102 mant ] [102])))
            (use (reg:V2DF 179))
            (clobber (reg:V2DF 23 xmm2))
            (clobber (reg:V2DF 24 xmm3))
        ]) ../../src-trunk/libiberty/floatformat.c:675 115 {*fixuns_truncdf_1}
     (expr_list:REG_EQUAL (unsigned_fix:SI (reg/v:DF 22 xmm1 [orig:102 mant ]
[102]))
        (nil)))
../../src-trunk/libiberty/floatformat.c:699:1: internal compiler error: in
reload_cse_simplify_operands, at postreload.c:403
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Revision 171729 is OK.

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |law at redhat dot com

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> 2011-03-30 19:36:49 UTC ---
It is cause by revision 171731:

http://gcc.gnu.org/ml/gcc-cvs/2011-03/msg01157.html

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

--- Comment #2 from H.J. Lu <hjl.tools at gmail dot com> 2011-03-30 19:38:24 UTC ---
Created attachment 23822
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23822
A testcase

[hjl@gnu-34 delta]$ /export/regression/rrs/171731/usr/bin/gcc  -S -m32 -msse2
-mfpmath=sse -fpic -O2 testcase.c
testcase.c: In function ‘__signbitf’:
testcase.c:11:1: warning: excess elements in union initializer [enabled by
default]
testcase.c:11:1: warning: (near initialization for ‘__u’) [enabled by default]
testcase.c: In function ‘__signbit’:
testcase.c:17:1: warning: excess elements in union initializer [enabled by
default]
testcase.c:17:1: warning: (near initialization for ‘__u’) [enabled by default]
testcase.c: At top level:
testcase.c:65:4: warning: excess elements in struct initializer [enabled by
default]
testcase.c:65:4: warning: (near initialization for ‘floatformat_ieee_half_big’)
[enabled by default]
testcase.c:65:4: warning: excess elements in struct initializer [enabled by
default]
testcase.c:65:4: warning: (near initialization for ‘floatformat_ieee_half_big’)
[enabled by default]
testcase.c:65:4: warning: excess elements in struct initializer [enabled by
default]
testcase.c:65:4: warning: (near initialization for ‘floatformat_ieee_half_big’)
[enabled by default]
testcase.c: In function ‘floatformat_from_double’:
testcase.c:129:2: error: insn does not satisfy its constraints:
(insn 36 34 98 4 (parallel [
            (set (reg:SI 21 xmm0)
                (unsigned_fix:SI (reg/v:DF 22 xmm1 [orig:67 mant ] [67])))
            (use (reg:V2DF 91))
            (clobber (reg:V2DF 23 xmm2))
            (clobber (reg:V2DF 24 xmm3))
        ]) testcase.c:123 153 {*fixuns_truncdf_1}
     (expr_list:REG_EQUIV (mem:SI (reg/f:SI 7 sp) [0 S4 A32])
        (expr_list:REG_EQUAL (unsigned_fix:SI (reg/v:DF 22 xmm1 [orig:67 mant ]
[67]))
            (nil))))
testcase.c:129:2: internal compiler error: in reload_cse_simplify_operands, at
postreload.c:403
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
[hjl@gnu-34 delta]$

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[law at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

--- Comment #3 from Jeffrey A. Law <law at redhat dot com> 2011-03-30 21:21:56 UTC ---
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 03/30/11 13:38, hjl.tools at gmail dot com wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371
> 
> --- Comment #2 from H.J. Lu <hjl.tools at gmail dot com> 2011-03-30 19:38:24 UTC ---
> Created attachment 23822
>   --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23822
> A testcase
> 
> [hjl@gnu-34 delta]$ /export/regression/rrs/171731/usr/bin/gcc  -S -m32 -msse2
> -mfpmath=sse -fpic -O2 testcase.c
> testcase.c: In function ‘__signbitf’:
> testcase.c:11:1: warning: excess elements in union initializer [enabled by
> default]
> testcase.c:11:1: warning: (near initialization for ‘__u’) [enabled by default]
> testcase.c: In function ‘__signbit’:
> testcase.c:17:1: warning: excess elements in union initializer [enabled by
> default]
> testcase.c:17:1: warning: (near initialization for ‘__u’) [enabled by default]
> testcase.c: At top level:
> testcase.c:65:4: warning: excess elements in struct initializer [enabled by
> default]
> testcase.c:65:4: warning: (near initialization for ‘floatformat_ieee_half_big’)
> [enabled by default]
> testcase.c:65:4: warning: excess elements in struct initializer [enabled by
> default]
> testcase.c:65:4: warning: (near initialization for ‘floatformat_ieee_half_big’)
> [enabled by default]
> testcase.c:65:4: warning: excess elements in struct initializer [enabled by
> default]
> testcase.c:65:4: warning: (near initialization for ‘floatformat_ieee_half_big’)
> [enabled by default]
> testcase.c: In function ‘floatformat_from_double’:
> testcase.c:129:2: error: insn does not satisfy its constraints:
> (insn 36 34 98 4 (parallel [
>             (set (reg:SI 21 xmm0)
>                 (unsigned_fix:SI (reg/v:DF 22 xmm1 [orig:67 mant ] [67])))
>             (use (reg:V2DF 91))
>             (clobber (reg:V2DF 23 xmm2))
>             (clobber (reg:V2DF 24 xmm3))
>         ]) testcase.c:123 153 {*fixuns_truncdf_1}
>      (expr_list:REG_EQUIV (mem:SI (reg/f:SI 7 sp) [0 S4 A32])
>         (expr_list:REG_EQUAL (unsigned_fix:SI (reg/v:DF 22 xmm1 [orig:67 mant ]
> [67]))
>             (nil))))
> testcase.c:129:2: internal compiler error: in reload_cse_simplify_operands, at
> postreload.c:403
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <http://gcc.gnu.org/bugs.html> for instructions.
Fix in testing.
jeff
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJNk57oAAoJEBRtltQi2kC77zIH/jrhxllibFAnmJ0M7w0p6s6o
Aw+LGY3qXvOsqQKYQjnR85BLPeS9T75NRiJsjDI78uG3RD44OztpSJ+ifxHNVOOp
ndVhAvc68WTlBmTtqGbuWWPuOBQ+NUZE9XjKju9pB1Njq79f7oKrnXYdcCK6xJWl
4zVexWkOsoY9xAKfZxA+StFuCJC4d7QoNqu9xXCJorN5TCznoSOIzMgr0Aerndkr
b6rRAVNPo6JPNUefhwnTLaVmbuaPcw5LOve4QpMBk14KxAnXRdiHGZ+aBTVyRLxt
3vzmEHrs/f+T/fheLp2q1g5zPJHjIlx7R3lhB3gSWAANSlZPQ+MyoM/FJde2ePA=
=3FMt
-----END PGP SIGNATURE-----

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[law at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

Jeffrey A. Law <law at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |zsojka at seznam dot cz

--- Comment #4 from Jeffrey A. Law <law at redhat dot com> 2011-03-30 21:26:31 UTC ---
*** Bug 48373 has been marked as a duplicate of this bug. ***

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

Hans-Peter Nilsson <hp at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.03.30 22:40:17
                 CC|                            |hp at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #5 from Hans-Peter Nilsson <hp at gcc dot gnu.org> 2011-03-30 22:40:17 UTC ---
FWIW, cris-elf too, worked:firstbreak 171723:171735 as follows:

/obj/dir/./gcc/xgcc -B/obj/dir/./gcc/ -nostdinc -B/obj/dir/cris-elf/newlib/
-isystem /obj/dir/cris-elf/newlib/targ-include -isystem
/src/gcc/newlib/libc/include -B/obj/dir/cris-elf/libgloss/cris
-L/obj/dir/cris-elf/libgloss/libnosys -L/src/gcc/libgloss/cris
-B/pre/fix/cris-elf/bin/ -B/pre/fix/cris-elf/lib/ -isystem
/pre/fix/cris-elf/include -isystem /pre/fix/cris-elf/sys-include    -g -O2
-march=v32 -mbest-lib-options -O2  -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
-D__GCC_FLOAT_NOT_NEEDED -fno-stack-protector -Dinhibit_libc  -I. -I.
-I../../.././gcc -I/src/gcc/libgcc -I/src/gcc/libgcc/. -I/src/gcc/libgcc/../gcc
-I/src/gcc/libgcc/../include  -DHAVE_CC_TLS -DUSE_EMUTLS -o _mulsc3.o -MT
_mulsc3.o -MD -MP -MF _mulsc3.dep -DL_mulsc3 -c
/src/gcc/libgcc/../gcc/libgcc2.c \

/src/gcc/libgcc/../gcc/libgcc2.c: In function '__mulsc3':
/src/gcc/libgcc/../gcc/libgcc2.c:1885:1: error: insn does not satisfy its
constraints:
(insn 22 21 24 2 (set (reg/v:SF 43 [ ac ])
        (reg:SF 10 r10)) /src/gcc/libgcc/../gcc/libgcc2.c:1827 47 {movsf}
     (expr_list:REG_EQUAL (mult:SF (reg/v:SF 5 r5 [orig:157 a ] [157])
            (reg/v:SF 4 r4 [orig:159 c ] [159]))
        (nil)))
/src/gcc/libgcc/../gcc/libgcc2.c:1885:1: internal compiler error: in
reload_cse_simplify_operands, at postreload.c:403

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

--- Comment #6 from Hans-Peter Nilsson <hp at gcc dot gnu.org> 2011-03-30 23:26:54 UTC ---
Created attachment 23828
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23828
Preprocessed libgcc2.i, repeat ICE for cris-elf with  -fpreprocessed libgcc2.i
-march=v32 -O2

In the odd event that you'd want to verify a fix on more than x86, here's a
test-case for cris-elf (and as it turns out, crisv32-elf).

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[law at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

Jeffrey A. Law <law at redhat dot com> changed:

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

--- Comment #7 from Jeffrey A. Law <law at redhat dot com> 2011-03-31 00:06:22 UTC ---
Fixed by follow up patch to reload1.c

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[law at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

--- Comment #8 from Jeffrey A. Law <law at redhat dot com> 2011-03-31 00:07:31 UTC ---
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 03/30/11 17:27, hp at gcc dot gnu.org wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371
> 
> --- Comment #6 from Hans-Peter Nilsson <hp at gcc dot gnu.org> 2011-03-30 23:26:54 UTC ---
> Created attachment 23828
>   --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23828
> Preprocessed libgcc2.i, repeat ICE for cris-elf with  -fpreprocessed libgcc2.i
> -march=v32 -O2
> 
> In the odd event that you'd want to verify a fix on more than x86, here's a
> test-case for cris-elf (and as it turns out, crisv32-elf).
I just checked in a fix for 48371 (reload1.c).  If you could verify that
change fixes the cris problems it would be greatly appreciated (that
patch wasn't supposed to change code generation at all :-)

jeff

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJNk8W6AAoJEBRtltQi2kC7DOkH/2kbHSHLxNO7ySsDR6EApZmH
biGUe4UFi+Rh4mgqOCB/ZahKD+ak0IP0/KJaolmGh6hDn9x+S1wlsRsimwjPdyUI
qaX2toZiKmiEkbadgKPV1PGWzHHeMfwuRUG1qtwiFOutFyitc8XDap5IO+V7sYpl
AjpM06/WbD/tbeAzFL8GVLXKpxDrRoZvog+GZ/ruPAoSf7xDfAFpvfmpuS8gVE57
MRBbJU1AEoN/JoCFLfV0NCMB2lmL99dEzsjz/H8GMEXaPPDlzWtc0Zpl0QucSNYr
tV/HASQj3zCdYj7k3rxXL/i8cNSRuiuAMtMjnLV2t/hb2k76qNnRs+21cDw+WYU=
=SMF8
-----END PGP SIGNATURE-----

[Bug bootstrap/48371] [4.7 Regression] bootstrap failure

[hp at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48371

--- Comment #9 from Hans-Peter Nilsson <hp at gcc dot gnu.org> 2011-03-31 04:55:07 UTC ---
(In reply to comment #8)
> If you could verify that
> change fixes the cris problems it would be greatly appreciated (that
> patch wasn't supposed to change code generation at all :-)

For closure, yes, all fine at r171766.