[Bug bootstrap/55089] New: ICE in varasm.c during bootstrap-lean on x86_64-pc-solaris2.10

[Bug bootstrap/55089] New: ICE in varasm.c during bootstrap-lean on x86_64-pc-solaris2.10

[Tim.Mooney at ndsu dot edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55089

             Bug #: 55089
           Summary: ICE in varasm.c during bootstrap-lean on
                    x86_64-pc-solaris2.10
    Classification: Unclassified
           Product: gcc
           Version: 4.7.2
            Status: UNCONFIRMED
          Severity: blocker
          Priority: P3
         Component: bootstrap
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: Tim.Mooney@ndsu.edu


Created attachment 28539
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28539
config.log from libgcc directory, where ICE occurs

I'm attempting to build gcc 4.7.2 on x86_64-sun-solaris2.10.  I'm using the
Oracle Studio 12.3 compiler suite to build the initial xgcc.  Relevant settings
are

CC="cc -m64"
CFLAGS=-m64
export CC 
CPPFLAGS="-I/local/gnu/include -I/local/include"
CFLAGS="-m64 -I/local/gnu/include -I/local/include"
LDFLAGS="-m64 -L/local/gnu/lib/64 -L/local/lib/64"
export CPPFLAGS CFLAGS LDFLAGS

BOOT_CFLAGS="-O -g"
export BOOT_CFLAGS

The build proceeds to where configure outputs

gmake[3]: Leaving directory `/local/src/RPM/BUILD/gcc-4.7.2-obj/gcc'
mkdir x86_64-pc-solaris2.10/libgcc
Checking multilib configuration for libgcc...
Configuring stage 1 in x86_64-pc-solaris2.10/libgcc
configure: creating cache ./config.cache
checking build system type... x86_64-pc-solaris2.10
checking host system type... x86_64-pc-solaris2.10
checking for --enable-version-specific-runtime-libs... no
checking for a BSD-compatible install... /local/gnu/bin/ginstall -c
checking for gawk... nawk
checking for x86_64-pc-solaris2.10-ar... ar
checking for x86_64-pc-solaris2.10-lipo... lipo
checking for x86_64-pc-solaris2.10-nm...
/local/src/RPM/BUILD/gcc-4.7.2-obj/./gc
c/nm
checking for x86_64-pc-solaris2.10-ranlib... ranlib
checking for x86_64-pc-solaris2.10-strip... strip
checking whether ln -s works... yes
checking for x86_64-pc-solaris2.10-gcc...
/local/src/RPM/BUILD/gcc-4.7.2-obj/./g
cc/xgcc -B/local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/
-B/local/gnu/x86_64-pc-solar
is2.10/bin/ -B/local/gnu/x86_64-pc-solaris2.10/lib/ -isystem
/local/gnu/x86_64-p
c-solaris2.10/include -isystem /local/gnu/x86_64-pc-solaris2.10/sys-include   
checking for suffix of object files... configure: error: in
`/local/src/RPM/BUIL
D/gcc-4.7.2-obj/x86_64-pc-solaris2.10/libgcc':
configure: error: cannot compute suffix of object files: cannot compile
See `config.log' for more details.


Examining
/local/src/RPM/BUILD/gcc-4.7.2-obj/x86_64-pc-solaris2.10/libgcc/config.log I
see

Thread model: posix
gcc version 4.7.2 (GCC)
configure:3355: $? = 0
configure:3344: /local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/xgcc
-B/local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/
-B/local/gnu/x86_64-pc-solaris2.10/bin/ -B/local/gnu/x86_64-pc-solaris2.10/lib/
-isystem /local/gnu/x86_64-pc-solaris2.10/include -isystem
/local/gnu/x86_64-pc-solaris2.10/sys-include    -V >&5
xgcc: error: unrecognized command line option '-V'
xgcc: fatal error: no input files
compilation terminated.
configure:3355: $? = 1
configure:3344: /local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/xgcc
-B/local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/
-B/local/gnu/x86_64-pc-solaris2.10/bin/ -B/local/gnu/x86_64-pc-solaris2.10/lib/
-isystem /local/gnu/x86_64-pc-solaris2.10/include -isystem
/local/gnu/x86_64-pc-solaris2.10/sys-include    -qversion >&5
xgcc: error: unrecognized command line option '-qversion'
xgcc: fatal error: no input files
compilation terminated.
configure:3355: $? = 1
configure:3371: /local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/xgcc
-B/local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/
-B/local/gnu/x86_64-pc-solaris2.10/bin/ -B/local/gnu/x86_64-pc-solaris2.10/lib/
-isystem /local/gnu/x86_64-pc-solaris2.10/include -isystem
/local/gnu/x86_64-pc-solaris2.10/sys-include    -o conftest -g -m64
-I/local/gnu/include -I/local/include   conftest.c  >&5
conftest.c:16:1: internal compiler error: in ?, at varasm.c:2555
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
configure:3374: $? = 1
configure:3562: checking for suffix of object files
configure:3584: /local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/xgcc
-B/local/src/RPM/BUILD/gcc-4.7.2-obj/./gcc/
-B/local/gnu/x86_64-pc-solaris2.10/bin/ -B/local/gnu/x86_64-pc-solaris2.10/lib/
-isystem /local/gnu/x86_64-pc-solaris2.10/include -isystem
/local/gnu/x86_64-pc-solaris2.10/sys-include    -c -g -m64 -I/local/gnu/include
-I/local/include  conftest.c >&5
conftest.c:16:1: internal compiler error: in ?, at varasm.c:2555
Please submit a full bug report,
with preprocessed source if appropriate.

I'm attaching two files; the entire log of what configure output from the
beginning, and the config.log from the libgcc directory.

Note that it's been quite a while since I last built gcc, so it's entirely
possible that something I'm doing is no longer the recommended way to build
gcc, but I would expect a make error rather than an internal compiler error in
that case.

I would be happy to provide any additional information that's needed.

[Bug bootstrap/55089] ICE in varasm.c during bootstrap-lean on x86_64-pc-solaris2.10

[Tim.Mooney at ndsu dot edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55089

--- Comment #1 from Tim Mooney <Tim.Mooney at ndsu dot edu> 2012-10-26 21:12:20 UTC ---
Created attachment 28540
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28540
output from the rpmbuild that was run to configure gcc

I use rpmbuild to package local software.  This attachment has the complete
output from running the rpmbuild command.  It's mainly useful because it shows
the important environment settings and all of the configure and compilation
output, up to the failure in libgcc.

[Bug bootstrap/55089] ICE in varasm.c during bootstrap-lean on x86_64-pc-solaris2.10

[Tim.Mooney at ndsu dot edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55089

--- Comment #2 from Tim Mooney <Tim.Mooney at ndsu dot edu> 2012-10-26 21:42:39 UTC ---
I just tried the older 4.7.0 and the same error happens, though the line number
in varasm.c is slightly different:

configure:3625: checking for suffix of object files
configure:3647: /local/src/RPM/BUILD/gcc-4.7.0-obj/./gcc/xgcc
-B/local/src/RPM/BUILD/gcc-4.7.0-obj/./gcc/
-B/local/gnu/x86_64-pc-solaris2.10/bin/ -B/local/gnu/x86_64-pc-solaris2.10/lib/
-isystem /local/gnu/x86_64-pc-solaris2.10/include -isystem
/local/gnu/x86_64-pc-solaris2.10/sys-include    -c -g -m64 -I/local/gnu/include
-I/local/include  conftest.c >&5
conftest.c:16:1: internal compiler error: in ?, at varasm.c:2528

[Bug target/55089] ICE in varasm.c during bootstrap-lean on x86_64-pc-solaris2.10

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55089

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2012-12-09
     Ever Confirmed|0                           |1

[Bug target/55089] ICE in varasm.c during bootstrap-lean on x86_64-pc-solaris2.10

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55089

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |build, ice-on-valid-code
          Component|bootstrap                   |target
               Host|                            |x86_64-sun-solaris2.10
              Build|                            |x86_64-sun-solaris2.10
           Severity|blocker                     |normal

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> 2012-12-09 02:42:36 UTC ---
CC=cc -m64
CFLAGS=-m64
+ export CC 
CPPFLAGS= -I/local/gnu/include -I/local/include
CFLAGS=-m64 -I/local/gnu/include -I/local/include
LDFLAGS= -m64 -L/local/gnu/lib/64 -L/local/lib/64
+ export CPPFLAGS CFLAGS LDFLAGS 
+ ../gcc-4.7.2/configure --prefix=/local/gnu --exec-prefix=/local/gnu 


Your stage1 compiler is producing an internal compiler error.
Either the compiler you are starting with is miscompiling the stage1 compiler
(most likely because I have seen other people compile on x86_64-pc-solaris2.10)
or your stage1 compiler has a bug in it.

> I'm using the Oracle Studio 12.3 compiler suite to build the initial xgcc.

Have you tried building it with a prebuilt GCC?  Maybe even trying to do
i386-pc-solaris2.10 compiler to begin with and then doing a
x86_64-pc-solaris2.10 compiler with that newly build gcc?  I am thinking
Oracle's compiler has bugs in it with respect of 64bit mode.