Yet another sparc-sun-solaris2.8 bootstrap failure

Yet another sparc-sun-solaris2.8 bootstrap failure

[Gerald Pfeifer]

It's rather disturbing to experience yet another bootstrap failure on
sparc-sun-solaris2.8.

This is about the fourth (or fifth?) *different* bootstrap failure on
this platform on the release branch. :-(

creating libffi_convenience.la
(cd .libs && rm -f libffi_convenience.la && ln -s ../libffi_convenience.la libffi_convenience.la)
/bin/ksh ./libtool --mode=link /files/pfeifer/OBJ-0408-1947/gcc/xgcc -B/files/pfeifer/OBJ-0408-1947/gcc/ -B/sw/gcc-current/sparc-sun-solaris2.8/bin/ -B/sw/gcc-current/sparc-sun-solaris2.8/lib/ -isystem /sw/gcc-current/sparc-sun-solaris2.8/include  -m64 -fexceptions -g -O2  -m64 -m64 -o libffi.la -rpath /sw/gcc-current/lib/sparcv9 -release 2.00-beta  src/debug.lo src/prep_cif.lo src/types.lo src/raw_api.lo src/java_raw_api.lo src/sparc/ffi.lo src/sparc/v8.lo src/sparc/v9.lo
/files/pfeifer/OBJ-0408-1947/gcc/xgcc -B/files/pfeifer/OBJ-0408-1947/gcc/ -B/sw/gcc-current/sparc-sun-solaris2.8/bin/ -B/sw/gcc-current/sparc-sun-solaris2.8/lib/ -isystem /sw/gcc-current/sparc-sun-solaris2.8/include  -m64 -shared -Wl,-h -Wl,libffi-2.00-beta.so -o .libs/libffi-2.00-beta.so  src/.libs/debug.o src/.libs/prep_cif.o src/.libs/types.o src/.libs/raw_api.o src/.libs/java_raw_api.o src/sparc/.libs/ffi.o src/sparc/.libs/v8.o src/sparc/.libs/v9.o  -lc
ld: fatal: relocation error: R_SPARC_RELATIVE: file src/sparc/.libs/v8.o: symbol <unknown>: offset 0x10298c is non-aligned
collect2: ld returned 1 exit status
gmake[7]: *** [libffi.la] Error 1
gmake[7]: Leaving directory `/files/pfeifer/OBJ-0408-1947/sparc-sun-solaris2.8/sparcv9/libffi'
gmake[6]: *** [all-recursive] Error 1
gmake[6]: Leaving directory `/files/pfeifer/OBJ-0408-1947/sparc-sun-solaris2.8/sparcv9/libffi'
gmake[5]: *** [all-recursive-am] Error 2
gmake[5]: Leaving directory `/files/pfeifer/OBJ-0408-1947/sparc-sun-solaris2.8/sparcv9/libffi'
gmake[4]: *** [multi-do] Error 1
gmake[4]: Leaving directory `/files/pfeifer/OBJ-0408-1947/sparc-sun-solaris2.8/libffi'
gmake[3]: *** [all-multi] Error 2
gmake[3]: Leaving directory `/files/pfeifer/OBJ-0408-1947/sparc-sun-solaris2.8/libffi'
gmake[2]: *** [all-recursive-am] Error 2
gmake[2]: Leaving directory `/files/pfeifer/OBJ-0408-1947/sparc-sun-solaris2.8/libffi'
gmake[1]: *** [all-target-libffi] Error 2
gmake[1]: Leaving directory `/files/pfeifer/OBJ-0408-1947'

Gerald

Re: Yet another sparc-sun-solaris2.8 bootstrap failure

[Jakub Jelinek]

On Mon, Apr 08, 2002 at 10:15:05PM +0200, Gerald Pfeifer wrote:
> It's rather disturbing to experience yet another bootstrap failure on
> sparc-sun-solaris2.8.
> 
> This is about the fourth (or fifth?) *different* bootstrap failure on
> this platform on the release branch. :-(

I've just commited a fix for this.

	Jakub

Re: Yet another sparc-sun-solaris2.8 bootstrap failure

[Gerald Pfeifer]

On Mon, 8 Apr 2002, Jakub Jelinek wrote:
>> This is about the fourth (or fifth?) *different* bootstrap failure on
>> this platform on the release branch. :-(
> I've just commited a fix for this.

Thanks!

Now we are back at the other libffi related bootstrap failure

  http://gcc.gnu.org/ml/gcc/2002-04/msg00029.html

  | checking size of double... 8
  | checking size of long double... 16
  | checking size of void *... 8
  | checking whether byte ordering is bigendian... cross-compiling... unknown
  | checking to probe for byte ordering... guessing bigendian ...  unknown
  | configure: error: unknown endianess - sorry
  | gmake[1]: *** [configure-target-libffi] Error 1
  | gmake[1]: Leaving directory `/files/pfeifer/OBJ-0402-1049'
  | gmake: *** [bootstrap-lean] Error 2

which occurs only on 32-bit sparc hosts:

  http://gcc.gnu.org/ml/gcc/2002-04/msg00080.html

  | I have it! The difference is not whether I'm testing trunk or branch,
  | the difference is whether the host is
  |
  |  % isainfo -v
  |  64-bit sparcv9 applications
  |  32-bit sparc applications
  |
  | which is okay, or
  |
  |  % isainfo -v
  |  32-bit sparc applications
  |
  | which is where this configure test breaks.

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Re: Yet another sparc-sun-solaris2.8 bootstrap failure

[Jakub Jelinek]

On Tue, Apr 09, 2002 at 03:10:41PM +0200, Gerald Pfeifer wrote:
> which occurs only on 32-bit sparc hosts:

s/sparc/sparc-sun-solaris/, sparc-linux has no problems.

>   http://gcc.gnu.org/ml/gcc/2002-04/msg00080.html

	Jakub

Re: Yet another sparc-sun-solaris2.8 bootstrap failure

[Tom Tromey]

>>>>> "Gerald" == Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at> writes:

Gerald> which occurs only on 32-bit sparc hosts:
Gerald>   http://gcc.gnu.org/ml/gcc/2002-04/msg00080.html

Unfortunately I don't think I have access to any such machine.  So I
don't know what I can do to help.

Right now I can't get 3.1 to build on Solaris 2.7 either.  The target
libiberty doesn't configure for me.  It could be some local problem
though; I haven't investigated yet.

Tom

Re: Yet another sparc-sun-solaris2.8 bootstrap failure

[Gerald Pfeifer]

On 9 Apr 2002, Tom Tromey wrote:
> Unfortunately I don't think I have access to any such machine.

You have (if you want). ;-)

> So I don't know what I can do to help.

I filed a high-priority PR for this problem this afternoon, and davem
already did a nice analysis:

  http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=6237

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Re: Yet another sparc-sun-solaris2.8 bootstrap failure

[David S. Miller]

   From: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
   Date: Wed, 10 Apr 2002 00:01:44 +0200 (CEST)
   
   I filed a high-priority PR for this problem this afternoon, and davem
   already did a nice analysis:
   
     http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=6237

My analysis is wrong btw, I though solaris-2.7+ targets would build
64-bit binaries by default, and this is not true.

Richard helped me figure this out and the issue really is that
it is bombing out trying to do 64-bit multilib of the target
libraries when the kernel can't run 64-bit binaries.
That is what fails, configuring the multilib when running on
a 32-bit only kernel.

Re: Yet another sparc-sun-solaris2.8 bootstrap failure

[Richard Henderson]

On Tue, Apr 09, 2002 at 03:22:13PM -0700, David S. Miller wrote:
> it is bombing out trying to do 64-bit multilib of the target
> libraries when the kernel can't run 64-bit binaries.

Incidentally, one can work around this with --disable-multilibs.


r~