This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: libjava build failure sol2.6 asm problem?


On Dec 19, 2001, Andreas Tobler <toa@pop.agri.ch> wrote:

> Alexandre Oliva wrote:

>> 
>> It's probably not the same problem as the one mentioned in the
>> URL above.  I tried a build a while ago using GNU as and ran into the
>> same problem.  It turned out that the label referenced in the
>> expression was indeed not present in the assembly code produced by
>> GCC.  It had been optimized away.  We do have a bug, indeed, but I
>> didn't have time to track it down.

> After a long search for the patch (only for contract customers) I got it
> and installed it.
> I made a make clean in the libjava dir and built it again. Same error.

Yup.  In this case, the bug is in GCC, not in Sun's assembler.

> Do you mean this mentioned label?

> sethi   %hi(_GLOBAL_OFFSET_TABLE_-(.LL306-.)), %o0

> The LL306?

Yup.

> It's just for my understanding. In the .s file I only see two
> occurences of this LL306. Both in the same way like above and not
> like others.

Right.  That's the problem.  The label should have been defined in the
file, with a line such as:

.LL306:

But this label was optimized away while there still were references to
it.  That's the GCC bug at hand.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]