This is the mail archive of the binutils@sources.redhat.com mailing list for the binutils 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: Can't build libjava in gcc cvs trunk 20050227 on sparc-linux: ar cq libgcj0_convenience.a ... Bus error: in bfd_hash_lookup () from /usr/lib/libbfd-2.15.92.0.2.so


On Tue, 1 Mar 2005 09:11:41 +1030, Alan Modra <amodra@bigpond.net.au> wrote:
> On Mon, Feb 28, 2005 at 02:37:14PM +0100, Christian Joensson wrote:
> > Program received signal SIGBUS, Bus error.
> > bfd_hash_lookup (table=0x611e1c, string=0x33117f1 ".debug_abbrev", create=1,
> >     copy=0) at /usr/local/src/trunk/src/bfd/hash.c:394
> 
> Please compile binutils with CFLAGS="-g -O" and see if you can figure
> out why bfd_hash_lookup is dying.

hmm, I think I will have to bail out of this now, debugging seems to
take "forever" and it might even be the case that it does not give me
anything.

Speculating about the resons for this problem, I would say it's either
my developmnet system and its development kernel that does memory
trashing in a way that affercs binutils ar like this, or it is a gcc
miscompilation (which seems to me far off but still possible) or it is
a binutils "bug".

For anyone interested in digging into this, I have put a tarred
achhive of the .o files togheter that you can run your favourite ar
capable of reading sparc-linux .o files and producing sparc-linux
archives together with a text file that describes my setup, the
verisons used, my test run of ar and a simple gdb back trace of the
Bus error I get.

It's available, for some time..., at 

ftp://ftp.j-son.org/pub/aurora/ar-problem-corona/

-- 
Cheers,

/ChJ


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