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: [PATCH] Fix objdump -d on ppc64 without dot syms (take 2)


On Sat, Aug 14, 2004 at 05:47:43PM +0200, Jakub Jelinek wrote:
> The previous patch unfortunately did not work for ET_REL objects,
> so here is new extended version of the patch which touches even objdump.c
> and various files in bfd/.

Please apply.

> BTW: with this and also
> http://sources.redhat.com/ml/binutils/2004-08/msg00030.html
> patch in the tree I'm getting a bunch of failres with gcc 3.4 on ppc64.
> With unpatched gcc 3.4.1 I'm getting:
> FAIL: visibility (hidden_undef_def) (non PIC)
> FAIL: visibility (hidden_undef_def) (non PIC, load offset)
> FAIL: visibility (hidden_undef_def)
> FAIL: visibility (hidden_undef_def) (PIC main, non PIC so)
> FAIL: visibility (hidden_undef_def) (PIC main)
> (see the above URL for details), when using gcc 3.4.1 with your dot syms
> patch these failures go away, but there are 4 new ones:

The fact that they go away indicates I have a problem in the linker.
I'll look into it today.

> FAIL: NOCROSSREFS 1

This one is a testsuite problem.  Getting rid of the .opd section spec
in cross1.t lets the test pass with both old and new gcc.

> FAIL: selective1
> FAIL: selective2

These are real linker problems too.

> FAIL: S-records

This one is also a linker problem, but I'm inclined to ignore it for the
time being.  The generic linker doesn't understand how to resolve
"bl foo" to the function code entry point.

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre


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