Re: MIPS multi-got link support

[Alexandre Oliva <aoliva at redhat dot com>]

On Jan 16, 2003, Hans-Peter Nilsson <hp@bitrange.com> wrote:

> Assuming the "reasonable size" just refers to the actual code
> size or number of functions: You don't mention why generating
> the test-case during test-time won't work to trigger the code in
> question.  For an example, see ld-elf/sec64k.exp.
> Just helping. :-)

Thanks.  I remembered there was a test like that, and I even
considered using it as inspiration.  The main problem is that we have
hashtables that use pointers to bfds as (part of) keys, and then the
process of bfd->got mapping, as well as the ordering of relocations
and got entries, ends up dependent on the exact addresses of bfds.  I
didn't think it was a problem until I got to the point of figuring out
I should create testcases.  And then, I realized I was definitely out
of luck.

> It seems too important to get silently broken every once in a while.

Agreed, especially given how seldom it's exercised in real life, but I
don't see an easy way out.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                 aoliva@{redhat.com, gcc.gnu.org}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist                Professional serial bug killer