This is the mail archive of the
libc-hacker@sourceware.cygnus.com
mailing list for the glibc project.
Re: Problem-glibc-2.0.96
> Date: Thu, 10 Sep 1998 21:57:21 -0700 (PDT)
> From: Matt McLean <keys@yikes.com>
> On 10 Sep 1998, Ulrich Drepper wrote:
>
> > Matt McLean <keys@yikes.com> writes:
> >
> > > It's not. But I think it IS, however, related to your new start.S. Did you
> > > try compiling your changes before you commited them.. :-( This looks like
> > > bad assembly.. Can you take a look?
> >
> > This problem was introduced by the major changes to ld.so I made at
> > the weekend. Yesterday Benjamin Kosnik made the necessary changes and
> > they are already checked in.
>
> this may let ld.so run again, but it's still not back to normal. now,
> rpcgen dies with a SIGILL (illegal instruction) when trying to create
> nlm_prot.T. this doesn't happen with glibc 2.0.95.
I have just sent in a fix for this. It only affected processors with
split instruction/data caches (that is, not my machine). As a bonus,
the fix makes ld.so faster. [The problem was always there, since I
first wrote the code. I'm amazed it never caused trouble before.]
--
Geoffrey Keating <geoffk@ozemail.com.au>