From: Richard Mortimer<richm@oldelvet.org.uk>
Date: Mon, 17 Jan 2011 23:34:03 +0000
I guess that points towards the binutils linker not doing the correct
thing.
Ok, it is in fact doing the correct thing.
I'm really surprised we never hit this before in all of these years
:-) I guess we've simply never hit this kind of expression in a module
before.
The issue is that modules aren't a "final link", it's really more like
an intermediate partial link.
So we do end up seeing the R_SPARC_LO10 + R_SPARC_13 sequences in the
final module object.
Therefore, we really should handle R_SPARC_13 in the sparc module loader.
Richard, I want you to get full credit for this since you did all of
the dirty work :-) Would you please cons up a formal patch with commit
message and signoff for this and I'll push it around?
Thanks a lot!