Re: Failures with exelib.exp testcase (was Re: minutes 2010-08-19)

[Roland McGrath <roland at redhat dot com>]

> It seemed more general than an e_machine type check. But it looks like I
> was wrong about that. I thought I read somewhere ia64 also used them,
> but now cannot find the reference.

ia64 also has function descriptors, but the details of what a function
descriptor looks like are very machine-specific and I don't think it would
be wise to have anything about function descriptor support not be specific
to the machine.

> So, after linking, using program header good, using section headers bad.

Correct.

> I am still unsure what to do about that section number hack we have in
> translate.cxx dump_unwindsyms, aka PR10088. Although now that I think
> about it, all this confusion in that PR might just be because .ko files
> are just not really linked yet.

It's indeed the case that Linux kernel modules (.ko) are outside the normal
rules of ELF things.  They are ET_REL files, not the product of a "final
link" in normal terms.  But they share a few characteristics with final
links.  They are in fact the product of an intermediate link, neither a
normal .o nor a final link.

I don't really know what issue you are talking about there.  Note that the
"relocation index" from libdwfl is not the same thing as an ELF section
index.


Thanks,
Roland