This is the mail archive of the
elfutils-devel@sourceware.org
mailing list for the elfutils project.
Re: [patch] Resolve ppc64 func descriptors as .func (via .opd)
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: elfutils-devel at lists dot fedorahosted dot org
- Date: Tue, 11 Dec 2012 22:11:11 +0100
- Subject: Re: [patch] Resolve ppc64 func descriptors as .func (via .opd)
On Tue, 11 Dec 2012 21:59:42 +0100, Mark Wielaard wrote:
> But you haven't convinced me that we need to mangle the symbol
> or the return string. Please just keep it simple.
Yes, this is what I try to do. BFD is proven to work and also one stays safe
when user interchangeably uses binutils<->elfutils input/output.
> but mangling the symbol name/return string shouldn't be part of it.
This would break the ppc64 ABI, I cannot do that as I have some professional
responsibility.
What is the next step?
(1) Submit an approval request to IBM as I have proposed in:
https://lists.fedorahosted.org/pipermail/elfutils-devel/2012-December/002824.html
Not sure how do you see that but IMNSHO they will say ".funcname".
(2) Discuss how to map name -> addr for "funcname"
(when there is no ".funcname") without writing such implementation.
(3) Discuss how to map name -> addr for "funcname"
(when there is no ".funcname") after I write the implementation.
If we return "funcname" I really do not understand how it can work.
"funcname" -> addr needs to report the function descriptor.
But then we return name which does not map back to the address for which we
reported the name.
Thanks,
Jan