[Bug 1669] [4.5/4.6 Regression] libgcc does not include t-ppccomm on rtems

[bugzilla-daemon at rtems.org]

https://www.rtems.org/bugzilla/show_bug.cgi?id=1669

--- Comment #1 from Ralf Corsepius <ralf.corsepius at rtems.org> 2010-08-15 02:00:16 CDT ---
(In reply to comment #0)
> Just a PR to track this on our side.
> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44793
> 
> What do you think of my patch?

This patch doesn't seem right to me.

AFAIU, what you are trying to achieve is to supply *crt* from inside of libgcc/
instead of gcc/ by adding libgcc/config/rs6000/t-ppccom instead of
gcc/config/rs6000/config/t-ppccom.

As a side-effect of your changes you however do not only replace *crt* but also
remove many other settings/variables etc. contained in
gcc/config/rs6000/config/t-ppccom, e.g. tramp.asm, several MULTILIB_* settings
and other details.

I don't fully understand what all these are used for, except that I am almost
100% sure that removing the MULTILIB_*-settings is wrong.

That said, I currently think the right approach would be to split out all of
the make-rules which currently are contained in both libgcc/*/t-ppccomm and in
gcc/*/t-ppccomm from gcc/*/t-ppccomm, or (alternatively) to make them
conditional in t-ppccomm (rsp. duplicate them to gcc/*/t-rtems)

What I also do not understand in this context, is how the linux configuration
(the only one which currently uses libgcc/*/t-ppccomm) works. It uses both
t-ppccomm's - This doesn't seem correct to me, also.

-- 
Configure bugmail: https://www.rtems.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.