Ralf Corsepius ralf.corsepius at rtems.org
Sat Nov 7 02:55:59 UTC 2009

On 11/06/2009 06:43 PM, Leontie Eugen wrote:
> Sorry for being vague. I am doing a RTEMS port to sparc64,
Wow, a 64bit target ... the "beyond bleeding edge"!

FWIW: I am not aware about anybody having successfully ported RTEMS to a 
64bit target before, nor am I sure if RTEMS is "64bit ready".
(I know, RTEMS is not 16bit clean, but I am working on this)

> so I have a new processor model, new bsp and the rtems is the latest CVS from a couple weeks back.
Upgrade to the very latest CVS-HEAD. A lot of "bit-wise" cleanup has 
gone into CVS-HEAD recently (and more will likely follow).

> I found this conversation online : http://old.nabble.com/-patch,rtems--Don't-build-__assert_func-for-RTEMS-td25554127.html
> and it is exactly what I need. Can you explain how should I configure NEWLIB for that, or how would I activate HAVE___ASSERT_FUNC ?

Try the very latest newlib*rtems4.10*.diff from 
contrib/crossrpms/patches in rtems CVS-HEAD and rebuild your GCC/newlib 
one-tree style with it.

It contains the very latest adaptation of the patch to the RTEMS-4.10 
toolchains you mention above.


More information about the users mailing list