RTEMS 4.10 Integration

[Yann E. MORIN]

Sebastion, All,

On Monday 25 October 2010 16:00:25 Sebastien Van Cauwenberghe wrote:
> I'd like to add RTEMS 4.10 kernel to CTNG, to build a custom toolchain for it.

Hey! That would be great!

Addind a new kernel should be quite easy, just look at how Linux and
mingw32 are doing, and try to imitate. There is also some documentation
in the file docs/8 - Internals.txt which is not really complete nor
absolutely up-to-date, but is a very good starting point ( and you can
also improve that file if you see fit ;-) )

Basicaly, what a kernel needs to do is publish:

- CT_DoKernelTupleValues
    a fundtion that computes the kernel-part of the tuple;

- do_kernel_get, do_kernel_extract
    a function for how to get the sources, and a function that extracts
    the sources (easy from tarballs);

- do_kernel_headers
    a function that installs the headers, if any;

- a config file that has options to set the version of the kernel.

And that's all, crostool-NG will detect the new kernel automatically, and
add it to the menuconfig.

> Is there a way to make conditional patching of GCC, Libc and so on, only when
> RTEMS is selected ?

No, there is no provision for that. And I would prefer that there not be such
direct possibility.

Well, in fact, there is a possibility, but it's quite ugly.

Paths and misc options  --->
  Patches origin  ---> Bundled, then local
  Local patch directory  ---> "${CT_LIB_DIR}/patches/rtems"

And put all your RTEMS-related patches in there, such as:
  patches/rtems
  patches/rtems/gcc/4.4.5/100-foo-bar.patch
  patches/rtems/gcc/4.4.5/120-buz-baz.patch
  patches/rtems/binutils/2.20.1/100-alpha.patch
and so on...

But really, I would prefer that it be avoided. From what I could see after
a quick glance at the gcc patch from RTEMS, the patch touches almost only
RTEMS-related files, touches configure-stuf so it recognises RTEMS in a
tuple, and a few other things. So I would hope this patch is generic enough
to be applied unconditionnally... But I only looked at it for 5 mins, we
have to clear that...

Regards,
Yann E. MORIN.

PS. I will be unavailable until Sunday, so don't expect any answer from my
    side before that...
YEM.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'



--
For unsubscribe information see http://sourceware.org/lists.html#faq