[Xorp-hackers] Building XORP on ARM Architecture

[Michael Birkholz]

On 4/10/08, Pavlin Radoslavov <pavlin at icsi.berkeley.edu> wrote:
>
> > I was wondering if anybody has been successful in building XORP for the
> ARM
> > architecture and if so, if they could provide me with the steps they
> went
> > through to get it to compile and run.
>
> Section 3.8. Cross-compilation in xorp/BUILD_NOTES contains
> information for cross-compiling XORP. It has a subsection for XScale
> processor which is basically ARM.
>

Thanks a lot for your reply, Pavlin. It got me started in the right
direction. The XScale configuration seemed to need very little modification
to build XORP for the ARM. The build process went smoothly for me after I
correctly set the host and the environment variables to point to my
toolchain.


> > I was able to compile it, but when I run xorp_rtrmgr or xorpsh I see no
> > output at all. Can you point me in the right direction as far as what I
> > should look for in the build log and also whether XORP generates any log
> > files when it runs that could provide me with debug information to try
> to
> > solve this problem.
>
> Some time ago I did few basic tests with that architecture (I think
> the OS was MontaVista), and from what I remember I think the log
> output was working (I was running XORP in foreground so the messages
> were printed on the terminal).
>
> You should be aware that the XORP log facility tries to send the
> output to one of the following devices (in order of preference).
>
> /dev/stderr
> /dev/console
> /dev/stdout
>
> E.g., if /dev/stderr exists on the system and is writeable, then
> the it will be used as the default output stream.
> (See xorp/libxorp/xlog.c and search for /dev/stderr to see the
> actual code that selects the output stream).
>
> One thing to verify is whether the selected output stream actually
> works. E.g., I have seen problems with broken symbolic links on
> embedded Linux where /dev/stderr exists, but

 "echo FOO > /dev/stderr" doesn't actually print the output.
>

You were exactly right about a broken output stream. Instead of connecting
via a serial modem application I was using, I tried to telnet into the board
and I immediately saw error messages printed on the screen. Now that I can
see the errors being generated, I think I have a chance at figuring out how
to fix the remaining bugs, but I'll definitely be back in contact if I need
some more help.

Thanks!

Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ICSI.Berkeley.EDU/pipermail/xorp-hackers/attachments/20080413/5e208464/attachment.html