This is the mail archive of the
crossgcc@sourceware.org
mailing list for the crossgcc project.
See the CrossGCC FAQ for lots
more information.
crosstool with MPC5200B powerpc processor
- From: "Andy Gibbs" <andyg1001 at hotmail dot co dot uk>
- To: <crossgcc at sourceware dot org>
- Date: Thu, 1 Jul 2010 12:18:19 +0200
- Subject: crosstool with MPC5200B powerpc processor
Hi Yann,
A quick email to thank you for all your work on crosstool. Using 1.7.1 I've
got a very stable cross-compiler for my system based around a Freescale
MPC5200B. For interest these are the pertinent details, in case you want to
add it to your list of successful builds on your website:
Target architecture : powerpc, 32-bit big-endian
CPU: 603e
linux: 2.6.29.1
binutils: 2.20
gcc: 4.4.4, with c++
libc: eglibc 2.10, with ports addon and nptl
extra target CFLAGS: -U_FORTIFY_SOURCE
companion libs: gmp 4.3.2, mpfr 3.0.0, ppl 0.10.2, cloog 0.15.9, mpc 0.8.2
additional: native and cross gdb 6.8 built, plus gcc test suite
The important thing is not to use 'sstrip' since it is very broken -- i.e.
it totally scrambles libraries! I also couldn't get glibc or eglibc to
compile without disabling the 'FORTIFY_SOURCE' flag. Everything else can be
set as the defaults.
Using the gcc test suite option worked mostly very well. The summaries were
as follows:
=== gcc Summary ===
# of expected passes 53409
# of unexpected failures 146
# of expected failures 193
# of unresolved testcases 76
# of unsupported tests 631
=== g++ Summary ===
# of expected passes 19467
# of unexpected failures 41
# of expected failures 157
# of unresolved testcases 42
# of unsupported tests 158
What was interesting and I haven't yet tracked down is that not all of the
tests that executed on the target were able to do so. Most worked fine, but
some (according to the detailed logs) were not found following the copy to
the target. I "found" them manually and could get them to execute from the
console, but they were in a totally different folder to where they should
have been. Therefore this may be a bug in some of the gcc test scripts.
But, needless to say, I'm very impressed how smooth an operation it has been
on the whole.
Many, many, many thanks!
Andy
--
For unsubscribe information see http://sourceware.org/lists.html#faq