On Sunday 01 March 2009, Michael Buesch wrote:
> My problem is that the kernel oops message is not that useful, as the crash
> happens inside of rtai. So it doesn't contain much information to locate
> the actual line of code that triggers the exception. (The line that's
> miscompiled).

Problem confirmed with 2.6.28 & RTAI-3.7(devel) and emc2 compiled with 
gcc-4.3. Compiling emc2 kernel modules with gcc-4.2 (and everything else with 
gcc-4.3) does not cause a crash.

Using EMC code (not to be confused with emc2) and gcc-4.3, no crash, kernel 
lock ups or other abnormal kernel messages.

> So it would be nice if somebody knows from the top of the head that "in arc
> moves we use this and that special gcc feature, you might like to look at
> it...". :)

As far as I am aware, ARM is the only architecture known to have issues with 
inline code and gcc-4.3. Using gcc-4.3 with i386 & x86_64 should not cause 
problems. What is more likely is one of the extended registers is being 
trashed as soon as a particular set of floating point instructions are 
executed in kernel space.

The answer is to ensure the affected registers are saved/restored at the 
appropriate times, or move floating point math out of kernel space, or ensure 
the compiler does not use the registers in question - Using the appropriate 
compiler flags resolves the issue with gcc-4.3/2.6.28.7/RTAI/emc2.


------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to