Mark Mitchell wrote:
Yes, the start-romvars.s is generates from start.S. Here's the version of it I get after the first arm-none-eabi-gcc -E call:42Bastian wrote:arm-none-eabi-gcc: Internal error: Segmentation fault (program as)It's probably start-romvars.s that's causing the fault, not start.S, right? If so, please post that file to the list.IMO the first instruction he posted generates start-romvars.s (gcc -E ...) from start.SGood point; sorry I missed that. However, I can't reproduce the crash on my system...
http://andrewpullin.org/tools/arm/start-romvars.s Here's an strace output of the second call: http://andrewpullin.org/tools/arm/strace.txtYou can't reproduce the segfault? I've tried in on an Ubuntu 32bit machine, running kernel 2.6.31-21-generic. I am using the latest 2010q1 release of the toolchain , although I tried it with the 2009q3 release, and it also segfaults. Maybe that indicates it's a specific compiler config, or something completely external?
Thanks, Andrew