Am 18.10.2010 16:06, Steven Rostedt wrote: > On Mon, 2010-10-18 at 14:33 +0200, Felix Jirka wrote: >> Hello, >> >> I'm trying to build a kernel-image based on 2.6.32.20 with RTAI-patch >> (3.8.1) and enabled tracing. >> >> During the first tests I noticed that when I try to enable the >> function trace via debugfs the machine reboots. >> Then I enabled FTRACE_STARTUP_TEST and now the kernel freezes during >> the boot process while doing the function trace test and it seems to >> be completely unresponsive (sysrq-key not working...). >> >> I can also observe this behavior with a vanilla 2.6.32.20 kernel >> running in a VMWARE-Machine. > > It is most likely that the virtual machine code is being traced and > causing an infinite recursion. The function tracer has basic recursion > protection, but there's some code that it can't protect against. If any > of the following is traced, then it will cause a recursion crash: > > Atomic incs, disabling interrupts, smp_processor_id(). > > >> >> Is this caused by some wrong configuration settings on my side or is >> there a bug involved? >> >> How can I collect more information about the problem? > > Unfortunately, you will need the developers of vmware or RTAI to fix > this for you. To close this off-topic issue for this list: Actually, none of both is involved. It's the Adeos/I-pipe kernel patch that RTAI uses (with some modifications). I've been hammering on those bits for a 2.6.35 over the last days. Works fine here with both ftrace and perf, but it's still experimental. Announcements on Adeos&Xenomai lists will follow later today. Jan PS: Adeos has it's own function tracer that should work with your kernel. -- Siemens AG, Corporate Technology, CT T DE IT 1 Corporate Competence Center Embedded Linux -- To unsubscribe from this list: send the line "unsubscribe linux-trace-users" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
|