Please tell me what you think about this new version. I still can not fallback to save the ip for archs that don't implement it as it can't be done in a generic way, but more work has been made to implement this support more easily. No more assembly in this new version, it also uses functions instead of macros and it implements the callers level skipping. Then instead of starting the callchain from perf/ftrace event callbacks, we start on the real event source. Thanks. Frederic Weisbecker (2): perf: Introduce new perf_save_regs() for hot regs snapshot perf: Take a hot regs snapshot for trace events arch/x86/kernel/cpu/perf_event.c | 12 ++++++++++ arch/x86/kernel/dumpstack.h | 15 ++++++++++++ include/linux/ftrace_event.h | 7 ++++- include/linux/perf_event.h | 42 +++++++++++++++++++++++++++++++++++- include/trace/ftrace.h | 6 ++++- kernel/perf_event.c | 14 ++++++------ kernel/trace/trace_event_profile.c | 3 +- kernel/trace/trace_kprobe.c | 5 ++- kernel/trace/trace_syscalls.c | 4 +- 9 files changed, 92 insertions(+), 16 deletions(-) -- To unsubscribe from this list: send the line "unsubscribe linux-arch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html