Hi Peter and Ingo, I have here two fixes for 4.1, both of which are basically results of not testing code properly with the NMI watchdog disabled, so one lesson learned here.
One bug results in trace data loss and unknown NMI warnings and the other one is a NULL pointer dereference. The latter might need to be revisited in the future as we're now handling PT PMIs from intel_pmu_handle_irq() and we might want to change that. On the other hand, PT and x86_pmu share the PMI status register, so from that angle it kind of makes sense. But until such time as we decide to sort it out, the proposed fix should do nicely. Alexander Shishkin (2): perf/x86/intel/bts: Fix DS area sharing with x86_pmu events perf/x86/intel: Fix PMI handling for Intel PT arch/x86/kernel/cpu/perf_event.c | 70 +++++++++++++++++++++--------- arch/x86/kernel/cpu/perf_event.h | 4 ++ arch/x86/kernel/cpu/perf_event_intel_bts.c | 9 ++++ 3 files changed, 63 insertions(+), 20 deletions(-) -- 2.1.4 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
