On Wed, 26 Feb 2020 11:27:58 +0100 Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote: > On Mon, Feb 24, 2020 at 05:02:31PM -0500, Steven Rostedt wrote: > > > The other is for the hwlat detector that measures the time it was in an > > NMI, as NMIs appear as a hardware latency too. > > Yeah,.. I hate that one. But I ended up with this patch. > > And yes, I know some of those notrace annotations are strictly > unnessecary due to Makefile crap, but having them is _SO_ much easier. > > --- > Subject: x86,tracing: Robustify ftrace_nmi_enter() > From: Peter Zijlstra <peterz@xxxxxxxxxxxxx> > Date: Mon Feb 24 23:40:29 CET 2020 > > ftrace_nmi_enter() > trace_hwlat_callback() > trace_clock_local() > sched_clock() > paravirt_sched_clock() > native_sched_clock() > > All must not be traced or kprobed, it will be called from do_debug() > before the kprobe handler. > Acked-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx> -- Steve > Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx> > --- > arch/x86/include/asm/paravirt.h | 2 +- > arch/x86/kernel/tsc.c | 7 +++++-- > include/linux/ftrace_irq.h | 4 ++-- > kernel/trace/trace_clock.c | 2 ++ > kernel/trace/trace_hwlat.c | 4 +++- > 5 files changed, 13 insertions(+), 6 deletions(-) >
