Messages in this thread | | | Date | Thu, 19 Aug 2010 14:39:26 +0200 | From | Robert Richter <> | Subject | Re: [PATCH -v3] perf, x86: try to handle unknown nmis with running perfctrs |
| |
On 19.08.10 06:45:53, Peter Zijlstra wrote: > I queued it with that part changed to: > > + this_nmi = percpu_read(irq_stat.__nmi_count); > + if ((handled > 1) || > + /* the next nmi could be a back-to-back nmi */ > + ((__get_cpu_var(nmi).marked == this_nmi) && > + (__get_cpu_var(nmi).handled > 1))) { > + /* > + * We could have two subsequent back-to-back nmis: The > + * first handles more than one counter, the 2nd > + * handles only one counter and the 3rd handles no > + * counter. > + * > + * This is the 2nd nmi because the previous was > + * handling more than one counter. We will mark the > + * next (3rd) and then drop it if unhandled. > + */ > + __get_cpu_var(nmi).marked = this_nmi + 1; > + __get_cpu_var(nmi).handled = handled; > + } > > return NOTIFY_STOP; > }
I am fine with this. Thanks Peter.
-Robert
-- Advanced Micro Devices, Inc. Operating System Research Center
| |