Messages in this thread | | | Date | Mon, 17 Aug 2009 10:49:01 -0400 (EDT) | From | Steven Rostedt <> | Subject | Re: 2.6.31-rc5 regression: x86 MCE malfunction on Thinkpad T42p |
| |
On Tue, 11 Aug 2009, Johannes Stezenbach wrote:
> On Tue, Aug 11, 2009 at 12:13:07AM +0200, Johannes Stezenbach wrote: > > On Mon, Aug 10, 2009 at 11:31:33PM +0200, Ingo Molnar wrote: > > > * Johannes Stezenbach <js@sig21.net> wrote: > > > > > > > > Could the warning be caused by the cpufreq ondemand governor? ISTR > > > > that one should switch to the performance governor before doing > > > > any profiling, but I forgot for this test. > > > > > > there might be a connection - it could in theory cause sched_clock() > > > transients and confuse the ring-buffer time-stamping. > > > > I'll try tomorrow after a fresh boot if the warning also appears > > with the performance governor. > > Nope, cpufreq was not the culprit. The "Delta way too big!" > warning in rb_reserve_next_event() happens after suspend-to-RAM.
Yeah, the ring buffer expects the time delta of two different events on the same page to be less than 2^59 nanosecs apart. But something must have screwed things up, since 2^59 nanosecs is 6671 days (18 years).
Perhaps coming out of suspend to ram changes the timestamp? Does it reset the result of sched_clock?
-- Steve
| |