Messages in this thread |  | | Date | Sat, 22 Mar 2008 14:34:49 +0100 | From | Gabriel C <> | Subject | Re: 2.6.25-rc5-git6: Reported regressions from 2.6.24 |
| |
Thomas Gleixner wrote: > On Fri, 21 Mar 2008, Thomas Gleixner wrote: >>> | 1.78 us, TSC-warps:0 | 19.27 us, TOD-warps:0 | 19.37 us, CLOCK-warps:0 >> Ok. So the watchdog trigger is a false positive. >> >> Thinking more about it, it looks like Andi's change triggers some >> hidden bug in the combination of NO_HZ and add_timer_on(), where the >> CPU on which the timer is added is likely in a long idle sleep. I look >> into this tomorrow. > > Ok. Here is what's happening: > > CPU0 runs the watchdog timer and schedules it on CPU1. > > With NO_HZ enabled CPU1 is in a long idle sleep. At this point of the > boot process there is probably no timer pending on CPU1, which means > the idle sleep is infinite. > > Now some time later CPU1 gets woken by an interrupt/IPI and runs the > timer wheel. At this point the pm_timer which is the reference clock > has already wrapped around, so the watchdog thinks that there is a > huge time difference and marks the TSC unstable. > > Aside of that watchdog issue this also affects the other users of > add_timer_on(): e.g. queue_delayed_work_on(). > > Can you please apply the patch below and verify it with Andi's > watchdog patch applied ?
Did that , git head , Andi's + your patch but TSC is still marked unstable.
> > Thanks, > > tglx >
Gabriel
|  |