On Wed, 2005-12-14 at 09:04 -0500, Steven Rostedt wrote:
> On Wed, 2005-12-14 at 08:48 -0500, Steven Rostedt wrote:
> 
> > And going into gdb, I get:
> > 
> > (gdb) li *0xc0136b98
> > 0xc0136b98 is in hrtimer_cancel (kernel/hrtimer.c:671).
> > 666     int hrtimer_cancel(struct hrtimer *timer)
> > 667     {
> > 668             for (;;) {
> > 669                     int ret = hrtimer_try_to_cancel(timer);
> > 670
> > 671                     if (ret >= 0)
> > 672                             return ret;
> > 673             }
> > 674     }
> > 675
> > 
> > So it may not really be locked, and if I waited a couple of hours, it
> > might actually finish (the test usually takes a couple of minutes to
> > run, and I let it run here for about 20 minutes).

Ingo,

Running my jitter test shows a slow down on the responsiveness of
nanosleep.

In 2.6.14-rt2 (the only 2.6.14-rtx laying around still) I get:

reported res = 0.000001000
starting calibrate
finished calibrate: 736.0509MHz 736050890

  Requested time         Max             error
  --------------         ---             -----
  0.000001000           0.000032218     0.000031218
  0.000010000           0.000033196     0.000023196
  0.000100000           0.000124324     0.000024324
  0.001000000           0.001027797     0.000027797
  0.010000000           0.010025151     0.000025151
  0.100000000           0.100020185     0.000020185
  1.000000000           1.000024266     0.000024266
  1.000001000           1.000028153     0.000027153
  1.000010000           1.000036685     0.000026685
  1.000100000           1.000122415     0.000022415
  1.001000000           1.001023463     0.000023463
  1.010000000           1.010021942     0.000021942
  1.100000000           1.100025656     0.000025656
  2.000000000           2.000028559     0.000028559

errors off by 20 - 32 usecs

On 2.6.15-rc5-rt1 (with my patches) on the same machine I get:

reported res = 0.000001000
starting calibrate
finished calibrate: 736.0510MHz 736051025

  Requested time         Max             error
  --------------         ---             -----
  0.000001000           0.000089476     0.000088476
  0.000010000           0.000045054     0.000035054
  0.000100000           0.000135034     0.000035034
  0.001000000           0.001041864     0.000041864
  0.010000000           0.010042513     0.000042513
  0.100000000           0.100031702     0.000031702
  1.000000000           1.000041268     0.000041268
  1.000001000           1.000044342     0.000043342
  1.000010000           1.000055852     0.000045852
  1.000100000           1.000157908     0.000057908
  1.001000000           1.001055853     0.000055853
  1.010000000           1.010036436     0.000036436
  1.100000000           1.100040173     0.000040173
  2.000000000           2.000060079     0.000060079


errors off by 35 - 89 usecs

Is this normal??

Attached is the jitter test that I ran (also the test that will crash
the system without my patches).

-- Steve