Re: rt20 scheduling latency testcase and failure data

[Darren Hart]

On Saturday 13 May 2006 11:21, Lee Revell wrote:
> On Sat, 2006-05-13 at 11:06 -0700, Darren Hart wrote:
> >      1 [softirq-timer/0]
>
> What happens if you set the softirq-timer threads to 99?
>

After setting all 4 softirq-timer threads to prio 99 I seemed to get only 2 
failures in 100 runs. #51 slept too long (10ms too long!), the latency 
appeared after the sleep in #90 (nearly 483ms worth).  Those latencies seem 
huge to me.

ITERATIONS 0-50 Passed

ITERATION 51
-------------------------------
Scheduling Latency
-------------------------------

Running 10000 iterations with a period of 5 ms
Expected running time: 50 s

ITERATION DELAY(US) MAX_DELAY(US) FAILURES
--------- --------- ------------- --------
     7000     10197         10197        1

PERIOD MISSED!
     scheduled delta:     4079 us
        actual delta:    14263 us
             latency:    10183 us
---------------------------------------
      previous start: 35010197 us
                 now: 35011117 us
     scheduled start: 35005000 us
next scheduled start is in the past!


Start Latency:  114 us: FAIL
Min Latency:      9 us: PASS
Avg Latency:      7 us: PASS
Max Latency:   10197 us: FAIL
Failed Iterations: 1

ITERATIONS 52-89 Passed

ITERATION 90
-------------------------------
Scheduling Latency
-------------------------------

Running 10000 iterations with a period of 5 ms
Expected running time: 50 s

ITERATION DELAY(US) MAX_DELAY(US) FAILURES
--------- --------- ------------- --------
     2747         9            20        0

PERIOD MISSED!
     scheduled delta:     4072 us
        actual delta:     4079 us
             latency:        6 us
---------------------------------------
      previous start: 13735009 us
                 now: 14218183 us
     scheduled start: 13740000 us
next scheduled start is in the past!


Start Latency:  112 us: FAIL
Min Latency:      8 us: PASS
Avg Latency:      2 us: PASS
Max Latency:     20 us: PASS
Failed Iterations: 0

ITERATIONS 91-99 Passed

Thanks,

-- 
Darren Hart
IBM Linux Technology Center
Realtime Linux Team
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/