Michael,
Thank you!
1. Did you notice the answer of Ivan? ----He solved the problem by logging
execution time in a RTFifo. I tried logging the execution time use a static
variable, and then check it as YOU did in the RTL cheduling accuracy
measuring example ( example/measurements/rt_process.c). How do you think
about it?
2. In rtl_time.h there is a function:
int rtl_setclockhandler (clockid_t h, clock_irq_handler_t fn);
Can I append a handler to the clock with the this function (or with
rtl_request_irq() ) after having set up the clock.
James...
>The simplest way seems to be to create a higher-priority thread with the
same
>period and an offset in start time. The thread would check if the original
>thread's computation is completed. It can also suspend or kill the task.
>
>Michael.
>
>Zaimin Zhong ([EMAIL PROTECTED]) wrote:
>> After having made a thread execute periodically with
>> pthread_make_periodic_np(3), how to find out whether the thread overruns
the
>> CPU within from this thread.
>>
>> To state my question more clearly: In my program, a RTL thread carries
out
>> periodic computation at fixed step. It is necessary for the thread to
find
>> out whether the last computation is finished when the given time slice
>> elapses.
>>
>> If I correctly understand, RTL thread will complete its computation even
if
>> the give time slice is too small to finish the job and give out logically
>> correct but NON-REAL-TIME results. Am I correct?
>>
>> James
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
---
For more information on Real-Time Linux see:
http://www.rtlinux.org/rtlinux/
