Following my previous e-mail with respect to detecting a 60-200 kHz TTL signal wich 
generates an interrupt throug an ISA based PIO-24 card: 

The response-time of one interrupt is 6-7 mirosecond  meaning a frequency of 150 kHz. 
However detecting a frequency above 25 kHz goes with a deviation. Probably caused bij 
the the timer interrupts. Studying the code and doing a few timing experiments leads 
me to these conclusion:

A real-time Timerinterrupt is invoked every 5 millisecond and lasts 9-13 microseconds. 
A part of the RT-timer ISR can be interrupted by other real-time interrupts. The Linux 
Timer interrupt is invoked every 10 ms (this is a mean, it fluctaties between 6 and 
400 ms) and lasts 30 -70 micoseconds.

Does anyone know whether linux interrupts (i.e. keyboard, hard-disk-IDE0, PS2/mouse, 
ehternet and above all the timer interrupts) always can be interrupted by real-time 
interrupts? Normally the linux timer ISR is started with the  SA-INTERRUPT flag which 
means NO interrupts.

- Has anyone experience with enabling interrupts in a real-time ISR (normally all 
interrupts are disabled). I fear deadlocks with the real-time Timer-interrupts or very 
long interrupt response times. 

Greetings,

Erwin Triepels



-- [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/

Reply via email to