Hi, can anybody answer the $Subject question? Consider a following situation. An interrupt on some IRQ occurs and the RTLinux executive has a RT handler for this IRQ, installed with rtl_request_irq() in init_module(). RTLinux, obviously, calls that handler. Then, while the handler is being executed, an interrupt occurs on another IRQ line, say, from keyboard or disk or network adapter (and there is no RT ISR for this IRQ). RTLinux executive gets the control and... Will RTLinux call Linux ISR immediately (of course, under assumption, that Linux has interrupts `enabled` and there are _NO_ real-time tasks/threads at all, except the RT ISR, executed at the moment when this second interrupt occured)? In other words, can the RT ISR be interrupted by Linux ISR (in the assumption, that there are no RT threads)? I am asking because I have a suspection that the answer to the question is YES. I measured the time that my ISR takes to complete and everything is OK while there is no much activity in the system (I am working on the SuperMicro 370DL with dual PIII). When the machine is loaded with disk i/o, network i/o, graphical i/o (X), this time sometimes become too long, compared with `normal`: `Normal` is 12-20 microsecs, and sometimes I got ISR execution times up to several milliseconds. Whereas while playing with RT-_threads_, I did not get such significant delays in periodic tasks, independently of how heavily machine is loaded. Regards, Eugene Zhiganov. -- [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/
