preemption of irq handler ??

[Philippe Gerum]

Philippe Gerum wrote:
 > 
 > As far as RTAI/x86 is concerned, your ISR won't be preempted by
 > another interrupt whatever its source is _unless_ your handler
 > explicitely re-enables the x86 interrupt flag (i.e. sti).

I forgot to mention that causing a task switch in your handler will
make the resumed task's interrupt mask reasserted. Since a task
usually runs with interrupts enabled, your source will be able to
interrupt the CPU right after it is switched in by the handler.

Philippe.