> At first, if nic_irq is not used by a RT_irq_handler, you do not need
> rt_free_global.
OK; is rt_free_global() potentially dangerous (in case there's nothing
to free)?
> Then after rt_request_global_irq , you have to do rt_enable_irq(nic_irq)
> to enble it.
Should I enable the interrupt just once, immediately after the call to
rt_request_global_irq(), or in the handler, immediately before calling
rt_pend_linux_irq()?
Here is an excerpt of a reply I just got from Paolo Mantegazza (I hope
it's not a problem to quote it). I am not sure about edge-enabled
vs. level-enabled interrupts (how do I check this?), and, again,
whether I should enable/unmask an interrupt immediately, every time
before pending.
> When you enter you handler you have interrupt disabled. If it is
> anedge triggered one it is already acknowledged. If it is level is
> masked and you must car of unmasking it before pending. The pending
> must be done within your interrupt handler when it is needed. I
> repeat look at fastick and timer examples.
Thanks a lot to both of you,
Aleks
--- [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/
