On Fri, 3 Mar 2000 yodaiken@fsmlabs.com wrote:
> I have not looked at the APIC docs recently, but doesn't this block
> irqs [...]
yes. While this obviously bad for RTLinux (and needs to be wrapped there),
it's good for just about anything else. It makes IRQs more 'atomic'
(nothing interrupts the fast interrupt), and filters out broken IRQ
handlers which somehow magically depend on IRQs getting delivered. Also,
the 'focus CPU' feature lets subsequent IRQs getting sent to the same CPU
if they occur while the IRQ is still being handled on one CPU.
In the long run IRQ handlers should become simple, fast, atomic functions,
just like most networking drivers look like currently. They set up soft
IRQs (see softnet's architecture), which then do the more nondeterminate
work.
-- mingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Mar 07 2000 - 21:00:14 EST