Messages in this thread | | | Date | Tue, 17 Oct 2000 18:37:13 +0200 | From | Andrea Arcangeli <> | Subject | Re: PATCH 2.4.0.10.3: pc_keyb and q40_keyb cleanup |
| |
On Tue, Oct 17, 2000 at 05:43:34PM +0200, Jes Sorensen wrote: > Which shouldn't matter as the irq source should be disabled. In fact I > thought we were guaranteed not to be re-interrupted in a handler > unless one explicitly does __sti(), has this changed?
A single irq handler won't be re-interrupted, correct. Not even if you do __sti().
(in the edge triggered IO-APIC case, the irq source is not disabled to avoid missing events but the highlevel irq logic makes sure that the irq _handler_ won't be run if it was just in-progress somewhere in the system, even if in another CPU)
But the fact the irq handler is single threaded with respect to itself is irrelevant with the keyboard_interrupt case because that irq handler will be recalled by _two_ indipendent irq lines (irq 2 for the keyboard and irq 12 for the PS/2 Mouse).
Andrea - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
| |