Re: Different semantics for enable/disable IRQs

yodaiken@chelm.cs.nmt.edu
Tue, 31 Mar 1998 09:30:15 -0700

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: Alan Cox: "Re: 2.1.pre92-1: SC_ENABLE_IP"
Previous message: Finn Arne Gangstad: "Compilation fix for capabilities in 2.1.92"

On Tue, Mar 31, 1998 at 08:04:13AM -0800, Linus Torvalds wrote:
>
>
> On Mon, 30 Mar 1998, Oleg Subbotin wrote:
> >
> > Why is there a different semantics for enable/disable irq (in the case
> > of SMP) for the old 8259A interrupt controller and for IOAPIC?.
>
> Broken implementation reasons.

So when we write emulate_enable_irq, we will ignore the more complex new
version and assume that "enable_irq" means enable irq and
not, (--disablecount[irq] == 0 ? really enable : nop )

Ok?

> However, the fact is that it is a bug for the caller to not serialize the
> enable/disable calls correctly anyway, so any code that shows this
> implementation difference is to be considered broken anyway..
>
> Linus

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu

Next message: Alan Cox: "Re: 2.1.pre92-1: SC_ENABLE_IP"
Previous message: Finn Arne Gangstad: "Compilation fix for capabilities in 2.1.92"