----- Original Message -----
From: Richard B. Johnson <email@example.com>
To: Bradley D. LaRonde <firstname.lastname@example.org>
Cc: <email@example.com>; <firstname.lastname@example.org>
Sent: Tuesday, September 28, 1999 1:33 PM
Subject: Re: SA_INTERRUPT
> On Tue, 28 Sep 1999, Bradley D. LaRonde wrote:
> > What is the current wisdom on SA_INTERRUPT?
> It is unfortunate that the same #define is used for signals and
> for kernel interrupts. They are not related. As you noticed,i
> SA_INTERRUPT with respect to signals is a no-op since it doesn't change
> the nature of signals. However, with respect to IRQ handling within the
> kernel, it does enable interrupts before your possible bottom-half ISR
> is checked to see if it exists and executed.
> See line 726 (about) in ../arch/i386/kernel/irq.c
It looks like it enables interrupts even before the top-half handler is run.
Still, though, I would like to know when/why one should/shouldn't use this