Re: [patch] severe softirq handling performance bug, fix, 2.4.5

[Andrea Arcangeli]

On Sat, May 26, 2001 at 04:55:06PM -0700, David S. Miller wrote:
> And looking at the x86 code, I don't even understand how your fixes
> can make a difference, what about the do_softirq() call in
> arch/i386/kernel/irq.c:do_IRQ()???  That should be taking care of all
> these "error cases" right?

Yes, except when a softirq is marked running again under do_sofitrq (it
is mostly an issue when the machine is otherwise idle that means we
will waste cpu if we don't run the sofitrq immediatly, this problem was
noticed by Manfred last month, but that is just a special case of the
generic case of a softirq marked running again under do_softirq), and
all those cases are supposed to be taken care by ksoftirqd.

Andrea