On Fri, 16 Sep 2011, Marc Zyngier wrote:

On 15/09/11 23:49, Thomas Gleixner wrote:

quoted

quoted

+
+	raw_spin_unlock_irqrestore(&desc->lock, flags);
+
+	unregister_handler_proc(irq, action);
+
+	/* Make sure it's not being used on another CPU: */
+	synchronize_irq(irq);

That's not helping w/o making synchronize_irq() aware of the percpu
stuff. Also there is the question whether we need the ability to
remove such interrupts in the first place. The target users are low
level arch interrupts not some random device drivers.

Again, there is no need for this at the moment (the timer code runs
running forever), and this is only there for completeness.

I'll see if I can come up with a synchronize_percpu_irq() without adding
too much bloat to irqdesc.

You'd need a PROGRESS flag per cpu, which is overkill. What you can do
is to check whether the percpu enabled cpumask is completely empty and
return with a WARN when not.

Thanks,

	tglx