On Tue, 30 Nov 2010, Jeremy Fitzhardinge wrote:
> On 11/30/2010 11:07 AM, Christoph Lameter wrote:
> > static irqreturn_t xen_timer_interrupt(int irq, void *dev_id)
> > {
> > - struct clock_event_device *evt = &__get_cpu_var(xen_clock_events);
> > irqreturn_t ret;
> >
> > ret = IRQ_NONE;
> > - if (evt->event_handler) {
> > - evt->event_handler(evt);
> > + if (__this_cpu_read(xen_clock_events.event_handler)) {
> > + __this_cpu_read(xen_clock_events.event_handler)(evt);
>
> Really? What code does this generate? If this is generating two
> segment-prefixed reads rather than getting the address and doing normal
> reads on it, then I don't think it is an improvement.
Lets drop that hunk. No point to do optimizations at that location then.
evt is also not defined then. Without the evt address determination via
__get_cpu_var we have at least 2 prefixed load and one address
calculation to figure out the parameter to pass. No win.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>