[PATCH] x86_64: Don't need to read PIT in timer handler when PM timer is used
authorAndi Kleen <ak@suse.de>
Sat, 25 Mar 2006 15:29:31 +0000 (16:29 +0100)
committerLinus Torvalds <torvalds@g5.osdl.org>
Sat, 25 Mar 2006 17:10:53 +0000 (09:10 -0800)
The PM timer path through main_timer_handler doesn't need
the delay variable because it figures it out in a different way.
Don't try to read it from the PIT. With stopped PIT timer
it is even useless.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
arch/x86_64/kernel/time.c

index ee5ce3d3cbc3e8c50c2bdbaba99070e3e64d9957..86039baaaf185872fe68e23a639b1d34105f557b 100644 (file)
@@ -354,7 +354,7 @@ void main_timer_handler(struct pt_regs *regs)
 {
        static unsigned long rtc_update = 0;
        unsigned long tsc;
-       int delay, offset = 0, lost = 0;
+       int delay = 0, offset = 0, lost = 0;
 
 /*
  * Here we are in the timer irq handler. We have irqs locally disabled (so we
@@ -375,7 +375,7 @@ void main_timer_handler(struct pt_regs *regs)
                 */
                offset = hpet_readl(HPET_T0_CMP) - hpet_tick;
                delay = hpet_readl(HPET_COUNTER) - offset;
-       } else {
+       } else if (!pmtmr_ioport) {
                spin_lock(&i8253_lock);
                outb_p(0x00, 0x43);
                delay = inb_p(0x40);