Skip to contentSkip to conversation

Issues in Linux HPET example

PATCH New Jaswinder Singh Rajput 2 participants 2 messages Sep 1, 2010 – Sep 6, 2010

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hello,

By executing Documentation/timers/hpet_example.c

for polling, I requested for 3 iterations but it seems iteration work for only 2 as first expired time is always very small.

[root@netbook timers]# ./hpet_example poll /dev/hpet 10 3 -hpet: executing poll
hpet_poll: info.hi_flags 0x0
hpet_poll: expired time = 0x13
hpet_poll: revents = 0x1
hpet_poll: data 0x1
hpet_poll: expired time = 0x1868c
hpet_poll: revents = 0x1
hpet_poll: data 0x1
hpet_poll: expired time = 0x18645
hpet_poll: revents = 0x1
hpet_poll: data 0x1
[root@netbook timers]#

for fasync, I requested for 3 iterations but it shows 4 counts.

[root@netbook timers]# ./hpet_example fasync /dev/hpet 10 3 -hpet: executing fasync
hpet_fasync: info.hi_flags 0x0
hpet_sigio: called
hpet_sigio: called
hpet_fasync: count = 2
hpet_sigio: called
hpet_fasync: count = 3
hpet_sigio: called
hpet_fasync: count = 4
[root@netbook timers]#

How can I fix these issues.

Thanks,
--
Jaswinder Singh.

maintainer
#1 lore raw

PATCH 1/1 hpet: fix unwanted interrupt due to stale irq status bit

Quoted 5 lines · Jaswinder Singh Rajput

By executing Documentation/timers/hpet_example.c

for polling, I requested for 3 iterations but it seems iteration work for only 2 as first expired time is always very small.

Quoted 12 lines
# ./hpet_example poll /dev/hpet 10 3
-hpet: executing poll
hpet_poll: info.hi_flags 0x0
hpet_poll: expired time = 0x13
hpet_poll: revents = 0x1
hpet_poll: data 0x1
hpet_poll: expired time = 0x1868c
hpet_poll: revents = 0x1
hpet_poll: data 0x1
hpet_poll: expired time = 0x18645
hpet_poll: revents = 0x1
hpet_poll: data 0x1

Clearing the HPET interrupt enable bit disables interrupt generation but does not disable the timer, so the interrupt status bit will still be set when the timer elapses. If another interrupt arrives before the timer has been correctly programmed (due to some other device on the same interrupt line, or CONFIG_DEBUG_SHIRQ), this results in an extra unwanted interrupt event because the status bit is likely to be set from comparator matches that happened before the device was opened.

Therefore, we have to ensure that the interrupt status bit is and stays cleared until we actually program the timer.

---
 drivers/char/hpet.c |   15 +++++++++++++++
 1 file changed, 15 insertions(+)
drivers/char/hpet.c +15 -0
@@ -514,6 +514,21 @@ static int hpet_ioctl_ieon(struct hpet_d
 	if (irq) {
 		unsigned long irq_flags;
 
+		if (devp->hd_flags & HPET_SHARED_IRQ) {
+			/*
+			 * To prevent the interrupt handler from seeing an
+			 * unwanted interrupt status bit, program the timer
+			 * so that it will not fire in the near future ...
+			 */
+			writel(readl(&timer->hpet_config) & ~Tn_TYPE_CNF_MASK,
+			       &timer->hpet_config);
+			write_counter(read_counter(&hpet->hpet_mc),
+				      &timer->hpet_compare);
+			/* ... and clear any left-over status. */
+			isr = 1 << (devp - devp->hd_hpets->hp_dev);
+			writel(isr, &hpet->hpet_isr);
+		}
+
 		sprintf(devp->hd_name, "hpet%d", (int)(devp - hpetp->hp_dev));
 		irq_flags = devp->hd_flags & HPET_SHARED_IRQ
 						? IRQF_SHARED : IRQF_DISABLED;