On Sunday 07 January 2007 3:19 am, Pavel Machek wrote:
>
> > Create /sys/power/alarm.
> > The way it works is exactly the same as /proc/acpi/alarm.
> > I.e. "#echo yyyy-mm-dd hh-mm-ss >/sys/power/alarm" supports existing absolute time.
> > And "#echo +yyyy-mm-dd hh-mm-ss >/sys/power/alarm" supports a duration.
>
> NAK. /proc/acpi/alarm is a mess, and this just moves it to /sysfs.
> 'One value per file', please.
Sort of like the appended patch, instead ... which doesn't need to know a
thing about ACPI. This is what I suggested in response to an earlier patch
from Paul Sokolovsky.
- Dave
================ CUT HERE
This adds a new "wakealarm" sysfs attribute to RTC class devices which
support alarm operations and are wakeup-capable:
- It reads as either empty, or the scheduled alarm time as seconds
since the POSIX epoch. (That time may already have passed, since
nothing currently enforces one-shot alarm semantics...)
- It can be written with an alarm time in the future, again seconds
since the POSIX epoch, which enables the alarm.
- It can be written with an alarm time not in the future (such as 0,
the start of the POSIX epoch) to disable the alarm.
Usage examples, after "cd /sys/class/rtc/rtcN":
alarm after 45 minutes:
# echo $(( $(cat since_epoch) + 45 * 60 )) > wakealarm
alarm next tuesday evening (using GNU date):
# date -d '10pm tuesday' "+%s" > wakealarm
disable alarm:
# echo 0 > wakealarm
This resembles the /proc/acpi/alarm file in that nothing happens when
the alarm triggers ... except possibly waking the system from sleep.
It's also like that in a nasty way: not much can be done to prevent
one task from clobbering another task's alarm settings.
It differs from that file in that there's no in-kernel date parser.
Note that a few RTCs ignore rtc_wkalrm.enabled when setting alarms, or
aren't set up correctly, so they won't yet behave with this attribute.
Signed-off-by: David Brownell <dbrownell at users.sourceforge.net>
Index: g26/drivers/rtc/rtc-sysfs.c
===================================================================
--- g26.orig/drivers/rtc/rtc-sysfs.c 2007-01-07 19:34:50.000000000 -0800
+++ g26/drivers/rtc/rtc-sysfs.c 2007-01-07 19:34:52.000000000 -0800
@@ -78,6 +78,92 @@ static struct attribute_group rtc_attr_g
.attrs = rtc_attrs,
};
+
+static ssize_t
+rtc_sysfs_show_wakealarm(struct class_device *dev, char *buf)
+{
+ ssize_t retval;
+ unsigned long alarm;
+ struct rtc_wkalrm alm;
+
+ /* Don't show disabled alarms; but the RTC could leave the
+ * alarm enabled after it's already triggered. Alarms are
+ * conceptually one-shot, even though some common hardware
+ * (PCs) doesn't actually work that way.
+ *
+ * REVISIT maybe we should require RTC implementations to
+ * disable the RTC alarm after it triggers, for uniformity.
+ */
+ retval = rtc_read_alarm(dev, &alm);
+ if (retval == 0 && alm.enabled) {
+ rtc_tm_to_time(&alm.time, &alarm);
+ retval = sprintf(buf, "%lu\n", alarm);
+ }
+
+ return retval;
+}
+
+static ssize_t
+rtc_sysfs_set_wakealarm(struct class_device *dev, const char *buf, size_t n)
+{
+ ssize_t retval;
+ unsigned long now, alarm;
+ struct rtc_wkalrm alm;
+
+ /* Only request alarms that trigger in the future. Disable them
+ * by writing another time, e.g. 0 meaning Jan 1 1970 UTC.
+ */
+ retval = rtc_read_time(dev, &alm.time);
+ if (retval < 0)
+ return retval;
+ rtc_tm_to_time(&alm.time, &now);
+
+ alarm = simple_strtoul(buf, NULL, 0);
+ if (alarm > now) {
+ /* Avoid accidentally clobbering active alarms; we can't
+ * entirely prevent that here, without even the minimal
+ * locking from the /dev/rtcN api.
+ */
+ retval = rtc_read_alarm(dev, &alm);
+ if (retval < 0)
+ return retval;
+ if (alm.enabled)
+ return -EBUSY;
+
+ alm.enabled = 1;
+ } else {
+ alm.enabled = 0;
+
+ /* Provide a valid future alarm time. Linux isn't EFI,
+ * this time won't be ignored when disabling the alarm.
+ */
+ alarm = now + 300;
+ }
+ rtc_time_to_tm(alarm, &alm.time);
+
+ retval = rtc_set_alarm(dev, &alm);
+ return (retval < 0) ? retval : n;
+}
+static const CLASS_DEVICE_ATTR(wakealarm, S_IRUGO | S_IWUSR,
+ rtc_sysfs_show_wakealarm, rtc_sysfs_set_wakealarm);
+
+
+/* The reason to trigger an alarm with no process watching it (via sysfs)
+ * is its side effect: waking from a system state like suspend-to-RAM or
+ * suspend-to-disk. So: no attribute unless that side effect is possible.
+ * (Userspace may disable that mechanism later.)
+ */
+static inline int rtc_does_wakealarm(struct class_device *class_dev)
+{
+ struct rtc_device *rtc;
+
+ if (!device_can_wakeup(class_dev->dev))
+ return 0;
+ rtc = to_rtc_device(class_dev);
+ return rtc->ops->set_alarm != NULL;
+}
+
+
static int __devinit rtc_sysfs_add_device(struct class_device *class_dev,
struct class_interface *class_intf)
{
@@ -87,8 +173,18 @@ static int __devinit rtc_sysfs_add_devic
err = sysfs_create_group(&class_dev->kobj, &rtc_attr_group);
if (err)
- dev_err(class_dev->dev,
- "failed to create sysfs attributes\n");
+ dev_err(class_dev->dev, "failed to create %s\n",
+ "sysfs attributes");
+ else if (rtc_does_wakealarm(class_dev)) {
+ /* not all RTCs support both alarms and wakeup */
+ err = class_device_create_file(class_dev,
+ &class_device_attr_wakealarm);
+ if (err) {
+ dev_err(class_dev->dev, "failed to create %s\n",
+ "alarm attribute");
+ sysfs_remove_group(&class_dev->kobj, &rtc_attr_group);
+ }
+ }
return err;
}
@@ -96,6 +192,9 @@ static int __devinit rtc_sysfs_add_devic
static void rtc_sysfs_remove_device(struct class_device *class_dev,
struct class_interface *class_intf)
{
+ if (rtc_does_wakealarm(class_dev))
+ class_device_remove_file(class_dev,
+ &class_device_attr_wakealarm);
sysfs_remove_group(&class_dev->kobj, &rtc_attr_group);
}
