timerfd: Allow timers to be cancelled when clock was set

Some applications must be aware of clock realtime being set
backward. A simple example is a clock applet which arms a timer for
the next minute display. If clock realtime is set backward then the
applet displays a stale time for the amount of time which the clock
was set backwards. Due to that applications poll the time because we
don't have an interface.

Extend the timerfd interface by adding a flag which puts the timer
onto a different internal realtime clock. All timers on this clock are
expired whenever the clock was set.

The timerfd core records the monotonic offset when the timer is
created. When the timer is armed, then the current offset is compared
to the previous recorded offset. When it has changed, then
timerfd_settime returns -ECANCELED. When a timer is read the offset is
compared and if it changed -ECANCELED returned to user space. Periodic
timers are not rearmed in the cancelation case.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: John Stultz <johnstul@us.ibm.com>
Cc: Chris Friesen <chris.friesen@genband.com>
Tested-by: Kay Sievers <kay.sievers@vrfy.org>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Davide Libenzi <davidel@xmailserver.org>
Reviewed-by: Alexander Shishkin <virtuoso@slind.org>
Link: http://lkml.kernel.org/r/%3Calpine.LFD.2.02.1104271359580.3323%40ionos%3E
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

1 files changed, 51 insertions, 6 deletions

diff --git a/fs/timerfd.c b/fs/timerfd.c
index 8c4fc1425b3e..7e14c9e7c4ee 100644
--- a/fs/timerfd.c
+++ b/fs/timerfd.c
@@ -26,10 +26,12 @@
 struct timerfd_ctx {
 	struct hrtimer tmr;
 	ktime_t tintv;
+	ktime_t moffs;
 	wait_queue_head_t wqh;
 	u64 ticks;
 	int expired;
 	int clockid;
+	bool might_cancel;
 };
 
 /*
@@ -59,24 +61,52 @@ static ktime_t timerfd_get_remaining(struct timerfd_ctx *ctx)
 	return remaining.tv64 < 0 ? ktime_set(0, 0): remaining;
 }
 
-static void timerfd_setup(struct timerfd_ctx *ctx, int flags,
-			  const struct itimerspec *ktmr)
+static bool timerfd_canceled(struct timerfd_ctx *ctx)
+{
+	ktime_t moffs;
+
+	if (!ctx->might_cancel)
+		return false;
+
+	moffs = ktime_get_monotonic_offset();
+
+	if (moffs.tv64 == ctx->moffs.tv64)
+		return false;
+
+	ctx->moffs = moffs;
+	return true;
+}
+
+static int timerfd_setup(struct timerfd_ctx *ctx, int flags,
+			 const struct itimerspec *ktmr)
 {
 	enum hrtimer_mode htmode;
 	ktime_t texp;
+	int clockid = ctx->clockid;
 
 	htmode = (flags & TFD_TIMER_ABSTIME) ?
 		HRTIMER_MODE_ABS: HRTIMER_MODE_REL;
 
+	ctx->might_cancel = false;
+	if (htmode == HRTIMER_MODE_ABS && ctx->clockid == CLOCK_REALTIME &&
+	    (flags & TFD_TIMER_CANCELON_SET)) {
+		clockid = CLOCK_REALTIME_COS;
+		ctx->might_cancel = true;
+	}
+
 	texp = timespec_to_ktime(ktmr->it_value);
 	ctx->expired = 0;
 	ctx->ticks = 0;
 	ctx->tintv = timespec_to_ktime(ktmr->it_interval);
-	hrtimer_init(&ctx->tmr, ctx->clockid, htmode);
+	hrtimer_init(&ctx->tmr, clockid, htmode);
 	hrtimer_set_expires(&ctx->tmr, texp);
 	ctx->tmr.function = timerfd_tmrproc;
-	if (texp.tv64 != 0)
+	if (texp.tv64 != 0) {
 		hrtimer_start(&ctx->tmr, texp, htmode);
+		if (timerfd_canceled(ctx))
+			return -ECANCELED;
+	}
+	return 0;
 }
 
 static int timerfd_release(struct inode *inode, struct file *file)
@@ -118,8 +148,21 @@ static ssize_t timerfd_read(struct file *file, char __user *buf, size_t count,
 		res = -EAGAIN;
 	else
 		res = wait_event_interruptible_locked_irq(ctx->wqh, ctx->ticks);
+
 	if (ctx->ticks) {
 		ticks = ctx->ticks;
+
+		/*
+		 * If clock has changed, we do not care about the
+		 * ticks and we do not rearm the timer. Userspace must
+		 * reevaluate anyway.
+		 */
+		if (timerfd_canceled(ctx)) {
+			ticks = 0;
+			ctx->expired = 0;
+			res = -ECANCELED;
+		}
+
 		if (ctx->expired && ctx->tintv.tv64) {
 			/*
 			 * If tintv.tv64 != 0, this is a periodic timer that
@@ -183,6 +226,7 @@ SYSCALL_DEFINE2(timerfd_create, int, clockid, int, flags)
 	init_waitqueue_head(&ctx->wqh);
 	ctx->clockid = clockid;
 	hrtimer_init(&ctx->tmr, clockid, HRTIMER_MODE_ABS);
+	ctx->moffs = ktime_get_monotonic_offset();
 
 	ufd = anon_inode_getfd("[timerfd]", &timerfd_fops, ctx,
 			       O_RDWR | (flags & TFD_SHARED_FCNTL_FLAGS));
@@ -199,6 +243,7 @@ SYSCALL_DEFINE4(timerfd_settime, int, ufd, int, flags,
 	struct file *file;
 	struct timerfd_ctx *ctx;
 	struct itimerspec ktmr, kotmr;
+	int ret;
 
 	if (copy_from_user(&ktmr, utmr, sizeof(ktmr)))
 		return -EFAULT;
@@ -240,14 +285,14 @@ SYSCALL_DEFINE4(timerfd_settime, int, ufd, int, flags,
 	/*
 	 * Re-program the timer to the new value ...
 	 */
-	timerfd_setup(ctx, flags, &ktmr);
+	ret = timerfd_setup(ctx, flags, &ktmr);
 
 	spin_unlock_irq(&ctx->wqh.lock);
 	fput(file);
 	if (otmr && copy_to_user(otmr, &kotmr, sizeof(kotmr)))
 		return -EFAULT;
 
-	return 0;
+	return ret;
 }
 
 SYSCALL_DEFINE2(timerfd_gettime, int, ufd, struct itimerspec __user *, otmr)