Merge tag 'timers-core-2023-02-20' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull timer updates from Thomas Gleixner:
 "Updates for timekeeping, timers and clockevent/source drivers:

  Core:

   - Yet another round of improvements to make the clocksource watchdog
     more robust:

       - Relax the clocksource-watchdog skew criteria to match the NTP
         criteria.

       - Temporarily skip the watchdog when high memory latencies are
         detected which can lead to false-positives.

       - Provide an option to enable TSC skew detection even on systems
         where TSC is marked as reliable.

     Sigh!

   - Initialize the restart block in the nanosleep syscalls to be
     directed to the no restart function instead of doing a partial
     setup on entry.

     This prevents an erroneous restart_syscall() invocation from
     corrupting user space data. While such a situation is clearly a
     user space bug, preventing this is a correctness issue and caters
     to the least suprise principle.

   - Ignore the hrtimer slack for realtime tasks in schedule_hrtimeout()
     to align it with the nanosleep semantics.

  Drivers:

   - The obligatory new driver bindings for Mediatek, Rockchip and
     RISC-V variants.

   - Add support for the C3STOP misfeature to the RISC-V timer to handle
     the case where the timer stops in deeper idle state.

   - Set up a static key in the RISC-V timer correctly before first use.

   - The usual small improvements and fixes all over the place"

* tag 'timers-core-2023-02-20' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (30 commits)
  clocksource/drivers/timer-sun4i: Add CLOCK_EVT_FEAT_DYNIRQ
  clocksource/drivers/em_sti: Mark driver as non-removable
  clocksource/drivers/sh_tmu: Mark driver as non-removable
  clocksource/drivers/riscv: Patch riscv_clock_next_event() jump before first use
  clocksource/drivers/timer-microchip-pit64b: Add delay timer
  clocksource/drivers/timer-microchip-pit64b: Select driver only on ARM
  dt-bindings: timer: sifive,clint: add comaptibles for T-Head's C9xx
  dt-bindings: timer: mediatek,mtk-timer: add MT8365
  clocksource/drivers/riscv: Get rid of clocksource_arch_init() callback
  clocksource/drivers/sh_cmt: Mark driver as non-removable
  clocksource/drivers/timer-microchip-pit64b: Drop obsolete dependency on COMPILE_TEST
  clocksource/drivers/riscv: Increase the clock source rating
  clocksource/drivers/timer-riscv: Set CLOCK_EVT_FEAT_C3STOP based on DT
  dt-bindings: timer: Add bindings for the RISC-V timer device
  RISC-V: time: initialize hrtimer based broadcast clock event device
  dt-bindings: timer: rk-timer: Add rktimer for rv1126
  time/debug: Fix memory leak with using debugfs_lookup()
  clocksource: Enable TSC watchdog checking of HPET and PMTMR only when requested
  posix-timers: Use atomic64_try_cmpxchg() in __update_gt_cputime()
  clocksource: Verify HPET and PMTMR when TSC unverified
  ...

7 files changed, 81 insertions, 34 deletions

diff --git a/kernel/time/Kconfig b/kernel/time/Kconfig
index a41753be1a2bf..bae8f11070bef 100644
--- a/kernel/time/Kconfig
+++ b/kernel/time/Kconfig
@@ -200,10 +200,14 @@ config CLOCKSOURCE_WATCHDOG_MAX_SKEW_US
 	int "Clocksource watchdog maximum allowable skew (in μs)"
 	depends on CLOCKSOURCE_WATCHDOG
 	range 50 1000
-	default 100
+	default 125
 	help
 	  Specify the maximum amount of allowable watchdog skew in
 	  microseconds before reporting the clocksource to be unstable.
+	  The default is based on a half-second clocksource watchdog
+	  interval and NTP's maximum frequency drift of 500 parts
+	  per million.	If the clocksource is good enough for NTP,
+	  it is good enough for the clocksource watchdog!
 
 endmenu
 endif
diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c
index 9cf32ccda715d..91836b727cef5 100644
--- a/kernel/time/clocksource.c
+++ b/kernel/time/clocksource.c
@@ -96,6 +96,11 @@ static int finished_booting;
 static u64 suspend_start;
 
 /*
+ * Interval: 0.5sec.
+ */
+#define WATCHDOG_INTERVAL (HZ >> 1)
+
+/*
  * Threshold: 0.0312s, when doubled: 0.0625s.
  * Also a default for cs->uncertainty_margin when registering clocks.
  */
@@ -106,11 +111,14 @@ static u64 suspend_start;
  * clocksource surrounding a read of the clocksource being validated.
  * This delay could be due to SMIs, NMIs, or to VCPU preemptions.  Used as
  * a lower bound for cs->uncertainty_margin values when registering clocks.
+ *
+ * The default of 500 parts per million is based on NTP's limits.
+ * If a clocksource is good enough for NTP, it is good enough for us!
  */
 #ifdef CONFIG_CLOCKSOURCE_WATCHDOG_MAX_SKEW_US
 #define MAX_SKEW_USEC	CONFIG_CLOCKSOURCE_WATCHDOG_MAX_SKEW_US
 #else
-#define MAX_SKEW_USEC	100
+#define MAX_SKEW_USEC	(125 * WATCHDOG_INTERVAL / HZ)
 #endif
 
 #define WATCHDOG_MAX_SKEW (MAX_SKEW_USEC * NSEC_PER_USEC)
@@ -140,11 +148,6 @@ static inline void clocksource_watchdog_unlock(unsigned long *flags)
 static int clocksource_watchdog_kthread(void *data);
 static void __clocksource_change_rating(struct clocksource *cs, int rating);
 
-/*
- * Interval: 0.5sec.
- */
-#define WATCHDOG_INTERVAL (HZ >> 1)
-
 static void clocksource_watchdog_work(struct work_struct *work)
 {
 	/*
@@ -257,8 +260,8 @@ static enum wd_read_status cs_watchdog_read(struct clocksource *cs, u64 *csnow,
 			goto skip_test;
 	}
 
-	pr_warn("timekeeping watchdog on CPU%d: %s read-back delay of %lldns, attempt %d, marking unstable\n",
-		smp_processor_id(), watchdog->name, wd_delay, nretries);
+	pr_warn("timekeeping watchdog on CPU%d: wd-%s-wd excessive read-back delay of %lldns vs. limit of %ldns, wd-wd read-back delay only %lldns, attempt %d, marking %s unstable\n",
+		smp_processor_id(), cs->name, wd_delay, WATCHDOG_MAX_SKEW, wd_seq_delay, nretries, cs->name);
 	return WD_READ_UNSTABLE;
 
 skip_test:
@@ -384,6 +387,15 @@ void clocksource_verify_percpu(struct clocksource *cs)
 }
 EXPORT_SYMBOL_GPL(clocksource_verify_percpu);
 
+static inline void clocksource_reset_watchdog(void)
+{
+	struct clocksource *cs;
+
+	list_for_each_entry(cs, &watchdog_list, wd_list)
+		cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
+}
+
+
 static void clocksource_watchdog(struct timer_list *unused)
 {
 	u64 csnow, wdnow, cslast, wdlast, delta;
@@ -391,6 +403,7 @@ static void clocksource_watchdog(struct timer_list *unused)
 	int64_t wd_nsec, cs_nsec;
 	struct clocksource *cs;
 	enum wd_read_status read_ret;
+	unsigned long extra_wait = 0;
 	u32 md;
 
 	spin_lock(&watchdog_lock);
@@ -410,13 +423,30 @@ static void clocksource_watchdog(struct timer_list *unused)
 
 		read_ret = cs_watchdog_read(cs, &csnow, &wdnow);
 
-		if (read_ret != WD_READ_SUCCESS) {
-			if (read_ret == WD_READ_UNSTABLE)
-				/* Clock readout unreliable, so give it up. */
-				__clocksource_unstable(cs);
+		if (read_ret == WD_READ_UNSTABLE) {
+			/* Clock readout unreliable, so give it up. */
+			__clocksource_unstable(cs);
 			continue;
 		}
 
+		/*
+		 * When WD_READ_SKIP is returned, it means the system is likely
+		 * under very heavy load, where the latency of reading
+		 * watchdog/clocksource is very big, and affect the accuracy of
+		 * watchdog check. So give system some space and suspend the
+		 * watchdog check for 5 minutes.
+		 */
+		if (read_ret == WD_READ_SKIP) {
+			/*
+			 * As the watchdog timer will be suspended, and
+			 * cs->last could keep unchanged for 5 minutes, reset
+			 * the counters.
+			 */
+			clocksource_reset_watchdog();
+			extra_wait = HZ * 300;
+			break;
+		}
+
 		/* Clocksource initialized ? */
 		if (!(cs->flags & CLOCK_SOURCE_WATCHDOG) ||
 		    atomic_read(&watchdog_reset_pending)) {
@@ -443,12 +473,20 @@ static void clocksource_watchdog(struct timer_list *unused)
 		/* Check the deviation from the watchdog clocksource. */
 		md = cs->uncertainty_margin + watchdog->uncertainty_margin;
 		if (abs(cs_nsec - wd_nsec) > md) {
+			u64 cs_wd_msec;
+			u64 wd_msec;
+			u32 wd_rem;
+
 			pr_warn("timekeeping watchdog on CPU%d: Marking clocksource '%s' as unstable because the skew is too large:\n",
 				smp_processor_id(), cs->name);
 			pr_warn("                      '%s' wd_nsec: %lld wd_now: %llx wd_last: %llx mask: %llx\n",
 				watchdog->name, wd_nsec, wdnow, wdlast, watchdog->mask);
 			pr_warn("                      '%s' cs_nsec: %lld cs_now: %llx cs_last: %llx mask: %llx\n",
 				cs->name, cs_nsec, csnow, cslast, cs->mask);
+			cs_wd_msec = div_u64_rem(cs_nsec - wd_nsec, 1000U * 1000U, &wd_rem);
+			wd_msec = div_u64_rem(wd_nsec, 1000U * 1000U, &wd_rem);
+			pr_warn("                      Clocksource '%s' skewed %lld ns (%lld ms) over watchdog '%s' interval of %lld ns (%lld ms)\n",
+				cs->name, cs_nsec - wd_nsec, cs_wd_msec, watchdog->name, wd_nsec, wd_msec);
 			if (curr_clocksource == cs)
 				pr_warn("                      '%s' is current clocksource.\n", cs->name);
 			else if (curr_clocksource)
@@ -512,7 +550,7 @@ static void clocksource_watchdog(struct timer_list *unused)
 	 * pair clocksource_stop_watchdog() clocksource_start_watchdog().
 	 */
 	if (!timer_pending(&watchdog_timer)) {
-		watchdog_timer.expires += WATCHDOG_INTERVAL;
+		watchdog_timer.expires += WATCHDOG_INTERVAL + extra_wait;
 		add_timer_on(&watchdog_timer, next_cpu);
 	}
 out:
@@ -537,14 +575,6 @@ static inline void clocksource_stop_watchdog(void)
 	watchdog_running = 0;
 }
 
-static inline void clocksource_reset_watchdog(void)
-{
-	struct clocksource *cs;
-
-	list_for_each_entry(cs, &watchdog_list, wd_list)
-		cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
-}
-
 static void clocksource_resume_watchdog(void)
 {
 	atomic_inc(&watchdog_reset_pending);
diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c
index 3ae661ab62603..e8c08292defcb 100644
--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -2089,7 +2089,7 @@ long hrtimer_nanosleep(ktime_t rqtp, const enum hrtimer_mode mode,
 	u64 slack;
 
 	slack = current->timer_slack_ns;
-	if (dl_task(current) || rt_task(current))
+	if (rt_task(current))
 		slack = 0;
 
 	hrtimer_init_sleeper_on_stack(&t, clockid, mode);
@@ -2126,6 +2126,7 @@ SYSCALL_DEFINE2(nanosleep, struct __kernel_timespec __user *, rqtp,
 	if (!timespec64_valid(&tu))
 		return -EINVAL;
 
+	current->restart_block.fn = do_no_restart_syscall;
 	current->restart_block.nanosleep.type = rmtp ? TT_NATIVE : TT_NONE;
 	current->restart_block.nanosleep.rmtp = rmtp;
 	return hrtimer_nanosleep(timespec64_to_ktime(tu), HRTIMER_MODE_REL,
@@ -2147,6 +2148,7 @@ SYSCALL_DEFINE2(nanosleep_time32, struct old_timespec32 __user *, rqtp,
 	if (!timespec64_valid(&tu))
 		return -EINVAL;
 
+	current->restart_block.fn = do_no_restart_syscall;
 	current->restart_block.nanosleep.type = rmtp ? TT_COMPAT : TT_NONE;
 	current->restart_block.nanosleep.compat_rmtp = rmtp;
 	return hrtimer_nanosleep(timespec64_to_ktime(tu), HRTIMER_MODE_REL,
@@ -2270,7 +2272,7 @@ void __init hrtimers_init(void)
 /**
  * schedule_hrtimeout_range_clock - sleep until timeout
  * @expires:	timeout value (ktime_t)
- * @delta:	slack in expires timeout (ktime_t)
+ * @delta:	slack in expires timeout (ktime_t) for SCHED_OTHER tasks
  * @mode:	timer mode
  * @clock_id:	timer clock to be used
  */
@@ -2297,6 +2299,13 @@ schedule_hrtimeout_range_clock(ktime_t *expires, u64 delta,
 		return -EINTR;
 	}
 
+	/*
+	 * Override any slack passed by the user if under
+	 * rt contraints.
+	 */
+	if (rt_task(current))
+		delta = 0;
+
 	hrtimer_init_sleeper_on_stack(&t, clock_id, mode);
 	hrtimer_set_expires_range_ns(&t.timer, *expires, delta);
 	hrtimer_sleeper_start_expires(&t, mode);
@@ -2316,7 +2325,7 @@ EXPORT_SYMBOL_GPL(schedule_hrtimeout_range_clock);
 /**
  * schedule_hrtimeout_range - sleep until timeout
  * @expires:	timeout value (ktime_t)
- * @delta:	slack in expires timeout (ktime_t)
+ * @delta:	slack in expires timeout (ktime_t) for SCHED_OTHER tasks
  * @mode:	timer mode
  *
  * Make the current task sleep until the given expiry time has
@@ -2324,7 +2333,8 @@ EXPORT_SYMBOL_GPL(schedule_hrtimeout_range_clock);
  * the current task state has been set (see set_current_state()).
  *
  * The @delta argument gives the kernel the freedom to schedule the
- * actual wakeup to a time that is both power and performance friendly.
+ * actual wakeup to a time that is both power and performance friendly
+ * for regular (non RT/DL) tasks.
  * The kernel give the normal best effort behavior for "@expires+@delta",
  * but may decide to fire the timer earlier, but no earlier than @expires.
  *
diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index cb925e8ef9a8b..2f5e9b34022c5 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -243,13 +243,12 @@ static void proc_sample_cputime_atomic(struct task_cputime_atomic *at,
  */
 static inline void __update_gt_cputime(atomic64_t *cputime, u64 sum_cputime)
 {
-	u64 curr_cputime;
-retry:
-	curr_cputime = atomic64_read(cputime);
-	if (sum_cputime > curr_cputime) {
-		if (atomic64_cmpxchg(cputime, curr_cputime, sum_cputime) != curr_cputime)
-			goto retry;
-	}
+	u64 curr_cputime = atomic64_read(cputime);
+
+	do {
+		if (sum_cputime <= curr_cputime)
+			return;
+	} while (!atomic64_try_cmpxchg(cputime, &curr_cputime, sum_cputime));
 }
 
 static void update_gt_cputime(struct task_cputime_atomic *cputime_atomic,
diff --git a/kernel/time/posix-stubs.c b/kernel/time/posix-stubs.c
index 90ea5f373e50e..828aeecbd1e8a 100644
--- a/kernel/time/posix-stubs.c
+++ b/kernel/time/posix-stubs.c
@@ -147,6 +147,7 @@ SYSCALL_DEFINE4(clock_nanosleep, const clockid_t, which_clock, int, flags,
 		return -EINVAL;
 	if (flags & TIMER_ABSTIME)
 		rmtp = NULL;
+	current->restart_block.fn = do_no_restart_syscall;
 	current->restart_block.nanosleep.type = rmtp ? TT_NATIVE : TT_NONE;
 	current->restart_block.nanosleep.rmtp = rmtp;
 	texp = timespec64_to_ktime(t);
@@ -240,6 +241,7 @@ SYSCALL_DEFINE4(clock_nanosleep_time32, clockid_t, which_clock, int, flags,
 		return -EINVAL;
 	if (flags & TIMER_ABSTIME)
 		rmtp = NULL;
+	current->restart_block.fn = do_no_restart_syscall;
 	current->restart_block.nanosleep.type = rmtp ? TT_COMPAT : TT_NONE;
 	current->restart_block.nanosleep.compat_rmtp = rmtp;
 	texp = timespec64_to_ktime(t);
diff --git a/kernel/time/posix-timers.c b/kernel/time/posix-timers.c
index 5dead89308b74..0c8a87a11b39d 100644
--- a/kernel/time/posix-timers.c
+++ b/kernel/time/posix-timers.c
@@ -1270,6 +1270,7 @@ SYSCALL_DEFINE4(clock_nanosleep, const clockid_t, which_clock, int, flags,
 		return -EINVAL;
 	if (flags & TIMER_ABSTIME)
 		rmtp = NULL;
+	current->restart_block.fn = do_no_restart_syscall;
 	current->restart_block.nanosleep.type = rmtp ? TT_NATIVE : TT_NONE;
 	current->restart_block.nanosleep.rmtp = rmtp;
 
@@ -1297,6 +1298,7 @@ SYSCALL_DEFINE4(clock_nanosleep_time32, clockid_t, which_clock, int, flags,
 		return -EINVAL;
 	if (flags & TIMER_ABSTIME)
 		rmtp = NULL;
+	current->restart_block.fn = do_no_restart_syscall;
 	current->restart_block.nanosleep.type = rmtp ? TT_COMPAT : TT_NONE;
 	current->restart_block.nanosleep.compat_rmtp = rmtp;
 
diff --git a/kernel/time/test_udelay.c b/kernel/time/test_udelay.c
index 13b11eb62685e..20d5df631570e 100644
--- a/kernel/time/test_udelay.c
+++ b/kernel/time/test_udelay.c
@@ -149,7 +149,7 @@ module_init(udelay_test_init);
 static void __exit udelay_test_exit(void)
 {
 	mutex_lock(&udelay_test_lock);
-	debugfs_remove(debugfs_lookup(DEBUGFS_FILENAME, NULL));
+	debugfs_lookup_and_remove(DEBUGFS_FILENAME, NULL);
 	mutex_unlock(&udelay_test_lock);
 }