Merge branch 'clocksource' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu into timers/core

Pull clocksource watchdog updates from Paul McKenney:

 - Avoid accidental unstable marking of clocksources by rejecting
   clocksource measurements where the source of the skew is the delay
   reading reference clocksource itself.  This change avoids many of the
   current false positives caused by epic cache-thrashing workloads.

 - Reduce the default clocksource_watchdog() retries to 2, thus offsetting
   the increased overhead due to #1 above rereading the reference
   clocksource.

Link: https://lore.kernel.org/lkml/20220105001723.GA536708@paulmck-ThinkPad-P17-Gen-1

7 files changed, 49 insertions, 12 deletions

diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index 9725c546a0d4..3ea934b034f7 100644
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -603,8 +603,8 @@
 	clocksource.max_cswd_read_retries= [KNL]
 			Number of clocksource_watchdog() retries due to
 			external delays before the clock will be marked
-			unstable.  Defaults to three retries, that is,
-			four attempts to read the clock under test.
+			unstable.  Defaults to two retries, that is,
+			three attempts to read the clock under test.
 
 	clocksource.verify_n_cpus= [KNL]
 			Limit the number of CPUs checked for clocksources
diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c
index b8a14d2fb5ba..b7e52a642948 100644
--- a/kernel/time/clocksource.c
+++ b/kernel/time/clocksource.c
@@ -107,7 +107,7 @@ static u64 suspend_start;
  * 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.
  */
-#define WATCHDOG_MAX_SKEW (50 * NSEC_PER_USEC)
+#define WATCHDOG_MAX_SKEW (100 * NSEC_PER_USEC)
 
 #ifdef CONFIG_CLOCKSOURCE_WATCHDOG
 static void clocksource_watchdog_work(struct work_struct *work);
@@ -199,23 +199,30 @@ void clocksource_mark_unstable(struct clocksource *cs)
 	spin_unlock_irqrestore(&watchdog_lock, flags);
 }
 
-ulong max_cswd_read_retries = 3;
+ulong max_cswd_read_retries = 2;
 module_param(max_cswd_read_retries, ulong, 0644);
 EXPORT_SYMBOL_GPL(max_cswd_read_retries);
 static int verify_n_cpus = 8;
 module_param(verify_n_cpus, int, 0644);
 
-static bool cs_watchdog_read(struct clocksource *cs, u64 *csnow, u64 *wdnow)
+enum wd_read_status {
+	WD_READ_SUCCESS,
+	WD_READ_UNSTABLE,
+	WD_READ_SKIP
+};
+
+static enum wd_read_status cs_watchdog_read(struct clocksource *cs, u64 *csnow, u64 *wdnow)
 {
 	unsigned int nretries;
-	u64 wd_end, wd_delta;
-	int64_t wd_delay;
+	u64 wd_end, wd_end2, wd_delta;
+	int64_t wd_delay, wd_seq_delay;
 
 	for (nretries = 0; nretries <= max_cswd_read_retries; nretries++) {
 		local_irq_disable();
 		*wdnow = watchdog->read(watchdog);
 		*csnow = cs->read(cs);
 		wd_end = watchdog->read(watchdog);
+		wd_end2 = watchdog->read(watchdog);
 		local_irq_enable();
 
 		wd_delta = clocksource_delta(wd_end, *wdnow, watchdog->mask);
@@ -226,13 +233,34 @@ static bool cs_watchdog_read(struct clocksource *cs, u64 *csnow, u64 *wdnow)
 				pr_warn("timekeeping watchdog on CPU%d: %s retried %d times before success\n",
 					smp_processor_id(), watchdog->name, nretries);
 			}
-			return true;
+			return WD_READ_SUCCESS;
 		}
+
+		/*
+		 * Now compute delay in consecutive watchdog read to see if
+		 * there is too much external interferences that cause
+		 * significant delay in reading both clocksource and watchdog.
+		 *
+		 * If consecutive WD read-back delay > WATCHDOG_MAX_SKEW/2,
+		 * report system busy, reinit the watchdog and skip the current
+		 * watchdog test.
+		 */
+		wd_delta = clocksource_delta(wd_end2, wd_end, watchdog->mask);
+		wd_seq_delay = clocksource_cyc2ns(wd_delta, watchdog->mult, watchdog->shift);
+		if (wd_seq_delay > WATCHDOG_MAX_SKEW/2)
+			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);
-	return false;
+	return WD_READ_UNSTABLE;
+
+skip_test:
+	pr_info("timekeeping watchdog on CPU%d: %s wd-wd read-back delay of %lldns\n",
+		smp_processor_id(), watchdog->name, wd_seq_delay);
+	pr_info("wd-%s-wd read-back delay of %lldns, clock-skew test skipped!\n",
+		cs->name, wd_delay);
+	return WD_READ_SKIP;
 }
 
 static u64 csnow_mid;
@@ -356,6 +384,7 @@ static void clocksource_watchdog(struct timer_list *unused)
 	int next_cpu, reset_pending;
 	int64_t wd_nsec, cs_nsec;
 	struct clocksource *cs;
+	enum wd_read_status read_ret;
 	u32 md;
 
 	spin_lock(&watchdog_lock);
@@ -373,9 +402,12 @@ static void clocksource_watchdog(struct timer_list *unused)
 			continue;
 		}
 
-		if (!cs_watchdog_read(cs, &csnow, &wdnow)) {
-			/* Clock readout unreliable, so give it up. */
-			__clocksource_unstable(cs);
+		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);
 			continue;
 		}
 
diff --git a/tools/testing/selftests/rcutorture/configs/rcu/SRCU-T b/tools/testing/selftests/rcutorture/configs/rcu/SRCU-T
index d6557c38dfe4..c70cf0405f24 100644
--- a/tools/testing/selftests/rcutorture/configs/rcu/SRCU-T
+++ b/tools/testing/selftests/rcutorture/configs/rcu/SRCU-T
@@ -2,6 +2,7 @@ CONFIG_SMP=n
 CONFIG_PREEMPT_NONE=y
 CONFIG_PREEMPT_VOLUNTARY=n
 CONFIG_PREEMPT=n
+CONFIG_PREEMPT_DYNAMIC=n
 #CHECK#CONFIG_TINY_SRCU=y
 CONFIG_RCU_TRACE=n
 CONFIG_DEBUG_LOCK_ALLOC=y
diff --git a/tools/testing/selftests/rcutorture/configs/rcu/SRCU-U b/tools/testing/selftests/rcutorture/configs/rcu/SRCU-U
index 6bc24e99862f..bc9eeabaa1b1 100644
--- a/tools/testing/selftests/rcutorture/configs/rcu/SRCU-U
+++ b/tools/testing/selftests/rcutorture/configs/rcu/SRCU-U
@@ -2,6 +2,7 @@ CONFIG_SMP=n
 CONFIG_PREEMPT_NONE=y
 CONFIG_PREEMPT_VOLUNTARY=n
 CONFIG_PREEMPT=n
+CONFIG_PREEMPT_DYNAMIC=n
 #CHECK#CONFIG_TINY_SRCU=y
 CONFIG_RCU_TRACE=n
 CONFIG_DEBUG_LOCK_ALLOC=n
diff --git a/tools/testing/selftests/rcutorture/configs/rcu/TINY01 b/tools/testing/selftests/rcutorture/configs/rcu/TINY01
index 6db705e55487..0953c52fcfd7 100644
--- a/tools/testing/selftests/rcutorture/configs/rcu/TINY01
+++ b/tools/testing/selftests/rcutorture/configs/rcu/TINY01
@@ -2,6 +2,7 @@ CONFIG_SMP=n
 CONFIG_PREEMPT_NONE=y
 CONFIG_PREEMPT_VOLUNTARY=n
 CONFIG_PREEMPT=n
+CONFIG_PREEMPT_DYNAMIC=n
 #CHECK#CONFIG_TINY_RCU=y
 CONFIG_HZ_PERIODIC=n
 CONFIG_NO_HZ_IDLE=y
diff --git a/tools/testing/selftests/rcutorture/configs/rcu/TINY02 b/tools/testing/selftests/rcutorture/configs/rcu/TINY02
index d8674264318d..30439f6fc20e 100644
--- a/tools/testing/selftests/rcutorture/configs/rcu/TINY02
+++ b/tools/testing/selftests/rcutorture/configs/rcu/TINY02
@@ -2,6 +2,7 @@ CONFIG_SMP=n
 CONFIG_PREEMPT_NONE=y
 CONFIG_PREEMPT_VOLUNTARY=n
 CONFIG_PREEMPT=n
+CONFIG_PREEMPT_DYNAMIC=n
 #CHECK#CONFIG_TINY_RCU=y
 CONFIG_HZ_PERIODIC=y
 CONFIG_NO_HZ_IDLE=n
diff --git a/tools/testing/selftests/rcutorture/configs/rcuscale/TINY b/tools/testing/selftests/rcutorture/configs/rcuscale/TINY
index fb05ef5279b4..64e9490c1f79 100644
--- a/tools/testing/selftests/rcutorture/configs/rcuscale/TINY
+++ b/tools/testing/selftests/rcutorture/configs/rcuscale/TINY
@@ -2,6 +2,7 @@ CONFIG_SMP=n
 CONFIG_PREEMPT_NONE=y
 CONFIG_PREEMPT_VOLUNTARY=n
 CONFIG_PREEMPT=n
+CONFIG_PREEMPT_DYNAMIC=n
 #CHECK#CONFIG_TINY_RCU=y
 CONFIG_HZ_PERIODIC=n
 CONFIG_NO_HZ_IDLE=y