Merge tag 'rcu.2023.06.22a' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu

Pull RCU updates from Paul McKenney:
 "Documentation updates

  Miscellaneous fixes, perhaps most notably:

   - Remove RCU_NONIDLE(). The new visibility of most of the idle loop
     to RCU has obsoleted this API.

   - Make the RCU_SOFTIRQ callback-invocation time limit also apply to
     the rcuc kthreads that invoke callbacks for CONFIG_PREEMPT_RT.

   - Add a jiffies-based callback-invocation time limit to handle
     long-running callbacks. (The local_clock() function is only invoked
     once per 32 callbacks due to its high overhead.)

   - Stop rcu_tasks_invoke_cbs() from using never-onlined CPUs, which
     fixes a bug that can occur on systems with non-contiguous CPU
     numbering.

  kvfree_rcu updates:

   - Eliminate the single-argument variant of k[v]free_rcu() now that
     all uses have been converted to k[v]free_rcu_mightsleep().

   - Add WARN_ON_ONCE() checks for k[v]free_rcu*() freeing callbacks too
     soon. Yes, this is closing the barn door after the horse has
     escaped, but Murphy says that there will be more horses.

  Callback-offloading updates:

   - Fix a number of bugs involving the shrinker and lazy callbacks.

  Tasks RCU updates

  Torture-test updates"

* tag 'rcu.2023.06.22a' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu: (32 commits)
  torture: Remove duplicated argument -enable-kvm for ppc64
  doc/rcutorture: Add description of rcutorture.stall_cpu_block
  rcu/rcuscale: Stop kfree_scale_thread thread(s) after unloading rcuscale
  rcu/rcuscale: Move rcu_scale_*() after kfree_scale_cleanup()
  rcutorture: Correct name of use_softirq module parameter
  locktorture: Add long_hold to adjust lock-hold delays
  rcu/nocb: Make shrinker iterate only over NOCB CPUs
  rcu-tasks: Stop rcu_tasks_invoke_cbs() from using never-onlined CPUs
  rcu: Make rcu_cpu_starting() rely on interrupts being disabled
  rcu: Mark rcu_cpu_kthread() accesses to ->rcu_cpu_has_work
  rcu: Mark additional concurrent load from ->cpu_no_qs.b.exp
  rcu: Employ jiffies-based backstop to callback time limit
  rcu: Check callback-invocation time limit for rcuc kthreads
  rcu: Remove RCU_NONIDLE()
  rcu: Add more RCU files to kernel-api.rst
  rcu-tasks: Clarify the cblist_init_generic() function's pr_info() output
  rcu-tasks: Avoid pr_info() with spin lock in cblist_init_generic()
  rcu/nocb: Recheck lazy callbacks under the ->nocb_lock from shrinker
  rcu/nocb: Fix shrinker race against callback enqueuer
  rcu/nocb: Protect lazy shrinker against concurrent (de-)offloading
  ...

kernel/locking/locktorture.c

@@ -33,24 +33,19 @@
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Paul E. McKenney <paulmck@linux.ibm.com>");
 
-torture_param(int, nwriters_stress, -1,
-	     "Number of write-locking stress-test threads");
-torture_param(int, nreaders_stress, -1,
-	     "Number of read-locking stress-test threads");
+torture_param(int, nwriters_stress, -1, "Number of write-locking stress-test threads");
+torture_param(int, nreaders_stress, -1, "Number of read-locking stress-test threads");
+torture_param(int, long_hold, 100, "Do occasional long hold of lock (ms), 0=disable");
 torture_param(int, onoff_holdoff, 0, "Time after boot before CPU hotplugs (s)");
-torture_param(int, onoff_interval, 0,
-	     "Time between CPU hotplugs (s), 0=disable");
-torture_param(int, shuffle_interval, 3,
-	     "Number of jiffies between shuffles, 0=disable");
+torture_param(int, onoff_interval, 0, "Time between CPU hotplugs (s), 0=disable");
+torture_param(int, shuffle_interval, 3, "Number of jiffies between shuffles, 0=disable");
 torture_param(int, shutdown_secs, 0, "Shutdown time (j), <= zero to disable.");
-torture_param(int, stat_interval, 60,
-	     "Number of seconds between stats printk()s");
+torture_param(int, stat_interval, 60, "Number of seconds between stats printk()s");
 torture_param(int, stutter, 5, "Number of jiffies to run/halt test, 0=disable");
 torture_param(int, rt_boost, 2,
-		"Do periodic rt-boost. 0=Disable, 1=Only for rt_mutex, 2=For all lock types.");
+		   "Do periodic rt-boost. 0=Disable, 1=Only for rt_mutex, 2=For all lock types.");
 torture_param(int, rt_boost_factor, 50, "A factor determining how often rt-boost happens.");
-torture_param(int, verbose, 1,
-	     "Enable verbose debugging printk()s");
+torture_param(int, verbose, 1, "Enable verbose debugging printk()s");
 torture_param(int, nested_locks, 0, "Number of nested locks (max = 8)");
 /* Going much higher trips "BUG: MAX_LOCKDEP_CHAIN_HLOCKS too low!" errors */
 #define MAX_NESTED_LOCKS 8
@@ -120,7 +115,7 @@ static int torture_lock_busted_write_lock(int tid __maybe_unused)
 
 static void torture_lock_busted_write_delay(struct torture_random_state *trsp)
 {
-	const unsigned long longdelay_ms = 100;
+	const unsigned long longdelay_ms = long_hold ? long_hold : ULONG_MAX;
 
 	/* We want a long delay occasionally to force massive contention.  */
 	if (!(torture_random(trsp) %
@@ -198,16 +193,18 @@ __acquires(torture_spinlock)
 static void torture_spin_lock_write_delay(struct torture_random_state *trsp)
 {
 	const unsigned long shortdelay_us = 2;
-	const unsigned long longdelay_ms = 100;
+	const unsigned long longdelay_ms = long_hold ? long_hold : ULONG_MAX;
+	unsigned long j;
 
 	/* We want a short delay mostly to emulate likely code, and
 	 * we want a long delay occasionally to force massive contention.
 	 */
-	if (!(torture_random(trsp) %
-	      (cxt.nrealwriters_stress * 2000 * longdelay_ms)))
+	if (!(torture_random(trsp) % (cxt.nrealwriters_stress * 2000 * longdelay_ms))) {
+		j = jiffies;
 		mdelay(longdelay_ms);
-	if (!(torture_random(trsp) %
-	      (cxt.nrealwriters_stress * 2 * shortdelay_us)))
+		pr_alert("%s: delay = %lu jiffies.\n", __func__, jiffies - j);
+	}
+	if (!(torture_random(trsp) % (cxt.nrealwriters_stress * 200 * shortdelay_us)))
 		udelay(shortdelay_us);
 	if (!(torture_random(trsp) % (cxt.nrealwriters_stress * 20000)))
 		torture_preempt_schedule();  /* Allow test to be preempted. */
@@ -322,7 +319,7 @@ __acquires(torture_rwlock)
 static void torture_rwlock_write_delay(struct torture_random_state *trsp)
 {
 	const unsigned long shortdelay_us = 2;
-	const unsigned long longdelay_ms = 100;
+	const unsigned long longdelay_ms = long_hold ? long_hold : ULONG_MAX;
 
 	/* We want a short delay mostly to emulate likely code, and
 	 * we want a long delay occasionally to force massive contention.
@@ -455,14 +452,12 @@ __acquires(torture_mutex)
 
 static void torture_mutex_delay(struct torture_random_state *trsp)
 {
-	const unsigned long longdelay_ms = 100;
+	const unsigned long longdelay_ms = long_hold ? long_hold : ULONG_MAX;
 
 	/* We want a long delay occasionally to force massive contention.  */
 	if (!(torture_random(trsp) %
 	      (cxt.nrealwriters_stress * 2000 * longdelay_ms)))
 		mdelay(longdelay_ms * 5);
-	else
-		mdelay(longdelay_ms / 5);
 	if (!(torture_random(trsp) % (cxt.nrealwriters_stress * 20000)))
 		torture_preempt_schedule();  /* Allow test to be preempted. */
 }
@@ -630,7 +625,7 @@ __acquires(torture_rtmutex)
 static void torture_rtmutex_delay(struct torture_random_state *trsp)
 {
 	const unsigned long shortdelay_us = 2;
-	const unsigned long longdelay_ms = 100;
+	const unsigned long longdelay_ms = long_hold ? long_hold : ULONG_MAX;
 
 	/*
 	 * We want a short delay mostly to emulate likely code, and
@@ -640,7 +635,7 @@ static void torture_rtmutex_delay(struct torture_random_state *trsp)
 	      (cxt.nrealwriters_stress * 2000 * longdelay_ms)))
 		mdelay(longdelay_ms);
 	if (!(torture_random(trsp) %
-	      (cxt.nrealwriters_stress * 2 * shortdelay_us)))
+	      (cxt.nrealwriters_stress * 200 * shortdelay_us)))
 		udelay(shortdelay_us);
 	if (!(torture_random(trsp) % (cxt.nrealwriters_stress * 20000)))
 		torture_preempt_schedule();  /* Allow test to be preempted. */
@@ -695,14 +690,12 @@ __acquires(torture_rwsem)
 
 static void torture_rwsem_write_delay(struct torture_random_state *trsp)
 {
-	const unsigned long longdelay_ms = 100;
+	const unsigned long longdelay_ms = long_hold ? long_hold : ULONG_MAX;
 
 	/* We want a long delay occasionally to force massive contention.  */
 	if (!(torture_random(trsp) %
 	      (cxt.nrealwriters_stress * 2000 * longdelay_ms)))
 		mdelay(longdelay_ms * 10);
-	else
-		mdelay(longdelay_ms / 10);
 	if (!(torture_random(trsp) % (cxt.nrealwriters_stress * 20000)))
 		torture_preempt_schedule();  /* Allow test to be preempted. */
 }
@@ -848,8 +841,8 @@ static int lock_torture_writer(void *arg)
 
 			lwsp->n_lock_acquired++;
 		}
-		cxt.cur_ops->write_delay(&rand);
 		if (!skip_main_lock) {
+			cxt.cur_ops->write_delay(&rand);
 			lock_is_write_held = false;
 			WRITE_ONCE(last_lock_release, jiffies);
 			cxt.cur_ops->writeunlock(tid);

kernel/rcu/Kconfig

@@ -314,4 +314,22 @@ config RCU_LAZY
 	  To save power, batch RCU callbacks and flush after delay, memory
 	  pressure, or callback list growing too big.
 
+config RCU_DOUBLE_CHECK_CB_TIME
+	bool "RCU callback-batch backup time check"
+	depends on RCU_EXPERT
+	default n
+	help
+	  Use this option to provide more precise enforcement of the
+	  rcutree.rcu_resched_ns module parameter in situations where
+	  a single RCU callback might run for hundreds of microseconds,
+	  thus defeating the 32-callback batching used to amortize the
+	  cost of the fine-grained but expensive local_clock() function.
+
+	  This option rounds rcutree.rcu_resched_ns up to the next
+	  jiffy, and overrides the 32-callback batching if this limit
+	  is exceeded.
+
+	  Say Y here if you need tighter callback-limit enforcement.
+	  Say N here if you are unsure.
+
 endmenu # "RCU Subsystem"

kernel/rcu/rcu.h

@@ -642,4 +642,10 @@ void show_rcu_tasks_trace_gp_kthread(void);
 static inline void show_rcu_tasks_trace_gp_kthread(void) {}
 #endif
 
+#ifdef CONFIG_TINY_RCU
+static inline bool rcu_cpu_beenfullyonline(int cpu) { return true; }
+#else
+bool rcu_cpu_beenfullyonline(int cpu);
+#endif
+
 #endif /* __LINUX_RCU_H */

kernel/rcu/rcuscale.c

@@ -522,89 +522,6 @@ rcu_scale_print_module_parms(struct rcu_scale_ops *cur_ops, const char *tag)
 		 scale_type, tag, nrealreaders, nrealwriters, verbose, shutdown);
 }
 
-static void
-rcu_scale_cleanup(void)
-{
-	int i;
-	int j;
-	int ngps = 0;
-	u64 *wdp;
-	u64 *wdpp;
-
-	/*
-	 * Would like warning at start, but everything is expedited
-	 * during the mid-boot phase, so have to wait till the end.
-	 */
-	if (rcu_gp_is_expedited() && !rcu_gp_is_normal() && !gp_exp)
-		SCALEOUT_ERRSTRING("All grace periods expedited, no normal ones to measure!");
-	if (rcu_gp_is_normal() && gp_exp)
-		SCALEOUT_ERRSTRING("All grace periods normal, no expedited ones to measure!");
-	if (gp_exp && gp_async)
-		SCALEOUT_ERRSTRING("No expedited async GPs, so went with async!");
-
-	if (torture_cleanup_begin())
-		return;
-	if (!cur_ops) {
-		torture_cleanup_end();
-		return;
-	}
-
-	if (reader_tasks) {
-		for (i = 0; i < nrealreaders; i++)
-			torture_stop_kthread(rcu_scale_reader,
-					     reader_tasks[i]);
-		kfree(reader_tasks);
-	}
-
-	if (writer_tasks) {
-		for (i = 0; i < nrealwriters; i++) {
-			torture_stop_kthread(rcu_scale_writer,
-					     writer_tasks[i]);
-			if (!writer_n_durations)
-				continue;
-			j = writer_n_durations[i];
-			pr_alert("%s%s writer %d gps: %d\n",
-				 scale_type, SCALE_FLAG, i, j);
-			ngps += j;
-		}
-		pr_alert("%s%s start: %llu end: %llu duration: %llu gps: %d batches: %ld\n",
-			 scale_type, SCALE_FLAG,
-			 t_rcu_scale_writer_started, t_rcu_scale_writer_finished,
-			 t_rcu_scale_writer_finished -
-			 t_rcu_scale_writer_started,
-			 ngps,
-			 rcuscale_seq_diff(b_rcu_gp_test_finished,
-					   b_rcu_gp_test_started));
-		for (i = 0; i < nrealwriters; i++) {
-			if (!writer_durations)
-				break;
-			if (!writer_n_durations)
-				continue;
-			wdpp = writer_durations[i];
-			if (!wdpp)
-				continue;
-			for (j = 0; j < writer_n_durations[i]; j++) {
-				wdp = &wdpp[j];
-				pr_alert("%s%s %4d writer-duration: %5d %llu\n",
-					scale_type, SCALE_FLAG,
-					i, j, *wdp);
-				if (j % 100 == 0)
-					schedule_timeout_uninterruptible(1);
-			}
-			kfree(writer_durations[i]);
-		}
-		kfree(writer_tasks);
-		kfree(writer_durations);
-		kfree(writer_n_durations);
-	}
-
-	/* Do torture-type-specific cleanup operations.  */
-	if (cur_ops->cleanup != NULL)
-		cur_ops->cleanup();
-
-	torture_cleanup_end();
-}
-
 /*
  * Return the number if non-negative.  If -1, the number of CPUs.
  * If less than -1, that much less than the number of CPUs, but
@@ -624,20 +541,6 @@ static int compute_real(int n)
 	return nr;
 }
 
-/*
- * RCU scalability shutdown kthread.  Just waits to be awakened, then shuts
- * down system.
- */
-static int
-rcu_scale_shutdown(void *arg)
-{
-	wait_event_idle(shutdown_wq, atomic_read(&n_rcu_scale_writer_finished) >= nrealwriters);
-	smp_mb(); /* Wake before output. */
-	rcu_scale_cleanup();
-	kernel_power_off();
-	return -EINVAL;
-}
-
 /*
  * kfree_rcu() scalability tests: Start a kfree_rcu() loop on all CPUs for number
  * of iterations and measure total time and number of GP for all iterations to complete.
@@ -874,6 +777,108 @@ unwind:
 	return firsterr;
 }
 
+static void
+rcu_scale_cleanup(void)
+{
+	int i;
+	int j;
+	int ngps = 0;
+	u64 *wdp;
+	u64 *wdpp;
+
+	/*
+	 * Would like warning at start, but everything is expedited
+	 * during the mid-boot phase, so have to wait till the end.
+	 */
+	if (rcu_gp_is_expedited() && !rcu_gp_is_normal() && !gp_exp)
+		SCALEOUT_ERRSTRING("All grace periods expedited, no normal ones to measure!");
+	if (rcu_gp_is_normal() && gp_exp)
+		SCALEOUT_ERRSTRING("All grace periods normal, no expedited ones to measure!");
+	if (gp_exp && gp_async)
+		SCALEOUT_ERRSTRING("No expedited async GPs, so went with async!");
+
+	if (kfree_rcu_test) {
+		kfree_scale_cleanup();
+		return;
+	}
+
+	if (torture_cleanup_begin())
+		return;
+	if (!cur_ops) {
+		torture_cleanup_end();
+		return;
+	}
+
+	if (reader_tasks) {
+		for (i = 0; i < nrealreaders; i++)
+			torture_stop_kthread(rcu_scale_reader,
+					     reader_tasks[i]);
+		kfree(reader_tasks);
+	}
+
+	if (writer_tasks) {
+		for (i = 0; i < nrealwriters; i++) {
+			torture_stop_kthread(rcu_scale_writer,
+					     writer_tasks[i]);
+			if (!writer_n_durations)
+				continue;
+			j = writer_n_durations[i];
+			pr_alert("%s%s writer %d gps: %d\n",
+				 scale_type, SCALE_FLAG, i, j);
+			ngps += j;
+		}
+		pr_alert("%s%s start: %llu end: %llu duration: %llu gps: %d batches: %ld\n",
+			 scale_type, SCALE_FLAG,
+			 t_rcu_scale_writer_started, t_rcu_scale_writer_finished,
+			 t_rcu_scale_writer_finished -
+			 t_rcu_scale_writer_started,
+			 ngps,
+			 rcuscale_seq_diff(b_rcu_gp_test_finished,
+					   b_rcu_gp_test_started));
+		for (i = 0; i < nrealwriters; i++) {
+			if (!writer_durations)
+				break;
+			if (!writer_n_durations)
+				continue;
+			wdpp = writer_durations[i];
+			if (!wdpp)
+				continue;
+			for (j = 0; j < writer_n_durations[i]; j++) {
+				wdp = &wdpp[j];
+				pr_alert("%s%s %4d writer-duration: %5d %llu\n",
+					scale_type, SCALE_FLAG,
+					i, j, *wdp);
+				if (j % 100 == 0)
+					schedule_timeout_uninterruptible(1);
+			}
+			kfree(writer_durations[i]);
+		}
+		kfree(writer_tasks);
+		kfree(writer_durations);
+		kfree(writer_n_durations);
+	}
+
+	/* Do torture-type-specific cleanup operations.  */
+	if (cur_ops->cleanup != NULL)
+		cur_ops->cleanup();
+
+	torture_cleanup_end();
+}
+
+/*
+ * RCU scalability shutdown kthread.  Just waits to be awakened, then shuts
+ * down system.
+ */
+static int
+rcu_scale_shutdown(void *arg)
+{
+	wait_event_idle(shutdown_wq, atomic_read(&n_rcu_scale_writer_finished) >= nrealwriters);
+	smp_mb(); /* Wake before output. */
+	rcu_scale_cleanup();
+	kernel_power_off();
+	return -EINVAL;
+}
+
 static int __init
 rcu_scale_init(void)
 {

kernel/rcu/tasks.h

@@ -241,7 +241,6 @@ static void cblist_init_generic(struct rcu_tasks *rtp)
 	if (rcu_task_enqueue_lim < 0) {
 		rcu_task_enqueue_lim = 1;
 		rcu_task_cb_adjust = true;
-		pr_info("%s: Setting adjustable number of callback queues.\n", __func__);
 	} else if (rcu_task_enqueue_lim == 0) {
 		rcu_task_enqueue_lim = 1;
 	}
@@ -272,7 +271,9 @@ static void cblist_init_generic(struct rcu_tasks *rtp)
 		raw_spin_unlock_rcu_node(rtpcp); // irqs remain disabled.
 	}
 	raw_spin_unlock_irqrestore(&rtp->cbs_gbl_lock, flags);
-	pr_info("%s: Setting shift to %d and lim to %d.\n", __func__, data_race(rtp->percpu_enqueue_shift), data_race(rtp->percpu_enqueue_lim));
+
+	pr_info("%s: Setting shift to %d and lim to %d rcu_task_cb_adjust=%d.\n", rtp->name,
+			data_race(rtp->percpu_enqueue_shift), data_race(rtp->percpu_enqueue_lim), rcu_task_cb_adjust);
 }
 
 // IRQ-work handler that does deferred wakeup for call_rcu_tasks_generic().
@@ -463,6 +464,7 @@ static void rcu_tasks_invoke_cbs(struct rcu_tasks *rtp, struct rcu_tasks_percpu
 {
 	int cpu;
 	int cpunext;
+	int cpuwq;
 	unsigned long flags;
 	int len;
 	struct rcu_head *rhp;
@@ -473,11 +475,13 @@ static void rcu_tasks_invoke_cbs(struct rcu_tasks *rtp, struct rcu_tasks_percpu
 	cpunext = cpu * 2 + 1;
 	if (cpunext < smp_load_acquire(&rtp->percpu_dequeue_lim)) {
 		rtpcp_next = per_cpu_ptr(rtp->rtpcpu, cpunext);
-		queue_work_on(cpunext, system_wq, &rtpcp_next->rtp_work);
+		cpuwq = rcu_cpu_beenfullyonline(cpunext) ? cpunext : WORK_CPU_UNBOUND;
+		queue_work_on(cpuwq, system_wq, &rtpcp_next->rtp_work);
 		cpunext++;
 		if (cpunext < smp_load_acquire(&rtp->percpu_dequeue_lim)) {
 			rtpcp_next = per_cpu_ptr(rtp->rtpcpu, cpunext);
-			queue_work_on(cpunext, system_wq, &rtpcp_next->rtp_work);
+			cpuwq = rcu_cpu_beenfullyonline(cpunext) ? cpunext : WORK_CPU_UNBOUND;
+			queue_work_on(cpuwq, system_wq, &rtpcp_next->rtp_work);
 		}
 	}
 

kernel/rcu/tree.c

@@ -2046,19 +2046,35 @@ rcu_check_quiescent_state(struct rcu_data *rdp)
 	rcu_report_qs_rdp(rdp);
 }
 
+/* Return true if callback-invocation time limit exceeded. */
+static bool rcu_do_batch_check_time(long count, long tlimit,
+				    bool jlimit_check, unsigned long jlimit)
+{
+	// Invoke local_clock() only once per 32 consecutive callbacks.
+	return unlikely(tlimit) &&
+	       (!likely(count & 31) ||
+		(IS_ENABLED(CONFIG_RCU_DOUBLE_CHECK_CB_TIME) &&
+		 jlimit_check && time_after(jiffies, jlimit))) &&
+	       local_clock() >= tlimit;
+}
+
 /*
  * Invoke any RCU callbacks that have made it to the end of their grace
  * period.  Throttle as specified by rdp->blimit.
  */
 static void rcu_do_batch(struct rcu_data *rdp)
 {
+	long bl;
+	long count = 0;
 	int div;
 	bool __maybe_unused empty;
 	unsigned long flags;
-	struct rcu_head *rhp;
+	unsigned long jlimit;
+	bool jlimit_check = false;
+	long pending;
 	struct rcu_cblist rcl = RCU_CBLIST_INITIALIZER(rcl);
-	long bl, count = 0;
-	long pending, tlimit = 0;
+	struct rcu_head *rhp;
+	long tlimit = 0;
 
 	/* If no callbacks are ready, just return. */
 	if (!rcu_segcblist_ready_cbs(&rdp->cblist)) {
@@ -2082,11 +2098,15 @@ static void rcu_do_batch(struct rcu_data *rdp)
 	div = READ_ONCE(rcu_divisor);
 	div = div < 0 ? 7 : div > sizeof(long) * 8 - 2 ? sizeof(long) * 8 - 2 : div;
 	bl = max(rdp->blimit, pending >> div);
-	if (in_serving_softirq() && unlikely(bl > 100)) {
+	if ((in_serving_softirq() || rdp->rcu_cpu_kthread_status == RCU_KTHREAD_RUNNING) &&
+	    (IS_ENABLED(CONFIG_RCU_DOUBLE_CHECK_CB_TIME) || unlikely(bl > 100))) {
+		const long npj = NSEC_PER_SEC / HZ;
 		long rrn = READ_ONCE(rcu_resched_ns);
 
 		rrn = rrn < NSEC_PER_MSEC ? NSEC_PER_MSEC : rrn > NSEC_PER_SEC ? NSEC_PER_SEC : rrn;
 		tlimit = local_clock() + rrn;
+		jlimit = jiffies + (rrn + npj + 1) / npj;
+		jlimit_check = true;
 	}
 	trace_rcu_batch_start(rcu_state.name,
 			      rcu_segcblist_n_cbs(&rdp->cblist), bl);
@@ -2126,21 +2146,23 @@ static void rcu_do_batch(struct rcu_data *rdp)
 			 * Make sure we don't spend too much time here and deprive other
 			 * softirq vectors of CPU cycles.
 			 */
-			if (unlikely(tlimit)) {
-				/* only call local_clock() every 32 callbacks */
-				if (likely((count & 31) || local_clock() < tlimit))
-					continue;
-				/* Exceeded the time limit, so leave. */
+			if (rcu_do_batch_check_time(count, tlimit, jlimit_check, jlimit))
 				break;
-			}
 		} else {
-			// In rcuoc context, so no worries about depriving
-			// other softirq vectors of CPU cycles.
+			// In rcuc/rcuoc context, so no worries about
+			// depriving other softirq vectors of CPU cycles.
 			local_bh_enable();
 			lockdep_assert_irqs_enabled();
 			cond_resched_tasks_rcu_qs();
 			lockdep_assert_irqs_enabled();
 			local_bh_disable();
+			// But rcuc kthreads can delay quiescent-state
+			// reporting, so check time limits for them.
+			if (rdp->rcu_cpu_kthread_status == RCU_KTHREAD_RUNNING &&
+			    rcu_do_batch_check_time(count, tlimit, jlimit_check, jlimit)) {
+				rdp->rcu_cpu_has_work = 1;
+				break;
+			}
 		}
 	}
 
@@ -2459,12 +2481,12 @@ static void rcu_cpu_kthread(unsigned int cpu)
 		*statusp = RCU_KTHREAD_RUNNING;
 		local_irq_disable();
 		work = *workp;
-		*workp = 0;
+		WRITE_ONCE(*workp, 0);
 		local_irq_enable();
 		if (work)
 			rcu_core();
 		local_bh_enable();
-		if (*workp == 0) {
+		if (!READ_ONCE(*workp)) {
 			trace_rcu_utilization(TPS("End CPU kthread@rcu_wait"));
 			*statusp = RCU_KTHREAD_WAITING;
 			return;
@@ -2756,7 +2778,7 @@ EXPORT_SYMBOL_GPL(call_rcu);
  */
 struct kvfree_rcu_bulk_data {
 	struct list_head list;
-	unsigned long gp_snap;
+	struct rcu_gp_oldstate gp_snap;
 	unsigned long nr_records;
 	void *records[];
 };
@@ -2773,6 +2795,7 @@ struct kvfree_rcu_bulk_data {
  * struct kfree_rcu_cpu_work - single batch of kfree_rcu() requests
  * @rcu_work: Let queue_rcu_work() invoke workqueue handler after grace period
  * @head_free: List of kfree_rcu() objects waiting for a grace period
+ * @head_free_gp_snap: Grace-period snapshot to check for attempted premature frees.
  * @bulk_head_free: Bulk-List of kvfree_rcu() objects waiting for a grace period
  * @krcp: Pointer to @kfree_rcu_cpu structure
  */
@@ -2780,6 +2803,7 @@ struct kvfree_rcu_bulk_data {
 struct kfree_rcu_cpu_work {
 	struct rcu_work rcu_work;
 	struct rcu_head *head_free;
+	struct rcu_gp_oldstate head_free_gp_snap;
 	struct list_head bulk_head_free[FREE_N_CHANNELS];
 	struct kfree_rcu_cpu *krcp;
 };
@@ -2900,6 +2924,9 @@ drain_page_cache(struct kfree_rcu_cpu *krcp)
 	struct llist_node *page_list, *pos, *n;
 	int freed = 0;
 
+	if (!rcu_min_cached_objs)
+		return 0;
+
 	raw_spin_lock_irqsave(&krcp->lock, flags);
 	page_list = llist_del_all(&krcp->bkvcache);
 	WRITE_ONCE(krcp->nr_bkv_objs, 0);
@@ -2920,24 +2947,25 @@ kvfree_rcu_bulk(struct kfree_rcu_cpu *krcp,
 	unsigned long flags;
 	int i;
 
-	debug_rcu_bhead_unqueue(bnode);
+	if (!WARN_ON_ONCE(!poll_state_synchronize_rcu_full(&bnode->gp_snap))) {
+		debug_rcu_bhead_unqueue(bnode);
+		rcu_lock_acquire(&rcu_callback_map);
+		if (idx == 0) { // kmalloc() / kfree().
+			trace_rcu_invoke_kfree_bulk_callback(
+				rcu_state.name, bnode->nr_records,
+				bnode->records);
 
-	rcu_lock_acquire(&rcu_callback_map);
-	if (idx == 0) { // kmalloc() / kfree().
-		trace_rcu_invoke_kfree_bulk_callback(
-			rcu_state.name, bnode->nr_records,
-			bnode->records);
+			kfree_bulk(bnode->nr_records, bnode->records);
+		} else { // vmalloc() / vfree().
+			for (i = 0; i < bnode->nr_records; i++) {
+				trace_rcu_invoke_kvfree_callback(
+					rcu_state.name, bnode->records[i], 0);
 
-		kfree_bulk(bnode->nr_records, bnode->records);
-	} else { // vmalloc() / vfree().
-		for (i = 0; i < bnode->nr_records; i++) {
-			trace_rcu_invoke_kvfree_callback(
-				rcu_state.name, bnode->records[i], 0);
-
-			vfree(bnode->records[i]);
+				vfree(bnode->records[i]);
+			}
 		}
+		rcu_lock_release(&rcu_callback_map);
 	}
-	rcu_lock_release(&rcu_callback_map);
 
 	raw_spin_lock_irqsave(&krcp->lock, flags);
 	if (put_cached_bnode(krcp, bnode))
@@ -2984,6 +3012,7 @@ static void kfree_rcu_work(struct work_struct *work)
 	struct rcu_head *head;
 	struct kfree_rcu_cpu *krcp;
 	struct kfree_rcu_cpu_work *krwp;
+	struct rcu_gp_oldstate head_gp_snap;
 	int i;
 
 	krwp = container_of(to_rcu_work(work),
@@ -2998,6 +3027,7 @@ static void kfree_rcu_work(struct work_struct *work)
 	// Channel 3.
 	head = krwp->head_free;
 	krwp->head_free = NULL;
+	head_gp_snap = krwp->head_free_gp_snap;
 	raw_spin_unlock_irqrestore(&krcp->lock, flags);
 
 	// Handle the first two channels.
@@ -3014,7 +3044,8 @@ static void kfree_rcu_work(struct work_struct *work)
 	 * queued on a linked list through their rcu_head structures.
 	 * This list is named "Channel 3".
 	 */
-	kvfree_rcu_list(head);
+	if (head && !WARN_ON_ONCE(!poll_state_synchronize_rcu_full(&head_gp_snap)))
+		kvfree_rcu_list(head);
 }
 
 static bool
@@ -3081,7 +3112,7 @@ kvfree_rcu_drain_ready(struct kfree_rcu_cpu *krcp)
 		INIT_LIST_HEAD(&bulk_ready[i]);
 
 		list_for_each_entry_safe_reverse(bnode, n, &krcp->bulk_head[i], list) {
-			if (!poll_state_synchronize_rcu(bnode->gp_snap))
+			if (!poll_state_synchronize_rcu_full(&bnode->gp_snap))
 				break;
 
 			atomic_sub(bnode->nr_records, &krcp->bulk_count[i]);
@@ -3146,6 +3177,7 @@ static void kfree_rcu_monitor(struct work_struct *work)
 			// objects queued on the linked list.
 			if (!krwp->head_free) {
 				krwp->head_free = krcp->head;
+				get_state_synchronize_rcu_full(&krwp->head_free_gp_snap);
 				atomic_set(&krcp->head_count, 0);
 				WRITE_ONCE(krcp->head, NULL);
 			}
@@ -3194,7 +3226,7 @@ static void fill_page_cache_func(struct work_struct *work)
 	nr_pages = atomic_read(&krcp->backoff_page_cache_fill) ?
 		1 : rcu_min_cached_objs;
 
-	for (i = 0; i < nr_pages; i++) {
+	for (i = READ_ONCE(krcp->nr_bkv_objs); i < nr_pages; i++) {
 		bnode = (struct kvfree_rcu_bulk_data *)
 			__get_free_page(GFP_KERNEL | __GFP_NORETRY | __GFP_NOMEMALLOC | __GFP_NOWARN);
 
@@ -3218,6 +3250,10 @@ static void fill_page_cache_func(struct work_struct *work)
 static void
 run_page_cache_worker(struct kfree_rcu_cpu *krcp)
 {
+	// If cache disabled, bail out.
+	if (!rcu_min_cached_objs)
+		return;
+
 	if (rcu_scheduler_active == RCU_SCHEDULER_RUNNING &&
 			!atomic_xchg(&krcp->work_in_progress, 1)) {
 		if (atomic_read(&krcp->backoff_page_cache_fill)) {
@@ -3272,7 +3308,7 @@ add_ptr_to_bulk_krc_lock(struct kfree_rcu_cpu **krcp,
 			// scenarios.
 			bnode = (struct kvfree_rcu_bulk_data *)
 				__get_free_page(GFP_KERNEL | __GFP_NORETRY | __GFP_NOMEMALLOC | __GFP_NOWARN);
-			*krcp = krc_this_cpu_lock(flags);
+			raw_spin_lock_irqsave(&(*krcp)->lock, *flags);
 		}
 
 		if (!bnode)
@@ -3285,7 +3321,7 @@ add_ptr_to_bulk_krc_lock(struct kfree_rcu_cpu **krcp,
 
 	// Finally insert and update the GP for this page.
 	bnode->records[bnode->nr_records++] = ptr;
-	bnode->gp_snap = get_state_synchronize_rcu();
+	get_state_synchronize_rcu_full(&bnode->gp_snap);
 	atomic_inc(&(*krcp)->bulk_count[idx]);
 
 	return true;
@@ -4283,7 +4319,6 @@ int rcutree_prepare_cpu(unsigned int cpu)
 	 */
 	rnp = rdp->mynode;
 	raw_spin_lock_rcu_node(rnp);		/* irqs already disabled. */
-	rdp->beenonline = true;	 /* We have now been online. */
 	rdp->gp_seq = READ_ONCE(rnp->gp_seq);
 	rdp->gp_seq_needed = rdp->gp_seq;
 	rdp->cpu_no_qs.b.norm = true;
@@ -4310,6 +4345,16 @@ static void rcutree_affinity_setting(unsigned int cpu, int outgoing)
 	rcu_boost_kthread_setaffinity(rdp->mynode, outgoing);
 }
 
+/*
+ * Has the specified (known valid) CPU ever been fully online?
+ */
+bool rcu_cpu_beenfullyonline(int cpu)
+{
+	struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
+
+	return smp_load_acquire(&rdp->beenonline);
+}
+
 /*
  * Near the end of the CPU-online process.  Pretty much all services
  * enabled, and the CPU is now very much alive.
@@ -4368,15 +4413,16 @@ int rcutree_offline_cpu(unsigned int cpu)
  * Note that this function is special in that it is invoked directly
  * from the incoming CPU rather than from the cpuhp_step mechanism.
  * This is because this function must be invoked at a precise location.
+ * This incoming CPU must not have enabled interrupts yet.
  */
 void rcu_cpu_starting(unsigned int cpu)
 {
-	unsigned long flags;
 	unsigned long mask;
 	struct rcu_data *rdp;
 	struct rcu_node *rnp;
 	bool newcpu;
 
+	lockdep_assert_irqs_disabled();
 	rdp = per_cpu_ptr(&rcu_data, cpu);
 	if (rdp->cpu_started)
 		return;
@@ -4384,7 +4430,6 @@ void rcu_cpu_starting(unsigned int cpu)
 
 	rnp = rdp->mynode;
 	mask = rdp->grpmask;
-	local_irq_save(flags);
 	arch_spin_lock(&rcu_state.ofl_lock);
 	rcu_dynticks_eqs_online();
 	raw_spin_lock(&rcu_state.barrier_lock);
@@ -4403,17 +4448,17 @@ void rcu_cpu_starting(unsigned int cpu)
 	/* An incoming CPU should never be blocking a grace period. */
 	if (WARN_ON_ONCE(rnp->qsmask & mask)) { /* RCU waiting on incoming CPU? */
 		/* rcu_report_qs_rnp() *really* wants some flags to restore */
-		unsigned long flags2;
+		unsigned long flags;
 
-		local_irq_save(flags2);
+		local_irq_save(flags);
 		rcu_disable_urgency_upon_qs(rdp);
 		/* Report QS -after- changing ->qsmaskinitnext! */
-		rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags2);
+		rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
 	} else {
 		raw_spin_unlock_rcu_node(rnp);
 	}
 	arch_spin_unlock(&rcu_state.ofl_lock);
-	local_irq_restore(flags);
+	smp_store_release(&rdp->beenonline, true);
 	smp_mb(); /* Ensure RCU read-side usage follows above initialization. */
 }
 

kernel/rcu/tree_exp.h

@@ -643,7 +643,7 @@ static void synchronize_rcu_expedited_wait(void)
 					"O."[!!cpu_online(cpu)],
 					"o."[!!(rdp->grpmask & rnp->expmaskinit)],
 					"N."[!!(rdp->grpmask & rnp->expmaskinitnext)],
-					"D."[!!(rdp->cpu_no_qs.b.exp)]);
+					"D."[!!data_race(rdp->cpu_no_qs.b.exp)]);
 			}
 		}
 		pr_cont(" } %lu jiffies s: %lu root: %#lx/%c\n",

kernel/rcu/tree_nocb.h

@@ -1319,13 +1319,22 @@ lazy_rcu_shrink_count(struct shrinker *shrink, struct shrink_control *sc)
 	int cpu;
 	unsigned long count = 0;
 
+	if (WARN_ON_ONCE(!cpumask_available(rcu_nocb_mask)))
+		return 0;
+
+	/*  Protect rcu_nocb_mask against concurrent (de-)offloading. */
+	if (!mutex_trylock(&rcu_state.barrier_mutex))
+		return 0;
+
 	/* Snapshot count of all CPUs */
-	for_each_possible_cpu(cpu) {
+	for_each_cpu(cpu, rcu_nocb_mask) {
 		struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
 
 		count +=  READ_ONCE(rdp->lazy_len);
 	}
 
+	mutex_unlock(&rcu_state.barrier_mutex);
+
 	return count ? count : SHRINK_EMPTY;
 }
 
@@ -1336,15 +1345,45 @@ lazy_rcu_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 	unsigned long flags;
 	unsigned long count = 0;
 
-	/* Snapshot count of all CPUs */
-	for_each_possible_cpu(cpu) {
-		struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
-		int _count = READ_ONCE(rdp->lazy_len);
+	if (WARN_ON_ONCE(!cpumask_available(rcu_nocb_mask)))
+		return 0;
+	/*
+	 * Protect against concurrent (de-)offloading. Otherwise nocb locking
+	 * may be ignored or imbalanced.
+	 */
+	if (!mutex_trylock(&rcu_state.barrier_mutex)) {
+		/*
+		 * But really don't insist if barrier_mutex is contended since we
+		 * can't guarantee that it will never engage in a dependency
+		 * chain involving memory allocation. The lock is seldom contended
+		 * anyway.
+		 */
+		return 0;
+	}
 
-		if (_count == 0)
+	/* Snapshot count of all CPUs */
+	for_each_cpu(cpu, rcu_nocb_mask) {
+		struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
+		int _count;
+
+		if (WARN_ON_ONCE(!rcu_rdp_is_offloaded(rdp)))
 			continue;
+
+		if (!READ_ONCE(rdp->lazy_len))
+			continue;
+
 		rcu_nocb_lock_irqsave(rdp, flags);
-		WRITE_ONCE(rdp->lazy_len, 0);
+		/*
+		 * Recheck under the nocb lock. Since we are not holding the bypass
+		 * lock we may still race with increments from the enqueuer but still
+		 * we know for sure if there is at least one lazy callback.
+		 */
+		_count = READ_ONCE(rdp->lazy_len);
+		if (!_count) {
+			rcu_nocb_unlock_irqrestore(rdp, flags);
+			continue;
+		}
+		WARN_ON_ONCE(!rcu_nocb_flush_bypass(rdp, NULL, jiffies, false));
 		rcu_nocb_unlock_irqrestore(rdp, flags);
 		wake_nocb_gp(rdp, false);
 		sc->nr_to_scan -= _count;
@@ -1352,6 +1391,9 @@ lazy_rcu_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 		if (sc->nr_to_scan <= 0)
 			break;
 	}
+
+	mutex_unlock(&rcu_state.barrier_mutex);
+
 	return count ? count : SHRINK_STOP;
 }
 

kernel/rcu/tree_plugin.h

@@ -257,6 +257,8 @@ static void rcu_preempt_ctxt_queue(struct rcu_node *rnp, struct rcu_data *rdp)
 	 * GP should not be able to end until we report, so there should be
 	 * no need to check for a subsequent expedited GP.  (Though we are
 	 * still in a quiescent state in any case.)
+	 *
+	 * Interrupts are disabled, so ->cpu_no_qs.b.exp cannot change.
 	 */
 	if (blkd_state & RCU_EXP_BLKD && rdp->cpu_no_qs.b.exp)
 		rcu_report_exp_rdp(rdp);
@@ -941,7 +943,7 @@ notrace void rcu_preempt_deferred_qs(struct task_struct *t)
 {
 	struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
 
-	if (rdp->cpu_no_qs.b.exp)
+	if (READ_ONCE(rdp->cpu_no_qs.b.exp))
 		rcu_report_exp_rdp(rdp);
 }