Merge tag 'rcu.2022.12.02a' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu

Pull RCU updates from Paul McKenney:

 - Documentation updates. This is the second in a series from an ongoing
   review of the RCU documentation.

 - Miscellaneous fixes.

 - Introduce a default-off Kconfig option that depends on RCU_NOCB_CPU
   that, on CPUs mentioned in the nohz_full or rcu_nocbs boot-argument
   CPU lists, causes call_rcu() to introduce delays.

   These delays result in significant power savings on nearly idle
   Android and ChromeOS systems. These savings range from a few percent
   to more than ten percent.

   This series also includes several commits that change call_rcu() to a
   new call_rcu_hurry() function that avoids these delays in a few
   cases, for example, where timely wakeups are required. Several of
   these are outside of RCU and thus have acks and reviews from the
   relevant maintainers.

 - Create an srcu_read_lock_nmisafe() and an srcu_read_unlock_nmisafe()
   for architectures that support NMIs, but which do not provide
   NMI-safe this_cpu_inc(). These NMI-safe SRCU functions are required
   by the upcoming lockless printk() work by John Ogness et al.

 - Changes providing minor but important increases in torture test
   coverage for the new RCU polled-grace-period APIs.

 - Changes to torturescript that avoid redundant kernel builds, thus
   providing about a 30% speedup for the torture.sh acceptance test.

* tag 'rcu.2022.12.02a' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu: (49 commits)
  net: devinet: Reduce refcount before grace period
  net: Use call_rcu_hurry() for dst_release()
  workqueue: Make queue_rcu_work() use call_rcu_hurry()
  percpu-refcount: Use call_rcu_hurry() for atomic switch
  scsi/scsi_error: Use call_rcu_hurry() instead of call_rcu()
  rcu/rcutorture: Use call_rcu_hurry() where needed
  rcu/rcuscale: Use call_rcu_hurry() for async reader test
  rcu/sync: Use call_rcu_hurry() instead of call_rcu
  rcuscale: Add laziness and kfree tests
  rcu: Shrinker for lazy rcu
  rcu: Refactor code a bit in rcu_nocb_do_flush_bypass()
  rcu: Make call_rcu() lazy to save power
  rcu: Implement lockdep_rcu_enabled for !CONFIG_DEBUG_LOCK_ALLOC
  srcu: Debug NMI safety even on archs that don't require it
  srcu: Explain the reason behind the read side critical section on GP start
  srcu: Warn when NMI-unsafe API is used in NMI
  arch/s390: Add ARCH_HAS_NMI_SAFE_THIS_CPU_OPS Kconfig option
  arch/loongarch: Add ARCH_HAS_NMI_SAFE_THIS_CPU_OPS Kconfig option
  rcu: Fix __this_cpu_read() lockdep warning in rcu_force_quiescent_state()
  rcu-tasks: Make grace-period-age message human-readable
  ...

7 files changed, 94 insertions, 9 deletions

diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 637a60607c7d..915142abdf76 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -416,7 +416,7 @@ static __always_inline void guest_context_enter_irqoff(void)
 	 */
 	if (!context_tracking_guest_enter()) {
 		instrumentation_begin();
-		rcu_virt_note_context_switch(smp_processor_id());
+		rcu_virt_note_context_switch();
 		instrumentation_end();
 	}
 }
diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index 08605ce7379d..4da98ca6273e 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -108,6 +108,15 @@ static inline int rcu_preempt_depth(void)
 
 #endif /* #else #ifdef CONFIG_PREEMPT_RCU */
 
+#ifdef CONFIG_RCU_LAZY
+void call_rcu_hurry(struct rcu_head *head, rcu_callback_t func);
+#else
+static inline void call_rcu_hurry(struct rcu_head *head, rcu_callback_t func)
+{
+	call_rcu(head, func);
+}
+#endif
+
 /* Internal to kernel */
 void rcu_init(void);
 extern int rcu_scheduler_active;
@@ -340,6 +349,11 @@ static inline int rcu_read_lock_any_held(void)
 	return !preemptible();
 }
 
+static inline int debug_lockdep_rcu_enabled(void)
+{
+	return 0;
+}
+
 #endif /* #else #ifdef CONFIG_DEBUG_LOCK_ALLOC */
 
 #ifdef CONFIG_PROVE_RCU
diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h
index 768196a5f39d..5c271bf3a1e7 100644
--- a/include/linux/rcutiny.h
+++ b/include/linux/rcutiny.h
@@ -142,12 +142,10 @@ static inline int rcu_needs_cpu(void)
  * Take advantage of the fact that there is only one CPU, which
  * allows us to ignore virtualization-based context switches.
  */
-static inline void rcu_virt_note_context_switch(int cpu) { }
+static inline void rcu_virt_note_context_switch(void) { }
 static inline void rcu_cpu_stall_reset(void) { }
 static inline int rcu_jiffies_till_stall_check(void) { return 21 * HZ; }
 static inline void rcu_irq_exit_check_preempt(void) { }
-#define rcu_is_idle_cpu(cpu) \
-	(is_idle_task(current) && !in_nmi() && !in_hardirq() && !in_serving_softirq())
 static inline void exit_rcu(void) { }
 static inline bool rcu_preempt_need_deferred_qs(struct task_struct *t)
 {
diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h
index 5efb51486e8a..4003bf6cfa1c 100644
--- a/include/linux/rcutree.h
+++ b/include/linux/rcutree.h
@@ -27,7 +27,7 @@ void rcu_cpu_stall_reset(void);
  * wrapper around rcu_note_context_switch(), which allows TINY_RCU
  * to save a few bytes. The caller must have disabled interrupts.
  */
-static inline void rcu_virt_note_context_switch(int cpu)
+static inline void rcu_virt_note_context_switch(void)
 {
 	rcu_note_context_switch(false);
 }
@@ -87,8 +87,6 @@ bool poll_state_synchronize_rcu_full(struct rcu_gp_oldstate *rgosp);
 void cond_synchronize_rcu(unsigned long oldstate);
 void cond_synchronize_rcu_full(struct rcu_gp_oldstate *rgosp);
 
-bool rcu_is_idle_cpu(int cpu);
-
 #ifdef CONFIG_PROVE_RCU
 void rcu_irq_exit_check_preempt(void);
 #else
diff --git a/include/linux/slab.h b/include/linux/slab.h
index 45efc6c553b8..e7534a7b2daf 100644
--- a/include/linux/slab.h
+++ b/include/linux/slab.h
@@ -76,6 +76,17 @@
  * rcu_read_lock before reading the address, then rcu_read_unlock after
  * taking the spinlock within the structure expected at that address.
  *
+ * Note that it is not possible to acquire a lock within a structure
+ * allocated with SLAB_TYPESAFE_BY_RCU without first acquiring a reference
+ * as described above.  The reason is that SLAB_TYPESAFE_BY_RCU pages
+ * are not zeroed before being given to the slab, which means that any
+ * locks must be initialized after each and every kmem_struct_alloc().
+ * Alternatively, make the ctor passed to kmem_cache_create() initialize
+ * the locks at page-allocation time, as is done in __i915_request_ctor(),
+ * sighand_ctor(), and anon_vma_ctor().  Such a ctor permits readers
+ * to safely acquire those ctor-initialized locks under rcu_read_lock()
+ * protection.
+ *
  * Note that SLAB_TYPESAFE_BY_RCU was originally named SLAB_DESTROY_BY_RCU.
  */
 /* Defer freeing slabs to RCU */
diff --git a/include/linux/srcu.h b/include/linux/srcu.h
index 01226e4d960a..f0814ffca34b 100644
--- a/include/linux/srcu.h
+++ b/include/linux/srcu.h
@@ -64,6 +64,20 @@ unsigned long get_state_synchronize_srcu(struct srcu_struct *ssp);
 unsigned long start_poll_synchronize_srcu(struct srcu_struct *ssp);
 bool poll_state_synchronize_srcu(struct srcu_struct *ssp, unsigned long cookie);
 
+#ifdef CONFIG_NEED_SRCU_NMI_SAFE
+int __srcu_read_lock_nmisafe(struct srcu_struct *ssp) __acquires(ssp);
+void __srcu_read_unlock_nmisafe(struct srcu_struct *ssp, int idx) __releases(ssp);
+#else
+static inline int __srcu_read_lock_nmisafe(struct srcu_struct *ssp)
+{
+	return __srcu_read_lock(ssp);
+}
+static inline void __srcu_read_unlock_nmisafe(struct srcu_struct *ssp, int idx)
+{
+	__srcu_read_unlock(ssp, idx);
+}
+#endif /* CONFIG_NEED_SRCU_NMI_SAFE */
+
 #ifdef CONFIG_SRCU
 void srcu_init(void);
 #else /* #ifdef CONFIG_SRCU */
@@ -104,6 +118,18 @@ static inline int srcu_read_lock_held(const struct srcu_struct *ssp)
 
 #endif /* #else #ifdef CONFIG_DEBUG_LOCK_ALLOC */
 
+#define SRCU_NMI_UNKNOWN	0x0
+#define SRCU_NMI_UNSAFE		0x1
+#define SRCU_NMI_SAFE		0x2
+
+#if defined(CONFIG_PROVE_RCU) && defined(CONFIG_TREE_SRCU)
+void srcu_check_nmi_safety(struct srcu_struct *ssp, bool nmi_safe);
+#else
+static inline void srcu_check_nmi_safety(struct srcu_struct *ssp,
+					 bool nmi_safe) { }
+#endif
+
+
 /**
  * srcu_dereference_check - fetch SRCU-protected pointer for later dereferencing
  * @p: the pointer to fetch and protect for later dereferencing
@@ -161,17 +187,36 @@ static inline int srcu_read_lock(struct srcu_struct *ssp) __acquires(ssp)
 {
 	int retval;
 
+	srcu_check_nmi_safety(ssp, false);
 	retval = __srcu_read_lock(ssp);
 	rcu_lock_acquire(&(ssp)->dep_map);
 	return retval;
 }
 
+/**
+ * srcu_read_lock_nmisafe - register a new reader for an SRCU-protected structure.
+ * @ssp: srcu_struct in which to register the new reader.
+ *
+ * Enter an SRCU read-side critical section, but in an NMI-safe manner.
+ * See srcu_read_lock() for more information.
+ */
+static inline int srcu_read_lock_nmisafe(struct srcu_struct *ssp) __acquires(ssp)
+{
+	int retval;
+
+	srcu_check_nmi_safety(ssp, true);
+	retval = __srcu_read_lock_nmisafe(ssp);
+	rcu_lock_acquire(&(ssp)->dep_map);
+	return retval;
+}
+
 /* Used by tracing, cannot be traced and cannot invoke lockdep. */
 static inline notrace int
 srcu_read_lock_notrace(struct srcu_struct *ssp) __acquires(ssp)
 {
 	int retval;
 
+	srcu_check_nmi_safety(ssp, false);
 	retval = __srcu_read_lock(ssp);
 	return retval;
 }
@@ -187,14 +232,32 @@ static inline void srcu_read_unlock(struct srcu_struct *ssp, int idx)
 	__releases(ssp)
 {
 	WARN_ON_ONCE(idx & ~0x1);
+	srcu_check_nmi_safety(ssp, false);
 	rcu_lock_release(&(ssp)->dep_map);
 	__srcu_read_unlock(ssp, idx);
 }
 
+/**
+ * srcu_read_unlock_nmisafe - unregister a old reader from an SRCU-protected structure.
+ * @ssp: srcu_struct in which to unregister the old reader.
+ * @idx: return value from corresponding srcu_read_lock().
+ *
+ * Exit an SRCU read-side critical section, but in an NMI-safe manner.
+ */
+static inline void srcu_read_unlock_nmisafe(struct srcu_struct *ssp, int idx)
+	__releases(ssp)
+{
+	WARN_ON_ONCE(idx & ~0x1);
+	srcu_check_nmi_safety(ssp, true);
+	rcu_lock_release(&(ssp)->dep_map);
+	__srcu_read_unlock_nmisafe(ssp, idx);
+}
+
 /* Used by tracing, cannot be traced and cannot call lockdep. */
 static inline notrace void
 srcu_read_unlock_notrace(struct srcu_struct *ssp, int idx) __releases(ssp)
 {
+	srcu_check_nmi_safety(ssp, false);
 	__srcu_read_unlock(ssp, idx);
 }
 
diff --git a/include/linux/srcutree.h b/include/linux/srcutree.h
index e3014319d1ad..c689a81752c9 100644
--- a/include/linux/srcutree.h
+++ b/include/linux/srcutree.h
@@ -23,8 +23,9 @@ struct srcu_struct;
  */
 struct srcu_data {
 	/* Read-side state. */
-	unsigned long srcu_lock_count[2];	/* Locks per CPU. */
-	unsigned long srcu_unlock_count[2];	/* Unlocks per CPU. */
+	atomic_long_t srcu_lock_count[2];	/* Locks per CPU. */
+	atomic_long_t srcu_unlock_count[2];	/* Unlocks per CPU. */
+	int srcu_nmi_safety;			/* NMI-safe srcu_struct structure? */
 
 	/* Update-side state. */
 	spinlock_t __private lock ____cacheline_internodealigned_in_smp;