6 files changed, 82 insertions, 55 deletions

diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index 6fe0363724e9..70331218e4b4 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -196,6 +196,8 @@ static inline void rcu_read_lock_sched(void)
         __acquire(RCU_SCHED);
         rcu_read_acquire();
 }
+
+/* Used by lockdep and tracing: cannot be traced, cannot call lockdep. */
 static inline notrace void rcu_read_lock_sched_notrace(void)
 {
         preempt_disable_notrace();
@@ -213,6 +215,8 @@ static inline void rcu_read_unlock_sched(void)
         __release(RCU_SCHED);
         preempt_enable();
 }
+
+/* Used by lockdep and tracing: cannot be traced, cannot call lockdep. */
 static inline notrace void rcu_read_unlock_sched_notrace(void)
 {
         __release(RCU_SCHED);
diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h
index 88109c87f29c..19a3b06943e0 100644
--- a/include/linux/rcutree.h
+++ b/include/linux/rcutree.h
@@ -90,10 +90,6 @@ extern long rcu_batches_completed(void);
 extern long rcu_batches_completed_bh(void);
 extern long rcu_batches_completed_sched(void);
 
-static inline void rcu_init_sched(void)
-{
-}
-
 #ifdef CONFIG_NO_HZ
 void rcu_enter_nohz(void);
 void rcu_exit_nohz(void);
@@ -106,7 +102,7 @@ static inline void rcu_exit_nohz(void)
 }
 #endif /* CONFIG_NO_HZ */
 
-/* A context switch is a grace period for rcutree. */
+/* A context switch is a grace period for RCU-sched and RCU-bh. */
 static inline int rcu_blocking_is_gp(void)
 {
         return num_online_cpus() == 1;
diff --git a/init/main.c b/init/main.c
index 34971becbd3c..833d675677d1 100644
--- a/init/main.c
+++ b/init/main.c
@@ -782,7 +782,6 @@ static void __init do_initcalls(void)
  */
 static void __init do_basic_setup(void)
 {
-        rcu_init_sched(); /* needed by module_init stage. */
         init_workqueues();
         cpuset_init_smp();
         usermodehelper_init();
diff --git a/kernel/rcutree.c b/kernel/rcutree.c
index f85b6842d1e1..53a5ef0ca911 100644
--- a/kernel/rcutree.c
+++ b/kernel/rcutree.c
@@ -81,24 +81,29 @@ DEFINE_PER_CPU(struct rcu_data, rcu_sched_data);
 struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh_state);
 DEFINE_PER_CPU(struct rcu_data, rcu_bh_data);
 
-extern long rcu_batches_completed_sched(void);
-static struct rcu_node *rcu_get_root(struct rcu_state *rsp);
-static void cpu_quiet_msk(unsigned long mask, struct rcu_state *rsp,
-                          struct rcu_node *rnp, unsigned long flags);
-static void cpu_quiet_msk_finish(struct rcu_state *rsp, unsigned long flags);
+/* Forward declarations for rcutree_plugin.h */
+static inline void rcu_bootup_announce(void);
+long rcu_batches_completed(void);
+static void rcu_preempt_note_context_switch(int cpu);
+static int rcu_preempted_readers(struct rcu_node *rnp);
+#ifdef CONFIG_RCU_CPU_STALL_DETECTOR
+static void rcu_print_task_stall(struct rcu_node *rnp);
+#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
+static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp);
 #ifdef CONFIG_HOTPLUG_CPU
-static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp);
+static void rcu_preempt_offline_tasks(struct rcu_state *rsp,
+                                      struct rcu_node *rnp,
+                                      struct rcu_data *rdp);
+static void rcu_preempt_offline_cpu(int cpu);
 #endif /* #ifdef CONFIG_HOTPLUG_CPU */
-static void __rcu_process_callbacks(struct rcu_state *rsp,
-                                    struct rcu_data *rdp);
-static void __call_rcu(struct rcu_head *head,
-                       void (*func)(struct rcu_head *rcu),
-                       struct rcu_state *rsp);
-static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp);
-static void __cpuinit rcu_init_percpu_data(int cpu, struct rcu_state *rsp,
-                                           int preemptable);
+static void rcu_preempt_check_callbacks(int cpu);
+static void rcu_preempt_process_callbacks(void);
+void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu));
+static int rcu_preempt_pending(int cpu);
+static int rcu_preempt_needs_cpu(int cpu);
+static void __cpuinit rcu_preempt_init_percpu_data(int cpu);
+static void __init __rcu_init_preempt(void);
 
-#include "rcutree_plugin.h"
 
 /*
  * Return true if an RCU grace period is in progress.  The ACCESS_ONCE()s
@@ -377,7 +382,7 @@ static long dyntick_recall_completed(struct rcu_state *rsp)
 /*
  * Snapshot the specified CPU's dynticks counter so that we can later
  * credit them with an implicit quiescent state.  Return 1 if this CPU
- * is already in a quiescent state courtesy of dynticks idle mode.
+ * is in dynticks idle mode, which is an extended quiescent state.
  */
 static int dyntick_save_progress_counter(struct rcu_data *rdp)
 {
@@ -624,9 +629,15 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags)
         note_new_gpnum(rsp, rdp);
 
         /*
-         * Because we are first, we know that all our callbacks will
-         * be covered by this upcoming grace period, even the ones
-         * that were registered arbitrarily recently.
+         * Because this CPU just now started the new grace period, we know
+         * that all of its callbacks will be covered by this upcoming grace
+         * period, even the ones that were registered arbitrarily recently.
+         * Therefore, advance all outstanding callbacks to RCU_WAIT_TAIL.
+         *
+         * Other CPUs cannot be sure exactly when the grace period started.
+         * Therefore, their recently registered callbacks must pass through
+         * an additional RCU_NEXT_READY stage, so that they will be handled
+         * by the next RCU grace period.
          */
         rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL];
         rdp->nxttail[RCU_WAIT_TAIL] = rdp->nxttail[RCU_NEXT_TAIL];
@@ -886,7 +897,7 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp)
 
         /*
          * Move callbacks from the outgoing CPU to the running CPU.
-         * Note that the outgoing CPU is now quiscent, so it is now
+         * Note that the outgoing CPU is now quiescent, so it is now
          * (uncharacteristically) safe to access its rcu_data structure.
          * Note also that we must carefully retain the order of the
          * outgoing CPU's callbacks in order for rcu_barrier() to work
@@ -1577,25 +1588,6 @@ do { \
         } \
 } while (0)
 
-#ifdef CONFIG_TREE_PREEMPT_RCU
-
-void __init __rcu_init_preempt(void)
-{
-        int i;                  /* All used by RCU_INIT_FLAVOR(). */
-        int j;
-        struct rcu_node *rnp;
-
-        RCU_INIT_FLAVOR(&rcu_preempt_state, rcu_preempt_data);
-}
-
-#else /* #ifdef CONFIG_TREE_PREEMPT_RCU */
-
-void __init __rcu_init_preempt(void)
-{
-}
-
-#endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */
-
 void __init __rcu_init(void)
 {
         int i;                  /* All used by RCU_INIT_FLAVOR(). */
@@ -1612,6 +1604,8 @@ void __init __rcu_init(void)
         open_softirq(RCU_SOFTIRQ, rcu_process_callbacks);
 }
 
+#include "rcutree_plugin.h"
+
 module_param(blimit, int, 0);
 module_param(qhimark, int, 0);
 module_param(qlowmark, int, 0);
diff --git a/kernel/rcutree.h b/kernel/rcutree.h
index 9aa8c8a160d8..a48d11f37b4c 100644
--- a/kernel/rcutree.h
+++ b/kernel/rcutree.h
@@ -79,15 +79,21 @@ struct rcu_dynticks {
  * Definition for node within the RCU grace-period-detection hierarchy.
  */
 struct rcu_node {
-        spinlock_t lock;
+        spinlock_t lock;        /* Root rcu_node's lock protects some */
+                                /*  rcu_state fields as well as following. */
         long    gpnum;          /* Current grace period for this node. */
                                 /*  This will either be equal to or one */
                                 /*  behind the root rcu_node's gpnum. */
         unsigned long qsmask;   /* CPUs or groups that need to switch in */
                                 /*  order for current grace period to proceed.*/
+                                /*  In leaf rcu_node, each bit corresponds to */
+                                /*  an rcu_data structure, otherwise, each */
+                                /*  bit corresponds to a child rcu_node */
+                                /*  structure. */
         unsigned long qsmaskinit;
                                 /* Per-GP initialization for qsmask. */
         unsigned long grpmask;  /* Mask to apply to parent qsmask. */
+                                /*  Only one bit will be set in this mask. */
         int     grplo;          /* lowest-numbered CPU or group here. */
         int     grphi;          /* highest-numbered CPU or group here. */
         u8      grpnum;         /* CPU/group number for next level up. */
@@ -95,6 +101,9 @@ struct rcu_node {
         struct rcu_node *parent;
         struct list_head blocked_tasks[2];
                                 /* Tasks blocked in RCU read-side critsect. */
+                                /*  Grace period number (->gpnum) x blocked */
+                                /*  by tasks on the (x & 0x1) element of the */
+                                /*  blocked_tasks[] array. */
 } ____cacheline_internodealigned_in_smp;
 
 /* Index values for nxttail array in struct rcu_data. */
@@ -126,19 +135,22 @@ struct rcu_data {
          * Any of the partitions might be empty, in which case the
          * pointer to that partition will be equal to the pointer for
          * the following partition.  When the list is empty, all of
-         * the nxttail elements point to nxtlist, which is NULL.
+         * the nxttail elements point to the ->nxtlist pointer itself,
+         * which in that case is NULL.
          *
-         * [*nxttail[RCU_NEXT_READY_TAIL], NULL = *nxttail[RCU_NEXT_TAIL]):
-         *      Entries that might have arrived after current GP ended
-         * [*nxttail[RCU_WAIT_TAIL], *nxttail[RCU_NEXT_READY_TAIL]):
-         *      Entries known to have arrived before current GP ended
-         * [*nxttail[RCU_DONE_TAIL], *nxttail[RCU_WAIT_TAIL]):
-         *      Entries that batch # <= ->completed - 1: waiting for current GP
          * [nxtlist, *nxttail[RCU_DONE_TAIL]):
          *      Entries that batch # <= ->completed
          *      The grace period for these entries has completed, and
          *      the other grace-period-completed entries may be moved
          *      here temporarily in rcu_process_callbacks().
+         * [*nxttail[RCU_DONE_TAIL], *nxttail[RCU_WAIT_TAIL]):
+         *      Entries that batch # <= ->completed - 1: waiting for current GP
+         * [*nxttail[RCU_WAIT_TAIL], *nxttail[RCU_NEXT_READY_TAIL]):
+         *      Entries known to have arrived before current GP ended
+         * [*nxttail[RCU_NEXT_READY_TAIL], *nxttail[RCU_NEXT_TAIL]):
+         *      Entries that might have arrived after current GP ended
+         *      Note that the value of *nxttail[RCU_NEXT_TAIL] will
+         *      always be NULL, as this is the end of the list.
          */
         struct rcu_head *nxtlist;
         struct rcu_head **nxttail[RCU_NEXT_SIZE];
@@ -216,6 +228,9 @@ struct rcu_state {
                                                 /* Force QS state. */
         long    gpnum;                          /* Current gp number. */
         long    completed;                      /* # of last completed gp. */
+
+        /* End  of fields guarded by root rcu_node's lock. */
+
         spinlock_t onofflock;                   /* exclude on/offline and */
                                                 /*  starting new GP. */
         spinlock_t fqslock;                     /* Only one task forcing */
diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h
index 8ff1ba7b3c43..65250219ab6d 100644
--- a/kernel/rcutree_plugin.h
+++ b/kernel/rcutree_plugin.h
@@ -419,6 +419,18 @@ static void __cpuinit rcu_preempt_init_percpu_data(int cpu)
 }
 
 /*
+ * Initialize preemptable RCU's state structures.
+ */
+static void __init __rcu_init_preempt(void)
+{
+        int i;                  /* All used by RCU_INIT_FLAVOR(). */
+        int j;
+        struct rcu_node *rnp;
+
+        RCU_INIT_FLAVOR(&rcu_preempt_state, rcu_preempt_data);
+}
+
+/*
  * Check for a task exiting while in a preemptable-RCU read-side
  * critical section, clean up if so.  No need to issue warnings,
  * as debug_check_no_locks_held() already does this if lockdep
@@ -518,7 +530,7 @@ static void rcu_preempt_offline_cpu(int cpu)
  * Because preemptable RCU does not exist, it never has any callbacks
  * to check.
  */
-void rcu_preempt_check_callbacks(int cpu)
+static void rcu_preempt_check_callbacks(int cpu)
 {
 }
 
@@ -526,7 +538,7 @@ void rcu_preempt_check_callbacks(int cpu)
  * Because preemptable RCU does not exist, it never has any callbacks
  * to process.
  */
-void rcu_preempt_process_callbacks(void)
+static void rcu_preempt_process_callbacks(void)
 {
 }
 
@@ -563,4 +575,11 @@ static void __cpuinit rcu_preempt_init_percpu_data(int cpu)
 {
 }
 
+/*
+ * Because preemptable RCU does not exist, it need not be initialized.
+ */
+static void __init __rcu_init_preempt(void)
+{
+}
+
 #endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */