1 files changed, 149 insertions, 50 deletions

diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index b1c3588a8f00..76f5f53a645f 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -751,8 +751,9 @@ enum scx_rq_flags {
 	 */
 	SCX_RQ_ONLINE		= 1 << 0,
 	SCX_RQ_CAN_STOP_TICK	= 1 << 1,
-	SCX_RQ_BAL_KEEP		= 1 << 2, /* balance decided to keep current */
-	SCX_RQ_BYPASSING	= 1 << 3,
+	SCX_RQ_BAL_PENDING	= 1 << 2, /* balance hasn't run yet */
+	SCX_RQ_BAL_KEEP		= 1 << 3, /* balance decided to keep current */
+	SCX_RQ_BYPASSING	= 1 << 4,
 
 	SCX_RQ_IN_WAKEUP	= 1 << 16,
 	SCX_RQ_IN_BALANCE	= 1 << 17,
@@ -1147,7 +1148,10 @@ struct rq {
 	 */
 	unsigned int		nr_uninterruptible;
 
-	struct task_struct __rcu	*curr;
+	union {
+		struct task_struct __rcu *donor; /* Scheduler context */
+		struct task_struct __rcu *curr;  /* Execution context */
+	};
 	struct sched_dl_entity	*dl_server;
 	struct task_struct	*idle;
 	struct task_struct	*stop;
@@ -1344,6 +1348,11 @@ DECLARE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues);
 #define cpu_curr(cpu)		(cpu_rq(cpu)->curr)
 #define raw_rq()		raw_cpu_ptr(&runqueues)
 
+static inline void rq_set_donor(struct rq *rq, struct task_struct *t)
+{
+	/* Do nothing */
+}
+
 #ifdef CONFIG_SCHED_CORE
 static inline struct cpumask *sched_group_span(struct sched_group *sg);
 
@@ -2085,34 +2094,6 @@ static inline const struct cpumask *task_user_cpus(struct task_struct *p)
 
 #endif /* CONFIG_SMP */
 
-#include "stats.h"
-
-#if defined(CONFIG_SCHED_CORE) && defined(CONFIG_SCHEDSTATS)
-
-extern void __sched_core_account_forceidle(struct rq *rq);
-
-static inline void sched_core_account_forceidle(struct rq *rq)
-{
-	if (schedstat_enabled())
-		__sched_core_account_forceidle(rq);
-}
-
-extern void __sched_core_tick(struct rq *rq);
-
-static inline void sched_core_tick(struct rq *rq)
-{
-	if (sched_core_enabled(rq) && schedstat_enabled())
-		__sched_core_tick(rq);
-}
-
-#else /* !(CONFIG_SCHED_CORE && CONFIG_SCHEDSTATS): */
-
-static inline void sched_core_account_forceidle(struct rq *rq) { }
-
-static inline void sched_core_tick(struct rq *rq) { }
-
-#endif /* !(CONFIG_SCHED_CORE && CONFIG_SCHEDSTATS) */
-
 #ifdef CONFIG_CGROUP_SCHED
 
 /*
@@ -2260,11 +2241,25 @@ static inline u64 global_rt_runtime(void)
 	return (u64)sysctl_sched_rt_runtime * NSEC_PER_USEC;
 }
 
+/*
+ * Is p the current execution context?
+ */
 static inline int task_current(struct rq *rq, struct task_struct *p)
 {
 	return rq->curr == p;
 }
 
+/*
+ * Is p the current scheduling context?
+ *
+ * Note that it might be the current execution context at the same time if
+ * rq->curr == rq->donor == p.
+ */
+static inline int task_current_donor(struct rq *rq, struct task_struct *p)
+{
+	return rq->donor == p;
+}
+
 static inline int task_on_cpu(struct rq *rq, struct task_struct *p)
 {
 #ifdef CONFIG_SMP
@@ -2292,6 +2287,7 @@ static inline int task_on_rq_migrating(struct task_struct *p)
 #define WF_SYNC			0x10 /* Waker goes to sleep after wakeup */
 #define WF_MIGRATED		0x20 /* Internal use, task got migrated */
 #define WF_CURRENT_CPU		0x40 /* Prefer to move the wakee to the current CPU. */
+#define WF_RQ_SELECTED		0x80 /* ->select_task_rq() was called */
 
 #ifdef CONFIG_SMP
 static_assert(WF_EXEC == SD_BALANCE_EXEC);
@@ -2334,6 +2330,7 @@ extern const u32		sched_prio_to_wmult[40];
  * ENQUEUE_HEAD      - place at front of runqueue (tail if not specified)
  * ENQUEUE_REPLENISH - CBS (replenish runtime and postpone deadline)
  * ENQUEUE_MIGRATED  - the task was migrated during wakeup
+ * ENQUEUE_RQ_SELECTED - ->select_task_rq() was called
  *
  */
 
@@ -2360,6 +2357,7 @@ extern const u32		sched_prio_to_wmult[40];
 #define ENQUEUE_INITIAL		0x80
 #define ENQUEUE_MIGRATING	0x100
 #define ENQUEUE_DELAYED		0x200
+#define ENQUEUE_RQ_SELECTED	0x400
 
 #define RETRY_TASK		((void *)-1UL)
 
@@ -2448,7 +2446,7 @@ struct sched_class {
 
 static inline void put_prev_task(struct rq *rq, struct task_struct *prev)
 {
-	WARN_ON_ONCE(rq->curr != prev);
+	WARN_ON_ONCE(rq->donor != prev);
 	prev->sched_class->put_prev_task(rq, prev, NULL);
 }
 
@@ -2612,7 +2610,7 @@ static inline cpumask_t *alloc_user_cpus_ptr(int node)
 
 static inline struct task_struct *get_push_task(struct rq *rq)
 {
-	struct task_struct *p = rq->curr;
+	struct task_struct *p = rq->donor;
 
 	lockdep_assert_rq_held(rq);
 
@@ -2692,6 +2690,7 @@ extern void init_sched_rt_class(void);
 extern void init_sched_fair_class(void);
 
 extern void resched_curr(struct rq *rq);
+extern void resched_curr_lazy(struct rq *rq);
 extern void resched_cpu(int cpu);
 
 extern void init_rt_bandwidth(struct rt_bandwidth *rt_b, u64 period, u64 runtime);
@@ -2766,8 +2765,6 @@ static inline void sub_nr_running(struct rq *rq, unsigned count)
 
 static inline void __block_task(struct rq *rq, struct task_struct *p)
 {
-	WRITE_ONCE(p->on_rq, 0);
-	ASSERT_EXCLUSIVE_WRITER(p->on_rq);
 	if (p->sched_contributes_to_load)
 		rq->nr_uninterruptible++;
 
@@ -2775,6 +2772,38 @@ static inline void __block_task(struct rq *rq, struct task_struct *p)
 		atomic_inc(&rq->nr_iowait);
 		delayacct_blkio_start();
 	}
+
+	ASSERT_EXCLUSIVE_WRITER(p->on_rq);
+
+	/*
+	 * The moment this write goes through, ttwu() can swoop in and migrate
+	 * this task, rendering our rq->__lock ineffective.
+	 *
+	 * __schedule()				try_to_wake_up()
+	 *   LOCK rq->__lock			  LOCK p->pi_lock
+	 *   pick_next_task()
+	 *     pick_next_task_fair()
+	 *       pick_next_entity()
+	 *         dequeue_entities()
+	 *           __block_task()
+	 *             RELEASE p->on_rq = 0	  if (p->on_rq && ...)
+	 *					    break;
+	 *
+	 *					  ACQUIRE (after ctrl-dep)
+	 *
+	 *					  cpu = select_task_rq();
+	 *					  set_task_cpu(p, cpu);
+	 *					  ttwu_queue()
+	 *					    ttwu_do_activate()
+	 *					      LOCK rq->__lock
+	 *					      activate_task()
+	 *					        STORE p->on_rq = 1
+	 *   UNLOCK rq->__lock
+	 *
+	 * Callers must ensure to not reference @p after this -- we no longer
+	 * own it.
+	 */
+	smp_store_release(&p->on_rq, 0);
 }
 
 extern void activate_task(struct rq *rq, struct task_struct *p, int flags);
@@ -3166,6 +3195,34 @@ extern void nohz_run_idle_balance(int cpu);
 static inline void nohz_run_idle_balance(int cpu) { }
 #endif
 
+#include "stats.h"
+
+#if defined(CONFIG_SCHED_CORE) && defined(CONFIG_SCHEDSTATS)
+
+extern void __sched_core_account_forceidle(struct rq *rq);
+
+static inline void sched_core_account_forceidle(struct rq *rq)
+{
+	if (schedstat_enabled())
+		__sched_core_account_forceidle(rq);
+}
+
+extern void __sched_core_tick(struct rq *rq);
+
+static inline void sched_core_tick(struct rq *rq)
+{
+	if (sched_core_enabled(rq) && schedstat_enabled())
+		__sched_core_tick(rq);
+}
+
+#else /* !(CONFIG_SCHED_CORE && CONFIG_SCHEDSTATS): */
+
+static inline void sched_core_account_forceidle(struct rq *rq) { }
+
+static inline void sched_core_tick(struct rq *rq) { }
+
+#endif /* !(CONFIG_SCHED_CORE && CONFIG_SCHEDSTATS) */
+
 #ifdef CONFIG_IRQ_TIME_ACCOUNTING
 
 struct irqtime {
@@ -3596,24 +3653,41 @@ static inline void mm_cid_put(struct mm_struct *mm)
 	__mm_cid_put(mm, mm_cid_clear_lazy_put(cid));
 }
 
-static inline int __mm_cid_try_get(struct mm_struct *mm)
+static inline int __mm_cid_try_get(struct task_struct *t, struct mm_struct *mm)
 {
-	struct cpumask *cpumask;
-	int cid;
+	struct cpumask *cidmask = mm_cidmask(mm);
+	struct mm_cid __percpu *pcpu_cid = mm->pcpu_cid;
+	int cid = __this_cpu_read(pcpu_cid->recent_cid);
 
-	cpumask = mm_cidmask(mm);
+	/* Try to re-use recent cid. This improves cache locality. */
+	if (!mm_cid_is_unset(cid) && !cpumask_test_and_set_cpu(cid, cidmask))
+		return cid;
+	/*
+	 * Expand cid allocation if the maximum number of concurrency
+	 * IDs allocated (max_nr_cid) is below the number cpus allowed
+	 * and number of threads. Expanding cid allocation as much as
+	 * possible improves cache locality.
+	 */
+	cid = atomic_read(&mm->max_nr_cid);
+	while (cid < READ_ONCE(mm->nr_cpus_allowed) && cid < atomic_read(&mm->mm_users)) {
+		if (!atomic_try_cmpxchg(&mm->max_nr_cid, &cid, cid + 1))
+			continue;
+		if (!cpumask_test_and_set_cpu(cid, cidmask))
+			return cid;
+	}
 	/*
+	 * Find the first available concurrency id.
 	 * Retry finding first zero bit if the mask is temporarily
 	 * filled. This only happens during concurrent remote-clear
 	 * which owns a cid without holding a rq lock.
 	 */
 	for (;;) {
-		cid = cpumask_first_zero(cpumask);
-		if (cid < nr_cpu_ids)
+		cid = cpumask_first_zero(cidmask);
+		if (cid < READ_ONCE(mm->nr_cpus_allowed))
 			break;
 		cpu_relax();
 	}
-	if (cpumask_test_and_set_cpu(cid, cpumask))
+	if (cpumask_test_and_set_cpu(cid, cidmask))
 		return -1;
 
 	return cid;
@@ -3631,7 +3705,8 @@ static inline void mm_cid_snapshot_time(struct rq *rq, struct mm_struct *mm)
 	WRITE_ONCE(pcpu_cid->time, rq->clock);
 }
 
-static inline int __mm_cid_get(struct rq *rq, struct mm_struct *mm)
+static inline int __mm_cid_get(struct rq *rq, struct task_struct *t,
+			       struct mm_struct *mm)
 {
 	int cid;
 
@@ -3641,13 +3716,13 @@ static inline int __mm_cid_get(struct rq *rq, struct mm_struct *mm)
 	 * guarantee forward progress.
 	 */
 	if (!READ_ONCE(use_cid_lock)) {
-		cid = __mm_cid_try_get(mm);
+		cid = __mm_cid_try_get(t, mm);
 		if (cid >= 0)
 			goto end;
 		raw_spin_lock(&cid_lock);
 	} else {
 		raw_spin_lock(&cid_lock);
-		cid = __mm_cid_try_get(mm);
+		cid = __mm_cid_try_get(t, mm);
 		if (cid >= 0)
 			goto unlock;
 	}
@@ -3667,7 +3742,7 @@ static inline int __mm_cid_get(struct rq *rq, struct mm_struct *mm)
 	 * all newcoming allocations observe the use_cid_lock flag set.
 	 */
 	do {
-		cid = __mm_cid_try_get(mm);
+		cid = __mm_cid_try_get(t, mm);
 		cpu_relax();
 	} while (cid < 0);
 	/*
@@ -3684,7 +3759,8 @@ end:
 	return cid;
 }
 
-static inline int mm_cid_get(struct rq *rq, struct mm_struct *mm)
+static inline int mm_cid_get(struct rq *rq, struct task_struct *t,
+			     struct mm_struct *mm)
 {
 	struct mm_cid __percpu *pcpu_cid = mm->pcpu_cid;
 	struct cpumask *cpumask;
@@ -3701,8 +3777,9 @@ static inline int mm_cid_get(struct rq *rq, struct mm_struct *mm)
 		if (try_cmpxchg(&this_cpu_ptr(pcpu_cid)->cid, &cid, MM_CID_UNSET))
 			__mm_cid_put(mm, mm_cid_clear_lazy_put(cid));
 	}
-	cid = __mm_cid_get(rq, mm);
+	cid = __mm_cid_get(rq, t, mm);
 	__this_cpu_write(pcpu_cid->cid, cid);
+	__this_cpu_write(pcpu_cid->recent_cid, cid);
 
 	return cid;
 }
@@ -3755,7 +3832,7 @@ static inline void switch_mm_cid(struct rq *rq,
 		prev->mm_cid = -1;
 	}
 	if (next->mm_cid_active)
-		next->last_mm_cid = next->mm_cid = mm_cid_get(rq, next->mm);
+		next->last_mm_cid = next->mm_cid = mm_cid_get(rq, next, next->mm);
 }
 
 #else /* !CONFIG_SCHED_MM_CID: */
@@ -3768,6 +3845,28 @@ static inline void init_sched_mm_cid(struct task_struct *t) { }
 
 extern u64 avg_vruntime(struct cfs_rq *cfs_rq);
 extern int entity_eligible(struct cfs_rq *cfs_rq, struct sched_entity *se);
+#ifdef CONFIG_SMP
+static inline
+void move_queued_task_locked(struct rq *src_rq, struct rq *dst_rq, struct task_struct *task)
+{
+	lockdep_assert_rq_held(src_rq);
+	lockdep_assert_rq_held(dst_rq);
+
+	deactivate_task(src_rq, task, 0);
+	set_task_cpu(task, dst_rq->cpu);
+	activate_task(dst_rq, task, 0);
+}
+
+static inline
+bool task_is_pushable(struct rq *rq, struct task_struct *p, int cpu)
+{
+	if (!task_on_cpu(rq, p) &&
+	    cpumask_test_cpu(cpu, &p->cpus_mask))
+		return true;
+
+	return false;
+}
+#endif
 
 #ifdef CONFIG_RT_MUTEXES
 
@@ -3797,7 +3896,7 @@ static inline int rt_effective_prio(struct task_struct *p, int prio)
 
 extern int __sched_setscheduler(struct task_struct *p, const struct sched_attr *attr, bool user, bool pi);
 extern int __sched_setaffinity(struct task_struct *p, struct affinity_context *ctx);
-extern void __setscheduler_prio(struct task_struct *p, int prio);
+extern const struct sched_class *__setscheduler_class(int policy, int prio);
 extern void set_load_weight(struct task_struct *p, bool update_load);
 extern void enqueue_task(struct rq *rq, struct task_struct *p, int flags);
 extern bool dequeue_task(struct rq *rq, struct task_struct *p, int flags);