Merge branch 'perf/core' into perf/uprobes

Merge in latest upstream (and the latest perf development tree),
to prepare for tooling changes, and also to pick up v3.4 MM
changes that the uprobes code needs to take care of.

Signed-off-by: Ingo Molnar <mingo@kernel.org>

1 files changed, 23 insertions, 34 deletions

diff --git a/kernel/padata.c b/kernel/padata.c
index b4525993151..89fe3d1b9ef 100644
--- a/kernel/padata.c
+++ b/kernel/padata.c
@@ -1,6 +1,8 @@
 /*
  * padata.c - generic interface to process data streams in parallel
  *
+ * See Documentation/padata.txt for an api documentation.
+ *
  * Copyright (C) 2008, 2009 secunet Security Networks AG
  * Copyright (C) 2008, 2009 Steffen Klassert <steffen.klassert@secunet.com>
  *
@@ -29,7 +31,6 @@
 #include <linux/sysfs.h>
 #include <linux/rcupdate.h>
 
-#define MAX_SEQ_NR (INT_MAX - NR_CPUS)
 #define MAX_OBJ_NUM 1000
 
 static int padata_index_to_cpu(struct parallel_data *pd, int cpu_index)
@@ -43,18 +44,19 @@ static int padata_index_to_cpu(struct parallel_data *pd, int cpu_index)
 	return target_cpu;
 }
 
-static int padata_cpu_hash(struct padata_priv *padata)
+static int padata_cpu_hash(struct parallel_data *pd)
 {
 	int cpu_index;
-	struct parallel_data *pd;
-
-	pd =  padata->pd;
 
 	/*
 	 * Hash the sequence numbers to the cpus by taking
 	 * seq_nr mod. number of cpus in use.
 	 */
-	cpu_index =  padata->seq_nr % cpumask_weight(pd->cpumask.pcpu);
+
+	spin_lock(&pd->seq_lock);
+	cpu_index =  pd->seq_nr % cpumask_weight(pd->cpumask.pcpu);
+	pd->seq_nr++;
+	spin_unlock(&pd->seq_lock);
 
 	return padata_index_to_cpu(pd, cpu_index);
 }
@@ -132,12 +134,7 @@ int padata_do_parallel(struct padata_instance *pinst,
 	padata->pd = pd;
 	padata->cb_cpu = cb_cpu;
 
-	if (unlikely(atomic_read(&pd->seq_nr) == pd->max_seq_nr))
-		atomic_set(&pd->seq_nr, -1);
-
-	padata->seq_nr = atomic_inc_return(&pd->seq_nr);
-
-	target_cpu = padata_cpu_hash(padata);
+	target_cpu = padata_cpu_hash(pd);
 	queue = per_cpu_ptr(pd->pqueue, target_cpu);
 
 	spin_lock(&queue->parallel.lock);
@@ -173,7 +170,7 @@ EXPORT_SYMBOL(padata_do_parallel);
 static struct padata_priv *padata_get_next(struct parallel_data *pd)
 {
 	int cpu, num_cpus;
-	int next_nr, next_index;
+	unsigned int next_nr, next_index;
 	struct padata_parallel_queue *queue, *next_queue;
 	struct padata_priv *padata;
 	struct padata_list *reorder;
@@ -189,14 +186,6 @@ static struct padata_priv *padata_get_next(struct parallel_data *pd)
 	cpu = padata_index_to_cpu(pd, next_index);
 	next_queue = per_cpu_ptr(pd->pqueue, cpu);
 
-	if (unlikely(next_nr > pd->max_seq_nr)) {
-		next_nr = next_nr - pd->max_seq_nr - 1;
-		next_index = next_nr % num_cpus;
-		cpu = padata_index_to_cpu(pd, next_index);
-		next_queue = per_cpu_ptr(pd->pqueue, cpu);
-		pd->processed = 0;
-	}
-
 	padata = NULL;
 
 	reorder = &next_queue->reorder;
@@ -205,8 +194,6 @@ static struct padata_priv *padata_get_next(struct parallel_data *pd)
 		padata = list_entry(reorder->list.next,
 				    struct padata_priv, list);
 
-		BUG_ON(next_nr != padata->seq_nr);
-
 		spin_lock(&reorder->lock);
 		list_del_init(&padata->list);
 		atomic_dec(&pd->reorder_objects);
@@ -230,6 +217,7 @@ out:
 
 static void padata_reorder(struct parallel_data *pd)
 {
+	int cb_cpu;
 	struct padata_priv *padata;
 	struct padata_serial_queue *squeue;
 	struct padata_instance *pinst = pd->pinst;
@@ -270,13 +258,14 @@ static void padata_reorder(struct parallel_data *pd)
 			return;
 		}
 
-		squeue = per_cpu_ptr(pd->squeue, padata->cb_cpu);
+		cb_cpu = padata->cb_cpu;
+		squeue = per_cpu_ptr(pd->squeue, cb_cpu);
 
 		spin_lock(&squeue->serial.lock);
 		list_add_tail(&padata->list, &squeue->serial.list);
 		spin_unlock(&squeue->serial.lock);
 
-		queue_work_on(padata->cb_cpu, pinst->wq, &squeue->work);
+		queue_work_on(cb_cpu, pinst->wq, &squeue->work);
 	}
 
 	spin_unlock_bh(&pd->lock);
@@ -367,13 +356,13 @@ static int padata_setup_cpumasks(struct parallel_data *pd,
 	if (!alloc_cpumask_var(&pd->cpumask.pcpu, GFP_KERNEL))
 		return -ENOMEM;
 
-	cpumask_and(pd->cpumask.pcpu, pcpumask, cpu_active_mask);
+	cpumask_and(pd->cpumask.pcpu, pcpumask, cpu_online_mask);
 	if (!alloc_cpumask_var(&pd->cpumask.cbcpu, GFP_KERNEL)) {
 		free_cpumask_var(pd->cpumask.cbcpu);
 		return -ENOMEM;
 	}
 
-	cpumask_and(pd->cpumask.cbcpu, cbcpumask, cpu_active_mask);
+	cpumask_and(pd->cpumask.cbcpu, cbcpumask, cpu_online_mask);
 	return 0;
 }
 
@@ -400,7 +389,7 @@ static void padata_init_squeues(struct parallel_data *pd)
 /* Initialize all percpu queues used by parallel workers */
 static void padata_init_pqueues(struct parallel_data *pd)
 {
-	int cpu_index, num_cpus, cpu;
+	int cpu_index, cpu;
 	struct padata_parallel_queue *pqueue;
 
 	cpu_index = 0;
@@ -415,9 +404,6 @@ static void padata_init_pqueues(struct parallel_data *pd)
 		INIT_WORK(&pqueue->work, padata_parallel_worker);
 		atomic_set(&pqueue->num_obj, 0);
 	}
-
-	num_cpus = cpumask_weight(pd->cpumask.pcpu);
-	pd->max_seq_nr = num_cpus ? (MAX_SEQ_NR / num_cpus) * num_cpus - 1 : 0;
 }
 
 /* Allocate and initialize the internal cpumask dependend resources. */
@@ -444,7 +430,7 @@ static struct parallel_data *padata_alloc_pd(struct padata_instance *pinst,
 	padata_init_pqueues(pd);
 	padata_init_squeues(pd);
 	setup_timer(&pd->timer, padata_reorder_timer, (unsigned long)pd);
-	atomic_set(&pd->seq_nr, -1);
+	pd->seq_nr = 0;
 	atomic_set(&pd->reorder_objects, 0);
 	atomic_set(&pd->refcnt, 0);
 	pd->pinst = pinst;
@@ -580,7 +566,7 @@ EXPORT_SYMBOL(padata_unregister_cpumask_notifier);
 static bool padata_validate_cpumask(struct padata_instance *pinst,
 				    const struct cpumask *cpumask)
 {
-	if (!cpumask_intersects(cpumask, cpu_active_mask)) {
+	if (!cpumask_intersects(cpumask, cpu_online_mask)) {
 		pinst->flags |= PADATA_INVALID;
 		return false;
 	}
@@ -694,7 +680,7 @@ static int __padata_add_cpu(struct padata_instance *pinst, int cpu)
 {
 	struct parallel_data *pd;
 
-	if (cpumask_test_cpu(cpu, cpu_active_mask)) {
+	if (cpumask_test_cpu(cpu, cpu_online_mask)) {
 		pd = padata_alloc_pd(pinst, pinst->cpumask.pcpu,
 				     pinst->cpumask.cbcpu);
 		if (!pd)
@@ -762,6 +748,9 @@ static int __padata_remove_cpu(struct padata_instance *pinst, int cpu)
 			return -ENOMEM;
 
 		padata_replace(pinst, pd);
+
+		cpumask_clear_cpu(cpu, pd->cpumask.cbcpu);
+		cpumask_clear_cpu(cpu, pd->cpumask.pcpu);
 	}
 
 	return 0;