1 files changed, 81 insertions, 48 deletions

diff --git a/kernel/bpf/percpu_freelist.c b/kernel/bpf/percpu_freelist.c
index 673fa6fe2d73..632762b57299 100644
--- a/kernel/bpf/percpu_freelist.c
+++ b/kernel/bpf/percpu_freelist.c
@@ -1,8 +1,5 @@
+// SPDX-License-Identifier: GPL-2.0-only
 /* Copyright (c) 2016 Facebook
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of version 2 of the GNU General Public
- * License as published by the Free Software Foundation.
  */
 #include "percpu_freelist.h"
 
@@ -17,7 +14,7 @@ int pcpu_freelist_init(struct pcpu_freelist *s)
 	for_each_possible_cpu(cpu) {
 		struct pcpu_freelist_head *head = per_cpu_ptr(s->freelist, cpu);
 
-		raw_spin_lock_init(&head->lock);
+		raw_res_spin_lock_init(&head->lock);
 		head->first = NULL;
 	}
 	return 0;
@@ -28,77 +25,113 @@ void pcpu_freelist_destroy(struct pcpu_freelist *s)
 	free_percpu(s->freelist);
 }
 
-static inline void __pcpu_freelist_push(struct pcpu_freelist_head *head,
-					struct pcpu_freelist_node *node)
+static inline void pcpu_freelist_push_node(struct pcpu_freelist_head *head,
+					   struct pcpu_freelist_node *node)
 {
-	raw_spin_lock(&head->lock);
 	node->next = head->first;
-	head->first = node;
-	raw_spin_unlock(&head->lock);
+	WRITE_ONCE(head->first, node);
+}
+
+static inline bool ___pcpu_freelist_push(struct pcpu_freelist_head *head,
+					 struct pcpu_freelist_node *node)
+{
+	if (raw_res_spin_lock(&head->lock))
+		return false;
+	pcpu_freelist_push_node(head, node);
+	raw_res_spin_unlock(&head->lock);
+	return true;
+}
+
+void __pcpu_freelist_push(struct pcpu_freelist *s,
+			struct pcpu_freelist_node *node)
+{
+	struct pcpu_freelist_head *head;
+	int cpu;
+
+	if (___pcpu_freelist_push(this_cpu_ptr(s->freelist), node))
+		return;
+
+	while (true) {
+		for_each_cpu_wrap(cpu, cpu_possible_mask, raw_smp_processor_id()) {
+			if (cpu == raw_smp_processor_id())
+				continue;
+			head = per_cpu_ptr(s->freelist, cpu);
+			if (raw_res_spin_lock(&head->lock))
+				continue;
+			pcpu_freelist_push_node(head, node);
+			raw_res_spin_unlock(&head->lock);
+			return;
+		}
+	}
 }
 
 void pcpu_freelist_push(struct pcpu_freelist *s,
 			struct pcpu_freelist_node *node)
 {
-	struct pcpu_freelist_head *head = this_cpu_ptr(s->freelist);
+	unsigned long flags;
 
-	__pcpu_freelist_push(head, node);
+	local_irq_save(flags);
+	__pcpu_freelist_push(s, node);
+	local_irq_restore(flags);
 }
 
 void pcpu_freelist_populate(struct pcpu_freelist *s, void *buf, u32 elem_size,
 			    u32 nr_elems)
 {
 	struct pcpu_freelist_head *head;
-	unsigned long flags;
-	int i, cpu, pcpu_entries;
+	unsigned int cpu, cpu_idx, i, j, n, m;
 
-	pcpu_entries = nr_elems / num_possible_cpus() + 1;
-	i = 0;
+	n = nr_elems / num_possible_cpus();
+	m = nr_elems % num_possible_cpus();
 
-	/* disable irq to workaround lockdep false positive
-	 * in bpf usage pcpu_freelist_populate() will never race
-	 * with pcpu_freelist_push()
-	 */
-	local_irq_save(flags);
+	cpu_idx = 0;
 	for_each_possible_cpu(cpu) {
-again:
 		head = per_cpu_ptr(s->freelist, cpu);
-		__pcpu_freelist_push(head, buf);
-		i++;
-		buf += elem_size;
-		if (i == nr_elems)
-			break;
-		if (i % pcpu_entries)
-			goto again;
+		j = n + (cpu_idx < m ? 1 : 0);
+		for (i = 0; i < j; i++) {
+			/* No locking required as this is not visible yet. */
+			pcpu_freelist_push_node(head, buf);
+			buf += elem_size;
+		}
+		cpu_idx++;
 	}
-	local_irq_restore(flags);
 }
 
-struct pcpu_freelist_node *pcpu_freelist_pop(struct pcpu_freelist *s)
+static struct pcpu_freelist_node *___pcpu_freelist_pop(struct pcpu_freelist *s)
 {
+	struct pcpu_freelist_node *node = NULL;
 	struct pcpu_freelist_head *head;
-	struct pcpu_freelist_node *node;
-	unsigned long flags;
-	int orig_cpu, cpu;
+	int cpu;
 
-	local_irq_save(flags);
-	orig_cpu = cpu = raw_smp_processor_id();
-	while (1) {
+	for_each_cpu_wrap(cpu, cpu_possible_mask, raw_smp_processor_id()) {
 		head = per_cpu_ptr(s->freelist, cpu);
-		raw_spin_lock(&head->lock);
+		if (!READ_ONCE(head->first))
+			continue;
+		if (raw_res_spin_lock(&head->lock))
+			continue;
 		node = head->first;
 		if (node) {
-			head->first = node->next;
-			raw_spin_unlock_irqrestore(&head->lock, flags);
+			WRITE_ONCE(head->first, node->next);
+			raw_res_spin_unlock(&head->lock);
 			return node;
 		}
-		raw_spin_unlock(&head->lock);
-		cpu = cpumask_next(cpu, cpu_possible_mask);
-		if (cpu >= nr_cpu_ids)
-			cpu = 0;
-		if (cpu == orig_cpu) {
-			local_irq_restore(flags);
-			return NULL;
-		}
+		raw_res_spin_unlock(&head->lock);
 	}
+	return node;
+}
+
+struct pcpu_freelist_node *__pcpu_freelist_pop(struct pcpu_freelist *s)
+{
+	return ___pcpu_freelist_pop(s);
+}
+
+struct pcpu_freelist_node *pcpu_freelist_pop(struct pcpu_freelist *s)
+{
+	struct pcpu_freelist_node *ret;
+	unsigned long flags;
+
+	local_irq_save(flags);
+	ret = __pcpu_freelist_pop(s);
+	local_irq_restore(flags);
+	return ret;
 }