Linux-libre 4.5.3-gnupck-4.5.3-gnu

7 files changed, 93 insertions, 32 deletions

diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 6a498daf2..355cd5f2b 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -2697,9 +2697,10 @@ static ssize_t __cgroup_procs_write(struct kernfs_open_file *of, char *buf,
 				    size_t nbytes, loff_t off, bool threadgroup)
 {
 	struct task_struct *tsk;
+	struct cgroup_subsys *ss;
 	struct cgroup *cgrp;
 	pid_t pid;
-	int ret;
+	int ssid, ret;
 
 	if (kstrtoint(strstrip(buf), 0, &pid) || pid < 0)
 		return -EINVAL;
@@ -2747,8 +2748,10 @@ out_unlock_rcu:
 	rcu_read_unlock();
 out_unlock_threadgroup:
 	percpu_up_write(&cgroup_threadgroup_rwsem);
+	for_each_subsys(ss, ssid)
+		if (ss->post_attach)
+			ss->post_attach();
 	cgroup_kn_unlock(of->kn);
-	cpuset_post_attach_flush();
 	return ret ?: nbytes;
 }
 
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 41989ab4d..df16d0c93 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -58,7 +58,6 @@
 #include <asm/uaccess.h>
 #include <linux/atomic.h>
 #include <linux/mutex.h>
-#include <linux/workqueue.h>
 #include <linux/cgroup.h>
 #include <linux/wait.h>
 
@@ -1016,7 +1015,7 @@ static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from,
 	}
 }
 
-void cpuset_post_attach_flush(void)
+static void cpuset_post_attach(void)
 {
 	flush_workqueue(cpuset_migrate_mm_wq);
 }
@@ -2087,6 +2086,7 @@ struct cgroup_subsys cpuset_cgrp_subsys = {
 	.can_attach	= cpuset_can_attach,
 	.cancel_attach	= cpuset_cancel_attach,
 	.attach		= cpuset_attach,
+	.post_attach	= cpuset_post_attach,
 	.bind		= cpuset_bind,
 	.legacy_cftypes	= files,
 	.early_init	= 1,
diff --git a/kernel/events/core.c b/kernel/events/core.c
index f0b4b328d..a0ef98b25 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -2402,14 +2402,24 @@ static void ctx_sched_out(struct perf_event_context *ctx,
 			cpuctx->task_ctx = NULL;
 	}
 
-	is_active ^= ctx->is_active; /* changed bits */
-
+	/*
+	 * Always update time if it was set; not only when it changes.
+	 * Otherwise we can 'forget' to update time for any but the last
+	 * context we sched out. For example:
+	 *
+	 *   ctx_sched_out(.event_type = EVENT_FLEXIBLE)
+	 *   ctx_sched_out(.event_type = EVENT_PINNED)
+	 *
+	 * would only update time for the pinned events.
+	 */
 	if (is_active & EVENT_TIME) {
 		/* update (and stop) ctx time */
 		update_context_time(ctx);
 		update_cgrp_time_from_cpuctx(cpuctx);
 	}
 
+	is_active ^= ctx->is_active; /* changed bits */
+
 	if (!ctx->nr_active || !(is_active & EVENT_ALL))
 		return;
 
@@ -8479,6 +8489,7 @@ SYSCALL_DEFINE5(perf_event_open,
 					f_flags);
 	if (IS_ERR(event_file)) {
 		err = PTR_ERR(event_file);
+		event_file = NULL;
 		goto err_context;
 	}
 
diff --git a/kernel/futex.c b/kernel/futex.c
index 5d6ce6413..11b502159 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -1212,10 +1212,20 @@ static int wake_futex_pi(u32 __user *uaddr, u32 uval, struct futex_q *this,
 	if (unlikely(should_fail_futex(true)))
 		ret = -EFAULT;
 
-	if (cmpxchg_futex_value_locked(&curval, uaddr, uval, newval))
+	if (cmpxchg_futex_value_locked(&curval, uaddr, uval, newval)) {
 		ret = -EFAULT;
-	else if (curval != uval)
-		ret = -EINVAL;
+	} else if (curval != uval) {
+		/*
+		 * If a unconditional UNLOCK_PI operation (user space did not
+		 * try the TID->0 transition) raced with a waiter setting the
+		 * FUTEX_WAITERS flag between get_user() and locking the hash
+		 * bucket lock, retry the operation.
+		 */
+		if ((FUTEX_TID_MASK & curval) == uval)
+			ret = -EAGAIN;
+		else
+			ret = -EINVAL;
+	}
 	if (ret) {
 		raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock);
 		return ret;
@@ -1442,8 +1452,8 @@ void requeue_futex(struct futex_q *q, struct futex_hash_bucket *hb1,
 	if (likely(&hb1->chain != &hb2->chain)) {
 		plist_del(&q->list, &hb1->chain);
 		hb_waiters_dec(hb1);
-		plist_add(&q->list, &hb2->chain);
 		hb_waiters_inc(hb2);
+		plist_add(&q->list, &hb2->chain);
 		q->lock_ptr = &hb2->lock;
 	}
 	get_futex_key_refs(key2);
@@ -2536,6 +2546,15 @@ retry:
 		if (ret == -EFAULT)
 			goto pi_faulted;
 		/*
+		 * A unconditional UNLOCK_PI op raced against a waiter
+		 * setting the FUTEX_WAITERS bit. Try again.
+		 */
+		if (ret == -EAGAIN) {
+			spin_unlock(&hb->lock);
+			put_futex_key(&key);
+			goto retry;
+		}
+		/*
 		 * wake_futex_pi has detected invalid state. Tell user
 		 * space.
 		 */
diff --git a/kernel/locking/mcs_spinlock.h b/kernel/locking/mcs_spinlock.h
index 5b9102a47..c835270f0 100644
--- a/kernel/locking/mcs_spinlock.h
+++ b/kernel/locking/mcs_spinlock.h
@@ -67,7 +67,13 @@ void mcs_spin_lock(struct mcs_spinlock **lock, struct mcs_spinlock *node)
 	node->locked = 0;
 	node->next   = NULL;
 
-	prev = xchg_acquire(lock, node);
+	/*
+	 * We rely on the full barrier with global transitivity implied by the
+	 * below xchg() to order the initialization stores above against any
+	 * observation of @node. And to provide the ACQUIRE ordering associated
+	 * with a LOCK primitive.
+	 */
+	prev = xchg(lock, node);
 	if (likely(prev == NULL)) {
 		/*
 		 * Lock acquired, don't need to set node->locked to 1. Threads
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 544c98aac..5699f2dd2 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7555,7 +7555,7 @@ void set_curr_task(int cpu, struct task_struct *p)
 /* task_group_lock serializes the addition/removal of task groups */
 static DEFINE_SPINLOCK(task_group_lock);
 
-static void free_sched_group(struct task_group *tg)
+static void sched_free_group(struct task_group *tg)
 {
 	free_fair_sched_group(tg);
 	free_rt_sched_group(tg);
@@ -7581,7 +7581,7 @@ struct task_group *sched_create_group(struct task_group *parent)
 	return tg;
 
 err:
-	free_sched_group(tg);
+	sched_free_group(tg);
 	return ERR_PTR(-ENOMEM);
 }
 
@@ -7601,17 +7601,16 @@ void sched_online_group(struct task_group *tg, struct task_group *parent)
 }
 
 /* rcu callback to free various structures associated with a task group */
-static void free_sched_group_rcu(struct rcu_head *rhp)
+static void sched_free_group_rcu(struct rcu_head *rhp)
 {
 	/* now it should be safe to free those cfs_rqs */
-	free_sched_group(container_of(rhp, struct task_group, rcu));
+	sched_free_group(container_of(rhp, struct task_group, rcu));
 }
 
-/* Destroy runqueue etc associated with a task group */
 void sched_destroy_group(struct task_group *tg)
 {
 	/* wait for possible concurrent references to cfs_rqs complete */
-	call_rcu(&tg->rcu, free_sched_group_rcu);
+	call_rcu(&tg->rcu, sched_free_group_rcu);
 }
 
 void sched_offline_group(struct task_group *tg)
@@ -8070,31 +8069,26 @@ cpu_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
 	if (IS_ERR(tg))
 		return ERR_PTR(-ENOMEM);
 
+	sched_online_group(tg, parent);
+
 	return &tg->css;
 }
 
-static int cpu_cgroup_css_online(struct cgroup_subsys_state *css)
+static void cpu_cgroup_css_released(struct cgroup_subsys_state *css)
 {
 	struct task_group *tg = css_tg(css);
-	struct task_group *parent = css_tg(css->parent);
 
-	if (parent)
-		sched_online_group(tg, parent);
-	return 0;
+	sched_offline_group(tg);
 }
 
 static void cpu_cgroup_css_free(struct cgroup_subsys_state *css)
 {
 	struct task_group *tg = css_tg(css);
 
-	sched_destroy_group(tg);
-}
-
-static void cpu_cgroup_css_offline(struct cgroup_subsys_state *css)
-{
-	struct task_group *tg = css_tg(css);
-
-	sched_offline_group(tg);
+	/*
+	 * Relies on the RCU grace period between css_released() and this.
+	 */
+	sched_free_group(tg);
 }
 
 static void cpu_cgroup_fork(struct task_struct *task)
@@ -8454,9 +8448,8 @@ static struct cftype cpu_files[] = {
 
 struct cgroup_subsys cpu_cgrp_subsys = {
 	.css_alloc	= cpu_cgroup_css_alloc,
+	.css_released	= cpu_cgroup_css_released,
 	.css_free	= cpu_cgroup_css_free,
-	.css_online	= cpu_cgroup_css_online,
-	.css_offline	= cpu_cgroup_css_offline,
 	.fork		= cpu_cgroup_fork,
 	.can_attach	= cpu_cgroup_can_attach,
 	.attach		= cpu_cgroup_attach,
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 7ff5dc7d2..9e82d0450 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -667,6 +667,35 @@ static void set_work_pool_and_clear_pending(struct work_struct *work,
 	 */
 	smp_wmb();
 	set_work_data(work, (unsigned long)pool_id << WORK_OFFQ_POOL_SHIFT, 0);
+	/*
+	 * The following mb guarantees that previous clear of a PENDING bit
+	 * will not be reordered with any speculative LOADS or STORES from
+	 * work->current_func, which is executed afterwards.  This possible
+	 * reordering can lead to a missed execution on attempt to qeueue
+	 * the same @work.  E.g. consider this case:
+	 *
+	 *   CPU#0                         CPU#1
+	 *   ----------------------------  --------------------------------
+	 *
+	 * 1  STORE event_indicated
+	 * 2  queue_work_on() {
+	 * 3    test_and_set_bit(PENDING)
+	 * 4 }                             set_..._and_clear_pending() {
+	 * 5                                 set_work_data() # clear bit
+	 * 6                                 smp_mb()
+	 * 7                               work->current_func() {
+	 * 8				      LOAD event_indicated
+	 *				   }
+	 *
+	 * Without an explicit full barrier speculative LOAD on line 8 can
+	 * be executed before CPU#0 does STORE on line 1.  If that happens,
+	 * CPU#0 observes the PENDING bit is still set and new execution of
+	 * a @work is not queued in a hope, that CPU#1 will eventually
+	 * finish the queued @work.  Meanwhile CPU#1 does not see
+	 * event_indicated is set, because speculative LOAD was executed
+	 * before actual STORE.
+	 */
+	smp_mb();
 }
 
 static void clear_work_data(struct work_struct *work)