Merge remote-tracking branch 'ovl/rename2' into for-linus

6 files changed, 99 insertions, 30 deletions

diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index d1c51b7f5221..5e8dab5bf9ad 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -6270,6 +6270,12 @@ void cgroup_sk_alloc(struct sock_cgroup_data *skcd)
 	if (cgroup_sk_alloc_disabled)
 		return;
 
+	/* Socket clone path */
+	if (skcd->val) {
+		cgroup_get(sock_cgroup_ptr(skcd));
+		return;
+	}
+
 	rcu_read_lock();
 
 	while (true) {
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 3cfabdf7b942..fc9bb2225291 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -2496,11 +2496,11 @@ static int __perf_event_stop(void *info)
 	return 0;
 }
 
-static int perf_event_restart(struct perf_event *event)
+static int perf_event_stop(struct perf_event *event, int restart)
 {
 	struct stop_event_data sd = {
 		.event		= event,
-		.restart	= 1,
+		.restart	= restart,
 	};
 	int ret = 0;
 
@@ -3549,10 +3549,18 @@ static int perf_event_read(struct perf_event *event, bool group)
 			.group = group,
 			.ret = 0,
 		};
-		ret = smp_call_function_single(event->oncpu, __perf_event_read, &data, 1);
-		/* The event must have been read from an online CPU: */
-		WARN_ON_ONCE(ret);
-		ret = ret ? : data.ret;
+		/*
+		 * Purposely ignore the smp_call_function_single() return
+		 * value.
+		 *
+		 * If event->oncpu isn't a valid CPU it means the event got
+		 * scheduled out and that will have updated the event count.
+		 *
+		 * Therefore, either way, we'll have an up-to-date event count
+		 * after this.
+		 */
+		(void)smp_call_function_single(event->oncpu, __perf_event_read, &data, 1);
+		ret = data.ret;
 	} else if (event->state == PERF_EVENT_STATE_INACTIVE) {
 		struct perf_event_context *ctx = event->ctx;
 		unsigned long flags;
@@ -3921,7 +3929,7 @@ static void exclusive_event_destroy(struct perf_event *event)
 
 static bool exclusive_event_match(struct perf_event *e1, struct perf_event *e2)
 {
-	if ((e1->pmu->capabilities & PERF_PMU_CAP_EXCLUSIVE) &&
+	if ((e1->pmu == e2->pmu) &&
 	    (e1->cpu == e2->cpu ||
 	     e1->cpu == -1 ||
 	     e2->cpu == -1))
@@ -4837,6 +4845,19 @@ static void ring_buffer_attach(struct perf_event *event,
 		spin_unlock_irqrestore(&rb->event_lock, flags);
 	}
 
+	/*
+	 * Avoid racing with perf_mmap_close(AUX): stop the event
+	 * before swizzling the event::rb pointer; if it's getting
+	 * unmapped, its aux_mmap_count will be 0 and it won't
+	 * restart. See the comment in __perf_pmu_output_stop().
+	 *
+	 * Data will inevitably be lost when set_output is done in
+	 * mid-air, but then again, whoever does it like this is
+	 * not in for the data anyway.
+	 */
+	if (has_aux(event))
+		perf_event_stop(event, 0);
+
 	rcu_assign_pointer(event->rb, rb);
 
 	if (old_rb) {
@@ -6112,7 +6133,7 @@ static void perf_event_addr_filters_exec(struct perf_event *event, void *data)
 	raw_spin_unlock_irqrestore(&ifh->lock, flags);
 
 	if (restart)
-		perf_event_restart(event);
+		perf_event_stop(event, 1);
 }
 
 void perf_event_exec(void)
@@ -6156,7 +6177,13 @@ static void __perf_event_output_stop(struct perf_event *event, void *data)
 
 	/*
 	 * In case of inheritance, it will be the parent that links to the
-	 * ring-buffer, but it will be the child that's actually using it:
+	 * ring-buffer, but it will be the child that's actually using it.
+	 *
+	 * We are using event::rb to determine if the event should be stopped,
+	 * however this may race with ring_buffer_attach() (through set_output),
+	 * which will make us skip the event that actually needs to be stopped.
+	 * So ring_buffer_attach() has to stop an aux event before re-assigning
+	 * its rb pointer.
 	 */
 	if (rcu_dereference(parent->rb) == rb)
 		ro->err = __perf_event_stop(&sd);
@@ -6670,7 +6697,7 @@ static void __perf_addr_filters_adjust(struct perf_event *event, void *data)
 	raw_spin_unlock_irqrestore(&ifh->lock, flags);
 
 	if (restart)
-		perf_event_restart(event);
+		perf_event_stop(event, 1);
 }
 
 /*
@@ -7859,7 +7886,7 @@ static void perf_event_addr_filters_apply(struct perf_event *event)
 	mmput(mm);
 
 restart:
-	perf_event_restart(event);
+	perf_event_stop(event, 1);
 }
 
 /*
diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
index ae9b90dc9a5a..257fa460b846 100644
--- a/kernel/events/ring_buffer.c
+++ b/kernel/events/ring_buffer.c
@@ -330,15 +330,22 @@ void *perf_aux_output_begin(struct perf_output_handle *handle,
 	if (!rb)
 		return NULL;
 
-	if (!rb_has_aux(rb) || !atomic_inc_not_zero(&rb->aux_refcount))
+	if (!rb_has_aux(rb))
 		goto err;
 
 	/*
-	 * If rb::aux_mmap_count is zero (and rb_has_aux() above went through),
-	 * the aux buffer is in perf_mmap_close(), about to get freed.
+	 * If aux_mmap_count is zero, the aux buffer is in perf_mmap_close(),
+	 * about to get freed, so we leave immediately.
+	 *
+	 * Checking rb::aux_mmap_count and rb::refcount has to be done in
+	 * the same order, see perf_mmap_close. Otherwise we end up freeing
+	 * aux pages in this path, which is a bug, because in_atomic().
 	 */
 	if (!atomic_read(&rb->aux_mmap_count))
-		goto err_put;
+		goto err;
+
+	if (!atomic_inc_not_zero(&rb->aux_refcount))
+		goto err;
 
 	/*
 	 * Nesting is not supported for AUX area, make sure nested
diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c
index 637389088b3f..26ba5654d9d5 100644
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -820,6 +820,8 @@ __irq_do_set_handler(struct irq_desc *desc, irq_flow_handler_t handle,
 	desc->name = name;
 
 	if (handle != handle_bad_irq && is_chained) {
+		unsigned int type = irqd_get_trigger_type(&desc->irq_data);
+
 		/*
 		 * We're about to start this interrupt immediately,
 		 * hence the need to set the trigger configuration.
@@ -828,8 +830,10 @@ __irq_do_set_handler(struct irq_desc *desc, irq_flow_handler_t handle,
 		 * chained interrupt. Reset it immediately because we
 		 * do know better.
 		 */
-		__irq_set_trigger(desc, irqd_get_trigger_type(&desc->irq_data));
-		desc->handle_irq = handle;
+		if (type != IRQ_TYPE_NONE) {
+			__irq_set_trigger(desc, type);
+			desc->handle_irq = handle;
+		}
 
 		irq_settings_set_noprobe(desc);
 		irq_settings_set_norequest(desc);
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 2a906f20fba7..44817c640e99 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -2016,6 +2016,28 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
 	success = 1; /* we're going to change ->state */
 	cpu = task_cpu(p);
 
+	/*
+	 * Ensure we load p->on_rq _after_ p->state, otherwise it would
+	 * be possible to, falsely, observe p->on_rq == 0 and get stuck
+	 * in smp_cond_load_acquire() below.
+	 *
+	 * sched_ttwu_pending()                 try_to_wake_up()
+	 *   [S] p->on_rq = 1;                  [L] P->state
+	 *       UNLOCK rq->lock  -----.
+	 *                              \
+	 *				 +---   RMB
+	 * schedule()                   /
+	 *       LOCK rq->lock    -----'
+	 *       UNLOCK rq->lock
+	 *
+	 * [task p]
+	 *   [S] p->state = UNINTERRUPTIBLE     [L] p->on_rq
+	 *
+	 * Pairs with the UNLOCK+LOCK on rq->lock from the
+	 * last wakeup of our task and the schedule that got our task
+	 * current.
+	 */
+	smp_rmb();
 	if (p->on_rq && ttwu_remote(p, wake_flags))
 		goto stat;
 
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index dade4c9559cc..7bc56762ca35 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5124,19 +5124,20 @@ tracing_read_pipe(struct file *filp, char __user *ubuf,
 	struct trace_iterator *iter = filp->private_data;
 	ssize_t sret;
 
-	/* return any leftover data */
-	sret = trace_seq_to_user(&iter->seq, ubuf, cnt);
-	if (sret != -EBUSY)
-		return sret;
-
-	trace_seq_init(&iter->seq);
-
 	/*
 	 * Avoid more than one consumer on a single file descriptor
 	 * This is just a matter of traces coherency, the ring buffer itself
 	 * is protected.
 	 */
 	mutex_lock(&iter->mutex);
+
+	/* return any leftover data */
+	sret = trace_seq_to_user(&iter->seq, ubuf, cnt);
+	if (sret != -EBUSY)
+		goto out;
+
+	trace_seq_init(&iter->seq);
+
 	if (iter->trace->read) {
 		sret = iter->trace->read(iter, filp, ubuf, cnt, ppos);
 		if (sret)
@@ -6163,9 +6164,6 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos,
 		return -EBUSY;
 #endif
 
-	if (splice_grow_spd(pipe, &spd))
-		return -ENOMEM;
-
 	if (*ppos & (PAGE_SIZE - 1))
 		return -EINVAL;
 
@@ -6175,6 +6173,9 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos,
 		len &= PAGE_MASK;
 	}
 
+	if (splice_grow_spd(pipe, &spd))
+		return -ENOMEM;
+
  again:
 	trace_access_lock(iter->cpu_file);
 	entries = ring_buffer_entries_cpu(iter->trace_buffer->buffer, iter->cpu_file);
@@ -6232,19 +6233,21 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos,
 	/* did we read anything? */
 	if (!spd.nr_pages) {
 		if (ret)
-			return ret;
+			goto out;
 
+		ret = -EAGAIN;
 		if ((file->f_flags & O_NONBLOCK) || (flags & SPLICE_F_NONBLOCK))
-			return -EAGAIN;
+			goto out;
 
 		ret = wait_on_pipe(iter, true);
 		if (ret)
-			return ret;
+			goto out;
 
 		goto again;
 	}
 
 	ret = splice_to_pipe(pipe, &spd);
+out:
 	splice_shrink_spd(&spd);
 
 	return ret;