5 files changed, 150 insertions, 48 deletions

diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index e5c063fc8ef9..3ec7df7dbeec 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -1828,7 +1828,7 @@ static struct pt_regs *get_bpf_raw_tp_regs(void)
 	struct bpf_raw_tp_regs *tp_regs = this_cpu_ptr(&bpf_raw_tp_regs);
 	int nest_level = this_cpu_inc_return(bpf_raw_tp_nest_level);
 
-	if (WARN_ON_ONCE(nest_level > ARRAY_SIZE(tp_regs->regs))) {
+	if (nest_level > ARRAY_SIZE(tp_regs->regs)) {
 		this_cpu_dec(bpf_raw_tp_nest_level);
 		return ERR_PTR(-EBUSY);
 	}
@@ -2932,6 +2932,9 @@ int bpf_kprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_prog *pr
 	if (sizeof(u64) != sizeof(void *))
 		return -EOPNOTSUPP;
 
+	if (attr->link_create.flags)
+		return -EINVAL;
+
 	if (!is_kprobe_multi(prog))
 		return -EINVAL;
 
@@ -3346,7 +3349,9 @@ int bpf_uprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_prog *pr
 	}
 
 	if (pid) {
+		rcu_read_lock();
 		task = get_pid_task(find_vpid(pid), PIDTYPE_TGID);
+		rcu_read_unlock();
 		if (!task) {
 			err = -ESRCH;
 			goto error_path_put;
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index baa5547e977a..6ab740d3185b 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -2796,6 +2796,12 @@ int ring_buffer_resize(struct trace_buffer *buffer, unsigned long size,
 	if (nr_pages < 2)
 		nr_pages = 2;
 
+	/*
+	 * Keep CPUs from coming online while resizing to synchronize
+	 * with new per CPU buffers being created.
+	 */
+	guard(cpus_read_lock)();
+
 	/* prevent another thread from changing buffer sizes */
 	mutex_lock(&buffer->mutex);
 	atomic_inc(&buffer->resizing);
@@ -2840,7 +2846,6 @@ int ring_buffer_resize(struct trace_buffer *buffer, unsigned long size,
 			cond_resched();
 		}
 
-		cpus_read_lock();
 		/*
 		 * Fire off all the required work handlers
 		 * We can't schedule on offline CPUs, but it's not necessary
@@ -2880,7 +2885,6 @@ int ring_buffer_resize(struct trace_buffer *buffer, unsigned long size,
 			cpu_buffer->nr_pages_to_update = 0;
 		}
 
-		cpus_read_unlock();
 	} else {
 		cpu_buffer = buffer->buffers[cpu_id];
 
@@ -2908,8 +2912,6 @@ int ring_buffer_resize(struct trace_buffer *buffer, unsigned long size,
 			goto out_err;
 		}
 
-		cpus_read_lock();
-
 		/* Can't run something on an offline CPU. */
 		if (!cpu_online(cpu_id))
 			rb_update_pages(cpu_buffer);
@@ -2928,7 +2930,6 @@ int ring_buffer_resize(struct trace_buffer *buffer, unsigned long size,
 		}
 
 		cpu_buffer->nr_pages_to_update = 0;
-		cpus_read_unlock();
 	}
 
  out:
@@ -6754,7 +6755,7 @@ int ring_buffer_subbuf_order_set(struct trace_buffer *buffer, int order)
 	old_size = buffer->subbuf_size;
 
 	/* prevent another thread from changing buffer sizes */
-	mutex_lock(&buffer->mutex);
+	guard(mutex)(&buffer->mutex);
 	atomic_inc(&buffer->record_disabled);
 
 	/* Make sure all commits have finished */
@@ -6859,7 +6860,6 @@ int ring_buffer_subbuf_order_set(struct trace_buffer *buffer, int order)
 	}
 
 	atomic_dec(&buffer->record_disabled);
-	mutex_unlock(&buffer->mutex);
 
 	return 0;
 
@@ -6868,7 +6868,6 @@ error:
 	buffer->subbuf_size = old_size;
 
 	atomic_dec(&buffer->record_disabled);
-	mutex_unlock(&buffer->mutex);
 
 	for_each_buffer_cpu(buffer, cpu) {
 		cpu_buffer = buffer->buffers[cpu];
@@ -7274,8 +7273,8 @@ consume:
 	/* Check if any events were dropped */
 	missed_events = cpu_buffer->lost_events;
 
-	if (cpu_buffer->reader_page != cpu_buffer->commit_page) {
-		if (missed_events) {
+	if (missed_events) {
+		if (cpu_buffer->reader_page != cpu_buffer->commit_page) {
 			struct buffer_data_page *bpage = reader->page;
 			unsigned int commit;
 			/*
@@ -7296,13 +7295,23 @@ consume:
 				local_add(RB_MISSED_STORED, &bpage->commit);
 			}
 			local_add(RB_MISSED_EVENTS, &bpage->commit);
+		} else if (!WARN_ONCE(cpu_buffer->reader_page == cpu_buffer->tail_page,
+				      "Reader on commit with %ld missed events",
+				      missed_events)) {
+			/*
+			 * There shouldn't be any missed events if the tail_page
+			 * is on the reader page. But if the tail page is not on the
+			 * reader page and the commit_page is, that would mean that
+			 * there's a commit_overrun (an interrupt preempted an
+			 * addition of an event and then filled the buffer
+			 * with new events). In this case it's not an
+			 * error, but it should still be reported.
+			 *
+			 * TODO: Add missed events to the page for user space to know.
+			 */
+			pr_info("Ring buffer [%d] commit overrun lost %ld events at timestamp:%lld\n",
+				cpu, missed_events, cpu_buffer->reader_page->page->time_stamp);
 		}
-	} else {
-		/*
-		 * There really shouldn't be any missed events if the commit
-		 * is on the reader page.
-		 */
-		WARN_ON_ONCE(missed_events);
 	}
 
 	cpu_buffer->lost_events = 0;
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 82da3ac14024..57e1af1d3e6d 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -1731,6 +1731,9 @@ extern int event_enable_register_trigger(char *glob,
 extern void event_enable_unregister_trigger(char *glob,
 					    struct event_trigger_data *test,
 					    struct trace_event_file *file);
+extern struct event_trigger_data *
+trigger_data_alloc(struct event_command *cmd_ops, char *cmd, char *param,
+		   void *private_data);
 extern void trigger_data_free(struct event_trigger_data *data);
 extern int event_trigger_init(struct event_trigger_data *data);
 extern int trace_event_trigger_enable_disable(struct trace_event_file *file,
@@ -1757,11 +1760,6 @@ extern bool event_trigger_check_remove(const char *glob);
 extern bool event_trigger_empty_param(const char *param);
 extern int event_trigger_separate_filter(char *param_and_filter, char **param,
 					 char **filter, bool param_required);
-extern struct event_trigger_data *
-event_trigger_alloc(struct event_command *cmd_ops,
-		    char *cmd,
-		    char *param,
-		    void *private_data);
 extern int event_trigger_parse_num(char *trigger,
 				   struct event_trigger_data *trigger_data);
 extern int event_trigger_set_filter(struct event_command *cmd_ops,
diff --git a/kernel/trace/trace_events_hist.c b/kernel/trace/trace_events_hist.c
index 4ebafc655223..3379e14d38e9 100644
--- a/kernel/trace/trace_events_hist.c
+++ b/kernel/trace/trace_events_hist.c
@@ -5249,17 +5249,94 @@ hist_trigger_actions(struct hist_trigger_data *hist_data,
 	}
 }
 
+/*
+ * The hist_pad structure is used to save information to create
+ * a histogram from the histogram trigger. It's too big to store
+ * on the stack, so when the histogram trigger is initialized
+ * a percpu array of 4 hist_pad structures is allocated.
+ * This will cover every context from normal, softirq, irq and NMI
+ * in the very unlikely event that a tigger happens at each of
+ * these contexts and interrupts a currently active trigger.
+ */
+struct hist_pad {
+	unsigned long		entries[HIST_STACKTRACE_DEPTH];
+	u64			var_ref_vals[TRACING_MAP_VARS_MAX];
+	char			compound_key[HIST_KEY_SIZE_MAX];
+};
+
+static struct hist_pad __percpu *hist_pads;
+static DEFINE_PER_CPU(int, hist_pad_cnt);
+static refcount_t hist_pad_ref;
+
+/* One hist_pad for every context (normal, softirq, irq, NMI) */
+#define MAX_HIST_CNT 4
+
+static int alloc_hist_pad(void)
+{
+	lockdep_assert_held(&event_mutex);
+
+	if (refcount_read(&hist_pad_ref)) {
+		refcount_inc(&hist_pad_ref);
+		return 0;
+	}
+
+	hist_pads = __alloc_percpu(sizeof(struct hist_pad) * MAX_HIST_CNT,
+				   __alignof__(struct hist_pad));
+	if (!hist_pads)
+		return -ENOMEM;
+
+	refcount_set(&hist_pad_ref, 1);
+	return 0;
+}
+
+static void free_hist_pad(void)
+{
+	lockdep_assert_held(&event_mutex);
+
+	if (!refcount_dec_and_test(&hist_pad_ref))
+		return;
+
+	free_percpu(hist_pads);
+	hist_pads = NULL;
+}
+
+static struct hist_pad *get_hist_pad(void)
+{
+	struct hist_pad *hist_pad;
+	int cnt;
+
+	if (WARN_ON_ONCE(!hist_pads))
+		return NULL;
+
+	preempt_disable();
+
+	hist_pad = per_cpu_ptr(hist_pads, smp_processor_id());
+
+	if (this_cpu_read(hist_pad_cnt) == MAX_HIST_CNT) {
+		preempt_enable();
+		return NULL;
+	}
+
+	cnt = this_cpu_inc_return(hist_pad_cnt) - 1;
+
+	return &hist_pad[cnt];
+}
+
+static void put_hist_pad(void)
+{
+	this_cpu_dec(hist_pad_cnt);
+	preempt_enable();
+}
+
 static void event_hist_trigger(struct event_trigger_data *data,
 			       struct trace_buffer *buffer, void *rec,
 			       struct ring_buffer_event *rbe)
 {
 	struct hist_trigger_data *hist_data = data->private_data;
 	bool use_compound_key = (hist_data->n_keys > 1);
-	unsigned long entries[HIST_STACKTRACE_DEPTH];
-	u64 var_ref_vals[TRACING_MAP_VARS_MAX];
-	char compound_key[HIST_KEY_SIZE_MAX];
 	struct tracing_map_elt *elt = NULL;
 	struct hist_field *key_field;
+	struct hist_pad *hist_pad;
 	u64 field_contents;
 	void *key = NULL;
 	unsigned int i;
@@ -5267,12 +5344,18 @@ static void event_hist_trigger(struct event_trigger_data *data,
 	if (unlikely(!rbe))
 		return;
 
-	memset(compound_key, 0, hist_data->key_size);
+	hist_pad = get_hist_pad();
+	if (!hist_pad)
+		return;
+
+	memset(hist_pad->compound_key, 0, hist_data->key_size);
 
 	for_each_hist_key_field(i, hist_data) {
 		key_field = hist_data->fields[i];
 
 		if (key_field->flags & HIST_FIELD_FL_STACKTRACE) {
+			unsigned long *entries = hist_pad->entries;
+
 			memset(entries, 0, HIST_STACKTRACE_SIZE);
 			if (key_field->field) {
 				unsigned long *stack, n_entries;
@@ -5296,26 +5379,31 @@ static void event_hist_trigger(struct event_trigger_data *data,
 		}
 
 		if (use_compound_key)
-			add_to_key(compound_key, key, key_field, rec);
+			add_to_key(hist_pad->compound_key, key, key_field, rec);
 	}
 
 	if (use_compound_key)
-		key = compound_key;
+		key = hist_pad->compound_key;
 
 	if (hist_data->n_var_refs &&
-	    !resolve_var_refs(hist_data, key, var_ref_vals, false))
-		return;
+	    !resolve_var_refs(hist_data, key, hist_pad->var_ref_vals, false))
+		goto out;
 
 	elt = tracing_map_insert(hist_data->map, key);
 	if (!elt)
-		return;
+		goto out;
 
-	hist_trigger_elt_update(hist_data, elt, buffer, rec, rbe, var_ref_vals);
+	hist_trigger_elt_update(hist_data, elt, buffer, rec, rbe, hist_pad->var_ref_vals);
 
-	if (resolve_var_refs(hist_data, key, var_ref_vals, true))
-		hist_trigger_actions(hist_data, elt, buffer, rec, rbe, key, var_ref_vals);
+	if (resolve_var_refs(hist_data, key, hist_pad->var_ref_vals, true)) {
+		hist_trigger_actions(hist_data, elt, buffer, rec, rbe,
+				     key, hist_pad->var_ref_vals);
+	}
 
 	hist_poll_wakeup();
+
+ out:
+	put_hist_pad();
 }
 
 static void hist_trigger_stacktrace_print(struct seq_file *m,
@@ -6160,6 +6248,9 @@ static int event_hist_trigger_init(struct event_trigger_data *data)
 {
 	struct hist_trigger_data *hist_data = data->private_data;
 
+	if (alloc_hist_pad() < 0)
+		return -ENOMEM;
+
 	if (!data->ref && hist_data->attrs->name)
 		save_named_trigger(hist_data->attrs->name, data);
 
@@ -6204,6 +6295,7 @@ static void event_hist_trigger_free(struct event_trigger_data *data)
 
 		destroy_hist_data(hist_data);
 	}
+	free_hist_pad();
 }
 
 static struct event_trigger_ops event_hist_trigger_ops = {
@@ -6219,9 +6311,7 @@ static int event_hist_trigger_named_init(struct event_trigger_data *data)
 
 	save_named_trigger(data->named_data->name, data);
 
-	event_hist_trigger_init(data->named_data);
-
-	return 0;
+	return event_hist_trigger_init(data->named_data);
 }
 
 static void event_hist_trigger_named_free(struct event_trigger_data *data)
@@ -6708,7 +6798,7 @@ static int event_hist_trigger_parse(struct event_command *cmd_ops,
 		return PTR_ERR(hist_data);
 	}
 
-	trigger_data = event_trigger_alloc(cmd_ops, cmd, param, hist_data);
+	trigger_data = trigger_data_alloc(cmd_ops, cmd, param, hist_data);
 	if (!trigger_data) {
 		ret = -ENOMEM;
 		goto out_free;
diff --git a/kernel/trace/trace_events_trigger.c b/kernel/trace/trace_events_trigger.c
index 27e21488d574..d5dbda9b0e4b 100644
--- a/kernel/trace/trace_events_trigger.c
+++ b/kernel/trace/trace_events_trigger.c
@@ -825,7 +825,7 @@ out:
 }
 
 /**
- * event_trigger_alloc - allocate and init event_trigger_data for a trigger
+ * trigger_data_alloc - allocate and init event_trigger_data for a trigger
  * @cmd_ops: The event_command operations for the trigger
  * @cmd: The cmd string
  * @param: The param string
@@ -836,14 +836,14 @@ out:
  * trigger_ops to assign to the event_trigger_data.  @private_data can
  * also be passed in and associated with the event_trigger_data.
  *
- * Use event_trigger_free() to free an event_trigger_data object.
+ * Use trigger_data_free() to free an event_trigger_data object.
  *
  * Return: The trigger_data object success, NULL otherwise
  */
-struct event_trigger_data *event_trigger_alloc(struct event_command *cmd_ops,
-					       char *cmd,
-					       char *param,
-					       void *private_data)
+struct event_trigger_data *trigger_data_alloc(struct event_command *cmd_ops,
+					      char *cmd,
+					      char *param,
+					      void *private_data)
 {
 	struct event_trigger_data *trigger_data;
 	struct event_trigger_ops *trigger_ops;
@@ -1010,13 +1010,13 @@ event_trigger_parse(struct event_command *cmd_ops,
 		return ret;
 
 	ret = -ENOMEM;
-	trigger_data = event_trigger_alloc(cmd_ops, cmd, param, file);
+	trigger_data = trigger_data_alloc(cmd_ops, cmd, param, file);
 	if (!trigger_data)
 		goto out;
 
 	if (remove) {
 		event_trigger_unregister(cmd_ops, file, glob+1, trigger_data);
-		kfree(trigger_data);
+		trigger_data_free(trigger_data);
 		ret = 0;
 		goto out;
 	}
@@ -1043,7 +1043,7 @@ event_trigger_parse(struct event_command *cmd_ops,
 
  out_free:
 	event_trigger_reset_filter(cmd_ops, trigger_data);
-	kfree(trigger_data);
+	trigger_data_free(trigger_data);
 	goto out;
 }
 
@@ -1814,7 +1814,7 @@ int event_enable_trigger_parse(struct event_command *cmd_ops,
 	enable_data->enable = enable;
 	enable_data->file = event_enable_file;
 
-	trigger_data = event_trigger_alloc(cmd_ops, cmd, param, enable_data);
+	trigger_data = trigger_data_alloc(cmd_ops, cmd, param, enable_data);
 	if (!trigger_data) {
 		kfree(enable_data);
 		goto out;