Re: [PATCH v4 0/7] tracing: 'hist' triggers

[Daniel Wagner]

Hi Tom,

On 04/21/2015 12:03 AM, Tom Zanussi wrote:
> On Mon, 2015-04-20 at 14:52 +0200, Daniel Wagner wrote:
>> On 04/10/2015 06:05 PM, Tom Zanussi wrote:
>> So I am wondering if the path from ftrace_trigger_soft_disabled()
>> to event_triggers_call() is supposed never to happen? The comment
>> on ftrace_trigger_soft_disabled() indicate this might happen on
>> normal operation:
>>
> 
> So I looked at this on the plane and you're right, this is a path that
> should never be taken in this case, since the hist trigger does set the
> post_trigger flag and should therefore TRIGGER_COND would be true and
> that block should never be entered.
> 
> However, the code that registers the trigger first enables the event,
> then sets the cond flag, which would allow this to happen - a trigger
> expecting data would be passed a NULL rec.
> 
> Can you try the patch below and see if it fixes the problem? (untested
> and I haven't even had a chance to compile-test it..)

Thanks for the explanation. With that and your patch I could get
it working. There are a couple of calling places missing in your
patch. With them it works nicely. Maybe moving update_cond_flag()
inside trace_event_trigger_enable_disable() would be an option?

cheers,
daniel


diff --git a/kernel/trace/trace_events_hist.c b/kernel/trace/trace_events_hist.c
index 4d8f5df..221ab1f 100644
--- a/kernel/trace/trace_events_hist.c
+++ b/kernel/trace/trace_events_hist.c
@@ -1471,11 +1471,12 @@ static int hist_register_trigger(char *glob, struct event_trigger_ops *ops,
 	list_add_rcu(&data->list, &file->triggers);
 	ret++;
 
+	update_cond_flag(file);
 	if (trace_event_trigger_enable_disable(file, 1) < 0) {
 		list_del_rcu(&data->list);
+		update_cond_flag(file);
 		ret--;
 	}
-	update_cond_flag(file);
  out:
 	return ret;
 }
diff --git a/kernel/trace/trace_events_trigger.c b/kernel/trace/trace_events_trigger.c
index 8c8f2ee..e47a229 100644
--- a/kernel/trace/trace_events_trigger.c
+++ b/kernel/trace/trace_events_trigger.c
@@ -1264,6 +1264,7 @@ int event_enable_trigger_func(struct event_command *cmd_ops,
 		goto out_free;
 	}
 
+	update_cond_flag(file);
 	ret = trace_event_enable_disable(event_enable_file, 1, 1);
 	if (ret < 0)
 		goto out_put;
@@ -1286,6 +1287,7 @@ int event_enable_trigger_func(struct event_command *cmd_ops,
  out_disable:
 	trace_event_enable_disable(event_enable_file, 0, 1);
  out_put:
+	update_cond_flag(file);
 	module_put(event_enable_file->event_call->mod);
  out_free:
 	if (cmd_ops->set_filter)



 
> Tom
> 
> [PATCH] tracing: Update cond flag before enabling or disabling a trigger
> 
> Before a trigger is enabled, the cond flag should be set beforehand,
> otherwise an trigger that's expecting to process a trace record
> (e.g. one with post_trigger set) could be invoked without one.
> 
> Likewise a trigger's cond flag should be reset after it's disabled,
> not before.
> 
> Signed-off-by: Tom Zanussi <tom.zanussi@xxxxxxxxxxxxxxx>
> ---
>  kernel/trace/trace_events_trigger.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)
> 
> diff --git a/kernel/trace/trace_events_trigger.c b/kernel/trace/trace_events_trigger.c
> index 7105f15..108e46c 100644
> --- a/kernel/trace/trace_events_trigger.c
> +++ b/kernel/trace/trace_events_trigger.c
> @@ -546,11 +546,12 @@ static int register_trigger(char *glob, struct event_trigger_ops *ops,
>  	list_add_rcu(&data->list, &file->triggers);
>  	ret++;
>  
> +	update_cond_flag(file);
>  	if (trace_event_trigger_enable_disable(file, 1) < 0) {
>  		list_del_rcu(&data->list);
> +		update_cond_flag(file);
>  		ret--;
>  	}
> -	update_cond_flag(file);
>  out:
>  	return ret;
>  }
> @@ -578,8 +579,8 @@ void unregister_trigger(char *glob, struct event_trigger_ops *ops,
>  		if (data->cmd_ops->trigger_type == test->cmd_ops->trigger_type) {
>  			unregistered = true;
>  			list_del_rcu(&data->list);
> -			update_cond_flag(file);
>  			trace_event_trigger_enable_disable(file, 0);
> +			update_cond_flag(file);
>  			break;
>  		}
>  	}
> @@ -1324,11 +1325,12 @@ int event_enable_register_trigger(char *glob,
>  	list_add_rcu(&data->list, &file->triggers);
>  	ret++;
>  
> +	update_cond_flag(file);
>  	if (trace_event_trigger_enable_disable(file, 1) < 0) {
>  		list_del_rcu(&data->list);
> +		update_cond_flag(file);
>  		ret--;
>  	}
> -	update_cond_flag(file);
>  out:
>  	return ret;
>  }
> @@ -1351,8 +1353,8 @@ void event_enable_unregister_trigger(char *glob,
>  		    (enable_data->file == test_enable_data->file)) {
>  			unregistered = true;
>  			list_del_rcu(&data->list);
> -			update_cond_flag(file);
>  			trace_event_trigger_enable_disable(file, 0);
> +			update_cond_flag(file);
>  			break;
>  		}
>  	}
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/