Re: ftrace/perf_event leak

[Peter Zijlstra]

On Wed, 2010-09-01 at 17:38 +0800, Li Zefan wrote:

> ---
> diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c
> index 8a2b73f..9d1d1f2 100644
> --- a/kernel/trace/trace_event_perf.c
> +++ b/kernel/trace/trace_event_perf.c
> @@ -161,6 +161,7 @@ void perf_trace_destroy(struct perf_event *p_event)
>  		}
>  	}
>  out:
> +	module_put(tp_event->mod);
>  	mutex_unlock(&event_mutex);
>  }

Ah, indeed, we so that try_module_get() for each reference to the
tracepoint, I guess we also should do the below, in case the
registration fails.

diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c
index 000e6e8..31cc4cb 100644
--- a/kernel/trace/trace_event_perf.c
+++ b/kernel/trace/trace_event_perf.c
@@ -91,6 +91,8 @@ int perf_trace_init(struct perf_event *p_event)
 		    tp_event->class && tp_event->class->reg &&
 		    try_module_get(tp_event->mod)) {
 			ret = perf_trace_event_init(tp_event, p_event);
+			if (ret)
+				module_put(tp_event->mod);
 			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/