Re: [PATCH for tip][bugfix] ftrace: fix deadlock when setting set_ftrace_pid

[Lai Jiangshan]

KOSAKI Motohiro wrote:
> Impact: fix deadlock bug
> 
> reproduce way:
> 	# cd /sys/kernel/debug/tracing
> 	# echo $$ > set_ftrace_pid
> 
> 	then, console become hangup.
> 
> detail:
> when writing set_ftracepid, kernel callstack is following
> 
> 	ftrace_pid_write()
> 		mutex_lock(&ftrace_lock);
> 		ftrace_update_pid_func()
> 			mutex_lock(&ftrace_lock);
> 			mutex_unlock(&ftrace_lock);
> 		mutex_unlock(&ftrace_lock);
> 
> then, system alaways deadlock when ftrace_pid_write() is called.
> 
> In past days, ftrace_pid_write() use ftrace_start_lock, but
> commit e6ea44e9b4c12325337cd1c06103cd515a1c02b2 consolidate
> ftrace_start_lock to ftrace_lock.
> 
> 
> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
> ---
> diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> index 5a3a06b..0d3b393 100644
> --- a/kernel/trace/ftrace.c
> +++ b/kernel/trace/ftrace.c
> @@ -218,10 +218,8 @@ static void ftrace_update_pid_func(void)
>  {
>  	ftrace_func_t func;
>  
> -	mutex_lock(&ftrace_lock);
> -
>  	if (ftrace_trace_function == ftrace_stub)
> -		goto out;
> +		return;
>  
>  	func = ftrace_trace_function;
>  
> @@ -238,9 +236,6 @@ static void ftrace_update_pid_func(void)
>  #else
>  	__ftrace_trace_function = func;
>  #endif
> -
> - out:
> -	mutex_unlock(&ftrace_lock);
>  }
>  
>  /* set when tracing only a pid */
> 
> 

Reviewed-By: Lai Jiangshan <laijs@xxxxxxxxxxxxxx>


--
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/