linux-kernel - Re: [PATCH] tracing: let user specify tracing_thresh after selecting function

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20140603185720.79552fb8@gandalf.local.home>
Date:	Tue, 3 Jun 2014 18:57:20 -0400
From:	Steven Rostedt <rostedt@...dmis.org>
To:	Stanislav Fomichev <stfomichev@...dex-team.ru>
Cc:	fweisbec@...il.com, mingo@...hat.com, jovi.zhangwei@...wei.com,
	namhyung.kim@....com, tom.zanussi@...ux.intel.com,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] tracing: let user specify tracing_thresh after
 selecting function_graph

On Tue,  3 Jun 2014 21:12:49 +0400
Stanislav Fomichev <stfomichev@...dex-team.ru> wrote:

> Currently, tracing_thresh works only if we specify it before selecting
> function_graph tracer. If we do the opposite, tracing_thresh will change
> it's value, but it will not be applied.
> To fix it, we need to always register handlers which take tracing_thresh into
> account.
> 
> Signed-off-by: Stanislav Fomichev <stfomichev@...dex-team.ru>
> ---
>  kernel/trace/trace_functions_graph.c | 8 ++------
>  1 file changed, 2 insertions(+), 6 deletions(-)
> 
> diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c
> index deff11200261..a894031986a3 100644
> --- a/kernel/trace/trace_functions_graph.c
> +++ b/kernel/trace/trace_functions_graph.c
> @@ -433,12 +433,8 @@ static int graph_trace_init(struct trace_array *tr)
>  	int ret;
>  
>  	set_graph_array(tr);
> -	if (tracing_thresh)
> -		ret = register_ftrace_graph(&trace_graph_thresh_return,
> -					    &trace_graph_thresh_entry);
> -	else
> -		ret = register_ftrace_graph(&trace_graph_return,
> -					    &trace_graph_entry);
> +	ret = register_ftrace_graph(&trace_graph_thresh_return,
> +				    &trace_graph_thresh_entry);
>  	if (ret)
>  		return ret;
>  	tracing_start_cmdline_record();

I would rather not endure the overhead of checking tracing_thresh
every time. I never use it, I didn't even know there was someone that
did ;-)

I'm not saying you can't fix this bug, I'm just saying that it
shouldn't be done this way.

Ideally there should be a callback for the tracer to deal with it, but
instead, you can for now just hardcode a function graph tracer callback
into tracing_max_lat_write(). Something like this:

	if (ptr == &tracing_thresh)
		ftrace_update_func_graph_thresh();

And then if the function tracer is active, stop it, and then restart it
with the tracing thresh function.

-- Steve
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/