linux-kernel - Re: [PATCH V3] tracing: Make latency tracers fully support the set_graph

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:	Tue, 28 Jun 2016 14:20:02 +0900
From:	Namhyung Kim <namhyung@...il.com>
To:	Chunyu Hu <chuhu@...hat.com>
Cc:	Steven Rostedt <rostedt@...dmis.org>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH V3] tracing: Make latency tracers fully support the set_graph_notrace

Hello,

On Fri, Jun 24, 2016 at 7:55 PM, Chunyu Hu <chuhu@...hat.com> wrote:
> latency tracers(wakeup, wakeup_rt, wakeup_dl, irqsoff) can use the
> function_graph trace when display_graph trace option is set by user
> via tracefs. And currently the set_graph_notrace filter is not fully
> supported in latency tracers, only the graph_ret event can be filtered,
> the graph_ent events will always be submitted to the trace ring buffer
> without respecting to the filter.
>
> The issue is that the submitted graph_entry event that matches the
> filter can be assigned with a negative depth(minuts FTRACE_NOTRACE_DEPTH)
> which will be used as the array index of fgraph_cpu_data when printing
> trace entries, as a result, an oops can be hit when accessing the array.
>
> Fully supporting the set_graph_notrace filter in latency tracers can
> avoid this oops and provide a small enhancement for these tracers at
> the same time.
>
> To reproduce the oops:
> echo 1 > options/display_graph
> echo schedule > set_graph_notrace
> echo wakeup > current_tracer
> cat trace (several times)

I'm unabled to reproduce the oops even after running 'cat trace' multiple times.

Anyway, the patch looks good to me.

Acked-by: Namhyung Kim <namhyung@...nel.org>

Thanks,
Namhyung


>
> Signed-off-by: Chunyu Hu <chuhu@...hat.com>
> ---
>  kernel/trace/trace_irqsoff.c      | 6 ++++++
>  kernel/trace/trace_sched_wakeup.c | 6 ++++++
>  2 files changed, 12 insertions(+)
>
> diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c
> index 03cdff8..a4ed46a 100644
> --- a/kernel/trace/trace_irqsoff.c
> +++ b/kernel/trace/trace_irqsoff.c
> @@ -175,6 +175,12 @@ static int irqsoff_graph_entry(struct ftrace_graph_ent *trace)
>         int ret;
>         int pc;
>
> +       if (trace->depth < 0)
> +               return 0;
> +
> +       if (ftrace_graph_notrace_addr(trace->func))
> +               return 1;
> +
>         if (!func_prolog_dec(tr, &data, &flags))
>                 return 0;
>
> diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c
> index 9d4399b..e54fff7 100644
> --- a/kernel/trace/trace_sched_wakeup.c
> +++ b/kernel/trace/trace_sched_wakeup.c
> @@ -239,6 +239,12 @@ static int wakeup_graph_entry(struct ftrace_graph_ent *trace)
>         unsigned long flags;
>         int pc, ret = 0;
>
> +       if (trace->depth < 0)
> +               return 0;
> +
> +       if (ftrace_graph_notrace_addr(trace->func))
> +               return 1;
> +
>         if (!func_prolog_preempt_disable(tr, &data, &pc))
>                 return 0;
>
> --
> 1.8.3.1
>