linux-kernel - Re: [PATCH] tracing: Avoid adding duplicated tracer options when update_tracer

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]

Message-ID: <20220323142129.4175-1-mark-pk.tsai@mediatek.com>
Date:   Wed, 23 Mar 2022 22:21:29 +0800
From:   Mark-PK Tsai <mark-pk.tsai@...iatek.com>
To:     <rostedt@...dmis.org>
CC:     <linux-arm-kernel@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>,
        <linux-mediatek@...ts.infradead.org>, <lkp@...el.com>,
        <lkp@...ts.01.org>, <mark-pk.tsai@...iatek.com>,
        <matthias.bgg@...il.com>, <mingo@...hat.com>,
        <oliver.sang@...el.com>, <yj.chiang@...iatek.com>
Subject: Re: [PATCH] tracing: Avoid adding duplicated tracer options when update_tracer_options is running in parallel

> On Wed, 23 Mar 2022 11:24:42 +0800
> Mark-PK Tsai <mark-pk.tsai@...iatek.com> wrote:
> 
> > When update_tracer_options is running in parallel,
> > tr->tops might be updated before the trace_types list traversal.
> > Let update_tracer_options traverse the trace_types list safely in
> > kernel init time and avoid the tr->tops update before it finish.
> 
> ??? Have you seen a bug here? I'm totally confused by this.

Sorry to make you confused.

After the below patch, update_tracer_options might be executed later than registering
hwlat_tracer, which is in late_initcall.

https://lore.kernel.org/lkml/20220316151639.9216-1-mark-pk.tsai@mediatek.com/

The init_hwlat_tracer initcall will put hwlat_tracer to tr->tops.
Then when the later arrived __update_tracer_options is trying to
update all the tracer options, create_trace_option_files show the
below warning because hwlat_tracer is already in the list.

[ 6.680068 ][ T7 ] WARNING: CPU: 0 PID: 7 at kernel/trace/trace.c:8899 create_trace_option_files (kernel/trace/trace.c:8899 (discriminator 1))

full log: https://lore.kernel.org/lkml/20220322133339.GA32582@xsang-OptiPlex-9020/


> 
> > 
> > Link: https://lore.kernel.org/lkml/20220322133339.GA32582@xsang-OptiPlex-9020/
> > Reported-by: kernel test robot <oliver.sang@...el.com>
> > Signed-off-by: Mark-PK Tsai <mark-pk.tsai@...iatek.com>
> > ---
> >  kernel/trace/trace.c | 7 +++++++
> >  1 file changed, 7 insertions(+)
> > 
> > diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
> > index adb37e437a05..2974ae056068 100644
> > --- a/kernel/trace/trace.c
> > +++ b/kernel/trace/trace.c
> > @@ -6317,12 +6317,18 @@ static void tracing_set_nop(struct trace_array *tr)
> >  	tr->current_trace = &nop_trace;
> >  }
> >  
> > +static bool tracer_options_updated;
> > +
> >  static void add_tracer_options(struct trace_array *tr, struct tracer *t)
> >  {
> >  	/* Only enable if the directory has been created already. */
> >  	if (!tr->dir)
> >  		return;
> >  
> > +	/* Only create trace option files after update_tracer_options finish */
> > +	if (!tracer_options_updated)
> > +		return;
> > +
> >  	create_trace_option_files(tr, t);
> >  }
> >  
> > @@ -9133,6 +9139,7 @@ static void update_tracer_options(struct trace_array *tr)
> >  {
> >  	mutex_lock(&trace_types_lock);
> 
> How is update_trace_options run in parallel?
> 
> There's a mutex that protects it. 
> 

Oh sorry.
What I trying to tell is that update_trace_options is run in parallel with
the initcall thread after:

https://lore.kernel.org/lkml/20220316151639.9216-1-mark-pk.tsai@mediatek.com/


> -- Steve
> 
> 
> >  	__update_tracer_options(tr);
> > +	tracer_options_updated = true;
> >  	mutex_unlock(&trace_types_lock);
> >  }
> >