Re: [PATCH v4 3/3] media: atomisp: Only use trace_printk if allowed

From: Nicolas Boichat <drinkcat@chromium.org>
To: Joe Perches <joe@perches.com>
Cc: Steven Rostedt <rostedt@goodmis.org>,
	Mauro Carvalho Chehab <mchehab@kernel.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	Sakari Ailus <sakari.ailus@linux.intel.com>,
	devel@driverdev.osuosl.org, lkml <linux-kernel@vger.kernel.org>,
	Linux Media Mailing List <linux-media@vger.kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Josh Poimboeuf <jpoimboe@redhat.com>,
	Douglas Anderson <dianders@chromium.org>,
	Guenter Roeck <groeck@chromium.org>
Subject: Re: [PATCH v4 3/3] media: atomisp: Only use trace_printk if allowed
Date: Fri, 21 Aug 2020 10:42:14 +0800	[thread overview]
Message-ID: <CANMq1KDYBbtrrCw6YUeoAPeHyet3L7qM3di9zmULDDRaQR_fZA@mail.gmail.com> (raw)
In-Reply-To: <f9d33bcaa2eda680752205d3c3690cb6bc421730.camel@perches.com>

On Fri, Aug 21, 2020 at 10:36 AM Joe Perches <joe@perches.com> wrote:
>
> On Thu, 2020-08-20 at 21:57 -0400, Steven Rostedt wrote:
> > On Fri, 21 Aug 2020 09:39:19 +0800
> > Nicolas Boichat <drinkcat@chromium.org> wrote:
> []
> > > Some other approaches/ideas:
> > >  1. Filter all lkml messages that contain trace_printk. Already found
> > > 1 instance, and I can easily reply to those with a semi-canned answer,
> > > if I remember to check that filter regularly (not sustainable in the
> > > long run...).
> >
> > Added Joe Perches to the thread.
> >
> > We can update checkpatch.pl to complain about a trace_printk() that it
> > finds in the added code.
>
> Why?
>
> I don't see much value in a trace_printk checkpatch warning.
> tracing is still dependent on CONFIG_TRACING otherwise
> trace_printk is an if (0)
>
> ELI5 please.

This is my "new" canned answer to this:

Please do not use trace_printk in production code [1,2], it is only
meant for debug use. Consider using trace events, or dev_dbg.
[1] https://elixir.bootlin.com/linux/v5.8/source/kernel/trace/trace.c#L3158
[2] https://elixir.bootlin.com/linux/v5.8/source/include/linux/kernel.h#L766

I also had arguments in patch 2/3 notes:

There's at least 3 reasons that I can come up with:
 1. trace_printk introduces some overhead. [some users, e.g.
Android/Chrome OS, want CONFIG_TRACING but _not_ that extra overhead]
 2. If the kernel keeps adding always-enabled trace_printk, it will be
much harder for developers to make use of trace_printk for debugging.
 3. People may assume that trace_printk is for debugging only, and may
accidentally output sensitive data (theoretical at this stage).

(we'll need to summarize that somehow if we want to add to checkpatch.pl)