Re: [PATCH 3/4] tracing: Allow mmio tracer to display trace_printk() and other events

[Steven Rostedt <rostedt@goodmis.org>]

On Thu, 2010-05-13 at 11:54 +0300, Pekka Paalanen wrote:
> On Wed, 12 May 2010 21:21:13 -0400
> Steven Rostedt <rostedt@goodmis.org> wrote:
> 
> > From: Steven Rostedt <srostedt@redhat.com>
> > 
> > The mmio tracer has its own function to handle reading of events.
> > But if it encounters an event that it does not understand it
> > ignores it instead of telling the calling function that it is not
> > processing it.
> > 
> > If someone adds trace_printk() or enables events along with the
> > mmio tracer, then these events will not be displayed in the trace
> > output.
> > 
> > Simple solution is to just have the mmio print return UNHANDLED to
> > let the caller know that it did not processes the event and the
> > caller can process the event further.
> 
> Does this not mean that the mmiotrace output may contain
> foreign lines? If it does, it will break the user space.
> The dump format is specified in
> Documentation/trace/mmiotrace.txt.
> 
> If you want to handle arbitrary messages, format them as
> MARK events, please.
> 
> If I understood correctly, then NAK for this patch. Otherwise,
> could you explain how this does not break the mmiotrace dump
> format?
> 
> Is the tracing infrastructure now supporting several
> active tracers at the same time? If yes, and if mmiotrace
> should be able to co-operate, we need a new revision of the
> dump format, or a tool that extracts the mmiotrace
> events in the current format.
> 

It only displays other events if the user enabled those events.

But that said, I don't want to break existing userspace tools. I can add
a mmiotrace option "mmiotrace_all_events", if the user wants to see all
events within the mmiotracer then they can just enable that option,
otherwise, the mmiotracer will act like it currently does.

How does that sound?

-- Steve

> Thanks.
> 
> > Reported-by: Larry Finger <Larry.Finger@lwfinger.net>
> > Cc: Pekka Paalanen <pq@iki.fi>
> > Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
> > ---
> >  kernel/trace/trace_mmiotrace.c |    3 ++-
> >  1 files changed, 2 insertions(+), 1 deletions(-)
> > 
> > diff --git a/kernel/trace/trace_mmiotrace.c
> > b/kernel/trace/trace_mmiotrace.c index 017fa37..592c00f 100644
> > --- a/kernel/trace/trace_mmiotrace.c
> > +++ b/kernel/trace/trace_mmiotrace.c
> > @@ -282,7 +282,8 @@ static enum print_line_t
> > mmio_print_line(struct trace_iterator *iter) case TRACE_PRINT:
> >  		return mmio_print_mark(iter);
> >  	default:
> > -		return TRACE_TYPE_HANDLED; /* ignore unknown
> > entries */
> > +		/* Not our event */
> > +		return TRACE_TYPE_UNHANDLED;
> >  	}
> >  }
> >  
> > -- 
> > 1.7.0
>