Re: [RFC][PATCH] tracing: Add __cpumask() macro to trace events to record cpumasks

[Steven Rostedt <rostedt@goodmis.org>]

On Tue, 6 May 2014 19:16:34 +0000 (UTC)
Mathieu Desnoyers <mathieu.desnoyers@efficios.com> wrote:

> ----- Original Message -----
> > From: "Steven Rostedt" <rostedt@goodmis.org>
> > To: "LKML" <linux-kernel@vger.kernel.org>
> > Cc: "Javi Merino" <javi.merino@arm.com>, "Andrew Morton" <akpm@linux-foundation.org>, "Mathieu Desnoyers"
> > <mathieu.desnoyers@efficios.com>, "Ingo Molnar" <mingo@kernel.org>, "Namhyung Kim" <namhyung@kernel.org>, "Jiri
> > Olsa" <jolsa@redhat.com>
> > Sent: Tuesday, May 6, 2014 1:22:38 PM
> > Subject: [RFC][PATCH] tracing: Add __cpumask() macro to trace events to record cpumasks
> > 
> > 
> > Being able to show a cpumask of events can be useful as some events
> > may affect only some CPUs. There is no standard way to record the
> > cpumask and converting it to a string is rather expensive during
> > the trace as traces happen in hotpaths. It would be better to record
> > the raw event mask and be able to parse it at print time.
> 
> Why name the type cpumask especially ? The type could be a "bitmask",
> and the specific field semantic in this case (field name) could
> be "cpus", thus implying that this is a cpu mask. This would allow
> using bitmasks for other things than cpu masks.

Hmm, sure, perhaps it would be better to call it bitmask instead. It
basically is already.

> 
> Moreover, I've been thinking about eventually adding a bitmask associated
> with an enumeration, so each bit of the bitmask could be associated with
> an entry from an enumeration (a string). So having a bitmask type would
> be a good step in that direction. For instance, the block layer
> instrumentation could use this for the rwbs field, instead of
> printing characters with blk_fill_rwbs(). This would save space and
> cpu time at trace collection.

Perhaps.

-- Steve