[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250324103000.479f83dd@gandalf.local.home>
Date: Mon, 24 Mar 2025 10:30:00 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: "Masami Hiramatsu (Google)" <mhiramat@...nel.org>
Cc: Douglas Raillard <douglas.raillard@....com>, Mathieu Desnoyers
<mathieu.desnoyers@...icios.com>, linux-kernel@...r.kernel.org,
linux-trace-kernel@...r.kernel.org
Subject: Re: [PATCH 3/3] tracing: Rename trace_synth() to
synth_event_trace2()
On Mon, 24 Mar 2025 15:29:45 +0900
Masami Hiramatsu (Google) <mhiramat@...nel.org> wrote:
> Yeah, because those are mainly for the tests, and we are expecting that if
> any modules wants to emit its events, it will define new trace-events and
> use it instead of synthetic events. The synthetic events are for
> programming via tracefs, not reporting from the kernel modules.
> It is confusing if any synthetic events are reported without any origin of
> real trace event. (so, it is an intermadiate event type.) IOW, We expect
> that synthetic event is reported by other events via event trigger.
> The current APIs are just for testing.
>
> Hmm, I should hide those by CONFIG_SYNTH_EVENT_TESTS.
Perhaps we should remove synth_event_trace() from the include/linux header
file, and move it to kernel/trace/trace.h? That way it is only used for
internal purposes, and not exposed to modules. You could wrap the export
with a #ifdef CONFIG_SYNTH_EVENT_TESTS.
-- Steve
Powered by blists - more mailing lists