Re: [PATCH] tracing: Export key trace event symbols

From: Christoph Hellwig <hch@infradead.org>
To: Ron Rechenmacher <ron@fnal.gov>
Cc: Christoph Hellwig <hch@infradead.org>,
	linux-kernel@vger.kernel.org,
	Steven Rostedt <rostedt@goodmis.org>
Subject: Re: [PATCH] tracing: Export key trace event symbols
Date: Tue, 21 Apr 2015 06:23:55 -0700	[thread overview]
Message-ID: <20150421132355.GA18161@infradead.org> (raw)
In-Reply-To: <55364CF4.2090600@fnal.gov>

On Tue, Apr 21, 2015 at 08:13:24AM -0500, Ron Rechenmacher wrote:
> Hi Christoph,
> 
> My module, and information about it, can be found at:
> https://cdcvs.fnal.gov/redmine/projects/trace
> https://cdcvs.fnal.gov/redmine/projects/trace/repository/show/src_module
> with the header at
> https://cdcvs.fnal.gov/redmine/projects/trace/repository/raw/include/trace.h

And now send a patch including a rationale on why we need it.  I haven't
looked deep enough, but so far I can't see any real value add in it.

> 
> It seems that you are the person that I have to convince in order to get
> my patch accepted. Is this true?

In the end you'll need to convince a maintainer to apply it.

> Can you tell me how:
> ./kernel/trace/power-traces.c:17:EXPORT_TRACEPOINT_SYMBOL_GPL(power_start);

I can't find a tracepoint with that name in the current tree.

> ./kernel/trace/power-traces.c:19:EXPORT_TRACEPOINT_SYMBOL_GPL(cpu_idle);

This one is used by the potentially modular intel_idle (now cpuids)
driver, so it's exported for a reason.

> should be allowed (how it fits with your definition of the Linux kernel below) and
> my proposed EXPORTS not?
> I hope that you will not suggest removing the above 2 exports.

It's perfectly fine to remove unused exports, and we do it regularly