From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1757158AbcC2OEy (ORCPT <rfc822;w@1wt.eu>);
	Tue, 29 Mar 2016 10:04:54 -0400
Received: from bombadil.infradead.org ([198.137.202.9]:33189 "EHLO
	bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1756814AbcC2OEw (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Tue, 29 Mar 2016 10:04:52 -0400
Date: Tue, 29 Mar 2016 16:04:39 +0200
From: Peter Zijlstra <peterz@infradead.org>
To: Wang Nan <wangnan0@huawei.com>
Cc: Alexei Starovoitov <ast@kernel.org>,
        Arnaldo Carvalho de Melo <acme@redhat.com>,
        linux-kernel@vger.kernel.org,
        Brendan Gregg <brendan.d.gregg@gmail.com>,
        He Kuang <hekuang@huawei.com>, Jiri Olsa <jolsa@kernel.org>,
        Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>,
        Namhyung Kim <namhyung@kernel.org>, pi3orama@163.com,
        Zefan Li <lizefan@huawei.com>
Subject: Re: [PATCH 4/4] perf core: Add backward attribute to perf event
Message-ID: <20160329140439.GK3408@twins.programming.kicks-ass.net>
References: <1459147292-239310-1-git-send-email-wangnan0@huawei.com>
 <1459147292-239310-5-git-send-email-wangnan0@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1459147292-239310-5-git-send-email-wangnan0@huawei.com>
User-Agent: Mutt/1.5.21 (2012-12-30)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Mon, Mar 28, 2016 at 06:41:32AM +0000, Wang Nan wrote:

Could you maybe write a perf/tests thingy for this so that _some_
userspace exists that exercises this new code?


>  int perf_output_begin(struct perf_output_handle *handle,
>  		      struct perf_event *event, unsigned int size)
>  {
> +	if (unlikely(is_write_backward(event)))
> +		return __perf_output_begin(handle, event, size, true);
>  	return __perf_output_begin(handle, event, size, false);
>  }

Would something like:

int perf_output_begin(...)
{
	if (unlikely(is_write_backward(event))
		return perf_output_begin_backward(...);
	return perf_output_begin_forward(...);
}

make sense; I'm not sure how much is still using this, but it seems
somewhat excessive to inline two copies of that thing into a single
function.

Alternatively; something like:

int perf_output_begin(...)
{
	return __perf_output_begin(..., unlikely(event->attr.backwards));
}

might make sense too.