Date: Wed, 2 Dec 2015 18:15:21 +0100
From: Peter Zijlstra <peterz@infradead.org>
To: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Chris Metcalf <cmetcalf@ezchip.com>, LKML <linux-kernel@vger.kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Luiz Capitulino <lcapitulino@redhat.com>,
        Christoph Lameter <cl@linux.com>, Ingo Molnar <mingo@kernel.org>,
        Viresh Kumar <viresh.kumar@linaro.org>, Rik van Riel <riel@redhat.com>
Subject: Re: [PATCH 3/7] perf: Migrate perf to use new tick dependency mask
 model
Message-ID: <20151202171521.GU3816@twins.programming.kicks-ass.net>
References: <1447424529-13671-1-git-send-email-fweisbec@gmail.com>
 <1447424529-13671-4-git-send-email-fweisbec@gmail.com>
 <56548E15.5050004@ezchip.com>
 <20151125123428.GD16609@lerouge>
 <20151202161758.GS3816@twins.programming.kicks-ass.net>
 <20151202170313.GB16408@lerouge>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20151202170313.GB16408@lerouge>
User-Agent: Mutt/1.5.21 (2012-12-30)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 792
Lines: 14

On Wed, Dec 02, 2015 at 06:03:14PM +0100, Frederic Weisbecker wrote:
> Right, I can do either that or I can move the dependency to the CPU level
> and count nr_freq to the cpu_ctx when any ctx gets scheduled in/out. Then
> everytime we inc and nr_freq == 1, we set the dependency (all that should
> be serialized as it only happens locally).

Doing it per CPU would, as you say, add accounting crap to the context
switch path. Now given that context switches with perf enabled are
already silly expensive that might just fall away into the noise.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/