Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Sat, 15 Oct 2022 15:53:19 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Joel Fernandes <joel@joelfernandes.org>
Cc:     Connor O'Brien <connoro@google.com>, linux-kernel@vger.kernel.org,
        kernel-team@android.com, John Stultz <jstultz@google.com>,
        Qais Yousef <qais.yousef@arm.com>,
        Ingo Molnar <mingo@redhat.com>,
        Juri Lelli <juri.lelli@redhat.com>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        Dietmar Eggemann <dietmar.eggemann@arm.com>,
        Steven Rostedt <rostedt@goodmis.org>,
        Ben Segall <bsegall@google.com>, Mel Gorman <mgorman@suse.de>,
        Daniel Bristot de Oliveira <bristot@redhat.com>,
        Valentin Schneider <vschneid@redhat.com>,
        Will Deacon <will@kernel.org>,
        Waiman Long <longman@redhat.com>,
        Boqun Feng <boqun.feng@gmail.com>,
        "Paul E . McKenney" <paulmck@kernel.org>, youssefesmat@google.com
Subject: Re: [RFC PATCH 07/11] sched: Add proxy execution
Message-ID: <Y0q7TyEWco/PfOAI@hirez.programming.kicks-ass.net>
References: <20221003214501.2050087-1-connoro@google.com>
 <20221003214501.2050087-8-connoro@google.com>
 <Y0YeUhDZWb49mpLo@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Y0YeUhDZWb49mpLo@google.com>
Precedence: bulk

On Wed, Oct 12, 2022 at 01:54:26AM +0000, Joel Fernandes wrote:

> Here, for CFS, doesn't this mean that the ->vruntime of the blocked task does
> not change while it is giving the owner a chance to run?
> 
> So that feels like if the blocked task is on the left-most node of the cfs
> rbtree, then it will always proxy to the owner while not letting other
> unrelated CFS tasks to run. That sounds like a breakage of fairness.
> 
> Or did I miss something?

So CFS will pick the leftmost task, if it happens to be the lock owner,
it will run as normal, owner time will advance.

When owner time is advanced enough to not be elegible anymore, it is
likely a blocked task is selected, at which time we'll advance it's
vruntime.

All this repeats until the whole block chain has advanced enough to make
the owner elegible again and the whole thing repeats.

From this it is easy to see that the critical secion executes with the
direct sum of the blockchain as a whole (all of them will have donated
their relative time to make the owner elegible again) -- provided the
critical section is of course long enough for this to matter (or it's
owner's weight small enough etc..).