Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Fri, 9 Oct 2020 11:12:18 +0200
From:   Michal Hocko <mhocko@suse.com>
To:     Peter Zijlstra <peterz@infradead.org>,
        Thomas Gleixner <tglx@linutronix.de>
Cc:     Frederic Weisbecker <fweisbecker@suse.de>,
        LKML <linux-kernel@vger.kernel.org>,
        Mel Gorman <mgorman@suse.de>, Ingo Molnar <mingo@redhat.com>
Subject: Re: [RFC PATCH] kernel: allow to configure PREEMPT_NONE,
 PREEMPT_VOLUNTARY on kernel command line
Message-ID: <20201009091218.GF4967@dhcp22.suse.cz>
References: <20201007120401.11200-1-mhocko@kernel.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20201007120401.11200-1-mhocko@kernel.org>
Precedence: bulk

On Wed 07-10-20 14:04:01, Michal Hocko wrote:
> From: Michal Hocko <mhocko@suse.com>
> 
> Many people are still relying on pre built distribution kernels and so
> distributions have to provide mutliple kernel flavors to offer different
> preemption models. Most of them are providing PREEMPT_NONE for typical
> server deployments and PREEMPT_VOLUNTARY for desktop users.
> 
> Having two different kernel binaries differing only by the preemption
> mode seems rather wasteful and inflexible. Especially when the difference
> between PREEMPT_NONE and PREEMPT_VOLUNTARY is really minimal. Both only
> allow explicit scheduling points while running in the kernel and it is
> only might_sleep which adds additional preemption points for
> PREEMPT_VOLUNTARY.
> 
> Add a kernel command line parameter preempt_mode=[none, voluntary] which
> allows to override the default compile time preemption mode
> (CONFIG_PREEMPT_NONE resp. CONFIG_PREEMPT_VOLUTARY). Voluntary preempt
> mode is guarded by a jump label to prevent any potential runtime overhead.
> 
> Signed-off-by: Michal Hocko <mhocko@suse.com>
> ---
> 
> Hi all,
> this is sent as an RFC because it still needs some work. E.g. move jump
> label changes into their own patch. They are needed mostly to get rid of
> header files dependencies (seen for CONFIG_JUMP_LABEL=n via atomic.h ->
> bug.h).
> 
> I wanted to make sure that the idea is sound for maintainers first. The
> next step would be extending the command line to support full preemption
> as well but there is much more work in that area. Frederic has promissed
> to look into that.
> 
> Thoughts?

Is there any additional feedback? Should I split up the patch and repost
for inclusion?
-- 
Michal Hocko
SUSE Labs