[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20171115081656.at4cllb53pprfbtz@gmail.com>
Date: Wed, 15 Nov 2017 09:16:56 +0100
From: Ingo Molnar <mingo@...nel.org>
To: "quan.xu04@...il.com" <quan.xu04@...il.com>
Cc: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
kvm <kvm@...r.kernel.org>, Quan Xu <quan.xu0@...il.com>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>
Subject: Re: [PATCH RFC v3 4/6] Documentation: Add three sysctls for smart
idle poll
* quan.xu04@...il.com <quan.xu04@...il.com> wrote:
> > > could I only leave paravirt_poll_threshold_ns parameter (the maximum poll time),
> > > which is as similar as "adaptive halt-polling" Wanpeng mentioned.. then user can
> > > turn it off, or find an appropriate threshold for some odd scenario..
> > That way lies utter madness. Maybe add it as a debugfs knob, but exposing it to
> > userspace: NAK.
> >
> .. so, I will make these 3 parameters by default in next v4.
> paravirt_poll_threshold_ns = 200000
> paravirt_poll_shrink = 2
> paravirt_poll_grow = 2
>
> neither touch the syscal ABI nor expose it to userspace again.
Note that you can still expose it to developers via debugfs knobs.
I.e. same development and testing flexibility, but no ABI promise and an implicit
expectation for this mechanism to be self-tuning.
Thanks,
Ingo
Powered by blists - more mailing lists