lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180625185037.GE129942@google.com>
Date:   Mon, 25 Jun 2018 11:50:37 -0700
From:   Matthias Kaehlcke <mka@...omium.org>
To:     Rob Herring <robh@...nel.org>
Cc:     MyungJoo Ham <myungjoo.ham@...sung.com>,
        Kyungmin Park <kyungmin.park@...sung.com>,
        Chanwoo Choi <cw00.choi@...sung.com>,
        Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Mark Rutland <mark.rutland@....com>, linux-pm@...r.kernel.org,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        Brian Norris <briannorris@...omium.org>,
        Douglas Anderson <dianders@...omium.org>,
        Enric Balletbo i Serra <enric.balletbo@...labora.com>,
        "Rafael J . Wysocki" <rjw@...ysocki.net>,
        Viresh Kumar <viresh.kumar@...aro.org>,
        Lee Jones <lee.jones@...aro.org>,
        Benson Leung <bleung@...omium.org>,
        Olof Johansson <olof@...om.net>
Subject: Re: [PATCH v4 09/12] dt-bindings: PM / OPP: add opp-throttlers
 property

Hi Rob,

On Mon, Jun 25, 2018 at 09:33:41AM -0600, Rob Herring wrote:
> On Wed, Jun 20, 2018 at 06:52:34PM -0700, Matthias Kaehlcke wrote:
> > The optional opp-throttlers property is used to configure the throttlers
> > (see drivers/misc/throttler/*) that use a given OPP to throttle the
> > corresponding device(s).
> > 
> > Signed-off-by: Matthias Kaehlcke <mka@...omium.org>
> > Reviewed-by: Brian Norris <briannorris@...omium.org>
> > ---
> > Changes in v4:
> > - added 'Reviewed-by: Brian Norris <briannorris@...omium.org>' tag
> > 
> > Changes in v3:
> > - none
> > 
> > Changes in v2:
> > - patch added to series
> > ---
> >  Documentation/devicetree/bindings/opp/opp.txt | 3 +++
> >  1 file changed, 3 insertions(+)
> > 
> > diff --git a/Documentation/devicetree/bindings/opp/opp.txt b/Documentation/devicetree/bindings/opp/opp.txt
> > index c396c4c0af92..747e79740c75 100644
> > --- a/Documentation/devicetree/bindings/opp/opp.txt
> > +++ b/Documentation/devicetree/bindings/opp/opp.txt
> > @@ -170,6 +170,9 @@ Optional properties:
> >    functioning of the current device at the current OPP (where this property is
> >    present).
> >  
> > +- opp-throttlers: Array with phandles of throttlers that use this OPP to
> > +  throttle the corresponding device(s).
> > +
> 
> I think it would be better to make this a boolean for each OPP entry and 
> then add "operating-points-v2" property to the EC node to point to the 
> OPP table.

Thanks for your suggestion. "operating-points-v2" would have to be an
array of phandles, a single thottler can have multiple throttling
devices.

> Unless there is some need for a different throttler for each OPP entry?

Having the option to use different OPPs per throttler would be my
preference. E.g. there could be configurations where one throttler
only interacts with certain throttling devices and another one
with others.

I see another option to achieve this, if you don't like the reference
to the throttlers in the OPPs. The throttler could have a list of OPPs
(as phandles, not frequencies as in v1). The main inconvenient I see
here is that the used OPPs would need a label, which they usually
don't have. Maybe this is no soooo bad, since the label could be added
at device level, only on devices that use a throttler, so it wouldn't
clutter the platform .dts files.

This could be a single array with all OPPs from different devices,
or multiple arrays, one for each throttling device:

throttler-opps-0 = <&cpu0_opp_03 &cpu0_opp_05>;
throttler-opps-1 = <&gpu_opp_02 &gpu_opp_04>;

My preference would be multiple arrays, because it's easier to read.

What do you think?

Thanks

Matthias

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ