linux-kernel - Re: [PATCH 1/2] dt-bindings: pfuze100: add optional pfuze-disable-sw binding

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: <20180713083039.4z3ulwtg2ekx5n23@pengutronix.de>
Date:   Fri, 13 Jul 2018 10:30:39 +0200
From:   Marco Felsch <m.felsch@...gutronix.de>
To:     Mark Brown <broonie@...nel.org>
Cc:     lgirdwood@...il.com, robh+dt@...nel.org, mark.rutland@....com,
        fabio.estevam@....com, Anson.Huang@....com, kernel@...gutronix.de,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] dt-bindings: pfuze100: add optional pfuze-disable-sw
 binding

Hi Mark,

On 18-07-12 16:31, Mark Brown wrote:
> On Thu, Jul 12, 2018 at 01:02:39PM +0200, Marco Felsch wrote:
> 
> > +Optional properties:
> > +- pfuze-disable-sw: Disable all unused switch regulators to save power
> > +  consumption. Attention, some platforms are using the switch regulators as DDR
> > +  ref or supply voltage. Mark these regulators as "regulator-always-on" to skip
> > +  disabling these regulators. If not specified, the driver simualtes the
> > +  disabling. This means the state of the regulator is set to 'disabled' but the
> > +  driver don't disable the regulator.
> 
> This is a bit of a confused way of specifying things that depends on the
> Linux implementation, and the property sounds like a double negative
> too.  I'd say something like "pfuze-support-disable" and then explicitly
> say that this is a workaround for backwards compatibility.

I can't find the double negative. Anyway your binding sounds better. So
I will use yours. Should we add a vendor prefix too to be clear? I will
also add some more informations to mark it as workaround.

> I'd also recommend changing the implementation patch to just register a
> different version of the desc and ops that just doesn't have the disable
> operation so that the framework knows what's going on.  While the
> current implementation works now there's the possibility that at some
> point in the future we might start relying on the disable actually
> having taken effect somehow and will get confused.  There's some
> existing drivers that optimize their resume paths if they know power
> wasn't removed.

Okay I will change that too. I didn't know that there are drivers with
optimized resume paths.

Thanks for your feedback.

Regards,
Marco