linux-kernel - Re: [PATCH 1/2] regulator: Add support for power budget

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: <8f5ce662-f8bc-449e-99cf-737066dabebd@sirena.org.uk>
Date: Mon, 13 Jan 2025 15:44:29 +0000
From: Mark Brown <broonie@...nel.org>
To: Kory Maincent <kory.maincent@...tlin.com>
Cc: Liam Girdwood <lgirdwood@...il.com>, Rob Herring <robh@...nel.org>,
	Krzysztof Kozlowski <krzk+dt@...nel.org>,
	Conor Dooley <conor+dt@...nel.org>,
	Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
	linux-kernel@...r.kernel.org, devicetree@...r.kernel.org,
	Oleksij Rempel <o.rempel@...gutronix.de>, kernel@...gutronix.de
Subject: Re: [PATCH 1/2] regulator: Add support for power budget

On Mon, Jan 13, 2025 at 04:26:20PM +0100, Kory Maincent wrote:
> Mark Brown <broonie@...nel.org> wrote:
> > On Mon, Jan 13, 2025 at 03:45:51PM +0100, Kory Maincent wrote:
> > > Mark Brown <broonie@...nel.org> wrote:  

> > but the others are legit.  Still, we should be able to map between the
> > two.

> We could have something like that in regulator_request_power_budget()?

> if (rdev->desc->ops->get_voltage && rdev->desc->ops->set_current_limit) {
> 	ret = regulator_get_voltage(rdev);
> 	if (ret < 0)
> 		return ret;
> 
> 	tmp_64 = pw_req;
> 	tmp_64 *= 1000000000ull;
> 	/* uA = mW * 1000000000 / uV */
> 	uA = DIV_ROUND_CLOSEST_ULL(tmp_64, uV);
> 	ret = regulator_set_current_limit(rdev, uA);
> 	if (ret)
> 		return ret;

Yup, indeed.  That said I am wondering if it's safer to just configure
the constraint in the hardware rather than the currently requested
limit, considering what might happen in the case where there's multiple
consumers that have only been partially updated.  If the hardware limits
or shuts down rather than warning it'll blow up badly so it might be
better to be conservative.  Unfortunately we don't distinguish in the
ops.  Possibly it should be a policy thing even but then that's better
at runtime...

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)