linux-kernel - Re: [PATCH 0/6] regulator: core: Add support for external outputs

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]

Date:   Wed, 4 May 2022 13:28:11 -0700
From:   Zev Weiss <zev@...ilderbeest.net>
To:     Mark Brown <broonie@...nel.org>
Cc:     Liam Girdwood <lgirdwood@...il.com>, linux-kernel@...r.kernel.org,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        devicetree@...r.kernel.org, openbmc@...ts.ozlabs.org
Subject: Re: [PATCH 0/6] regulator: core: Add support for external outputs

On Wed, May 04, 2022 at 05:56:13AM PDT, Mark Brown wrote:
>On Tue, May 03, 2022 at 11:50:35PM -0700, Zev Weiss wrote:
>
>> The DT bindings changes (patches 1 and 2) consist of a boolean
>> regulator property to mark it as supplying an external output, and a
>> reg-external-output binding to act as a downstream device representing
>> that output.  The redundancy between the two maybe isn't entirely
>> ideal, but it was the cleanest approach I've been able to come up with
>> so far in terms of working with the regulator subsystem; I'm certainly
>> open to suggestions for better ways of going about this.
>
>Nothing in the series articulates what the purpose of the redundancy is
>- your description of this is a consumer, why would the regulator itself
>care what's connected to it?

Hi Mark, thanks for the review.

To some extent that was an additional (maybe excessive) protective 
measure against regulators ending up userspace-controllable when they 
shouldn't be, since I had previously gotten the impression that there 
was some concern about that possibility.

More functionally though, it was also basically a hack to allow the 
'state' sysfs attribute's mode to get set properly in 
regulator_register(), before the consumer is known.  Though if things 
are rearranged based on what you said in another message about putting 
the enable/disable control in a consumer driver instead of the regulator 
itself, it should be easy to get rid of.

Thanks,
Zev