linux-kernel - Re: [PATCHv2 1/3] software node: Power management operations for software nodes

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:   Fri, 30 Oct 2020 12:27:02 +0200
From:   Heikki Krogerus <heikki.krogerus@...ux.intel.com>
To:     "Rafael J. Wysocki" <rafael@...nel.org>
Cc:     "Rafael J. Wysocki" <rjw@...ysocki.net>,
        Felipe Balbi <balbi@...nel.org>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Sakari Ailus <sakari.ailus@...ux.intel.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        "open list:ULTRA-WIDEBAND (UWB) SUBSYSTEM:" 
        <linux-usb@...r.kernel.org>,
        ACPI Devel Maling List <linux-acpi@...r.kernel.org>
Subject: Re: [PATCHv2 1/3] software node: Power management operations for
 software nodes

Hi Rafael,

On Thu, Oct 29, 2020 at 06:10:59PM +0100, Rafael J. Wysocki wrote:
> Well, this basically implements a wrapper PM domain that is somewhat
> more generic, as a concept, then software nodes PM.
> 
> At least it is not specific to software nodes, so I'd prefer it to be
> defined generically.

I don't think we should generalize it like that. I do not think the
power domains should have any links between each other at the general
level (just like we probable should not link fwnodes together anymore
like we do now with the "secondary" fwnode). That is why I have
confined this to software nodes only for now.

I think ideally devices could belong to multiple power domains. That
would be the general solution. I did not think that trying to figure
out how to do that would be reasonable as the first approach (maybe I
should have done exactly that?). But would it be acceptable to allow
devices to belong to multiple power domains?

> Moreover, IIUC, this breaks if the "primary" PM domain callbacks try
> to get to the original PM domain via the dev->pm_domain pointer, which
> the genpd callbacks do.

Ouch, that is true.

> Do we want to wrap the ACPI PM domain only, by any chance?  If so, it
> may be more straightforward to invoke swnode callbacks directly from
> there, if any.

The software node can still be the only "primary" fwnode. I don't
think we should limit this to only platforms (and kernels) that
support ACPI.

thanks,

-- 
heikki