Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
MIME-Version: 1.0
References: <20220425134204.149042-1-sebastianene@google.com>
 <20220425134204.149042-2-sebastianene@google.com> <YmbonypWxzZJbjQ1@robh.at.kernel.org>
 <Ymqksfth+sj5JOWo@google.com>
In-Reply-To: <Ymqksfth+sj5JOWo@google.com>
From:   Rob Herring <robh@kernel.org>
Date:   Fri, 29 Apr 2022 15:46:05 -0500
Message-ID: <CAL_JsqKwR9Dq_zUhG7-KLJpHF=H2yLQBA8x09yfgSRscv2bhhg@mail.gmail.com>
Subject: Re: [PATCH v3 1/2] dt-bindings: vm-wdt: Add qemu,vm-watchdog compatible
To:     Sebastian Ene <sebastianene@google.com>
Cc:     "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Derek Kiernan <derek.kiernan@xilinx.com>,
        Dragan Cvetic <dragan.cvetic@xilinx.com>,
        Arnd Bergmann <arnd@arndb.de>, devicetree@vger.kernel.org,
        Quentin Perret <qperret@google.com>,
        Will Deacon <will@kernel.org>, Marc Zyngier <maz@kernel.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Apr 28, 2022 at 9:29 AM Sebastian Ene <sebastianene@google.com> wrote:
>
> On Mon, Apr 25, 2022 at 01:29:51PM -0500, Rob Herring wrote:
> > On Mon, Apr 25, 2022 at 01:42:05PM +0000, Sebastian Ene wrote:
> > > The stall detection mechanism allows to configure the expiration
> > > duration and the internal counter clock frequency measured in Hz.
> > > Add these properties in the schema.
> > >
> > > Signed-off-by: Sebastian Ene <sebastianene@google.com>
> > > ---
> > >  .../devicetree/bindings/misc/vm-wdt.yaml      | 44 +++++++++++++++++++
> > >  1 file changed, 44 insertions(+)
> > >  create mode 100644 Documentation/devicetree/bindings/misc/vm-wdt.yaml
> > >
> > > diff --git a/Documentation/devicetree/bindings/misc/vm-wdt.yaml b/Documentation/devicetree/bindings/misc/vm-wdt.yaml
> > > new file mode 100644
> > > index 000000000000..cb7665a0c5af
> > > --- /dev/null
> > > +++ b/Documentation/devicetree/bindings/misc/vm-wdt.yaml
> > > @@ -0,0 +1,44 @@
> > > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> > > +%YAML 1.2
> > > +---
> > > +$id: http://devicetree.org/schemas/misc/vm-wdt.yaml#
> > > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > > +
> > > +title: VM watchdog
> > > +
> > > +description: |
> > > +  This binding describes a CPU stall detector mechanism for virtual cpus.
> > > +
> > > +maintainers:
> > > +  - Sebastian Ene <sebastianene@google.com>
> > > +
> > > +properties:
> > > +  compatible:
> > > +    enum:
> > > +      - qemu,vm-watchdog
> > > +  clock:
>
> Hi,
>
> >
> > 'clocks' is already a defined property and 'clock' is too close. It's
> > also ambiguous what it is. 'clock-frequency' instead perhaps.
> >
>
> Yes, I think 'clock-frequency' is a better name. I will update it.

You are defining the register interface, so why not define a register
containing the frequency? Again, make your interface discoverable.

> > > +    $ref: /schemas/types.yaml#/definitions/uint32
> > > +    description: |
> > > +      The watchdog internal clock measure in Hz used to decrement the
> > > +      watchdog counter register on each tick.
> > > +      Defaults to 10 if unset.
> > > +  timeout-sec:
> > > +    $ref: /schemas/types.yaml#/definitions/uint32
> > > +    description: |
> > > +      The watchdog expiration timeout measured in seconds.
> > > +      Defaults to 8 if unset.
> > > +
> > > +required:
> > > +  - compatible
> > > +
> > > +additionalProperties: false
> > > +
> > > +examples:
> > > +  - |
> > > +    watchdog {
> > > +      compatible = "qemu,vm-watchdog";
> > > +      clock = <10>;
> > > +      timeout-sec = <8>;
> >
> > How does one access this 'hardware'?
> >
>
> This is a MMIO device.

Then how do you discover its address? You need 'reg'.

> > Why does this need to be in DT?
> >
> > We have DT because h/w designers are incapable of making h/w
> > discoverable. Why repeat that problem with s/w interfaces?
> >
>
> We need to have this one in the DT because in a secure VM we only load
> trusted DT components.

How does using DT make something trusted vs. any other mechanism the
hypervisor controls?

I would like to know from the virtualization folks (Xen/KVM) how they
would implement this feature? Certainly this could be reused, but then
we need an ACPI binding too for non-DT systems.

Why not use virtio?

Rob