Re: [PATCH v3 05/11] dt-bindings: i3c: Document core bindings

From: Boris Brezillon
Date: Mon Mar 26 2018 - 07:20:32 EST


Hi Geert,

On Mon, 26 Mar 2018 12:22:24 +0200
Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:

> Hi Boris,
>
> On Fri, Mar 23, 2018 at 12:00 PM, Boris Brezillon
> <boris.brezillon@xxxxxxxxxxx> wrote:
> > From: Boris Brezillon <boris.brezillon@xxxxxxxxxxxxxxxxxx>
> >
> > A new I3C subsystem has been added and a generic description has been
> > created to represent the I3C bus and the devices connected on it.
> >
> > Document this generic representation.
> >
> > Signed-off-by: Boris Brezillon <boris.brezillon@xxxxxxxxxxxxxxxxxx>
>
> Thanks for your patch!
>
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/i3c/i3c.txt
> > @@ -0,0 +1,140 @@
>
> > +I3C devices
> > +===========
> > +
> > +All I3C devices are supposed to support DAA (Dynamic Address Assignment), and
> > +are thus discoverable. So, by default, I3C devices do not have to be described
> > +in the device tree.
>
> But if they're described, they should have a compatible value, no?

What's the point of having a compatible here? I mean, I3C devices are
already attached to the relevant drivers using the Provisional ID, why
would we need a compatible if we don't parse it?

>
> > +This being said, one might want to attach extra resources to these devices,
> > +and those resources may have to be described in the device tree, which in turn
> > +means we have to describe I3C devices.
> > +
> > +Another use case for describing an I3C device in the device tree is when this
> > +I3C device has a static address and we want to assign it a specific dynamic
> > +address before the DAA takes place (so that other devices on the bus can't
> > +take this dynamic address).
> > +
> > +The I3C device should be names <device-type>@<static-address>,<i3c-pid>,
>
> named

Oops. Will fix the typo.

>
> So the i3c-pid in the unit address is represented as a 64-bit number, not as two
> comma-separated 32-bit numbers?

Right. I've changed my mind so many times that I ended up mixing the 2
representations I have considered.

Here are the choices we have:

1/ expose the raw ProvisionalID which is a 48-bit integer formed by the
concatenation of the vendor ID, part ID and instance ID:
ProvisionalID = VendorID << 33 | PartID << 16 | InstanceID << 12 |
ExtraInfo
The I3C dev node name should in this case be something like
<device-type>@<static-address>,<i3c-pid>

2/ split the fields we are really interested in in different cells:
2nd cell: vendorID
3rd cell: partID
4th cell: instanceID
In this case, the node name should be
<device-type>@<static-address>,<i3c-vendor>,<i3c-partid>,<i3c-instanceid>

Note that we can still change for the second representation if Rob is
okay.

>
> > +Example:
> > +
> > + i3c-master@d040000 {
> > + compatible = "cdns,i3c-master";
> > + clocks = <&coreclock>, <&i3csysclock>;
> > + clock-names = "pclk", "sysclk";
> > + interrupts = <3 0>;
> > + reg = <0x0d040000 0x1000>;
> > + #address-cells = <3>;
> > + #size-cells = <0>;
> > +
> > + status = "okay";
> > + i2c-scl-frequency = <100000>;
> > +
> > + /* I2C device. */
> > + nunchuk: nunchuk@52 {
>
> @52,8000001000000000?

Well, I2C devs is a special case, and I'm not sure we want to add the
extra LVR information + the IS_I2C_DEV bit in the node name.

>
> > + compatible = "nintendo,nunchuk";
> > + reg = <0x52 0x80000010 0x0>;

One more detail: people are unlikely to define reg using raw values: I
provide macros to do that in patch 6.

> > + };
> > +
> > + /* I3C device with a static address. */
> > + thermal_sensor: sensor@68,39200144004 {
>
> No compatible value?

No, as said above, it's not needed.

>
> > + reg = <0x68 0x392 0x144004>;
> > + assigned-address = <0xa>;
> > + };
> > +
> > + /*
> > + * I3C device without a static address but requiring resources
> > + * described in the DT.
> > + */
> > + sensor@0,39200154004 {
>
> No compatible value?

Ditto.

Thanks for your review.

Boris

--
Boris Brezillon, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com