by Randy Dunlap

[permalink] [raw]

Subject: Re: [PATCH V0 4/7] usb: common: eud: Added the driver support for Embedded USB Debugger(EUD)

On 10/4/21 4:16 AM, Souradeep Chowdhury wrote:
> Add support for control peripheral of EUD (Embedded USB Debugger) to
> listen to events such as USB attach/detach, pet EUD to indicate software

I don't quite understand: what pets the EUD? how does it do that?

> is functional.Reusing the platform device kobj, sysfs entry 'enable' is

functional. Reusing

> created to enable or disable EUD.
>
> To enable the eud the following needs to be done
> echo 1 >/sys/bus/platform/.../enable
>
> To disable eud, following is the command
> echo 0 >/sys/bus/platform/.../enable
>
> Signed-off-by: Souradeep Chowdhury<[email protected]>
> ---
> Documentation/ABI/testing/sysfs-driver-eud | 7 +
> drivers/usb/common/Kconfig | 9 +
> drivers/usb/common/Makefile | 1 +
> drivers/usb/common/qcom_eud.c | 256 +++++++++++++++++++++++++++++
> 4 files changed, 273 insertions(+)
> create mode 100644 Documentation/ABI/testing/sysfs-driver-eud
> create mode 100644 drivers/usb/common/qcom_eud.c
>
> diff --git a/Documentation/ABI/testing/sysfs-driver-eud b/Documentation/ABI/testing/sysfs-driver-eud
> new file mode 100644
> index 0000000..14a02da
> --- /dev/null
> +++ b/Documentation/ABI/testing/sysfs-driver-eud
> @@ -0,0 +1,7 @@
> +What: /sys/bus/platform/.../enable
> +Date: October 2021
> +Contact: Souradeep Chowdhury<[email protected]>
> +Description:
> + The Enable/Disable sysfs interface for Embedded
> + USB Debugger(EUD).This enables and disables the

Debugger (EUD). This enables

> + EUD based on a 1 or a 0 value.

--
~Randy

2021-10-05 04:44:45

On Thu 07 Oct 04:25 CDT 2021, [email protected] wrote:

> On 2021-10-05 22:07, Bjorn Andersson wrote:
> > On Tue 05 Oct 06:11 PDT 2021, [email protected] wrote:
> >
> > > On 2021-10-04 22:07, Rob Herring wrote:
> > > > On Mon, Oct 04, 2021 at 04:46:19PM +0530, Souradeep Chowdhury wrote:
> > > > > Added the property for EUD(Embedded USB Debug) connector.Added
> > > > > the "reg" and "interrupts" property which is needed for EUD.
> > > >
> > > > You are going to need a better explanation of this h/w.
> > >
> > > Ack. Will update this with the detailed hardware description
> > > in the next version.
> > >
> > > >
> > > > >
> > > > > Signed-off-by: Souradeep Chowdhury <[email protected]>
> > > > > ---
> > > > > .../devicetree/bindings/connector/usb-connector.yaml | 15
> > > > > +++++++++++++++
> > > > > 1 file changed, 15 insertions(+)
> > > > >
> > > > > diff --git
> > > > > a/Documentation/devicetree/bindings/connector/usb-connector.yaml
> > > > > b/Documentation/devicetree/bindings/connector/usb-connector.yaml
> > > > > index 7eb8659..908129f 100644
> > > > > --- a/Documentation/devicetree/bindings/connector/usb-connector.yaml
> > > > > +++ b/Documentation/devicetree/bindings/connector/usb-connector.yaml
> > > > > @@ -30,6 +30,21 @@ properties:
> > > > > - const: samsung,usb-connector-11pin
> > > > > - const: usb-b-connector
> > > > >
> > > > > + - items:
> > > > > + - enum:
> > > > > + - qcom,sc7280-usb-connector-eud
> > > > > + - const: qcom,usb-connector-eud
> > > > > + - const: usb-c-connector
> > > > > +
> > > > > + reg:
> > > > > + items:
> > > > > + - description: EUD Base Register Region
> > > > > + - description: EUD Mode Manager Region
> > > >
> > > > A connector node represents the physical connector on a board. That
> > > > can't really be an MMIO peripheral. Maybe you need a node for EUD and
> > > > then it should have a connector child node? Don't really know without
> > > > understanding this h/w.
> > >
> > > As per the previous discussion on the EUD, it was agreed upon to map
> > > EUD
> > > as a type C connector and use Role-Switch to change the USB role
> > > instead
> > > of extcon interface that was being used previously. The link for the
> > > same
> > > is as follows:-
> > >
> > > https://lore.kernel.org/lkml/[email protected]/
> > >
> >
> > Not using extcon is the right thing, but perhaps we should make the EUD
> > a role_switch provider and client, so that we can describe how it sits
> > inbetween the connector and the controller.
> >
> > That way it has the power to pass through or override requests from the
> > upstream role-switcher, based on the status of EUD.
> >
> >
> > That said, I'm still curious to what happens if I renegotiate the roles
> > dynamically in a Type-C environment, while enabling EUD. How would the
> > device on the other end of the cable know that it's supposed to be a
> > host? Or there's simply a reset of the link when this happens?
> >
> > Thanks,
> > Bjorn
>
> Hi Bjorn,
>

Hi Souradeep

> By making EUD Role-Switch provider and client do you mean that
> we should have a EUD node which will have a connector node as
> child and this connector node will have a port that points towards
> the drd role-switch?
>
> So that my device tree node will look like the following in that case
>
> eud@88e0000 {
> compatible = "qcom,usb-connector-eud";
> reg = <0 0x88e0000 0 0x2000>,
> <0 0x88e2000 0 0x1000>;
> interrupt-parent = <&pdc>;
> interrupts = <11 IRQ_TYPE_LEVEL_HIGH>;
> usb_con: connector {
> compatible = "usb-c-connector";
> label = "USB-C";
> port {
> eud_usb_output: endpoint {
> remote-endpoint = <&eud_usb3_input>;
> };
> };
>
> };
>
>
> @usb2 {
> dwc3 {
> usb-role-switch;
> port {
> eud_usb3_input: endpoint {
> remote-endpoint = <&eud_usb_output>;
> };
> };
> };

While your "output" and "input" matches the direction of the role
switching, I think the connection should be describe in the other
direction.

Also my suggestion was that EUD is both connector for the dwc3 and has a
reference to the connector described in the TypeC controller - to
properly describe the relationship:

DWC -> EUD -> connector

With the role switching request going from the connector (pmic_glink
driver) to DWC through the EUD, which can override the vote.

That said, this is just my suggestion. You need to ensure that Rob
understands the hardware design well enough to approve your proposed
binding.

E.g. The connector in the EUD isn't a usb-c-connector, it's some
type of internal connection, the next step in that chain is the actual
usb-c-connector.

Regards,
Bjorn

>
> Also EUD functions only in device mode, so when the role-switch is done by
> the controller
> to set the device mode, the PC on the other end becomes the host.
>
> Thanks,
> Souradeep
>
> >
> > > >
> > > > > +
> > > > > + interrupts:
> > > > > + description:
> > > > > + EUD interrupt
> > > > > +
> > > > > label:
> > > > > description: Symbolic name for the connector.
> > > > >
> > > > > --
> > > > > QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a
> > > > > member
> > > > > of Code Aurora Forum, hosted by The Linux Foundation
> > > > >
> > > > >

2021-11-02 09:23:07

by Souradeep Chowdhury

[permalink] [raw]

Subject: Re: [PATCH V0 1/7] dt-bindings: connector: Add property for eud type c connector

On 2021-10-12 08:52, Bjorn Andersson wrote:
> On Thu 07 Oct 04:25 CDT 2021, [email protected] wrote:
>
>> On 2021-10-05 22:07, Bjorn Andersson wrote:
>> > On Tue 05 Oct 06:11 PDT 2021, [email protected] wrote:
>> >
>> > > On 2021-10-04 22:07, Rob Herring wrote:
>> > > > On Mon, Oct 04, 2021 at 04:46:19PM +0530, Souradeep Chowdhury wrote:
>> > > > > Added the property for EUD(Embedded USB Debug) connector.Added
>> > > > > the "reg" and "interrupts" property which is needed for EUD.
>> > > >
>> > > > You are going to need a better explanation of this h/w.
>> > >
>> > > Ack. Will update this with the detailed hardware description
>> > > in the next version.
>> > >
>> > > >
>> > > > >
>> > > > > Signed-off-by: Souradeep Chowdhury <[email protected]>
>> > > > > ---
>> > > > > .../devicetree/bindings/connector/usb-connector.yaml | 15
>> > > > > +++++++++++++++
>> > > > > 1 file changed, 15 insertions(+)
>> > > > >
>> > > > > diff --git
>> > > > > a/Documentation/devicetree/bindings/connector/usb-connector.yaml
>> > > > > b/Documentation/devicetree/bindings/connector/usb-connector.yaml
>> > > > > index 7eb8659..908129f 100644
>> > > > > --- a/Documentation/devicetree/bindings/connector/usb-connector.yaml
>> > > > > +++ b/Documentation/devicetree/bindings/connector/usb-connector.yaml
>> > > > > @@ -30,6 +30,21 @@ properties:
>> > > > > - const: samsung,usb-connector-11pin
>> > > > > - const: usb-b-connector
>> > > > >
>> > > > > + - items:
>> > > > > + - enum:
>> > > > > + - qcom,sc7280-usb-connector-eud
>> > > > > + - const: qcom,usb-connector-eud
>> > > > > + - const: usb-c-connector
>> > > > > +
>> > > > > + reg:
>> > > > > + items:
>> > > > > + - description: EUD Base Register Region
>> > > > > + - description: EUD Mode Manager Region
>> > > >
>> > > > A connector node represents the physical connector on a board. That
>> > > > can't really be an MMIO peripheral. Maybe you need a node for EUD and
>> > > > then it should have a connector child node? Don't really know without
>> > > > understanding this h/w.
>> > >
>> > > As per the previous discussion on the EUD, it was agreed upon to map
>> > > EUD
>> > > as a type C connector and use Role-Switch to change the USB role
>> > > instead
>> > > of extcon interface that was being used previously. The link for the
>> > > same
>> > > is as follows:-
>> > >
>> > > https://lore.kernel.org/lkml/[email protected]/
>> > >
>> >
>> > Not using extcon is the right thing, but perhaps we should make the EUD
>> > a role_switch provider and client, so that we can describe how it sits
>> > inbetween the connector and the controller.
>> >
>> > That way it has the power to pass through or override requests from the
>> > upstream role-switcher, based on the status of EUD.
>> >
>> >
>> > That said, I'm still curious to what happens if I renegotiate the roles
>> > dynamically in a Type-C environment, while enabling EUD. How would the
>> > device on the other end of the cable know that it's supposed to be a
>> > host? Or there's simply a reset of the link when this happens?
>> >
>> > Thanks,
>> > Bjorn
>>
>> Hi Bjorn,
>>
>
> Hi Souradeep
>
>> By making EUD Role-Switch provider and client do you mean that
>> we should have a EUD node which will have a connector node as
>> child and this connector node will have a port that points towards
>> the drd role-switch?
>>
>> So that my device tree node will look like the following in that case
>>
>> eud@88e0000 {
>> compatible = "qcom,usb-connector-eud";
>> reg = <0 0x88e0000 0 0x2000>,
>> <0 0x88e2000 0 0x1000>;
>> interrupt-parent = <&pdc>;
>> interrupts = <11 IRQ_TYPE_LEVEL_HIGH>;
>> usb_con: connector {
>> compatible = "usb-c-connector";
>> label = "USB-C";
>> port {
>> eud_usb_output: endpoint {
>> remote-endpoint = <&eud_usb3_input>;
>> };
>> };
>>
>> };
>>
>>
>> @usb2 {
>> dwc3 {
>> usb-role-switch;
>> port {
>> eud_usb3_input: endpoint {
>> remote-endpoint = <&eud_usb_output>;
>> };
>> };
>> };
>
> While your "output" and "input" matches the direction of the role
> switching, I think the connection should be describe in the other
> direction.
>
> Also my suggestion was that EUD is both connector for the dwc3 and has
> a
> reference to the connector described in the TypeC controller - to
> properly describe the relationship:
>
> DWC -> EUD -> connector
>
> With the role switching request going from the connector (pmic_glink
> driver) to DWC through the EUD, which can override the vote.
>
>
> That said, this is just my suggestion. You need to ensure that Rob
> understands the hardware design well enough to approve your proposed
> binding.
>
>
> E.g. The connector in the EUD isn't a usb-c-connector, it's some
> type of internal connection, the next step in that chain is the actual
> usb-c-connector.
>
> Regards,
> Bjorn
>
>>
>> Also EUD functions only in device mode, so when the role-switch is
>> done by
>> the controller
>> to set the device mode, the PC on the other end becomes the host.
>>
>> Thanks,
>> Souradeep

Hi Bjorn/Robb,

Following is how my EUD design looks like

My device tree node will look like as follows

eud@88e0000 {
compatible = "qcom,usb-connector-eud";
reg = <0 0x88e0000 0 0x2000>,
<0 0x88e2000 0 0x1000>;
interrupt-parent = <&pdc>;
interrupts = <11 IRQ_TYPE_LEVEL_HIGH>;
ports {
#address-cells = <2>;
#size-cells = <2>;
port@0 {
reg = <0>;
usb2_eud: endpoint {
remote-endpoint = <&eud_ep>;
};
};
};
};

@usb2 {
dwc3 {
usb-role-switch;
usb_con: eud_usb_connector {
compatible = “usb-connector-eud", "usb-c-connector";
ports {
#address-cells = <2>;
#size-cells = <2>;
port@0 {
reg = <0>;
eud_ep: endpoint {
remote-endpoint = <&usb2_eud>;
};
};
};
};

So, I am putting a connector node as a child of dwc3 node which can be
connected to the EUD node via ports. I am keeping EUD as a separate
device tree node.

I can get the device equivalent to the connector by doing the following
in EUD driver

dev = usb_of_get_companion_dev(&pdev->dev);

Once I get the device, I can then do

role_sw = usb_role_switch_get(dev);

This will give me the role switch reference which can be used to set
device or host role.

The advantage here is that

-> I am keeping the existing yamls untouched which won't be the case if
I map EUD as child of dwc3.

->This is also consistent with EUD hardware since EUD is not a connector
in itself but it needs a connector to function
because it can work only in device mode so the role-switch needs to be
done.

Let me know your thoughts regarding this.

Thanks,
Souradeep

>>
>> >
>> > > >
>> > > > > +
>> > > > > + interrupts:
>> > > > > + description:
>> > > > > + EUD interrupt
>> > > > > +
>> > > > > label:
>> > > > > description: Symbolic name for the connector.
>> > > > >
>> > > > > --
>> > > > > QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a
>> > > > > member
>> > > > > of Code Aurora Forum, hosted by The Linux Foundation
>> > > > >
>> > > > >