Message-ID: <2dab1111-49fd-43b7-8624-2d61b3d546b1@google.com>
Date: Mon, 16 Sep 2024 18:59:47 -0700
From: Amit Sunil Dhamne <amitsd@...gle.com>
To: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>, robh+dt@...nel.org,
 krzysztof.kozlowski+dt@...aro.org, conor+dt@...nel.org,
 heikki.krogerus@...ux.intel.com, gregkh@...uxfoundation.org
Cc: linux-kernel@...r.kernel.org, kyletso@...gle.com, rdbabiera@...gle.com,
 Badhri Jagan Sridharan <badhri@...gle.com>, linux-usb@...r.kernel.org,
 devicetree@...r.kernel.org
Subject: Re: [RFC 1/2] dt-bindings: connector: Add property to set pd timer
 values

Hi Krzysztof,

Thanks for the review!

On 9/16/24 9:05 AM, Krzysztof Kozlowski wrote:
> On 11/09/2024 02:07, Amit Sunil Dhamne wrote:
>> This commit adds a new property "pd-timers" to enable setting of
>> platform/board specific pd timer values for timers that have a range of
>> acceptable values.
>>
>> Cc: Badhri Jagan Sridharan <badhri@...gle.com>
>> Cc: linux-usb@...r.kernel.org
>> Cc: devicetree@...r.kernel.org
>> Signed-off-by: Amit Sunil Dhamne <amitsd@...gle.com>
> Please work on mainline, not ancient tree. You cannot get my CC address
> like that from mainline.
I was working off gregkh's tree on usb-next branch as that's suggested 
for USB development.


> It's not possible. So either you don't develop
> on mainline or you don't use get_maintainers.pl/b4/patman.
>
The above branch and even the tree on Linus' master branch has you
listed as a maintainer
(https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/MAINTAINERS#n17181).
I guess that's why the get_maintainers script probably returned your
email id when I ran it. Please let me know if I missed something :).


>> ---
>>   .../bindings/connector/usb-connector.yaml     | 23 +++++++++++++++++++
>>   include/dt-bindings/usb/pd.h                  |  8 +++++++
>>   2 files changed, 31 insertions(+)
>>
>> diff --git a/Documentation/devicetree/bindings/connector/usb-connector.yaml b/Documentation/devicetree/bindings/connector/usb-connector.yaml
>> index fb216ce68bb3..9be4ed12f13c 100644
>> --- a/Documentation/devicetree/bindings/connector/usb-connector.yaml
>> +++ b/Documentation/devicetree/bindings/connector/usb-connector.yaml
>> @@ -253,6 +253,16 @@ properties:
>>   
>>       additionalProperties: false
>>   
>> +  pd-timers:
>> +    description: An array of u32 integers, where an even index (i) is the timer (referenced in
>> +      dt-bindings/usb/pd.h) and the odd index (i+1) is the timer value in ms (refer
> timer of what? OS behavior?
In the context of USB Type C Power Delivery (PD), timers are run on the 
typec protocol driver
(usb/typec/tcpm/tcpm.c).
These are used to keep track of min/max or range of time required to 
enter a PD state with the
goal of a successful USB typec capabilities negotiation.  Eg., the timer 
PD_TIMER_SINK_WAIT_CAP (referred to as SinkWaitCapTimer in spec)would be 
responsible to keep track of whether a power source sent us (as sink) PD 
source capabilities pkts within 600ms (say), if yes, then we would 
transition to the next state or do a state machine reset. USB PD 3.1 
spec refers to these elements as timers and therefore referred to as 
such here.


>> +      "Table 6-68 Time Values" of "USB Power Delivery Specification Revision 3.0, Version 1.2 " for
>> +      the appropriate value). For certain timers the PD spec defines a range rather than a fixed
>> +      value. The timers may need to be tuned based on the platform. This dt property allows the user
> Do not describe what DT is. We all know what DT properties allow.
> Instead describe how this relates to hardware or boards.
>
> All this is wrongly wrapped. See Coding style (and I am not telling you
> the value on purpose, so you will read the coding style) .


Ack. Thanks for pointing it out, I will fix both the above in the next 
revision.


>
>> +      to assign specific values based on the platform. If these values are not explicitly defined,
>> +      TCPM will use a valid default value for such timers.
> And what is the default?

Defaults are given in (include/linux/usb/pd.h). But I guess I should 
have probably mentioned
that here.


>
>> +    $ref: /schemas/types.yaml#/definitions/uint32-array
> I guess you want matrix here.

Yes, I should have. Though, I will be re-implementing this such that 
each timer is represented
as a separate property based on Rob and Dmitry's suggestion in
https://lore.kernel.org/lkml/20240916163328.GA394032-robh@kernel.org/ .

>> +
>>   dependencies:
>>     sink-vdos-v1: [ sink-vdos ]
>>     sink-vdos: [ sink-vdos-v1 ]
>> @@ -478,3 +488,16 @@ examples:
>>               };
>>           };
>>       };
>> +
>> +  # USB-C connector with PD timers
>> +  - |
>> +    #include <dt-bindings/usb/pd.h>
>> +    usb {
>> +        connector {
>> +            compatible = "usb-c-connector";
>> +            label = "USB-C";
>> +            pd-timers =
>> +                <PD_TIMER_SINK_WAIT_CAP 600>,
>> +                <PD_TIMER_CC_DEBOUNCE 170>;
> Incorporate it into existing example.
>
Ack.


>> +        };
>> +    };
>> diff --git a/include/dt-bindings/usb/pd.h b/include/dt-bindings/usb/pd.h
>> index e6526b138174..6c58c30f3f39 100644
>> --- a/include/dt-bindings/usb/pd.h
>> +++ b/include/dt-bindings/usb/pd.h
>> @@ -465,4 +465,12 @@
>>   	 | ((vbm) & 0x3) << 15 | (curr) << 14 | ((vbi) & 0x3f) << 7	\
>>   	 | ((gi) & 0x3f) << 1 | (ct))
>>   
>> +/* PD Timer definitions */
>> +/* tTypeCSinkWaitCap (Table 6-68 Time Values, USB PD3.1 Spec) */
> Please expand this a bit, so we won't have to reach to external sources.

Ack.

I will incorporate all of your review comments.

Since you are no longer maintaining the
"OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" component, please let 
me know
if you'd still like to be CC'ed in the subsequent revisions.


Thanks,

Amit


>> +#define PD_TIMER_SINK_WAIT_CAP		0
>> +/* tPSSourceOff (Table 6-68 Time Values, USB PD3.1 Spec) */
>> +#define PD_TIMER_PS_SOURCE_OFF		1
>> +/* tCCDebounce (Table 4-33 CC Timing, USB Type-C Cable & Connector Spec Rel2.2) */
>> +#define PD_TIMER_CC_DEBOUNCE		2
>
> Best regards,
> Krzysztof
>

Powered by blists - more mailing lists