Re: [PATCH 1/2] wifi: ath12k: extend the link capable flag

[Kalle Valo <kvalo@kernel.org>]

Jeff Johnson <quic_jjohnson@quicinc.com> writes:

> On 4/2/2024 2:32 AM, Karthikeyan Periyasamy wrote:
>
>> 
>> 
>> On 4/1/2024 10:24 PM, Jeff Johnson wrote:
>>> On 3/28/2024 6:23 PM, Karthikeyan Periyasamy wrote:
>>>> Link capability categorized as Single Link Operation (SLO) and Multi Link
>>>> Operation (MLO).
>>>>
>>>> 	* Intra-chip SLO/MLO refers to links present within a chip
>>>> 	* Inter-chip SLO/MLO refers to links present across multiple chips
>>>
>>> Is "chip" the correct term?
>>>
>>> I'm thinking that this should be called "device" since that is the unit of
>>> hardware that is detected by a bus probe function and which is handled by a
>>> *device* driver.
>>>
>>> Doesn't this make more sense if the references to chip and SoC are changed to
>>> device?
>>>
>> 
>> In the QMI, SLO/MLO parameter exposed as chip only not device. So 
>> followed the same terminology to avoid confusion for code readability.
>> 
>> struct wlfw_host_mlo_chip_info_s_v01 {
>>          u8 chip_id;
>>          u8 num_local_links;
>>          u8 hw_link_id[QMI_WLFW_MAX_NUM_MLO_LINKS_PER_CHIP_V01];
>>          u8 valid_mlo_link_id[QMI_WLFW_MAX_NUM_MLO_LINKS_PER_CHIP_V01];
>> };
>> 
>> struct qmi_wlanfw_host_cap_req_msg_v01 {
>> 
>> ...
>> 
>> u8 mlo_num_chips;
>> 
>> u8 mlo_chip_info_valid;
>> 
>> struct wlfw_host_mlo_chip_info_s_v01 
>> mlo_chip_info[QMI_WLFW_MAX_NUM_MLO_CHIPS_V01];
>> 
>> ...
>> 
>> }
>> 
>
> Please don't let firmware interface naming drive host driver code naming.
> And push back on the firmware team when they've introduced poor naming.
>
> As many Software Engineering experts stress, naming is probably the single
> most important thing we do. So we need to make sure we are using the correct
> names for all of the software objects that comprise the driver, especially
> with this multi-device MLO feature where we now have to represent a multitude
> of individual devices as a single logical wiphy.
>
> Lack of a single common term for each object in the architecture makes the
> code far less maintainable.

Amen to that. I think we should come up with a terminology list or
something, otherwise it's hard to keep up with all teams having their
own terminology.

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches