Re: [PATCH 1/1] wifi: ath12k: support calibration-variant from device tree

[Andrew LaMarche]

I will prepare v2 this week.

> On Jun 29, 2026, at 4:21 AM, Ernest Van Hoecke <ernestvanhoecke@xxxxxxxxx> wrote:
> 
> On Mon, Jun 15, 2026 at 03:36:50PM +0200, Ernest Van Hoecke wrote:
>> On Wed, May 27, 2026 at 05:01:51PM -0400, Andrew LaMarche wrote:
>>> Ack on the deprecation of qcom,ath12k-calibration-variant in favor of the
>>> generic qcom,calibration-variant. 
>>> 
>>> However, drivers/net/wireless/ath/ath12k/core.c still misses the logic to 
>>> actually load in the BDF, which this patch also accomplishes. I don’t see that
>>> in the Qualcomm-authored series you linked. Perhaps a v2 to address this?
>>> 
>>> Andrew
>>> 
>>>> On May 27, 2026, at 3:52 PM, Jeff Johnson <jeff.johnson@xxxxxxxxxxxxxxxx> wrote:
>>>> 
>>>> On 5/27/2026 9:12 AM, Andrew LaMarche wrote:
>>>>> Hi,
>>>>> 
>>>>> A kind ping here. I’m not sure why this functionality is missing in the first place, but it is needed for loading caldata from the device tree.
>>>> 
>>>> Your patch duplicates functionality in the Qualcomm authored series:
>>>> https://msgid.link/20250228184214.337119-1-quic_rajkbhag@xxxxxxxxxxx
>>>> 
>>>> And note the upstream device bindings for ath10k and ath11k only support the
>>>> generic binding qcom,calibration-variant.
>>>> 
>>>> There are no longer any generation-specific bindings, see:
>>>> https://msgid.link/20250225-b-wifi-qcom-calibration-variant-v1-0-3b2aa3f89c53@xxxxxxxxxx
>>>> 
>>>> /jeff
>>> 
>> 
>> Hi Jeff,
>> Hi Andrew,
>> 
>> We also need this functionality, and as Andrew noted, the patch series
>> from Qualcomm does not actually address this.
>> 
>> I believe some confusion arose because the patch series mentions the
>> "qcom,calibration-variant" DT property, but it does nothing to implement
>> this in the driver. In that series it only exists within a schema file:
>> Documentation/devicetree/bindings/net/wireless/qcom,ipq5332-wifi.yaml
>> 
>> Our need for this property was previously discussed here, and for now I
>> don't see a way forward without it:
>> https://lore.kernel.org/all/77vowy4ax4cl6dlc45i2q3fjmwn3q676wqghq267tmbix7773b@27h5t66mflur/
>> 
>> Thanks a lot for your work here.
>> 
>> Best regards,
>> Ernest
>> 
> 
> Hi Andrew,
> 
> Are you planning to send a v2 of this?
> 
> If not, I'm considering sending a version of this with the
> qcom,calibration-variant binding to move this forward.
> 
> Thanks for your work here!
> 
> Kind regards,
> Ernest
>