Re: [PATCH] wifi: ath11k: fix txpower in ap mode for 6 Ghz

Am 13.05.2026 um 12:12 schrieb Baochen Qiang:
>
> On 5/13/2026 1:20 PM, Sebastian Gottschall wrote:
>> Am 13.05.2026 um 03:56 schrieb Baochen Qiang:
>>> On 5/13/2026 4:48 AM, Sebastian Gottschall wrote:
>>>> there is an issue which has been discovered a while ago while testing
>>>> which killed txpower to 0dbm once scanning is triggered in AP mode
>>>> on QCN9074 based chipsets if 6 Ghz is in use.
>>>> ath11k_wmi_send_vdev_set_tpc_power must be set in AP mode in the same
>>>> way is it is for STA as it is implemented in ath12k and in the
>>>> qsdk version for ath11k.
>>>> this patch must be considered to be backported (i discovered that issue
>>>> already 2 years ago)
>>>>
>>>> Signed-off-by: Sebastian Gottschall <[email protected]>
>>>> ---
>>>>    drivers/net/wireless/ath/ath11k/mac.c | 9 +++++----
>>>>    1 file changed, 5 insertions(+), 4 deletions(-)
>>>>
>>>> diff --git a/drivers/net/wireless/ath/ath11k/mac.c b/drivers/net/wireless/ath/ath11k/mac.c
>>>> index a48b6bf1f29a..dc192152b2e8 100644
>>>> --- a/drivers/net/wireless/ath/ath11k/mac.c
>>>> +++ b/drivers/net/wireless/ath/ath11k/mac.c
>>>> @@ -3497,13 +3497,14 @@ static int ath11k_mac_config_obss_pd(struct ath11k *ar,
>>>>        return 0;
>>>>    }
>>>>    -static bool ath11k_mac_supports_station_tpc(struct ath11k *ar,
>>>> +static bool ath11k_mac_supports_tpc(struct ath11k *ar,
>>>>                            struct ath11k_vif *arvif,
>>>>                            const struct cfg80211_chan_def *chandef)
>>>>    {
>>>>        return ath11k_wmi_supports_6ghz_cc_ext(ar) &&
>>>>            test_bit(WMI_TLV_SERVICE_EXT_TPC_REG_SUPPORT, ar->ab->wmi_ab.svc_map) &&
>>>> -        arvif->vdev_type == WMI_VDEV_TYPE_STA &&
>>>> +        (arvif->vdev_type == WMI_VDEV_TYPE_STA ||
>>>> +        arvif->vdev_type == WMI_VDEV_TYPE_AP) &&
>>>>            arvif->vdev_subtype == WMI_VDEV_SUBTYPE_NONE &&
>>>>            chandef->chan &&
>>>>            chandef->chan->band == NL80211_BAND_6GHZ;
>>>> @@ -7647,8 +7648,8 @@ ath11k_mac_vdev_start_restart(struct ath11k_vif *arvif,
>>>>        /* TODO: For now we only set TPC power here. However when
>>>>         * channel changes, say CSA, it should be updated again.
>>>>         */
>>>> -    if (ath11k_mac_supports_station_tpc(ar, arvif, chandef)) {
>>>> -        ath11k_mac_fill_reg_tpc_info(ar, arvif->vif, &arvif->chanctx);
>>>> +    if (ath11k_mac_supports_tpc(ar, arvif, chandef)) {
>>>> +        ath11k_mac_fill_reg_tpc_info(ar, arvif->vif, ctx);
>>> is this intentional change? I doubt it even doesn't compile.
>> in addition. i compile tested it. and if you talk about the change from arvid->chanctx to
>> ctx. this valid
>> arvif->chanctx is not always set. the only trustworthy source is ctx and represents the
>> current channel context.
>> there was never a reason to use arvif->chanctx
> my bad, I was referring to a wrong codebase.
>
> Yes, you are correct. Even the upstream ath12k uses ctx instead of arvif->chanctx.
yes. i found that out while reviewing. just had scratching why 
arvif->chanctx was used initially.
for me it looks this patch was ported from qsdk patches but modified to 
support station only.
reason unknown. regarding your power type comment (LPI) i have to review 
the current ath11k upstream codebase first
since i developed this patch initially for kernel 6.12.
>
>> Sebastian
>>
>>>>            ath11k_wmi_send_vdev_set_tpc_power(ar, arvif->vdev_id,
>>>>                               &arvif->reg_tpc_info);
>>>>        }
>>> the ath12k patch has additional changes to hardcode power_type to LPI for an AP vdev,
>>> should it be done for ath11k as well?
>>>
>>> commit a44958c83e52238996a48e7d604167c9b9d0a529
>>> Author: Hariharan Basuthkar <[email protected]>
>>> Date:   Fri Jun 6 16:53:16 2025 +0530
>>>
>>>       wifi: ath12k: Send WMI_VDEV_SET_TPC_POWER_CMD for AP vdev
>>>
>>>       [...]
>>>
>>>       For a AP vdev, hardcode ap_power_type as IEEE80211_REG_LPI_AP, which is
>>>       filled in the WMI_VDEV_SET_TPC_POWER_CMDID sent to the target.
>>>       Currently, LPI is the only supported AP power mode. Since there is
>>>       no method currently to set the AP power mode from user-space, the
>>>       power mode is hard-coded to LPI.
>>>
>>>
>