[PATCH ath-next v4 11/11] wifi: ath12k: use link_id_map to resolve RX link id for all devices

[Rameshkumar Sundaram]

On 11/10/2026 12:06 PM, Baochen Qiang wrote:
> 
> 
> On 11/10/2026 2:07 PM, Rameshkumar Sundaram wrote:
>> On 11/2/2026 12:08 PM, Baochen Qiang wrote:
>>> Commit a2fe9dc70f3b ("wifi: ath12k: Fix low MLO RX throughput on
>>> WCN7850") added a set_rx_link_id() op so that each device could resolve
>>> the RX link id in its own way: QCN9274/IPQ5332 index
>>> dp_peer->link_id_map[] with the hw_link_id, while WCN7850/QCC2072 walk
>>> dp_peer->link_peers[] and match by peer_id, tracked with a link_peers_map
>>> bitmap.
>>>
>>> Now that ath12k_skb_rxcb::drv_link_id carries the driver link id for all
>>> devices (the hw_link_id on QCN9274/IPQ5332 and the logical link id on
>>> WCN7850/QCC2072) and ath12k_dp_peer::link_id_map maps it to the IEEE80211
>>> link id, the RX path can resolve the link id uniformly.
>>>
>>> Revert the set_rx_link_id() machinery and the link_peers_map bitmap, and
>>> derive the link id directly from link_id_map[drv_link_id] in
>>> ath12k_dp_rx_deliver_msdu() for all devices. The NULL dp_peer guard added
>>> to ath12k_mac_peer_cleanup_all() is kept as it is an independent fix.
>>>
>>
>> Fixes: a2fe9dc70f3b ("wifi: ath12k: Fix low MLO RX throughput on WCN7850") ??
> 
> hmm, not sure here, since it would need to backport the whole series which is too large
> for a stable backport.
> 


Yeah, makes sense. Let's leave out the Fixes tag.

-- 
Ramesh