Re: [PATCH ath-next] wifi: ath12k: fix MLO 4-way handshake timeout on QCC2072

[Jeff Johnson]

On 8/14/2026 6:43 PM, Miaoqing Pan wrote:
> In MLO connections, ath12k_wifi7_dp_rx_set_link_id_wcn7850() attempts
> to resolve the IEEE 802.11 link ID for received frames by matching
> rxcb->peer_id against the link_peer->peer_id entries in
> dp_peer->link_peers[].
> 
> This fails for EAPOL frames because the firmware delivers them via
> the MLD peer path. In this case, rxcb->peer_id contains an MLD peer
> ID while link_peer->peer_id contains a link peer ID. The mismatch
> prevents status->link_valid from being set.
> 
> When mac80211 receives a data frame with link_valid=0 for an MLO
> station, it calls link_sta_info_get_bss() with hdr->addr2 to resolve
> the link ID. If hdr->addr2 contains the AP MLD address rather than a
> link address, the lookup fails because link_sta_info_get_bss()
> indexes by link address, and the frame is dropped in
> __ieee80211_rx_handle_packet(). This is observed on QCC2072 where
> EAPOL M3 uses the MLD address,
> 
>   deauthenticated (Reason: 15=4WAY_HANDSHAKE_TIMEOUT)
> 
> Fix this by resolving the link peer using the current operating
> frequency associated with each link peer.
> 
> For 6 GHz links, the received frequency corresponds to the operating
> center frequency, while for other bands it corresponds to the primary
> channel frequency. Match the received frame against associated links
> using the appropriate frequency for each band.
> 
> The frequency-based lookup is used only when exactly one associated
> link matches the received frequency. If multiple links share the same
> frequency, the match becomes ambiguous and no link ID is assigned. In
> that case, leave status->link_valid unset and let mac80211 fall back
> to its own link resolution logic.
> 
> Tested-on: QCC2072 hw1.0 PCI WLAN.COL.1.0.c2-00228-QCACOLSWPL_V1_TO_SILICON-1
> Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c7-00108-QCAHMTSWPL_V1.0_V2.0_SILICONZ_UPSTREAM-3
> 
> Fixes: a2fe9dc70f3b ("wifi: ath12k: Fix low MLO RX throughput on WCN7850")
> Signed-off-by: Miaoqing Pan <miaoqing.pan@xxxxxxxxxxxxxxxx>
> ---

FYI this no longer applies to ath/main, can you please rebase?

Applying: wifi: ath12k: fix MLO 4-way handshake timeout on QCC2072
Using index info to reconstruct a base tree...
M       drivers/net/wireless/ath/ath12k/wifi7/dp_rx.c
Falling back to patching base and 3-way merge...
Auto-merging drivers/net/wireless/ath/ath12k/wifi7/dp_rx.c
CONFLICT (content): Merge conflict in drivers/net/wireless/ath/ath12k/wifi7/dp_rx.c