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

[Burke Cates]

On Mon, Nov 02, 2026 at 04:02:56PM +0800, Baochen Qiang wrote:
> Hi Burke, can you please try if below series can fix your issue as well?
>
> https://lore.kernel.org/all/20261102-ath12k-logical-link-id-v4-0-f4c7807c299b@xxxxxxxxxxxxxxxx/

Yes, it fixes it here too. I took Miaoqing's patch out and applied
v4 1-11 instead, on v7.4-rc1 plus three ath12k commits from ath.git
(c752394249f8 and 02121aa4a012 from ath-next, 1e678e37ff13 from
ath-current) and out-of-tree Glymur platform/board patches; the tree
also carries Miaoqing's dp_peer reconnect fix (8b6b871332b5). All 11
applied as posted.

On the tri-band MLO SSID that failed with reason 15 (SAE, PMF
required, 2.4+5+6 GHz):

- 12 of 12 connect/disconnect cycles completed the 4-way handshake and
  came up as MLO with three links (2.4 and 6 GHz active), ping OK each
  time; no reason 15 in 23 handshakes over the session
- 5 of 5 rejoins after s2idle resume
- non-MLO WPA2 on 5 GHz still fine

This firmware sends the PEER_EXTENDED event (two per association), so
the host allocation path in 8/11 was not exercised. I'll send a
Tested-by on the cover letter.

Tested-on: QCC2072 hw1.0 PCI WLAN.COL.1.0.c2-00277-QCACOLSWPL_V1_TO_SILICONZ-1

AI coding assistants helped run these tests and draft this reply; the
results are from the hardware runs.

Thanks,
Burke