ath11k multicast action frame RX

[James Prestwood]

On 1/24/24 5:06 AM, James Prestwood wrote:

> Hi,
>
> I recently added support to ath10k, for the QCA6174, to receive 
> multicast action frames in order to support DPP. I'm trying to do this 
> for the ath11k QCNFA765/WCN6855. I took the same approach as ath10k 
> which was actually quite simple but I'm unable to see any multicast 
> frames coming over even when I enable RX/DATA debugging.
>
> What I've done so far is:
>
>  - Add FIF_MCAST_ACTION to the supported filters list
>
>  - Created/started a monitor vdev from within 
> ath11k_mac_op_configure_filter()
>
>     ath11k_mac_monitor_vdev_create(ar);
>
>     ath11k_mac_monitor_start(ar);
>
> - Also tried adding
>
>     ath11k_mac_config_mon_status_default(ar, true);
>
> I can successfully create/start the monitor vdev. I see now as I'm 
> writing this email that my specific hardware does not support monitor 
> mode outright... Is this a dead end or is there hope with purely 
> driver changes? or is it some filtering at the firmware level 
> preventing this?

I was able to finally see my broadcast frame after some additional 
testing. It seems the driver/firmware/something gets a bit hung up when 
I attempt to put it in monitor mode. It will start printing these 
messages for about a minute:

[   51.093034] ath11k_pci 0000:00:06.0: data dest_rx: new ppdu_id a1a2 
!= status ppdu_id a1a1 dest_mon_not_reaped = 650 dest_mon_stuck = 12
[   51.094461] ath11k_pci 0000:00:06.0: data dest_rx: new ppdu_id a1a3 
!= status ppdu_id a1a2 dest_mon_not_reaped = 651 dest_mon_stuck = 12
[   51.096281] ath11k_pci 0000:00:06.0: data dest_rx: new ppdu_id a1a5 
!= status ppdu_id a1a3 dest_mon_not_reaped = 652 dest_mon_stuck = 12

In most cases these ppdu ID's are a one-off from the expected ID which 
seemed weird to me. But if I let it sit long enough I all of a sudden 
get a ton of "data rx" messages, which is what I would expect if I 
enabled a monitor vdev. Once this flurry starts I see the action frame 
from my other client, sometimes at least. The firmware generally crashes 
after a while so I've still got a ways to go but it seems within the 
realm of possibility that this device could receive broadcast action frames?

Hoping for some breadcrumbs to follow here...

Thanks,

James

>
> Thanks,
>
> James
>
>