Re: [PATCH] ath11k: workaround firmware bug where peer_id=0

[Baochen Qiang <baochen.qiang@oss.qualcomm.com>]

On 4/14/2026 8:54 PM, Matthew Leach wrote:
> Hi Baochen,
> 
> Baochen Qiang <baochen.qiang@oss.qualcomm.com> writes:
> 
>> On 3/30/2026 3:57 PM, Matthew Leach wrote:
>>> Hello,
>>>
>>> Matthew Leach <matthew.leach@collabora.com> writes:
>>>
> 
> [...]
> 
>> for chips like QCA2066 and WCN6855 etc 0 is a valid value, however
>> this is not for chips like QCN9074 etc.
>>
>> so a possible fix would be to add hardware ops based on chips: for
>> QCN9074 we keep the existing validation on 0 in the ops, while for
>> QCA2066 the ops is a null func. Or even simper we can remove the
>> validation for all chips.
> 
> In that case, does it make sense to remove the condition check
> 
> if (rxcb->peer_id)
> 
> in ath11k_dp_rx_h_find_peer()? It looks like this has been used as a
> small optimisation, where if peer_id isn't valid it skips checking for
> it in the peer hash table. However, if on newer chips peer_id=0 is
> valid, we should remove this?

yeah, I think so. This check was also based on the non-zero peer id assumption.

> 
> Regards,