Wen Gong <quic_wgong@xxxxxxxxxxx> writes:
> On 10/11/2022 11:53 AM, Wen Gong wrote:
>
>> On 10/11/2022 2:09 AM, Kalle Valo wrote:
>>> Wen Gong <quic_wgong@xxxxxxxxxxx> wrote:
>>>
>>>> When switch to connect to a new AP for station which is already
>>>> connected
>>>> to an AP, the time cost is too long, it arrives 10 seconds.
>>>>
>>>> The reason is when switch connection, disconnect operation happened on
>>>> the 1st AP, then 11d scan start command sent to firmware, and then a
>>>> new hw scan arrived for the 2nd AP. The 11d scan is running at this
>>>> moment, so the hw scan can not start immediately, it needs to wait
>>>> the 11d scan finished, it increased the time cost of switch AP and
>>>> even happened scan fail as log below after apply the incoming patch.
>>>>
>>>> [ 1194.815104] ath11k_pci 0000:06:00.0: failed to start hw scan: -110
>>>> [ 1196.864157] ath11k_pci 0000:06:00.0: failed to start hw scan: -110
>>>> [ 1198.911926] ath11k_pci 0000:06:00.0: failed to start hw scan: -110
>>>>
>>>> Change to set 11d state while disconnect, and the 11d scan will be
>>>> started after the new hw scan in ath11k_mac_op_hw_scan(). Then the
>>>> time cost of switching AP is small and not happened scan fail.
>>>>
>>>> Tested-on: WCN6855 hw2.0 PCI
>>>> WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3
>>>>
>>>> Fixes: 9dcf6808b253 ("ath11k: add 11d scan offload support")
>>>> Signed-off-by: Wen Gong <quic_wgong@xxxxxxxxxxx>
>>>> Signed-off-by: Kalle Valo <quic_kvalo@xxxxxxxxxxx>
>>> I still see warnings:
>>>
>>> [ 3627.691737] ath11k_pci 0000:06:00.0: failed to start hw scan: -110
>>> [ 3715.435608] ath11k_pci 0000:06:00.0: failed to start hw scan: -110
>>> [ 3860.333214] ath11k_pci 0000:06:00.0: failed to start hw scan: -110
>>> [ 4542.000955] ath11k_pci 0000:06:00.0: failed to start hw scan: -110
>>> [ 4604.018451] ath11k_pci 0000:06:00.0: failed to start hw scan: -110
>> The failed log is now happened frequently as before, I will change
>> it again to suppress the failed log.
> change typo.
> The failed log is NOT happened frequently as before, I will change it
> again to suppress the failed log.
Please think about this more, ignoring errors doesn't sound like a good
solution.
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
