6GHz AP association failure on WCN7850

[Kang Yang]

On 3/6/2024 10:25 AM, Jun Nie wrote:
> Kang Yang <quic_kangyang at quicinc.com> 于2024年3月5日周二 16:30写道：
>>
>>
>>
>> On 3/4/2024 6:14 PM, Jun Nie wrote:
>>> Kang Yang <quic_kangyang at quicinc.com> 于2024年3月1日周五 19:42写道：
>>>>
>>>>
>>>>
>>>> On 3/1/2024 5:14 PM, Jun Nie wrote:
>>>>> Hi,
>>>>>
>>>>> I see there is association failure many times when I try to connect my
>>>>> 6GHz AP with latest mainline Linux. Does anyone also observe it too?
>>>>> Any newer firmware than WLAN.HMT.2.0-01734 is available? Or we
>>>>> recommend any specific version firmware?
>>>>>
>>>>> The issue only happens with this version firmware:
>>>>>      [    9.441624] ath12k_pci 0000:01:00.0: Hardware name: wcn7850 hw2.0
>>>>>       [   10.058822] ath12k_pci 0000:01:00.0: chip_id 0x2 chip_family 0x4
>>>>> board_id 0x3d soc_id 0x40170200
>>>>>       [   10.082876] ath12k_pci 0000:01:00.0: fw_version 0x200a06c6
>>>>> fw_build_timestamp 2023-09-20 03:43 fw_build_id
>>>>> QC_IMAGE_VERSION_STRING=WLAN.HMT.2.0-01734-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1
>>>>>
>>>>
>>>> Wrong firmware.
>>>>
>>>> We use WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3 and
>>>> WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4,
>>>>
>>>> https://github.com/kvalo/ath11k-firmware/tree/master/ath12k-testing/WCN7850/hw2.0
>>>>
>>>>> I do not see the issue with firmware:
>>>>> [    7.896737] ath12k_pci 0000:01:00.0: fw_version 0x1005bc12
>>>>> fw_build_timestamp 2022-10-11 12:13 fw_build_id
>>>>> QC_IMAGE_VERSION_STRING=WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4
>>>>>
>>>>
>>>> That's it, i recommend you to use
>>>> WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3. It's new one.
>>>>
>>>>
>>>>
>>>>> Detail log can be found at:
>>>>> https://www.dropbox.com/scl/fi/jgn1zbnw44i1vypbjxl0v/ath12k-6GHz-association-fail-log.tgz?rlkey=pg2ifnzcttzimz7vaxb7tuc6w&dl=0
>>>>>
>>>>> Jun
>>>
>>> Hi Kang,
>>>
>>> I see wlan0 failure with below log with firmware version
>>> WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4.
>>
>>
>> Can you check if you completely replaces our upstream version?
>>
>> It is best not to mix it with other versions
>> (board.bin/board.bin/regdb.bin/amss.bin/m3.bin).
>>
>> Mixed use may also have an impact on throughput.
>>
> Yeah, tried to replace all binaries from below path, but no help.
> https://github.com/kvalo/ath11k-firmware/tree/master/ath12k-testing/WCN7850/hw2.0
> https://github.com/kvalo/ath11k-firmware/tree/master/ath12k-testing/WCN7850/hw2.0/WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3
> 

So, both WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4. and 
WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3 have this problem? 
Their log is the same?




> What's your test platform? I use sm8650-hdk. Not sure whether there is something
> platform dependent.

Our test platform is x86 devices(normal x86 PC or x86 board).

These binary can only be uploaded after detailed testing on x86 platform.


> These binary was tested OK with Linux mainline on some QCOM platform, right?
> 




Your platform is arm. Actually, we never test whether upstream firmware 
of ath12k can work well on other platforms. Any errors could happen.



Did you download this firmware by git clone? Or just download manually.

Please retry with git clone if you download manually.







>>
>>
>>
>>
>>> Do I need any patch outside mainline? Actually, I add below 2 patches
>>> for 6GHz band.
>>> afb7a22d9e22 wifi: ath12k: build 6 GHz regd based on vdev type and 6
>>> GHz power type
>>> 6c60469f0bae wifi: ath12k: add support to select 6 GHz Regulatory type
>>>
>>
>>
>> If problem still exist, please revert these patches and check again.
> 
> Reverting them does not help as expected.
>>
>>
>>
>>
>>>
>>> [    6.194535] ath12k_pci 0000:01:00.0: setting mhi state: INIT(0)
>>> [    6.195067] ath12k_pci 0000:01:00.0: setting mhi state: POWER_ON(2)
>>> [    6.195069] mhi mhi0: Requested to power ON
>>> [    6.195083] mhi mhi0: Power on setup success
>>> [    6.196541] mhi mhi0: Image transfer failed
>>> [    6.196559] mhi mhi0: Reg: ERROR_CODE value: 0xef128100
>>> [    6.196574] mhi mhi0: Reg: ERROR_DBG1 value: 0xff7fc000
>>> [    6.196588] mhi mhi0: Reg: ERROR_DBG2 value: 0x0
>>> [    6.196601] mhi mhi0: Reg: ERROR_DBG3 value: 0x1988
>>> [    6.196631] mhi mhi0: MHI did not load image over BHI, ret: -5
>>>
>>
>> BTW, is it a probability problem?
> 
> It happens 100%.
>>
>>> Kalle,
>>> Thanks for the reminder of attachment size limitation!
>>>
>>> Jun