ath11k: qca6390 firmware crashes while connecting to 160 MHz channel

[Wen Gong]

On 2021-03-01 15:54, Kalle Valo wrote:
> Wen Gong <wgong at codeaurora.org> writes:
> 
>> On 2021-02-14 02:36, Justin Mazzola Paluska wrote:
>>> On 2/12/21 1:16 AM, Kalle Valo wrote:
>>>> Justin Mazzola Paluska <jmp at justinmp.com> writes:
>>>> 
>>>>> On 2/11/21 6:10 AM, Kalle Valo wrote:
>>>>>> (changing the title to find this easier)
>>>>>> 
>>>>>> Justin Mazzola Paluska <jmp at justinmp.com> writes:
>>>>>> 
>>>>>>> On 12/21/20 12:11 PM, Kalle Valo wrote:
>>>>>>>>> My router supports 160 MHz channels in the 5 GHz bands [1].  I
>>>>>>>>> switched the router to advertising 80 MHz channels and the
>>>>>>>>> card could
>>>>>>>>> connect.
>>>>>>>> Thanks for the detailed debugging, this should help us
>>>>>>>> reproducing the
>>>>>>>> problem. What's the exact openwrt version you are using?
>>>>>>> The OpenWRT version is:
>>>>>>> 
>>>>>>> OpenWrt 19.07.5 r11257-5090152ae3 / LuCI openwrt-19.07 branch
>>>>>>> git-20.341.57626-51f55b5
>>>>>>> 
>>>>>>> (This is the latest stable release.)
>>>>>> Thanks. I got a report that this might be due to commit 
>>>>>> 3579994476b6
>>>>>> ("wireless: fix wrong 160/80+80 MHz setting"). Can you try
>>>>>> reverting the
>>>>>> commit just to confirm we are seeing the same issue?
>>>>> I will try over the weekend.  Shall I still use the
>>>>> ath11k-qca6390-bringup branch or something else?  (I'm still 
>>>>> running
>>>>> the ath11k-qca6390-bringup branch on my laptop).
>>>> Yes, use the same branch. Least amount of changes that way.
>>> 
>>> I reverted 3579994476b6 and re-enabled 160 MHz channels on my 
>>> router. 
>>> I was able to connect to my router using the 5 GHz band at 80 MHz (as
>>> expected).
>>> 
>>> It seems that reverting 3579994476b6 fixed the firmware crash with 
>>> 160
>>> MHz channels!
>>> 
>>> I'll leave my router in 160 MHz mode for a while to see if there are
>>> any problems that crop up over time.
>> 
>> Does anyone trying revert this patch(3579994476b6 wireless: fix wrong
>> 160/80+80 MHz setting) upstream?
> 
> Yes, Justin reports above that he reverted commit 3579994476b6 and the
> firmware did not crash after that.
I know that.
Has someone sent patch to revert it in public?