ath11k-qca6390-bringup-202012140938: rebate to v5.10 and MHI M2 workaround

[Kalle Valo]

Thomas Krause <thomaskrause at posteo.de> writes:

> Am 14.12.20 um 10:56 schrieb Kalle Valo:
>> (Bcc: people reporting qca6390 problems)
>>
>> Hi,
>>
>> there are updates in ath11k-qca6390-bringup branch[1] and the new tag
>> is:
>>
>> ath11k-qca6390-bringup-202012140938

[...]

> just compiled the kernel with the attached settings and even with the
> messages in dmesg, it seems to work extremely well on my Dell XPS
> 9310. In addition to use the updated bringup-branch, I also updated my
> BIOS to version 1.1.4 and disabled the "Fastboot" mode in the BIOS
> setup.

Unfortunately the new BIOS still has the DMAR bug:

[    0.050968] DMAR: [Firmware Bug]: Your BIOS is broken; DMAR reported at address 0!
               BIOS vendor: Dell Inc.; Ver: 1.1.4; Product Version: 

And ath11k only gets 1 MSI vector:

[   16.085779] ath11k_pci 0000:56:00.0: MSI vectors: 1

>> In the reports try to include at least:
>>
>> * name of the bringup branch tag you used
>
> ath11k-qca6390-bringup-202012140938
>
>> * if there are other changes in the kernel, patches, reverts etc
>
> No code changes to the kernel, just some kernel configs enabled as
> previously suggested (see attached config).
>
>> * do you use the memmap kernel parameter
>
> This kernel parameter is not set explicitly in the boot options
>
>> * make and model of the laptop/computer (eg. Dell XPS 13 9310)
>
> It is the Dell XPS 13 9310
>
>> * how much testing you did
>
> I tested both the Wifi and also connected some Bluetooth headsets
> simultaneously (and streaming some music and downloads to stress test
> both connections). My tests ran for about 45 minutes, in which I put
> the system to standby three times and also did some wandering around
> in the hope to have a handover between different access points. During
> all this time, the connection seemed to be stable, even if the Gnome
> Network Manager sometimes reported a question mark as status. Also,
> both Wifi an Bluetooth were available after each standby.

Great, thanks for testing. This is good info.

> Thank everybody involved for their work on this, if there are any out
> debug outputs or configurations I should share I would be happy to.

A big thank you goes to wink as he has been doing extensive debugging
and found the workaround to disable M3 MHI state. Great work!

Though I doubt we can upstream the MHI workaround so we still need to
find the root cause for the problems with 1 MHI vector patch. So expect
more testing requests from me in the future :)

But I have now ordered Dell XPS 13 laptop and I should get it in three
weeks, keeping fingers crossed. That should make things easier as I can
test patches myself.

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches