ath11k: QCA6390 support on Linux

[Kalle Valo]

Kenny Johansson <wirehell at gmail.com> writes:

> Short story:
> * Firmware matches
> * Reverting 7fef431be9c9 does not work
> * fd5ad4d1e980 works
> * 3650b228f83a does not work

An excellent summary, makes it a lot easier to understand what's
happening :)

> Details:
>
> Firmware matches:
> ~/qca$ md5sum /lib/firmware/ath11k/QCA6390/hw2.0/*
> a101dc90f8e876f39383b60c9da64ec4  /lib/firmware/ath11k/QCA6390/hw2.0/amss.bin
> 4c0781f659d2b7d6bef10a2e3d457728  /lib/firmware/ath11k/QCA6390/hw2.0/board-2.bin
> d4c912a3501a3694a3f460d13de06d28  /lib/firmware/ath11k/QCA6390/hw2.0/m3.bin

Good, we can rule out firmware corruption then.

> The commit 3650b228f83a worked
>
> [    7.452884] ath11k_pci 0000:6c:00.0: WARNING: ath11k PCI support is
> experimental!
> [    7.452926] ath11k_pci 0000:6c:00.0: BAR 0: assigned [mem
> 0xb4200000-0xb42fffff 64bit]
> [    7.452939] ath11k_pci 0000:6c:00.0: enabling device (0000 -> 0002)
> [    7.941932] ath11k_pci 0000:6c:00.0: Respond mem req failed,
> result: 1, err: 0
> [    7.941933] ath11k_pci 0000:6c:00.0: qmi failed to respond fw mem req:-22
> [    7.952897] ath11k_pci 0000:6c:00.0: chip_id 0x0 chip_family 0xb
> board_id 0xff soc_id 0xffffffff
> [    7.952897] ath11k_pci 0000:6c:00.0: fw_version 0x101c06cc
> fw_build_timestamp 2020-06-24 19:50 fw_build_id
> [    8.156172] ath11k_pci 0000:6c:00.0 wlp108s0: renamed from wlan0

I think you have a typo here and you mean that commit fd5ad4d1e980
worked, right? The one which is based on v5.9-rc6:

fd5ad4d1e980 wcn36xx: Advertise beacon filtering support in bmps

> I will try to use fd5ad4d1e980 a bit more to make sure it is not
> something intermittent, though I suspect not as the bringup-branch
> seemed to work consistently, while the latest seems consistently
> broken for me.

Yes, it's a good idea to rule out all intermittent issues. This is a
weird problem and any randomness might confuse the testing results.

Even if your symptoms are different I suspect this is related to Pavel's
problem, where the number one suspect is high physical addresses are
causing problems:

http://lists.infradead.org/pipermail/ath11k/2020-November/000556.html

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

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