Re: Ordering Dell XPS 13 9310 to test QCA6390

From: Kalle Valo <kvalo@codeaurora.org>
To: wi nk <wink@technolu.st>
Cc: "ath11k@lists.infradead.org" <ath11k@lists.infradead.org>,
	Mitchell Nordine <mail@mitchellnordine.com>
Subject: Re: Ordering Dell XPS 13 9310 to test QCA6390
Date: Mon, 22 Feb 2021 18:10:08 +0200	[thread overview]
Message-ID: <87czwshyn3.fsf@codeaurora.org> (raw)
In-Reply-To: <CAHUdJJVk3jDY9FGjPTuDTF2+QK+B8f8s6ZqMA+8uH-7Yd==Mmw@mail.gmail.com> (wi nk's message of "Wed, 13 Jan 2021 13:02:22 +0100")

wi nk <wink@technolu.st> writes:

> On Wed, Jan 13, 2021 at 12:54 PM Kalle Valo <kvalo@codeaurora.org> wrote:
>>
>> wi nk <wink@technolu.st> writes:
>>
>> >> I got the XPS 13 9310 now and it indeed had AX500. But it also has an
>> >> updated BIOS:
>> >>
>> >> [    0.000000] DMI: Dell Inc. XPS 13 9310/0MRT12, BIOS 1.1.4 11/23/2020
>> >>
>> >> And with that I can get 32 MSI vectors:
>> >>
>> >> [ 2.758219] ath11k_pci 0000:56:00.0: WARNING: ath11k PCI support
>> >> is experimental!
>> >> [ 2.758778] ath11k_pci 0000:56:00.0: BAR 0: assigned [mem
>> >> 0x8c300000-0x8c3fffff 64bit]
>> >> [    2.758797] ath11k_pci 0000:56:00.0: enabling device (0000 -> 0002)
>> >> [    2.759086] ath11k_pci 0000:56:00.0: MSI vectors: 32
>> >> [    3.208447] ath11k_pci 0000:56:00.0: qmi req mem_seg[0] 0x3f000000 3522560 1
>> >> [    3.208449] ath11k_pci 0000:56:00.0: qmi req mem_seg[1] 0x3f500000 884736 4
>> >> [ 3.215355] ath11k_pci 0000:56:00.0: chip_id 0x0 chip_family 0xb
>> >> board_id 0xff soc_id 0xffffffff
>> >> [ 3.215358] ath11k_pci 0000:56:00.0: fw_version 0x101c06cc
>> >> fw_build_timestamp 2020-06-24 19:50 fw_build_id
>> >
>> >   I did see the updated BIOS from a few weeks ago, but I've been a
>> > little hesitant to update since I wasn't sure what would happen or if
>> > I could downgrade.
>>
>> Makes sense, better to be careful.
>>
>> > So to confirm, the new BIOS is allowing the driver to obtain all 32
>> > msi vectors removing the need for the single MSI patching? Does the
>> > MHI / M2 behavior also change to work properly?
>>
>> I have not tested these yet as I only did a quick 10 minute test. I was
>> offline over two weeks so I have a lot of catching up to do, but will
>> get back to you in next few days on these. My plan is to also test this
>> new BIOS with the official 5.10 release from Linus (tag v5.10).
>>
>
> Awesome, let me know how it goes.  I've got the 9310 branch running
> 5.10.2 with no issues (16 days uptime) so I wouldn't expect you to see
> anything super new there, good luck :D.

Heh, "next few days" ended to be a month :) But anyway, I have been
using the laptop quite a lot during the last month. Windows seems to
update the BIOS without asking me and now I'm at "BIOS 1.2.5
12/10/2020". I reliably get 32 MSI vectors and I haven't see any of the
M2 kernel crashes.

I mostly use master branch of my ath.git tree, which contains latest
ath11k and other wireless components. It's working fine for me and I
haven't had crashes or anything like, but there seems to some issues
with roaming which I need to debug.

I also briefly tested tag ath11k-qca6390-bringup-202012140938 and was
able to force use of 1 MSI vectors by disabling VT-d from the BIOS.
Everything was still working fine, but after I reverted the M2
workaround kernel started crashing and I was not able to even log in to
the system. This is very good as now I have a way to test the M2 problem
even with latest BIOS.

Unfortunately 1 MSI vector support didn't make it to v5.12 as it was
crashing for me[1], even with just 32 MSI vectors, and it has to wait
v5.13.

[1] https://patchwork.kernel.org/project/linux-wireless/cover/20201223030225.2345-1-cjhuang@codeaurora.org/

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

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

-- 
ath11k mailing list
ath11k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath11k