Bug 219282

Summary:	ath12k: firmware regression with update 82318c966fd1
Product:	Drivers	Reporter:	Takashi Iwai (tiwai)
Component:	network-wireless	Assignee:	Kalle Valo (kvalo)
Status:	RESOLVED CODE_FIX
Severity:	normal	CC:	christian, kvalo, mario.limonciello, pbrobinson, rezon.anselm
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:		Subsystem:
Regression:	No	Bisected commit-id:
Attachments:	output of dmesg \| grep ath12k when wifi is working output of dmesg \| grep ath12k when wifi is not working output of hwinfo output of lspci -mnn

Description Takashi Iwai 2024-09-16 13:12:57 UTC

We've received a bug report on openSUSE Tumbleweed about the regression of ath12k with the recent firmware update:
  https://bugzilla.suse.com/show_bug.cgi?id=1230596
Also for Arch Linux:
  https://bbs.archlinux.org/viewtopic.php?id=299286

The change seems relevant with linux-firmware git commit 82318c966fd1.
Reverting the change fixes the problem according to the original bug reporter.

Comment 1 Takashi Iwai 2024-09-16 13:13:53 UTC

The original report on linux-wireless ML:
  https://lore.kernel.org/87bk0nc1yb.wl-tiwai@suse.de

Comment 2 Kalle Valo 2024-09-16 13:15:15 UTC

Can someone send dmesg output?

https://wireless.wiki.kernel.org/en/users/drivers/ath12k/bugreport

Comment 3 Haak 2024-09-16 13:40:55 UTC

Created attachment 306883 [details]
output of dmesg | grep ath12k when wifi is working

Comment 4 Haak 2024-09-16 13:41:38 UTC

Created attachment 306884 [details]
output of dmesg | grep ath12k when wifi is not working

Comment 5 Haak 2024-09-16 13:42:24 UTC

Created attachment 306885 [details]
output of hwinfo

Comment 6 Haak 2024-09-16 13:42:55 UTC

Created attachment 306886 [details]
output of lspci -mnn

Comment 7 Haak 2024-09-16 13:50:53 UTC

I made the original report on bugzilla.suse. Attached the outputs of dmesg, hwinfo and lspci above.
System is a Thinkpad 14 Gen 5 (AMD) and I'm running openSUSE Tumbleweed. I also tried booting from a Tumbleweed live stick last week and Wifi didn't work there, either. Let me know if you need anything else.

Comment 8 Kalle Valo 2024-09-23 16:21:36 UTC

Jeff just uploaded a new version of board-2.bin:

https://git.codelinaro.org/clo/ath-firmware/ath12k-firmware/-/commit/6de0474a14f8eee790b9f1166346b01c3556a11b

Please test and let us know if it fixes the issue or not.

Comment 9 Peter Robinson 2024-09-24 20:22:54 UTC

> Please test and let us know if it fixes the issue or not.

Had confirmation from a Fedora user it fixes the issue for them

Comment 10 Haak 2024-09-25 07:38:21 UTC

How do I test this?
I just tried downloading the new version, using xz to compress it and replaced the board-2.bin.xz in my old, working version by it. After restarting, wifi was NOT showing up again. (Going back to the old version fixed it.)
Was this even the right way to test?

Comment 11 Peter Robinson 2024-09-25 09:00:00 UTC

(In reply to Haak from comment #10)
> How do I test this?
> I just tried downloading the new version, using xz to compress it and
> replaced the board-2.bin.xz in my old, working version by it. After
> restarting, wifi was NOT showing up again. (Going back to the old version
> fixed it.)
> Was this even the right way to test?

You don't need to compress it, just move the compressed one out of the way and put it in /lib/firmware/ath12k/WCN7850/hw2.0/, the kernel expects the xz compress with particular settings so that may be the issue, but it doesn't need to be compressed, the kernel will search for either.

Comment 12 Haak 2024-09-25 15:10:47 UTC

I see, thank you!
Renaming the original file and putting the new file in that location uncompressed worked - wifi is working this way.

Comment 13 Mario Limonciello (AMD) 2024-10-14 19:24:25 UTC

The fixed firmware is landed in upstream linux-firmware repo.

https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/commit/ath12k/WCN7850/hw2.0?id=34d181d737527d903ebb363de6b2e3b1dcd42683