Bug 28429 - Wifi on Intel Atom system not working with M8 of 22/2
Summary: Wifi on Intel Atom system not working with M8 of 22/2
Status: RESOLVED FIXED
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 8
Hardware: x86_64 Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Kernel and Drivers maintainers
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on: 28470
Blocks:
  Show dependency treegraph
 
Reported: 2021-02-24 13:44 CET by Herman Viaene
Modified: 2021-03-08 16:26 CET (History)
2 users (show)

See Also:
Source RPM: kernel-5.10.16-1.mga8.src.rpm
CVE:
Status comment:


Attachments
lspcidrake output from working M7 (2.00 KB, text/plain)
2021-02-24 13:45 CET, Herman Viaene
Details
lspcidrake output from failing M8 (2.09 KB, text/plain)
2021-02-24 13:46 CET, Herman Viaene
Details
messages from start till failed wifi connection attempt (90.40 KB, text/plain)
2021-02-25 10:53 CET, Herman Viaene
Details
M7 messages with successful wifi connection (88.38 KB, text/plain)
2021-02-25 11:39 CET, Herman Viaene
Details
Succesful boot with M7 kernel 5.10.16, wifi OK (79.99 KB, text/plain)
2021-02-25 17:05 CET, Herman Viaene
Details
Succesful boot with M8 kernel 5.10.18, wifi OK (82.34 KB, text/plain)
2021-02-26 14:09 CET, Herman Viaene
Details

Description Herman Viaene 2021-02-24 13:44:20 CET
Description of problem:
PeaqC1011 (UEFI) with intel Atom had its wifi working OK in M7 and all previous test ISO's of M8, including the "final" of 17/2.
Installing (not upgrading!!) the 22/2 M8 final over the 17/2 version breaks this device.
I will add the output of lspcidrake of this notebook from the M7 installation and the latest failing M8.

Version-Release number of selected component (if applicable):


How reproducible:
Allways

Steps to Reproduce:
1.
2.
3.
Comment 1 Herman Viaene 2021-02-24 13:45:58 CET
Created attachment 12379 [details]
lspcidrake output from working M7
Comment 2 Herman Viaene 2021-02-24 13:46:37 CET
Created attachment 12380 [details]
lspcidrake output from failing M8
Comment 3 Thomas Backlund 2021-02-24 13:49:18 CET
please provide dmesg and journal from boot until it has tried to connect to a wireless net

also is iwlwifi-firmware installed ?
Comment 4 Thomas Backlund 2021-02-24 13:51:07 CET
(In reply to Herman Viaene from comment #1)
> Created attachment 12379 [details]
> lspcidrake output from working M7

Was it actually working here ?
It does not show up at all
Comment 5 Herman Viaene 2021-02-24 13:57:16 CET
To comment 4: yes it works perfectly.
Comment 6 Thomas Backlund 2021-02-24 14:45:11 CET
ah, it does not need to even be an intel wifi on that rig... and by the looks of it, not connected directly to either pci or usb ...

so I really need dmesg from both mga7 and mga8

and if you install 5.10.16-1.mga7 from mga7 updates_testing... does that stop the wifi from working too ?
Aurelien Oudelet 2021-02-24 17:11:52 CET

CC: (none) => ouaurelien

Comment 7 Herman Viaene 2021-02-25 10:52:30 CET
@comment 3 and 6
Yes, iwlwifi-firmware is installed.
I am unsure what you mean by "dmesg" as this file does not exist.
Found https://wiki.mageia.org/en/Dmesgs_compared and executed command
journalctl -k --boot=0 > m8wififmsg.txt
Attaching here.
Doing same on M7 in a minute.
Comment 8 Herman Viaene 2021-02-25 10:53:30 CET
Created attachment 12387 [details]
messages from start till failed wifi connection attempt
Comment 9 Herman Viaene 2021-02-25 11:39:11 CET
Created attachment 12388 [details]
M7 messages with successful wifi connection
Comment 10 Herman Viaene 2021-02-25 11:40:04 CET
M7 journal attached. Test M7 with new kernel coming up.
Comment 11 Herman Viaene 2021-02-25 11:46:11 CET
I don't see that 5.10.16-1.mga7 kernel in mga7 updates_testing
Comment 13 Herman Viaene 2021-02-25 14:19:46 CET
I check in http://madb.mageia.org/tools/updates and just now still not visible there.
Comment 14 Thomas Backlund 2021-02-25 14:37:52 CET
yeah, I haven't assigned it to QA, so it wont show up there...

but it's in testing to be able to point users on it for tests to see if it's already fixed in newer kernels...

the next one actually ending up in QA will probably be something like 5.10.19 or so...  I'm still looking at some patches for other  issues
Comment 15 Herman Viaene 2021-02-25 15:13:52 CET
OK, downloaded from that link, installed on M7, rebooted, wifi is OK.
Comment 16 Herman Viaene 2021-02-25 15:16:44 CET
Just my 2c: I think concentrating on M7 might be a waste of time. The issue was not there in the M8 of 17/2, it appeared on the M8 of 22/2.
Comment 17 Thomas Backlund 2021-02-25 15:31:07 CET
the reason I wanted the mga7 test was to see if it was the kernel that broke...

as the 5.10.16-1.mga7 matches exactly the mga8 release kernel (just recompiled on mga7)

Interestingly I see the same WARNING: splats in both mga7 and mga8 outputs, but only mga8 fails to work

and seems to be a RTL8723BS wireless in that system
Comment 18 Thomas Backlund 2021-02-25 15:43:33 CET
can you grab the same log when booting 5.10.16-1.mga7 on mga7 so I can compare exact same kernel on both mga7 and mga8
Comment 19 Herman Viaene 2021-02-25 17:05:29 CET
Created attachment 12390 [details]
Succesful boot with M7 kernel 5.10.16, wifi OK
Comment 20 Thomas Backlund 2021-02-25 20:03:37 CET
thanks.


Can you on Mageia 8 try 5.10.18-3.mga8 from:
https://tmb.nu/Mageia/8/bugs/28429/

I have added fixes that should hopefully fix the uvcvideo splat and updates the rtl8723bs driver to 5.12-rc1 level
Comment 21 Dave Hodgins 2021-02-25 21:10:06 CET
(In reply to Herman Viaene from comment #19)
> Created attachment 12390 [details]
> Succesful boot with M7 kernel 5.10.16, wifi OK

From that log ...
Feb 25 16:14:41 mach7.hviaene.thuis kernel: rtl8723bs: acquire FW from file:rtlwifi/rtl8723bs_nic.bin
$ rpm -q -f /usr/lib/firmware/rtlwifi/rtl8723bs_nic.bin
rtlwifi-firmware-20201118-1.mga7.nonfree

The installer does not use nonfree firmware as per the release notes.

CC: (none) => davidwhodgins

Comment 22 Thomas Backlund 2021-02-26 06:23:26 CET
(In reply to Herman Viaene from comment #0)
> Description of problem:
> PeaqC1011 (UEFI) with intel Atom had its wifi working OK in M7 and all
> previous test ISO's of M8, including the "final" of 17/2.
> Installing (not upgrading!!) the 22/2 M8 final over the 17/2 version breaks
> this device.

Reading this part again...

you state "final" of 17/2 works...
and "final" of 22/2 dont work...

but between those, there were only 4 fixes.
- fixed release notes
- fixed installer left background, logo and selection text
- fixed non-english/french translations of sddm "username"
- fix pulseaudio xinit script to match newer startplasma-x11 in SESSSION


so the questions are:
- what did you install with (both times), what DE, what iso ?
- did you do anything diffently while configuring ?
Comment 23 Dave Hodgins 2021-02-26 08:09:10 CET
I strongly suspect the wifi on previous installs was configured after booting
to the installed system, instead of trying to get updates at the media
selection screen near the beginning of the install.
Comment 24 Herman Viaene 2021-02-26 11:29:27 CET
@Thomas Comment 20
Downloaded that kernel rpm, used urpmi to install it, rebooted and BINGO, I could connect wlan0.
Just I will do another boot to check it now automatically connects, at boot, I'll let you know soon.
Comment 25 Herman Viaene 2021-02-26 12:08:12 CET
Did a warm restart, and had to use MCC to get the wifi started, that gave a strange situation as the net applet in MATE and the MCC - netwerk center showed the wifi OK, but a ping to www.google.be or any connection in Firefox failed. Used MCC to change the wifi connection anyway as I want it to be (fixed IP, connecting to router as gateway, to own DNS server and provider"s DNS server and defining full FQDN), that went OK.
Made another warm restart, again net applet in MATE and the MCC - netwerk center give OK, but not really working.
Finally made a cold restart and automatically getting the wifi connection working OK and configured as I want it.
I've had this kind of situation in the past where I really needed a cold restart to get things working. This seems to be one of the quirks of this thing since I got this some months ago.
So, if you have no further questions or info you want, you may close this issue.
Tx a lot.
Comment 26 Thomas Backlund 2021-02-26 13:46:24 CET
@Herman, can you grab the log from running with 5.10.18-3.mga8 too ?

I'd like to see if I got all the WARNING splats matched or if there is still some to fix up
Comment 27 Herman Viaene 2021-02-26 14:09:09 CET
Created attachment 12399 [details]
Succesful boot with M8 kernel 5.10.18, wifi OK
Comment 28 Aurelien Oudelet 2021-02-28 21:12:45 CET
(In reply to Herman Viaene from comment #27)
> Created attachment 12399 [details]
> Succesful boot with M8 kernel 5.10.18, wifi OK

This seems to be near completion.
Final update in Bug 28471.

Assignee: bugsquad => kernel
Depends on: (none) => 28471

Aurelien Oudelet 2021-02-28 21:13:00 CET

Source RPM: (none) => kernel-5.10.16-1.mga8.src.rpm

Aurelien Oudelet 2021-03-01 14:49:05 CET

Depends on: (none) => 28470

Aurelien Oudelet 2021-03-01 14:49:24 CET

Depends on: 28471 => (none)

Comment 29 Thomas Backlund 2021-03-04 13:52:20 CET
An update for this issue has been pushed to the Mageia Updates repository.

https://advisories.mageia.org/MGASA-2021-0101.html

Resolution: (none) => FIXED
Status: NEW => RESOLVED

Comment 30 Herman Viaene 2021-03-08 09:05:43 CET
The problem appears again on kernel 5.10.20

Resolution: FIXED => (none)
Status: RESOLVED => REOPENED

Comment 31 Thomas Backlund 2021-03-08 09:10:37 CET
well, we have not changed anything regarding that hw.

does it help with a cold boot ?

meaning if you shutdown your computer completely, then wait a bit and start it again.
Comment 32 Herman Viaene 2021-03-08 09:46:33 CET
I do that all the time.
Comment 33 Herman Viaene 2021-03-08 16:26:25 CET
I cann't explain, but leaving the system alone for some minutes after starting, has the wifi coming up OK. But when trying to force it sooner, no joy. Let it be......

Status: REOPENED => RESOLVED
Resolution: (none) => FIXED


Note You need to log in before you can comment on or make changes to this bug.