[Bug 1221734] New: RTL8723BE cannot initialize mac after suspend or loss of AP
https://bugzilla.suse.com/show_bug.cgi?id=1221734 Bug ID: 1221734 Summary: RTL8723BE cannot initialize mac after suspend or loss of AP Classification: openSUSE Product: openSUSE Distribution Version: Leap 15.5 Hardware: x86-64 OS: openSUSE Leap 15.5 Status: NEW Severity: Normal Priority: P5 - None Component: Kernel:Drivers Assignee: kernel-bugs@suse.de Reporter: owinebar@gmail.com QA Contact: qa-bugs@suse.de Target Milestone: --- Found By: --- Blocker: --- I am using Leap 15.5 on an old ASUS laptop, model R541U. The wifi was working fine when I initially installed Leap 15.4 as the sole OS around April 2023. The wifi started acting flaky in September 2023, but I didn't pay much attention because it has been working. I finally took the time to try to get it working again. I set the power save parameters off, the debug_mask to 0xFFFFFFFFFFFFFFFF, and the debug_level to 5. I also set NetworkManager to use a permanent MAC address for wlan0 because there the error seemed to be associated with a message "Init MAC" failed. The device still fails in 2 situations, and reports the "Init MAC" error. The first is after suspending the laptop. Presumably the power save from the suspend is not dependent on the driver power save parameters, so it attempts and fails to reinitialize the wlan0 device. With debug level at 5, I see rtlwifi report "init MAC Fail as power on failure", the the "Init MAC failed" message from rtl8723be, then the message "Failed to config hardware!" from rtlwifi. The other failure arises when the connection to the AP is lost for 10s (at least that's what it reports) after rebooting with no intervening suspend. In the latter case, there are a lot of messages "Waiting for FW read clear HMEBox(0)!!! 0x130 = ff". The "Init MAC failed" message first appears in my messages file dated 2023-09-11. The previous rolled messages file is dated 2023-09-08. Looking at the zypp/history log, I evidently updated the kernel from 5.14.21-150400.24.74 to 5.14.21-150400.24.81, somewhere around August 31 or September 6 of 2023. I only updated the distribution on the laptop to 15.5 recently. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1221734 https://bugzilla.suse.com/show_bug.cgi?id=1221734#c1 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |owinebar@gmail.com, | |tiwai@suse.com Flags| |needinfo?(owinebar@gmail.co | |m) --- Comment #1 from Takashi Iwai <tiwai@suse.com> --- You should be still able to use the old Leap 15.4 kernel on your Leap 15.5 system. Could you try to install the old Leap 15.4 kernel and identify which kernel version started showing the problem? The Leap 15.4 (update) packages should be available at http://download.opensuse.org/update/leap/15.4/sle/ And, if you figure out the working / non-working kernel versions, please give the dmesg outputs from both (with debugging options) to compare. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1221734 https://bugzilla.suse.com/show_bug.cgi?id=1221734#c2 --- Comment #2 from Lynn Winebarger <owinebar@gmail.com> --- Created attachment 873704 --> https://bugzilla.suse.com/attachment.cgi?id=873704&action=edit /var/log/messages with debug_level=5 for kernel 5.14.21.69 /var/log/messages with last kernel not exhibiting an rtl8723 "Init MAC failed" message after invoking "systemctl shutdown" and then waking the laptop up. Note left module parameters at defaults, other than debug_level and debug_mask. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1221734 https://bugzilla.suse.com/show_bug.cgi?id=1221734#c3 Lynn Winebarger <owinebar@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(owinebar@gmail.co | |m) | --- Comment #3 from Lynn Winebarger <owinebar@gmail.com> --- Created attachment 873705 --> https://bugzilla.suse.com/attachment.cgi?id=873705&action=edit /var/log/messages with first kenerl version exhibiting rtl8723be fail after waking from suspend, 5.14.21-150400-24.74 /var/log/messages with last kernel not exhibiting an rtl8723 "Init MAC failed" message after invoking "systemctl shutdown" and then waking the laptop up. Note left module parameters at defaults, other than debug_level and debug_mask. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1221734 https://bugzilla.suse.com/show_bug.cgi?id=1221734#c4 --- Comment #4 from Lynn Winebarger <owinebar@gmail.com> --- (In reply to Takashi Iwai from comment #1)
You should be still able to use the old Leap 15.4 kernel on your Leap 15.5 system. Could you try to install the old Leap 15.4 kernel and identify which kernel version started showing the problem? The Leap 15.4 (update) packages should be available at http://download.opensuse.org/update/leap/15.4/sle/
And, if you figure out the working / non-working kernel versions, please give the dmesg outputs from both (with debugging options) to compare.
Ok, I have identified the versions. I do not get a rtl8723 failure after waking from suspend with kernel 5.14.21-150400-24.69, but do with 5.14.21-150400-24.74. I left the rtl8723be parameters unspecified except for debug_mask and debug_level (set to 5). Boot parameters from /var/log/boot.msg: Command line: BOOT_IMAGE=/boot/vmlinuz-5.14.21-150400.24.74-default root=/dev/mapper/system-root splash=silent pci=noaer preempt=full mitigations=auto quiet security=apparmor intremap=no_x2apic_optout -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1221734 https://bugzilla.suse.com/show_bug.cgi?id=1221734#c5 --- Comment #5 from Takashi Iwai <tiwai@suse.com> --- Thanks! Interestingly, through a quick glance, there is no obvious changes in the relevant code (either in Realtek WiFi driver or wireless net code). Is the problem reproducible reliably, and can switching the kernel between those versions confirm the first bad kernel? If yes, you can try the following: - Rename the first broken kernel modules: cd /lib/modules/5.14.21-150400.24.74-default/ mv kernel kernel.dist - Copy the last good working kernel modules to there: cp -a /lib/modules/5.14.21-150400.24.69-default/kernel /lib/modules/5.14.21-150400.24.74-default/ depmod -a 5.14.21-150400.24.74-default dracut -f --kver 5.14.21-150400.24.74-default Then retest 5.14.21-150400.24.74-default. There can be some hiccups, but the only interesting thing is the test of this bug. If the bug still appears after this, it means that the code built in the kernel got some bugs. OTOH, if this "fixes" the issue on *-74, it means some of the copied modules received the fix. Then you can go back to the original modules, and try to copy some of them, and identify which module are affected. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1221734 https://bugzilla.suse.com/show_bug.cgi?id=1221734#c6 --- Comment #6 from Lynn Winebarger <owinebar@gmail.com> --- Initially the downgrade to kernel 5.14.21-150400.24.69 appeared to resolve the issue, at least with regard to the "Init MAC failed" error. However, now I'm getting error reports from the kernel without even suspending. It may well be a hardware failure, given the age of the laptop and that I had to partially disassemble it to replace the battery last year. So maybe this should be closed. Thanks for looking into it. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1221734 https://bugzilla.suse.com/show_bug.cgi?id=1221734#c7 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |INVALID --- Comment #7 from Takashi Iwai <tiwai@suse.com> --- OK, then let's close. If any new fact appears, feel free to reopen. Thanks. -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠suse.com