WiFi hangs periodically #226
Comments
|
One problem with HP laptops is where the BIOS handling of PCI devices is wrong. To test if this is your problem, create (as root) a file named /usr/lib/modprobe.d/70-rtw89.conf. To this file, add the lines options rtw89_pci disable_aspm_l1-y disable_aspm_l1ss=y The first line handles the in-kernel version, and the second handles the driver from this repo. After this file is created, run the sequence If you have done it correctly, the second command should produce output like The critical parts are the disable stuff at the end of the first line. |
|
Should be: Thanks for the feedback. I have applied your advices. I will return in couple of days with the result. By the way - after modules reload was done WiFi failed to reconnect to my access point though it could scan wireless networks. But after reboot it managed to connect. There was no relevant information in dmesg. May be this is an issue of NetworkManager and friends. |
|
@lwfinger thanks for your advise. WiFi connection became stable, all the "timed out to flush" messages have gone. But the following errors remain: They happen from time to time 10-20 times a day. It seems that they don't affect WiFi stability but if you can tell me the way to avoid them I will appreciate it, since I need to look into logs quite often and these errors generate a lot noise for me. |
|
You are getting unexpected interrupts. Edit the options lines we created earlier, and add 'disable_clkreq=y' to each of those lines. I do not know if it will help or hurt, but it is our last option available. |
|
FYI, I have the same problem with my HP Pro Mini 260 G9. (PCI device ID |
|
@leiless: Have you applied the module parameters described in the March 14 postings? (I wish that GitHub would number the comments.) |
I did the same config as of now, and I need some time to check if it'll hang periodically. |
|
Those two lines with ...l1-y were a typo as shown in the comment following the one that I wrote. Each line should have ....l1=y instead. |
|
Okay, conf updated! $ cat /usr/lib/modprobe.d/70-rtw89.conf
options rtw89_pci disable_aspm_l1=y disable_aspm_l1ss=y disable_clkreq=y
options rtw89pci disable_aspm_l1=y disable_aspm_l1ss=y disable_clkreq=y
$ sudo modprobe -rv rtw_8852be
rmmod rtw_8852be
rmmod rtw89pci
rmmod rtw_8852b
rmmod rtw89core
rmmod mac80211
rmmod libarc4
rmmod cfg80211
$ sudo modprobe -v rtw_8852be
insmod /lib/modules/5.19.0-38-generic/kernel/lib/crypto/libarc4.ko
insmod /lib/modules/5.19.0-38-generic/kernel/net/wireless/cfg80211.ko
insmod /lib/modules/5.19.0-38-generic/kernel/net/mac80211/mac80211.ko
insmod /lib/modules/5.19.0-38-generic/kernel/drivers/net/wireless/realtek/rtw89/rtw89core.ko
insmod /lib/modules/5.19.0-38-generic/kernel/drivers/net/wireless/realtek/rtw89/rtw89pci.ko disable_aspm_l1=y disable_aspm_l1ss=y disable_clkreq=y
insmod /lib/modules/5.19.0-38-generic/kernel/drivers/net/wireless/realtek/rtw89/rtw_8852b.ko
insmod /lib/modules/5.19.0-38-generic/kernel/drivers/net/wireless/realtek/rtw89/rtw_8852be.ko |
|
Hey! It is been a while since I opened this issue. I've applied disable_clkreq=y to the config but it seems that nothing is changed. Still I get this logs in dmesg. Though the FW status has changed to an another value |
|
Did you add the other two options? Please do the following: sudo modprobe -rv rtw_8852be Post the output of all 3 commands. |
Yep! |
|
I also have the same problem with a hp 645 g9 (8852be) You can find attached a full dmesg dump. I also just tried disabling a pcie power saving feature in the BIOS. I'll post here again if it works. |
|
That was NOT a complete dmesg - you selected only a portion. What I did see is that you seem to be missing the firmware. Please run the following command 'md5sum /lib/firmware/rtw89' and post the output. |
|
Sorry about the dmesg, I thought everything was there ...
But disabling the pcie power saving feature in the BIOS seems to have worked. My wifi hasn't dropped once since I applied it (approx 3h usage). (It used to drop after 15 minutes in average). |
@lwfinger, after I updated the config, the Wi-Fi could still hang sometimes. Logs: https://pastebin.com/ehZHyNTF The 8852BE driver was built upon the commit $ cat /usr/lib/modprobe.d/70-rtw89.conf
options rtw89_pci disable_aspm_l1=y disable_aspm_l1ss=y disable_clkreq=y
options rtw89pci disable_aspm_l1=y disable_aspm_l1ss=y disable_clkreq=y
$ lscpu | grep -w Model
Model name: 12th Gen Intel(R) Core(TM) i5-1235U
$ md5sum /lib/firmware/rtw89/rt*
49951f548b87ea0258d128195ef6e0cf /lib/firmware/rtw89/rtl8852au_config.bin
2dc64f98c75ec454a1b18962eefe7b2f /lib/firmware/rtw89/rtl8852au_fw.bin
6edd77ccced6eda33850d302853fe665 /lib/firmware/rtw89/rtw8852a_fw.bin
a3317e9646c8e669c7c2d1fc19a6109d /lib/firmware/rtw89/rtw8852b_fw.bin
ace8c45cf09d44fe99e2480a94abc2ff /lib/firmware/rtw89/rtw8852c_fw.bin |
|
The firmware you have for rtw8852a and rtw8852b are the latest,but you are missing rtw8852b_fw_1.bin. That is a later version with some improvements. Your firmware for rtw8852c is not the latest, but that should not make any difference. My md5sum values are as follows: 6edd77ccced6eda33850d302853fe665 rtw89/rtw8852a_fw.bin |
|
I just pushed commit 6169c9b. I contains all the changes that will be in kernel 6.4. |
Can I just download the latest firmware from https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/rtw89 and replace those Will it work? |
|
Yes. That is where I get mine. |
|
@lwfinger, I updated the firmware. BTW, still using the driver built from commit See if it'll still hang. $ md5sum /lib/firmware/rtw89/*
49951f548b87ea0258d128195ef6e0cf /lib/firmware/rtw89/rtl8852au_config.bin
2dc64f98c75ec454a1b18962eefe7b2f /lib/firmware/rtw89/rtl8852au_fw.bin
6edd77ccced6eda33850d302853fe665 /lib/firmware/rtw89/rtw8852a_fw.bin
97cc27ddbaacfcfd9f4b4ad425d114d3 /lib/firmware/rtw89/rtw8852b_fw-1.bin
a3317e9646c8e669c7c2d1fc19a6109d /lib/firmware/rtw89/rtw8852b_fw.bin
ceed5cd7b2066d7a8b99e73e9056d3c1 /lib/firmware/rtw89/rtw8852c_fw.bin |
|
I contacted Realtek and was pointed to a patch that has been submitted but not yet merged. It was incorporated into this repo and a new version was pushed. Please pull and test. |
|
I pulled the changes, reinstalled, removed the BIOS changes I did before (now running again with enabled pcie power savings) and it seems to work fine. No drop in 2-3 hours. I also never thanked you for maintaining this driver, so thank you very much 😃 🎉 !!!! |
|
Glad to hear. I am going to close the issue. If there are any changes, re-open it. |
…l page Though SER can recover this case, traffic can get stuck for a while. Fix it by adjusting page quota to avoid hardware access null page of CMAC/DMAC. Fixes: a1cb097 ("wifi: rtw89: 8852b: configure DLE mem") Fixes: 3e870b4 ("wifi: rtw89: 8852b: add HFC quota arrays") Cc: stable@vger.kernel.org Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Link: lwfinger/rtw89#226 (comment) Link: lwfinger/rtw89#240 Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
…l page [ Upstream commit c0426c4 ] Though SER can recover this case, traffic can get stuck for a while. Fix it by adjusting page quota to avoid hardware access null page of CMAC/DMAC. Fixes: a1cb097 ("wifi: rtw89: 8852b: configure DLE mem") Fixes: 3e870b4 ("wifi: rtw89: 8852b: add HFC quota arrays") Cc: stable@vger.kernel.org Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Link: lwfinger/rtw89#226 (comment) Link: lwfinger/rtw89#240 Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20230426034737.24870-1-pkshih@realtek.com Signed-off-by: Sasha Levin <sashal@kernel.org>
…l page BugLink: https://bugs.launchpad.net/bugs/2028979 [ Upstream commit c0426c446d92023d344131d01d929bc25db7a24e ] Though SER can recover this case, traffic can get stuck for a while. Fix it by adjusting page quota to avoid hardware access null page of CMAC/DMAC. Fixes: a1cb097 ("wifi: rtw89: 8852b: configure DLE mem") Fixes: 3e870b4 ("wifi: rtw89: 8852b: add HFC quota arrays") Cc: stable@vger.kernel.org Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Link: lwfinger/rtw89#226 (comment) Link: lwfinger/rtw89#240 Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20230426034737.24870-1-pkshih@realtek.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
…l page BugLink: https://bugs.launchpad.net/bugs/2028979 [ Upstream commit c0426c446d92023d344131d01d929bc25db7a24e ] Though SER can recover this case, traffic can get stuck for a while. Fix it by adjusting page quota to avoid hardware access null page of CMAC/DMAC. Fixes: a1cb097 ("wifi: rtw89: 8852b: configure DLE mem") Fixes: 3e870b4 ("wifi: rtw89: 8852b: add HFC quota arrays") Cc: stable@vger.kernel.org Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Link: lwfinger/rtw89#226 (comment) Link: lwfinger/rtw89#240 Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20230426034737.24870-1-pkshih@realtek.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
…l page BugLink: https://bugs.launchpad.net/bugs/2028979 [ Upstream commit c0426c446d92023d344131d01d929bc25db7a24e ] Though SER can recover this case, traffic can get stuck for a while. Fix it by adjusting page quota to avoid hardware access null page of CMAC/DMAC. Fixes: a1cb097 ("wifi: rtw89: 8852b: configure DLE mem") Fixes: 3e870b4 ("wifi: rtw89: 8852b: add HFC quota arrays") Cc: stable@vger.kernel.org Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Link: lwfinger/rtw89#226 (comment) Link: lwfinger/rtw89#240 Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20230426034737.24870-1-pkshih@realtek.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
…l page
Though SER can recover this case, traffic can get stuck for a while. Fix it
by adjusting page quota to avoid hardware access null page of CMAC/DMAC.
Fixes: a1cb097168fa ("wifi: rtw89: 8852b: configure DLE mem")
Fixes: 3e870b481733 ("wifi: rtw89: 8852b: add HFC quota arrays")
Cc: stable@vger.kernel.org
Tested-by: Larry Finger <Larry.Finger@lwfinger.net>
Link: #226 (comment)
Link: #240
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20230426034737.24870-1-pkshih@realtek.com
Hi, I have HP ProBook 450 G9 running Linux kernel 6.1 with the latest rtw89 driver from this repo and firmware from linux-firmware repository. Approx. twice a day wifi on laptop hangs. I mean by that that it doesn't show up disconnect but there is no traffic between laptop and access point. This can be fixed if I tell NetworkManager to connect to another BSS (or disconnect from current one). After some timeout disconnect happens and then I can successfuly reconnect to my access point.
PCI device ID 10ec:b852
When this happens I see following in dmesg:
The text was updated successfully, but these errors were encountered: