Re: rtw89_8852bte (RTL8852BE-VT): recurring SER 0x1000/0x4000/0x1001/0x1002 on fw 0.29.122.1/.2; question on rtw8852bt_fw-1.bin

[Bitterblue Smith <rtl8821cerfe2@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 3197 bytes --]

On 02/11/2026 03:50, Ping-Ke Shih wrote:
> Chris Vitalos <cvitalos@gmail.com> wrote:
>> Hello Ping-Ke, all,
>>
>> This is a follow-up to the July thread "[BUG] rtw89_8852bte: Recurring
>> firmware crashes (LBC watchdog / SER) on RTL8852BE-VT despite ASPM and
>> PS disabled". I see the same signature on a different laptop and
>> distro, and I have a question about the newer firmware format.
> 
> I sent attached debug patch to the people who created the thread, but I have
> not received log.
> 
> To collect more detail info, please apply attached patch and run test. 
> When SER occurs, firmware will lock itself and driver dump a lot of
> info for debugging.
> 
> Since we want to collect three logs at least, please reboot before collecting
> 2nd and 3th logs.
> 
>> Firmware tested
>> - 0.29.122.1 (2ac94d05), as shipped: SER burst at 2150 s uptime.
>> - 0.29.122.2 (42816c59), upstream linux-firmware, installed via
>> /lib/firmware/updates: two SER bursts on one boot, at 18:39:25 and
>> 19:44:31 local (boot was 17:39:31).
> 
> It means no obvious difference from the two firmware, right?
> 
>>
>> Symptoms
>> Every burst has the same order: SER catches error 0x1000, "firmware
>> failed to ack for leaving ps mode", 0x4000, 0x1001, 0x1002, with many
>> "[ERR]fw PC = 0xb898..." and "0xb892..." lines in between. On the
>> 7.0.0-42 boot, disable_ps_mode is N and iw reports power save on
>> (defaults). 
> 
> Please turn off power save entirely by NetworkManager or iw command
>    iw wlan0 set power_save off
> 
>> I have not yet tested the rtw89_pci
>> disable_clkreq/disable_aspm_l1/disable_aspm_l1ss options or
>> disable_ps_mode on this unit; I understand the July reporter saw
>> crashes with them set.
> 
> It is worth to give it a try.
> 
>>
>> Impact
>> - First event: connectivity (ping, DNS) failed until I restarted the
>> Wi-Fi interface.
>> - 18:39:25 event: Firefox browser VPN reported it could not reach its
>> servers; I did not need to restart Wi-Fi.
>> - wlo1 stayed associated: after the 17:56:52 resume association, the
>> kernel log shows no deauth/auth/assoc lines through either burst.
>> NetworkManager/wpa_supplicant logged only a DHCPv6 lease line at
>> 18:40:01 between 18:38 and 18:42.
>> - I do not have a measured outage duration.
> 
> Can you try another AP to see if anything different? 
> 
>>
>> Question
>> Ubuntu's linux-firmware package ships both rtw8852bt_fw.bin and
>> rtw8852bt_fw-1.bin, but the 7.0 driver declares and loads only
>> rtw89/rtw8852bt_fw.bin (modinfo rtw89_8852bt shows a single firmware
>> entry). Which kernel adds support for the -1 format, and does that
>> firmware change the power-save-exit behavior behind SER 0x1000 with
>> "firmware failed to ack for leaving ps mode"? If it might help, I can
>> test a kernel that supports it.
> 
> As I know, the main purpose of -1 format is the changes of BT-coexistence, 
> so it might not help.
> 
> Ping-Ke

Someone else ran into the same problem (I think) and captured four
crashes with your patch. dmesg output is attached. They used this
commit, which is pretty much the same as current rtw-next:

https://github.com/morrownr/rtw89/commit/8019af4b9ed57a9042e139432b88687770f155b5

[-- Attachment #2: rtw89-debug.7z --]
[-- Type: application/x-7z-compressed, Size: 3647172 bytes --]