On 3/17/23 13:12, Tim K wrote:
Hello again all,
I've been running a few rtw8822bu dongles in monitor mode doing a
packet capture on an STM32, running Linux 5.15. It's been fairly
stable over a few days, but I checked back in on it and found this in
the log.
Any thoughts on what could have caused this? This is from lwfinger's
GitHub repo at 4bede29 (I believe.)
[406207.728602] 8<--- cut here ---
[406207.730533] Unable to handle kernel NULL pointer dereference at
virtual address 00000004
[406207.740137] pgd = 40e434d7
[406207.741665] [00000004] *pgd=00000000
[406207.745300] Internal error: Oops: 5 [#1] PREEMPT SMP ARM
[406207.801313] CPU: 0 PID: 23145 Comm: kworker/u4:2 Tainted: G
O 5.15.24 #13
[406207.809516] Hardware name: STM32 (Device Tree Support)
[406207.814783] Workqueue: rtw88_usb: rx wq rtw_usb_rx_handler [rtw88_usb]
[406207.821397] PC is at rtw_rx_fill_rx_status+0x40/0x278 [rtw_core]
[406207.827604] LR is at rtw_rx_fill_rx_status+0x3c/0x278 [rtw_core]
[406207.833666] pc : [<bf324cac>] lr : [<bf324ca8>] psr: 400f0113
[406207.840043] sp : c5379e20 ip : c5379eb8 fp : 00000001
...
[406208.084918] [<bf324cac>] (rtw_rx_fill_rx_status [rtw_core]) from
[<bf3c6b44>] (rtw8822b_query_rx_desc+0x10c/0x19c [rtw_8822b])
[406208.096524] [<bf3c6b44>] (rtw8822b_query_rx_desc [rtw_8822b]) from
[<bf416920>] (rtw_usb_rx_handler+0x6c/0x16c [rtw88_usb])
[406208.107820] [<bf416920>] (rtw_usb_rx_handler [rtw88_usb]) from
[<c01455cc>] (process_one_work+0x1dc/0x588)
[406208.117570] [<c01455cc>] (process_one_work) from [<c0145e04>]
(worker_thread+0x5c/0x568)
[406208.125790] [<c0145e04>] (worker_thread) from [<c014c230>]
(kthread+0x144/0x160)
[406208.133310] [<c014c230>] (kthread) from [<c0100130>]
(ret_from_fork+0x14/0x24)
...
[406208.186520] ---[ end trace 140c234232a74c11 ]---
Tim,
I can tell from your dump that you are not running x86_64 hardware. Is it ARM?
In any case, please run the command 'gdb rtw_core.ko' while in the rtw88
directory, and once gdb gets started, enter the command
l *rtw_rx_fill_rx_status+0x40
Report the output. That will show the line of code that is failing. We should be
able to figure it out from there.
Larry
