Created attachment 258400 [details] panic screenshot Kernal panic after few packets over WiFi root@minipc:~ # pciconf -lv rtw890@pci0:3:0:0 rtw890@pci0:3:0:0: class=0x028000 rev=0x00 hdr=0x00 vendor=0x10ec device=0xb852 subvendor=0x10ec subdevice=0xb852 vendor = 'Realtek Semiconductor Co., Ltd.' device = 'RTL8852BE PCIe 802.11ax Wireless Network Controller' class = network root@minipc:~ # freebsd-version -uk 14.2-RELEASE-p1 14.2-RELEASE-p2 root@minipc:~ # pkg info | grep wifi wifi-firmware-rtw89-kmod-20241017.1402000_1 Firmware modules for the rtw89 (rtw89) WiFi NIC driver wifi-firmware-rtw89-kmod-rtw8852b-20241017.1401000_1 Firmware modules for the rtw89 (rtw8852b) WiFi NIC driver /etc/rc.conf # wifi wlans_rtw890="wlan0" ifconfig_wlan0="WPA DHCP SYNC" create_args_wlan0="country France" /boot/loader.conf compat.linuxkpi.skb.mem_limit=1 root@minipc:~ # kldstat Id Refs Address Size Name 1 74 0xffffffff80200000 1f3c6c0 kernel 2 1 0xffffffff8213d000 36c8 coretemp.ko 3 1 0xffffffff82141000 77d8 cryptodev.ko 4 1 0xffffffff82149000 5da658 zfs.ko 5 1 0xffffffff82725000 21338 geom_mirror.ko 6 1 0xffffffff831e5000 3390 acpi_wmi.ko 7 1 0xffffffff831e9000 4250 ichsmb.ko 8 1 0xffffffff831ee000 2178 smbus.ko 9 1 0xffffffff83200000 24fee0 if_rtw89.ko 10 1 0xffffffff831f1000 3360 uhid.ko 11 1 0xffffffff831f5000 3360 wmt.ko 12 1 0xffffffff831f9000 4d20 ng_ubt.ko 13 6 0xffffffff83450000 abb8 netgraph.ko 14 2 0xffffffff8345b000 a250 ng_hci.ko 15 4 0xffffffff83466000 2670 ng_bluetooth.ko 16 1 0xffffffff83469000 e5b0 snd_uaudio.ko 17 1 0xffffffff83478000 33c0 uchcom.ko 18 1 0xffffffff8347c000 4e10 ucom.ko 19 1 0xffffffff83481000 e268 ng_l2cap.ko 20 1 0xffffffff83490000 1bf68 ng_btsocket.ko 21 1 0xffffffff834ac000 38f8 ng_socket.ko 22 1 0xffffffff834b0000 2a68 mac_ntpd.ko root@minipc:~ # dmesg | grep rtw rtw890: <rtw89_8852be> port 0x4000-0x40ff mem 0x80500000-0x805fffff at device 0.0 on pci3 rtw890: successfully loaded firmware image 'rtw89/rtw8852b_fw-1.bin' rtw890: loaded firmware rtw89/rtw8852b_fw-1.bin rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 5 rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 3 rtw890: MAC has already powered on rtw890: chip rfe_type is 5 rtw890: <rtw89_8852be> port 0x4000-0x40ff mem 0x80500000-0x805fffff at device 0.0 on pci3 rtw890: successfully loaded firmware image 'rtw89/rtw8852b_fw-1.bin' rtw890: loaded firmware rtw89/rtw8852b_fw-1.bin rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 5 rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 3 rtw890: MAC has already powered on rtw890: chip rfe_type is 5 current process = 0 (rtw890) rtw890: <rtw89_8852be> port 0x4000-0x40ff mem 0x80500000-0x805fffff at device 0.0 on pci3 rtw890: successfully loaded firmware image 'rtw89/rtw8852b_fw-1.bin' rtw890: loaded firmware rtw89/rtw8852b_fw-1.bin rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 5 rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 3 rtw890: MAC has already powered on rtw890: chip rfe_type is 5 current process = 0 (rtw89_tx_wq_3) #6 0xffffffff83204e1f at rtw89_core_txq_work+0x14f rtw890: <rtw89_8852be> port 0x4000-0x40ff mem 0x80500000-0x805fffff at device 0.0 on pci3 rtw890: successfully loaded firmware image 'rtw89/rtw8852b_fw-1.bin' rtw890: loaded firmware rtw89/rtw8852b_fw-1.bin rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 5 rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 3 rtw890: MAC has already powered on rtw890: chip rfe_type is 5 current process = 0 (rtw89_tx_wq_3) #6 0xffffffff83204e1f at rtw89_core_txq_work+0x14f rtw890: <rtw89_8852be> port 0x4000-0x40ff mem 0x80500000-0x805fffff at device 0.0 on pci3 rtw890: successfully loaded firmware image 'rtw89/rtw8852b_fw-1.bin' rtw890: loaded firmware rtw89/rtw8852b_fw-1.bin rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 5 rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 3 rtw890: MAC has already powered on rtw890: chip rfe_type is 5 Server side on LAN (iperf3 -s -p 9000) [ 5] local 192.168.78.254 port 9000 connected to 192.168.78.154 port 13909 [ ID] Interval Transfer Bitrate Retr Cwnd [ 5] 0.00-1.00 sec 3.00 MBytes 25.1 Mbits/sec 583 419 KBytes [ 5] 1.00-2.03 sec 0.00 Bytes 0.00 bits/sec 8 744 KBytes [ 5] 2.03-3.06 sec 2.38 MBytes 19.5 Mbits/sec 15 4.53 KBytes [ 5] 3.06-4.00 sec 1.75 MBytes 15.5 Mbits/sec 0 6.97 KBytes [ 5] 4.00-5.03 sec 2.50 MBytes 20.4 Mbits/sec 0 9.61 KBytes [ 5] 5.03-6.00 sec 128 KBytes 1.07 Mbits/sec 3 1.41 KBytes [ 5] 6.00-7.00 sec 0.00 Bytes 0.00 bits/sec 1 1.41 KBytes Client side panic after 6 seconds (iperf3 -c server -w 1024K -p 9000 -R --get-server-output) screen shot in attachment. JC
(In reply to Jean-Claude MICHOT from comment #0) > wifi-firmware-rtw89-kmod-rtw8852b-20241017.1401000_1 Looks like your kmod is built for 14.1. You need to build it from ports or reinstall from kmod_latest_2 repo.
From ports (head): same PANIC with wifi-firmware-rtw89-kmod-rtw8852b-20241017.1402000_2 after add from kmod_latest_2 repo seem more reliable, the first start of iperf is fine, with 22 Mbits/sec but i restart a 2nd test with iperf and get same PANIC.
fixed with 14.3 upgrade. but performance are low (23 Mbits/s) with 5 Ghz AP @ 5 meters.
(In reply to Jean-Claude MICHOT from comment #3) For me, things have not changed on 14.3 - I still see quite stable 22 MB/s, maybe a bit longer than on 14.2, but at some moment it panics. I have not collected the stack trace yet, as it usually happens when I am in a graphics console.
(In reply to Phil Krylov from comment #4) Can you enable crash dumps? Might have to compile a debug kernel. Then you may be able to type 'dump' into the frozen screen and still get a crashdump. Are you also on an 8852BE?
(In reply to Phil Krylov from comment #4) That seems "normal" to be stuck at 22 Mbit/s with the rtw8852b on FreeBSD 14.3. root@minipc:~ # ifconfig wlan0 mode 11ac ifconfig: SIOCSIFMEDIA (media): Device not configured I haven't found any way to switch to 802.11ac mode. root@minipc:~ # ifconfig wlan0 list caps drivercaps=581c001<STA,SHSLOT,SHPREAMBLE,MONITOR,WPA1,WPA2,WME> cryptocaps=b<WEP,TKIP,AES_CCM> No HTCAP (High Throughput) or VHTCAP (Very High Throughput) capabilities are listed. Only basic 802.11a/b/g modes are supported - no MIMO (40 MHz), no VHT (80 MHz). I'm stuck with a stable but limited Wi-Fi throughput of around 22 Mbit/s. The hardware is capable of much more, but the current driver doesn't seem to support it.
(In reply to Bjoern A. Zeeb from comment #5) Yes rtw89_8852be rtw890: <rtw89_8852be> port 0x4000-0x40ff mem 0x80500000-0x805fffff at device 0.0 on pci3 rtw890: successfully loaded firmware image 'rtw89/rtw8852b_fw-1.bin' rtw890: loaded firmware rtw89/rtw8852b_fw-1.bin
(In reply to Jean-Claude MICHOT from comment #7) For you I know form #1, that question was for Phil.
(In reply to Jean-Claude MICHOT from comment #7) I can tell you the switch but it will only buy you unhappiness. rtw88 and rtw89 need extra code to be done. Sorry, the much I'd love to my time needs to be devoted elsewhere currently.
Created attachment 261249 [details] 14.3-RELEASE debug kernel If a PANIC with 14.3-RELEASE with kernel debug could be usefull.
re-open. Thanks a lot! Helps a lot!
Created attachment 261259 [details] core.txt from the panic
(In reply to Bjoern A. Zeeb from comment #5) Yes, it's an RTL8852BE in a Lenovo E14 G6 (AMD) laptop. I have built a debug kernel, achieved getting crash dumps from it, but somehow this crash is much harder to reproduce on the debug kernel. So I rebooted to the packaged 14.3-RELEASE kernel, and tried to reproduce it again: # speedtest-go -t 20 It runs great on the RX part of the test, but then on running TX part is panics, like, every second time. And, luckily, the panics on the 14.3-RELEASE kernel now produce crash dumps. Attaching core.txt.
(In reply to Jean-Claude MICHOT from comment #6) > The hardware is capable of much more, but the current driver doesn't seem to support it. On Linux, though, the same driver gives me 83 Mbps on 2.4GHz and 569 Mbps on 5GHz and is very stable. So hopefully fixing the FreeBSD side will be sufficient.
(In reply to Phil Krylov from comment #14) it's not just fixing for stabily; there are compat functions that need implementing for rtw88 and rtw89 to be able to do hw crypto offload, 11n and 11ac. But we'll get there. It seems it is the same panic you both are seeing. So thank you for the extra information from the crash dump. I'll try to have a look this week. Once things are stable the other bits are simply sitting down and filling in the gaps. So getting it stable would be really good.
After asking for help on the discord I've been recommended to share my information about the crash here. I receive the kernel panic whenever I run `ifconfig wlan0 up`. The wireless networking had no issues when I ran Fedora Silverblue if that is relevant. At one point I was able to scan for networks around me and I successfully connected to my home network once, but received the kernel panic quickly after. After that initial crash, this happens any time I try to start the wireless card. I have only attempted FreeBSD 15.0 RELEASE, and have no experience with previous releases. I am also on a Lenovo ThinkPad E14 Gen 6 with an AMD processor. I'm not sure what information you need about my system so I tried to be fairly exhaustive and tried to include any potentially related information. Let me know if there is anything else that can help. # uname -v FreeBSD 15.0-RELEASE releng/15.0-n280995-7aedc8de6446 MYKERNEL # ifconfig wlan0 wlan0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 options=0 ether // Removed groups: wlan ssid "" channel 1 (2412 MHz 11b) regdomain FCC country US authmode OPEN privacy OFF txpower 30 bmiss 7 scanvalid 60 wme bintval 0 parent interface: rtw890 media: IEEE 802.11 Wireless Ethernet autoselect (autoselect) status: no carrier nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL> # pkg info wifi-firmware-rtw89-kmod-rtw8852b wifi-firmware-rtw89-kmod-rtw8852b-20241017.1500068_2 Name : wifi-firmware-rtw89-kmod-rtw8852b Version : 20241017.1500068_2 Installed on : Fri Dec 12 13:02:27 2025 EST Origin : net/wifi-firmware-rtw89-kmod Architecture : FreeBSD:15:amd64 Prefix : /usr/local Categories : kld net Licenses : primary, whence Maintainer : bz@FreeBSD.org WWW : https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/rtw89 Comment : Firmware modules for the rtw89 (rtw8852b) WiFi NIC driver Annotations : FreeBSD_version: 1500068 flavor : rtw8852b repo_type : binary repository : FreeBSD-ports-kmods Flat size : 1.19MiB Description : Firmware files for Realtek WiFi NICs supported by the rtw89 driver. Currently 11ax supported chipsets: - RTL8851B - RTL8852A - RTL8852B - RTL8852C Currently 11be supported chipsets: - RTL8922A # cat /etc/rc.conf hostname="fbsdthinkpad" ifconfig_ue0="DHCP" ifconfig_ue0_ipv6="inet6 accept_rtadv" sshd_enable="YES" ntpd_enable="YES" ntpd_sync_on_start="YES" powerd_enable="YES" moused_nondefault_enable="NO" # Set dumpdev to "AUTO" to enable crash dumps, "NO" to disable dumpdev="AUTO" zfs_enable="YES" wlans_rtw890="wlan0" create_args_wlan0="country US regdomain fcc" dumpdir="/var/crash" jail_enable="YES" # cat /boot/loader.conf aesni_load="YES" geom_eli_load="YES" kern.geom.label.disk_ident.enable="0" kern.geom.label.gptid.enable="0" zfs_load="YES" # kldstat Id Refs Address Size Name 1 80 0xffffffff80200000 1ef7dc8 kernel 2 1 0xffffffff820f8000 1c270 geom_eli.ko 3 1 0xffffffff82117000 619c40 zfs.ko 4 1 0xffffffff838da000 3390 acpi_wmi.ko 5 1 0xffffffff83a00000 2fd4a8 if_rtw89.ko 6 1 0xffffffff838de000 6210 ig4.ko 7 1 0xffffffff838e5000 3220 intpm.ko 8 1 0xffffffff838e9000 2178 smbus.ko 9 1 0xffffffff838ec000 58c0 ng_ubt.ko 10 4 0xffffffff838f2000 abb8 netgraph.ko 11 3 0xffffffff838fd000 a330 ng_hci.ko 12 2 0xffffffff83908000 2670 ng_bluetooth.ko 13 1 0xffffffff8390b000 3218 iichid.ko 14 1 0xffffffff8390f000 21e8 hms.ko 15 1 0xffffffff83912000 30a8 hidmap.ko 16 1 0xffffffff83916000 3355 hmt.ko 17 1 0xffffffff8391a000 22cc hconf.ko 18 1 0xffffffff8391d000 2a80 mac_ntpd.ko 19 1 0xffffffff83920000 2630 nlsysevent.ko # dmesg | grep rtw rtw890: <rtw89_8852be> port 0x2000-0x20ff mem 0x80800000-0x808fffff at device 0.0 on pci2 rtw890: successfully loaded firmware image 'rtw89/rtw8852b_fw-1.bin' rtw890: loaded firmware rtw89/rtw8852b_fw-1.bin rtw890: MAC has already powered on rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 5 rtw890: Firmware version 0.29.29.5 (da87cccd), cmd version 0, type 3 rtw890: chip rfe_type is 1 rtw890: rfkill hardware state changed to enable # pciconf -lv ... rtw890@pci0:2:0:0: class=0x028000 rev=0x00 hdr=0x00 vendor=0x10ec device=0xb852 subvendor=0x17aa subdevice=0x4853 vendor = 'Realtek Semiconductor Co., Ltd.' device = 'RTL8852BE PCIe 802.11ax Wireless Network Controller' class = network ... - From Lenovo's Website WIFI Realtek Wi-Fi 6 RTL8852BE 2x2 AX & Bluetooth 5.1 or above # kgdb /boot/kernel/kernel /var/crash/vmcore.0 Fatal trap 12: page fault while in kernel mode cpuid = 8; apic id = 08 fault virtual address = 0x8 fault code = supervisor write data, page not present instruction pointer = 0x20:0xffffffff83a50049 stack pointer = 0x28:0xfffffe00cbb3cd80 frame pointer = 0x28:0xfffffe00cbb3cdb0 code segment = base rx0, limit 0xfffff, type 0x1b = DPL 0, pres 1, long 1, def32 0, gran 1 processor eflags = interrupt enabled, resume, IOPL = 0 current process = 0 (linuxkpi_short_wq_1) rdi: fffffe00b9751800 rsi: fffffe0123087c40 rdx: fffff800040f8190 rcx: 0000000000000000 r8: 0000000000000003 r9: 00000000000000ff rax: 0000000000000000 rbx: fffffe0122387580 rbp: fffffe00cbb3cdb0 r10: 0000000000000000 r11: 000000000000ffff r12: fffffe0122388808 r13: 0000000000000000 r14: fffffe0123087530 r15: fffff800027da000 trap number = 12 panic: page fault cpuid = 8 time = 1765935606 KDB: stack backtrace: #0 0xffffffff80bbe1ed at kdb_backtrace+0x5d #1 0xffffffff80b71576 at vpanic+0x136 #2 0xffffffff80b71433 at panic+0x43 #3 0xffffffff81079f69 at trap_pfault+0x3c9 #4 0xffffffff8104ff18 at calltrap+0x8 #5 0xffffffff80de5d47 at lkpi_wiphy_work+0x157 #6 0xffffffff80e07e64 at linux_work_fn+0xe4 #7 0xffffffff80bd4de2 at taskqueue_run_locked+0x182 #8 0xffffffff80bd5fb2 at taskqueue_thread_loop+0xc2 #9 0xffffffff80b2786b at fork_exit+0x7b #10 0xffffffff81050f3e at fork_trampoline+0xe Uptime: 2m23s Dumping 740 out of 15076 MB:..3%..11%..22%..31%..42%..52%..61%..72%..83%..91% (kgdb) list *0xffffffff83a50049 0xffffffff83a50049 is in rtw89_fw_c2h_work (/usr/src/sys/compat/linuxkpi/common/include/linux/skbuff.h:656). 651 SKB_TRACE2(skb, q); 652 653 WRITE_ONCE(q->qlen, q->qlen - 1); 654 p = skb->prev; 655 n = skb->next; 656 WRITE_ONCE(n->prev, p); 657 WRITE_ONCE(p->next, n); 658 skb->prev = skb->next = NULL; 659 } 660 (kgdb) backtrace #0 __curthread () at /usr/src/sys/amd64/include/pcpu_aux.h:57 #1 doadump (textdump=<optimized out>) at /usr/src/sys/kern/kern_shutdown.c:399 #2 0xffffffff80b710f9 in kern_reboot (howto=260) at /usr/src/sys/kern/kern_shutdown.c:519 #3 0xffffffff80b71607 in vpanic (fmt=0xffffffff811d12d5 "%s", ap=ap@entry=0xfffffe00cbb3cc40) at /usr/src/sys/kern/kern_shutdown.c:974 #4 0xffffffff80b71433 in panic (fmt=<unavailable>) at /usr/src/sys/kern/kern_shutdown.c:887 #5 0xffffffff81079f69 in trap_fatal (frame=<optimized out>, eva=<optimized out>) at /usr/src/sys/amd64/amd64/trap.c:969 #6 0xffffffff81079f69 in trap_pfault (frame=0xfffffe00cbb3ccc0, usermode=false, signo=<optimized out>, ucode=<optimized out>) #7 <signal handler called> #8 __skb_unlink (skb=0xfffffe0123087530, q=0xfffffe01223887d8) at /usr/src/sys/compat/linuxkpi/common/include/linux/skbuff.h:656 #9 skb_unlink (skb=0xfffffe0123087530, q=0xfffffe01223887d8) at /usr/src/sys/compat/linuxkpi/common/include/linux/skbuff.h:668 #10 rtw89_fw_c2h_work (wiphy=<optimized out>, work=<optimized out>) at /usr/src/sys/contrib/dev/rtw89/fw.c:6827 #11 0xffffffff80de5d47 in lkpi_wiphy_work (work=<optimized out>) at /usr/src/sys/compat/linuxkpi/common/src/linux_80211.c:7764 #12 0xffffffff80e07e64 in linux_work_fn (context=0xfffffe0122387008, pending=<optimized out>) at /usr/src/sys/compat/linuxkpi/common/src/linux_work.c:308 #13 0xffffffff80bd4de2 in taskqueue_run_locked (queue=queue@entry=0xfffff800013cd700) at /usr/src/sys/kern/subr_taskqueue.c:517 #14 0xffffffff80bd5fb2 in taskqueue_thread_loop (arg=arg@entry=0xfffff80001335a40) at /usr/src/sys/kern/subr_taskqueue.c:829 #15 0xffffffff80b2786b in fork_exit (callout=0xffffffff80bd5ef0 <taskqueue_thread_loop>, arg=0xfffff80001335a40, frame=0xfffffe00cbb3cf40) at /usr/src/sys/kern/kern_fork.c:1153 #16 <signal handler called> #17 0xd5a1de13c387474c in ?? () Backtrace stopped: Cannot access memory at address 0x89449276e0e7c08
Ok, Jean-Claude MICHOT, the panic from comment #10 should be fixed in main and stable/15; see PR 290636. Phil, I thought the panic from comment #12 was fixed earlier this year but I need to go and dig. Miles, the panic from comment #16 (skb_unlink) I have seen twice during debugging this month myself but haven't made it to a reproducer yet; I still have to go and see.
(In reply to Bjoern A. Zeeb from comment #17) Thanks Bjoern! I can only add that with 15.0-RELEASE, as well as with 15-STABLE, both with the old and the recent firmware packages, the kernel panics on this machine as soon as if_rtw89.ko is loaded, with the following stack trace. Should I file a separate bug? Fatal trap 12: page fault while in kernel mode cpuid = 6; apic id = 06 fault virtual address = 0x8 fault code = supervisor write data, page not present instruction pointer = 0x20:0xffffffff84850049 stack pointer = 0x28:0xfffffe00fb792d80 frame pointer = 0x28:0xfffffe00fb792db0 code segment = base rx0, limit 0xfffff, type 0x1b = DPL 0, pres 1, long 1, def32 0, gran 1 processor eflags = interrupt enabled, resume, IOPL = 0 current process = 0 (linuxkpi_short_wq_1) rdi: fffffe002dac4000 rsi: fffffe01f76ce710 rdx: fffff8041fe19d20 rcx: 0000000000000000 r8: 0000000000000001 r9: 00000000000000ff rax: 0000000000000000 rbx: fffffe01f6cdb580 rbp: fffffe00fb792db0 r10: 0000000000000000 r11: 000000000000ffff r12: fffffe01f6cdc808 r13: 0000000000000000 r14: fffffe01f76ce000 r15: fffff801028b3780 trap number = 12 panic: page fault cpuid = 6 time = 1766273203 KDB: stack backtrace: #0 0xffffffff80bbe1ed at kdb_backtrace+0x5d #1 0xffffffff80b71576 at vpanic+0x136 #2 0xffffffff80b71433 at panic+0x43 #3 0xffffffff81079f69 at trap_pfault+0x3c9 #4 0xffffffff8104ff18 at calltrap+0x8 #5 0xffffffff80de5d47 at lkpi_wiphy_work+0x157 #6 0xffffffff80e07e64 at linux_work_fn+0xe4 #7 0xffffffff80bd4de2 at taskqueue_run_locked+0x182 #8 0xffffffff80bd5fb2 at taskqueue_thread_loop+0xc2 #9 0xffffffff80b2786b at fork_exit+0x7b #10 0xffffffff81050f3e at fork_trampoline+0xe
(In reply to Phil Krylov from comment #18) No need for a separate PR. This is the same as Miles from #16. If you can reliably reproduce it and are willing to apply a patch (or patches to the system, recompile your own kernel and test), let me know. I have no confirmed way to reproduce it yet.
(In reply to Bjoern A. Zeeb from comment #19) I'm more than happy to apply patches, recompile the kernel, etc. just let me know here or over email whatever I need to do. I get the crash every time I attempt to bring wlan0 up in ifconfig so I can consistently reproduce the error.
I have the same call stack as the one mentioned in comment 16. I'm using the release build from 15.0 This is on a machine with 24 CPU's. This is my procedure to test the realtech wireless card: sysctl net.wlan.devices ifconfig wlan0 create wlandev rtw890 ifconfig wlan0 ssid <whatever> ifconfig wlan0 up I instrumented the kernel (rtw89.ko) and verified that the probe function is completed successfully. When the device is brought up, the crash happens. The problem is related to the rtwdev->c2h_queue. Looking at the call stacks of the different threads I noticed two interesting stacks: Stack 1: Tracing command "kernel", '\000' <repeats 13 times> pid 0 tid 100885 (CPU 5) 298 #0 __curthread () at /usr/src/sys/amd64/include/pcpu_aux.h:57 299 #1 doadump (textdump=<optimized out>) at /usr/src/sys/kern/kern_shutdown.c:399 300 #2 0xffffffff80b710f9 in kern_reboot (howto=260) 301 at /usr/src/sys/kern/kern_shutdown.c:519 302 #3 0xffffffff80b71607 in vpanic (fmt=0xffffffff811d12d5 "%s", 303 ap=ap@entry=0xfffffe0183d4ab20) at /usr/src/sys/kern/kern_shutdown.c:974 304 #4 0xffffffff80b71433 in panic (fmt=<unavailable>) 305 at /usr/src/sys/kern/kern_shutdown.c:887 306 #5 0xffffffff81079f69 in trap_fatal (frame=<optimized out>, 307 eva=<optimized out>) at /usr/src/sys/amd64/amd64/trap.c:969 308 #6 0xffffffff81079f69 in trap_pfault (frame=0xfffffe0183d4aba0, 309 usermode=false, signo=<optimized out>, ucode=<optimized out>) 310 #7 <signal handler called> 311 #8 __skb_unlink (skb=0xfffffe0185a99710, q=0xfffffe0184f317d8) 312 at /usr/src/sys/compat/linuxkpi/common/include/linux/skbuff.h:656 313 #9 skb_unlink (skb=0xfffffe0185a99710, q=0xfffffe0184f317d8) 314 at /usr/src/sys/compat/linuxkpi/common/include/linux/skbuff.h:668 315 #10 rtw89_fw_c2h_purge_obsoleted_scan_events (rtwdev=0xfffffe0184f30580) 316 at /usr/src/sys/contrib/dev/rtw89/fw.c:6862 317 #11 0xffffffff83453347 in rtw89_mac_scan_offload (rtwdev=0xfffffe0184f30580, 318 option=0xfffffe0183d4acb8, rtwvif_link=0xfffffe018402b570, wowlan=false) 319 at /usr/src/sys/contrib/dev/rtw89/mac.h:1606 320 #12 rtw89_hw_scan_offload (rtwdev=rtwdev@entry=0xfffffe0184f30580, 321 rtwvif_link=rtwvif_link@entry=0xfffffe018402b570, enable=true) 322 at /usr/src/sys/contrib/dev/rtw89/fw.c:8169 323 #13 0xffffffff83419d1c in rtw89_ops_hw_scan (hw=<optimized out>, 324 vif=<optimized out>, req=0xfffff80001f70000) 325 at /usr/src/sys/contrib/dev/rtw89/mac80211.c:1243 326 #14 0xffffffff80ded87d in lkpi_80211_mo_hw_scan ( 327 hw=hw@entry=0xfffffe0184f304c0, vif=vif@entry=0xfffffe018402af00, 328 sr=sr@entry=0xfffff80001f70000) 329 at /usr/src/sys/compat/linuxkpi/common/src/linux_80211_macops.c:235 330 #15 0xffffffff80de099f in lkpi_ic_scan_start (ic=0xfffffe0184bba000) 331 at /usr/src/sys/compat/linuxkpi/common/src/linux_80211.c:4837 332 #16 0xffffffff80d1e853 in scan_start (arg=<optimized out>, 333 pending=<optimized out>) at /usr/src/sys/net80211/ieee80211_scan_sw.c:678 334 #17 0xffffffff80bd4de2 in taskqueue_run_locked ( 335 queue=queue@entry=0xfffff800017b8e00) 336 at /usr/src/sys/kern/subr_taskqueue.c:517 337 #18 0xffffffff80bd5fb2 in taskqueue_thread_loop ( 338 arg=arg@entry=0xfffffe0184bba110) at /usr/src/sys/kern/subr_taskqueue.c:829 339 #19 0xffffffff80b2786b in fork_exit ( 340 callout=0xffffffff80bd5ef0 <taskqueue_thread_loop>, 341 arg=0xfffffe0184bba110, frame=0xfffffe0183d4af40) 342 at /usr/src/sys/kern/kern_fork.c:1153 343 #20 <signal handler called> 344 #21 0xc20fc30fc20fc20f in ?? () Stack 2: Tracing command "kernel", '\000' <repeats 13 times> pid 0 tid 100111 (CPU 17) 535 #0 cpustop_handler () at /usr/src/sys/x86/x86/mp_x86.c:1528 536 #1 0xffffffff81047638 in ipi_nmi_handler () 537 at /usr/src/sys/x86/x86/mp_x86.c:1485 538 #2 0xffffffff8103cf8f in nmi_handle_intr (frame=<unavailable>) 539 at /usr/src/sys/x86/x86/cpu_machdep.c:964 540 #3 <signal handler called> 541 #4 lock_delay (la=la@entry=0xfffffe00d742dd10) 542 at /usr/src/sys/kern/subr_lock.c:124 543 #5 0xffffffff80b4c8d8 in __mtx_lock_sleep (c=c@entry=0xfffffe0184f31808, 544 v=<optimized out>) at /usr/src/sys/kern/kern_mutex.c:624 545 #6 0xffffffff834506b5 in skb_unlink (skb=0xfffffe0185a98170, 546 q=0xfffffe0184f317d8) 547 at /usr/src/sys/compat/linuxkpi/common/include/linux/skbuff.h:667 548 #7 rtw89_fw_c2h_work (wiphy=<optimized out>, work=<optimized out>) 549 at /usr/src/sys/contrib/dev/rtw89/fw.c:6833 550 #8 0xffffffff80de5d47 in lkpi_wiphy_work (work=<optimized out>) 551 at /usr/src/sys/compat/linuxkpi/common/src/linux_80211.c:7764 552 #9 0xffffffff80e07e64 in linux_work_fn (context=0xfffffe0184f30008, 553 pending=<optimized out>) 554 at /usr/src/sys/compat/linuxkpi/common/src/linux_work.c:308 555 #10 0xffffffff80bd4de2 in taskqueue_run_locked ( 556 queue=queue@entry=0xfffff80001482900) 557 at /usr/src/sys/kern/subr_taskqueue.c:517 558 #11 0xffffffff80bd5fb2 in taskqueue_thread_loop ( 559 arg=arg@entry=0xfffff8000141acc0) at /usr/src/sys/kern/subr_taskqueue.c:829 560 #12 0xffffffff80b2786b in fork_exit ( 561 callout=0xffffffff80bd5ef0 <taskqueue_thread_loop>, 562 arg=0xfffff8000141acc0, frame=0xfffffe00d742df40) 563 at /usr/src/sys/kern/kern_fork.c:1153 564 #13 <signal handler called> 565 So it looks like we have two stacks operating at the same time on the rtwdev->c2h_queue. I can't see any locking in the function itself and in its callers. It might be locked higher in the stack. As an experiment I changed the rtw89_fw_c2h_purge_obsoleted_scan_events function to immediately return. With this change, the kernel no longer crashes, and the device is in up state. I'm able to scan other wireless networks now with: ifconfig wlan0 up scan So it looks to me like, the queue operations on rtwdev->c2h_queue are not correctly locked. The more CPU's you have the more likely you are to run into that problem. However, the command dhclient wlan0 still fails with "no link". I can see the following in /var/log/messages rtw890: ERROR: lkpi_ic_scan_start: hw_scan returned -16
As a better solution I introduced a new mutex for all operations on rtwdev->c2h_queue. In my tests that works reliably and I don't see any more panics. I can see an additional message in messages file: rtw890: Update probe request failed: ret -16 This seems to come from the function rtw89_wait_for_cond, which returns -EBUSY in case it cannot do the atomic exchange: int rtw89_wait_for_cond(struct rtw89_wait_info *wait, unsigned int cond) 4910 { 4911 › struct completion *cmpl = &wait->completion; 4912 › unsigned long time_left; 4913 › unsigned int cur; 4914 4915 › cur = atomic_cmpxchg(&wait->cond, RTW89_WAIT_COND_IDLE, cond); 4916 › if (cur != RTW89_WAIT_COND_IDLE) 4927 › › return -EBUSY;
The wait event in question is rtwdev->mac.fw_ofld_wait. I assume this is used to wait for the scan offload to complete.
(In reply to Stefan Roesch from comment #21) Just for the records, which chipset do you have?
Chipset information: rtw890@pci0:2:0:0: class=0x028000 rev=0x00 hdr=0x00 vendor=0x10ec device=0xb852 subvendor=0x103c subdevice=0x88e3 vendor = 'Realtek Semiconductor Co., Ltd.' device = 'RTL8852BE PCIe 802.11ax Wireless Network Controller' class = network
I searched for previous fixes to the linux kernel driver code. I found the following fix, which was merged on Nov 21 2025. https://github.com/morrownr/rtw89/commit/fbc39b25da9c3cd2928e7e6b50b12eead62628c1 The fix is not part of the release build.
If I cherrypick that fix, the page fault crash is also gone.
(In reply to Stefan Roesch from comment #27) That change (b47d748110fd2c2911b59b169b10cc14fd5d3648) if part of the 6.19 RC series and there's follow-up changes which prevent the firmware to get stuck in a certain condition. I'll prepare to pull v6.19-rc6 into FreeBSD the next days. I had waited for another mt76 fix end of last year but it seems we are good currently. Should make the update to a final 6.19 easy then.
*** Bug 292847 has been marked as a duplicate of this bug. ***
Given the vendor imports were done, there's a few reviews pending before I can push 6.19-rc6 (or maybe 6.19) into the tree which should fix this. In case any committer wants to help with the reviews pending, I believe it is these two which would be blocking rtw89: https://reviews.freebsd.org/D54808 https://reviews.freebsd.org/D55029
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/src/commit/?id=10f91d9ff7b827aac6035e224e6ddc2c79f3a0cd commit 10f91d9ff7b827aac6035e224e6ddc2c79f3a0cd Merge: bf603dbbb2c9 f5eb24d531de Author: Bjoern A. Zeeb <bz@FreeBSD.org> AuthorDate: 2025-12-05 20:48:02 +0000 Commit: Bjoern A. Zeeb <bz@FreeBSD.org> CommitDate: 2026-02-10 08:35:34 +0000 rtw89: update Realtek's rtw89 driver This version is based on git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git 24d479d26b25bce5faea3ddd9fa8f3a6c3129ea7 ( tag: v6.19-rc6 ). Sponsored by: The FreeBSD Foundation MFC after: 3 days PR: 285228 sys/contrib/dev/rtw89/Kconfig | 22 + sys/contrib/dev/rtw89/Makefile | 6 + sys/contrib/dev/rtw89/cam.c | 173 +++--- sys/contrib/dev/rtw89/cam.h | 446 ++++------------ sys/contrib/dev/rtw89/chan.c | 11 +- sys/contrib/dev/rtw89/chan.h | 10 +- sys/contrib/dev/rtw89/coex.c | 5 +- sys/contrib/dev/rtw89/core.c | 911 ++++++++++++++++++++++++++++---- sys/contrib/dev/rtw89/core.h | 242 ++++++++- sys/contrib/dev/rtw89/debug.c | 424 ++++++++++++++- sys/contrib/dev/rtw89/debug.h | 1 + sys/contrib/dev/rtw89/fw.c | 353 ++++++++++--- sys/contrib/dev/rtw89/fw.h | 144 ++++- sys/contrib/dev/rtw89/mac.c | 270 ++++++++-- sys/contrib/dev/rtw89/mac.h | 115 +++- sys/contrib/dev/rtw89/mac80211.c | 124 ++++- sys/contrib/dev/rtw89/mac_be.c | 10 +- sys/contrib/dev/rtw89/pci.c | 486 +++++++++++++---- sys/contrib/dev/rtw89/pci.h | 132 ++++- sys/contrib/dev/rtw89/pci_be.c | 18 +- sys/contrib/dev/rtw89/phy.c | 541 +++++++++++++++++-- sys/contrib/dev/rtw89/phy.h | 24 +- sys/contrib/dev/rtw89/phy_be.c | 13 + sys/contrib/dev/rtw89/ps.c | 26 +- sys/contrib/dev/rtw89/reg.h | 80 ++- sys/contrib/dev/rtw89/regd.c | 22 +- sys/contrib/dev/rtw89/rtw8851b.c | 7 + sys/contrib/dev/rtw89/rtw8851b_rfk.c | 167 +++--- sys/contrib/dev/rtw89/rtw8851be.c | 4 + sys/contrib/dev/rtw89/rtw8851bu.c | 27 + sys/contrib/dev/rtw89/rtw8852a.c | 129 ++++- sys/contrib/dev/rtw89/rtw8852a_rfk.c | 16 +- sys/contrib/dev/rtw89/rtw8852ae.c | 4 + sys/contrib/dev/rtw89/rtw8852au.c (new) | 79 +++ sys/contrib/dev/rtw89/rtw8852b.c | 7 + sys/contrib/dev/rtw89/rtw8852b_common.c | 6 +- sys/contrib/dev/rtw89/rtw8852b_rfk.c | 6 +- sys/contrib/dev/rtw89/rtw8852be.c | 4 + sys/contrib/dev/rtw89/rtw8852bt.c | 4 + sys/contrib/dev/rtw89/rtw8852bt_rfk.c | 14 +- sys/contrib/dev/rtw89/rtw8852bte.c | 4 + sys/contrib/dev/rtw89/rtw8852bu.c | 26 + sys/contrib/dev/rtw89/rtw8852c.c | 172 +++++- sys/contrib/dev/rtw89/rtw8852c.h | 2 +- sys/contrib/dev/rtw89/rtw8852c_rfk.c | 69 ++- sys/contrib/dev/rtw89/rtw8852ce.c | 4 + sys/contrib/dev/rtw89/rtw8852cu.c (new) | 69 +++ sys/contrib/dev/rtw89/rtw8922a.c | 26 +- sys/contrib/dev/rtw89/rtw8922ae.c | 4 + sys/contrib/dev/rtw89/sar.c | 15 + sys/contrib/dev/rtw89/sar.h | 1 + sys/contrib/dev/rtw89/ser.c | 5 +- sys/contrib/dev/rtw89/txrx.h | 45 +- sys/contrib/dev/rtw89/usb.c | 115 ++-- sys/contrib/dev/rtw89/usb.h | 12 + sys/contrib/dev/rtw89/wow.c | 87 ++- sys/contrib/dev/rtw89/wow.h | 6 + sys/modules/rtw89/Makefile | 7 +- 58 files changed, 4651 insertions(+), 1101 deletions(-)
I will merge the driver update to stable/15 in a few days. I do not know what I will do about stable/14 yet. For 14.4 we are too late as the other intermediate changes which would need to be merged are still huge and there was no feedback on the request for testing.
A commit in branch stable/15 references this bug: URL: https://cgit.FreeBSD.org/src/commit/?id=0d9a154fe73ba61951708f277bf9aee992588545 commit 0d9a154fe73ba61951708f277bf9aee992588545 Author: Bjoern A. Zeeb <bz@FreeBSD.org> AuthorDate: 2025-12-05 20:48:02 +0000 Commit: Bjoern A. Zeeb <bz@FreeBSD.org> CommitDate: 2026-02-26 23:06:04 +0000 rtw89: update Realtek's rtw89 driver This version is based on git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git 24d479d26b25bce5faea3ddd9fa8f3a6c3129ea7 ( tag: v6.19-rc6 ). Sponsored by: The FreeBSD Foundation PR: 285228 (cherry picked from commit 10f91d9ff7b827aac6035e224e6ddc2c79f3a0cd) sys/contrib/dev/rtw89/Kconfig | 22 + sys/contrib/dev/rtw89/Makefile | 6 + sys/contrib/dev/rtw89/cam.c | 173 +++--- sys/contrib/dev/rtw89/cam.h | 446 ++++------------ sys/contrib/dev/rtw89/chan.c | 11 +- sys/contrib/dev/rtw89/chan.h | 10 +- sys/contrib/dev/rtw89/coex.c | 5 +- sys/contrib/dev/rtw89/core.c | 911 ++++++++++++++++++++++++++++---- sys/contrib/dev/rtw89/core.h | 242 ++++++++- sys/contrib/dev/rtw89/debug.c | 424 ++++++++++++++- sys/contrib/dev/rtw89/debug.h | 1 + sys/contrib/dev/rtw89/fw.c | 353 ++++++++++--- sys/contrib/dev/rtw89/fw.h | 144 ++++- sys/contrib/dev/rtw89/mac.c | 270 ++++++++-- sys/contrib/dev/rtw89/mac.h | 115 +++- sys/contrib/dev/rtw89/mac80211.c | 124 ++++- sys/contrib/dev/rtw89/mac_be.c | 10 +- sys/contrib/dev/rtw89/pci.c | 486 +++++++++++++---- sys/contrib/dev/rtw89/pci.h | 132 ++++- sys/contrib/dev/rtw89/pci_be.c | 18 +- sys/contrib/dev/rtw89/phy.c | 541 +++++++++++++++++-- sys/contrib/dev/rtw89/phy.h | 24 +- sys/contrib/dev/rtw89/phy_be.c | 13 + sys/contrib/dev/rtw89/ps.c | 26 +- sys/contrib/dev/rtw89/reg.h | 80 ++- sys/contrib/dev/rtw89/regd.c | 22 +- sys/contrib/dev/rtw89/rtw8851b.c | 7 + sys/contrib/dev/rtw89/rtw8851b_rfk.c | 167 +++--- sys/contrib/dev/rtw89/rtw8851be.c | 4 + sys/contrib/dev/rtw89/rtw8851bu.c | 27 + sys/contrib/dev/rtw89/rtw8852a.c | 129 ++++- sys/contrib/dev/rtw89/rtw8852a_rfk.c | 16 +- sys/contrib/dev/rtw89/rtw8852ae.c | 4 + sys/contrib/dev/rtw89/rtw8852au.c (new) | 79 +++ sys/contrib/dev/rtw89/rtw8852b.c | 7 + sys/contrib/dev/rtw89/rtw8852b_common.c | 6 +- sys/contrib/dev/rtw89/rtw8852b_rfk.c | 6 +- sys/contrib/dev/rtw89/rtw8852be.c | 4 + sys/contrib/dev/rtw89/rtw8852bt.c | 4 + sys/contrib/dev/rtw89/rtw8852bt_rfk.c | 14 +- sys/contrib/dev/rtw89/rtw8852bte.c | 4 + sys/contrib/dev/rtw89/rtw8852bu.c | 26 + sys/contrib/dev/rtw89/rtw8852c.c | 172 +++++- sys/contrib/dev/rtw89/rtw8852c.h | 2 +- sys/contrib/dev/rtw89/rtw8852c_rfk.c | 69 ++- sys/contrib/dev/rtw89/rtw8852ce.c | 4 + sys/contrib/dev/rtw89/rtw8852cu.c (new) | 69 +++ sys/contrib/dev/rtw89/rtw8922a.c | 26 +- sys/contrib/dev/rtw89/rtw8922ae.c | 4 + sys/contrib/dev/rtw89/sar.c | 15 + sys/contrib/dev/rtw89/sar.h | 1 + sys/contrib/dev/rtw89/ser.c | 5 +- sys/contrib/dev/rtw89/txrx.h | 45 +- sys/contrib/dev/rtw89/usb.c | 115 ++-- sys/contrib/dev/rtw89/usb.h | 12 + sys/contrib/dev/rtw89/wow.c | 87 ++- sys/contrib/dev/rtw89/wow.h | 6 + sys/modules/rtw89/Makefile | 7 +- 58 files changed, 4651 insertions(+), 1101 deletions(-)
I do not know what to do about stable/14 yet; I'll investigate over the weekend.
Not sure if it's the same issue but as soon as I try to authenticate to a selected network the system reboots immediately and panics with a page fault.
(In reply to Jordan from comment #35) Jordan, which version of FreeBSD are you running? If you say page_fault--do you have a core.txt or a backtrace?
Created attachment 268837 [details] Panic screenshot
Bjoern, I'm on 15.0 RELEASE and I have attached the panic picture
(In reply to Jordan from comment #38) Hi Jordan, (a) 15.0-R does not have the updated driver with the fix, 15.1-R will have and stable/15 has. (b) The panic backtrace you sent looks like a non-WiFi problem. I'd suggest to report it on the stable mailing list as it doesn't have much detail.
Hey Bjoern, Thank you, I agree, it looks more kind of a hardware fault, yet it happens as soon as I authenticate to a network. Also, I had a previous panic about a page fault and the rtw89 driver, always in the same situation, but unfortunately I couldn't get a screen, the thing seems very erratic. I'll see if I can use wifibox for now.
(In reply to Jordan from comment #40) Hi, you could try to setup automated crashdumps upon panic; that way we'd get a full picture also for rtw89. If you want to try wifibox, you could also try to download a FreeBSD stable/15 snapshot and install it in a VM with rtw89 as passthru and see if there it works. That way we'd know that you'd at least be fine for 15.1-R.