pfSense Plus crash after adding OVPN as interface

andrei-z

Hi, everyone!

I've got a pfSense Plus 22.05-RELEASE VM running under Proxmox VE 7.2-11. It's been running excellently up until recently. I'm trying to get Bonjour running over my VPN, however that's not the issue currently at hand.

I went to Interfaces > Assignments and I removed an older HE.net IPv6 GIF tunnel interface that I'd intended to use, but ended up not doing so. Afterwards, I'd also deleted the GIF tunnel itself. Then, I went back into interface assignments and added ovpns1 (OVPN1) as an interface. Then, I clicked on OPT7 (its initial name), changed it to OPENVPN and enabled the interface. I saved the changes and applied them. After clicking the Apply Changes button, the WebUI locked up completely and I'd lost all internet connectivity on all of my devices on the network. I didn't try pinging pfSense itself, unfortunately.

Within 30-40 seconds, connectivity was established again, however everything was very slow. On the system dashboard, I got a message saying Netgate pfSense Plus has detected a crash report or programming bug. Click here for more information.. Therefore, I'd figured something had crashed. I've downloaded both of the files pfSense Plus has provided me with and I'm attaching them here:

I rebooted the pfSense Plus VM and now it's back to its original speed, so that issue is fixed. I don't have the technical skill to figure out what went wrong beyond reading Fatal trap 12: page fault while in kernel mode among the last lines in the second file, which I'm assuming is the error.

I've tried removing the OpenVPN interface and adding it again and pfSense didn't crash this time.

If anyone has any idea as to whether why the crash occurred and/or can suggest any fixes, I'd be very grateful.

Thanks,
Andrei

stephenw10

Back trace:

db:0:kdb.enter.default>  bt
Tracing pid 0 tid 100022 td 0xfffff800052b8740
kdb_enter() at kdb_enter+0x37/frame 0xfffffe00259f9040
vpanic() at vpanic+0x194/frame 0xfffffe00259f9090
panic() at panic+0x43/frame 0xfffffe00259f90f0
trap_fatal() at trap_fatal+0x38f/frame 0xfffffe00259f9150
trap_pfault() at trap_pfault+0x4f/frame 0xfffffe00259f91b0
calltrap() at calltrap+0x8/frame 0xfffffe00259f91b0
--- trap 0xc, rip = 0xffffffff80ed2d4d, rsp = 0xfffffe00259f9280, rbp = 0xfffffe00259f9350 ---
rtsock_msg_mbuf() at rtsock_msg_mbuf+0xed/frame 0xfffffe00259f9350
rt_newmaddrmsg() at rt_newmaddrmsg+0x102/frame 0xfffffe00259f9400
if_delmulti_locked() at if_delmulti_locked+0x2a/frame 0xfffffe00259f9430
if_delmulti_ifma_flags() at if_delmulti_ifma_flags+0x54/frame 0xfffffe00259f9460
in6m_release_task() at in6m_release_task+0x16d/frame 0xfffffe00259f94a0
taskqueue_run_locked() at taskqueue_run_locked+0x144/frame 0xfffffe00259f9500
taskqueue_thread_loop() at taskqueue_thread_loop+0xd2/frame 0xfffffe00259f9530
fork_exit() at fork_exit+0x7e/frame 0xfffffe00259f9570
fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe00259f9570
--- trap 0, rip = 0, rsp = 0, rbp = 0 ---

Panic:

<6>gif0: link state changed to DOWN


Fatal trap 12: page fault while in kernel mode
cpuid = 11; apic id = 0b
fault virtual address	= 0x68
fault code		= supervisor read data, page not present
instruction pointer	= 0x20:0xffffffff80ed2d4d
stack pointer	        = 0x28:0xfffffe00259f9280
frame pointer	        = 0x28:0xfffffe00259f9350
code segment		= base 0x0, limit 0xfffff, type 0x1b
			= DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags	= interrupt enabled, resume, IOPL = 0
current process		= 0 (in6m_free taskq)
trap number		= 12
panic: page fault
cpuid = 11
time = 1665857437
KDB: enter: panic

That's not a familiar crash to me. It sure looks like it could be from the removal of the gif interface though. Was that assigned? Perhaps it was still assigned with an IP when it was removed somehow? That shouldn't be allowed normally.

Steve

andrei-z

@stephenw10 Thank you for the information and your swift reply. I believe the interface did have an IP assigned when I'd removed it. Nothing stopped me from removing it. Could be a bug?

stephenw10

Yes, it could be. I'll try to replicate and open something if there isn't anything already open.