pf_test: kif == NULL, if_xname on multi-WAN and "Reset all states if WAN IP Address changes"

w0w

It's looks like I've different issue, caused by option "Reset all states if WAN IP Address changes".
Even having gif/IPv6 interfaces deleted this option triggers those "pf_test: ..." messages if I manipulate with WAN interface or playing with media where WAN (PPPoE) based on.
Looks like having this option not enabled makes the system stable again.

w0w

Tested 2.4.5 — not affected.

@w0w said in pf_test: kif == NULL, if_xname:

Tested 2.4.5 — not affected.

If you can provide me how to set this up and can see if my environment would permit me to test it on my version of 12-STABLE. Its current and just finished rebasing it.

I'm currently using using a PPPOE connection and using the stf to implement ipv6 via 6RD, The stf has been modified with commits that I pulled forward from 2.4.x.

w0w

@rschell
Actually I have multiple WANs, but I can test it in single WAN config later, to see if bug still there or not.
If you have WAN PPPoE, go to the System-Advanced-Networking and tick the option "Reset all states if WAN IP Address changes" reboot your firewall to make things clean and just try to re-plug the ethernet cable on PPPoE WAN port, unplug, wait for a 15 sec and put it back. If you don't see those pf_test messages in console or in system logs, system does not crashing, it's good.
Tested, it fails only with multiple WANs and Gateway group configured. So may be I need to test your 12-STABLE version by myself, can I download it somewhere?

@w0w I currently don't have any file sharing set up.

w0w

@rschell
https://filebin.net/ without registration.

@w0w said in pf_test: kif == NULL, if_xname on multi-WAN and "Reset all states if WAN IP Address changes":

https://filebin.net/ without registration.

Uploaded files@filebin.net

w0w

Sad to say but I have same error on the same config. Regardless of option "Reset all states if WAN IP Address changes" enabled or not after known sequence I get those pf_test: kif == NULL, if_xname and then Fatal trap 12 a bit later.

@w0w did you try again after rebooting? I seen a page fault occur on my bare metal instance but not VM, but much earlier in the boot process. Second boot it doesn't occur.

Not likely to fix the original error however.

w0w

It was rebooted multiple times before I did test sequence to crash it , so it is some bug. I'll wait for Netgate version, when it is out or if somebody else find something. It can be also related to ix driver, not sure, will test it also a bit later

w0w

On the igb0 I have similar situation, but different message on 2.5.0-DEVELOPMENT (amd64)
built on Fri Mar 13. Instead of pf_test spam I have this one...

kernel: igb0: TX(3) desc avail = 1024, pidx = 0

Looks like it's this one
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=239240

rschell do you have the patch mentioned in the thread included in your firewall version?

@w0w The patches mentioned in 239240 (MFC 356310) and have been incorporated. If its posted in 12-stable branch of pfsense/FreeBSD-src, as of the build date, its in there .

Impatient

I had that fatal trap 12: page fault message happen on a test machine with 2.5.

I tested mb,drive,memory,switched cpu everything tested good.

I finally swapped memory even though the previous mem passed memtest for 24 hrs
and the machine has been running fine since.

w0w

@Impatient
I do not think it's memory, could be Intel card driver, I don't have any problem until I do known sequences on known configuration. Also 2.4.5 is not affected by this bug on the same hardware.

Impatient

If it was the network card would it not mention that in the error?

It does mention what is not in memory (supervisor read data)

Then the result of that is the (panic: page fault).

I run dual wan PPPoE with Reset all states if WAN IP Address changes enabled
and get no error's running 2.5.0.a.20200313.1633 as of today on above system.
Previously in 2.5.0.a.20200312.something.

It has only been running for about 12 hrs. with 1 reboot.

w0w

@Impatient
My firewall is overall stable on 2.5.0, I can run it for a month until something triggers the fatal trap or "pf_test: kif == NULL", sometimes it crash on boot and 1 of 3 times crashing or/and this "pf_test: kif == NULL" message spamming, when there is the problem on the PPPoE ethernet port — for example ISP router is out of power or I have unplugged the cable or changed the link speed.
I repeat, I do not have any of the mentioned problems on 2.4.5 version, so I think it's software problem.

BlakeThoennes

Hello all,

As was announced 2.5 to be soon mainstreamed, I jumped onto the dev branch.

15 minutes after successful upgrade net went down, I couldnt ping or load pfsense web.
Error streaming on screen without stop was

igb1: TX(2) desc avail = 1024, pidx = 0

Running 2.5.0.a.20200330.1049 (latest x64)
Potential Solution based on power settings posted here? Not sure how to issue these commands. (from terminal at firewall, thru web, ssh...???)
https://forum.netgate.com/topic/149643/igbx-bug-tx-1-desc-avail-no-carrier

Bug report mentioned elsewhere here:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=239240

Had been running stable until now and never had a single issue.
Only way to get the firewall working (tried ifconfig up down, bridge up down, reset web config, etc) was to reboot the firewall. Has been working for over 50 minutes now without issue...

Heres the logs at time of event.
Mar 30 15:15:15 kernel igb1: TX(2) desc avail = 1024, pidx = 0
Mar 30 15:15:15 check_reload_status 356 Reloading filter
Mar 30 15:15:15 php-fpm 343 /rc.linkup: Hotplug event detected for LAN(lan) static IP (192.168.1.1 )
Mar 30 15:15:14 kernel igb1: link state changed to DOWN
Mar 30 15:15:14 check_reload_status 356 Linkup starting $igb1
Mar 30 15:15:14 kernel igb1: TX(1) desc avail = 42, pidx = 712

Additional Log Hardware Information for comparison to yours if you suffer from this too...
Mar 30 15:21:04 kernel wlan: mac acl policy registered
Mar 30 15:21:04 kernel random: entropy device external interface
Mar 30 15:21:04 kernel Timecounter "TSC" frequency 1600051300 Hz quality 1000
Mar 30 15:21:04 kernel Launching APs: 3 2 1
Mar 30 15:21:04 kernel ioapic0 <Version 2.0> irqs 0-114 on motherboard
Mar 30 15:21:04 kernel random: unblocking device.
Mar 30 15:21:04 kernel FreeBSD/SMP: 1 package(s) x 4 core(s)
Mar 30 15:21:04 kernel FreeBSD/SMP: Multiprocessor System Detected: 4 CPUs
Mar 30 15:21:04 kernel WARNING: L1 data cache covers fewer APIC IDs than a core (0 < 1)
Mar 30 15:21:04 kernel ACPI APIC Table: <ALASKA A M I >
Mar 30 15:21:04 kernel Event timer "LAPIC" quality 600
Mar 30 15:21:04 kernel avail memory = 8161169408 (7783 MB)
Mar 30 15:21:04 kernel real memory = 8589934592 (8192 MB)
Mar 30 15:21:04 kernel TSC: P-state invariant, performance statistics
Mar 30 15:21:04 kernel VT-x: PAT,HLT,MTF,PAUSE,EPT,UG,VPID
Mar 30 15:21:04 kernel Structured Extended Features3=0xc000400<IBPB,STIBP>
Mar 30 15:21:04 kernel Structured Extended Features=0x2282<TSCADJ,SMEP,ERMS,NFPUSG>
Mar 30 15:21:04 kernel AMD Features2=0x101<LAHF,Prefetch>
Mar 30 15:21:04 kernel AMD Features=0x28100800<SYSCALL,NX,RDTSCP,LM>
Mar 30 15:21:04 kernel Features2=0x43d8e3bf<SSE3,PCLMULQDQ,DTES64,MON,DS_CPL,VMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,SSE4.1,SSE4.2,MOVBE,POPCNT,TSCDLT,AESNI,RDRAND>
Mar 30 15:21:04 kernel Features=0xbfebfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE>
Mar 30 15:21:04 kernel Origin="GenuineIntel" Id=0x406c4 Family=0x6 Model=0x4c Stepping=4
Mar 30 15:21:04 kernel CPU: Intel(R) Celeron(R) CPU J3160 @ 1.60GHz (1600.05-MHz K8-class CPU)
Mar 30 15:21:04 kernel VT(efifb): resolution 800x600
Mar 30 15:21:04 kernel FreeBSD clang version 6.0.1 (tags/RELEASE_601/final 335540) (based on LLVM 6.0.1)
Mar 30 15:21:04 kernel FreeBSD 12.0-RELEASE-p10 ce9563d5729(RELENG_2_5) pfSense amd64

Suggestions? Recommend I go back to stable?

w0w

I am waiting for 2.5.0 based on STABLE. Expected to be soon. I hope that it will be released before the end of the world.

BlakeThoennes

Good news is it hasnt failed on me again since the one time 15m after original upgrade.

So... scary start, but stable for nearly 24hours now.

Hopefully that release is SOON.

w0w

Re-tested on the latest build
2.5.0-DEVELOPMENT (amd64)
built on Fri Apr 03 19:36:42 EDT 2020
FreeBSD 12.0-RELEASE-p10
Don't know what exactly was merged/patched, whatever else, but I can not crash my system anymore with my test sequence. I'll check later what changes are made on github.