virtualbox fatal trap 12 on 11.2-Release

romanaOne

romanaOne

I installed virtualbox-ose and now my system will not boot! As soon as the vboxblahblah.ko modules load, it says something about fatal trap 12 and the system reboots.

I can't boot single user either; the same thing happens.

The only way I know to fix it is to boot the install disk, choose live-cd, mount bootpool on /mnt, and delete /boot/modules/vbox*. Just commenting out vboxdrv_load="YES" in /boot/loader.conf does not allow me to reboot. The vboxblahblah.ko modules load anyway?

Is there an easier way to mount bootpool on the rescue/install CD? This is pretty cumbersome and there is no way to create the mountpoint it expects /bootpool

The system is HP Elitebook 8470p i5 with Intel HD3000 and I've just upgraded from 11.2-RC3 which I installed a few days ago, so the system has not been heavily customized.
 
check /etc/rc.conf for a line like kld_list= and make sure the vbox drivers are not being loaded there too. What you are describing tells me that they are being installed by multiple entities in your startup scripts. Check the meaning of trap 12. It is a user defined signal, but old freebsd posts say it is being trapped by the kernel to indicate an unsupported syscall(). Are you sure you are running vbox on a supported CPU?
 
In /etc/rc.conf I have vboxnet_enable="YES" and in /boot/loader.conf, I had vboxdrv_load="YES"

Is there someplace else?
Code: 
>grep vbox /boot/loader.*
loader.conf:#vboxdrv_load="YES"
>grep vbox /defaults/loader.*

VT-x and VT-d are enabled in the BIOS (or whatever you call it on UEFI systems.)

/var/run/dmesg.boot says about the processor:

Code: 
CPU: Intel(R) Core(TM) i5-3340M CPU @ 2.70GHz (2693.94-MHz K8-class CPU)

  Origin="GenuineIntel"  Id=0x306a9  Family=0x6  Model=0x3a  Stepping=9

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>

  Features2=0x7fbae3ff<SSE3,PCLMULQDQ,DTES64,MON,DS_CPL,VMX,SMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,PCID,SSE4.1,SSE4.2,x2APIC,POPCNT,TSCDLT,AESNI,XSAVE,OSXSAVE,AVX,F16C,RDRAND>

  AMD Features=0x28100800<SYSCALL,NX,RDTSCP,LM>

  AMD Features2=0x1<LAHF>

  Structured Extended Features=0x281<FSGSBASE,SMEP,ERMS>

  XSAVE Features=0x1<XSAVEOPT>

  VT-x: PAT,HLT,MTF,PAUSE,EPT,UG,VPID

  TSC: P-state invariant, performance statistics
 
romanaOne said:
In /etc/rc.conf I have vboxnet_enable="YES" and in /boot/loader.conf, I had vboxdrv_load="YES"

Is there someplace else?
Code: 
>grep vbox /boot/loader.*
loader.conf:#vboxdrv_load="YES"
>grep vbox /defaults/loader.*
Click to expand...

Well it seems based on scottro post that 11-2 is not ready for primetime so all bets are off.
 
I now uninstalled virtualbox and removed all the settings but something is botched up somewhere cause I can't upgrade at all as in my other thread.
 
  • Thanks
Reactions: Oko
Yes, these posts have convinced me to wait. On a couple of laptops, (no NVidia, no VirtualBox) it went without problems, but I'm not sure about my workstations, which have both NVidia and VirtualBox.

I do wonder if a fresh install, as opposed to an upgrade, would have the same problems. If someone buys me a workstation, I'll be happy to test it for them. :)
 
scottro said:
Yes, these posts have convinced me to wait. On a couple of laptops, (no NVidia, no VirtualBox) it went without problems, but I'm not sure about my workstations, which have both NVidia and VirtualBox.

I do wonder if a fresh install, as opposed to an upgrade, would have the same problems. If someone buys me a workstation, I'll be happy to test it for them. :)
Click to expand...

Historically I shun "updates". They are by nature more complicated to validate than a fresh install.
 
I'd also be curious (and hopefully someone who has done it will post in this thread) if it went more smoothly upgrading from source rather than FreeBSD update. At least, judging from drhowarddrfine's post in the other thread, on servers it seems to have gone smoothly, trouble coming with VirtualBox and/or NVidia both of which are 3rd party ports.
 
scottro said:
I'd also be curious (and hopefully someone who has done it will post in this thread) if it went more smoothly upgrading from source rather than FreeBSD update. At least, judging from drhowarddrfine's post in the other thread, on servers it seems to have gone smoothly, trouble coming with VirtualBox and/or NVidia both of which are 3rd party ports.
Click to expand...

and both vbox and nvidia are EXTREMELY complicated pieces of software, with lots of OS version specific hooks.
 
I upgraded my VirtualBox VM this morning with trepidation after reading this, but it worked just fine. I've not got anything VBox related in my /boot/loader.conf, but have the following in my /etc/rc.conf:
Code: 
vboxguest_enable="YES"
vboxservice_enable="YES"

The VM sits on a Dell Latitude E7470 with an Intel Core i7.
 
I built virtualbox-ose and vitrualbox-ose-kmod 5.2.12 from ports and it appears to work fine now. This is a newer version than the binary package (5.2.8_1), so maybe whatever it was got fixed.
 
romanaOne said:
I built virtualbox-ose and vitrualbox-ose-kmod 5.2.12 from ports and it appears to work fine now. This is a newer version than the binary package (5.2.8_1), so maybe whatever it was got fixed.
Click to expand...
That's probably safer anyway, as when kernel modules are involved, versioning is more important.
 
I checked this out on a different system (Skylake i7 6700HQ), now that there's a newer version (5.2.14) is in the latest binary repo:

First I tried installing virtualbox-ose 5.2.14 from pkg. Sure enough, I got the fatal trap 12 boot loop again.

After booting from a rescue drive and removing /boot/modules/vbox*, I built it from ports and it works fine.

So on two different computers, installing virtualbox-ose from binary pkgs causes a boot loop.

btw, the reason I'm using a rescue disk is because I can't get beadm to work on the Skylake system: I've got no boot menu (just a black screen) but the system seems to boot up ok after a slight 10sec pause.
 
romanaOne said:
So on two different computers, installing virtualbox-ose from binary pkgs causes a boot loop.
Click to expand...
Yes, this is a known issue. The 'problem' is that the binary packages are built for 11.1 because that's the lowest, still supported, version. Both the NVidia driver and the VirtualBox guest addons use a kernel driver. This kernel driver needs to be specifically built for the kernel it runs on. So the reason it's crashing is because the packages contain a 11.1 kernel module and you're trying to load that module on 11.2. If you build from ports both will use the correct kernel version and everything will work out just fine.

The "problem" is only temporary, as soon as 11.1 is EoL packages will be built for 11.2 and this issue will resolve itself.

https://www.freebsd.org/releases/11.2R/errata.html#open-issues
 
SirDice said:
Yes, this is a known issue. The 'problem' is that the binary packages are built for 11.1 because that's the lowest, still supported, version. Both the NVidia driver and the VirtualBox guest addons use a kernel driver. This kernel driver needs to be specifically built for the kernel it runs on. So the reason it's crashing is because the packages contain a 11.1 kernel module and you're trying to load that module on 11.2. If you build from ports both will use the correct kernel version and everything will work out just fine.

The "problem" is only temporary, as soon as 11.1 is EoL packages will be built for 11.2 and this issue will resolve itself.

https://www.freebsd.org/releases/11.2R/errata.html#open-issues
Click to expand...
This now makes me wonder if I’ve done something wrong, since I’ve upgraded to 11.2-RELEASE inside VirtualBox without issue, and I’m using Packages provided by FreeBSD repos.
 
You must log in or register to reply here.
Back
Top