Fatal trap 12 immediately at boot for install

terminalvark

New Member


Messages: 4

First post:rolleyes:

I wanted to try FreeBSD on my PC but it's giving me trouble.

The PC is relatively new, Gigabyte GA-Z170X-Gaming 7, 32GB DDR4 with Core i5-6600K. BIOS is set to defaults, no overclock. Memory has been tested overnight, 4 complete passes no faults. System has been stresstested with several tools, including prime. System is stable in Arch, Ubuntu and Windows 8.1.

It boots in UEFI from an M.2 SSD (NVMe/PCI-E), first SATA disk is my previous SSD Samsung 830, there are 3 HGST Ultrastars 4TB with a ZFS volume in RAIDZ1 and there's a LG bluray/DVD drive.

Arch Linux is my main system, it "controls" boot with systemd-boot for the other OS's. Preferably FreeBSD should be an option here as well, if that is possible.

I created a memstick from the download page, checked the image before dumping it to the stick, I followed the guide and created it with dd.

When it boots, I get a fatal trap 12: page fault while in kernel mode with fault code supervisor read data, page not present. The stack trace dumps 12 addresses with "??+0" appended to them. Uptime 1s :eek:.

I've been searching like crazy but I find so much, seems like this type of kernel panic is really common and generally associated with bad hardware or faulty memory. Hence I mentioned the fact that the system is very stable and checks out fine during stress tests and memory checks.

While I'm reasonably familiar with Linux (RHCSA 7 :p, RHCE 7 coming in march :D) I have never worked with *BSD. Wanted to give it a try ...

Hope someone can help me get this going. Would be greatly appreciated! :beer:
 

SirDice

Administrator
Staff member
Administrator
Moderator

Reaction score: 9,283
Messages: 33,826

Which image did you use?
 
OP
T

terminalvark

New Member


Messages: 4

Thanks for the reply.

I used the FreeBSD-10.2-RELEASE-amd64-uefi-memstick.img, as it is a UEFI machine and I'm using a USB memory stick.
 

Terry_Kennedy

Aspiring Daemon

Reaction score: 321
Messages: 941

When it boots, I get a fatal trap 12: page fault while in kernel mode with fault code supervisor read data, page not present. The stack trace dumps 12 addresses with "??+0" appended to them. Uptime 1s :eek:.

I've been searching like crazy but I find so much, seems like this type of kernel panic is really common and generally associated with bad hardware or faulty memory. Hence I mentioned the fact that the system is very stable and checks out fine during stress tests and memory checks.
Summary: If the trap 12 is always at a different address, it is almost certainly a hardware problem. If it happens at the same address, it is a bug.

Different addresses normally indicate a hardware problem. Sometimes it can be the system tripping over a bug that happened long enough ago that it appears random. The classic presentation of this is an mbuf getting clobbered, causing a NULL dereference later on. With an uptime of 1 second, this is unlikely to be what is happening here.

Crashes at the same address are almost definitely a bug. The kernel has encountered an extremely unexpected condition. If it was an expected error condition that was non-recoverable, you'd get something like "Panic: <thing> is not <condition>" or something similar.

In order to track this down, you'll need the complete panic message and backtrace. If the system doesn't have some sort of remote console support (common on servers, unusual on desktops) you may need to use something like a serial console, or maybe a video of the boot process and screengrabs (if they're clear enough to be readable).
 
OP
T

terminalvark

New Member


Messages: 4

Thanks.

It's always the same message.

I'm calling it quits though, not worth the trouble for me.

I'll install it in a virtual machine and play with that instead.
 

kpa

Beastie's Twin

Reaction score: 1,826
Messages: 6,318

UEFI support is very new in FreeBSD. You might have better luck in setting the machine to legacy (BIOS) boot and use the non-UEFI image for installing.
 
OP
T

terminalvark

New Member


Messages: 4

It's booting 4 different OS'es which all depend on UEFI already. Not sure if I can boot a legacy BIOS OS. I wIll have to figure it out if that works alongside the other UEFI OS'es.

In the meantime, FreeBSD gives me the exact same message in a Virtual Machine under VirtualBox when UEFI is enabled. See attached screenshot.
 

Attachments

bipbip

New Member


Messages: 11

Hi,

I'm front of the same issue. I upgraded my FreeBSD 10.2 (everything worked well before) and observed the same issue when I restarted.
It is always the same address with the trap. I tried to deactivate USB, network ... but nothing solve the problem. :(
So I tested with the latest image 10.2 from the website on an USB stick. Same issue like on my SSD.
Second chance, I tested with the FreeBSD 11-CURRENT (latest build: 20151130-r291495) and my server start.
The problem seems not present with FreeBSD 11-CURRENT

How can I downgrade (I suppose the kernel) or maybe upgrade to FreeBSD 11.0 (is it enough stable ?)

Or if the problem could be fixed :rolleyes:

My hardware is :
Asus AM1-A
AMD Athlon 5350
2 x 4 GB
1 x SSD 60 GB
2 x WD 1TB

Thanks in advance.

Bip
 

protocelt

Daemon

Reaction score: 413
Messages: 1,253

Hi,

I'm front of the same issue. I upgraded my FreeBSD 10.2 (everything worked well before) and observed the same issue when I restarted.
It is always the same address with the trap. I tried to deactivate USB, network ... but nothing solve the problem. :(
So I tested with the latest image 10.2 from the website on an USB stick. Same issue like on my SSD.
Second chance, I tested with the FreeBSD 11-CURRENT (latest build: 20151130-r291495) and my server start.
The problem seems not present with FreeBSD 11-CURRENT

How can I downgrade (I suppose the kernel) or maybe upgrade to FreeBSD 11.0 (is it enough stable ?)

Or if the problem could be fixed :rolleyes:

My hardware is :
Asus AM1-A
AMD Athlon 5350
2 x 4 GB
1 x SSD 60 GB
2 x WD 1TB

Thanks in advance.

Bip
If your not a developer or and advanced user, using CURRENT is almost always the wrong choice. Try a recent snapshot of 10-STABLE and see if that works for you first. STABLE has some quirks every now and again but is generally pretty stable. Keep in mind if you use STABLE or CURRENT you will have to upgrade the OS from source to keep it patched and up to date.
 

bipbip

New Member


Messages: 11

Test done with the latest snapshot of 10 STABLE and I can boot from the usk sitck without trap issue.

Now I have to upgrade my server kernel with the same from the snapshot or wait the next release ?
 

bipbip

New Member


Messages: 11

The solution has been for me to boot from the usb key (10.2 version 20151130-r291495) and copy the directory /boot/kernel to my SSD.
I don't know what was the problem but it was in the kernel.

++
 

SirDice

Administrator
Staff member
Administrator
Moderator

Reaction score: 9,283
Messages: 33,826

Now I have to upgrade my server kernel with the same from the snapshot or wait the next release ?
Keep in mind FreeBSD is a complete OS. Always keep your kernel and world in sync and on the same version.
 

bipbip

New Member


Messages: 11

I understand but if you can tell how I can find more information during the boot.
When the server boot, I can only see "Trap ..... " with memory address error.
 

bipbip

New Member


Messages: 11

After investigation and a lot of search, I definitively found my problem ... it was due to the motherboard.
I just disabled the C6 mode in the bios and now everything is back. :)
 
Top