[Bug]: Linux kernel 6.12.63 - VMs start, 6.18.20 they do not. Same kernel config.

[LeGrosFromage]

Version

trunk/main

Host OS Type

Linux

Host OS name + version

Gentoo, current everything

Host Architecture

x86

Guest OS Type

Windows

Guest Architecture

x86

Guest OS name + version

Windows 11

Component

Other

What happened?

Using linux kernel 6.12.63, my VMs power up and run just fine.

Using 6.18.0 or higher, the VMs instantly fail upon power on.
This happens with an existing VM or trying to create a new one.

There is nothing in dmesg and nothing obvious (to me) in the vbox.log.

The kernel config between the two is the same fundamentally, most new options are set to "N". I've tried various versions of 6.18 with no success. The modules build against 6.18 just fine and successfully load, it's just the VM refuses to power on (well, it does for one second then immediately aborts).

I can't believe it's an issue with VBox itself as I can't find anyone else having this problem, so it must be something on my setup, but I have absolutely no idea what. Sadly I can't stick with kernel 6.12.63 as I would like to start using bcachefs so need a much newer kernel.

Pointers, tips, advice appreciated.

VBox.log

How can we reproduce this?

Use linux kernel 6.12.63 - success
kernel 6.18.x - failure

Did you upload all of your necessary log files, screenshots, etc.?

• Yes, I've uploaded all pertinent files to this issue.

[galitsyn-oracle]

@LeGrosFromage ,

There is no sign that VM aborts in the VBox.log. Have you attached correct log file?

[LeGrosFromage]

Yes. It's the Vbox.log for the VM I tried to start. As soon as it tries to start, it fails, like I said I don't see anything obvious in the log - probably because nothing is making it to the log. But yes it's the correct log. Is there anything else I can provide to try to help? Sent with [Proton Mail](https://proton.me/mail/home) secure email.

…

On Wednesday, April 1st, 2026 at 06:02, Vadim Galitsin ***@***.***> wrote: galitsyn-oracle left a comment [(VirtualBox/virtualbox#609)](#609 (comment)) ***@***.***(https://github.com/LeGrosFromage) , There is no sign that VM aborts in the VBox.log. Have you attached correct log file? — Reply to this email directly, [view it on GitHub](#609?email_source=notifications&email_token=AXJNCAXS4B446YWGK3AZ25T4TTZMJA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJWHEZDMMZQGAZKM4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4169263002), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AXJNCARSAK2JLZA76PFFMND4TTZMJAVCNFSM6AAAAACXIJT6UWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNRZGI3DGMBQGI). You are receiving this because you were mentioned.Message ID: ***@***.***>

[galitsyn-oracle]

@LeGrosFromage , VM process gets terminated or VM stucks on boot? Is there anything suspicious in journalctl -b0?

[LeGrosFromage]

The process is terminated, the machine shows as "Aborted" with the "no entry" sign on it in the GUI. I'm not at that computer right now so I can't check the journalctl log at the moment but I'll get you that info as soon as I can. Sent with [Proton Mail](https://proton.me/mail/home) secure email.

…

On Wednesday, April 1st, 2026 at 10:23, Vadim Galitsin ***@***.***> wrote: galitsyn-oracle left a comment [(VirtualBox/virtualbox#609)](#609 (comment)) ***@***.***(https://github.com/LeGrosFromage) , VM process gets terminated or VM stucks on boot? Is there anything suspicious in journalctl -b0? — Reply to this email directly, [view it on GitHub](#609?email_source=notifications&email_token=AXJNCAXO6TPMM3BCPUM2LAT4TUX5VA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJXGA4DKOJWGM2KM4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4170859634), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AXJNCARUELYBYS35PTK7OOL4TUX5VAVCNFSM6AAAAACXIJT6UWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNZQHA2TSNRTGQ). You are receiving this because you were mentioned.Message ID: ***@***.***>

[galitsyn-oracle]

@LeGrosFromage , sounds like either something in the host system kills VM process or it crashes itself. Would be interesting to have a look at the backtrace. Could you please run VM process in gdb and attach backtrace?

$ gdb /usr/lib64/virtualbox/VirtualBoxVM
(gdb) run --startvm "Windows 11"
...
(gdb) thread apply all bt

[LeGrosFromage]

I will see what I can do. This is why I'm confused - using kernel 6.12.63 the VMs run fine, which leads me to believe it's not a systems issue but a kernel issue. However if it was kernel, then I'd expect to see multiple reports about it, but I'm not. I will work on the backtrace and the journalctl logs ASAP. Sent with [Proton Mail](https://proton.me/mail/home) secure email.

…

On Wednesday, April 1st, 2026 at 11:17, Vadim Galitsin ***@***.***> wrote: galitsyn-oracle left a comment [(VirtualBox/virtualbox#609)](#609 (comment)) ***@***.***(https://github.com/LeGrosFromage) , sounds like either something in the host system kills VM process or it crashes itself. Would be interesting to have a look at the backtrace. Could you please run VM process in gdb and attach backtrace? $ gdb /usr/lib64/virtualbox/VirtualBoxVM (gdb) run --startvm "Windows 11" ... (gdb) threads apply all bt — Reply to this email directly, [view it on GitHub](#609?email_source=notifications&email_token=AXJNCAQMZKG5DLUYIKIWNDL4TU6JPA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJXGEZDEMZTGI2KM4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4171223324), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AXJNCAV63IBI3YAFJ7GWFN34TU6JPAVCNFSM6AAAAACXIJT6UWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNZRGIZDGMZSGQ). You are receiving this because you were mentioned.Message ID: ***@***.***>

[LeGrosFromage]

Here is the gdb output... I was not expecting this... Still not sure why it works on one kernel and not the other though. ***@***.***:~$ gdb /usr/lib64/virtualbox/VirtualBoxVM GNU gdb (Gentoo 17.1 vanilla) 17.1 Copyright (C) 2025 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "x86_64-pc-linux-gnu". Type "show configuration" for configuration details. For bug reporting instructions, please see: <https://bugs.gentoo.org/>. Find the GDB manual and other documentation resources online at: <http://www.gnu.org/software/gdb/documentation/>. For help, type "help". Type "apropos word" to search for commands related to "word"... Reading symbols from /usr/lib64/virtualbox/VirtualBoxVM... (No debugging symbols found in /usr/lib64/virtualbox/VirtualBoxVM) (gdb) run --startvm "Windows 11" Starting program: /usr/lib64/virtualbox/VirtualBoxVM --startvm "Windows 11" [Thread debugging using libthread_db enabled] Using host libthread_db library "/usr/lib64/libthread_db.so.1". VirtualBoxVM: Error -10 in SUPR3HardenedMain! VirtualBoxVM: Effective UID is not root (euid=1026 egid=1000 uid=1026 gid=1000) VirtualBoxVM: Tip! It may help to reinstall VirtualBox. [Detaching after fork from child process 13213] [Inferior 1 (process 13210) exited with code 01] (gdb) QSettings::value: Empty key passed <html><b>Effective UID is not root (euid=1026 egid=1000 uid=1026 gid=1000) (rc=-10)</b><br/><br/>Please try reinstalling VirtualBox.<br><br><!--EOM-->where: SUPR3HardenedMain what: 2 VERR_PERMISSION_DENIED (-10) - Permission denied. </html> (gdb) threads apply all bt ❌ Undefined command: "threads". Try "help". (gdb) Sent with [Proton Mail](https://proton.me/mail/home) secure email.

…

On Wednesday, April 1st, 2026 at 11:27, Phil ***@***.***> wrote: I will see what I can do. This is why I'm confused - using kernel 6.12.63 the VMs run fine, which leads me to believe it's not a systems issue but a kernel issue. However if it was kernel, then I'd expect to see multiple reports about it, but I'm not. I will work on the backtrace and the journalctl logs ASAP. Sent with [Proton Mail](https://proton.me/mail/home) secure email. On Wednesday, April 1st, 2026 at 11:17, Vadim Galitsin ***@***.***> wrote: > galitsyn-oracle left a comment [(VirtualBox/virtualbox#609)](#609 (comment)) > > ***@***.***(https://github.com/LeGrosFromage) , sounds like either something in the host system kills VM process or it crashes itself. Would be interesting to have a look at the backtrace. Could you please run VM process in gdb and attach backtrace? > > $ gdb /usr/lib64/virtualbox/VirtualBoxVM > (gdb) run --startvm "Windows 11" > ... > (gdb) threads apply all bt > > — > Reply to this email directly, [view it on GitHub](#609?email_source=notifications&email_token=AXJNCAQMZKG5DLUYIKIWNDL4TU6JPA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJXGEZDEMZTGI2KM4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4171223324), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AXJNCAV63IBI3YAFJ7GWFN34TU6JPAVCNFSM6AAAAACXIJT6UWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNZRGIZDGMZSGQ). > You are receiving this because you were mentioned.Message ID: ***@***.***>

[galitsyn-oracle]

It should be (gdb) thread apply all bt (not threads, fixed typo in my previous comment). But I see that /usr/lib64/virtualbox/VirtualBoxVM seem does not have suid bit set (or has incorrect ownership) which looks to me as a build or packaging issue. In the end VM process cannot access /dev/vbox* nodes which seem to be the issue. You are using VirtualBox provided to your distribution, not by us. Does the issue persist if you install .run package from virtualbox.org?

[LeGrosFromage]

Im using Gentoo so using the package from the website may not be so straightforward, however I will try it. But first - I'm going to see if I can fix the permissions issue, now I've seen that at least it's a "real" error to work with and resolve before going any further. I'll redo the thread command when I'm back in front of that machine. Thank you for your guidance this far. Sent with [Proton Mail](https://proton.me/mail/home) secure email.

…

On Thursday, April 2nd, 2026 at 07:10, Vadim Galitsin ***@***.***> wrote: galitsyn-oracle left a comment [(VirtualBox/virtualbox#609)](#609 (comment)) It should be (gdb) thread apply all bt (not threads, fixed typo in my previous comment). But I see that /usr/lib64/virtualbox/VirtualBoxVM seem does not have suid bit set (or has incorrect ownership) which looks to me as a build or packaging issue. In the end VM process cannot access /dev/vbox* nodes which seem to be the issue. You are using VirtualBox provided to your distribution, not by us. Does the issue persist if you install .run packages from virtualbox.org? — Reply to this email directly, [view it on GitHub](#609?email_source=notifications&email_token=AXJNCATQ3SKIIIBCMRDY2OT4TZKCFA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJXG4ZTENRVGMY2M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4177326531), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AXJNCAVGYBFDOP5TXW3EDLD4TZKCFAVCNFSM6AAAAACXIJT6UWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNZXGMZDMNJTGE). You are receiving this because you were mentioned.Message ID: ***@***.***>

[LeGrosFromage]

Sorry this took so long. So the backtrace produced nothing - no information whatsoever. I've done a reinstall. But, I can confirm VirtualBox does have setuid: ls -l /usr/lib64/virtualbox/VirtualBoxVM -rwsr-x--- 1 root vboxusers 141424 Apr 4 05:50 /usr/lib64/virtualbox/VirtualBoxVM The /dev/vbox* access: ls -l /dev/vbox* crw------- 1 root root 10, 259 Apr 5 21:56 /dev/vboxdrv crw------- 1 root root 10, 260 Apr 5 21:56 /dev/vboxdrvu crw------- 1 root root 10, 261 Apr 5 21:56 /dev/vboxnetctl /dev/vboxusb: total 0 drwxr-x--- 2 root vboxusers 80 Apr 5 21:56 001 drwxr-x--- 2 root vboxusers 60 Apr 5 21:56 003 drwxr-x--- 2 root vboxusers 60 Apr 5 21:56 005 drwxr-x--- 2 root vboxusers 60 Apr 5 21:56 007 I'm in the vboxusers group: more /etc/group | grep vbox vboxusers:x:304:phil I asked Copilot a few questions about this, it told me there were security changes made in the kernel around 6.17 - 6.18 which could affect VB launching, but I feel if it were a major thing I wouldn't be the only one (apparently) with this problem. If it helps I can attach my working 6.12 kernel config and the non-working 6.18 one. Sent with [Proton Mail](https://proton.me/mail/home) secure email.

…

On Thursday, April 2nd, 2026 at 07:14, Phil ***@***.***> wrote: Im using Gentoo so using the package from the website may not be so straightforward, however I will try it. But first - I'm going to see if I can fix the permissions issue, now I've seen that at least it's a "real" error to work with and resolve before going any further. I'll redo the thread command when I'm back in front of that machine. Thank you for your guidance this far. Sent with [Proton Mail](https://proton.me/mail/home) secure email. On Thursday, April 2nd, 2026 at 07:10, Vadim Galitsin ***@***.***> wrote: > galitsyn-oracle left a comment [(VirtualBox/virtualbox#609)](#609 (comment)) > > It should be (gdb) thread apply all bt (not threads, fixed typo in my previous comment). But I see that /usr/lib64/virtualbox/VirtualBoxVM seem does not have suid bit set (or has incorrect ownership) which looks to me as a build or packaging issue. In the end VM process cannot access /dev/vbox* nodes which seem to be the issue. You are using VirtualBox provided to your distribution, not by us. Does the issue persist if you install .run packages from virtualbox.org? > > — > Reply to this email directly, [view it on GitHub](#609?email_source=notifications&email_token=AXJNCATQ3SKIIIBCMRDY2OT4TZKCFA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJXG4ZTENRVGMY2M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4177326531), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AXJNCAVGYBFDOP5TXW3EDLD4TZKCFAVCNFSM6AAAAACXIJT6UWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNZXGMZDMNJTGE). > You are receiving this because you were mentioned.Message ID: ***@***.***>