VBox 6.1.6: Win 10 VM freezing with kernel backtrace

Discussions related to using VirtualBox on Linux hosts.
Post Reply
JannisSEA
Posts: 25
Joined: 7. Oct 2013, 14:41

VBox 6.1.6: Win 10 VM freezing with kernel backtrace

Post by JannisSEA »

Hi, I'm running virtualbox v6.1.6 (also tried v6.1.4, same behaviour) on a Gentoo Linux machine with kernel version v5.4.38 (also tried v5.4.28, same behaviour). The Extension Pack is NOT installed. The guest OS is Windows 10 64bit. I connect remotely to the VM using VNC (using VirtualBox' integrated capabilities, no VNC server running in the geust machine).
The VM boots up fine but gets stuck seconds to minutes of usage. VirtualBox still reports the machine as "Running" but is unable to stop it using the "poweroff" command. The machine then stays stuck in "Stopping" and I have to reboot the host.
In the host's kernel log, there is a backtrace following a page fault error. I've tried to find the root cause by searching for the affected function names but nothing seems to be related to or relevant for the problem I am experiencing.

Kernel backtrace:

Code: Select all

[  295.761327] BUG: unable to handle page fault for address: 00007f48dc560000
[  295.761333] #PF: supervisor read access in kernel mode
[  295.761335] #PF: error_code(0x0001) - permissions violation
...
[  295.761365] Call Trace:
[  295.761372]  ? __switch_to_asm+0x34/0x70
[  295.761375]  ? down+0xd/0x50
[  295.761377]  ? up+0xd/0x50
[  295.761383]  ? VBoxHost_RTSemFastMutexRelease+0x3e/0x50 [vboxdrv]
[  295.761385]  ? _cond_resched+0x14/0x30
[  295.761387]  ? down+0xd/0x50
[  295.761390]  ? futex_wake+0x8b/0x160
[  295.761391]  ? _cond_resched+0x14/0x30
[  295.761394]  ? __kmalloc+0x122/0x210
[  295.761397]  ? supdrvIOCtl+0xfec/0x3300 [vboxdrv]
[  295.761401]  ? rtR0MemAllocEx+0x14b/0x230 [vboxdrv]
[  295.761402]  ? do_futex+0x8b0/0xcf0
[  295.761405]  ? __check_object_size+0x153/0x164
[  295.761408]  ? _copy_from_user+0x37/0x60
[  295.761411]  ? SUPR0Printf+0x206/0x340 [vboxdrv]
[  295.761413]  ? do_vfs_ioctl+0x431/0x6a0
[  295.761415]  ? ksys_ioctl+0x46/0x90
[  295.761416]  ? __x64_sys_ioctl+0x11/0x20
[  295.761419]  ? do_syscall_64+0x43/0x110
[  295.761421]  ? entry_SYSCALL_64_after_hwframe+0x44/0xa9
I'm attaching the following files:
- VBox.log
- host's kernel log
- host's kernel configuration

I'd be happy for any hints that help me solve that problem. Thanks!
Attachments
VirtualBoxFreeze.zip
(60.07 KiB) Downloaded 29 times
mpack
Site Moderator
Posts: 39156
Joined: 4. Sep 2008, 17:09
Primary OS: MS Windows 10
VBox Version: PUEL
Guest OSses: Mostly XP

Re: VBox 6.1.6: Win 10 VM freezing with kernel backtrace

Post by mpack »

00:00:00.012672 VirtualBox VM 6.1.6_Gentoo r137129 linux.amd64 (May 7 2020 11:57:32) release log
Not VirtualBox. We don't support self-builds and third party builds here.

I will tell you that you have 5.2.22 Guest Additions installed, which naturally will not recognize new hardware, e.g. the graphics controllers introduced in v6 (VMSVGA).
JannisSEA
Posts: 25
Joined: 7. Oct 2013, 14:41

Re: VBox 6.1.6: Win 10 VM freezing with kernel backtrace

Post by JannisSEA »

Thanks mpack for your answer.

About the version of the guest additions: I agree that it's v5.x. I would also love to upgrade them. However, this is not possible as the guest is freezing.

About the version of VirtualBox used: I agree that the user-space application is not the official binary distributed by Oracle. Howeer, as far as I understand, Oracle doesn't compile and distribute the kernel driver vboxdrv as a binary. It's up to the distributions to do so since the module needs to match the kernel version and configuration. The error I am reporting is obviously in the kernel driver. I would try an Oracle-provided build if there was one.
Horwaa
Posts: 3
Joined: 11. Jul 2021, 01:25

Re: VBox 6.1.6: Win 10 VM freezing with kernel backtrace

Post by Horwaa »

I'm encountering the same issue as the OP, but am running the build from the Virtualbox repository. I recently upgraded from 16.04 (VB 5.2.22) to Ubuntu 20.04 LTS, 5.4.0-77-generic and running Virtualbox 6.1.22. Virtualbox is being run entirely in headless mode with phpVirtualBox managing VMs via web services.

I'm able to boot my Linux and Win10 guests ok, but I'm having trouble with a particular Win10 Guest. The VM boots and I'm able to access via RDP. However, when I launch chrome in the guest VM, the Linux host segfaults with the below. I've tried booting the VM with a null graphics adapter, but the same problem ensues. Could the old guest extensions be really causing an issue like this?

Any suggestions appreciated.



...
[ 440.596915] BUG: unable to handle page fault for address: 00007fcf6b140000
[ 440.596918] #PF: supervisor read access in kernel mode
[ 440.596919] #PF: error_code(0x0001) - permissions violation
[ 440.596919] PGD 8000000efffa5067 P4D 8000000efffa5067 PUD ed2467067 PMD ed2399067 PTE 8000000ec8340067
[ 440.596922] Oops: 0001 [#1] SMP PTI
[ 440.596924] CPU: 1 PID: 9674 Comm: EMT-0 Tainted: P OE 5.4.0-77-generic #86-Ubuntu
[ 440.596925] Hardware name: System manufacturer System Product Name/TUF Z390-PRO GAMING, BIOS 2820 01/06/2021
[ 440.596927] RIP: 0010:0xffffac828ef5515a
[ 440.596928] Code: 04 74 01 a5 f6 c2 02 74 02 66 a5 f6 c2 01 74 01 a4 c3 cc cc cc cc 66 2e 0f 1f 84 00 00 00 00 00 fc 48 89 d1 48 c1 e9 03 31 c0 <f3> 48 a7 75 1a f6 c2 04 74 03 a7 75 2c f6 c2 02 74 04 66 a7 75 34
[ 440.596929] RSP: 0018:ffffac828ff8fa28 EFLAGS: 00010246
[ 440.596930] RAX: 0000000000000000 RBX: ffffac828ff8fb08 RCX: 0000000000000200
[ 440.596931] RDX: 0000000000001000 RSI: 00007fcf61069000 RDI: 00007fcf6b140000
[ 440.596932] RBP: ffffac828ff8faa0 R08: 0000000000001388 R09: ffff9792123d6010
[ 440.596933] R10: ffffac82a34b2800 R11: 00000000fffff99b R12: ffffac828d091010
[ 440.596933] R13: ffffac828ff99000 R14: 0000000000000140 R15: ffff978fd5432010
[ 440.596934] FS: 00007fcfc041b700(0000) GS:ffff97936da40000(0000) knlGS:0000000000000000
[ 440.596935] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 440.596936] CR2: 00007fcf6b140000 CR3: 0000000effe62001 CR4: 00000000003626e0
[ 440.596937] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 440.596937] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 440.596938] Call Trace:
[ 440.596944] ? up+0x37/0x70
[ 440.596955] ? VBoxHost_RTSemFastMutexRelease+0x43/0x50 [vboxdrv]
[ 440.596957] ? __schedule+0x2eb/0x740
[ 440.596960] ? futex_wait+0x241/0x290
[ 440.596961] ? _cond_resched+0x19/0x30
[ 440.596964] ? __kmalloc+0x194/0x290
[ 440.596969] ? supdrvIOCtl+0xfbb/0x3660 [vboxdrv]
[ 440.596975] ? rtR0MemAllocEx+0x152/0x240 [vboxdrv]
[ 440.596976] ? __check_object_size+0x13f/0x150
[ 440.596979] ? _copy_from_user+0x3e/0x60
[ 440.596983] ? VBoxDrvLinuxIOCtl_6_1_22+0x16a/0x260 [vboxdrv]
[ 440.596986] ? do_vfs_ioctl+0x407/0x670
[ 440.596987] ? __schedule+0x2eb/0x740
[ 440.596989] ? ksys_ioctl+0x67/0x90
[ 440.596991] ? __x64_sys_ioctl+0x1a/0x20
[ 440.596993] ? do_syscall_64+0x57/0x190
[ 440.596994] ? entry_SYSCALL_64_after_hwframe+0x44/0xa9
...
Horwaa
Posts: 3
Joined: 11. Jul 2021, 01:25

Re: VBox 6.1.6: Win 10 VM freezing with kernel backtrace

Post by Horwaa »

FYI: I was able to replicate the issue on other Win10 VMs by turning "Page Fusion" on. Turning "Page Fusion" off stabilized the VMs and I upgraded the Guest Additions to match the host version. Re-enabling "Page Fusion" and using the associated Win10 guests continued to result in kernel crashes on the host. Guess I'll leave page fusion off for now...
Post Reply