You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Windows 11 VM blue-screens at boot after updating to 0.108 — fixed by disabling the VirtIO GPU driver #100
After VMPal auto-updated to 0.108 and I restarted, my Windows 11 VM blue-screened during startup and dropped to the Windows recovery screen ("Choose an option"). Around the time of the update, a menu bar notification mentioned enabling DirectX.
The guest GPU driver itself does not appear to have changed: it is still 100.6.101.58000 (dated 30/08/2026, the same version reported in #25), and "Roll Back Driver" is unavailable. So the crash looks like it comes from a host-side change in 0.108 (possibly the DirectX change) interacting with the existing driver.
Steps to Reproduce
Windows 11 VM (imported from Parallels Desktop) running fine on the previous build
VMPal updates to 0.108; restart the VM
Windows blue-screens during boot and goes to recovery
Current Behavior
Blue screen on every boot; the VM never reaches the desktop
No BugCheck event found in the System event log
Workaround: recovery → Troubleshoot → Advanced options → Startup Settings → Safe Mode, then disable "Red Hat VirtIO GPU 3D controller" in Device Manager. Windows then boots normally on the Microsoft Basic Display Adapter, without GPU acceleration.
Expected Behavior
Windows boots normally after the update with GPU acceleration working
If the GPU path fails, the VM falls back to basic display instead of blue-screening
Environment
App version: VMPal 0.108 (Build 108)
Mac: MacBook Pro 16-inch (Nov 2024), Apple M4 Max, 64 GB
macOS version: 27.0.1
VM's OS: Windows 11 25H2, 4 CPU, 12 GB RAM, imported from Parallels Desktop
Display adapter: Red Hat VirtIO GPU 3D controller, driver 100.6.101.58000 (30/08/2026)
Log
Happy to provide a minidump from C:\Windows\Minidump if one was written, or to test a preview build.
Update: Still happening after updating to the latest build. Enabling the Red Hat VirtIO GPU 3D controller from a running Windows session causes an immediate BSOD — no reboot needed to trigger it. After the restart, the device came back disabled automatically and Windows booted normally on the basic display adapter.
Bugcheck:0x0000000A IRQL_NOT_LESS_OR_EQUAL — (0x0, 0x2, 0x0, 0xfffff800ac2585c4), i.e. a NULL read at DISPATCH_LEVEL. Guest is Windows 11 ARM64, build 26100.
From the triage minidump (raw stack scan, no symbols):
The faulting address is in ntoskrnl.exe+0x2585c4, but viogpu3d.sys frames sit directly beneath it on the stack (viogpu3d+0xc480, +0xc578, +0xc5a0).
viogpu3d.sys is the last driver in the loaded-module list.
So it looks like viogpu3d.sys passes a NULL or uninitialised object into a kernel routine as it starts up.
Possibly relevant: this VM was imported from Parallels, and Parallels Tools drivers are still loaded (prl_tg.sys, prl_pl011.sys, prl_strg.sys, prl_fs.sys, prl_virt_disk_balloon.sys).
The minidump is available on request. I'm not attaching it publicly because this is a work VM.
Update: I've uninstalled Parallels Tools and rebooted. Enabling the Red Hat VirtIO GPU 3D controller still causes an immediate BSOD, so the leftover Parallels drivers aren't the cause.
Hi @aussiemike-au can you test if the issue persists on the latest VMPal build (112)? We updated the driver so hopefully that should have fixed the issue
Activity
aussiemike-au commentedon Oct 29, 2026
Update: Still happening after updating to the latest build. Enabling the Red Hat VirtIO GPU 3D controller from a running Windows session causes an immediate BSOD — no reboot needed to trigger it. After the restart, the device came back disabled automatically and Windows booted normally on the basic display adapter.
Bugcheck:
0x0000000AIRQL_NOT_LESS_OR_EQUAL —(0x0, 0x2, 0x0, 0xfffff800ac2585c4), i.e. a NULL read at DISPATCH_LEVEL. Guest is Windows 11 ARM64, build 26100.From the triage minidump (raw stack scan, no symbols):
ntoskrnl.exe+0x2585c4, butviogpu3d.sysframes sit directly beneath it on the stack (viogpu3d+0xc480,+0xc578,+0xc5a0).viogpu3d.sysis the last driver in the loaded-module list.So it looks like viogpu3d.sys passes a NULL or uninitialised object into a kernel routine as it starts up.
Possibly relevant: this VM was imported from Parallels, and Parallels Tools drivers are still loaded (
prl_tg.sys,prl_pl011.sys,prl_strg.sys,prl_fs.sys,prl_virt_disk_balloon.sys).The minidump is available on request. I'm not attaching it publicly because this is a work VM.
aussiemike-au commentedon Oct 29, 2026
Update: I've uninstalled Parallels Tools and rebooted. Enabling the Red Hat VirtIO GPU 3D controller still causes an immediate BSOD, so the leftover Parallels drivers aren't the cause.
nangtrongvuon commentedon Oct 29, 2026
Thanks for the report, we're investigating the issue, should be fixed soon.
nangtrongvuon commentedon Oct 30, 2026
Hi @aussiemike-au can you test if the issue persists on the latest VMPal build (112)? We updated the driver so hopefully that should have fixed the issue