I've just updated from 14.1-RELEASE-p5 to 14.1-RELEASE-p7. I use freebsd-update to manage updates and I use the official FreeBSD pkg repository, following latest. Today I upgraded packages after ~2 weeks of no upgrades. I also patched today immediately after that.
I use vm-bhyve to manage my VMs.
I recently (~2 weeks) added a GPU and pass it through to VM in question. This has worked fine since installation after an initial error43.
Full config is at the end of this post.
After rebooting after
It takes many minutes to log in (10+). After logging in, doing almost anything in the VM causes another bhyve process CPU usage spike up, and the UI completely locks up. For example, launching a browser takes 30s-60s and the entire VM is locked up. I cannot launch other programs or interact with the OS in any way. Activities that should not take 16 vCPU in the VM nevertheless spike the bhyve process to max.
Watching resource utilization in the VM with Windows Resource Monitor shows spikes of CPU, Disk, and Network that align with the frozen UI and bhyve process spike, but the Resource Monitor spike is much less time than the bhyve process spike.
I have turned off passthrough temporarily as part of my troubleshooting. I will continue troubleshooting on my own, but wonder if anyone has thoughts or similar experience to guide my troubleshooting or useful things I should explore.
Thanks!
edit: update title based on current findings
I use vm-bhyve to manage my VMs.
I recently (~2 weeks) added a GPU and pass it through to VM in question. This has worked fine since installation after an initial error43.
Full config is at the end of this post.
After rebooting after
freebsd-update, I now find that my Windows 10 VM is behaving oddly. There are 16 host threads for the VM. The VM drive is encrypted with BitLocker. It boots and decrypts without issue. Upon attempting to log in to the VM through VNC or RDP (I use remmina and xfreerdp), the bhyve process shoots to 1500%-1600% (as observed in top, always blocked on kqread). Note that this matches the host threads for the VM.It takes many minutes to log in (10+). After logging in, doing almost anything in the VM causes another bhyve process CPU usage spike up, and the UI completely locks up. For example, launching a browser takes 30s-60s and the entire VM is locked up. I cannot launch other programs or interact with the OS in any way. Activities that should not take 16 vCPU in the VM nevertheless spike the bhyve process to max.
Watching resource utilization in the VM with Windows Resource Monitor shows spikes of CPU, Disk, and Network that align with the frozen UI and bhyve process spike, but the Resource Monitor spike is much less time than the bhyve process spike.
I have turned off passthrough temporarily as part of my troubleshooting. I will continue troubleshooting on my own, but wonder if anyone has thoughts or similar experience to guide my troubleshooting or useful things I should explore.
Thanks!
Code:
loader="uefi"
graphics="yes"
xhci_mouse="yes"
cpu=16
cpu_sockets=1
cpu_cores=8
cpu_threads=2
memory=32G
# put up to 8 disks on a single ahci controller.
# without this, adding a disk pushes the following network devices onto higher slot numbers,
# which causes windows to see them as a new interface
ahci_device_limit="8"
# ideally this should be changed to virtio-net and drivers installed in the guest
# e1000 works out-of-the-box
network0_type="e1000"
network0_switch="public"
disk0_type="nvme"
disk0_name="disk0.img"
disk0_dev="zvol"
#disk1_type="ahci-cd"
#disk1_dev="custom"
#disk1_name="/zroot/vms/isos/windows-server-2019.iso"
# windows expects the host to expose localtime by default, not UTC
utctime="no"
uuid="b87bf82b-b896-4dff-b0b7-4b203670b37d"
network0_mac="58:9c:fc:0b:6c:53"
graphics_port="5900"
graphics_listen="0.0.0.0"
graphics_res="1920x1080"
graphics_wait="auto"
#passthru0="35/0/0=2:0"edit: update title based on current findings
Last edited: