pve显卡虚拟化
Tag pve, 显卡, 虚拟化, on by view 59

家里闲置一段时间了的x79洋垃圾被我安装pve当做服务器了,但是它上面插了一张 nvidia GTX1060 6G 显卡,一直没用上,最近看到云游戏架构介绍之后,了解到显卡也可以虚拟化,决定自己动手将这张显卡用起来,在pve上虚拟化,这样一来可以在linux系统使用同时也可以在windows系统使用,让显卡发挥它的作用,以免浪费。

环境介绍

  • pve: 7.3
  • 显卡: nvidia GTX1060 6G
  • 主机平台: x79 E5 双路, 64G内存

一开始我参考这篇文章,首先在宿主机上安装必须软件以及显卡驱动。注意其中显卡驱动和mdevctl是核心服务软件。

apt update && apt install dkms  git build-essential pve-kernel-5.15 pve-headers-5.15 dkms cargo jq uuid-runtime -y
wget -P /opt/ http://ftp.br.debian.org/debian/pool/main/m/mdevctl/mdevctl_0.81-1_all.deb
dpkg -i /opt/mdevctl_0.81-1_all.deb

然后,配置内核

echo vfio >> /etc/modules 
echo vfio_iommu_type1 >> /etc/modules 
echo vfio_pci >> /etc/modules 
echo vfio_virqfd >> /etc/modules
echo  "blacklist nouveau" >> /etc/modprobe.d/blacklist.conf
echo "options kvm ignore_msrs=1" > /etc/modprobe.d/kvm.conf 

# 更新 initramfs
update-initramfs -k all -u

配置引导

#编辑grub,请不要盲目改。根据自己的环境,选择设置
nano /etc/default/grub 
#在里面找到:
GRUB_CMDLINE_LINUX_DEFAULT="quiet" 
#然后修改为:
GRUB_CMDLINE_LINUX_DEFAULT="quiet intel_iommu=on" 
#如果是amd cpu请改为:
GRUB_CMDLINE_LINUX_DEFAULT="quiet amd_iommu=on"
#更新引导
update-grub

重启宿主机一次

检查iommu是否开启成功

出现有如下iommu group说明成功
root@pve3:~# dmesg |grep iommu
[    0.000000] Command line: BOOT_IMAGE=/boot/vmlinuz-5.11.22-7-pve root=/dev/mapper/pve-root ro quiet iommu=pt intel_iommu=on
[    0.075784] Kernel command line: BOOT_IMAGE=/boot/vmlinuz-5.11.22-7-pve root=/dev/mapper/pve-root ro quiet iommu=pt intel_iommu=on
[    0.352588] iommu: Default domain type: Passthrough (set via kernel command line)
[    1.373583] pci 0000:00:00.0: Adding to iommu group 0
[    1.373592] pci 0000:00:02.0: Adding to iommu group 1
[    1.373605] pci 0000:00:14.0: Adding to iommu group 2
[    1.373613] pci 0000:00:17.0: Adding to iommu group 3
[    1.373623] pci 0000:00:1c.0: Adding to iommu group 4
[    1.373637] pci 0000:00:1d.0: Adding to iommu group 5
[    1.373647] pci 0000:00:1d.2: Adding to iommu group 6
[    1.373656] pci 0000:00:1d.3: Adding to iommu group 7
[    1.373675] pci 0000:00:1f.0: Adding to iommu group 8
[    1.373683] pci 0000:00:1f.2: Adding to iommu group 8
[    1.373691] pci 0000:00:1f.3: Adding to iommu group 8
[    1.373699] pci 0000:00:1f.4: Adding to iommu group 8
[    1.373707] pci 0000:00:1f.6: Adding to iommu group 9
[    1.373717] pci 0000:01:00.0: Adding to iommu group 10
[    1.373726] pci 0000:03:00.0: Adding to iommu group 11
[    1.373735] pci 0000:05:00.0: Adding to iommu group 12
[    1.656483]     intel_iommu=on

注意,上面日志中必须出现iommu group相关内容,否则是iommu开启失败,具体原因很可能是机器的bios中没有开启或者没有完全开启VT-d,请将VT-d及相关bios选项设置为enabled状态,重启机器后再执行上述dmesg命令查看日志是否正常。

安装驱动

# 将驱动下载至/opt目录
wget https://foxi.buduanwang.vip/pan/foxi/Virtualization/vGPU/NVIDIA-Linux-x86_64-460.73.01-grid-vgpu-kvm-v5-5.15.run -P /opt

# 给驱动添加可执行权限
chmod +x /opt/NVIDIA-Linux-x86_64-460.73.01-grid-vgpu-kvm-v5-5.15.run

# 安装
sh -c /opt/NVIDIA-Linux-x86_64-460.73.01-grid-vgpu-kvm-v5-5.15.run 

安装过程可以参考上文中所指出的引用文章

配置 vgpu_unlock

vgpu_unlock 是为消费级显卡虚拟化功能解锁,nvidia消费级显卡默认是不能开启vgpu虚拟化的,想要专业的支持虚拟化的显卡,需要购买 nvidia Tesla 等型号的显卡。我们这里是 GTX1060 所以需要使用 vgpu_unlock 进行 vpgu 解锁。

# 下载vgpu_unlock-rs版本
cd /opt/ && git clone https://github.com/mbilker/vgpu_unlock-rs.git 

# 编译
cd /opt/vgpu_unlock-rs && git checkout v2.0.1 && cargo build --release

# 安装 vgpu_unlock
cp /opt/vgpu_unlock-rs/target/release/libvgpu_unlock_rs.so /lib/nvidia/libvgpu_unlock_rs.so

重启宿主机一次

验证显卡驱动及 vgpu_unlock 是否成功。重启之后,使用nvidia-smi 确认是否如下,显示GPU信息。

➜  ~ nvidia-smi
Tue Nov 29 09:31:35 2022       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.73.01    Driver Version: 460.73.01    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  GeForce GTX 106...  On   | 00000000:03:00.0 Off |                  N/A |
| 10%   53C    P8     9W / 120W |   4084MiB /  6143MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

使用mdevctl types 验证是否出现mdev设备

➜  ~ mdevctl types
0000:03:00.0
  nvidia-156
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-2B
    Description: num_heads=4, frl_config=45, framebuffer=2048M, max_resolution=5120x2880, max_instance=12
  nvidia-215
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-2B4
    Description: num_heads=4, frl_config=45, framebuffer=2048M, max_resolution=5120x2880, max_instance=12
  nvidia-241
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-1B4
    Description: num_heads=4, frl_config=45, framebuffer=1024M, max_resolution=5120x2880, max_instance=24
  nvidia-283
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-4C
    Description: num_heads=1, frl_config=60, framebuffer=4096M, max_resolution=4096x2160, max_instance=6
  nvidia-284
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-6C
    Description: num_heads=1, frl_config=60, framebuffer=6144M, max_resolution=4096x2160, max_instance=4
  nvidia-285
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-8C
    Description: num_heads=1, frl_config=60, framebuffer=8192M, max_resolution=4096x2160, max_instance=3
  nvidia-286
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-12C
    Description: num_heads=1, frl_config=60, framebuffer=12288M, max_resolution=4096x2160, max_instance=2
  nvidia-287
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-24C
    Description: num_heads=1, frl_config=60, framebuffer=24576M, max_resolution=4096x2160, max_instance=1
  nvidia-46
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-1Q
    Description: num_heads=4, frl_config=60, framebuffer=1024M, max_resolution=5120x2880, max_instance=24
  nvidia-47
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-2Q
    Description: num_heads=4, frl_config=60, framebuffer=2048M, max_resolution=7680x4320, max_instance=12
  nvidia-48
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-3Q
    Description: num_heads=4, frl_config=60, framebuffer=3072M, max_resolution=7680x4320, max_instance=8
  nvidia-49
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-4Q
    Description: num_heads=4, frl_config=60, framebuffer=4096M, max_resolution=7680x4320, max_instance=6
  nvidia-50
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-6Q
    Description: num_heads=4, frl_config=60, framebuffer=6144M, max_resolution=7680x4320, max_instance=4
  nvidia-51
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-8Q
    Description: num_heads=4, frl_config=60, framebuffer=8192M, max_resolution=7680x4320, max_instance=3
  nvidia-52
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-12Q
    Description: num_heads=4, frl_config=60, framebuffer=12288M, max_resolution=7680x4320, max_instance=2
  nvidia-53
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-24Q
    Description: num_heads=4, frl_config=60, framebuffer=24576M, max_resolution=7680x4320, max_instance=1
  nvidia-54
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-1A
    Description: num_heads=1, frl_config=60, framebuffer=1024M, max_resolution=1280x1024, max_instance=24
  nvidia-55
    Available instances: 10
    Device API: vfio-pci
    Name: GRID P40-2A
    Description: num_heads=1, frl_config=60, framebuffer=2048M, max_resolution=1280x1024, max_instance=12
  nvidia-56
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-3A
    Description: num_heads=1, frl_config=60, framebuffer=3072M, max_resolution=1280x1024, max_instance=8
  nvidia-57
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-4A
    Description: num_heads=1, frl_config=60, framebuffer=4096M, max_resolution=1280x1024, max_instance=6
  nvidia-58
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-6A
    Description: num_heads=1, frl_config=60, framebuffer=6144M, max_resolution=1280x1024, max_instance=4
  nvidia-59
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-8A
    Description: num_heads=1, frl_config=60, framebuffer=8192M, max_resolution=1280x1024, max_instance=3
  nvidia-60
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-12A
    Description: num_heads=1, frl_config=60, framebuffer=12288M, max_resolution=1280x1024, max_instance=2
  nvidia-61
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-24A
    Description: num_heads=1, frl_config=60, framebuffer=24576M, max_resolution=1280x1024, max_instance=1
  nvidia-62
    Available instances: 0
    Device API: vfio-pci
    Name: GRID P40-1B
    Description: num_heads=4, frl_config=45, framebuffer=1024M, max_resolution=5120x2880, max_instance=24

如果上面两个验证不正常,请检查 nvidia-vgpud 服务和 nvidia-vgpu-mgr 服务的日志输出,命令如下

journalctl -u  nvidia-vgpud
journalctl -u  nvidia-vgpu-mgr

请根据日志错误自行谷歌,然后尝试修复。

配置 vgpu 参数,即前面安装的 vgpu_unlock-rs 的配置文件,路径 /etc/vgpu_unlock/profile_override.toml。内容如下

➜  ~ cat /etc/vgpu_unlock/profile_override.toml
[profile.nvidia-55]
num_displays = 1
display_width = 1920
display_height = 1080
max_pixels = 2073600
cuda_enabled = 1
frl_enabled = 0

注意:framebufferpci_idpci_device_id 这三个选项不要配置,因为这三个参数的值你拿不准,会导致后续创建虚拟机虚拟显卡后,虚拟机无法启动,常见报错如下 注意:我这里选择 nvidia-55 这个虚拟显卡,它显存是2G,我6G显卡可以虚拟出3个这个型号的显卡,你也可以选择其他显卡,具体显卡对应参数,参见上面mdevctl types的输出内容。

Input/output error Verify all devices in group 29 are bound to vfio-<bus> or pci-stub and not already in use

具体配置可以参考 vgpu_unlock-rs 项目主页

接下来就可以在pve-web上创建虚拟显卡了,首先设备选择 GTX 1060

0of8azqz

然后 MDev 选择前面 vgpu_unlock/profile_override.toml 里配置的 nvidia-55

eshvg6wq

显卡创建完毕,虚拟机开机,windows系统中设备管理器可以看到未知的显示设备,安装驱动,从这里下载,选择grid_win10原生驱动即可,也可以去NVIDIA官网下载grid驱动,但是据说虚拟机里的驱动版本不能比宿主机驱动版本高,具体我没验证。安装完毕如下

2wc85256

接下来我们在我们的 ubuntu-20.04 虚拟机上也添加一个显卡,添加步骤相同,看结果

root@ubuntu-gpu:~# lspci
00:00.0 Host bridge: Intel Corporation 440FX - 82441FX PMC [Natoma] (rev 02)
00:01.0 ISA bridge: Intel Corporation 82371SB PIIX3 ISA [Natoma/Triton II]
00:01.1 IDE interface: Intel Corporation 82371SB PIIX3 IDE [Natoma/Triton II]
00:01.2 USB controller: Intel Corporation 82371SB PIIX3 USB [Natoma/Triton II] (rev 01)
00:01.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 03)
00:02.0 VGA compatible controller: Device 1234:1111 (rev 02)
00:03.0 Unclassified device [00ff]: Red Hat, Inc. Virtio memory balloon
00:05.0 PCI bridge: Red Hat, Inc. QEMU PCI-PCI bridge
00:10.0 VGA compatible controller: NVIDIA Corporation GP102GL [Tesla P40] (rev a1)
00:12.0 Ethernet controller: Red Hat, Inc. Virtio network device
00:1e.0 PCI bridge: Red Hat, Inc. QEMU PCI-PCI bridge
00:1f.0 PCI bridge: Red Hat, Inc. QEMU PCI-PCI bridge
01:01.0 SCSI storage controller: Red Hat, Inc. Virtio SCSI

可以看到我们的NVIDIA显卡,表示虚拟显卡添加正常,驱动可以自己去官网下载安装Grid驱动。

可惜的是,显卡虚拟化成功了,但是需要买 nvidia 的 License 才能在虚拟机中用。老黄的刀法,不得不服啊……😅