# Ubuntu 18.10安装NVIDIA驱动

在MacBook Pro 2015笔记本使用的是Nvidia GeForce GT 750M显卡，默认Ubuntu 18.10安装的显卡驱动是开源的 `nouveau`。

```
01:00.0 VGA compatible controller: NVIDIA Corporation GK107M [GeForce GT 750M Mac Edition] (rev a1) (prog-if 00 [VGA controller])
    Subsystem: Apple Inc. GK107M [GeForce GT 750M Mac Edition]
    Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
    Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
    Latency: 0, Cache Line Size: 256 bytes
    Interrupt: pin A routed to IRQ 50
    Region 0: Memory at c0000000 (32-bit, non-prefetchable) [size=16M]
    Region 1: Memory at 80000000 (64-bit, prefetchable) [size=256M]
    Region 3: Memory at 90000000 (64-bit, prefetchable) [size=32M]
    Region 5: I/O ports at 1000 [size=128]
    Expansion ROM at c1000000 [disabled] [size=512K]
    Capabilities: <access denied>
    Kernel driver in use: nouveau
    Kernel modules: nvidiafb, nouveau
```

虽然`nouveau`开源驱动使用上没有太大问题，但是我发现系统日志会大量滚动输出:

```
3月 06 23:45:28 xcloud kernel: nouveau 0000:01:00.0: fifo: PBDMA0: 80006000 [GPFIFO GPPTR SIGNATURE] ch 3 [007f849000
3月 06 23:45:28 xcloud systemd-journald[563]: Missed 8 kernel messages
```

并且在使用挂起到磁盘的Hibernate功能时，图形界面会冻结无响应。

> 从 [The NVIDIA vs. Open-Source Nouveau Linux Driver Benchmarks For Summer 2018](https://www.phoronix.com/scan.php?page=article\&item=nouveau-summer-2018\&num=4) 来看，Nvdia的官方闭源驱动性能要远好于开源的nouveau。此外， [Mesa vs Nouveau vs Nvidia Proprietary](https://www.linuxquestions.org/questions/debian-26/mesa-vs-nouveau-vs-nvidia-proprietary-4175609592/) 讨论中可以看到，nouveau的性能不佳，而要使用OpenGL之类的显卡加速只能采用Nvidia官方闭源驱动。

## 自动从标准Ubuntu软件仓库安装

* 检查硬件驱动

最简单且常用的安装驱动方法是检查硬件型号并安装建议驱动，在Ubuntu中只需要执行:

```
ubuntu-drivers devices
```

此时会显示输出可用的驱动（闭源和开源），并提供了建议:

```
== /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0 ==
modalias : pci:v000010DEd00000FE9sv0000106Bsd00000130bc03sc00i00
vendor   : NVIDIA Corporation
model    : GK107M [GeForce GT 750M Mac Edition]
driver   : nvidia-driver-390 - distro non-free recommended
driver   : nvidia-340 - distro non-free
driver   : xserver-xorg-video-nouveau - distro free builtin

== /sys/devices/pci0000:00/0000:00:1c.2/0000:03:00.0 ==
modalias : pci:v000014E4d000043A0sv0000106Bsd00000134bc02sc80i00
vendor   : Broadcom Limited
model    : BCM4360 802.11ac Wireless Network Adapter
driver   : bcmwl-kernel-source - distro non-free
```

### 安装最新的驱动

* 采用 `graphics-drivers` PPA仓库，提供了beta版本的显卡驱动（nvidia-driver-418）:

> 这个软件仓库也是由ubuntu维护，但是提供了最新的Nvidia驱动

```
sudo add-apt-repository ppa:graphics-drivers/ppa
ubuntu-drivers devices
```

* 自动安装

`ubuntu-drivers`命令也提供了安装所有建议驱动的方法：

```
sudo ubuntu-drivers autoinstall
```

> 不过，目前测试升级到418版本nvidia驱动，依然没有实现 `vbetool dpms off` ，需要再探索。

安装完成后，再次检查 `lspci -vvv` 可以看到驱动替换成闭源驱动

```
01:00.0 VGA compatible controller: NVIDIA Corporation GK107M [GeForce GT 750M Mac Edition] (rev a1) (prog-if 00 [VGA controller])
    Subsystem: Apple Inc. GK107M [GeForce GT 750M Mac Edition]
    Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
    Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
    Latency: 0
    Interrupt: pin A routed to IRQ 52
    Region 0: Memory at c0000000 (32-bit, non-prefetchable) [size=16M]
    Region 1: Memory at 80000000 (64-bit, prefetchable) [size=256M]
    Region 3: Memory at 90000000 (64-bit, prefetchable) [size=32M]
    Region 5: I/O ports at 1000 [size=128]
    [virtual] Expansion ROM at c1000000 [disabled] [size=512K]
    Capabilities: <access denied>
    Kernel driver in use: nvidia
    Kernel modules: nvidiafb, nouveau, nvidia_drm, nvidia

03:00.0 Network controller: Broadcom Inc. and subsidiaries BCM4360 802.11ac Wireless Network Adapter (rev 03)
    Subsystem: Apple Inc. BCM4360 802.11ac Wireless Network Adapter
    Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
    Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort+ <TAbort- <MAbort- >SERR- <PERR- INTx-
    Latency: 0, Cache Line Size: 256 bytes
    Interrupt: pin A routed to IRQ 18
    Region 0: Memory at c1a00000 (64-bit, non-prefetchable) [size=32K]
    Region 2: Memory at c1800000 (64-bit, non-prefetchable) [size=2M]
    Capabilities: <access denied>
    Kernel driver in use: wl
    Kernel modules: bcma, wl
```

## Hibernate

在安装了Nvidia的驱动之后，尝试使用Hibernate，发现 `s2disk` 之后，再次按下电源按钮resume系统，在启动恢复磁盘镜像之后，操作系统确实不死机了（因为可以通过ssh远程登陆），但是屏幕黑屏，从远程ssh登陆检查 `top` 可以看到 Xorg出现异常的100%消耗CPU：

```
top - 10:12:11 up 18 min,  3 users,  load average: 0.97, 0.57, 0.26
Tasks: 359 total,   2 running, 357 sleeping,   0 stopped,   0 zombie
%Cpu(s): 12.5 us,  0.0 sy,  0.0 ni, 87.4 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
MiB Mem :  15948.8 total,  14183.9 free,    976.7 used,    788.2 buff/cache
MiB Swap:   9216.0 total,   9216.0 free,      0.0 used.  14640.8 avail Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
 1919 root      20   0  906564 223816  78632 R 100.0   1.4   3:18.98 Xorg
 4326 huatai    20   0   34500   4276   3396 R   0.3   0.0   0:00.28 top
    1 root      20   0  195088   9268   6660 S   0.0   0.1   0:02.72 systemd
```

另外，发现黑屏时候，键盘无响应，所以推测是Xorg相关的键盘输入设备驱动存在问题。

从 `/var/log/Xorg.0.log` 日志来看也许和modset有关

```
[    39.491] (WW) Falling back to old probe method for modesetting
[    39.491] (WW) Falling back to old probe method for fbdev
```

* 尝试关闭modeset

编辑 /etc/default/grub 添加

```
GRUB_CMDLINE_LINUX="nomodeset"
```

然后执行

```
sudo update-grub
```

不过这个方法没有生效

> Nvidia驱动还支持NVIDIA SLI FrameRendering和NVDIS Multi-GPU FrameRendering（在一块网卡上集成多个GPU核心）

另外可以尝试一下 `nvidia-settings`

### 排查Hibernate

目前没有解决NVIDIA的图形界面Hibernate，所以改为默认切换到字符界面，当需要Hibernate时候，通过 `s2disk` 来实现。

如果需要图形界面，则手工启动 startx

```
sudo systemctl set-default multi-user.target

echo "export XDG_CURRENT_DESKTOP=Budgie:GNOME" > ~/.xinitrc
echo "exec budgie-desktop" >> ~/.xinitrc
```

不过，存在的问题是启动X图形界面之后，如果使用 `s2disk` 再恢复以后Xorg会一直100%无法恢复。只有字符界面能够实现Hibernate。这个问题可能和Ubuntu 18.10的Xorg有关，我改为采用Ubuntu 18.4来避免这个问题。

#### 排查过程启用字符界面的方法

* 修改 `/etc/default/grub` 设置启动时显示终端内容

```
#GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"
#GRUB_CMDLINE_LINUX=""
GRUB_CMDLINE_LINUX_DEFAULT=""
GRUB_CMDLINE_LINUX="no_console_suspend initcall_debug"
```

* 执行更新grub::

```
update-grub
```

由于图形界面冻结，所以怀疑和显卡驱动或Xorg相关。虽然 [nouveau FAQ](https://nouveau.freedesktop.org/wiki/FAQ/#index18h3) 提到Suspend to RAM可以工作，suspend to disk应该也能够工作。

[这个问答](https://superuser.com/questions/205632/with-nouveau-driver-resume-from-hibernate-or-suspend-results-in-wrong-color-dis?rq=1) 提到nouveau驱动需要卸载掉nvidia blob并且在启动时nouveau驱动没有取得控制钱不要进行modeset。

参考 [Suspend-resume problems on Ubuntu 18.04](https://www.dell.com/community/Inspiron/Suspend-resume-problems-on-Ubuntu-18-04/td-p/6072410) 可以采用如下方法关闭nouveau的modeset

```
GRUB_CMDLINE_LINUX="no_console_suspend initcall_debug nouveau.modeset=0"
```

从 [The NVIDIA vs. Open-Source Nouveau Linux Driver Benchmarks For Summer 2018](https://www.phoronix.com/scan.php?page=article\&item=nouveau-summer-2018\&num=4) 来看，Nvdia的官方闭源驱动性能要远好于开源的nouveau。

此外， [Mesa vs Nouveau vs Nvidia Proprietary](https://www.linuxquestions.org/questions/debian-26/mesa-vs-nouveau-vs-nvidia-proprietary-4175609592/) 讨论中可以看到，nouveau的性能不佳，而要使用OpenGL之类的显卡加速只能采用Nvidia官方闭源驱动。

安装Nvida官方闭源驱动（会自动build相关nvdia模块并屏蔽掉nouveau）:

```
#列出建议驱动
ubuntu-drivers devices
#自动安装所有建议驱动
sudo ubuntu-drivers autoinstall
```

安装后重启系统。

此时Hibernate依然会导致Xorg负载100%和黑屏，重新生成Xorg配置::

```
nvidia-settings
```

并且关闭内核的modeset功能，即内核参数传递 `nomodeset`。不过，实际我没有解决，还是回退Ubuntu 18.4。

## NVIDIA Optimus

[NVIDIA Optimus](https://wiki.archlinux.org/index.php/NVIDIA_Optimus#Using_nvidia-xrun) 是一种允许Intel集成GPU和独立的NVIDIA GPU切换的技术。

不过，默认没有开启切换显卡的功能，所以以下命令

```
lspci | egrep 'VGA|3D'
```

只显示

```
01:00.0 VGA compatible controller: NVIDIA Corporation GK107M [GeForce GT 750M Mac Edition] (rev a1)
```

## 参考

* [How to install the NVIDIA drivers on Ubuntu 18.10 Cosmic Cuttlefish Linux](https://linuxconfig.org/how-to-install-the-nvidia-drivers-on-ubuntu-18-10-cosmic-cuttlefish-linux)
* [BinaryDriverHowto/Nvidia](https://help.ubuntu.com/community/BinaryDriverHowto/Nvidia)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://huataihuang.gitbook.io/cloud-atlas-draft/os/linux/ubuntu/install/install_nvidia_drivers_on_ubuntu_18_10.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.