我正在运行搭载 Ubuntu 18.04 和 NVidia Quadro T2000 的 Dell Precision 5540。
从lspci：
01:00.0 3D controller: NVIDIA Corporation TU117GLM [Quadro T2000 Mobile / Max-Q] (rev a1)
从ubuntu-drivers devices：
== /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0 ==
modalias : pci:v000010DEd00001FB8sv00001028sd00000906bc03sc02i00
vendor : NVIDIA Corporation
driver : nvidia-driver-430 - distro non-free
driver : nvidia-driver-435 - distro non-free recommended
driver : xserver-xorg-video-nouveau - distro free builtin
我的内核版本来自uname -r：
4.15.0-1079-oem
我的 Ubuntu 版本信息lsb_release -a：
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 18.04.4 LTS
Release: 18.04
Codename: bionic
当我登录并被nvidia选中nvidia-prime时（即，当加载 nvidia 驱动程序时），我可以登录，但过了一段时间后，X 服务器似乎崩溃了，冻结了 GUI。我说“冻结 GUI”，是因为如果我不管它，日志就会dmesg在后台继续增长。dmesg以下错误与冻结同时发生：
Apr 18 15:08:17 artificer kernel: NVRM: GPU at PCI:0000:01:00: GPU-54fd0cf3-82c7-1f02-af0a-5555977a0327
Apr 18 15:08:17 artificer kernel: NVRM: GPU Board Serial Number:
Apr 18 15:08:17 artificer kernel: NVRM: Xid (PCI:0000:01:00): 62, pid=617, 203c(3090) 00000000 00000000
Apr 18 15:08:17 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=617, Ch 00000000
Apr 18 15:08:17 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=617, Ch 00000001
Apr 18 15:08:17 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE) NVIDIA(0): The NVIDIA X driver has encountered an error; attempting to
Apr 18 15:08:17 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE) NVIDIA(0): recover...
Apr 18 15:08:17 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=617, Ch 00000008
Apr 18 15:08:19 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3247, Ch 00000009
Apr 18 15:08:19 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3247, Ch 0000000a
Apr 18 15:08:19 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3247, Ch 00000010
Apr 18 15:08:20 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3445, Ch 00000011
Apr 18 15:08:21 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3445, Ch 00000012
Apr 18 15:08:21 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3445, Ch 00000013
Apr 18 15:08:21 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3445, Ch 00000014
Apr 18 15:08:21 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3445, Ch 00000015
Apr 18 15:08:21 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=3445, Ch 00000018
Apr 18 15:08:21 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=4331, Ch 00000019
Apr 18 15:08:21 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=4331, Ch 0000001a
Apr 18 15:08:21 artificer kernel: NVRM: Xid (PCI:0000:01:00): 45, pid=4331, Ch 0000001b
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE) NVIDIA(GPU-0): Failed to allocate notification memory.
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE) NVIDIA(0): Failed to allocate push buffer
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE) NVIDIA(0): Error recovery failed.
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE) NVIDIA(0): *** Aborting ***
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE)
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: Fatal server error:
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE) Failed to recover from error!
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE)
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: Please consult the The X.Org Foundation support
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: at http://wiki.x.org
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: for help.
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE) Please also check the log file at "/home/brad/.local/share/xorg/Xorg.0.log" for additional information.
Apr 18 15:08:21 artificer /usr/lib/gdm3/gdm-x-session[3245]: (EE)
日志Xorg.0.log并没有提供更多信息。
[ 139.284] (EE) NVIDIA(0): The NVIDIA X driver has encountered an error; attempting to
[ 139.284] (EE) NVIDIA(0): recover...
[ 143.875] (EE) NVIDIA(GPU-0): Failed to allocate notification memory.
[ 143.875] (EE) NVIDIA(0): Failed to allocate push buffer
[ 143.875] (EE) NVIDIA(0): Error recovery failed.
[ 143.875] (EE) NVIDIA(0): *** Aborting ***
[ 143.875] (EE)
Fatal server error:
[ 143.875] (EE) Failed to recover from error!
[ 143.875] (EE)
[ 143.875] (EE)
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
[ 143.875] (EE) Please also check the log file at "/home/brad/.local/share/xorg/Xorg.0.log" for additional information.
[ 143.875] (EE)
我尝试了很多不同的解决方案，包括：
- 安装 440 驱动程序
- 安装 430 驱动程序
- 安装 435 驱动程序
- 通过安装驱动程序
sudo ubuntu-drivers autoinstall apt --purge remove 'nvidia-*每次安装前执行。
每次我做这些事情时，登录后大约两分钟内都会出现一致的“冻结”。我通常会利用这段时间在不可避免的崩溃之前运行prime-select intel，因为选择 GPU 后系统不会崩溃intel。
有人有其他想法吗？
答案1
在排除了一系列潜在的软件原因后，发现此行为是由 GPU 的物理故障引起的。完全恢复出厂设置后，冻结问题仍然存在。更换了硬件。