最近我的 nVIDIA 卡在 ubuntu 21.04 中停止工作。我注意到 CPU 持续使用,并且 top 显示 modprobe 始终使用 44.5% 的 CPU。
进一步的调查又udevadm monitor
反复出现了以下场景:
UDEV [601.231746] add /bus/pci/drivers/nvidia-nvswitch (drivers)
UDEV [601.234197] remove /bus/pci/drivers/nvidia-nvswitch (drivers)
UDEV [601.314627] add /bus/pci/drivers/nvidia (drivers)
UDEV [601.332429] remove /bus/pci/drivers/nvidia (drivers)
KERNEL[601.590924] add /bus/pci/drivers/nvidia-nvswitch (drivers)
KERNEL[601.591569] add /bus/pci/drivers/nvidia (drivers)
KERNEL[601.591594] remove /bus/pci/drivers/nvidia (drivers)
KERNEL[601.592254] remove /bus/pci/drivers/nvidia-nvswitch (drivers)
dmesg 显示以下内容:
[ 942.115453] nvidia-nvlink: Nvlink Core is being initialized, major device number 511
[ 942.115458] NVRM: This is a 64-bit BAR mapped above 4GB by the system
NVRM: BIOS or the Linux kernel, but the PCI bridge
NVRM: immediately upstream of this GPU does not define
NVRM: a matching prefetchable memory window.
[ 942.116045] NVRM: This may be due to a known Linux kernel bug. Please
NVRM: see the README section on 64-bit BARs for additional
NVRM: information.
[ 942.116047] nvidia: probe of 0000:01:00.0 failed with error -1
[ 942.116069] NVRM: The NVIDIA probe routine failed for 1 device(s).
[ 942.116070] NVRM: None of the NVIDIA devices were initialized.
[ 942.116556] nvidia-nvlink: Unregistered the Nvlink Core, major device number 511
再次,这种情况不断重复。
我真的不知道该如何修复它。GPU 在 Windows 10 中运行完美。
答案1
你找到解决办法了吗?
我感觉这只会影响戴尔 G3 3590,因为我找到的所有解决方案都不起作用。我有完全相同的型号,也有同样的问题。它在 Ubuntu 20.04 上也失败了。我尝试了很多方法,花了一个多星期的时间尝试了各种方法。我发现 Mint 19.3 可以与 450 版本的驱动程序配合使用。我猜 Ubuntu 19 也会这样。在我看来,这看起来像是一个内核错误,https://ubuntu-bugs.narkive.com/eINzSDzw/bug-1742112-new-nvidia-graphics-card-failed-to-initialized
我真的很希望它能够正常工作。