Bug#972514: Newest update breaks CUDA

[Joel Ray Holveck]

Thanks for the new version!  But it seems to introduce a regression.

After installing it on my bullseye box, I couldn't run CUDA programs 
anymore.  Normally, CUDA requires the nvidia-uvm module to be loaded.  
However, loading the uvm module reports the following errors in dmesg:

nvidia_uvm: module uses symbols from proprietary module nvidia, inheriting taint.
nvidia_uvm: Unknown symbol radix_tree_preloads (err -2)
nvidia_uvm: Unknown symbol set_cpus_allowed_ptr (err -2)
nvidia_uvm: Unknown symbol mmu_notifier_unregister (err -2)
nvidia_uvm: Unknown symbol __mmu_notifier_register (err -2)

I can run CUDA programs if I boot from a 580 kernel.

Do you know if I'm doing something wrong, or do I just need to wait for 
Nvidia to release a long-lived driver that supports the 5.9.x kernels?

Thanks,

joelh

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/pkg-nvidia-devel/attachments/20201103/4468a3a8/attachment.html>