Maybe this belongs in the "Experimental" forum, but if you have the experimental siduction repo enabled, then this issue will bite you in a dist-upgrade, so ....
This report is for one system only, so it is possible the problem is not for everyone. On this system:
CPU: Quad core Intel Core i7-4770 (-HT-MCP-) cache: 8192 KB
clock speeds: max: 3900 MHz 1: 3400 MHz 2: 3811 MHz 3: 3621 MHz 4: 3637 MHz 5: 3766 MHz 6: 3541 MHz
7: 3578 MHz 8: 3714 MHz
Graphics: Card: NVIDIA GM107 [GeForce GTX 750 Ti]
Display Server: X.org 1.17.1 driver: nvidia tty size: 142x68 Advanced Data: N/A for root
The new driver appeared to build correctly during the d-u, but the module failed to load when I restarted the X server (and after removing the prior module). The kernel message in dmesg output was this:
Nvidia: unknown symbol swiotlb (err 0)
I removed the nvidia-kernel-dkms package and reinstalled it, and the module again appeared to build correctly, but would not load.
I commented out the experimental repo, removed all the nvidia-related packages, and then reinstalled 346.72-1.siduction.2, and it built and loaded correctly.
The user/experimental repo is not a repo, which should be used for normal work.
I put new drivers there, for testing.
fixes/unstable and fixes/experimental are the repos, where the working drivers can be found.
And yes, i know, the modules can't be loaded, i investigate it at the moment.
OK thanks towo. I just wanted to alert anyone who lives on the edge like me. ;D