Dear Maintainer,

I tried switching to the open nvidia drivers, after reading that they were now considered mature enough, and only they would support unified_shared_memory in openmp gpu offload. And indeed it mostly worked flawlessly (in the past it worked but caused overheating on my laptop). However, they reproducibly cause my computer to always hang when resuming after a suspend, with an infinite sequence of messages like this on the console:

NVRM: kbusVerifyBar2_GM107: MMUTest BAR0 window offset 0x70f000 returned garbage 0x0

The proprietary nvidia drivers suspend and resume flawlessly, and I since I have both installed in parallel thanks to the debian alternatives system, I just switched back to them for the time being. But I would be happy to be able to switch for good to the open ones, and I am willing to help providing useful info to debug this issue.

Best regards, thanks in advance
Giacomo Mulas