[__jal]

Yeah, Nvidia rather sucks at cleaning up after themselves, for whatever reason.

The main reason I'm not buying another card from them is their stance towards Linux in general and their stance towards PCI virtualization in particular.

I'm currently not upgrading drivers due to the latter; they're trying to disable the capabilities (dedicating a card to a VM) that are the reason I bought a second card in the first place.

I can usually deal with either lazy or greedy, but both in conjunction is infuriating. Screw those guys.

[bonzini]

If you're talking about GPU pass-through, they're just trying to make it hard, but they're not making it impossible.

If you're talking about virtual GPUs, where one card is split across multiple VMs, however, unfortunately that's Tesla only. That said, I worked (as the maintainer of KVM) with the nVidia driver people working on vGPU, and I was very impressed. They were very knowledgeable and professional, and they managed to contribute a generic Linux framework for virtualizing PCI devices rather than a one-off hack specific to nVidia. Intel is using the same framework now, in fact.

[__jal]

I know it is not currently impossible, but I don't trust that the current "policy" will continue.

Personally, not talking about virtual GPUs. (I mean, that's cool stuff, but that's not my use case.) I'm glad you found the driver engineers to be solid. I doubt, however, the engineers are driving the decisions on what passthrough functionality Nvidia feels like allowing consumers to have this week.

Or driving decisions like this: https://devtalk.nvidia.com/default/topic/579449/linux/basemo...

I personally consider removing functionality after I purchase something to be a form of fraud. And Nvidia doesn't seem very shy about doing it. Thus, I don't trust them, and don't do business with untrustworthy vendors.

[yuhong]

I think they want to limit it to Quadros to be more precise.

[bonzini]

Yes, exactly. But they don't try _that_ hard.