Re: [PATCH 13/13] vfio/type1: Mark follow_pfn as unsafe

On Wed, Oct 07, 2020 at 08:14:06PM +0200, Daniel Vetter wrote:

Hm, but wouldn't need that the semi-nasty vma_open trick to make sure that vma doesn't untimely disappear? Or is the idea to look up the underlying vfio object, and refcount that directly?

Ah, the patches Alex was working on had the refcount I think, it does need co-ordination across multiple VFIO instances IIRC.

At least a simple check would guarentee we only have exposed PCI BAR pages which is not as bad security wise as the other stuff.

Jason