On 20.10.20 21:44, Peter Xu wrote: > On Thu, Sep 24, 2020 at 06:04:20PM +0200, David Hildenbrand wrote: >> Implement support for sparse RAM, to be used by virtio-mem. Handling >> is somewhat-similar to memory_region_is_iommu() handling, which also >> notifies on changes. >> >> Instead of mapping the whole region, we only map selected pieces (and >> unmap previously selected pieces) when notified by the SparseRAMHandler. > > It works with vIOMMU too, right? :) As long as vfio_get_vaddr() works as > expected with virtio-mem plugging new memories, then I think the answer should > be yes. >
I haven't tried, but I guess as it's simply mapped into &address_space_memory, it should work just fine. > If it's true, maybe worth mention it somewhere either in the commit message or > in the code comment, because it seems not that obvious. I will test and add, thanks for the hint! > > So if you have plan to do some real IOs in your future tests, may also worth > try with the "-device intel-iommu" and intel_iommu=on in the guest against the > same test. Thanks ... but I have an AMD system. Will try to find out how to get that running with AMD :) -- Thanks, David / dhildenb
