On Tue, May 16, 2023 at 02:35:21PM +0000, Shameerali Kolothum Thodi wrote: > Ok. Got it. So it depends on what SMMU does for that mapping and is not > related to migration per se and has the potential to crash the system if > SMMU go ahead with that memory access. Isn't it a more generic problem > then when we have multiple devices attached to the VM?
It is, and I am hoping to solve it when I get P2P support into iommufd > I need to check if there is anything in SMMU spec that forbids this > access. There isn't, it is up to the integration how it handles 'hairpin' traffic. A good implementation will work correctly as Intel and other CPUs do. A very bad implementation will crash. Medium is to return error tlps. Jason
