On Thu, 29 Sep 2022 15:32:03 +0100, Peter Xu <peterx@xxxxxxxxxx> wrote: > > > If I'm correct, Marc may be talking about SMMU, which is emulated in host > > instead of QEMU. In this case, the DMA target pages are similar to those > > pages for vgic/its tables. Both sets of pages are invisible from QEMU. > > OK, I'm not aware the SMMU can also be emulated in the host kernel, thanks > Gavin. If so then we really need to think as I mentioned above since we > will only sync this when switchover, and if the DMA range can cover a large > portion of guest mem (assuming the physical pages to DMA can be randomly > allocated by guest device drivers) then it may be another problem to cause > drastic downtime. I don't know where this is coming from. The kernel has no SMMU emulation at all. M. -- Without deviation from the norm, progress is not possible.
