> -----Original Message----- > From: Eric Auger <[email protected]> > Sent: Thursday, November 28, 2024 8:07 AM > To: Zhangfei Gao <[email protected]> > Cc: Jason Gunthorpe <[email protected]>; Shameerali Kolothum Thodi > <[email protected]>; Donald Dutile > <[email protected]>; Nicolin Chen <[email protected]>; qemu- > [email protected]; [email protected]; [email protected]; > Linuxarm <[email protected]>; Wangzhou (B) > <[email protected]>; jiangkunkun <[email protected]>; > Jonathan Cameron <[email protected]> > Subject: Re: [RFC PATCH 2/5] hw/arm/smmuv3: Add initial support for > SMMUv3 Nested device > > > > On 11/28/24 04:25, Zhangfei Gao wrote: > > Hi, Eric > > > > On Thu, 28 Nov 2024 at 00:06, Eric Auger <[email protected]> wrote: > > > >>> Yeah, there is no live migration support yet in the SMMU qmeu driver, > >>> AFAIK? > >> the non accelerated SMMU QEMU device does support migration. > > Could you clarify more about this? > > The migration is not supported if using viommu (SMMU QEMU device), > isn't it? > No this is not correct. Current QEMU SMMU device *does* support > migration (see VMStateDescription) as well as qemu virtio-iommu device. > so for instance if you run a guest with smmuv3 and protected virtio-pci > devices this is supposed to be migratable. If it does not work this is > bug and this should be fixed ;-)
I think if I am right Zhangfei was testing with vfio-pci device assigned on his vSVA branch. But migration with vfio device is currently explicitly blocked if vIOMMU is present. I think Joao is working on it here[1]. But we may require additional support when we have vSVA to handle any in-flight page fault handling gracefully. Thanks, Shameer 1. https://lore.kernel.org/all/[email protected]/
