Re: [RFC PATCH 5/5] hw/arm/virt-acpi-build: Add IORT RMR regions to handle MSI nested binding

[Nicolin Chen <nicolinc@nvidia.com>]

On Thu, Nov 14, 2024 at 11:41:58AM +0100, Eric Auger wrote:
> Hi Shameer,
> 
> On 11/14/24 09:48, Shameerali Kolothum Thodi wrote:
> >
> >> -----Original Message-----
> >> From: Nicolin Chen <nicolinc@nvidia.com>
> >> Sent: Wednesday, November 13, 2024 6:31 PM
> >> To: Shameerali Kolothum Thodi <shameerali.kolothum.thodi@huawei.com>
> >> Cc: qemu-arm@nongnu.org; qemu-devel@nongnu.org;
> >> eric.auger@redhat.com; peter.maydell@linaro.org; jgg@nvidia.com;
> >> ddutile@redhat.com; Linuxarm <linuxarm@huawei.com>; Wangzhou (B)
> >> <wangzhou1@hisilicon.com>; jiangkunkun <jiangkunkun@huawei.com>;
> >> Jonathan Cameron <jonathan.cameron@huawei.com>;
> >> zhangfei.gao@linaro.org
> >> Subject: Re: [RFC PATCH 5/5] hw/arm/virt-acpi-build: Add IORT RMR regions
> >> to handle MSI nested binding
> >>
> >> On Fri, Nov 08, 2024 at 12:52:42PM +0000, Shameer Kolothum wrote:
> >>> From: Eric Auger <eric.auger@redhat.com>
> >>>
> >>> To handle SMMUv3 nested stage support it is practical to expose the
> >>> guest with reserved memory regions (RMRs) covering the IOVAs used by
> >>> the host kernel to map physical MSI doorbells.
> >> There has been an ongoing solution for MSI alternative:
> >> https://lore.kernel.org/kvm/cover.1731130093.git.nicolinc@nvidia.com/
> >>
> >> So, I think we should keep this patch out of this series, instead put it on top
> >> of the testing branch.
> > Yes. I think then we can support DT solution as well. 
> >
> > On that MSI RFC above, have you seen Eric's earlier/initial proposal to bind the Guest MSI in
> > nested cases. IIRC, it was providing an IOCTL and then creating a mapping in the host.
> >
> > I think this is the latest on that.
> > https://lore.kernel.org/linux-iommu/20210411114659.15051-4-eric.auger@redhat.com/
> yes this is the latest before I stopped my VFIO integration efforts.
> >
> > But not sure, why we then moved to RMR approach. Eric?
> 
> This was indeed the 1st integration approach. Using RMR instead was
> suggested by Jean-Philippe and I considered it as simpler (because we
> needed the SET_MSI_BINDING iotcl) so I changed the approach.

Oh, I didn't realized Eric had this..

Now, Robin wanted it back (in iommufd though), against the RMR :-/

Nicolin