Re: [PATCH] iommufd: Enforce IOMMU_RESV_SW_MSI upon hwpt_paging allocation

From: Nicolin Chen
Date: Thu Aug 01 2024 - 13:39:46 EST


On Thu, Aug 01, 2024 at 10:28:48AM -0300, Jason Gunthorpe wrote:
> On Wed, Jul 31, 2024 at 11:23:24AM -0700, Nicolin Chen wrote:
> > > If we really want to go this route I wonder whether a per-IOMMU
> > > get_resv_regions() API should first exist then there is a clear definition
> > > what types are per-IOMMU and what are not.
> >
> > Probably it'd be necessary.
>
> But domains can cross IOMMU's as well, it is not per-iommu you are
> after but hardwired per-IOMMU driver type.
>
> What was the issue with trying to fix this on the nesting attach side?

No, seems that we all agreed on enforcing IOVAs to the parent
HWPT, so we will Just go on the nesting attach side.

Thanks
Nicolin