RE: [PATCH 2/5] iommu: Call helper function to get assigned pasid value

[Tian, Kevin]

> From: Jason Gunthorpe <jgg@xxxxxxxx>
> Sent: Wednesday, August 9, 2023 8:35 PM
> 
> On Wed, Aug 09, 2023 at 09:49:16AM +0000, Tian, Kevin wrote:
> > > From: Baolu Lu <baolu.lu@xxxxxxxxxxxxxxx>
> > > Sent: Wednesday, August 9, 2023 8:22 AM
> > >
> > > On 2023/8/8 15:49, Tina Zhang wrote:
> > > > Use the helper function mm_get_pasid() to get the mm assigned pasid
> > > > value.
> > >
> > > For internal iommu drivers, perhaps we should use another helper.
> > > Something like sva_domain_get_pasid()?
> > >
> > > Suppose that the iommu drivers should have no idea about the "mm".
> > >
> >
> > Aren't all touched functions accept a struct mm_struct pointer?
> 
> It is wrong for the driver to even ask this question.
> 
> Domains, regardless of what they are, get attached to PASIDs. Maybe
> many PASIDs, driver doesn't get to care. SVA isn't special. Stop
> making it special.

Agree. I didn't think that far for what this series intends to achieve.

> 
> The driver should rely on there being exactly one iommu_domain for SVA
> per mm so it can hang the mm_notifier off the iommu_domain

I'm confused. Isn't this series trying to allow multiple domains per mm?

> 
> But otherwise invalidation for a SVA domain should be *exactly the
> same flow* as invalidation for a paging domain. It iterates over the
> attachments and generates the correct list of PASIDs and ATCs.
> 
> Jason