Messages in this thread | | | From | "Tian, Kevin" <> | Subject | RE: [PATCH 7/7] iommu: Add iommu_domain::domain_ops | Date | Tue, 25 Jan 2022 00:58:04 +0000 |
| |
> From: Jason Gunthorpe via iommu > Sent: Tuesday, January 25, 2022 1:18 AM > > On Mon, Jan 24, 2022 at 09:58:18AM +0000, Tian, Kevin wrote: > > > From: Lu Baolu <baolu.lu@linux.intel.com> > > > Sent: Monday, January 24, 2022 3:11 PM > > > +/** > > > + * struct domain_ops - per-domain ops > > > + * @attach_dev: attach an iommu domain to a device > > > + * @detach_dev: detach an iommu domain from a device > > > > What is the criteria about whether an op should be iommu_ops or > domain_ops > > when it requires both domain and device pointers like above two (and > future > > PASID-based attach)? > > > > Other examples include: > > @apply_resv_region > > For apply_resv_region the 'dev' argument is really selecting a device > that is already attached to the domain, so it should be in the domain > ops.
Looks this one can be just removed. No iommu driver implements today.
> > > @is_attach_deferred > > Only two drivers implement this and neither use the domain > argument. Remove the domain argument and keep it in the iommu_ops >
Yes, especially when none of the two drivers actually uses domain.
Generally it makes sense to me by putting all attach related callbacks in iommu_ops, since they are all for building connections between device and domain while such connection is device specific (just like how a vfio device is connected to iommufd).
This in concept also applies to @enable_nesting which you mentioned in another thread. Nesting is a per-device configuration between multiple domains attached by a same device.
Thanks Kevin
| |