On Mon, Aug 11, 2025 at 7:04 PM Eugenio Perez Martin <epere...@redhat.com> wrote: > > On Mon, Aug 11, 2025 at 5:11 AM Jason Wang <jasow...@redhat.com> wrote: > > > > On Thu, Aug 7, 2025 at 7:58 PM Eugenio Pérez <epere...@redhat.com> wrote: > > > > > > Return the internal struct that represents the vq group as virtqueue map > > > token, instead of the device. > > > > Note that Michael prefers to use the iova domain. This indeed seems to > > be better. > > > > Well iova domain would delete an indirection in the pointer chase, but > it would be problematic to store the token in the caller. > > And we need to add some way to protect that the ASID of a vq group is > not changed in the middle of the operation by an ioctl. IOW, the > vq_group_internal struct pointer is constant for all the lifetime of > the device, while iova_domain is not.
I will post a new version of DMA rework and switch to using the iova domain there. Let's see if it works then. > > > > This allows the DMA functions to access > > > > s/DMA/map/ > > > > Ouch, thanks for the catch! > > > > the information per group. > > > > > > At this moment all the virtqueues share the same vq group, that only > > > can point to ASID 0. This change prepares the infrastructure for actual > > > per-group address space handling > > > > > > Signed-off-by: Eugenio Pérez <epere...@redhat.com> > > > > Thanks > > > Thanks
