Re: [PATCH v4 00/18] PCI/P2PDMA: Fix ACS egress control handling

[Jason Gunthorpe]

On Sun, Aug 30, 2026 at 12:02:23PM +0300, Leon Romanovsky wrote:
> > I don't know how the driver facing API should work, but at least real
> > devices have options to use ATS or not, use RO or not, and can make
> > use of information from the P2P subsytem to make the right choice.
> 
> Do you see a function like mlx5_umem_needs_ats() being implemented as part of
> the PCI P2P logic?
> https://lore.kernel.org/all/4-v1-bd147097458e+ede-umem_dmabuf_jgg@xxxxxxxxxx/

Maybe someday, virtualization makes it hard since the information
needed gets wiped away, but lets start by imagining bare metal working
right.

Ideally	if the P2P route requires ATS then the driver should know and
it should setup ATS. Similarly if it doesn't then it shouldn't. The
only way to know is to evaluate the paths through the ACS controls to
figure what works.

Jason