Re: [RFC PATCH 3/5] mm/vma: add support for peer to peer to device vma

[Jerome Glisse <jglisse@redhat.com>]

On Wed, Jan 30, 2019 at 08:11:19PM +0000, Jason Gunthorpe wrote:
> On Wed, Jan 30, 2019 at 01:00:02PM -0700, Logan Gunthorpe wrote:
> 
> > We never changed SGLs. We still use them to pass p2pdma pages, only we
> > need to be a bit careful where we send the entire SGL. I see no reason
> > why we can't continue to be careful once their in userspace if there's
> > something in GUP to deny them.
> > 
> > It would be nice to have heterogeneous SGLs and it is something we
> > should work toward but in practice they aren't really necessary at the
> > moment.
> 
> RDMA generally cannot cope well with an API that requires homogeneous
> SGLs.. User space can construct complex MRs (particularly with the
> proposed SGL MR flow) and we must marshal that into a single SGL or
> the drivers fall apart.
> 
> Jerome explained that GPU is worse, a single VMA may have a random mix
> of CPU or device pages..
> 
> This is a pretty big blocker that would have to somehow be fixed.

Note that HMM takes care of that RDMA ODP with my ODP to HMM patch,
so what you get for an ODP umem is just a list of dma address you
can program your device to. The aim is to avoid the driver to care
about that. The access policy when the UMEM object is created by
userspace through verbs API should however ascertain that for mmap
of device file it is only creating a UMEM that is fully covered by
one and only one vma. GPU device driver will have one vma per logical
GPU object. I expect other kind of device do that same so that they
can match a vma to a unique object in their driver.

> 
> > That doesn't even necessarily need to be the case. For HMM, I
> > understand, struct pages may not point to any accessible memory and the
> > memory that backs it (or not) may change over the life time of it. So
> > they don't have to be strictly tied to BARs addresses. p2pdma pages are
> > strictly tied to BAR addresses though.
> 
> No idea, but at least for this case I don't think we need magic HMM
> pages to make simple VMA ops p2p_map/umap work..

Yes, you do not need page for simple driver, if we start creating struct
page for all PCIE BAR we are gonna waste a lot of memory and resources
for no good reason. I doubt all of the PCIE BAR of a device enabling p2p
will ever be map as p2p. So simple driver do not need struct page, GPU
driver that do not use HMM (all GPU that are more than 2 years old) do
not need struct page. Struct page is a burden here more than anything
else. I have not seen one good thing the struct page gives you.

Cheers,
Jérôme