[PATCH v6 1/4] RDMA/umem: Support importing dma-buf as user memory region

[Xiong, Jianxin]

> -----Original Message-----
> From: Christoph Hellwig <hch at infradead.org>
> Sent: Tuesday, October 27, 2020 1:08 AM
> To: Jason Gunthorpe <jgg at ziepe.ca>
> Cc: Christoph Hellwig <hch at infradead.org>; Daniel Vetter <daniel at ffwll.ch>; Xiong, Jianxin <jianxin.xiong at intel.com>; linux-
> rdma at vger.kernel.org; dri-devel at lists.freedesktop.org; Leon Romanovsky <leon at kernel.org>; Doug Ledford <dledford at redhat.com>;
> Vetter, Daniel <daniel.vetter at intel.com>; Christian Koenig <christian.koenig at amd.com>
> Subject: Re: [PATCH v6 1/4] RDMA/umem: Support importing dma-buf as user memory region
> 
> On Mon, Oct 26, 2020 at 09:26:37AM -0300, Jason Gunthorpe wrote:
> > On Sat, Oct 24, 2020 at 08:48:07AM +0100, Christoph Hellwig wrote:
> > > On Fri, Oct 23, 2020 at 03:20:05PM -0300, Jason Gunthorpe wrote:
> > > > The problem is we have RDMA drivers that assume SGL's have a valid
> > > > struct page, and these hacky/wrong P2P sgls that DMABUF creates
> > > > cannot be passed into those drivers.
> > >
> > > RDMA drivers do not assume scatterlist have a valid struct page,
> > > scatterlists are defined to have a valid struct page.  Any
> > > scatterlist without a struct page is completely buggy.
> >
> > It is not just having the struct page, it needs to be a CPU accessible
> > one for memcpy/etc. They aren't correct with the
> > MEMORY_DEVICE_PCI_P2PDMA SGLs either.
> 
> Exactly.

In the function ib_umem_dmabuf_sgt_slice() (part of this patch) we could generate
a dma address array instead of filling the scatterlist 'umem->sg_head'. The array
would be handled similar to 'umem_odp->dma_list'. With such change, the RDMA
drivers wouldn't see incorrectly formed scatterlist. The check for dma_virt_ops here
wouldn't be needed either.

Would such proposal address the concern here?

-Jianxin