Re: [PATCH RFC 3/8] memory-provider: dmabuf devmem memory provider

[Jason Gunthorpe <jgg@nvidia.com>]

On Thu, Nov 16, 2023 at 07:10:01PM +0800, Yunsheng Lin wrote:
> On 2023/11/15 21:38, Jason Gunthorpe wrote:
> > On Wed, Nov 15, 2023 at 05:21:02PM +0800, Yunsheng Lin wrote:
> > 
> >>>>> I would expect net stack, page pool, driver still see the 'struct page',
> >>>>> only memory provider see the specific struct for itself, for the above,
> >>>>> devmem memory provider sees the 'struct page_pool_iov'.
> >>>>>
> >>>>> The reason I still expect driver to see the 'struct page' is that driver
> >>>>> will still need to support normal memory besides devmem.
> >>>
> >>> I wouldn't say this approach is unreasonable, but it does have to be
> >>> done carefully to isolate the mm. Keeping the struct page in the API
> >>> is going to make this very hard.
> >>
> >> I would expect that most of the isolation is done in page pool, as far as
> >> I can see:
> > 
> > It is the sort of thing that is important enough it should have
> > compiler help via types to prove that it is being done
> > properly. Otherwise it will be full of mistakes over time.
> 
> Yes, agreed.
> 
> I have done something similar as willy has done for some of
> folio conversion as below:

That is not at all what I mean, I mean you should not use
struct page * types at all in code that flows from the _iov version
except via limited accessors that can be audited and have appropriate
assertions.

Just releasing struct page * that is not a struct page * everywhere
without type safety will never be correct long term.

Jason