Re: [PATCH v2 5/6] mm: track gup pages with page->dma_pinned_* fields

[John Hubbard <jhubbard@nvidia.com>]

On 07/02/2018 05:08 PM, Christopher Lameter wrote:
> On Mon, 2 Jul 2018, John Hubbard wrote:
> 
>>>
>>> These two are just wrong. You cannot make any page reference for
>>> PageDmaPinned() account against a pin count. First, it is just conceptually
>>> wrong as these references need not be long term pins, second, you can
>>> easily race like:
>>>
>>> Pinner				Random process
>>> 				get_page(page)
>>> pin_page_for_dma()
>>> 				put_page(page)
>>> 				 -> oops, page gets unpinned too early
>>>
>>
>> I'll drop this approach, without mentioning any of the locking that is hiding in
>> there, since that was probably breaking other rules anyway. :) Thanks for your
>> patience in reviewing this.
> 
> Mayb the following would work:
> 
> If you establish a reference to a page then increase the page count. If
> the reference is a dma pin action also then increase the pinned count.
> 
> That way you know how many of the references to the page are dma
> pins and you can correctly manage the state of the page if the dma pins go
> away.
> 

I think this sounds like what this patch already does, right? See:
__put_page_for_pinned_dma(), __get_page_for_pinned_dma(), and 
pin_page_for_dma(). The locking seems correct to me, but I suspect it's 
too heavyweight for such a hot path. But without adding a new put_user_page()
call, that was the best I could come up with.

What I'm hearing now from Jan and Michal is that the desired end result is
a separate API call, put_user_pages(), so that we can explicitly manage
these pinned pages.

thanks,
-- 
John Hubbard
NVIDIA