[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210210175913.GO4718@ziepe.ca>
Date: Wed, 10 Feb 2021 13:59:13 -0400
From: Jason Gunthorpe <jgg@...pe.ca>
To: John Hubbard <jhubbard@...dia.com>
Cc: Daniel Vetter <daniel@...ll.ch>,
Alistair Popple <apopple@...dia.com>,
Linux MM <linux-mm@...ck.org>,
Nouveau Dev <nouveau@...ts.freedesktop.org>,
Ben Skeggs <bskeggs@...hat.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Linux Doc Mailing List <linux-doc@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
kvm-ppc@...r.kernel.org,
dri-devel <dri-devel@...ts.freedesktop.org>,
Ralph Campbell <rcampbell@...dia.com>,
Jerome Glisse <jglisse@...hat.com>
Subject: Re: [PATCH 0/9] Add support for SVM atomics in Nouveau
On Tue, Feb 09, 2021 at 12:53:27PM -0800, John Hubbard wrote:
> This direction sounds at least...possible. Using MMU notifiers instead of pins
> is definitely appealing. I'm not quite clear on the callback idea above, but
> overall it seems like taking advantage of the ZONE_DEVICE tracking of pages
> (without having to put anything additional in each struct page), could work.
It isn't the ZONE_DEVICE page that needs to be tracked.
Really what you want to do here is leave the CPU page in the VMA and
the page tables where it started and deny CPU access to the page. Then
all the proper machinery will continue to work.
IMHO "migration" is the wrong idea if the data isn't actually moving.
Jason
Powered by blists - more mailing lists