From mboxrd@z Thu Jan  1 00:00:00 1970
From: John Hubbard <jhubbard@nvidia.com>
Date: Thu, 11 Feb 2021 02:26:15 +0000
Subject: Re: [PATCH 0/9] Add support for SVM atomics in Nouveau
Message-Id: <2906f445-babb-5f4e-2d99-dc004ae1face@nvidia.com>
List-Id: <kvm-ppc.vger.kernel.org>
References: <20210209010722.13839-1-apopple@nvidia.com>
 <CAKMK7uGwg2-DTU7Zrco=TSkcR4yTqN1AF0hvVYEAbuj4BUYi5Q@mail.gmail.com>
 <3426910.QXTomnrpqD@nvdebian>
 <CAKMK7uHp+BzHF1=JhKjv5HYm_j0SVqsGdRqjUxVFYx4GSEPucg@mail.gmail.com>
 <57fe0deb-8bf6-d3ee-3545-11109e946528@nvidia.com>
 <YCPYtNeYCuu6i2/d@phenom.ffwll.local>
In-Reply-To: <YCPYtNeYCuu6i2/d@phenom.ffwll.local>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: Alistair Popple <apopple@nvidia.com>, Linux MM <linux-mm@kvack.org>, Nouveau Dev <nouveau@lists.freedesktop.org>, Ben Skeggs <bskeggs@redhat.com>, Andrew Morton <akpm@linux-foundation.org>, Linux Doc Mailing List <linux-doc@vger.kernel.org>, Linux Kernel Mailing List <linux-kernel@vger.kernel.org>, kvm-ppc@vger.kernel.org, dri-devel <dri-devel@lists.freedesktop.org>, Ralph Campbell <rcampbell@nvidia.com>, Jerome Glisse <jglisse@redhat.com>, Jason Gunthorpe <jgg@ziepe.ca>

On 2/10/21 4:59 AM, Daniel Vetter wrote:
...
>> GPU atomic operations to sysmem are hard to categorize, because because application
>> programmers could easily write programs that do a long series of atomic operations.
>> Such a program would be a little weird, but it's hard to rule out.
> 
> Yeah, but we can forcefully break this whenever we feel like by revoking
> the page, moving it, and then reinstating the gpu pte again and let it
> continue.

Oh yes, that's true.

> 
> If that's no possible then what we need here instead is an mlock() type of
> thing I think.
No need for that, then.


thanks,
-- 
John Hubbard
NVIDIA