[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aKNYwu+2xMXjRkiU@lstrano-desk.jf.intel.com>
Date: Mon, 18 Aug 2025 09:45:54 -0700
From: Matthew Brost <matthew.brost@...el.com>
To: Thomas Hellström <thomas.hellstrom@...ux.intel.com>
CC: Jason Gunthorpe <jgg@...pe.ca>, <intel-xe@...ts.freedesktop.org>, "Andrew
Morton" <akpm@...ux-foundation.org>, Simona Vetter <simona.vetter@...ll.ch>,
Dave Airlie <airlied@...il.com>, <dri-devel@...ts.freedesktop.org>,
<linux-mm@...ck.org>, <linux-kernel@...r.kernel.org>, Christian
König <christian.koenig@....com>
Subject: Re: [RFC PATCH 1/6] mm/mmu_notifier: Allow multiple struct
mmu_interval_notifier passes
On Mon, Aug 18, 2025 at 06:42:36PM +0200, Thomas Hellström wrote:
> On Mon, 2025-08-18 at 13:36 -0300, Jason Gunthorpe wrote:
> > On Mon, Aug 18, 2025 at 09:25:20AM -0700, Matthew Brost wrote:
> > > I think this choice makes sense: it allows embedding the wait state
> > > from
> > > the initial notifier call into the pass structure. Patch [6] shows
> > > this
> > > by attaching the issued TLB invalidation fences to the pass. Since
> > > a
> > > single notifier may be invoked multiple times with different ranges
> > > but
> > > the same seqno,
> >
> > That should be explained, but also seems to be a bit of a different
> > issue..
> >
> > If the design is really to only have two passes and this linked list
> > is about retaining state then there should not be so much freedom to
> > have more passes.
>
> Actually the initial suggestion was two passes only. Then I thought I
> saw a use-case for even three passes and added the multi-pass thing,
> but I think it turned out we didn't have such a use-case. IMO we could
> restrict it to two-pass. Matthew, that should be completely OK for the
> SVM use-case, right?
>
Yea, I just replied that 2 passes should be sufficient.
Matt
> /Thomas
>
>
> >
> > Jason
>
Powered by blists - more mailing lists