Organization: Red Hat UK Ltd. Registered Address: Red Hat UK Ltd, Amberley
        Place, 107-111 Peascod Street, Windsor, Berkshire, SI4 1TE, United
        Kingdom.
        Registered in England and Wales under Company Registration No. 3798903
From:   David Howells <dhowells@redhat.com>
In-Reply-To: <Y8/njtHRF/dXM+cx@infradead.org>
References: <Y8/njtHRF/dXM+cx@infradead.org> <Y8/lEVirzumLn4OG@infradead.org> <Y8/hhvfDtVcsgQd6@nvidia.com> <Y8/ZekMEAfi8VeFl@nvidia.com> <20230123173007.325544-1-dhowells@redhat.com> <20230123173007.325544-11-dhowells@redhat.com> <31f7d71d-0eb9-2250-78c0-2e8f31023c66@nvidia.com> <84721e8d-d40e-617c-b75e-ead51c3e1edf@nvidia.com> <852117.1674567983@warthog.procyon.org.uk> <852914.1674568628@warthog.procyon.org.uk> <859186.1674569545@warthog.procyon.org.uk>
To:     Christoph Hellwig <hch@infradead.org>
Cc:     dhowells@redhat.com, Jason Gunthorpe <jgg@nvidia.com>,
        John Hubbard <jhubbard@nvidia.com>,
        Al Viro <viro@zeniv.linux.org.uk>,
        Matthew Wilcox <willy@infradead.org>,
        Jens Axboe <axboe@kernel.dk>, Jan Kara <jack@suse.cz>,
        Jeff Layton <jlayton@kernel.org>,
        Logan Gunthorpe <logang@deltatee.com>,
        linux-fsdevel@vger.kernel.org, linux-block@vger.kernel.org,
        linux-kernel@vger.kernel.org, Christoph Hellwig <hch@lst.de>,
        linux-mm@kvack.org
Subject: Re: [PATCH v8 10/10] mm: Renumber FOLL_PIN and FOLL_GET down
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <863856.1674570312.1@warthog.procyon.org.uk>
Date:   Tue, 24 Jan 2023 14:25:12 +0000
Message-ID: <863857.1674570312@warthog.procyon.org.uk>
Precedence: bulk

Christoph Hellwig <hch@infradead.org> wrote:

> > > With the latest series we don't need PAGE_CLEANUP_PUT at all.
> > 
> > We will when it comes to skbuffs.
> 
> I'm a little doubtful of that.

skbuff fragments will be and will have to be a mixture of allocated pages that
need putting and pinned or non-ref'd and non-pinned zerocopy stuff.  I have
posted a patch that works for the limited amount of driverage that I use on my
test machine.

Think network filesystem messages where you have a mixture of protocol bits
generated by the kernel and data provided by direct I/O being sent by zerocopy
(libceph, for example).

David