Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230406094245.3633290-1-dhowells@redhat.com> <20230406094245.3633290-7-dhowells@redhat.com>
In-Reply-To: <20230406094245.3633290-7-dhowells@redhat.com>
From:   Willem de Bruijn <willemdebruijn.kernel@gmail.com>
Date:   Thu, 6 Apr 2023 22:01:42 -0400
Message-ID: <CAF=yD-LAC4QCfoGVKaW-GzU26=xp-6Wuq3jxAhJK1+KV0M+q2A@mail.gmail.com>
Subject: Re: [PATCH net-next v5 06/19] tcp: Make sendmsg(MSG_SPLICE_PAGES)
 copy unspliceable data
To:     David Howells <dhowells@redhat.com>
Cc:     netdev@vger.kernel.org, "David S. Miller" <davem@davemloft.net>,
        Eric Dumazet <edumazet@google.com>,
        Jakub Kicinski <kuba@kernel.org>,
        Paolo Abeni <pabeni@redhat.com>,
        Matthew Wilcox <willy@infradead.org>,
        Al Viro <viro@zeniv.linux.org.uk>,
        Christoph Hellwig <hch@infradead.org>,
        Jens Axboe <axboe@kernel.dk>, Jeff Layton <jlayton@kernel.org>,
        Christian Brauner <brauner@kernel.org>,
        Chuck Lever III <chuck.lever@oracle.com>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-mm@kvack.org, David Ahern <dsahern@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Apr 6, 2023 at 5:43=E2=80=AFAM David Howells <dhowells@redhat.com> =
wrote:
>
> If sendmsg() with MSG_SPLICE_PAGES encounters a page that shouldn't be
> spliced - a slab page, for instance, or one with a zero count - make
> tcp_sendmsg() copy it.
>
> Signed-off-by: David Howells <dhowells@redhat.com>
> cc: Eric Dumazet <edumazet@google.com>
> cc: "David S. Miller" <davem@davemloft.net>
> cc: David Ahern <dsahern@kernel.org>
> cc: Jakub Kicinski <kuba@kernel.org>
> cc: Paolo Abeni <pabeni@redhat.com>
> cc: Jens Axboe <axboe@kernel.dk>
> cc: Matthew Wilcox <willy@infradead.org>
> cc: netdev@vger.kernel.org
> ---
>  net/ipv4/tcp.c | 28 +++++++++++++++++++++++++---
>  1 file changed, 25 insertions(+), 3 deletions(-)
>
> diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
> index 510bacc7ce7b..238a8ad6527c 100644
> --- a/net/ipv4/tcp.c
> +++ b/net/ipv4/tcp.c
> @@ -1418,10 +1418,10 @@ int tcp_sendmsg_locked(struct sock *sk, struct ms=
ghdr *msg, size_t size)
>                                 goto do_error;
>                         copy =3D err;
>                 } else if (zc =3D=3D 2) {
> -                       /* Splice in data. */
> +                       /* Splice in data if we can; copy if we can't. */
>                         struct page *page =3D NULL, **pages =3D &page;
>                         size_t off =3D 0, part;
> -                       bool can_coalesce;
> +                       bool can_coalesce, put =3D false;
>                         int i =3D skb_shinfo(skb)->nr_frags;
>
>                         copy =3D iov_iter_extract_pages(&msg->msg_iter, &=
pages,
> @@ -1448,12 +1448,34 @@ int tcp_sendmsg_locked(struct sock *sk, struct ms=
ghdr *msg, size_t size)
>                                 goto wait_for_space;
>                         copy =3D part;
>
> +                       if (!sendpage_ok(page)) {
> +                               const void *p =3D kmap_local_page(page);
> +                               void *q;
> +
> +                               q =3D page_frag_memdup(NULL, p + off, cop=
y,
> +                                                    sk->sk_allocation, U=
LONG_MAX);
> +                               kunmap_local(p);
> +                               if (!q) {
> +                                       iov_iter_revert(&msg->msg_iter, c=
opy);
> +                                       err =3D copy ?: -ENOMEM;
> +                                       goto do_error;
> +                               }
> +                               page =3D virt_to_page(q);
> +                               off =3D offset_in_page(q);
> +                               put =3D true;
> +                               can_coalesce =3D false;
> +                       }
> +

This is almost identical in the later udp and unix implementations.
Could this be a wrapper, something like

    page =3D sendpage_copy_if_needed(&page, &off, copy, gfp, &put));

(it seems page is never needed if it would return NULL)