diff options
| author | Frederik Deweerdt <deweerdt.lkml@gmail.com> | 2024-12-09 21:06:48 -0800 |
|---|---|---|
| committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2024-12-19 18:11:21 +0100 |
| commit | bc6d8cc2c5638cb7ab44f1a0717887b431616885 (patch) | |
| tree | cf724439d44c2e59e4dbe20d85f8226603cdb94c | |
| parent | 450a844c045ff0895d41b05a1cbe8febd1acfcfd (diff) | |
splice: do not checksum AF_UNIX sockets
commit 6bd8614fc2d076fc21b7488c9f279853960964e2 upstream.
When `skb_splice_from_iter` was introduced, it inadvertently added
checksumming for AF_UNIX sockets. This resulted in significant
slowdowns, for example when using sendfile over unix sockets.
Using the test code in [1] in my test setup (2G single core qemu),
the client receives a 1000M file in:
- without the patch: 1482ms (+/- 36ms)
- with the patch: 652.5ms (+/- 22.9ms)
This commit addresses the issue by marking checksumming as unnecessary in
`unix_stream_sendmsg`
Cc: stable@vger.kernel.org
Signed-off-by: Frederik Deweerdt <deweerdt.lkml@gmail.com>
Fixes: 2e910b95329c ("net: Add a function to splice pages into an skbuff for MSG_SPLICE_PAGES")
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Joe Damato <jdamato@fastly.com>
Link: https://patch.msgid.link/Z1fMaHkRf8cfubuE@xiberoa
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
| -rw-r--r-- | net/unix/af_unix.c | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index dca4429014db..ab23c8d72122 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -2219,6 +2219,7 @@ static int unix_stream_sendmsg(struct socket *sock, struct msghdr *msg, fds_sent = true; if (unlikely(msg->msg_flags & MSG_SPLICE_PAGES)) { + skb->ip_summed = CHECKSUM_UNNECESSARY; err = skb_splice_from_iter(skb, &msg->msg_iter, size, sk->sk_allocation); if (err < 0) { |