From: Paolo Abeni <pabeni@redhat.com>
To: Dmytro Shytyi <dmytro@shytyi.net>, mptcp@lists.linux.dev
Subject: Re: [RFC PATCH mptcp-next v12 3/7] mptcp: add subflow_v(4,6)_send_synack()
Date: Wed, 28 Sep 2022 11:23:50 +0200 [thread overview]
Message-ID: <3bd2e2ed85386ea1716ff12d54704dc160733be2.camel@redhat.com> (raw)
In-Reply-To: <20220927225341.14165-4-dmytro@shytyi.net>
On Tue, 2022-09-27 at 22:53 +0000, Dmytro Shytyi wrote:
> In this patch we add skb to the msk, dequeue it from sk, remove TSs and
> do skb mapping.
>
> Signed-off-by: Dmytro Shytyi <dmytro@shytyi.net>
> ---
> net/ipv4/tcp_fastopen.c | 19 ++++++++++++------
> net/mptcp/fastopen.c | 43 +++++++++++++++++++++++++++++++++++++++++
> net/mptcp/protocol.c | 2 +-
> net/mptcp/protocol.h | 3 +++
> net/mptcp/subflow.c | 42 ++++++++++++++++++++++++++++++++++++++++
> 5 files changed, 102 insertions(+), 7 deletions(-)
>
> diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c
> index 45cc7f1ca296..d6b1380525ea 100644
> --- a/net/ipv4/tcp_fastopen.c
> +++ b/net/ipv4/tcp_fastopen.c
> @@ -356,13 +356,20 @@ struct sock *tcp_try_fastopen(struct sock *sk, struct sk_buff *skb,
> if (foc->len == 0) /* Client requests a cookie */
> NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPFASTOPENCOOKIEREQD);
>
> - if (!((tcp_fastopen & TFO_SERVER_ENABLE) &&
> - (syn_data || foc->len >= 0) &&
> - tcp_fastopen_queue_check(sk))) {
> - foc->len = -1;
> - return NULL;
> + if (sk_is_mptcp(sk)) {
> + if (((syn_data || foc->len >= 0) &&
> + tcp_fastopen_queue_check(sk))) {
> + foc->len = -1;
> + return NULL;
> + }
> + } else {
> + if (!((tcp_fastopen & TFO_SERVER_ENABLE) &&
> + (syn_data || foc->len >= 0) &&
> + tcp_fastopen_queue_check(sk))) {
> + foc->len = -1;
> + return NULL;
> + }
> }
This should really not be needed; with a proper setup, sock_net(sk)-
>ipv4.sysctl_tcp_fastopen/tcp_fastopen should already be
TFO_SERVER_ENABLE at this point. You can double check that with the
'perf' tool adding a probe on the relevant LoC dumping the sysctl value
(or with a possibly easier way but required a rebuild, with a temporary
printk)
> -
> if (tcp_fastopen_no_cookie(sk, dst, TFO_SERVER_COOKIE_NOT_REQD))
> goto fastopen;
>
> diff --git a/net/mptcp/fastopen.c b/net/mptcp/fastopen.c
> index d6fb45e6be4f..1d824a4d9606 100644
> --- a/net/mptcp/fastopen.c
> +++ b/net/mptcp/fastopen.c
> @@ -17,3 +17,46 @@ void mptcp_gen_msk_ackseq_fastopen(struct mptcp_sock *msk, struct mptcp_subflow_
> pr_debug("ack_seq=%llu sndr_key=%llu", msk->ack_seq, mp_opt.sndr_key);
> atomic64_set(&msk->rcv_wnd_sent, ack_seq);
> }
> +
> +void subflow_fastopen_send_synack_set_params(struct mptcp_subflow_context *subflow,
> + struct request_sock *req)
> +{
> + struct tcp_request_sock *tcp_r_sock = tcp_rsk(req);
The conventional name for a tcp_request_sock variable is 'treq'
> + struct sock *ssk = subflow->tcp_sock;
> + struct sock *sk = subflow->conn;
> + struct mptcp_sock *msk;
> + struct sk_buff *skb;
> + struct tcp_sock *tp;
> +
> + msk = mptcp_sk(sk);
> + tp = tcp_sk(ssk);
> +
> + /* <mark subflow/msk as "mptfo"> */
Please remove the '<' '>' from the comments, here and below.
> + msk->is_mptfo = 1;
> +
> + skb = skb_peek(&ssk->sk_receive_queue);
> +
> + /* <dequeue the skb from sk receive queue> */
> + __skb_unlink(skb, &ssk->sk_receive_queue);
> + skb_ext_reset(skb);
> + skb_orphan(skb);
> +
> + /* <set the skb mapping> */
> + tp->copied_seq += tp->rcv_nxt - tcp_r_sock->rcv_isn - 1;
> + subflow->map_seq = mptcp_subflow_get_mapped_dsn(subflow);
> + subflow->ssn_offset = tp->copied_seq - 1;
> +
> + /* <acquire the msk data lock> */
the above comment and the next 3 are really not needed.
> + mptcp_data_lock(sk);
> +
> + /* <move skb into msk receive queue> */
> + mptcp_set_owner_r(skb, sk);
> + __skb_queue_tail(&msk->receive_queue, skb);
> + atomic64_set(&msk->rcv_wnd_sent, mptcp_subflow_get_mapped_dsn(subflow));
I think the above statement is not needed here? we already have it in
mptcp_gen_msk_ackseq_fastopen() in the previous patch?
Instead I think we need to initialize the MPTCP_CB for skb, to avoid
random coalescing from later skbs.
> +
> + /* <call msk data_ready> */
> + (sk)->sk_data_ready(sk);
> +
> + /* <release the msk data lock> */
> + mptcp_data_unlock(sk);
> +}
> diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
> index 0db50712bad7..7e63b414011c 100644
> --- a/net/mptcp/protocol.c
> +++ b/net/mptcp/protocol.c
> @@ -206,7 +206,7 @@ static void mptcp_rfree(struct sk_buff *skb)
> mptcp_rmem_uncharge(sk, len);
> }
>
> -static void mptcp_set_owner_r(struct sk_buff *skb, struct sock *sk)
> +void mptcp_set_owner_r(struct sk_buff *skb, struct sock *sk)
> {
> skb_orphan(skb);
> skb->sk = sk;
> diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
> index a9708a2eb2bc..9c46e802a73a 100644
> --- a/net/mptcp/protocol.h
> +++ b/net/mptcp/protocol.h
> @@ -842,6 +842,9 @@ bool mptcp_userspace_pm_active(const struct mptcp_sock *msk);
> // Fast Open Mechanism functions begin
> void mptcp_gen_msk_ackseq_fastopen(struct mptcp_sock *msk, struct mptcp_subflow_context *subflow,
> struct mptcp_options_received mp_opt);
> +void mptcp_set_owner_r(struct sk_buff *skb, struct sock *sk);
> +void subflow_fastopen_send_synack_set_params(struct mptcp_subflow_context *subflow,
> + struct request_sock *req);
> // Fast Open Mechanism functions end
>
> static inline bool mptcp_pm_should_add_signal(struct mptcp_sock *msk)
> diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
> index 07dd23d0fe04..c48143bff114 100644
> --- a/net/mptcp/subflow.c
> +++ b/net/mptcp/subflow.c
> @@ -307,6 +307,46 @@ static struct dst_entry *subflow_v4_route_req(const struct sock *sk,
> return NULL;
> }
>
> +static int subflow_v4_send_synack(const struct sock *sk, struct dst_entry *dst,
> + struct flowi *fl,
> + struct request_sock *req,
> + struct tcp_fastopen_cookie *foc,
> + enum tcp_synack_type synack_type,
> + struct sk_buff *syn_skb)
> +{
> + struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
> + struct inet_request_sock *ireq = inet_rsk(req);
> +
> + /* <evenutally clear tstamp_ok, as needed depending on cookie size> */
> + if (foc)
> + ireq->tstamp_ok = 0;
I guess you really need to check the cookie size, too? If the cookie
size is small enough, stripping the timestamp should not be needed.
Also you can move the above 2 lines in
subflow_fastopen_send_synack_set_params()
Cheers,
Paolo
next prev parent reply other threads:[~2022-09-28 9:23 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-27 22:53 [RFC PATCH mptcp-next v12 0/7] mptcp: Fast Open Mechanism Dmytro Shytyi
2022-09-27 22:53 ` [RFC PATCH mptcp-next v12 1/7] mptcp: introduce MSG_FASTOPEN flag Dmytro Shytyi
2022-09-28 9:01 ` Paolo Abeni
2022-10-01 3:08 ` Dmytro Shytyi
2022-10-03 8:02 ` Paolo Abeni
2022-10-03 11:26 ` Paolo Abeni
2022-09-27 22:53 ` [RFC PATCH mptcp-next v12 2/7] mptcp: fix retrans., add mptfo vars to msk Dmytro Shytyi
2022-09-28 9:05 ` Paolo Abeni
2022-10-01 3:01 ` Dmytro Shytyi
2022-09-27 22:53 ` [RFC PATCH mptcp-next v12 3/7] mptcp: add subflow_v(4,6)_send_synack() Dmytro Shytyi
2022-09-28 9:23 ` Paolo Abeni [this message]
2022-10-01 2:59 ` Dmytro Shytyi
2022-09-27 22:53 ` [RFC PATCH mptcp-next v12 4/7] mptcp: sockopt: make 'tcp_fastopen_connect' generic Dmytro Shytyi
2022-09-28 9:26 ` Paolo Abeni
2022-10-01 2:50 ` Dmytro Shytyi
2022-09-27 22:53 ` [RFC PATCH mptcp-next v12 5/7] mptcp: add TCP_FASTOPEN_NO_COOKIE support Dmytro Shytyi
2022-09-27 22:53 ` [RFC PATCH mptcp-next v12 6/7] mptcp: add TCP_FASTOPEN option Dmytro Shytyi
2022-09-28 9:28 ` Paolo Abeni
2022-10-01 2:49 ` Dmytro Shytyi
2022-09-27 22:53 ` [RFC PATCH mptcp-next v12 7/7] selftests: mptfo initiator/listener Dmytro Shytyi
2022-09-27 23:23 ` selftests: mptfo initiator/listener: Build Failure MPTCP CI
2022-09-28 0:38 ` selftests: mptfo initiator/listener: Tests Results MPTCP CI
2022-09-28 9:45 ` [RFC PATCH mptcp-next v12 7/7] selftests: mptfo initiator/listener Paolo Abeni
2022-10-01 3:03 ` Dmytro Shytyi
2022-10-03 8:31 ` Paolo Abeni
2022-10-03 10:48 ` Dmytro Shytyi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3bd2e2ed85386ea1716ff12d54704dc160733be2.camel@redhat.com \
--to=pabeni@redhat.com \
--cc=dmytro@shytyi.net \
--cc=mptcp@lists.linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).