From: Jason Xing <kerneljasonxing@gmail.com>
To: Mat Martineau <martineau@kernel.org>
Cc: edumazet@google.com, mhiramat@kernel.org,
mathieu.desnoyers@efficios.com, rostedt@goodmis.org,
kuba@kernel.org, pabeni@redhat.com, davem@davemloft.net,
matttbe@kernel.org, geliang@kernel.org, mptcp@lists.linux.dev,
netdev@vger.kernel.org, linux-trace-kernel@vger.kernel.org,
Jason Xing <kernelxing@tencent.com>
Subject: Re: [PATCH net-next v2 5/6] mptcp: support rstreason for passive reset
Date: Fri, 5 Apr 2024 08:12:55 +0800 [thread overview]
Message-ID: <CAL+tcoCDTEov0YkeZD7B0v=TQEsfs9LtGiOge71UxUaPzWA9kQ@mail.gmail.com> (raw)
In-Reply-To: <d8fe5d37-e317-59a5-9a01-d7c6ae43be7b@kernel.org>
Hello Mat,
On Fri, Apr 5, 2024 at 4:33 AM Mat Martineau <martineau@kernel.org> wrote:
>
> On Thu, 4 Apr 2024, Jason Xing wrote:
>
> > From: Jason Xing <kernelxing@tencent.com>
> >
> > It relys on what reset options in MPTCP does as rfc8684 says. Reusing
> > this logic can save us much energy. This patch replaces all the prior
> > NOT_SPECIFIED reasons.
> >
> > Signed-off-by: Jason Xing <kernelxing@tencent.com>
> > ---
> > net/mptcp/subflow.c | 26 ++++++++++++++++++++------
> > 1 file changed, 20 insertions(+), 6 deletions(-)
> >
> > diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
> > index a68d5d0f3e2a..24668d3020aa 100644
> > --- a/net/mptcp/subflow.c
> > +++ b/net/mptcp/subflow.c
> > @@ -304,7 +304,10 @@ static struct dst_entry *subflow_v4_route_req(const struct sock *sk,
> >
> > dst_release(dst);
> > if (!req->syncookie)
> > - tcp_request_sock_ops.send_reset(sk, skb, SK_RST_REASON_NOT_SPECIFIED);
> > + /* According to RFC 8684, 3.2. Starting a New Subflow,
> > + * we should use an "MPTCP specific error" reason code.
> > + */
> > + tcp_request_sock_ops.send_reset(sk, skb, SK_RST_REASON_MPTCP_RST_EMPTCP);
>
> Hi Jason -
>
> In this case, the MPTCP reset reason is set in subflow_check_req(). Looks
> like it uses EMPTCP but that isn't guaranteed to stay the same. I think it
> would be better to extract the reset reason from the skb extension or the
> subflow context "reset_reason" field.
Good suggestions :)
>
>
> > return NULL;
> > }
> >
> > @@ -371,7 +374,10 @@ static struct dst_entry *subflow_v6_route_req(const struct sock *sk,
> >
> > dst_release(dst);
> > if (!req->syncookie)
> > - tcp6_request_sock_ops.send_reset(sk, skb, SK_RST_REASON_NOT_SPECIFIED);
> > + /* According to RFC 8684, 3.2. Starting a New Subflow,
> > + * we should use an "MPTCP specific error" reason code.
> > + */
> > + tcp6_request_sock_ops.send_reset(sk, skb, SK_RST_REASON_MPTCP_RST_EMPTCP);
>
> Same issue here.
Got it.
>
> > return NULL;
> > }
> > #endif
> > @@ -778,6 +784,7 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> > bool fallback, fallback_is_fatal;
> > struct mptcp_sock *owner;
> > struct sock *child;
> > + int reason;
> >
> > pr_debug("listener=%p, req=%p, conn=%p", listener, req, listener->conn);
> >
> > @@ -833,7 +840,8 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> > */
> > if (!ctx || fallback) {
> > if (fallback_is_fatal) {
> > - subflow_add_reset_reason(skb, MPTCP_RST_EMPTCP);
> > + reason = MPTCP_RST_EMPTCP;
> > + subflow_add_reset_reason(skb, reason);
> > goto dispose_child;
> > }
> > goto fallback;
> > @@ -861,7 +869,8 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> > } else if (ctx->mp_join) {
> > owner = subflow_req->msk;
> > if (!owner) {
> > - subflow_add_reset_reason(skb, MPTCP_RST_EPROHIBIT);
> > + reason = MPTCP_RST_EPROHIBIT;
> > + subflow_add_reset_reason(skb, reason);
> > goto dispose_child;
> > }
> >
> > @@ -875,13 +884,18 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> > ntohs(inet_sk((struct sock *)owner)->inet_sport));
> > if (!mptcp_pm_sport_in_anno_list(owner, sk)) {
> > SUBFLOW_REQ_INC_STATS(req, MPTCP_MIB_MISMATCHPORTACKRX);
> > + reason = MPTCP_RST_EUNSPEC;
>
> I think the MPTCP code here should have been using MPTCP_RST_EPROHIBIT.
I'll update in the V2 of the patch.
Thanks,
Jason
next prev parent reply other threads:[~2024-04-05 0:13 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-04 7:20 [PATCH net-next 0/6] Implement reset reason mechanism to detect Jason Xing
2024-04-04 7:20 ` [PATCH net-next v2 1/6] net: introduce rstreason to detect why the RST is sent Jason Xing
2024-04-04 7:20 ` [PATCH net-next v2 2/6] rstreason: prepare for passive reset Jason Xing
2024-04-04 7:20 ` [PATCH net-next v2 3/6] rstreason: prepare for active reset Jason Xing
2024-04-04 7:20 ` [PATCH net-next v2 4/6] tcp: support rstreason for passive reset Jason Xing
2024-04-04 7:20 ` [PATCH net-next v2 5/6] mptcp: " Jason Xing
2024-04-04 20:33 ` Mat Martineau
2024-04-05 0:12 ` Jason Xing [this message]
2024-04-04 7:20 ` [PATCH net-next v2 6/6] rstreason: make it work in trace world Jason Xing
2024-04-04 8:16 ` [PATCH net-next 0/6] Implement reset reason mechanism to detect MPTCP CI
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='CAL+tcoCDTEov0YkeZD7B0v=TQEsfs9LtGiOge71UxUaPzWA9kQ@mail.gmail.com' \
--to=kerneljasonxing@gmail.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=geliang@kernel.org \
--cc=kernelxing@tencent.com \
--cc=kuba@kernel.org \
--cc=linux-trace-kernel@vger.kernel.org \
--cc=martineau@kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=matttbe@kernel.org \
--cc=mhiramat@kernel.org \
--cc=mptcp@lists.linux.dev \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=rostedt@goodmis.org \
/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 read-only IMAP folder(s) and NNTP newsgroup(s).