[PATCH mptcp-next v3 0/2] mptcp: shrink struct mptcp_options_received

[PATCH mptcp-next v3 0/2] mptcp: shrink struct mptcp_options_received

[Quanye Yang via B4 Relay]

struct mptcp_options_received is the per-packet decode buffer used while
parsing incoming MPTCP options. It is larger than needed because several
suboptions cannot be combined.

This series splits the FASTCLOSE key from rcvr_key, then overlaps the
remaining mutually exclusive payloads in an anonymous union. DSS and
rm_list stay outside the union. On x86_64 the structure shrinks from
136 to 72 bytes.

---
Changes in v3:
- Split the FASTCLOSE key change into its own commit
- Add comments linking union members to their suboptions
- Link to v2: https://patch.msgid.link/20260904-mptcp-shrink-opt-rx-v2-v2-1-d8662bb18ef2@proton.me

Changes in v2:
- Group fields by suboption and move join_id into the MP_JOIN group
- Split rcvr_key into rcvr_key (MP_CAPABLE) and fc_recv_key (MP_FASTCLOSE)
- Overlap MP_JOIN token, thmac and hmac
- Drop the dependency on the invalid-option series
- Link to v1: https://patch.msgid.link/20260904-mptcp-shrink-opt-rx-v1-1-2d87ab9505ff@proton.me

---
Quanye Yang (2):
      mptcp: split FASTCLOSE key from rcvr_key
      mptcp: shrink struct mptcp_options_received

 net/mptcp/options.c  |  6 +++---
 net/mptcp/protocol.h | 44 +++++++++++++++++++++++++++++---------------
 2 files changed, 32 insertions(+), 18 deletions(-)
---
base-commit: d5a4e55ccfb893614646d0f0e5c440daa92b7467
change-id: 20260904-mptcp-shrink-opt-rx-v2-d8e0c6951294

Best regards,
--  
Quanye Yang <quanyeyang@proton.me>

[PATCH mptcp-next v3 1/2] mptcp: split FASTCLOSE key from rcvr_key

[Quanye Yang via B4 Relay]

From: Quanye Yang <quanyeyang@proton.me>

MP_CAPABLE and MP_FASTCLOSE both stored their key in rcvr_key. Give
FASTCLOSE a dedicated fc_recv_key overlapped with rcvr_key in a union.

This helps shrink struct mptcp_options_received.

Signed-off-by: Quanye Yang <quanyeyang@proton.me>
---
 net/mptcp/options.c  | 6 +++---
 net/mptcp/protocol.h | 5 ++++-
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/net/mptcp/options.c b/net/mptcp/options.c
index f87707110c75..943b2b37eaae 100644
--- a/net/mptcp/options.c
+++ b/net/mptcp/options.c
@@ -374,10 +374,10 @@ static void mptcp_parse_option(const struct sk_buff *skb,
 			break;
 
 		ptr += 2;
-		mp_opt->rcvr_key = get_unaligned_be64(ptr);
+		mp_opt->fc_recv_key = get_unaligned_be64(ptr);
 		ptr += 8;
 		mp_opt->suboptions |= OPTION_MPTCP_FASTCLOSE;
-		pr_debug("MP_FASTCLOSE: recv_key=%llu\n", mp_opt->rcvr_key);
+		pr_debug("MP_FASTCLOSE: fc_recv_key=%llu\n", mp_opt->fc_recv_key);
 		break;
 
 	case MPTCPOPT_RST:
@@ -1259,7 +1259,7 @@ bool mptcp_incoming_options(struct sock *sk, struct sk_buff *skb)
 
 	if (unlikely(mp_opt.suboptions != OPTION_MPTCP_DSS)) {
 		if ((mp_opt.suboptions & OPTION_MPTCP_FASTCLOSE) &&
-		    READ_ONCE(msk->local_key) == mp_opt.rcvr_key) {
+		    READ_ONCE(msk->local_key) == mp_opt.fc_recv_key) {
 			WRITE_ONCE(msk->rcv_fastclose, true);
 			mptcp_schedule_work((struct sock *)msk);
 			MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_MPFASTCLOSERX);
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index b3121c8c766b..04e1180012b9 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -146,7 +146,10 @@ static inline bool before64(__u64 seq1, __u64 seq2)
 
 struct mptcp_options_received {
 	u64	sndr_key;
-	u64	rcvr_key;
+	union {
+		u64	rcvr_key;
+		u64	fc_recv_key;
+	};
 	u64	data_ack;
 	u64	data_seq;
 	u32	subflow_seq;

-- 
2.55.0

[PATCH mptcp-next v3 2/2] mptcp: shrink struct mptcp_options_received

[Quanye Yang via B4 Relay]

From: Quanye Yang <quanyeyang@proton.me>

struct mptcp_options_received is allocated on the stack while parsing
incoming MPTCP options. Several suboptions are mutually exclusive, as
enforced by mptcp_parse_option(), so their payloads can overlap.

Group fields by suboption and place the mutually exclusive payloads in
an anonymous union. Keep DSS and rm_list outside the union: they can
be combined with other suboptions. Move join_id into the MP_JOIN
group, and overlap token, thmac and hmac inside that group.

Further shrinking would require changing the parser so currently
coexisting fields (DSS mapping vs ACK, rm_list, status flags) can
overlap. That adds complexity for little gain, since the outer union is
already dominated by the MP_JOIN / ADD_ADDR members.

This reduces the structure size from 136 to 72 bytes on x86_64.

Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/625
Signed-off-by: Quanye Yang <quanyeyang@proton.me>
---
 net/mptcp/protocol.h | 45 ++++++++++++++++++++++++++++-----------------
 1 file changed, 28 insertions(+), 17 deletions(-)

diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 04e1180012b9..534d052ec007 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -145,16 +145,13 @@ static inline bool before64(__u64 seq1, __u64 seq2)
 #define after64(seq2, seq1)	before64(seq1, seq2)
 
 struct mptcp_options_received {
-	u64	sndr_key;
-	union {
-		u64	rcvr_key;
-		u64	fc_recv_key;
+	struct { /* DSS, also used by MP_CAPABLE with data */
+		u64	data_ack;
+		u64	data_seq;
+		u32	subflow_seq;
+		u16	data_len;
+		__sum16	csum;
 	};
-	u64	data_ack;
-	u64	data_seq;
-	u32	subflow_seq;
-	u16	data_len;
-	__sum16	csum;
 	struct_group(status,
 		u16 suboptions;
 		u16 use_map:1,
@@ -170,15 +167,29 @@ struct mptcp_options_received {
 		    deny_join_id0:1,
 		    __unused:2;
 	);
-	u8	join_id;
-	u32	token;
-	u32	nonce;
-	u64	thmac;
-	u8	hmac[MPTCPOPT_HMAC_LEN];
-	struct mptcp_addr_info addr;
 	struct mptcp_rm_list rm_list;
-	u64	ahmac;
-	u64	fail_seq;
+	/* Options below are mutually exclusive, see mptcp_parse_option() */
+	union {
+		struct { /* MP_CAPABLE */
+			u64	sndr_key;
+			u64	rcvr_key;
+		};
+		struct { /* MP_JOIN */
+			u32	nonce;
+			u8	join_id;
+			union {
+				u32	token;			/* SYN */
+				u64	thmac;			/* SYN + ACK */
+				u8	hmac[MPTCPOPT_HMAC_LEN];	/* ACK */
+			};
+		};
+		struct { /* ADD_ADDR */
+			struct mptcp_addr_info addr;
+			u64	ahmac;
+		};
+		u64	fail_seq;	/* MP_FAIL */
+		u64	fc_recv_key;	/* MP_FASTCLOSE */
+	};
 };
 
 static inline __be32 mptcp_option(u8 subopt, u8 len, u8 nib, u8 field)

-- 
2.55.0

Re: [PATCH mptcp-next v3 0/2] mptcp: shrink struct mptcp_options_received

[MPTCP CI]

Hi Quanye,

Thank you for your modifications, that's great!

Our CI did some validations and here is its report:

- KVM Validation: normal (except selftest_mptcp_join): Success! ✅
- KVM Validation: normal (only selftest_mptcp_join): Success! ✅
- KVM Validation: debug (except selftest_mptcp_join): Success! ✅
- KVM Validation: debug (only selftest_mptcp_join): Success! ✅
- KVM Validation: btf-normal (only bpftest_all): Success! ✅
- KVM Validation: btf-debug (only bpftest_all): Success! ✅
- Perf: 
- Task: https://github.com/multipath-tcp/mptcp_net-next/actions/runs/34305335821

Initiator: Patchew Applier
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/43ca63ab73b0
Patchwork: https://patchwork.kernel.org/project/mptcp/list/?series=1160956


If there are some issues, you can reproduce them using the same environment as
the one used by the CI thanks to a docker image, e.g.:

    $ cd [kernel source code]
    $ docker run -v "${PWD}:${PWD}:rw" -w "${PWD}" --privileged --rm -it \
        --pull always mptcp/mptcp-upstream-virtme-docker:latest \
        auto-normal

For more details:

    https://github.com/multipath-tcp/mptcp-upstream-virtme-docker


Please note that despite all the efforts that have been already done to have a
stable tests suite when executed on a public CI like here, it is possible some
reported issues are not due to your modifications. Still, do not hesitate to
help us improve that ;-)

Cheers,
MPTCP GH Action bot
Bot operated by Matthieu Baerts (NGI0 Core)

Re: [PATCH mptcp-next v3 0/2] mptcp: shrink struct mptcp_options_received

[Matthieu Baerts]

Hi Quanye,

On 09/09/2026 04:40, Quanye Yang via B4 Relay wrote:
> struct mptcp_options_received is the per-packet decode buffer used while
> parsing incoming MPTCP options. It is larger than needed because several
> suboptions cannot be combined.
> 
> This series splits the FASTCLOSE key from rcvr_key, then overlaps the
> remaining mutually exclusive payloads in an anonymous union. DSS and
> rm_list stay outside the union. On x86_64 the structure shrinks from
> 136 to 72 bytes.
> 
> ---
> Changes in v3:
> - Split the FASTCLOSE key change into its own commit
> - Add comments linking union members to their suboptions
> - Link to v2: https://patch.msgid.link/20260904-mptcp-shrink-opt-rx-v2-v2-1-d8662bb18ef2@proton.me
Thank you for the v3, it looks good to me:

Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.

Re: [PATCH mptcp-next v3 0/2] mptcp: shrink struct mptcp_options_received

[Matthieu Baerts]

Hi Quanye,

On 09/09/2026 04:40, Quanye Yang via B4 Relay wrote:
> struct mptcp_options_received is the per-packet decode buffer used while
> parsing incoming MPTCP options. It is larger than needed because several
> suboptions cannot be combined.
> 
> This series splits the FASTCLOSE key from rcvr_key, then overlaps the
> remaining mutually exclusive payloads in an anonymous union. DSS and
> rm_list stay outside the union. On x86_64 the structure shrinks from
> 136 to 72 bytes.
Now in our tree (with small changes to silence checkpatch warnings):

New patches for t/upstream:
- f23ba7b8cca7: mptcp: split FASTCLOSE key from rcvr_key
- 6ae8ca66c6ca: mptcp: shrink struct mptcp_options_received
- Results: c692d107c31c..ca7966735328 (export)

Tests are now in progress:

- export:
https://github.com/multipath-tcp/mptcp_net-next/commit/162cbd03ff40db2152751a2fd401abd33b5ba9dc/checks

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.