[RFC mptcp-next v20 00/15] MPTCP KTLS support

[RFC mptcp-next v20 00/15] MPTCP KTLS support

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

v20:
- Patch 1:
  - Use kfree_rcu() instead of synchronous kfree() in tls_proto_cleanup()
    and tls_proto_put() to prevent use-after-free for RCU readers
  - Add duplicate check under tls_proto_lock before inserting new tls_proto
    to list to prevent race condition
  - Fix reference counting: drop the initial reference from tls_build_proto()
    after successful tls_ctx_create() to prevent refcount leak
- Patch 2:
  - Replace spin_lock() with spin_lock_bh() in tls_register_prot_ops()
    and tls_unregister_prot_ops() to avoid Lockdep splat (softirq context
    usage)
- Patch 3:
  - Add memory barrier ordering in tls_sk_poll() to prevent NULL pointer
    dereference during concurrent setsockopt(TCP_ULP)
  - Move try_module_get() inside tls_prot_ops_find() while holding the lock
    to prevent ops structure from being freed before module reference is
    acquired
- Patch 8:
  - Add missing fallback check __mptcp_check_fallback(msk) before calling
    mptcp_setsockopt_tcp_ulp() to prevent TLS being attached to fallback
    sockets
  - Re-check fallback status inside mptcp_setsockopt_tcp_ulp() after acquiring
    socket lock to eliminate TOCTOU race
- Patch 10:
  - Remove unused _metadata parameter from is_mptcp_enable()

v19:
- Remove RCU from tls_proto and tls_prot_ops management
  Replace RCU with spinlock (tls_proto_lock) for all list operations.
  Remove rcu_head from struct tls_proto, eliminate call_rcu() and
  synchronize_rcu() calls. Use list_del + kfree directly in
  tls_proto_cleanup() and tls_proto_put().
- Add memory barrier in update_sk_prot()
  Use smp_store_release() when setting sk->sk_socket->ops to ensure it is
  not visible before icsk_ulp_data. Add explanatory comment to prevent
  NULL pointer dereference in concurrent tls_sk_poll() due to CPU
  reordering.
- Add module owner field to tls_prot_ops
  Add .owner = THIS_MODULE to tls_tcp_ops (and tls_mptcp_ops in MPTCP
  side). Add try_module_get(ops->owner) in tls_build_proto() before
  caching ops pointer. Add module_put(proto->ops->owner) in
  tls_proto_put() when releasing proto.
- Fix tls_toe_sk_destruct use-after-free
  Restore sk->sk_prot before calling original destructor to prevent UAF
  when __sk_destruct() accesses sk->sk_prot->destroy.
- Optimize indirect call performance
  Delay inq and get_skb_seq indirect calls in tls_read_flush_backlog()
  and tls_rx_msg_size() to avoid unnecessary retpoline overhead on fast
  path.
- Fix shutdown_reuse test flakiness
  Replace MPTCP-specific wait with generic bind retry loop. Retry bind()
  on EINVAL up to 1000 times (1 second) to handle asynchronous state
  transitions for both TCP and MPTCP.
- Remove redundant tls_ctx->proto->ops NULL checks
  Rely on memory barrier to guarantee ordering, eliminate unnecessary
  NULL checks that could introduce other issues.
- Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1779355169.git.tanggeliang@kylinos.cn/

v18:
 - add ip_ver field to tls_proto and pass it to tls_proto_find to
   distinguish IPv4/IPv6 instances   
 - convert tls_proto_mutex to spinlock, and convert tls_proto_cleanup
   to use call_rcu and rcu_barrier for softirq context compatibility
 - add tls_proto_put() for atomic refcount decrement + list_del +
   call_rcu, and simplify tls_ctx_free and tls_init error paths to use it
 - use GFP_ATOMIC for tls_proto allocation in tls_build_proto
 - reorder tls_register cleanup labels to unwind correctly on error
 - implement new mptcp_inq
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1777459066.git.tanggeliang@kylinos.cn/

v17:
 - remove owner from tls_prot_ops since both TCP and MPTCP are
   built-in and cannot be unloaded
 - add rcu_head to tls_proto for delayed freeing
 - replace synchronize_rcu with call_rcu in tls_ctx_free
 - remove module refcounting (try_module_get / module_put) from
   tls_build_proto and tls_init
 - simplify tls_proto_cleanup to directly free all protos without
   refcnt check
 - change tls_proto refcnt from 2 to 1 (only socket reference, no
   list reference)
 - move synchronize_rcu outside spinlock in tls_unregister_prot_ops
 - use kzalloc_obj instead of kzalloc to fix the checkpatch warning
 - update commit logs
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1777026753.git.tanggeliang@kylinos.cn/

v16:
 - drop rcu_head from struct tls_proto, use refcnt for lifecycle
   management.
 - add back TLS_NUM_PROTS to handle IPv4/IPv6 separately.
 - add .owner field to tls_tcp_ops and tls_mptcp_ops (THIS_MODULE).
 - add module refcounting (try_module_get / module_put) in tls_build_proto
   and tls_init.
 - add missing NULL check for tls_ctx->proto->ops in tls_sk_poll.
 - add RCU read lock protection in tls_register_prot_ops.
 - add error handling for tls_register_prot_ops calls in tls_register
   (with rollback on failure).
 - adjust MPTCP cleanup: move tcp_cleanup_ulp from mptcp_destroy_common
   to mptcp_destroy.
 - remove increase_rlimit from selftest and fix fd check.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1776924681.git.tanggeliang@kylinos.cn/

v15:
 - patch 1: add proto parameter for tls_toe_bypass.
 - patch 1: add a proto null-check in update_sk_prot.
 - patch 1: hold mutex_lock in tls_proto_cleanup.
 - patch 14: raise the limit of file descriptor values to 4096 to avoid
   test failures.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1776469068.git.tanggeliang@kylinos.cn/

v14:
 - address review comments from sashiko
 - patch 1: add rcu for tls_proto, add tls_proto_cleanup.
 - patch 2: add unregister helper.
 - patch 3: add tls_prot_ops pointer to tls_proto, instead of
   tls_context
 - patch 5: update mptcp_get_skb_seq, using map_seq - offset, then the
   patch "tls: add skb offset check for mptcp" can be dropped.
 - patch 7: check len < 0.
 - patch 8: call tcp_cleanup_ulp in mptcp_destroy_common.
 - patch 9: replace all "tls" as "espintcp" in sock_test_tcpulp.
 - patch 10: add is_mptcp_enable helper.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1775476921.git.tanggeliang@kylinos.cn/

v13:
 - patch 1: Add new patch "add per-protocol cache" to address AI review.
 - patch 2: Hold RCU read lock in tls_prot_ops_find().
 - patch 3: Set icsk_ulp_data to NULL in error path.
 - patch 6: Use spin_is_locked() instead of lockdep_is_held() to fix
   build errors.
 - patch 9: Drop tcp_sock_set_ulp().
 - patch 11: Remove the "return" statement in ulp_sock_pair and check
   the return values of socket().
 - patch 14: Update wait_for_tcp_close().
 - patch 16: Add a max argument to init() and set it to '0' to disable
   multipath testing, so that this series does not depend on the "mptcp:
   fix stall because of data_ready" series. Multipath testing will be
   re‑enabled together with that series later, as a squash‑to patch.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1775227717.git.tanggeliang@kylinos.cn/

v12:
 - Thanks for the help from Paolo and Gang Yan, I finally solved the
   deadlock issue in read_sock. As a result, the patch "mptcp: avoid
   sleeping in read_sock path under softirq" in v11 has been dropped,
   and instead a lock_is_held interface has been added to struct
   tls_prot_ops. When MPTCP implements this interface, it not only
   checks sock_owned_by_user_nocheck(sk) as TCP does, but also needs
   to check whether the MPTCP data lock is held.
 - Update selftests to make them more stable.
 - Fix shellcheck errors for the selftests.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1775115102.git.tanggeliang@kylinos.cn/

v11:
 - Fix memory leak errors reported by CI. In v10, these occurred in
   the shutdown_reuse test and "usleep(500000)" caused the memory
   leaks. In v11, a dedicated helper wait_for_tcp_close() has been
   added to provide an appropriate delay.
 - Drop the code that used mptcp_data_trylock() in mptcp_move_skbs()
   to fix a deadlock issue, as that deadlock no longer occurs in v11.
 - Do not add "mptcp" variable for the "tls_err" tests, adding it for
   the "tls" tests is sufficient.
 - No longer increase timeout values for poll/epoll tests, as they
   are no longer needed.
 - Add ns1 definition in mptcp_tls.sh to fix "ns1 is referenced but
   not assigned" error.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1773911536.git.tanggeliang@kylinos.cn/

v10:
 - Address comments by ai review:
   - patch 2: call tls_ctx_free(sk, ctx) and clear icsk_ulp_data before
     goto out.
   - patch 3: update commit log as "validate each SKB's offset except
     the first".
   - patch 5: add sock_owned_by_user() checks.
   - patch 7: disable device offload for MPTCP sockets.
   - patch 9: use TCP_ULP_NAME_MAX in mptcp_setsockopt_tcp_ulp(), drop
     SOL_TLS in mptcp_supported_sockopt().
 - Make .get_skb_off optional instead of mandatory, TCP does not need
   to define it.
 - Test "espintcp" ULP instead of "smc" in patch 10. "smc" ULP is
   removed recently.
 - With Gang Yan's "mptcp: fix stall because of data_ready" v3, mptcp
   tls selftests can run without failures. Now add them in this set.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1773737371.git.tanggeliang@kylinos.cn/

v9:
 - add a new patch to "add MPTCP SKB offset check in strp queue walk",
   thanks to Gang Yan for the fix.
 - add a new patch to "avoid deadlocks in read_sock path", replacing the
   "in_softirq()" check used in v8.
 - update the selftests.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1773365606.git.tanggeliang@kylinos.cn/

v8:
 - do not hold tls_prot_ops_lock in tls_init(); otherwise, a deadlock
   occurs.
 - change return value of mptcp_stream_is_readable() as 'bool' to fix the
   "expected restricted __poll_t" warning reported by CI.
 - fixed other CI checkpatch warnings regarding excessively long lines.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1768294706.git.tanggeliang@kylinos.cn/

v7:
 - Passing an MPTCP socket to tcp_sock_rate_check_app_limited() causes a
   crash. In v7, an MPTCP version of check_app_limited() is implemented,
   which calls tcp_sock_rate_check_app_limited() for each subflow.
 - Register tls_tcp_ops and tls_mptcp_ops in tls_register() rather than in
   tls_init().
 - Set ctx->ops in tls_init() instead of in do_tls_setsockopt_conf().
 - Keep tls_device.c unchanged. MPTCP TLS_HW mode has not been implemented
   yet, so EOPNOTSUPP is returned in this case.
 - Also add TCP TLS tests in mptcp_join.sh.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1768284047.git.tanggeliang@kylinos.cn/

v6:
 - register each ops as Matt suggested.
 - drop sk_is_msk().
 - add tcp_sock_get_ulp/tcp_sock_set_ulp helpers.
 - set another ULP in sock_test_tcpulp as Matt suggested.
 - add tls tests using multiple subflows in mptcp_join.sh.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1767518836.git.tanggeliang@kylinos.cn/

v5:
 - As suggested by Mat and Matt, this set introduces struct tls_prot_ops
   for TLS.
 - Includes Gang Yan's patches to add MPTCP support to the TLS selftests.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1766372799.git.tanggeliang@kylinos.cn/

v4:
 - split "tls: add MPTCP protocol support" into smaller, more
   focused patches.
 - a new mptcp_inq helper has been implemented instead of directly
   using mptcp_inq_hint to fix the issue mentioned in [1].
 - add sk_is_msk helper.
 - the 'expect' parameter will no longer be added to sock_test_tcpulp.
   Instead, SOCK_TEST_TCPULP items causing the tests failure will be
   directly removed.
 - remove the "TCP KTLS" tests, keeping only the MPTCP-related ones.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1765505775.git.tanggeliang@kylinos.cn/

[1]
https://patchwork.kernel.org/project/mptcp/patch/ce74452f4c095a1761ef493b767b4bd9f9c14359.1764333805.git.tanggeliang@kylinos.cn/

v3:
 - mptcp_read_sock() and mptcp_poll() are not exported, as mptcp_sockopt
   test does not use read_sock/poll interfaces. They will be exported when
   new tests are added in the future.
 - call mptcp_inq_hint in tls_device_rx_resync_new_rec(),
   tls_device_core_ctrl_rx_resync() and tls_read_flush_backlog() too.
 - update selftests.
 - Link: https://patchwork.kernel.org/project/mptcp/cover/cover.1763800601.git.tanggeliang@kylinos.cn/

v2:
 - fix disconnect.
 - update selftests.

This series adds KTLS support for MPTCP. Since the ULP of msk is not being
used, ULP KTLS can be directly configured onto msk without affecting its
communication.

Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/480

Gang Yan (1):
  mptcp: update mptcp_check_readable

Geliang Tang (14):
  tls: add per-protocol cache to support mptcp
  tls: introduce struct tls_prot_ops
  tls: add tls_prot_ops pointer to tls_proto
  mptcp: implement tls_mptcp_ops
  tls: disable device offload for mptcp sockets
  mptcp: update ulp getsockopt for tls support
  mptcp: enable ulp setsockopt for tls support
  selftests: mptcp: connect: use espintcp for ulp test
  selftests: tls: add mptcp variant for testing
  selftests: tls: increase pollin timeouts for mptcp
  selftests: tls: increase nonblocking data size for mptcp
  selftests: tls: retry bind on EINVAL in shutdown_reuse
  selftests: tls: add mptcp test cases
  selftests: mptcp: cover mptcp tls tests

 include/linux/tcp.h                           |   1 +
 include/net/mptcp.h                           |   2 +
 include/net/tcp.h                             |   1 +
 include/net/tls.h                             |  32 ++
 include/net/tls_toe.h                         |   3 +-
 net/ipv4/tcp.c                                |  45 ++-
 net/mptcp/protocol.c                          | 131 ++++++++-
 net/mptcp/protocol.h                          |   1 +
 net/mptcp/sockopt.c                           |  58 +++-
 net/tls/tls.h                                 |   3 +-
 net/tls/tls_device.c                          |   6 +
 net/tls/tls_main.c                            | 278 +++++++++++++++---
 net/tls/tls_strp.c                            |  33 ++-
 net/tls/tls_sw.c                              |   6 +-
 net/tls/tls_toe.c                             |   7 +-
 tools/testing/selftests/net/mptcp/.gitignore  |   1 +
 tools/testing/selftests/net/mptcp/Makefile    |   2 +
 tools/testing/selftests/net/mptcp/config      |   6 +
 .../selftests/net/mptcp/mptcp_connect.c       |   4 +-
 .../testing/selftests/net/mptcp/mptcp_tls.sh  |  64 ++++
 tools/testing/selftests/net/mptcp/tls.c       |   1 +
 tools/testing/selftests/net/tls.c             | 155 +++++++++-
 22 files changed, 748 insertions(+), 92 deletions(-)
 create mode 100755 tools/testing/selftests/net/mptcp/mptcp_tls.sh
 create mode 120000 tools/testing/selftests/net/mptcp/tls.c

-- 
2.53.0

[RFC mptcp-next v20 01/15] tls: add per-protocol cache to support mptcp

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

The TLS ULP uses a single global array to cache base protocol operations.
When MPTCP sockets enable TLS, they overwrite this global cache with
mptcp_prot, causing active TCP TLS sockets to use MPTCP-specific ops.
This leads to type confusion and kernel panics.

Fix by replacing the global cache with a per-protocol linked list.
Each protocol (TCP, MPTCP, etc.) now has its own cached operations,
stored in struct tls_proto and referenced from tls_context.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 include/net/tls.h     |  12 ++++
 include/net/tls_toe.h |   3 +-
 net/tls/tls.h         |   3 +-
 net/tls/tls_main.c    | 137 ++++++++++++++++++++++++++++++------------
 net/tls/tls_toe.c     |   7 ++-
 5 files changed, 120 insertions(+), 42 deletions(-)

diff --git a/include/net/tls.h b/include/net/tls.h
index ebd2550280ae..65c4c2c72e06 100644
--- a/include/net/tls.h
+++ b/include/net/tls.h
@@ -220,6 +220,16 @@ struct tls_prot_info {
 	u16 tail_size;
 };
 
+struct tls_proto {
+	struct rcu_head			rcu;
+	refcount_t			refcnt;
+	struct list_head		list;
+	int				ip_ver;
+	const struct proto		*prot;
+	struct proto prots[TLS_NUM_CONFIG][TLS_NUM_CONFIG];
+	struct proto_ops proto_ops[TLS_NUM_CONFIG][TLS_NUM_CONFIG];
+};
+
 struct tls_context {
 	/* read-only cache line */
 	struct tls_prot_info prot_info;
@@ -257,6 +267,8 @@ struct tls_context {
 	struct proto *sk_proto;
 	struct sock *sk;
 
+	struct tls_proto *proto;
+
 	void (*sk_destruct)(struct sock *sk);
 
 	union tls_crypto_context crypto_send;
diff --git a/include/net/tls_toe.h b/include/net/tls_toe.h
index b3aa7593ce2c..b73029364b2c 100644
--- a/include/net/tls_toe.h
+++ b/include/net/tls_toe.h
@@ -69,7 +69,8 @@ struct tls_toe_device {
 	struct kref kref;
 };
 
-int tls_toe_bypass(struct sock *sk);
+int tls_toe_bypass(struct sock *sk,
+		   struct tls_proto *proto);
 int tls_toe_hash(struct sock *sk);
 void tls_toe_unhash(struct sock *sk);
 
diff --git a/net/tls/tls.h b/net/tls/tls.h
index 12f44cb649c9..3a16047c3a8f 100644
--- a/net/tls/tls.h
+++ b/net/tls/tls.h
@@ -136,7 +136,8 @@ struct tls_rec {
 int __net_init tls_proc_init(struct net *net);
 void __net_exit tls_proc_fini(struct net *net);
 
-struct tls_context *tls_ctx_create(struct sock *sk);
+struct tls_context *tls_ctx_create(struct sock *sk,
+				   struct tls_proto *proto);
 void tls_ctx_free(struct sock *sk, struct tls_context *ctx);
 void update_sk_prot(struct sock *sk, struct tls_context *ctx);
 
diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
index fd39acf41a61..30b50ea708ee 100644
--- a/net/tls/tls_main.c
+++ b/net/tls/tls_main.c
@@ -119,23 +119,61 @@ CHECK_CIPHER_DESC(TLS_CIPHER_SM4_CCM, tls12_crypto_info_sm4_ccm);
 CHECK_CIPHER_DESC(TLS_CIPHER_ARIA_GCM_128, tls12_crypto_info_aria_gcm_128);
 CHECK_CIPHER_DESC(TLS_CIPHER_ARIA_GCM_256, tls12_crypto_info_aria_gcm_256);
 
-static const struct proto *saved_tcpv6_prot;
-static DEFINE_MUTEX(tcpv6_prot_mutex);
-static const struct proto *saved_tcpv4_prot;
-static DEFINE_MUTEX(tcpv4_prot_mutex);
-static struct proto tls_prots[TLS_NUM_PROTS][TLS_NUM_CONFIG][TLS_NUM_CONFIG];
-static struct proto_ops tls_proto_ops[TLS_NUM_PROTS][TLS_NUM_CONFIG][TLS_NUM_CONFIG];
+static LIST_HEAD(tls_proto_list);
+static DEFINE_SPINLOCK(tls_proto_lock);
 static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
 			 const struct proto *base);
 
+static struct tls_proto *tls_proto_find(const struct proto *prot,
+					int ip_ver)
+{
+	struct tls_proto *proto, *ret = NULL;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(proto, &tls_proto_list, list) {
+		if (proto->prot == prot && proto->ip_ver == ip_ver &&
+		    refcount_inc_not_zero(&proto->refcnt)) {
+			ret = proto;
+			break;
+		}
+	}
+	rcu_read_unlock();
+	return ret;
+}
+
+static void tls_proto_free(struct rcu_head *rcu)
+{
+	struct tls_proto *proto = container_of(rcu, struct tls_proto, rcu);
+
+	kfree(proto);
+}
+
+static void tls_proto_cleanup(void)
+{
+	struct tls_proto *prot, *tmp;
+
+	spin_lock_bh(&tls_proto_lock);
+	list_for_each_entry_safe(prot, tmp, &tls_proto_list, list) {
+		list_del_rcu(&prot->list);
+		call_rcu(&prot->rcu, tls_proto_free);
+	}
+	spin_unlock_bh(&tls_proto_lock);
+}
+
 void update_sk_prot(struct sock *sk, struct tls_context *ctx)
 {
-	int ip_ver = sk->sk_family == AF_INET6 ? TLSV6 : TLSV4;
+	struct tls_proto *proto = ctx->proto;
+
+	if (!proto)
+		return;
 
+	/* Ensure sk->sk_socket->ops is not visible before icsk_ulp_data.
+	 * Pairs with the rcu_assign_pointer() release in tls_ctx_create().
+	 */
+	smp_store_release(&sk->sk_socket->ops,
+			  &proto->proto_ops[ctx->tx_conf][ctx->rx_conf]);
 	WRITE_ONCE(sk->sk_prot,
-		   &tls_prots[ip_ver][ctx->tx_conf][ctx->rx_conf]);
-	WRITE_ONCE(sk->sk_socket->ops,
-		   &tls_proto_ops[ip_ver][ctx->tx_conf][ctx->rx_conf]);
+		   &proto->prots[ctx->tx_conf][ctx->rx_conf]);
 }
 
 int wait_on_pending_writer(struct sock *sk, long *timeo)
@@ -314,6 +352,16 @@ static void tls_write_space(struct sock *sk)
 	ctx->sk_write_space(sk);
 }
 
+static void tls_proto_put(struct tls_proto *proto)
+{
+	if (refcount_dec_and_test(&proto->refcnt)) {
+		spin_lock_bh(&tls_proto_lock);
+		list_del_rcu(&proto->list);
+		spin_unlock_bh(&tls_proto_lock);
+		call_rcu(&proto->rcu, tls_proto_free);
+	}
+}
+
 /**
  * tls_ctx_free() - free TLS ULP context
  * @sk:  socket to with @ctx is attached
@@ -327,6 +375,11 @@ void tls_ctx_free(struct sock *sk, struct tls_context *ctx)
 	if (!ctx)
 		return;
 
+	if (ctx->proto) {
+		tls_proto_put(ctx->proto);
+		ctx->proto = NULL;
+	}
+
 	memzero_explicit(&ctx->crypto_send, sizeof(ctx->crypto_send));
 	memzero_explicit(&ctx->crypto_recv, sizeof(ctx->crypto_recv));
 	mutex_destroy(&ctx->tx_lock);
@@ -910,7 +963,8 @@ static int tls_disconnect(struct sock *sk, int flags)
 	return -EOPNOTSUPP;
 }
 
-struct tls_context *tls_ctx_create(struct sock *sk)
+struct tls_context *tls_ctx_create(struct sock *sk,
+				   struct tls_proto *proto)
 {
 	struct inet_connection_sock *icsk = inet_csk(sk);
 	struct tls_context *ctx;
@@ -921,6 +975,7 @@ struct tls_context *tls_ctx_create(struct sock *sk)
 
 	mutex_init(&ctx->tx_lock);
 	ctx->sk_proto = READ_ONCE(sk->sk_prot);
+	ctx->proto = proto;
 	ctx->sk = sk;
 	/* Release semantic of rcu_assign_pointer() ensures that
 	 * ctx->sk_proto is visible before changing sk->sk_prot in
@@ -968,35 +1023,34 @@ static void build_proto_ops(struct proto_ops ops[TLS_NUM_CONFIG][TLS_NUM_CONFIG]
 #endif
 }
 
-static void tls_build_proto(struct sock *sk)
+static struct tls_proto *tls_build_proto(struct sock *sk)
 {
 	int ip_ver = sk->sk_family == AF_INET6 ? TLSV6 : TLSV4;
 	struct proto *prot = READ_ONCE(sk->sk_prot);
+	struct tls_proto *proto, *cache;
 
-	/* Build IPv6 TLS whenever the address of tcpv6 _prot changes */
-	if (ip_ver == TLSV6 &&
-	    unlikely(prot != smp_load_acquire(&saved_tcpv6_prot))) {
-		mutex_lock(&tcpv6_prot_mutex);
-		if (likely(prot != saved_tcpv6_prot)) {
-			build_protos(tls_prots[TLSV6], prot);
-			build_proto_ops(tls_proto_ops[TLSV6],
-					sk->sk_socket->ops);
-			smp_store_release(&saved_tcpv6_prot, prot);
-		}
-		mutex_unlock(&tcpv6_prot_mutex);
-	}
+	proto = kzalloc_obj(*proto, GFP_KERNEL);
+	if (!proto)
+		return NULL;
 
-	if (ip_ver == TLSV4 &&
-	    unlikely(prot != smp_load_acquire(&saved_tcpv4_prot))) {
-		mutex_lock(&tcpv4_prot_mutex);
-		if (likely(prot != saved_tcpv4_prot)) {
-			build_protos(tls_prots[TLSV4], prot);
-			build_proto_ops(tls_proto_ops[TLSV4],
-					sk->sk_socket->ops);
-			smp_store_release(&saved_tcpv4_prot, prot);
-		}
-		mutex_unlock(&tcpv4_prot_mutex);
+	spin_lock_bh(&tls_proto_lock);
+	cache = tls_proto_find(prot, ip_ver);
+	if (cache) {
+		spin_unlock_bh(&tls_proto_lock);
+		kfree(proto);
+		return cache;
 	}
+
+	proto->ip_ver = ip_ver;
+	proto->prot = prot;
+	refcount_set(&proto->refcnt, 1);
+	build_protos(proto->prots, prot);
+	build_proto_ops(proto->proto_ops,
+			sk->sk_socket->ops);
+	list_add_rcu(&proto->list, &tls_proto_list);
+
+	spin_unlock_bh(&tls_proto_lock);
+	return proto;
 }
 
 static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
@@ -1046,13 +1100,16 @@ static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
 
 static int tls_init(struct sock *sk)
 {
+	struct tls_proto *proto;
 	struct tls_context *ctx;
 	int rc = 0;
 
-	tls_build_proto(sk);
+	proto = tls_build_proto(sk);
+	if (!proto)
+		return -ENOMEM;
 
 #ifdef CONFIG_TLS_TOE
-	if (tls_toe_bypass(sk))
+	if (tls_toe_bypass(sk, proto))
 		return 0;
 #endif
 
@@ -1062,13 +1119,16 @@ static int tls_init(struct sock *sk)
 	 * to modify the accept implementation to clone rather then
 	 * share the ulp context.
 	 */
-	if (sk->sk_state != TCP_ESTABLISHED)
+	if (sk->sk_state != TCP_ESTABLISHED) {
+		tls_proto_put(proto);
 		return -ENOTCONN;
+	}
 
 	/* allocate tls context */
 	write_lock_bh(&sk->sk_callback_lock);
-	ctx = tls_ctx_create(sk);
+	ctx = tls_ctx_create(sk, proto);
 	if (!ctx) {
+		tls_proto_put(proto);
 		rc = -ENOMEM;
 		goto out;
 	}
@@ -1265,6 +1325,7 @@ static int __init tls_register(void)
 static void __exit tls_unregister(void)
 {
 	tcp_unregister_ulp(&tcp_tls_ulp_ops);
+	tls_proto_cleanup();
 	tls_strp_dev_exit();
 	tls_device_cleanup();
 	unregister_pernet_subsys(&tls_proc_ops);
diff --git a/net/tls/tls_toe.c b/net/tls/tls_toe.c
index 825669e1ab47..d4cd3974414a 100644
--- a/net/tls/tls_toe.c
+++ b/net/tls/tls_toe.c
@@ -48,13 +48,16 @@ static void tls_toe_sk_destruct(struct sock *sk)
 	struct inet_connection_sock *icsk = inet_csk(sk);
 	struct tls_context *ctx = tls_get_ctx(sk);
 
+	WRITE_ONCE(sk->sk_prot, ctx->sk_proto);
+
 	ctx->sk_destruct(sk);
 	/* Free ctx */
 	rcu_assign_pointer(icsk->icsk_ulp_data, NULL);
 	tls_ctx_free(sk, ctx);
 }
 
-int tls_toe_bypass(struct sock *sk)
+int tls_toe_bypass(struct sock *sk,
+		   struct tls_proto *proto)
 {
 	struct tls_toe_device *dev;
 	struct tls_context *ctx;
@@ -63,7 +66,7 @@ int tls_toe_bypass(struct sock *sk)
 	spin_lock_bh(&device_spinlock);
 	list_for_each_entry(dev, &device_list, dev_list) {
 		if (dev->feature && dev->feature(dev)) {
-			ctx = tls_ctx_create(sk);
+			ctx = tls_ctx_create(sk, proto);
 			if (!ctx)
 				goto out;
 
-- 
2.53.0

[RFC mptcp-next v20 02/15] tls: introduce struct tls_prot_ops

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

To extend MPTCP support based on TCP TLS, a tls_prot_ops structure has
been introduced for TLS, encapsulating TCP-specific helpers within this
structure.

Add registering, validating and finding functions for this structure to
add, validate and find a tls_prot_ops on the global list tls_prot_ops_list.

Register TCP-specific structure tls_tcp_ops in tls_register().

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 include/net/tls.h  |  19 ++++++++
 net/tls/tls_main.c | 105 ++++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 123 insertions(+), 1 deletion(-)

diff --git a/include/net/tls.h b/include/net/tls.h
index 65c4c2c72e06..eab4e68642ad 100644
--- a/include/net/tls.h
+++ b/include/net/tls.h
@@ -220,6 +220,25 @@ struct tls_prot_info {
 	u16 tail_size;
 };
 
+struct tls_prot_ops {
+	struct module		*owner;
+	int			protocol;
+	struct list_head	list;
+
+	int (*inq)(struct sock *sk);
+	int (*sendmsg_locked)(struct sock *sk, struct msghdr *msg, size_t size);
+	struct sk_buff *(*recv_skb)(struct sock *sk, u32 *off);
+	bool (*lock_is_held)(struct sock *sk);
+	int (*read_sock)(struct sock *sk, read_descriptor_t *desc,
+			 sk_read_actor_t recv_actor);
+	void (*read_done)(struct sock *sk, size_t len);
+	u32 (*get_skb_seq)(struct sk_buff *skb);
+	__poll_t (*poll)(struct file *file, struct socket *sock,
+			 struct poll_table_struct *wait);
+	bool (*epollin_ready)(const struct sock *sk);
+	void (*check_app_limited)(struct sock *sk);
+};
+
 struct tls_proto {
 	struct rcu_head			rcu;
 	refcount_t			refcnt;
diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
index 30b50ea708ee..ca51f5159eda 100644
--- a/net/tls/tls_main.c
+++ b/net/tls/tls_main.c
@@ -120,6 +120,7 @@ CHECK_CIPHER_DESC(TLS_CIPHER_ARIA_GCM_128, tls12_crypto_info_aria_gcm_128);
 CHECK_CIPHER_DESC(TLS_CIPHER_ARIA_GCM_256, tls12_crypto_info_aria_gcm_256);
 
 static LIST_HEAD(tls_proto_list);
+static LIST_HEAD(tls_prot_ops_list);
 static DEFINE_SPINLOCK(tls_proto_lock);
 static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
 			 const struct proto *base);
@@ -160,6 +161,22 @@ static void tls_proto_cleanup(void)
 	spin_unlock_bh(&tls_proto_lock);
 }
 
+static struct tls_prot_ops *tls_prot_ops_find(int protocol)
+{
+	struct tls_prot_ops *ops, *ret = NULL;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(ops, &tls_prot_ops_list, list) {
+		if (ops->protocol == protocol) {
+			ret = ops;
+			break;
+		}
+	}
+	rcu_read_unlock();
+
+	return ret;
+}
+
 void update_sk_prot(struct sock *sk, struct tls_context *ctx)
 {
 	struct tls_proto *proto = ctx->proto;
@@ -1296,6 +1313,85 @@ static struct tcp_ulp_ops tcp_tls_ulp_ops __read_mostly = {
 	.get_info_size		= tls_get_info_size,
 };
 
+static int tls_validate_prot_ops(const struct tls_prot_ops *ops)
+{
+	if (!ops->inq || !ops->sendmsg_locked ||
+	    !ops->recv_skb || !ops->lock_is_held ||
+	    !ops->read_sock || !ops->read_done ||
+	    !ops->get_skb_seq ||
+	    !ops->poll || !ops->epollin_ready ||
+	    !ops->check_app_limited) {
+		pr_err("%d does not implement required ops\n", ops->protocol);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int tls_register_prot_ops(struct tls_prot_ops *ops)
+{
+	int ret;
+
+	ret = tls_validate_prot_ops(ops);
+	if (ret)
+		return ret;
+
+	spin_lock_bh(&tls_proto_lock);
+	if (tls_prot_ops_find(ops->protocol)) {
+		spin_unlock_bh(&tls_proto_lock);
+		return -EEXIST;
+	}
+
+	list_add_tail_rcu(&ops->list, &tls_prot_ops_list);
+	spin_unlock_bh(&tls_proto_lock);
+
+	pr_debug("tls_prot_ops %d registered\n", ops->protocol);
+	return 0;
+}
+
+static void tls_unregister_prot_ops(struct tls_prot_ops *ops)
+{
+	spin_lock_bh(&tls_proto_lock);
+	list_del_rcu(&ops->list);
+	spin_unlock_bh(&tls_proto_lock);
+	synchronize_rcu();
+}
+
+static struct sk_buff *tls_tcp_recv_skb(struct sock *sk, u32 *off)
+{
+	return tcp_recv_skb(sk, tcp_sk(sk)->copied_seq, off);
+}
+
+static bool tls_tcp_lock_is_held(struct sock *sk)
+{
+	return sock_owned_by_user_nocheck(sk);
+}
+
+static u32 tls_tcp_get_skb_seq(struct sk_buff *skb)
+{
+	return TCP_SKB_CB(skb)->seq;
+}
+
+static bool tls_tcp_epollin_ready(const struct sock *sk)
+{
+	return tcp_epollin_ready(sk, INT_MAX);
+}
+
+static struct tls_prot_ops tls_tcp_ops = {
+	.owner			= THIS_MODULE,
+	.protocol		= IPPROTO_TCP,
+	.inq			= tcp_inq,
+	.sendmsg_locked		= tcp_sendmsg_locked,
+	.recv_skb		= tls_tcp_recv_skb,
+	.lock_is_held		= tls_tcp_lock_is_held,
+	.read_sock		= tcp_read_sock,
+	.read_done		= tcp_read_done,
+	.get_skb_seq		= tls_tcp_get_skb_seq,
+	.poll			= tcp_poll,
+	.epollin_ready		= tls_tcp_epollin_ready,
+	.check_app_limited	= tcp_rate_check_app_limited,
+};
+
 static int __init tls_register(void)
 {
 	int err;
@@ -1308,13 +1404,19 @@ static int __init tls_register(void)
 	if (err)
 		goto err_pernet;
 
-	err = tls_device_init();
+	err = tls_register_prot_ops(&tls_tcp_ops);
 	if (err)
 		goto err_strp;
 
+	err = tls_device_init();
+	if (err)
+		goto err_ops;
+
 	tcp_register_ulp(&tcp_tls_ulp_ops);
 
 	return 0;
+err_ops:
+	tls_unregister_prot_ops(&tls_tcp_ops);
 err_strp:
 	tls_strp_dev_exit();
 err_pernet:
@@ -1325,6 +1427,7 @@ static int __init tls_register(void)
 static void __exit tls_unregister(void)
 {
 	tcp_unregister_ulp(&tcp_tls_ulp_ops);
+	tls_unregister_prot_ops(&tls_tcp_ops);
 	tls_proto_cleanup();
 	tls_strp_dev_exit();
 	tls_device_cleanup();
-- 
2.53.0

[RFC mptcp-next v20 03/15] tls: add tls_prot_ops pointer to tls_proto

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

A pointer to struct tls_prot_ops, named 'ops', has been added to struct
tls_proto. The places originally calling TLS-specific helpers have now
been modified to indirectly invoke them via 'ops' pointer in tls_proto.

In tls_build_proto(), proto->ops is assigned either 'tls_mptcp_ops' or
'tls_tcp_ops' based on the socket protocol.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 include/net/tls.h  |  1 +
 net/tls/tls_main.c | 25 +++++++++++++++++++++----
 net/tls/tls_strp.c | 33 ++++++++++++++++++++++-----------
 net/tls/tls_sw.c   |  6 ++++--
 4 files changed, 48 insertions(+), 17 deletions(-)

diff --git a/include/net/tls.h b/include/net/tls.h
index eab4e68642ad..1ebcdcd27ea7 100644
--- a/include/net/tls.h
+++ b/include/net/tls.h
@@ -245,6 +245,7 @@ struct tls_proto {
 	struct list_head		list;
 	int				ip_ver;
 	const struct proto		*prot;
+	const struct tls_prot_ops	*ops;
 	struct proto prots[TLS_NUM_CONFIG][TLS_NUM_CONFIG];
 	struct proto_ops proto_ops[TLS_NUM_CONFIG][TLS_NUM_CONFIG];
 };
diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
index ca51f5159eda..c840a228af0d 100644
--- a/net/tls/tls_main.c
+++ b/net/tls/tls_main.c
@@ -243,13 +243,13 @@ int tls_push_sg(struct sock *sk,
 	ctx->splicing_pages = true;
 	while (1) {
 		/* is sending application-limited? */
-		tcp_rate_check_app_limited(sk);
+		ctx->proto->ops->check_app_limited(sk);
 		p = sg_page(sg);
 retry:
 		bvec_set_page(&bvec, p, size, offset);
 		iov_iter_bvec(&msg.msg_iter, ITER_SOURCE, &bvec, 1, size);
 
-		ret = tcp_sendmsg_locked(sk, &msg, size);
+		ret = ctx->proto->ops->sendmsg_locked(sk, &msg, size);
 
 		if (ret != size) {
 			if (ret > 0) {
@@ -375,6 +375,7 @@ static void tls_proto_put(struct tls_proto *proto)
 		spin_lock_bh(&tls_proto_lock);
 		list_del_rcu(&proto->list);
 		spin_unlock_bh(&tls_proto_lock);
+		module_put(proto->ops->owner);
 		call_rcu(&proto->rcu, tls_proto_free);
 	}
 }
@@ -479,14 +480,21 @@ static __poll_t tls_sk_poll(struct file *file, struct socket *sock,
 	u8 shutdown;
 	int state;
 
-	mask = tcp_poll(file, sock, wait);
+	/* Paired with smp_store_release() in update_sk_prot().
+	 * Ensures that the read of icsk_ulp_data (tls_ctx) is ordered after
+	 * the read of sk->sk_socket->ops inside tls_ctx->proto->ops->poll().
+	 * Prevents seeing a new ops pointer but a stale (NULL) ulp_data.
+	 */
+	smp_rmb();
+
+	tls_ctx = tls_get_ctx(sk);
+	mask = tls_ctx->proto->ops->poll(file, sock, wait);
 
 	state = inet_sk_state_load(sk);
 	shutdown = READ_ONCE(sk->sk_shutdown);
 	if (unlikely(state != TCP_ESTABLISHED || shutdown & RCV_SHUTDOWN))
 		return mask;
 
-	tls_ctx = tls_get_ctx(sk);
 	ctx = tls_sw_ctx_rx(tls_ctx);
 	psock = sk_psock_get(sk);
 
@@ -1045,6 +1053,7 @@ static struct tls_proto *tls_build_proto(struct sock *sk)
 	int ip_ver = sk->sk_family == AF_INET6 ? TLSV6 : TLSV4;
 	struct proto *prot = READ_ONCE(sk->sk_prot);
 	struct tls_proto *proto, *cache;
+	struct tls_prot_ops *ops;
 
 	proto = kzalloc_obj(*proto, GFP_KERNEL);
 	if (!proto)
@@ -1058,8 +1067,16 @@ static struct tls_proto *tls_build_proto(struct sock *sk)
 		return cache;
 	}
 
+	ops = tls_prot_ops_find(sk->sk_protocol);
+	if (!ops || !try_module_get(ops->owner)) {
+		spin_unlock_bh(&tls_proto_lock);
+		kfree(proto);
+		return NULL;
+	}
+
 	proto->ip_ver = ip_ver;
 	proto->prot = prot;
+	proto->ops = ops;
 	refcount_set(&proto->refcnt, 1);
 	build_protos(proto->prots, prot);
 	build_proto_ops(proto->proto_ops,
diff --git a/net/tls/tls_strp.c b/net/tls/tls_strp.c
index c72e88317627..869e2916db5a 100644
--- a/net/tls/tls_strp.c
+++ b/net/tls/tls_strp.c
@@ -120,6 +120,7 @@ struct sk_buff *tls_strp_msg_detach(struct tls_sw_context_rx *ctx)
 int tls_strp_msg_cow(struct tls_sw_context_rx *ctx)
 {
 	struct tls_strparser *strp = &ctx->strp;
+	struct tls_context *tls_ctx = tls_get_ctx(strp->sk);
 	struct sk_buff *skb;
 
 	if (strp->copy_mode)
@@ -132,7 +133,7 @@ int tls_strp_msg_cow(struct tls_sw_context_rx *ctx)
 	tls_strp_anchor_free(strp);
 	strp->anchor = skb;
 
-	tcp_read_done(strp->sk, strp->stm.full_len);
+	tls_ctx->proto->ops->read_done(strp->sk, strp->stm.full_len);
 	strp->copy_mode = 1;
 
 	return 0;
@@ -376,6 +377,7 @@ static int tls_strp_copyin(read_descriptor_t *desc, struct sk_buff *in_skb,
 
 static int tls_strp_read_copyin(struct tls_strparser *strp)
 {
+	struct tls_context *ctx = tls_get_ctx(strp->sk);
 	read_descriptor_t desc;
 
 	desc.arg.data = strp;
@@ -383,13 +385,14 @@ static int tls_strp_read_copyin(struct tls_strparser *strp)
 	desc.count = 1; /* give more than one skb per call */
 
 	/* sk should be locked here, so okay to do read_sock */
-	tcp_read_sock(strp->sk, &desc, tls_strp_copyin);
+	ctx->proto->ops->read_sock(strp->sk, &desc, tls_strp_copyin);
 
 	return desc.error;
 }
 
 static int tls_strp_read_copy(struct tls_strparser *strp, bool qshort)
 {
+	struct tls_context *ctx = tls_get_ctx(strp->sk);
 	struct skb_shared_info *shinfo;
 	struct page *page;
 	int need_spc, len;
@@ -398,7 +401,7 @@ static int tls_strp_read_copy(struct tls_strparser *strp, bool qshort)
 	 * to read the data out. Otherwise the connection will stall.
 	 * Without pressure threshold of INT_MAX will never be ready.
 	 */
-	if (likely(qshort && !tcp_epollin_ready(strp->sk, INT_MAX)))
+	if (likely(qshort && !ctx->proto->ops->epollin_ready(strp->sk)))
 		return 0;
 
 	shinfo = skb_shinfo(strp->anchor);
@@ -434,12 +437,13 @@ static int tls_strp_read_copy(struct tls_strparser *strp, bool qshort)
 static bool tls_strp_check_queue_ok(struct tls_strparser *strp)
 {
 	unsigned int len = strp->stm.offset + strp->stm.full_len;
+	struct tls_context *ctx = tls_get_ctx(strp->sk);
 	struct sk_buff *first, *skb;
 	u32 seq;
 
 	first = skb_shinfo(strp->anchor)->frag_list;
 	skb = first;
-	seq = TCP_SKB_CB(first)->seq;
+	seq = ctx->proto->ops->get_skb_seq(first);
 
 	/* Make sure there's no duplicate data in the queue,
 	 * and the decrypted status matches.
@@ -449,7 +453,7 @@ static bool tls_strp_check_queue_ok(struct tls_strparser *strp)
 		len -= skb->len;
 		skb = skb->next;
 
-		if (TCP_SKB_CB(skb)->seq != seq)
+		if (ctx->proto->ops->get_skb_seq(skb) != seq)
 			return false;
 		if (skb_cmp_decrypted(first, skb))
 			return false;
@@ -460,11 +464,11 @@ static bool tls_strp_check_queue_ok(struct tls_strparser *strp)
 
 static void tls_strp_load_anchor_with_queue(struct tls_strparser *strp, int len)
 {
-	struct tcp_sock *tp = tcp_sk(strp->sk);
+	struct tls_context *ctx = tls_get_ctx(strp->sk);
 	struct sk_buff *first;
 	u32 offset;
 
-	first = tcp_recv_skb(strp->sk, tp->copied_seq, &offset);
+	first = ctx->proto->ops->recv_skb(strp->sk, &offset);
 	if (WARN_ON_ONCE(!first))
 		return;
 
@@ -483,6 +487,7 @@ static void tls_strp_load_anchor_with_queue(struct tls_strparser *strp, int len)
 
 bool tls_strp_msg_load(struct tls_strparser *strp, bool force_refresh)
 {
+	struct tls_context *ctx = tls_get_ctx(strp->sk);
 	struct strp_msg *rxm;
 	struct tls_msg *tlm;
 
@@ -490,7 +495,8 @@ bool tls_strp_msg_load(struct tls_strparser *strp, bool force_refresh)
 	DEBUG_NET_WARN_ON_ONCE(!strp->stm.full_len);
 
 	if (!strp->copy_mode && force_refresh) {
-		if (unlikely(tcp_inq(strp->sk) < strp->stm.full_len)) {
+		if (unlikely(ctx->proto->ops->inq(strp->sk) <
+			     strp->stm.full_len)) {
 			WRITE_ONCE(strp->msg_ready, 0);
 			memset(&strp->stm, 0, sizeof(strp->stm));
 			return false;
@@ -511,9 +517,10 @@ bool tls_strp_msg_load(struct tls_strparser *strp, bool force_refresh)
 /* Called with lock held on lower socket */
 static int tls_strp_read_sock(struct tls_strparser *strp)
 {
+	struct tls_context *ctx = tls_get_ctx(strp->sk);
 	int sz, inq;
 
-	inq = tcp_inq(strp->sk);
+	inq = ctx->proto->ops->inq(strp->sk);
 	if (inq < 1)
 		return 0;
 
@@ -556,6 +563,8 @@ void tls_strp_check_rcv(struct tls_strparser *strp)
 /* Lower sock lock held */
 void tls_strp_data_ready(struct tls_strparser *strp)
 {
+	struct tls_context *ctx = tls_get_ctx(strp->sk);
+
 	/* This check is needed to synchronize with do_tls_strp_work.
 	 * do_tls_strp_work acquires a process lock (lock_sock) whereas
 	 * the lock held here is bh_lock_sock. The two locks can be
@@ -563,7 +572,7 @@ void tls_strp_data_ready(struct tls_strparser *strp)
 	 * allows a thread in BH context to safely check if the process
 	 * lock is held. In this case, if the lock is held, queue work.
 	 */
-	if (sock_owned_by_user_nocheck(strp->sk)) {
+	if (ctx->proto->ops->lock_is_held(strp->sk)) {
 		queue_work(tls_strp_wq, &strp->work);
 		return;
 	}
@@ -583,10 +592,12 @@ static void tls_strp_work(struct work_struct *w)
 
 void tls_strp_msg_done(struct tls_strparser *strp)
 {
+	struct tls_context *ctx = tls_get_ctx(strp->sk);
+
 	WARN_ON(!strp->stm.full_len);
 
 	if (likely(!strp->copy_mode))
-		tcp_read_done(strp->sk, strp->stm.full_len);
+		ctx->proto->ops->read_done(strp->sk, strp->stm.full_len);
 	else
 		tls_strp_flush_anchor_copy(strp);
 
diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index 964ebc268ee4..9517f5cb091b 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -1987,7 +1987,8 @@ tls_read_flush_backlog(struct sock *sk, struct tls_prot_info *prot,
 		return false;
 
 	max_rec = prot->overhead_size - prot->tail_size + TLS_MAX_PAYLOAD_SIZE;
-	if (done - *flushed_at < SZ_128K && tcp_inq(sk) > max_rec)
+	if (done - *flushed_at < SZ_128K &&
+	    tls_get_ctx(sk)->proto->ops->inq(sk) > max_rec)
 		return false;
 
 	*flushed_at = done;
@@ -2516,7 +2517,8 @@ int tls_rx_msg_size(struct tls_strparser *strp, struct sk_buff *skb)
 	}
 
 	tls_device_rx_resync_new_rec(strp->sk, data_len + TLS_HEADER_SIZE,
-				     TCP_SKB_CB(skb)->seq + strp->stm.offset);
+				     tls_ctx->proto->ops->get_skb_seq(skb) +
+				     strp->stm.offset);
 	return data_len + TLS_HEADER_SIZE;
 
 read_failure:
-- 
2.53.0

[RFC mptcp-next v20 04/15] mptcp: update mptcp_check_readable

[Geliang Tang]

From: Gang Yan <yangang@kylinos.cn>

This patch makes mptcp_check_readable() aligned with TCP, and renames it to
mptcp_stream_is_readable(). It will be used in the case of KTLS, because
'prot' will be modified, tls_sw_sock_is_readable() is expected to be called
from prot->sock_is_readable().

Co-developed-by: Geliang Tang <geliang@kernel.org>
Signed-off-by: Geliang Tang <geliang@kernel.org>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
---
 net/mptcp/protocol.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index ce8372fb3c6a..a36ef97155a7 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -3276,9 +3276,11 @@ void __mptcp_unaccepted_force_close(struct sock *sk)
 	__mptcp_destroy_sock(sk);
 }
 
-static __poll_t mptcp_check_readable(struct sock *sk)
+static bool mptcp_stream_is_readable(struct sock *sk)
 {
-	return mptcp_epollin_ready(sk) ? EPOLLIN | EPOLLRDNORM : 0;
+	if (mptcp_epollin_ready(sk))
+		return true;
+	return sk_is_readable(sk);
 }
 
 static void mptcp_check_listen_stop(struct sock *sk)
@@ -4342,7 +4344,8 @@ static __poll_t mptcp_poll(struct file *file, struct socket *sock,
 		mask |= EPOLLIN | EPOLLRDNORM | EPOLLRDHUP;
 
 	if (state != TCP_SYN_SENT && state != TCP_SYN_RECV) {
-		mask |= mptcp_check_readable(sk);
+		if (mptcp_stream_is_readable(sk))
+			mask |= EPOLLIN | EPOLLRDNORM;
 		if (shutdown & SEND_SHUTDOWN)
 			mask |= EPOLLOUT | EPOLLWRNORM;
 		else
-- 
2.53.0

[RFC mptcp-next v20 05/15] mptcp: implement tls_mptcp_ops

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

This patch implements the MPTCP-specific struct tls_prot_ops, named
'tls_mptcp_ops'.

Passing an MPTCP socket to tcp_sock_rate_check_app_limited() can
trigger a crash. Here, an MPTCP version of check_app_limited() is
implemented, which calls tcp_sock_rate_check_app_limited() for each
subflow.

When MPTCP implements lock_is_held interface, it not only checks
sock_owned_by_user_nocheck(sk) as TCP does, but also needs to check
whether the MPTCP data lock is held. This is required because TLS
may call lock_is_held from softirq context with bh_lock_sock held.
Checking both conditions ensures TLS always defers to workqueue when
the MPTCP data lock is held, avoiding deadlock.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 include/net/mptcp.h  |   2 +
 include/net/tcp.h    |   1 +
 net/ipv4/tcp.c       |   9 +++-
 net/mptcp/protocol.c | 121 +++++++++++++++++++++++++++++++++++++++++--
 net/mptcp/protocol.h |   1 +
 net/tls/tls_main.c   |  13 +++++
 6 files changed, 141 insertions(+), 6 deletions(-)

diff --git a/include/net/mptcp.h b/include/net/mptcp.h
index 4cf59e83c1c5..02564eceeb7e 100644
--- a/include/net/mptcp.h
+++ b/include/net/mptcp.h
@@ -132,6 +132,8 @@ struct mptcp_pm_ops {
 	void (*release)(struct mptcp_sock *msk);
 } ____cacheline_aligned_in_smp;
 
+extern struct tls_prot_ops tls_mptcp_ops;
+
 #ifdef CONFIG_MPTCP
 void mptcp_init(void);
 
diff --git a/include/net/tcp.h b/include/net/tcp.h
index f063eccbbba3..1c8201f69ef1 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -849,6 +849,7 @@ static inline int tcp_bound_to_half_wnd(struct tcp_sock *tp, int pktsize)
 
 /* tcp.c */
 void tcp_get_info(struct sock *, struct tcp_info *);
+void tcp_sock_rate_check_app_limited(struct tcp_sock *tp);
 void tcp_rate_check_app_limited(struct sock *sk);
 
 /* Read 'sendfile()'-style from a TCP socket */
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index a058f350a759..bdad459e6605 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1097,9 +1097,9 @@ int tcp_sendmsg_fastopen(struct sock *sk, struct msghdr *msg, int *copied,
 }
 
 /* If a gap is detected between sends, mark the socket application-limited. */
-void tcp_rate_check_app_limited(struct sock *sk)
+void tcp_sock_rate_check_app_limited(struct tcp_sock *tp)
 {
-	struct tcp_sock *tp = tcp_sk(sk);
+	struct sock *sk = (struct sock *)tp;
 
 	if (/* We have less than one packet to send. */
 	    tp->write_seq - tp->snd_nxt < tp->mss_cache &&
@@ -1112,6 +1112,11 @@ void tcp_rate_check_app_limited(struct sock *sk)
 		tp->app_limited =
 			(tp->delivered + tcp_packets_in_flight(tp)) ? : 1;
 }
+
+void tcp_rate_check_app_limited(struct sock *sk)
+{
+	tcp_sock_rate_check_app_limited(tcp_sk(sk));
+}
 EXPORT_SYMBOL_GPL(tcp_rate_check_app_limited);
 
 int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index a36ef97155a7..505eac23f35d 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -24,6 +24,7 @@
 #include <net/mptcp.h>
 #include <net/hotdata.h>
 #include <net/xfrm.h>
+#include <net/tls.h>
 #include <asm/ioctls.h>
 #include "protocol.h"
 #include "mib.h"
@@ -1909,7 +1910,7 @@ static void mptcp_rps_record_subflows(const struct mptcp_sock *msk)
 	}
 }
 
-static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
+static int mptcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t len)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
 	struct page_frag *pfrag;
@@ -1921,8 +1922,6 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 	msg->msg_flags &= MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL |
 			  MSG_FASTOPEN | MSG_EOR;
 
-	lock_sock(sk);
-
 	mptcp_rps_record_subflows(msk);
 
 	if (unlikely(inet_test_bit(DEFER_CONNECT, sk) ||
@@ -2038,7 +2037,6 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 	}
 
 out:
-	release_sock(sk);
 	return copied;
 
 do_error:
@@ -2049,6 +2047,17 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 	goto out;
 }
 
+static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
+{
+	int ret;
+
+	lock_sock(sk);
+	ret = mptcp_sendmsg_locked(sk, msg, len);
+	release_sock(sk);
+
+	return ret;
+}
+
 static void mptcp_rcv_space_adjust(struct mptcp_sock *msk, int copied);
 
 static void mptcp_eat_recv_skb(struct sock *sk, struct sk_buff *skb)
@@ -4726,3 +4735,107 @@ int __init mptcp_proto_v6_init(void)
 	return err;
 }
 #endif
+
+static int mptcp_inq(struct sock *sk)
+{
+	const struct mptcp_sock *msk = mptcp_sk(sk);
+	const struct sk_buff *skb;
+
+	if ((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV))
+		return 0;
+
+	skb = skb_peek(&sk->sk_receive_queue);
+	if (skb) {
+		u64 answ = READ_ONCE(msk->ack_seq) - MPTCP_SKB_CB(skb)->map_seq;
+
+		if (answ >= INT_MAX)
+			answ = INT_MAX;
+
+		/* Subtract 1, if FIN was received */
+		if (answ &&
+		    (sk->sk_state == TCP_CLOSE ||
+		     (sk->sk_shutdown & RCV_SHUTDOWN)))
+			answ--;
+
+		return (int)answ;
+	}
+
+	return 0;
+}
+
+static bool mptcp_lock_is_held(struct sock *sk)
+{
+	return sock_owned_by_user_nocheck(sk) ||
+	       mptcp_data_is_locked(sk);
+}
+
+static void mptcp_read_done(struct sock *sk, size_t len)
+{
+	struct mptcp_sock *msk = mptcp_sk(sk);
+	struct sk_buff *skb;
+	size_t left;
+	u32 offset;
+
+	msk_owned_by_me(msk);
+
+	if (sk->sk_state == TCP_LISTEN)
+		return;
+
+	left = len;
+	while (left && (skb = mptcp_recv_skb(sk, &offset)) != NULL) {
+		int used;
+
+		used = min_t(size_t, skb->len - offset, left);
+		msk->bytes_consumed += used;
+		MPTCP_SKB_CB(skb)->offset += used;
+		MPTCP_SKB_CB(skb)->map_seq += used;
+		left -= used;
+
+		if (skb->len > offset + used)
+			break;
+
+		mptcp_eat_recv_skb(sk, skb);
+	}
+
+	mptcp_rcv_space_adjust(msk, len - left);
+
+	/* Clean up data we have read: This will do ACK frames. */
+	if (left != len)
+		mptcp_cleanup_rbuf(msk, len - left);
+}
+
+static u32 mptcp_get_skb_seq(struct sk_buff *skb)
+{
+	return MPTCP_SKB_CB(skb)->map_seq - MPTCP_SKB_CB(skb)->offset;
+}
+
+static void mptcp_check_app_limited(struct sock *sk)
+{
+	struct mptcp_sock *msk = mptcp_sk(sk);
+	struct mptcp_subflow_context *subflow;
+
+	mptcp_for_each_subflow(msk, subflow) {
+		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
+		bool slow;
+
+		slow = lock_sock_fast(ssk);
+		tcp_sock_rate_check_app_limited(tcp_sk(ssk));
+		unlock_sock_fast(ssk, slow);
+	}
+}
+
+struct tls_prot_ops tls_mptcp_ops = {
+	.owner			= THIS_MODULE,
+	.protocol		= IPPROTO_MPTCP,
+	.inq			= mptcp_inq,
+	.sendmsg_locked		= mptcp_sendmsg_locked,
+	.recv_skb		= mptcp_recv_skb,
+	.lock_is_held		= mptcp_lock_is_held,
+	.read_sock		= mptcp_read_sock,
+	.read_done		= mptcp_read_done,
+	.get_skb_seq		= mptcp_get_skb_seq,
+	.poll			= mptcp_poll,
+	.epollin_ready		= mptcp_epollin_ready,
+	.check_app_limited	= mptcp_check_app_limited,
+};
+EXPORT_SYMBOL(tls_mptcp_ops);
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 661600f8b573..1c604a1ded6f 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -380,6 +380,7 @@ struct mptcp_sock {
 
 #define mptcp_data_lock(sk) spin_lock_bh(&(sk)->sk_lock.slock)
 #define mptcp_data_unlock(sk) spin_unlock_bh(&(sk)->sk_lock.slock)
+#define mptcp_data_is_locked(sk) spin_is_locked(&(sk)->sk_lock.slock)
 
 #define mptcp_for_each_subflow(__msk, __subflow)			\
 	list_for_each_entry(__subflow, &((__msk)->conn_list), node)
diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
index c840a228af0d..fbb60dc1832e 100644
--- a/net/tls/tls_main.c
+++ b/net/tls/tls_main.c
@@ -1425,6 +1425,12 @@ static int __init tls_register(void)
 	if (err)
 		goto err_strp;
 
+#ifdef CONFIG_MPTCP
+	err = tls_register_prot_ops(&tls_mptcp_ops);
+	if (err)
+		goto err_tcp;
+#endif
+
 	err = tls_device_init();
 	if (err)
 		goto err_ops;
@@ -1433,6 +1439,10 @@ static int __init tls_register(void)
 
 	return 0;
 err_ops:
+#ifdef CONFIG_MPTCP
+	tls_unregister_prot_ops(&tls_mptcp_ops);
+err_tcp:
+#endif
 	tls_unregister_prot_ops(&tls_tcp_ops);
 err_strp:
 	tls_strp_dev_exit();
@@ -1444,6 +1454,9 @@ static int __init tls_register(void)
 static void __exit tls_unregister(void)
 {
 	tcp_unregister_ulp(&tcp_tls_ulp_ops);
+#ifdef CONFIG_MPTCP
+	tls_unregister_prot_ops(&tls_mptcp_ops);
+#endif
 	tls_unregister_prot_ops(&tls_tcp_ops);
 	tls_proto_cleanup();
 	tls_strp_dev_exit();
-- 
2.53.0

[RFC mptcp-next v20 06/15] tls: disable device offload for mptcp sockets

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

MPTCP TLS hardware offload is not yet implemented. Return -EOPNOTSUPP
when attempting to enable device offload on MPTCP sockets.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/tls/tls_device.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/tls/tls_device.c b/net/tls/tls_device.c
index 741aef09bfd3..06f45edffb5f 100644
--- a/net/tls/tls_device.c
+++ b/net/tls/tls_device.c
@@ -1074,6 +1074,9 @@ int tls_set_device_offload(struct sock *sk)
 	ctx = tls_get_ctx(sk);
 	prot = &ctx->prot_info;
 
+	if (sk->sk_protocol == IPPROTO_MPTCP)
+		return -EOPNOTSUPP;
+
 	if (ctx->priv_ctx_tx)
 		return -EEXIST;
 
@@ -1196,6 +1199,9 @@ int tls_set_device_offload_rx(struct sock *sk, struct tls_context *ctx)
 	struct net_device *netdev;
 	int rc = 0;
 
+	if (sk->sk_protocol == IPPROTO_MPTCP)
+		return -EOPNOTSUPP;
+
 	if (ctx->crypto_recv.info.version != TLS_1_2_VERSION)
 		return -EOPNOTSUPP;
 
-- 
2.53.0

[RFC mptcp-next v20 07/15] mptcp: update ulp getsockopt for tls support

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

This patch extracts TCP_ULP getsockopt operation into a tcp_sock_get_ulp()
helper so that it can also be used in MPTCP.

TCP_ULP was obtained by calling mptcp_getsockopt_first_sf_only() to get
ULP of the first subflow. Now that the mechanism has changed, a new helper
mptcp_getsockopt_tcp_ulp() is added to get ULP of msk.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 include/linux/tcp.h |  1 +
 net/ipv4/tcp.c      | 36 ++++++++++++++++++++++--------------
 net/mptcp/sockopt.c | 18 ++++++++++++++++++
 3 files changed, 41 insertions(+), 14 deletions(-)

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 8a6807082672..0026b6798b2f 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -653,6 +653,7 @@ void tcp_sock_set_quickack(struct sock *sk, int val);
 int tcp_sock_set_syncnt(struct sock *sk, int val);
 int tcp_sock_set_user_timeout(struct sock *sk, int val);
 int tcp_sock_set_maxseg(struct sock *sk, int val);
+int tcp_sock_get_ulp(struct sock *sk, sockptr_t optval, sockptr_t optlen);
 
 static inline bool dst_tcp_usec_ts(const struct dst_entry *dst)
 {
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index bdad459e6605..8e41ce029985 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -4486,6 +4486,27 @@ struct sk_buff *tcp_get_timestamping_opt_stats(const struct sock *sk,
 	return stats;
 }
 
+int tcp_sock_get_ulp(struct sock *sk, sockptr_t optval, sockptr_t optlen)
+{
+	struct inet_connection_sock *icsk = inet_csk(sk);
+	int len;
+
+	if (copy_from_sockptr(&len, optlen, sizeof(int)))
+		return -EFAULT;
+	len = min_t(unsigned int, len, TCP_ULP_NAME_MAX);
+	if (!icsk->icsk_ulp_ops) {
+		len = 0;
+		if (copy_to_sockptr(optlen, &len, sizeof(int)))
+			return -EFAULT;
+		return 0;
+	}
+	if (copy_to_sockptr(optlen, &len, sizeof(int)))
+		return -EFAULT;
+	if (copy_to_sockptr(optval, icsk->icsk_ulp_ops->name, len))
+		return -EFAULT;
+	return 0;
+}
+
 int do_tcp_getsockopt(struct sock *sk, int level,
 		      int optname, sockptr_t optval, sockptr_t optlen)
 {
@@ -4595,20 +4616,7 @@ int do_tcp_getsockopt(struct sock *sk, int level,
 		return 0;
 
 	case TCP_ULP:
-		if (copy_from_sockptr(&len, optlen, sizeof(int)))
-			return -EFAULT;
-		len = min_t(unsigned int, len, TCP_ULP_NAME_MAX);
-		if (!icsk->icsk_ulp_ops) {
-			len = 0;
-			if (copy_to_sockptr(optlen, &len, sizeof(int)))
-				return -EFAULT;
-			return 0;
-		}
-		if (copy_to_sockptr(optlen, &len, sizeof(int)))
-			return -EFAULT;
-		if (copy_to_sockptr(optval, icsk->icsk_ulp_ops->name, len))
-			return -EFAULT;
-		return 0;
+		return tcp_sock_get_ulp(sk, optval, optlen);
 
 	case TCP_FASTOPEN_KEY: {
 		u64 key[TCP_FASTOPEN_KEY_BUF_LENGTH / sizeof(u64)];
diff --git a/net/mptcp/sockopt.c b/net/mptcp/sockopt.c
index 87b5796d0135..ee1e71c178fa 100644
--- a/net/mptcp/sockopt.c
+++ b/net/mptcp/sockopt.c
@@ -1403,6 +1403,23 @@ static int mptcp_put_int_option(struct mptcp_sock *msk, char __user *optval,
 	return 0;
 }
 
+static int mptcp_getsockopt_tcp_ulp(struct sock *sk, char __user *optval,
+				    int __user *optlen)
+{
+	int ret, len;
+
+	if (copy_from_sockptr(&len, USER_SOCKPTR(optlen), sizeof(int)))
+		return -EFAULT;
+
+	if (len < 0)
+		return -EINVAL;
+
+	lock_sock(sk);
+	ret = tcp_sock_get_ulp(sk, USER_SOCKPTR(optval), USER_SOCKPTR(optlen));
+	release_sock(sk);
+	return ret;
+}
+
 static int mptcp_getsockopt_sol_tcp(struct mptcp_sock *msk, int optname,
 				    char __user *optval, int __user *optlen)
 {
@@ -1410,6 +1427,7 @@ static int mptcp_getsockopt_sol_tcp(struct mptcp_sock *msk, int optname,
 
 	switch (optname) {
 	case TCP_ULP:
+		return mptcp_getsockopt_tcp_ulp(sk, optval, optlen);
 	case TCP_CONGESTION:
 	case TCP_INFO:
 	case TCP_CC_INFO:
-- 
2.53.0

[RFC mptcp-next v20 08/15] mptcp: enable ulp setsockopt for tls support

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

Allow MPTCP sockets to set the TCP_ULP socket option to enable TLS.
Add mptcp_setsockopt_tcp_ulp() which validates the socket state (must
not be CLOSE or LISTEN), only accepts "tls" as the ULP name, and then
calls tcp_set_ulp().

Include TCP_ULP in the list of supported options in supported_sockopt(),
and handle it in setsockopt_sol_tcp() instead of returning -EOPNOTSUPP.

Call tcp_cleanup_ulp() in mptcp_destroy_common() to release ULP module's
reference count.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/protocol.c |  1 +
 net/mptcp/sockopt.c  | 40 +++++++++++++++++++++++++++++++++++++++-
 2 files changed, 40 insertions(+), 1 deletion(-)

diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index 505eac23f35d..06e037fe8807 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -3640,6 +3640,7 @@ static void mptcp_destroy(struct sock *sk)
 	/* allow the following to close even the initial subflow */
 	msk->free_first = 1;
 	mptcp_destroy_common(msk);
+	tcp_cleanup_ulp(sk);
 	sk_sockets_allocated_dec(sk);
 }
 
diff --git a/net/mptcp/sockopt.c b/net/mptcp/sockopt.c
index ee1e71c178fa..a9fa3af9efcc 100644
--- a/net/mptcp/sockopt.c
+++ b/net/mptcp/sockopt.c
@@ -12,6 +12,7 @@
 #include <net/protocol.h>
 #include <net/tcp.h>
 #include <net/mptcp.h>
+#include <net/tls.h>
 #include "protocol.h"
 
 #define MIN_INFO_OPTLEN_SIZE		16
@@ -573,6 +574,7 @@ static bool mptcp_supported_sockopt(int level, int optname)
 		case TCP_FASTOPEN_CONNECT:
 		case TCP_FASTOPEN_KEY:
 		case TCP_FASTOPEN_NO_COOKIE:
+		case TCP_ULP:
 			return true;
 		}
 
@@ -825,6 +827,42 @@ static int mptcp_setsockopt_all_sf(struct mptcp_sock *msk, int level,
 	return ret;
 }
 
+static int mptcp_setsockopt_tcp_ulp(struct sock *sk, sockptr_t optval,
+				    unsigned int optlen)
+{
+	struct mptcp_sock *msk = mptcp_sk(sk);
+	char name[TCP_ULP_NAME_MAX];
+	int err = 0;
+	size_t len;
+	int val;
+
+	if (optlen < 1)
+		return -EINVAL;
+
+	len = min_t(long, TCP_ULP_NAME_MAX - 1, optlen);
+	val = strncpy_from_sockptr(name, optval, len);
+	if (val < 0)
+		return -EFAULT;
+	name[val] = 0;
+
+	if (strcmp(name, "tls"))
+		return -EOPNOTSUPP;
+
+	sockopt_lock_sock(sk);
+	if (__mptcp_check_fallback(msk)) {
+		err = -EOPNOTSUPP;
+		goto out;
+	}
+	if ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)) {
+		err = -ENOTCONN;
+		goto out;
+	}
+	err = tcp_set_ulp(sk, name);
+out:
+	sockopt_release_sock(sk);
+	return err;
+}
+
 static int mptcp_setsockopt_sol_tcp(struct mptcp_sock *msk, int optname,
 				    sockptr_t optval, unsigned int optlen)
 {
@@ -833,7 +871,7 @@ static int mptcp_setsockopt_sol_tcp(struct mptcp_sock *msk, int optname,
 
 	switch (optname) {
 	case TCP_ULP:
-		return -EOPNOTSUPP;
+		return mptcp_setsockopt_tcp_ulp(sk, optval, optlen);
 	case TCP_CONGESTION:
 		return mptcp_setsockopt_sol_tcp_congestion(msk, optval, optlen);
 	case TCP_DEFER_ACCEPT:
-- 
2.53.0

[RFC mptcp-next v20 09/15] selftests: mptcp: connect: use espintcp for ulp test

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

With KTLS being implemented, "tls" should no longer be used in
sock_test_tcpulp(), it breaks mptcp_connect.sh tests. Another ULP
name, "espintcp", is set instead in this patch.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 tools/testing/selftests/net/mptcp/config          | 1 +
 tools/testing/selftests/net/mptcp/mptcp_connect.c | 4 ++--
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/net/mptcp/config b/tools/testing/selftests/net/mptcp/config
index 59051ee2a986..b9960f227796 100644
--- a/tools/testing/selftests/net/mptcp/config
+++ b/tools/testing/selftests/net/mptcp/config
@@ -34,3 +34,4 @@ CONFIG_NFT_SOCKET=m
 CONFIG_NFT_TPROXY=m
 CONFIG_SYN_COOKIES=y
 CONFIG_VETH=y
+CONFIG_XFRM_ESPINTCP=m
diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.c b/tools/testing/selftests/net/mptcp/mptcp_connect.c
index cbe573c4ab3a..299a7a02d6f5 100644
--- a/tools/testing/selftests/net/mptcp/mptcp_connect.c
+++ b/tools/testing/selftests/net/mptcp/mptcp_connect.c
@@ -285,11 +285,11 @@ static void sock_test_tcpulp(int sock, int proto, unsigned int line)
 	if (buflen > 0) {
 		if (strcmp(buf, "mptcp") != 0)
 			xerror("unexpected ULP '%s' for proto %d at line %u", buf, proto, line);
-		ret = do_ulp_so(sock, "tls");
+		ret = do_ulp_so(sock, "espintcp");
 		if (ret == 0)
 			X("setsockopt");
 	} else if (proto == IPPROTO_MPTCP) {
-		ret = do_ulp_so(sock, "tls");
+		ret = do_ulp_so(sock, "espintcp");
 		if (ret != -1)
 			X("setsockopt");
 	}
-- 
2.53.0

[RFC mptcp-next v20 10/15] selftests: tls: add mptcp variant for testing

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

To enable easy MPTCP socket creation in MPTCP TLS tests, two protocol
parameters (cli_proto and srv_proto) have been added to ulp_sock_pair().
These are passed as third arguments of socket(): 0 creates TCP sockets,
IPPROTO_MPTCP creates MPTCP sockets.

A new variant "mptcp" is added both in FIXTURE_VARIANT(tls) to control
whether to create MPTCP sockets or not for tests.

Add is_mptcp_enable() helper to check MPTCP support. Used in
FIXTURE_SETUP(tls) to skip MPTCP variants when MPTCP is not enabled.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 tools/testing/selftests/net/tls.c | 44 +++++++++++++++++++++++++++----
 1 file changed, 39 insertions(+), 5 deletions(-)

diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c
index 9b9a3cb2700d..5f8382a51558 100644
--- a/tools/testing/selftests/net/tls.c
+++ b/tools/testing/selftests/net/tls.c
@@ -26,6 +26,10 @@
 #define TLS_PAYLOAD_MAX_LEN 16384
 #define SOL_TLS 282
 
+#ifndef IPPROTO_MPTCP
+#define IPPROTO_MPTCP 262
+#endif
+
 static int fips_enabled;
 
 struct tls_crypto_info_keys {
@@ -108,8 +112,9 @@ static void memrnd(void *s, size_t n)
 		*byte++ = rand();
 }
 
-static void ulp_sock_pair(struct __test_metadata *_metadata,
-			  int *fd, int *cfd, bool *notls)
+static void __ulp_sock_pair(struct __test_metadata *_metadata,
+			    int *fd, int *cfd, bool *notls,
+			    int cli_proto, int srv_proto)
 {
 	struct sockaddr_in addr;
 	socklen_t len;
@@ -122,8 +127,8 @@ static void ulp_sock_pair(struct __test_metadata *_metadata,
 	addr.sin_addr.s_addr = htonl(INADDR_ANY);
 	addr.sin_port = 0;
 
-	*fd = socket(AF_INET, SOCK_STREAM, 0);
-	sfd = socket(AF_INET, SOCK_STREAM, 0);
+	*fd = socket(AF_INET, SOCK_STREAM, cli_proto);
+	sfd = socket(AF_INET, SOCK_STREAM, srv_proto);
 
 	ret = bind(sfd, &addr, sizeof(addr));
 	ASSERT_EQ(ret, 0);
@@ -153,6 +158,12 @@ static void ulp_sock_pair(struct __test_metadata *_metadata,
 	ASSERT_EQ(ret, 0);
 }
 
+static void ulp_sock_pair(struct __test_metadata *_metadata,
+			  int *fd, int *cfd, bool *notls)
+{
+	__ulp_sock_pair(_metadata, fd, cfd, notls, 0, 0);
+}
+
 /* Produce a basic cmsg */
 static int tls_send_cmsg(int fd, unsigned char record_type,
 			 void *data, size_t len, int flags)
@@ -310,6 +321,7 @@ FIXTURE_VARIANT(tls)
 	uint16_t tls_version;
 	uint16_t cipher_type;
 	bool nopad, fips_non_compliant;
+	bool mptcp;
 };
 
 FIXTURE_VARIANT_ADD(tls, 12_aes_gcm)
@@ -395,6 +407,23 @@ FIXTURE_VARIANT_ADD(tls, 12_aria_gcm_256)
 	.cipher_type = TLS_CIPHER_ARIA_GCM_256,
 };
 
+static bool is_mptcp_enable()
+{
+	char buf[16] = { 0 };
+	ssize_t n;
+	int fd;
+
+	fd = open("/proc/sys/net/mptcp/enabled", O_RDONLY);
+	if (fd < 0)
+		return false;
+
+	n = read(fd, buf, sizeof(buf) - 1);
+	close(fd);
+	if (n <= 0)
+		return false;
+	return (atoi(buf) == 1);
+}
+
 FIXTURE_SETUP(tls)
 {
 	struct tls_crypto_info_keys tls12;
@@ -404,10 +433,15 @@ FIXTURE_SETUP(tls)
 	if (fips_enabled && variant->fips_non_compliant)
 		SKIP(return, "Unsupported cipher in FIPS mode");
 
+	if (variant->mptcp && !is_mptcp_enable())
+		SKIP(return, "no MPTCP support");
+
 	tls_crypto_info_init(variant->tls_version, variant->cipher_type,
 			     &tls12, 0);
 
-	ulp_sock_pair(_metadata, &self->fd, &self->cfd, &self->notls);
+	__ulp_sock_pair(_metadata, &self->fd, &self->cfd, &self->notls,
+			variant->mptcp ? IPPROTO_MPTCP : 0,
+			variant->mptcp ? IPPROTO_MPTCP : 0);
 
 	if (self->notls)
 		return;
-- 
2.53.0

[RFC mptcp-next v20 11/15] selftests: tls: increase pollin timeouts for mptcp

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

MPTCP requires longer timeouts in pollin test due to subflow establishment
delays and slower state transitions. Increase timeout values to prevent
false failures:

 #  RUN           tls.13_sm4_ccm_mptcp.pollin ...
 # tls.c:1411:pollin:Expected poll(&fd, 1, 20) (0) == 1 (1)
 # tls.c:1412:pollin:Expected fd.revents & POLLIN (0) == 1 (1)
 # pollin: Test failed
 #          FAIL  tls.13_sm4_ccm_mptcp.pollin
 not ok 357 tls.13_sm4_ccm_mptcp.pollin

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 tools/testing/selftests/net/tls.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c
index 5f8382a51558..95875e979ea7 100644
--- a/tools/testing/selftests/net/tls.c
+++ b/tools/testing/selftests/net/tls.c
@@ -1362,6 +1362,7 @@ TEST_F(tls, bidir)
 
 TEST_F(tls, pollin)
 {
+	int timeout = variant->mptcp ? 100 : 20;
 	char const *test_str = "test_poll";
 	struct pollfd fd = { 0, 0, 0 };
 	char buf[10];
@@ -1371,11 +1372,11 @@ TEST_F(tls, pollin)
 	fd.fd = self->cfd;
 	fd.events = POLLIN;
 
-	EXPECT_EQ(poll(&fd, 1, 20), 1);
+	EXPECT_EQ(poll(&fd, 1, timeout), 1);
 	EXPECT_EQ(fd.revents & POLLIN, 1);
 	EXPECT_EQ(recv(self->cfd, buf, send_len, MSG_WAITALL), send_len);
 	/* Test timing out */
-	EXPECT_EQ(poll(&fd, 1, 20), 0);
+	EXPECT_EQ(poll(&fd, 1, timeout), 0);
 }
 
 TEST_F(tls, poll_wait)
-- 
2.53.0

[RFC mptcp-next v20 12/15] selftests: tls: increase nonblocking data size for mptcp

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

Increase the data size in nonblocking tests to accommodate MPTCP's
multi-subflow behavior and ensure sufficient data for testing,
avoiding the following errors:

 #  RUN           tls.12_aria_gcm_mptcp.nonblocking ...
 # tls.c:1534:nonblocking:Expected 0 (0) != eagain (0)
 # nonblocking: Test failed

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 tools/testing/selftests/net/tls.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c
index 95875e979ea7..1d2a5661785f 100644
--- a/tools/testing/selftests/net/tls.c
+++ b/tools/testing/selftests/net/tls.c
@@ -1468,6 +1468,9 @@ TEST_F(tls, nonblocking)
 	int flags;
 	int res;
 
+	if (variant->mptcp)
+		data *= 4;
+
 	flags = fcntl(self->fd, F_GETFL, 0);
 	fcntl(self->fd, F_SETFL, flags | O_NONBLOCK);
 	fcntl(self->cfd, F_SETFL, flags | O_NONBLOCK);
-- 
2.53.0

[RFC mptcp-next v20 13/15] selftests: tls: retry bind on EINVAL in shutdown_reuse

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

In the shutdown_reuse test, after shutdown and close, bind() may fail
with EINVAL for MPTCP sockets due to asynchronous state transition of
the top-level MPTCP socket. The subflow may have reached TCP_CLOSE,
but the MPTCP socket state hasn't been updated yet.

Retry bind() on EINVAL for up to 1000 iterations (1 second) to allow
the MPTCP socket to complete its state transition.

This fixes the following intermittent failures:

 #  RUN           tls.12_aes_gcm_mptcp.shutdown_reuse ...
 # tls.c:1790:shutdown_reuse:Expected ret (-1) == 0 (0)
 # shutdown_reuse: Test failed
 #          FAIL  tls.12_aes_gcm_mptcp.shutdown_reuse
 not ok 14 tls.12_aes_gcm_mptcp.shutdown_reuse
 #  RUN           tls.13_aes_gcm_mptcp.shutdown_reuse ...
 # tls.c:1790:shutdown_reuse:Expected ret (-1) == 0 (0)
 # shutdown_reuse: Test failed
 #          FAIL  tls.13_aes_gcm_mptcp.shutdown_reuse
 not ok 15 tls.13_aes_gcm_mptcp.shutdown_reuse
 #  RUN           tls.12_chacha_mptcp.shutdown_reuse ...
 #            OK  tls.12_chacha_mptcp.shutdown_reuse
 ok 16 tls.12_chacha_mptcp.shutdown_reuse
 #  RUN           tls.13_chacha_mptcp.shutdown_reuse ...
 #            OK  tls.13_chacha_mptcp.shutdown_reuse
 ok 17 tls.13_chacha_mptcp.shutdown_reuse
 #  RUN           tls.13_sm4_gcm_mptcp.shutdown_reuse ...
 # tls.c:1790:shutdown_reuse:Expected ret (-1) == 0 (0)
 # shutdown_reuse: Test failed
 #          FAIL  tls.13_sm4_gcm_mptcp.shutdown_reuse
 not ok 18 tls.13_sm4_gcm_mptcp.shutdown_reuse

This is only done for MPTCP variants to avoid slowing down plain TCP
tests.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 tools/testing/selftests/net/tls.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c
index 1d2a5661785f..c0737c410db8 100644
--- a/tools/testing/selftests/net/tls.c
+++ b/tools/testing/selftests/net/tls.c
@@ -1744,6 +1744,7 @@ TEST_F(tls, shutdown_unsent)
 TEST_F(tls, shutdown_reuse)
 {
 	struct sockaddr_in addr;
+	int i = 0;
 	int ret;
 
 	shutdown(self->fd, SHUT_RDWR);
@@ -1754,7 +1755,13 @@ TEST_F(tls, shutdown_reuse)
 	addr.sin_addr.s_addr = htonl(INADDR_ANY);
 	addr.sin_port = 0;
 
+retry:
 	ret = bind(self->fd, &addr, sizeof(addr));
+	if (variant->mptcp &&
+	    ret < 0 && errno == EINVAL && i++ < 1000) {
+		usleep(1000);
+		goto retry;
+	}
 	EXPECT_EQ(ret, 0);
 	ret = listen(self->fd, 10);
 	EXPECT_EQ(ret, -1);
-- 
2.53.0

[RFC mptcp-next v20 14/15] selftests: tls: add mptcp test cases

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

This patch introduces MPTCP test cases for the TLS fixture. These "mptcp"
variants are configured to create MPTCP sockets specifically for MPTCP TLS
testing purposes.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 tools/testing/selftests/net/tls.c | 96 +++++++++++++++++++++++++++++++
 1 file changed, 96 insertions(+)

diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c
index c0737c410db8..d41b81ba256f 100644
--- a/tools/testing/selftests/net/tls.c
+++ b/tools/testing/selftests/net/tls.c
@@ -407,6 +407,102 @@ FIXTURE_VARIANT_ADD(tls, 12_aria_gcm_256)
 	.cipher_type = TLS_CIPHER_ARIA_GCM_256,
 };
 
+FIXTURE_VARIANT_ADD(tls, 12_aes_gcm_mptcp)
+{
+	.tls_version = TLS_1_2_VERSION,
+	.cipher_type = TLS_CIPHER_AES_GCM_128,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 13_aes_gcm_mptcp)
+{
+	.tls_version = TLS_1_3_VERSION,
+	.cipher_type = TLS_CIPHER_AES_GCM_128,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 12_chacha_mptcp)
+{
+	.tls_version = TLS_1_2_VERSION,
+	.cipher_type = TLS_CIPHER_CHACHA20_POLY1305,
+	.fips_non_compliant = true,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 13_chacha_mptcp)
+{
+	.tls_version = TLS_1_3_VERSION,
+	.cipher_type = TLS_CIPHER_CHACHA20_POLY1305,
+	.fips_non_compliant = true,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 13_sm4_gcm_mptcp)
+{
+	.tls_version = TLS_1_3_VERSION,
+	.cipher_type = TLS_CIPHER_SM4_GCM,
+	.fips_non_compliant = true,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 13_sm4_ccm_mptcp)
+{
+	.tls_version = TLS_1_3_VERSION,
+	.cipher_type = TLS_CIPHER_SM4_CCM,
+	.fips_non_compliant = true,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 12_aes_ccm_mptcp)
+{
+	.tls_version = TLS_1_2_VERSION,
+	.cipher_type = TLS_CIPHER_AES_CCM_128,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 13_aes_ccm_mptcp)
+{
+	.tls_version = TLS_1_3_VERSION,
+	.cipher_type = TLS_CIPHER_AES_CCM_128,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 12_aes_gcm_256_mptcp)
+{
+	.tls_version = TLS_1_2_VERSION,
+	.cipher_type = TLS_CIPHER_AES_GCM_256,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 13_aes_gcm_256_mptcp)
+{
+	.tls_version = TLS_1_3_VERSION,
+	.cipher_type = TLS_CIPHER_AES_GCM_256,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 13_nopad_mptcp)
+{
+	.tls_version = TLS_1_3_VERSION,
+	.cipher_type = TLS_CIPHER_AES_GCM_128,
+	.nopad = true,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 12_aria_gcm_mptcp)
+{
+	.tls_version = TLS_1_2_VERSION,
+	.cipher_type = TLS_CIPHER_ARIA_GCM_128,
+	.mptcp = true,
+};
+
+FIXTURE_VARIANT_ADD(tls, 12_aria_gcm_256_mptcp)
+{
+	.tls_version = TLS_1_2_VERSION,
+	.cipher_type = TLS_CIPHER_ARIA_GCM_256,
+	.mptcp = true,
+};
+
 static bool is_mptcp_enable()
 {
 	char buf[16] = { 0 };
-- 
2.53.0

[RFC mptcp-next v20 15/15] selftests: mptcp: cover mptcp tls tests

[Geliang Tang]

From: Geliang Tang <tanggeliang@kylinos.cn>

The mptcp tests for tls.c is available now, this patch adds mptcp_tls.sh
to test it in the MPTCP CI by default.

Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 tools/testing/selftests/net/mptcp/.gitignore  |  1 +
 tools/testing/selftests/net/mptcp/Makefile    |  2 +
 tools/testing/selftests/net/mptcp/config      |  5 ++
 .../testing/selftests/net/mptcp/mptcp_tls.sh  | 64 +++++++++++++++++++
 tools/testing/selftests/net/mptcp/tls.c       |  1 +
 5 files changed, 73 insertions(+)
 create mode 100755 tools/testing/selftests/net/mptcp/mptcp_tls.sh
 create mode 120000 tools/testing/selftests/net/mptcp/tls.c

diff --git a/tools/testing/selftests/net/mptcp/.gitignore b/tools/testing/selftests/net/mptcp/.gitignore
index 833279fb34e2..f6defec6eeb5 100644
--- a/tools/testing/selftests/net/mptcp/.gitignore
+++ b/tools/testing/selftests/net/mptcp/.gitignore
@@ -4,4 +4,5 @@ mptcp_diag
 mptcp_inq
 mptcp_sockopt
 pm_nl_ctl
+tls
 *.pcap
diff --git a/tools/testing/selftests/net/mptcp/Makefile b/tools/testing/selftests/net/mptcp/Makefile
index 22ba0da2adb8..f7c959a25b3b 100644
--- a/tools/testing/selftests/net/mptcp/Makefile
+++ b/tools/testing/selftests/net/mptcp/Makefile
@@ -14,6 +14,7 @@ TEST_PROGS := \
 	mptcp_connect_splice.sh \
 	mptcp_join.sh \
 	mptcp_sockopt.sh \
+	mptcp_tls.sh \
 	pm_netlink.sh \
 	simult_flows.sh \
 	userspace_pm.sh \
@@ -25,6 +26,7 @@ TEST_GEN_FILES := \
 	mptcp_inq \
 	mptcp_sockopt \
 	pm_nl_ctl \
+	tls \
 # end of TEST_GEN_FILES
 
 TEST_FILES := \
diff --git a/tools/testing/selftests/net/mptcp/config b/tools/testing/selftests/net/mptcp/config
index b9960f227796..395f1d5b020a 100644
--- a/tools/testing/selftests/net/mptcp/config
+++ b/tools/testing/selftests/net/mptcp/config
@@ -35,3 +35,8 @@ CONFIG_NFT_TPROXY=m
 CONFIG_SYN_COOKIES=y
 CONFIG_VETH=y
 CONFIG_XFRM_ESPINTCP=m
+CONFIG_TLS=m
+CONFIG_CRYPTO_ARIA=m
+CONFIG_CRYPTO_CCM=m
+CONFIG_CRYPTO_CHACHA20POLY1305=m
+CONFIG_CRYPTO_SM4_GENERIC=m
diff --git a/tools/testing/selftests/net/mptcp/mptcp_tls.sh b/tools/testing/selftests/net/mptcp/mptcp_tls.sh
new file mode 100755
index 000000000000..04a0cb6ab0f3
--- /dev/null
+++ b/tools/testing/selftests/net/mptcp/mptcp_tls.sh
@@ -0,0 +1,64 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+
+. "$(dirname "${0}")/mptcp_lib.sh"
+
+ret=0
+ns1=""
+pid=""
+
+# This function is used in the cleanup trap
+#shellcheck disable=SC2317,SC2329
+cleanup()
+{
+	if [ -n "$pid" ] && kill -0 "$pid" 2>/dev/null; then
+		kill "$pid" 2>/dev/null
+		wait "$pid" 2>/dev/null
+	fi
+
+	mptcp_lib_ns_exit "$ns1"
+}
+
+init()
+{
+	local max="${1:-4}"
+
+	mptcp_lib_ns_init ns1
+
+	mptcp_lib_pm_nl_set_limits "$ns1" "$max" "$max"
+
+	local i
+	for i in $(seq 1 "$max"); do
+		mptcp_lib_pm_nl_add_endpoint "$ns1" \
+			"127.0.0.1" flags signal port 1000"$i"
+	done
+}
+
+mptcp_lib_check_mptcp
+
+trap cleanup EXIT
+
+# Temporarily set max to '0' to disable multipath testing,
+# as it depends on "mptcp: fix stall because of data_ready" series of fixes.
+# It will be re-enabled together with that series later as a squash-to patch.
+init 0
+
+ip netns exec "$ns1" ./tls -v 12_aes_gcm_mptcp \
+			   -v 13_aes_gcm_mptcp \
+			   -v 12_chacha_mptcp \
+			   -v 13_chacha_mptcp \
+			   -v 13_sm4_gcm_mptcp \
+			   -v 13_sm4_ccm_mptcp \
+			   -v 12_aes_ccm_mptcp \
+			   -v 13_aes_ccm_mptcp \
+			   -v 12_aes_gcm_256_mptcp \
+			   -v 13_aes_gcm_256_mptcp \
+			   -v 13_nopad_mptcp \
+			   -v 12_aria_gcm_mptcp \
+			   -v 12_aria_gcm_256_mptcp &
+pid=$!
+wait $pid
+ret=$?
+
+mptcp_lib_result_print_all_tap
+exit $ret
diff --git a/tools/testing/selftests/net/mptcp/tls.c b/tools/testing/selftests/net/mptcp/tls.c
new file mode 120000
index 000000000000..724b1f047c89
--- /dev/null
+++ b/tools/testing/selftests/net/mptcp/tls.c
@@ -0,0 +1 @@
+../tls.c
\ No newline at end of file
-- 
2.53.0

Re: [RFC mptcp-next v20 00/15] MPTCP KTLS support

[Geliang Tang]

Hi Matt,

On Tue, 2026-05-26 at 17:46 +0800, Geliang Tang wrote:
> From: Geliang Tang <tanggeliang@kylinos.cn>
> 
> v20:

This version did not trigger sashiko. I'm not sure if it's because the
patches couldn't be applied, as some code was merged into the export
branch yesterday. If possible, please help me trigger sashiko again.

Thanks,
-Geliang

> - Patch 1:
>   - Use kfree_rcu() instead of synchronous kfree() in
> tls_proto_cleanup()
>     and tls_proto_put() to prevent use-after-free for RCU readers
>   - Add duplicate check under tls_proto_lock before inserting new
> tls_proto
>     to list to prevent race condition
>   - Fix reference counting: drop the initial reference from
> tls_build_proto()
>     after successful tls_ctx_create() to prevent refcount leak
> - Patch 2:
>   - Replace spin_lock() with spin_lock_bh() in
> tls_register_prot_ops()
>     and tls_unregister_prot_ops() to avoid Lockdep splat (softirq
> context
>     usage)
> - Patch 3:
>   - Add memory barrier ordering in tls_sk_poll() to prevent NULL
> pointer
>     dereference during concurrent setsockopt(TCP_ULP)
>   - Move try_module_get() inside tls_prot_ops_find() while holding
> the lock
>     to prevent ops structure from being freed before module reference
> is
>     acquired
> - Patch 8:
>   - Add missing fallback check __mptcp_check_fallback(msk) before
> calling
>     mptcp_setsockopt_tcp_ulp() to prevent TLS being attached to
> fallback
>     sockets
>   - Re-check fallback status inside mptcp_setsockopt_tcp_ulp() after
> acquiring
>     socket lock to eliminate TOCTOU race
> - Patch 10:
>   - Remove unused _metadata parameter from is_mptcp_enable()
> 
> v19:
> - Remove RCU from tls_proto and tls_prot_ops management
>   Replace RCU with spinlock (tls_proto_lock) for all list operations.
>   Remove rcu_head from struct tls_proto, eliminate call_rcu() and
>   synchronize_rcu() calls. Use list_del + kfree directly in
>   tls_proto_cleanup() and tls_proto_put().
> - Add memory barrier in update_sk_prot()
>   Use smp_store_release() when setting sk->sk_socket->ops to ensure
> it is
>   not visible before icsk_ulp_data. Add explanatory comment to
> prevent
>   NULL pointer dereference in concurrent tls_sk_poll() due to CPU
>   reordering.
> - Add module owner field to tls_prot_ops
>   Add .owner = THIS_MODULE to tls_tcp_ops (and tls_mptcp_ops in MPTCP
>   side). Add try_module_get(ops->owner) in tls_build_proto() before
>   caching ops pointer. Add module_put(proto->ops->owner) in
>   tls_proto_put() when releasing proto.
> - Fix tls_toe_sk_destruct use-after-free
>   Restore sk->sk_prot before calling original destructor to prevent
> UAF
>   when __sk_destruct() accesses sk->sk_prot->destroy.
> - Optimize indirect call performance
>   Delay inq and get_skb_seq indirect calls in
> tls_read_flush_backlog()
>   and tls_rx_msg_size() to avoid unnecessary retpoline overhead on
> fast
>   path.
> - Fix shutdown_reuse test flakiness
>   Replace MPTCP-specific wait with generic bind retry loop. Retry
> bind()
>   on EINVAL up to 1000 times (1 second) to handle asynchronous state
>   transitions for both TCP and MPTCP.
> - Remove redundant tls_ctx->proto->ops NULL checks
>   Rely on memory barrier to guarantee ordering, eliminate unnecessary
>   NULL checks that could introduce other issues.
> - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1779355169.git.tanggeliang@kylinos.cn/
> 
> v18:
>  - add ip_ver field to tls_proto and pass it to tls_proto_find to
>    distinguish IPv4/IPv6 instances   
>  - convert tls_proto_mutex to spinlock, and convert tls_proto_cleanup
>    to use call_rcu and rcu_barrier for softirq context compatibility
>  - add tls_proto_put() for atomic refcount decrement + list_del +
>    call_rcu, and simplify tls_ctx_free and tls_init error paths to
> use it
>  - use GFP_ATOMIC for tls_proto allocation in tls_build_proto
>  - reorder tls_register cleanup labels to unwind correctly on error
>  - implement new mptcp_inq
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1777459066.git.tanggeliang@kylinos.cn/
> 
> v17:
>  - remove owner from tls_prot_ops since both TCP and MPTCP are
>    built-in and cannot be unloaded
>  - add rcu_head to tls_proto for delayed freeing
>  - replace synchronize_rcu with call_rcu in tls_ctx_free
>  - remove module refcounting (try_module_get / module_put) from
>    tls_build_proto and tls_init
>  - simplify tls_proto_cleanup to directly free all protos without
>    refcnt check
>  - change tls_proto refcnt from 2 to 1 (only socket reference, no
>    list reference)
>  - move synchronize_rcu outside spinlock in tls_unregister_prot_ops
>  - use kzalloc_obj instead of kzalloc to fix the checkpatch warning
>  - update commit logs
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1777026753.git.tanggeliang@kylinos.cn/
> 
> v16:
>  - drop rcu_head from struct tls_proto, use refcnt for lifecycle
>    management.
>  - add back TLS_NUM_PROTS to handle IPv4/IPv6 separately.
>  - add .owner field to tls_tcp_ops and tls_mptcp_ops (THIS_MODULE).
>  - add module refcounting (try_module_get / module_put) in
> tls_build_proto
>    and tls_init.
>  - add missing NULL check for tls_ctx->proto->ops in tls_sk_poll.
>  - add RCU read lock protection in tls_register_prot_ops.
>  - add error handling for tls_register_prot_ops calls in tls_register
>    (with rollback on failure).
>  - adjust MPTCP cleanup: move tcp_cleanup_ulp from
> mptcp_destroy_common
>    to mptcp_destroy.
>  - remove increase_rlimit from selftest and fix fd check.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1776924681.git.tanggeliang@kylinos.cn/
> 
> v15:
>  - patch 1: add proto parameter for tls_toe_bypass.
>  - patch 1: add a proto null-check in update_sk_prot.
>  - patch 1: hold mutex_lock in tls_proto_cleanup.
>  - patch 14: raise the limit of file descriptor values to 4096 to
> avoid
>    test failures.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1776469068.git.tanggeliang@kylinos.cn/
> 
> v14:
>  - address review comments from sashiko
>  - patch 1: add rcu for tls_proto, add tls_proto_cleanup.
>  - patch 2: add unregister helper.
>  - patch 3: add tls_prot_ops pointer to tls_proto, instead of
>    tls_context
>  - patch 5: update mptcp_get_skb_seq, using map_seq - offset, then
> the
>    patch "tls: add skb offset check for mptcp" can be dropped.
>  - patch 7: check len < 0.
>  - patch 8: call tcp_cleanup_ulp in mptcp_destroy_common.
>  - patch 9: replace all "tls" as "espintcp" in sock_test_tcpulp.
>  - patch 10: add is_mptcp_enable helper.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1775476921.git.tanggeliang@kylinos.cn/
> 
> v13:
>  - patch 1: Add new patch "add per-protocol cache" to address AI
> review.
>  - patch 2: Hold RCU read lock in tls_prot_ops_find().
>  - patch 3: Set icsk_ulp_data to NULL in error path.
>  - patch 6: Use spin_is_locked() instead of lockdep_is_held() to fix
>    build errors.
>  - patch 9: Drop tcp_sock_set_ulp().
>  - patch 11: Remove the "return" statement in ulp_sock_pair and check
>    the return values of socket().
>  - patch 14: Update wait_for_tcp_close().
>  - patch 16: Add a max argument to init() and set it to '0' to
> disable
>    multipath testing, so that this series does not depend on the
> "mptcp:
>    fix stall because of data_ready" series. Multipath testing will be
>    re‑enabled together with that series later, as a squash‑to patch.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1775227717.git.tanggeliang@kylinos.cn/
> 
> v12:
>  - Thanks for the help from Paolo and Gang Yan, I finally solved the
>    deadlock issue in read_sock. As a result, the patch "mptcp: avoid
>    sleeping in read_sock path under softirq" in v11 has been dropped,
>    and instead a lock_is_held interface has been added to struct
>    tls_prot_ops. When MPTCP implements this interface, it not only
>    checks sock_owned_by_user_nocheck(sk) as TCP does, but also needs
>    to check whether the MPTCP data lock is held.
>  - Update selftests to make them more stable.
>  - Fix shellcheck errors for the selftests.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1775115102.git.tanggeliang@kylinos.cn/
> 
> v11:
>  - Fix memory leak errors reported by CI. In v10, these occurred in
>    the shutdown_reuse test and "usleep(500000)" caused the memory
>    leaks. In v11, a dedicated helper wait_for_tcp_close() has been
>    added to provide an appropriate delay.
>  - Drop the code that used mptcp_data_trylock() in mptcp_move_skbs()
>    to fix a deadlock issue, as that deadlock no longer occurs in v11.
>  - Do not add "mptcp" variable for the "tls_err" tests, adding it for
>    the "tls" tests is sufficient.
>  - No longer increase timeout values for poll/epoll tests, as they
>    are no longer needed.
>  - Add ns1 definition in mptcp_tls.sh to fix "ns1 is referenced but
>    not assigned" error.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1773911536.git.tanggeliang@kylinos.cn/
> 
> v10:
>  - Address comments by ai review:
>    - patch 2: call tls_ctx_free(sk, ctx) and clear icsk_ulp_data
> before
>      goto out.
>    - patch 3: update commit log as "validate each SKB's offset except
>      the first".
>    - patch 5: add sock_owned_by_user() checks.
>    - patch 7: disable device offload for MPTCP sockets.
>    - patch 9: use TCP_ULP_NAME_MAX in mptcp_setsockopt_tcp_ulp(),
> drop
>      SOL_TLS in mptcp_supported_sockopt().
>  - Make .get_skb_off optional instead of mandatory, TCP does not need
>    to define it.
>  - Test "espintcp" ULP instead of "smc" in patch 10. "smc" ULP is
>    removed recently.
>  - With Gang Yan's "mptcp: fix stall because of data_ready" v3, mptcp
>    tls selftests can run without failures. Now add them in this set.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1773737371.git.tanggeliang@kylinos.cn/
> 
> v9:
>  - add a new patch to "add MPTCP SKB offset check in strp queue
> walk",
>    thanks to Gang Yan for the fix.
>  - add a new patch to "avoid deadlocks in read_sock path", replacing
> the
>    "in_softirq()" check used in v8.
>  - update the selftests.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1773365606.git.tanggeliang@kylinos.cn/
> 
> v8:
>  - do not hold tls_prot_ops_lock in tls_init(); otherwise, a deadlock
>    occurs.
>  - change return value of mptcp_stream_is_readable() as 'bool' to fix
> the
>    "expected restricted __poll_t" warning reported by CI.
>  - fixed other CI checkpatch warnings regarding excessively long
> lines.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1768294706.git.tanggeliang@kylinos.cn/
> 
> v7:
>  - Passing an MPTCP socket to tcp_sock_rate_check_app_limited()
> causes a
>    crash. In v7, an MPTCP version of check_app_limited() is
> implemented,
>    which calls tcp_sock_rate_check_app_limited() for each subflow.
>  - Register tls_tcp_ops and tls_mptcp_ops in tls_register() rather
> than in
>    tls_init().
>  - Set ctx->ops in tls_init() instead of in do_tls_setsockopt_conf().
>  - Keep tls_device.c unchanged. MPTCP TLS_HW mode has not been
> implemented
>    yet, so EOPNOTSUPP is returned in this case.
>  - Also add TCP TLS tests in mptcp_join.sh.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1768284047.git.tanggeliang@kylinos.cn/
> 
> v6:
>  - register each ops as Matt suggested.
>  - drop sk_is_msk().
>  - add tcp_sock_get_ulp/tcp_sock_set_ulp helpers.
>  - set another ULP in sock_test_tcpulp as Matt suggested.
>  - add tls tests using multiple subflows in mptcp_join.sh.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1767518836.git.tanggeliang@kylinos.cn/
> 
> v5:
>  - As suggested by Mat and Matt, this set introduces struct
> tls_prot_ops
>    for TLS.
>  - Includes Gang Yan's patches to add MPTCP support to the TLS
> selftests.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1766372799.git.tanggeliang@kylinos.cn/
> 
> v4:
>  - split "tls: add MPTCP protocol support" into smaller, more
>    focused patches.
>  - a new mptcp_inq helper has been implemented instead of directly
>    using mptcp_inq_hint to fix the issue mentioned in [1].
>  - add sk_is_msk helper.
>  - the 'expect' parameter will no longer be added to
> sock_test_tcpulp.
>    Instead, SOCK_TEST_TCPULP items causing the tests failure will be
>    directly removed.
>  - remove the "TCP KTLS" tests, keeping only the MPTCP-related ones.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1765505775.git.tanggeliang@kylinos.cn/
> 
> [1]
> https://patchwork.kernel.org/project/mptcp/patch/ce74452f4c095a1761ef493b767b4bd9f9c14359.1764333805.git.tanggeliang@kylinos.cn/
> 
> v3:
>  - mptcp_read_sock() and mptcp_poll() are not exported, as
> mptcp_sockopt
>    test does not use read_sock/poll interfaces. They will be exported
> when
>    new tests are added in the future.
>  - call mptcp_inq_hint in tls_device_rx_resync_new_rec(),
>    tls_device_core_ctrl_rx_resync() and tls_read_flush_backlog() too.
>  - update selftests.
>  - Link:
> https://patchwork.kernel.org/project/mptcp/cover/cover.1763800601.git.tanggeliang@kylinos.cn/
> 
> v2:
>  - fix disconnect.
>  - update selftests.
> 
> This series adds KTLS support for MPTCP. Since the ULP of msk is not
> being
> used, ULP KTLS can be directly configured onto msk without affecting
> its
> communication.
> 
> Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/480
> 
> Gang Yan (1):
>   mptcp: update mptcp_check_readable
> 
> Geliang Tang (14):
>   tls: add per-protocol cache to support mptcp
>   tls: introduce struct tls_prot_ops
>   tls: add tls_prot_ops pointer to tls_proto
>   mptcp: implement tls_mptcp_ops
>   tls: disable device offload for mptcp sockets
>   mptcp: update ulp getsockopt for tls support
>   mptcp: enable ulp setsockopt for tls support
>   selftests: mptcp: connect: use espintcp for ulp test
>   selftests: tls: add mptcp variant for testing
>   selftests: tls: increase pollin timeouts for mptcp
>   selftests: tls: increase nonblocking data size for mptcp
>   selftests: tls: retry bind on EINVAL in shutdown_reuse
>   selftests: tls: add mptcp test cases
>   selftests: mptcp: cover mptcp tls tests
> 
>  include/linux/tcp.h                           |   1 +
>  include/net/mptcp.h                           |   2 +
>  include/net/tcp.h                             |   1 +
>  include/net/tls.h                             |  32 ++
>  include/net/tls_toe.h                         |   3 +-
>  net/ipv4/tcp.c                                |  45 ++-
>  net/mptcp/protocol.c                          | 131 ++++++++-
>  net/mptcp/protocol.h                          |   1 +
>  net/mptcp/sockopt.c                           |  58 +++-
>  net/tls/tls.h                                 |   3 +-
>  net/tls/tls_device.c                          |   6 +
>  net/tls/tls_main.c                            | 278 +++++++++++++++-
> --
>  net/tls/tls_strp.c                            |  33 ++-
>  net/tls/tls_sw.c                              |   6 +-
>  net/tls/tls_toe.c                             |   7 +-
>  tools/testing/selftests/net/mptcp/.gitignore  |   1 +
>  tools/testing/selftests/net/mptcp/Makefile    |   2 +
>  tools/testing/selftests/net/mptcp/config      |   6 +
>  .../selftests/net/mptcp/mptcp_connect.c       |   4 +-
>  .../testing/selftests/net/mptcp/mptcp_tls.sh  |  64 ++++
>  tools/testing/selftests/net/mptcp/tls.c       |   1 +
>  tools/testing/selftests/net/tls.c             | 155 +++++++++-
>  22 files changed, 748 insertions(+), 92 deletions(-)
>  create mode 100755 tools/testing/selftests/net/mptcp/mptcp_tls.sh
>  create mode 120000 tools/testing/selftests/net/mptcp/tls.c

Re: [RFC mptcp-next v20 00/15] MPTCP KTLS support

[Matthieu Baerts]

Hi Geliang,

On 27/05/2026 16:03, Geliang Tang wrote:
> Hi Matt,
> 
> On Tue, 2026-05-26 at 17:46 +0800, Geliang Tang wrote:
>> From: Geliang Tang <tanggeliang@kylinos.cn>
>>
>> v20:
> 
> This version did not trigger sashiko. I'm not sure if it's because the
> patches couldn't be applied, as some code was merged into the export
> branch yesterday. If possible, please help me trigger sashiko again.
It was:

  https://sashiko.dev/#/patchset/cover.1779788090.git.tanggeliang@kylinos.cn

But not sure why GitHub Actions didn't start (quite a few issues there,
so, probably just unlucky). I manually restarted it by overriding the
tag after a rebase.

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

Re: [RFC mptcp-next v20 05/15] mptcp: implement tls_mptcp_ops

[gang.yan]

May 26, 2026 at 5:46 PM, "Geliang Tang" <geliang@kernel.org mailto:geliang@kernel.org?to=%22Geliang%20Tang%22%20%3Cgeliang%40kernel.org%3E > wrote:


> 
> From: Geliang Tang <tanggeliang@kylinos.cn>
> 
> This patch implements the MPTCP-specific struct tls_prot_ops, named
> 'tls_mptcp_ops'.
> 
> Passing an MPTCP socket to tcp_sock_rate_check_app_limited() can
> trigger a crash. Here, an MPTCP version of check_app_limited() is
> implemented, which calls tcp_sock_rate_check_app_limited() for each
> subflow.
> 
> When MPTCP implements lock_is_held interface, it not only checks
> sock_owned_by_user_nocheck(sk) as TCP does, but also needs to check
> whether the MPTCP data lock is held. This is required because TLS
> may call lock_is_held from softirq context with bh_lock_sock held.
> Checking both conditions ensures TLS always defers to workqueue when
> the MPTCP data lock is held, avoiding deadlock.
> 
> Co-developed-by: Gang Yan <yangang@kylinos.cn>
> Signed-off-by: Gang Yan <yangang@kylinos.cn>
> Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
> ---
>  include/net/mptcp.h | 2 +
>  include/net/tcp.h | 1 +
>  net/ipv4/tcp.c | 9 +++-
>  net/mptcp/protocol.c | 121 +++++++++++++++++++++++++++++++++++++++++--
>  net/mptcp/protocol.h | 1 +
>  net/tls/tls_main.c | 13 +++++
>  6 files changed, 141 insertions(+), 6 deletions(-)
> 
> diff --git a/include/net/mptcp.h b/include/net/mptcp.h
> index 4cf59e83c1c5..02564eceeb7e 100644
> --- a/include/net/mptcp.h
> +++ b/include/net/mptcp.h
> @@ -132,6 +132,8 @@ struct mptcp_pm_ops {
>  void (*release)(struct mptcp_sock *msk);
>  } ____cacheline_aligned_in_smp;
>  
> +extern struct tls_prot_ops tls_mptcp_ops;
> +
>  #ifdef CONFIG_MPTCP
>  void mptcp_init(void);
>  
> diff --git a/include/net/tcp.h b/include/net/tcp.h
> index f063eccbbba3..1c8201f69ef1 100644
> --- a/include/net/tcp.h
> +++ b/include/net/tcp.h
> @@ -849,6 +849,7 @@ static inline int tcp_bound_to_half_wnd(struct tcp_sock *tp, int pktsize)
>  
>  /* tcp.c */
>  void tcp_get_info(struct sock *, struct tcp_info *);
> +void tcp_sock_rate_check_app_limited(struct tcp_sock *tp);
>  void tcp_rate_check_app_limited(struct sock *sk);
>  
>  /* Read 'sendfile()'-style from a TCP socket */
> diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
> index a058f350a759..bdad459e6605 100644
> --- a/net/ipv4/tcp.c
> +++ b/net/ipv4/tcp.c
> @@ -1097,9 +1097,9 @@ int tcp_sendmsg_fastopen(struct sock *sk, struct msghdr *msg, int *copied,
>  }
>  
>  /* If a gap is detected between sends, mark the socket application-limited. */
> -void tcp_rate_check_app_limited(struct sock *sk)
> +void tcp_sock_rate_check_app_limited(struct tcp_sock *tp)
>  {
> - struct tcp_sock *tp = tcp_sk(sk);
> + struct sock *sk = (struct sock *)tp;
>  
>  if (/* We have less than one packet to send. */
>  tp->write_seq - tp->snd_nxt < tp->mss_cache &&
> @@ -1112,6 +1112,11 @@ void tcp_rate_check_app_limited(struct sock *sk)
>  tp->app_limited =
>  (tp->delivered + tcp_packets_in_flight(tp)) ? : 1;
>  }
> +
> +void tcp_rate_check_app_limited(struct sock *sk)
> +{
> + tcp_sock_rate_check_app_limited(tcp_sk(sk));
> +}
>  EXPORT_SYMBOL_GPL(tcp_rate_check_app_limited);
>  
>  int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
> diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
> index a36ef97155a7..505eac23f35d 100644
> --- a/net/mptcp/protocol.c
> +++ b/net/mptcp/protocol.c
> @@ -24,6 +24,7 @@
>  #include <net/mptcp.h>
>  #include <net/hotdata.h>
>  #include <net/xfrm.h>
> +#include <net/tls.h>
>  #include <asm/ioctls.h>
>  #include "protocol.h"
>  #include "mib.h"
> @@ -1909,7 +1910,7 @@ static void mptcp_rps_record_subflows(const struct mptcp_sock *msk)
>  }
>  }
>  
> -static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
> +static int mptcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t len)
>  {
>  struct mptcp_sock *msk = mptcp_sk(sk);
>  struct page_frag *pfrag;
> @@ -1921,8 +1922,6 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
>  msg->msg_flags &= MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL |
>  MSG_FASTOPEN | MSG_EOR;
>  
> - lock_sock(sk);
> -
>  mptcp_rps_record_subflows(msk);
>  
>  if (unlikely(inet_test_bit(DEFER_CONNECT, sk) ||
> @@ -2038,7 +2037,6 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
>  }
>  
>  out:
> - release_sock(sk);
>  return copied;
>  
>  do_error:
> @@ -2049,6 +2047,17 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
>  goto out;
>  }
>  
> +static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
> +{
> + int ret;
> +
> + lock_sock(sk);
> + ret = mptcp_sendmsg_locked(sk, msg, len);
> + release_sock(sk);
> +
> + return ret;
> +}
> +
>  static void mptcp_rcv_space_adjust(struct mptcp_sock *msk, int copied);
>  
>  static void mptcp_eat_recv_skb(struct sock *sk, struct sk_buff *skb)
> @@ -4726,3 +4735,107 @@ int __init mptcp_proto_v6_init(void)
>  return err;
>  }
>  #endif
> +
> +static int mptcp_inq(struct sock *sk)
> +{
> + const struct mptcp_sock *msk = mptcp_sk(sk);
> + const struct sk_buff *skb;
> +
> + if ((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV))
> + return 0;
> +
> + skb = skb_peek(&sk->sk_receive_queue);
> + if (skb) {
> + u64 answ = READ_ONCE(msk->ack_seq) - MPTCP_SKB_CB(skb)->map_seq;
> +
> + if (answ >= INT_MAX)
> + answ = INT_MAX;
> +
> + /* Subtract 1, if FIN was received */
> + if (answ &&
> + (sk->sk_state == TCP_CLOSE ||
> + (sk->sk_shutdown & RCV_SHUTDOWN)))
> + answ--;
> +
> + return (int)answ;
> + }
> +
> + return 0;
> +}
> +
> +static bool mptcp_lock_is_held(struct sock *sk)
> +{
> + return sock_owned_by_user_nocheck(sk) ||
> + mptcp_data_is_locked(sk);
> +}
> +
> +static void mptcp_read_done(struct sock *sk, size_t len)
> +{
> + struct mptcp_sock *msk = mptcp_sk(sk);
> + struct sk_buff *skb;
> + size_t left;
> + u32 offset;
> +
> + msk_owned_by_me(msk);
> +
> + if (sk->sk_state == TCP_LISTEN)
> + return;
> +
> + left = len;
> + while (left && (skb = mptcp_recv_skb(sk, &offset)) != NULL) {
> + int used;
> +
> + used = min_t(size_t, skb->len - offset, left);
> + msk->bytes_consumed += used;
> + MPTCP_SKB_CB(skb)->offset += used;
> + MPTCP_SKB_CB(skb)->map_seq += used;
> + left -= used;
> +
> + if (skb->len > offset + used)
> + break;
> +
> + mptcp_eat_recv_skb(sk, skb);
> + }
> +
> + mptcp_rcv_space_adjust(msk, len - left);
> +
> + /* Clean up data we have read: This will do ACK frames. */
> + if (left != len)
> + mptcp_cleanup_rbuf(msk, len - left);
> +}
> +
> +static u32 mptcp_get_skb_seq(struct sk_buff *skb)
> +{
> + return MPTCP_SKB_CB(skb)->map_seq - MPTCP_SKB_CB(skb)->offset;
> +}
> +
> +static void mptcp_check_app_limited(struct sock *sk)
> +{
> + struct mptcp_sock *msk = mptcp_sk(sk);
> + struct mptcp_subflow_context *subflow;
> +
> + mptcp_for_each_subflow(msk, subflow) {
> + struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
> + bool slow;
> +
> + slow = lock_sock_fast(ssk);
> + tcp_sock_rate_check_app_limited(tcp_sk(ssk));
> + unlock_sock_fast(ssk, slow);
> + }
> +}
> +
> +struct tls_prot_ops tls_mptcp_ops = {
> + .owner = THIS_MODULE,
> + .protocol = IPPROTO_MPTCP,
> + .inq = mptcp_inq,
> + .sendmsg_locked = mptcp_sendmsg_locked,
> + .recv_skb = mptcp_recv_skb,
> + .lock_is_held = mptcp_lock_is_held,
> + .read_sock = mptcp_read_sock,
> + .read_done = mptcp_read_done,
> + .get_skb_seq = mptcp_get_skb_seq,

Hi Geliang,

I think we need a get_skb_off callback, something like:

+static u32 mptcp_get_skb_off(struct sk_buff *skb)
+{
+       return MPTCP_SKB_CB(skb)->offset;
+}
+
 struct tls_prot_ops tls_mptcp_ops = {
        .owner                  = THIS_MODULE,
        .protocol               = IPPROTO_MPTCP,
@@ -5156,6 +5181,7 @@ struct tls_prot_ops tls_mptcp_ops = {
        .read_sock              = mptcp_read_sock,
        .read_done              = mptcp_read_done,
        .get_skb_seq            = mptcp_get_skb_seq,
+       .get_skb_off            = mptcp_get_skb_off,

The reason is that some functions in TLS still call skb_copy_bits directly,
which completely ignores the MPTCP-level offset stored in the SKB. That causes
data corruption / transmission errors.

Here is one concrete example:

The expected TLS record header is 17 03 03. We get 17 03 from the first skb, and the final 03 should come from the next skb at offset 2. But skb_copy_bits reads from offset 0 of that next skb, so we end up with 17 03 17 — and that causes error -22:
'''
  [  180.277464][ T2744] tls_rx_msg_size: bad TLS header: 17 03 17, offset=30 anchor_len=4590 anchor_dlen=4590  map_seq=12790065556595467861 end_seq=12790065556595467863frag_list=00000000683d062a fl_len=32 next=00000000d63e8c38 next_len=510          
  map_seq=12790065556595467863                                                                                                                                                                                                                            
  [  180.279420][ T2744] tls_rx_msg_size: next_skb first 8 bytes: 17 03 03 00 19 00 00 00                                                                                                                                                                 
  # # tls.c:629:multi_chunk_sendfile:Expected recv(self->cfd, buf, test_payload_size, MSG_WAITALL) (3945) == test_payload_size (4097)                                                                                                                     
  [  180.279828][ T2744] YGYG return: copied:3945, err:-22      
'''

In other places the same class of problem also triggers a -74 error.

After looking into this, here is a work-around patch I came up with:
diff --git a/net/tls/tls_strp.c b/net/tls/tls_strp.c
index c24aa4ec2687..bcd05f15bc68 100644
--- a/net/tls/tls_strp.c
+++ b/net/tls/tls_strp.c
@@ -452,7 +452,8 @@ static bool tls_strp_check_queue_ok(struct tls_strparser *strp)
                seq += skb->len;
                len -= skb->len;
                skb = skb->next;
-
+               if (ctx->proto->ops->get_skb_off(skb))
+                       return false;
                if (ctx->proto->ops->get_skb_seq(skb) != seq)
                        return false;
                if (skb_cmp_decrypted(first, skb))
@@ -531,6 +532,18 @@ static int tls_strp_read_sock(struct tls_strparser *strp)
                return tls_strp_read_copy(strp, true);
 
        tls_strp_load_anchor_with_queue(strp, inq);
+
+       /* If the next skb has non-zero MPTCP offset, skb_copy_bits
+        * would read stale data. Fallback to copy mode before
+        * tls_rx_msg_size touches the data.
+        */
+       if (ctx->proto->ops->get_skb_off) {
+               struct sk_buff *first = skb_shinfo(strp->anchor)->frag_list;
+
+               if (first->next && ctx->proto->ops->get_skb_off(first->next))
+                       return tls_strp_read_copy(strp, false);
+       }
+

The first chunk prevents the -74 error by forcing TLS into copy mode
(no direct skb_copy_bits on those SKBs for decryption) when it detects an MPTCP
offset.

The second chunk is a work-around that avoids the bad skb_copy_bits read inside
tls_rx_msg_size by falling back to copy mode when the next SKB in the frag_list
has a non-zero offset.

Longer term I think we should refactor tls_rx_msg_size to properly handle the
MPTCP offset, rather than relying on these fallback paths.

Thanks
Gang


> + .poll = mptcp_poll,
> + .epollin_ready = mptcp_epollin_ready,
> + .check_app_limited = mptcp_check_app_limited,
> +};
> +EXPORT_SYMBOL(tls_mptcp_ops);
> diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
> index 661600f8b573..1c604a1ded6f 100644
> --- a/net/mptcp/protocol.h
> +++ b/net/mptcp/protocol.h
> @@ -380,6 +380,7 @@ struct mptcp_sock {
>  
>  #define mptcp_data_lock(sk) spin_lock_bh(&(sk)->sk_lock.slock)
>  #define mptcp_data_unlock(sk) spin_unlock_bh(&(sk)->sk_lock.slock)
> +#define mptcp_data_is_locked(sk) spin_is_locked(&(sk)->sk_lock.slock)
>  
>  #define mptcp_for_each_subflow(__msk, __subflow) \
>  list_for_each_entry(__subflow, &((__msk)->conn_list), node)
> diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
> index c840a228af0d..fbb60dc1832e 100644
> --- a/net/tls/tls_main.c
> +++ b/net/tls/tls_main.c
> @@ -1425,6 +1425,12 @@ static int __init tls_register(void)
>  if (err)
>  goto err_strp;
>  
> +#ifdef CONFIG_MPTCP
> + err = tls_register_prot_ops(&tls_mptcp_ops);
> + if (err)
> + goto err_tcp;
> +#endif
> +
>  err = tls_device_init();
>  if (err)
>  goto err_ops;
> @@ -1433,6 +1439,10 @@ static int __init tls_register(void)
>  
>  return 0;
>  err_ops:
> +#ifdef CONFIG_MPTCP
> + tls_unregister_prot_ops(&tls_mptcp_ops);
> +err_tcp:
> +#endif
>  tls_unregister_prot_ops(&tls_tcp_ops);
>  err_strp:
>  tls_strp_dev_exit();
> @@ -1444,6 +1454,9 @@ static int __init tls_register(void)
>  static void __exit tls_unregister(void)
>  {
>  tcp_unregister_ulp(&tcp_tls_ulp_ops);
> +#ifdef CONFIG_MPTCP
> + tls_unregister_prot_ops(&tls_mptcp_ops);
> +#endif
>  tls_unregister_prot_ops(&tls_tcp_ops);
>  tls_proto_cleanup();
>  tls_strp_dev_exit();
> -- 
> 2.53.0
>

Re: [RFC mptcp-next v20 00/15] MPTCP KTLS support

[MPTCP CI]

Hi Geliang,

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): Unstable: 1 failed test(s): selftest_mptcp_connect_mmap ⚠️ 
- KVM Validation: debug (only selftest_mptcp_join): Success! ✅
- KVM Validation: btf-normal (only bpftest_all): Success! ✅
- KVM Validation: btf-debug (only bpftest_all): Success! ✅
- Task: https://github.com/multipath-tcp/mptcp_net-next/actions/runs/26494777306

Initiator: Matthieu Baerts (NGI0)
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/22b32594b480
Patchwork: https://patchwork.kernel.org/project/mptcp/list/?series=1100835


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)