From: Geliang Tang tanggeliang@kylinos.cn

Just the same as userspace PM, a new parameter needs_id is added for in-kernel PM mptcp_pm_nl_append_new_local_addr() too.

Add a new helper mptcp_pm_has_addr_attr_id() to check whether an address ID is set from PM or not.

In mptcp_pm_nl_get_local_id(), needs_id is always true, but in mptcp_pm_nl_add_addr_doit(), pass mptcp_pm_has_addr_attr_id() to needs_it.

Fixes: efd5a4c04e18 ("mptcp: add the address ID assignment bitmap") Cc: stable@vger.kernel.org Signed-off-by: Geliang Tang tanggeliang@kylinos.cn Reviewed-by: Mat Martineau martineau@kernel.org Signed-off-by: Matthieu Baerts (NGI0) matttbe@kernel.org --- net/mptcp/pm_netlink.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)

diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c index 287a60381eae..a24c9128dee9 100644 --- a/net/mptcp/pm_netlink.c +++ b/net/mptcp/pm_netlink.c @@ -901,7 +901,8 @@ static void __mptcp_pm_release_addr_entry(struct mptcp_pm_addr_entry *entry) }

static int mptcp_pm_nl_append_new_local_addr(struct pm_nl_pernet *pernet, - struct mptcp_pm_addr_entry *entry) + struct mptcp_pm_addr_entry *entry, + bool needs_id) { struct mptcp_pm_addr_entry *cur, *del_entry = NULL; unsigned int addr_max; @@ -949,7 +950,7 @@ static int mptcp_pm_nl_append_new_local_addr(struct pm_nl_pernet *pernet, } }

- if (!entry->addr.id) { + if (!entry->addr.id && needs_id) { find_next: entry->addr.id = find_next_zero_bit(pernet->id_bitmap, MPTCP_PM_MAX_ADDR_ID + 1, @@ -960,7 +961,7 @@ static int mptcp_pm_nl_append_new_local_addr(struct pm_nl_pernet *pernet, } }

- if (!entry->addr.id) + if (!entry->addr.id && needs_id) goto out;

__set_bit(entry->addr.id, pernet->id_bitmap); @@ -1092,7 +1093,7 @@ int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc entry->ifindex = 0; entry->flags = MPTCP_PM_ADDR_FLAG_IMPLICIT; entry->lsk = NULL; - ret = mptcp_pm_nl_append_new_local_addr(pernet, entry); + ret = mptcp_pm_nl_append_new_local_addr(pernet, entry, true); if (ret < 0) kfree(entry);

@@ -1285,6 +1286,18 @@ static int mptcp_nl_add_subflow_or_signal_addr(struct net *net) return 0; }

+static bool mptcp_pm_has_addr_attr_id(const struct nlattr *attr, + struct genl_info *info) +{ + struct nlattr *tb[MPTCP_PM_ADDR_ATTR_MAX + 1]; + + if (!nla_parse_nested_deprecated(tb, MPTCP_PM_ADDR_ATTR_MAX, attr, + mptcp_pm_address_nl_policy, info->extack) && + tb[MPTCP_PM_ADDR_ATTR_ID]) + return true; + return false; +} + int mptcp_pm_nl_add_addr_doit(struct sk_buff *skb, struct genl_info *info) { struct nlattr *attr = info->attrs[MPTCP_PM_ENDPOINT_ADDR]; @@ -1326,7 +1339,8 @@ int mptcp_pm_nl_add_addr_doit(struct sk_buff *skb, struct genl_info *info) goto out_free; } } - ret = mptcp_pm_nl_append_new_local_addr(pernet, entry); + ret = mptcp_pm_nl_append_new_local_addr(pernet, entry, + !mptcp_pm_has_addr_attr_id(attr, info)); if (ret < 0) { GENL_SET_ERR_MSG_FMT(info, "too many addresses or duplicate one: %d", ret); goto out_free;

From: Paolo Abeni pabeni@redhat.com

The local address id is accessed lockless by the NL PM, add all the required ONCE annotation. There is a caveat: the local id can be initialized late in the subflow life-cycle, and its validity is controlled by the local_id_valid flag.

Remove such flag and encode the validity in the local_id field itself with negative value before initialization. That allows accessing the field consistently with a single read operation.

Fixes: 0ee4261a3681 ("mptcp: implement mptcp_pm_remove_subflow") Cc: stable@vger.kernel.org Signed-off-by: Paolo Abeni pabeni@redhat.com Reviewed-by: Mat Martineau martineau@kernel.org Signed-off-by: Matthieu Baerts (NGI0) matttbe@kernel.org --- net/mptcp/diag.c | 2 +- net/mptcp/pm_netlink.c | 6 +++--- net/mptcp/pm_userspace.c | 2 +- net/mptcp/protocol.c | 2 +- net/mptcp/protocol.h | 15 ++++++++++++--- net/mptcp/subflow.c | 9 +++++---- 6 files changed, 23 insertions(+), 13 deletions(-)

diff --git a/net/mptcp/diag.c b/net/mptcp/diag.c index e57c5f47f035..6ff6f14674aa 100644 --- a/net/mptcp/diag.c +++ b/net/mptcp/diag.c @@ -65,7 +65,7 @@ static int subflow_get_info(struct sock *sk, struct sk_buff *skb) sf->map_data_len) || nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_FLAGS, flags) || nla_put_u8(skb, MPTCP_SUBFLOW_ATTR_ID_REM, sf->remote_id) || - nla_put_u8(skb, MPTCP_SUBFLOW_ATTR_ID_LOC, sf->local_id)) { + nla_put_u8(skb, MPTCP_SUBFLOW_ATTR_ID_LOC, subflow_get_local_id(sf))) { err = -EMSGSIZE; goto nla_failure; } diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c index a24c9128dee9..912e25077437 100644 --- a/net/mptcp/pm_netlink.c +++ b/net/mptcp/pm_netlink.c @@ -800,7 +800,7 @@ static void mptcp_pm_nl_rm_addr_or_subflow(struct mptcp_sock *msk, mptcp_for_each_subflow_safe(msk, subflow, tmp) { struct sock *ssk = mptcp_subflow_tcp_sock(subflow); int how = RCV_SHUTDOWN | SEND_SHUTDOWN; - u8 id = subflow->local_id; + u8 id = subflow_get_local_id(subflow);

if (rm_type == MPTCP_MIB_RMADDR && subflow->remote_id != rm_id) continue; @@ -809,7 +809,7 @@ static void mptcp_pm_nl_rm_addr_or_subflow(struct mptcp_sock *msk,

pr_debug(" -> %s rm_list_ids[%d]=%u local_id=%u remote_id=%u mpc_id=%u", rm_type == MPTCP_MIB_RMADDR ? "address" : "subflow", - i, rm_id, subflow->local_id, subflow->remote_id, + i, rm_id, id, subflow->remote_id, msk->mpc_endpoint_id); spin_unlock_bh(&msk->pm.lock); mptcp_subflow_shutdown(sk, ssk, how); @@ -1994,7 +1994,7 @@ static int mptcp_event_add_subflow(struct sk_buff *skb, const struct sock *ssk) if (WARN_ON_ONCE(!sf)) return -EINVAL;

- if (nla_put_u8(skb, MPTCP_ATTR_LOC_ID, sf->local_id)) + if (nla_put_u8(skb, MPTCP_ATTR_LOC_ID, subflow_get_local_id(sf))) return -EMSGSIZE;

if (nla_put_u8(skb, MPTCP_ATTR_REM_ID, sf->remote_id)) diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c index e582b3b2d174..d396a5973429 100644 --- a/net/mptcp/pm_userspace.c +++ b/net/mptcp/pm_userspace.c @@ -234,7 +234,7 @@ static int mptcp_userspace_pm_remove_id_zero_address(struct mptcp_sock *msk,

lock_sock(sk); mptcp_for_each_subflow(msk, subflow) { - if (subflow->local_id == 0) { + if (READ_ONCE(subflow->local_id) == 0) { has_id_0 = true; break; } diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 8ef2927ebca2..948606a537da 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -85,7 +85,7 @@ static int __mptcp_socket_create(struct mptcp_sock *msk) subflow->subflow_id = msk->subflow_id++;

/* This is the first subflow, always with id 0 */ - subflow->local_id_valid = 1; + WRITE_ONCE(subflow->local_id, 0); mptcp_sock_graft(msk->first, sk->sk_socket); iput(SOCK_INODE(ssock));

diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h index ed50f2015dc3..631a7f445f34 100644 --- a/net/mptcp/protocol.h +++ b/net/mptcp/protocol.h @@ -491,10 +491,9 @@ struct mptcp_subflow_context { remote_key_valid : 1, /* received the peer key from */ disposable : 1, /* ctx can be free at ulp release time */ stale : 1, /* unable to snd/rcv data, do not use for xmit */ - local_id_valid : 1, /* local_id is correctly initialized */ valid_csum_seen : 1, /* at least one csum validated */ is_mptfo : 1, /* subflow is doing TFO */ - __unused : 9; + __unused : 10; bool data_avail; bool scheduled; u32 remote_nonce; @@ -505,7 +504,7 @@ struct mptcp_subflow_context { u8 hmac[MPTCPOPT_HMAC_LEN]; /* MPJ subflow only */ u64 iasn; /* initial ack sequence number, MPC subflows only */ }; - u8 local_id; + s16 local_id; /* if negative not initialized yet */ u8 remote_id; u8 reset_seen:1; u8 reset_transient:1; @@ -556,6 +555,7 @@ mptcp_subflow_ctx_reset(struct mptcp_subflow_context *subflow) { memset(&subflow->reset, 0, sizeof(subflow->reset)); subflow->request_mptcp = 1; + WRITE_ONCE(subflow->local_id, -1); }

static inline u64 @@ -1022,6 +1022,15 @@ int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc); int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc); int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc);

+static inline u8 subflow_get_local_id(const struct mptcp_subflow_context *subflow) +{ + int local_id = READ_ONCE(subflow->local_id); + + if (local_id < 0) + return 0; + return local_id; +} + void __init mptcp_pm_nl_init(void); void mptcp_pm_nl_work(struct mptcp_sock *msk); void mptcp_pm_nl_rm_subflow_received(struct mptcp_sock *msk, diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c index c34ecadee120..015184bbf06c 100644 --- a/net/mptcp/subflow.c +++ b/net/mptcp/subflow.c @@ -577,8 +577,8 @@ static void subflow_finish_connect(struct sock *sk, const struct sk_buff *skb)

static void subflow_set_local_id(struct mptcp_subflow_context *subflow, int local_id) { - subflow->local_id = local_id; - subflow->local_id_valid = 1; + WARN_ON_ONCE(local_id < 0 || local_id > 255); + WRITE_ONCE(subflow->local_id, local_id); }

static int subflow_chk_local_id(struct sock *sk) @@ -587,7 +587,7 @@ static int subflow_chk_local_id(struct sock *sk) struct mptcp_sock *msk = mptcp_sk(subflow->conn); int err;

- if (likely(subflow->local_id_valid)) + if (likely(subflow->local_id >= 0)) return 0;

err = mptcp_pm_get_local_id(msk, (struct sock_common *)sk); @@ -1731,6 +1731,7 @@ static struct mptcp_subflow_context *subflow_create_ctx(struct sock *sk, pr_debug("subflow=%p", ctx);

ctx->tcp_sock = sk; + WRITE_ONCE(ctx->local_id, -1);

return ctx; } @@ -1966,7 +1967,7 @@ static void subflow_ulp_clone(const struct request_sock *req, new_ctx->idsn = subflow_req->idsn;

/* this is the first subflow, id is always 0 */ - new_ctx->local_id_valid = 1; + subflow_set_local_id(new_ctx, 0); } else if (subflow_req->mp_join) { new_ctx->ssn_offset = subflow_req->ssn_offset; new_ctx->mp_join = 1;

From: Paolo Abeni pabeni@redhat.com

Fullmesh endpoints could end-up unexpectedly generating duplicate subflows - same local and remote addresses - when multiple incoming ADD_ADDR are processed before the PM creates the subflow for the local endpoints.

Address the issue explicitly checking for duplicates at subflow creation time.

To avoid a quadratic computational complexity, track the unavailable remote address ids in a temporary bitmap and initialize such bitmap with the remote ids of all the existing subflows matching the local address currently processed.

The above allows additionally replacing the existing code checking for duplicate entry in the current set with a simple bit test operation.

Fixes: 2843ff6f36db ("mptcp: remote addresses fullmesh") Cc: stable@vger.kernel.org Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/435 Signed-off-by: Paolo Abeni pabeni@redhat.com Reviewed-by: Mat Martineau martineau@kernel.org Signed-off-by: Matthieu Baerts (NGI0) matttbe@kernel.org --- net/mptcp/pm_netlink.c | 33 ++++++++++++++++++--------------- 1 file changed, 18 insertions(+), 15 deletions(-)

diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c index ed6983af1ab2..58d17d9604e7 100644 --- a/net/mptcp/pm_netlink.c +++ b/net/mptcp/pm_netlink.c @@ -396,19 +396,6 @@ void mptcp_pm_free_anno_list(struct mptcp_sock *msk) } }

-static bool lookup_address_in_vec(const struct mptcp_addr_info *addrs, unsigned int nr, - const struct mptcp_addr_info *addr) -{ - int i; - - for (i = 0; i < nr; i++) { - if (addrs[i].id == addr->id) - return true; - } - - return false; -} - /* Fill all the remote addresses into the array addrs[], * and return the array size. */ @@ -440,6 +427,16 @@ static unsigned int fill_remote_addresses_vec(struct mptcp_sock *msk, msk->pm.subflows++; addrs[i++] = remote; } else { + DECLARE_BITMAP(unavail_id, MPTCP_PM_MAX_ADDR_ID + 1); + + /* Forbid creation of new subflows matching existing + * ones, possibly already created by incoming ADD_ADDR + */ + bitmap_zero(unavail_id, MPTCP_PM_MAX_ADDR_ID + 1); + mptcp_for_each_subflow(msk, subflow) + if (READ_ONCE(subflow->local_id) == local->id) + __set_bit(subflow->remote_id, unavail_id); + mptcp_for_each_subflow(msk, subflow) { ssk = mptcp_subflow_tcp_sock(subflow); remote_address((struct sock_common *)ssk, &addrs[i]); @@ -447,11 +444,17 @@ static unsigned int fill_remote_addresses_vec(struct mptcp_sock *msk, if (deny_id0 && !addrs[i].id) continue;

+ if (test_bit(addrs[i].id, unavail_id)) + continue; + if (!mptcp_pm_addr_families_match(sk, local, &addrs[i])) continue;

- if (!lookup_address_in_vec(addrs, i, &addrs[i]) && - msk->pm.subflows < subflows_max) { + if (msk->pm.subflows < subflows_max) { + /* forbid creating multiple address towards + * this id + */ + __set_bit(addrs[i].id, unavail_id); msk->pm.subflows++; i++; }

Since the 'Fixes' commit mentioned below, and if the kernel being tested doesn't support the 'fullmesh' flag, this error will be printed:

netlink error -22 (Invalid argument) ./pm_nl_ctl: bailing out due to netlink error[s]

But that can be normal if the kernel doesn't support the feature, no need to print this worrying error message while everything else looks OK. So we can mute stderr. Failures will still be detected if any.

Fixes: 1dc88d241f92 ("selftests: mptcp: pm_nl_ctl: always look for errors") Cc: stable@vger.kernel.org Reviewed-by: Geliang Tang geliang@kernel.org Signed-off-by: Matthieu Baerts (NGI0) matttbe@kernel.org --- tools/testing/selftests/net/mptcp/pm_netlink.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/net/mptcp/pm_netlink.sh b/tools/testing/selftests/net/mptcp/pm_netlink.sh index 79e83a2c95de..71899a3ffa7a 100755 --- a/tools/testing/selftests/net/mptcp/pm_netlink.sh +++ b/tools/testing/selftests/net/mptcp/pm_netlink.sh @@ -183,7 +183,7 @@ check "ip netns exec $ns1 ./pm_nl_ctl dump" "id 1 flags \ subflow 10.0.1.1" " (nobackup)"

# fullmesh support has been added later -ip netns exec $ns1 ./pm_nl_ctl set id 1 flags fullmesh +ip netns exec $ns1 ./pm_nl_ctl set id 1 flags fullmesh 2>/dev/null if ip netns exec $ns1 ./pm_nl_ctl dump | grep -q "fullmesh" || mptcp_lib_expect_all_features; then check "ip netns exec $ns1 ./pm_nl_ctl dump" "id 1 flags \

The selftest was correctly recording all the results, but the 'reverse direction' part was missing in the name when needed.

It is important to have a unique (sub)test name in TAP, because some CI environments drop tests with duplicated name.

Fixes: 675d99338e7a ("selftests: mptcp: simult flows: format subtests results in TAP") Cc: stable@vger.kernel.org Reviewed-by: Geliang Tang geliang@kernel.org Signed-off-by: Matthieu Baerts (NGI0) matttbe@kernel.org --- tools/testing/selftests/net/mptcp/simult_flows.sh | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/net/mptcp/simult_flows.sh b/tools/testing/selftests/net/mptcp/simult_flows.sh index 0cc964e6f2c1..8f9ddb3ad4fe 100755 --- a/tools/testing/selftests/net/mptcp/simult_flows.sh +++ b/tools/testing/selftests/net/mptcp/simult_flows.sh @@ -250,7 +250,8 @@ run_test() [ $bail -eq 0 ] || exit $ret fi

- printf "%-60s" "$msg - reverse direction" + msg+=" - reverse direction" + printf "%-60s" "${msg}" do_transfer $large $small $time lret=$? mptcp_lib_result_code "${lret}" "${msg}"

It is important to have a unique (sub)test name in TAP, because some CI environments drop tests with duplicated name.

Some 'in use' subtests from the diag selftest had the same names, e.g.:

chk 0 msk in use after flush

Now the previous value is taken, to have different names, e.g.:

chk 2->0 msk in use after flush

While at it, avoid repeating the full message, declare it once in the helper.

Fixes: ce9902573652 ("selftests: mptcp: diag: format subtests results in TAP") Cc: stable@vger.kernel.org Reviewed-by: Geliang Tang geliang@kernel.org Signed-off-by: Matthieu Baerts (NGI0) matttbe@kernel.org --- tools/testing/selftests/net/mptcp/diag.sh | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/tools/testing/selftests/net/mptcp/diag.sh b/tools/testing/selftests/net/mptcp/diag.sh index e0615c6ffb8d..266656a16229 100755 --- a/tools/testing/selftests/net/mptcp/diag.sh +++ b/tools/testing/selftests/net/mptcp/diag.sh @@ -166,9 +166,13 @@ chk_msk_listen() chk_msk_inuse() { local expected=$1 - local msg="$2" + local msg="....chk ${2:-${expected}} msk in use" local listen_nr

+ if [ "${expected}" -eq 0 ]; then + msg+=" after flush" + fi + listen_nr=$(ss -N "${ns}" -Ml | grep -c LISTEN) expected=$((expected + listen_nr))

@@ -179,7 +183,7 @@ chk_msk_inuse() sleep 0.1 done

- __chk_nr get_msk_inuse $expected "$msg" 0 + __chk_nr get_msk_inuse $expected "${msg}" 0 }

# $1: cestab nr @@ -227,11 +231,11 @@ wait_connected $ns 10000 chk_msk_nr 2 "after MPC handshake " chk_msk_remote_key_nr 2 "....chk remote_key" chk_msk_fallback_nr 0 "....chk no fallback" -chk_msk_inuse 2 "....chk 2 msk in use" +chk_msk_inuse 2 chk_msk_cestab 2 flush_pids

-chk_msk_inuse 0 "....chk 0 msk in use after flush" +chk_msk_inuse 0 "2->0" chk_msk_cestab 0

echo "a" | \ @@ -247,11 +251,11 @@ echo "b" | \ 127.0.0.1 >/dev/null & wait_connected $ns 10001 chk_msk_fallback_nr 1 "check fallback" -chk_msk_inuse 1 "....chk 1 msk in use" +chk_msk_inuse 1 chk_msk_cestab 1 flush_pids

-chk_msk_inuse 0 "....chk 0 msk in use after flush" +chk_msk_inuse 0 "1->0" chk_msk_cestab 0

NR_CLIENTS=100 @@ -273,11 +277,11 @@ for I in `seq 1 $NR_CLIENTS`; do done

wait_msk_nr $((NR_CLIENTS*2)) "many msk socket present" -chk_msk_inuse $((NR_CLIENTS*2)) "....chk many msk in use" +chk_msk_inuse $((NR_CLIENTS*2)) "many" chk_msk_cestab $((NR_CLIENTS*2)) flush_pids

-chk_msk_inuse 0 "....chk 0 msk in use after flush" +chk_msk_inuse 0 "many->0" chk_msk_cestab 0

mptcp_lib_result_print_all_tap