mptcp: add more sockopts
Subsystems (2)
Geliang Tang
reviewerv2 cover mptcp: add more sockopts
From: Geliang Tang <[email protected]>
v2:
- Use and extend mptcp_setsockopt_all_sf to remove duplicate code in v1
Depends on:
- refactor MPTCP socket option handling
Based-on: <[email protected]>
v1:
https://patchwork.kernel.org/project/mptcp/cover/[email protected]/
Add sockopt support for IPV6_TCLASS, IP_TTL and IPV6_UNICAST_HOPS.
Geliang Tang (3):
mptcp: add IPV6_TCLASS socket option support
mptcp: add IP_TTL socket option support
mptcp: add IPV6_UNICAST_HOPS socket option support
net/mptcp/sockopt.c | 54 +++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 50 insertions(+), 4 deletions(-)
Signature · 1 line
2.48.1
Geliang Tang
reviewerPATCH v2 1/3 mptcp: add IPV6_TCLASS socket option support
From: Geliang Tang <[email protected]>
Implement support for IPV6_TCLASS socket option in MPTCP:
1. setsockopt(IPV6_TCLASS):
- Sets traffic class on main MPTCP socket
- Propagates value to all IPv6 subflows
2. getsockopt(IPV6_TCLASS):
- Returns current value from inet6_sk(sk)->tclass
3. Syncs traffic class to new subflows
Mirrors existing IPv4 TOS handling and ensures consistent QoS across subflows. Fixes reported issues with traffic class propagation in MPTCP.
--- net/mptcp/sockopt.c | 27 +++++++++++++++++++++++---- 1 file changed, 23 insertions(+), 4 deletions(-)
net/mptcp/sockopt.c +23 -4
@@ -458,6 +458,9 @@ static int mptcp_setsockopt_v6(struct mptcp_sock *msk, int optname,case IPV6_FREEBIND :return mptcp_setsockopt_first_sf_only (msk,SOL_IPV6 , optname, optval, optlen); +case IPV6_TCLASS : +return mptcp_setsockopt_all_sf (msk,SOL_IPV6 , optname, + optval, optlen); }return -EOPNOTSUPP ;
@@ -765,7 +768,7 @@ static int mptcp_setsockopt_sol_ip_first_sf_only(struct mptcp_sock *msk, int optreturn 0 ; } -static int mptcp_setsockopt_sol_ip_all_sf (struct mptcp_sock *msk, +static int mptcp_setsockopt_sol_ip_all_sf (struct mptcp_sock *msk,int level,int optname,sockptr_t optval,unsigned int optlen) {
@@ -773,10 +776,16 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk,struct sock *sk = (struct sock *)msk;int err, val; -if (optname !=IP_TOS ) +if (optname !=IP_TOS && + optname !=IPV6_TCLASS )return -EOPNOTSUPP ; - err =ip_setsockopt (sk,SOL_IP , optname, optval, optlen); +if (level ==SOL_IP ) + err =ip_setsockopt (sk, level, optname, optval, optlen); +#if IS_ENABLED (CONFIG_MPTCP_IPV6 ) + else + err =ipv6_setsockopt (sk, level, optname, optval, optlen); +#endif if (err !=0 )return err;
@@ -785,6 +794,8 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk,sockopt_seq_inc (msk);if (optname ==IP_TOS ) val =READ_ONCE (inet_sk (sk)->tos); +else if (optname ==IPV6_TCLASS ) + val =READ_ONCE (inet6_sk (sk)->tclass);mptcp_for_each_subflow (msk, subflow) {struct sock *ssk =mptcp_subflow_tcp_sock (subflow);bool slow;
@@ -792,6 +803,8 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk, slow =lock_sock_fast (ssk);if (optname ==IP_TOS )__ip_sock_set_tos (ssk, val); +else if (optname ==IPV6_TCLASS ) +WRITE_ONCE (inet6_sk (ssk)->tclass, val);unlock_sock_fast (ssk, slow); }release_sock (sk);
@@ -885,7 +898,8 @@ static int mptcp_setsockopt_all_sf(struct mptcp_sock *msk, int level,return mptcp_setsockopt_sol_tcp_all_sf (msk, optname, optval, optlen);case SOL_IP : -return mptcp_setsockopt_sol_ip_all_sf (msk, optname, +case SOL_IPV6 : +return mptcp_setsockopt_sol_ip_all_sf (msk, level, optname, optval, optlen); }
@@ -1569,6 +1583,9 @@ static int mptcp_getsockopt_v6(struct mptcp_sock *msk, int optname,case IPV6_FREEBIND :return mptcp_put_int_option (msk, optval, optlen,inet_test_bit (FREEBIND , sk)); +case IPV6_TCLASS : +return mptcp_put_int_option (msk, optval, optlen, +inet6_sk (sk)->tclass); }return -EOPNOTSUPP ;
@@ -1677,6 +1694,8 @@ static void sync_socket_options(struct mptcp_sock *msk, struct sock *ssk)inet_assign_bit (FREEBIND , ssk,inet_test_bit (FREEBIND , sk));inet_assign_bit (BIND_ADDRESS_NO_PORT , ssk,inet_test_bit (BIND_ADDRESS_NO_PORT , sk));WRITE_ONCE (inet_sk (ssk)->local_port_range,READ_ONCE (inet_sk (sk)->local_port_range)); +if (ssk->sk_family ==AF_INET6 ) +WRITE_ONCE (inet6_sk (ssk)->tclass,READ_ONCE (inet6_sk (sk)->tclass)); }void mptcp_sockopt_sync_locked (struct mptcp_sock *msk,struct sock *ssk)
Signature · 1 line
2.48.1
Geliang Tang
reviewerPATCH v2 2/3 mptcp: add IP_TTL socket option support
In reply to Geliang TangFrom: Geliang Tang <[email protected]>
Implement setsockopt/getsockopt for IP_TTL in MPTCP:
1. setsockopt(IP_TTL):
- Applies TTL to MPTCP socket
- Propagates TTL to all subflows
2. getsockopt(IP_TTL):
- Returns current TTL from inet_sk(sk)->uc_ttl
- Falls back to sysctl_ip_default_ttl when unset
3. Syncs TTL to new subflows
4. Maintains option sequence counters
Enables applications to control IPv4 TTL settings consistently across all subflows. Resolves outstanding requests for TTL support in MPTCP.
--- net/mptcp/sockopt.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-)
net/mptcp/sockopt.c +13 -1
@@ -776,7 +776,7 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk, int level,struct sock *sk = (struct sock *)msk;int err, val; -if (optname !=IP_TOS && +if (optname !=IP_TOS && optname !=IP_TTL && optname !=IPV6_TCLASS )return -EOPNOTSUPP ;
@@ -794,6 +794,8 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk, int level,sockopt_seq_inc (msk);if (optname ==IP_TOS ) val =READ_ONCE (inet_sk (sk)->tos); +else if (optname ==IP_TTL ) + val =READ_ONCE (inet_sk (sk)->uc_ttl);else if (optname ==IPV6_TCLASS ) val =READ_ONCE (inet6_sk (sk)->tclass);mptcp_for_each_subflow (msk, subflow) {
@@ -803,6 +805,8 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk, int level, slow =lock_sock_fast (ssk);if (optname ==IP_TOS )__ip_sock_set_tos (ssk, val); +else if (optname ==IP_TTL ) +WRITE_ONCE (inet_sk (ssk)->uc_ttl, val);else if (optname ==IPV6_TCLASS )WRITE_ONCE (inet6_sk (ssk)->tclass, val);unlock_sock_fast (ssk, slow);
@@ -822,6 +826,7 @@ static int mptcp_setsockopt_v4(struct mptcp_sock *msk, int optname,case IP_LOCAL_PORT_RANGE :return mptcp_setsockopt_first_sf_only (msk,SOL_IP , optname, optval, optlen);case IP_TOS : +case IP_TTL :return mptcp_setsockopt_all_sf (msk,SOL_IP , optname, optval, optlen); }
@@ -1547,6 +1552,7 @@ static int mptcp_getsockopt_v4(struct mptcp_sock *msk, int optname, char __user *optval, int __user *optlen) { struct sock *sk = (void *)msk; +int val;switch (optname) {case IP_TOS :
@@ -1563,6 +1569,11 @@ static int mptcp_getsockopt_v4(struct mptcp_sock *msk, int optname,case IP_LOCAL_PORT_RANGE :return mptcp_put_int_option (msk, optval, optlen,READ_ONCE (inet_sk (sk)->local_port_range)); +case IP_TTL : + val =READ_ONCE (inet_sk (sk)->uc_ttl); +if (val <0 ) + val =READ_ONCE (sock_net (sk)->ipv4.sysctl_ip_default_ttl); +return mptcp_put_int_option (msk, optval, optlen, val); }return -EOPNOTSUPP ;
@@ -1694,6 +1705,7 @@ static void sync_socket_options(struct mptcp_sock *msk, struct sock *ssk)inet_assign_bit (FREEBIND , ssk,inet_test_bit (FREEBIND , sk));inet_assign_bit (BIND_ADDRESS_NO_PORT , ssk,inet_test_bit (BIND_ADDRESS_NO_PORT , sk));WRITE_ONCE (inet_sk (ssk)->local_port_range,READ_ONCE (inet_sk (sk)->local_port_range)); +WRITE_ONCE (inet_sk (ssk)->uc_ttl,READ_ONCE (inet_sk (sk)->uc_ttl));if (ssk->sk_family ==AF_INET6 )WRITE_ONCE (inet6_sk (ssk)->tclass,READ_ONCE (inet6_sk (sk)->tclass)); }
Signature · 1 line
2.48.1
Geliang Tang
reviewerPATCH v2 3/3 mptcp: add IPV6_UNICAST_HOPS socket option support
In reply to Geliang TangFrom: Geliang Tang <[email protected]>
Implement setsockopt/getsockopt for IPV6_UNICAST_HOPS in MPTCP:
1. setsockopt(IPV6_UNICAST_HOPS):
- Sets hop limit on MPTCP socket
- Propagates value to all IPv6 subflows
2. getsockopt(IPV6_UNICAST_HOPS):
- Returns current hop limit (from inet6_sk or system default)
3. Syncs hop limit to new subflows
4. Uses conditional compilation for IPv6 support
Provides standard IPv6 hop limit control for MPTCP connections with consistent behavior across subflows. Completes IPv6 socket option parity.
--- net/mptcp/sockopt.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-)
net/mptcp/sockopt.c +17 -2
@@ -459,6 +459,7 @@ static int mptcp_setsockopt_v6(struct mptcp_sock *msk, int optname,return mptcp_setsockopt_first_sf_only (msk,SOL_IPV6 , optname, optval, optlen);case IPV6_TCLASS : +case IPV6_UNICAST_HOPS :return mptcp_setsockopt_all_sf (msk,SOL_IPV6 , optname, optval, optlen); }
@@ -777,7 +778,7 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk, int level,int err, val;if (optname !=IP_TOS && optname !=IP_TTL && - optname !=IPV6_TCLASS ) + optname !=IPV6_TCLASS && optname !=IPV6_UNICAST_HOPS )return -EOPNOTSUPP ;if (level ==SOL_IP )
@@ -798,6 +799,8 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk, int level, val =READ_ONCE (inet_sk (sk)->uc_ttl); else if (optname == IPV6_TCLASS) val =READ_ONCE (inet6_sk (sk)->tclass); + else if (optname == IPV6_UNICAST_HOPS) + val =READ_ONCE (inet6_sk (sk)->hop_limit);mptcp_for_each_subflow (msk, subflow) {struct sock *ssk =mptcp_subflow_tcp_sock (subflow);bool slow;
@@ -809,6 +812,8 @@ static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk, int level,WRITE_ONCE (inet_sk (ssk)->uc_ttl, val); else if (optname == IPV6_TCLASS) WRITE_ONCE(inet6_sk(ssk)->tclass, val); + else if (optname == IPV6_UNICAST_HOPS) + WRITE_ONCE(inet6_sk(ssk)->hop_limit, val);unlock_sock_fast (ssk ,slow ); }release_sock (sk);
@@ -1583,6 +1588,7 @@ static int mptcp_getsockopt_v6(struct mptcp_sock *msk, int optname, char __user *optval, int __user *optlen) { struct sock *sk = (void *)msk; +int val;switch (optname) {case IPV6_V6ONLY :
@@ -1597,6 +1603,13 @@ static int mptcp_getsockopt_v6(struct mptcp_sock *msk, int optname,case IPV6_TCLASS :return mptcp_put_int_option (msk, optval, optlen,inet6_sk (sk)->tclass); +case IPV6_UNICAST_HOPS : + val =READ_ONCE (inet6_sk (sk)->hop_limit); +#if IS_ENABLED (CONFIG_MPTCP_IPV6 ) +if (val <0 ) + val =READ_ONCE (sock_net (sk)->ipv6.devconf_all->hop_limit); +#endif +return mptcp_put_int_option (msk, optval, optlen, val); }return -EOPNOTSUPP ;
@@ -1706,8 +1719,10 @@ static void sync_socket_options(struct mptcp_sock *msk, struct sock *ssk)inet_assign_bit (BIND_ADDRESS_NO_PORT , ssk,inet_test_bit (BIND_ADDRESS_NO_PORT , sk));WRITE_ONCE (inet_sk (ssk)->local_port_range,READ_ONCE (inet_sk (sk)->local_port_range));WRITE_ONCE (inet_sk (ssk)->uc_ttl,READ_ONCE (inet_sk (sk)->uc_ttl)); -if (ssk->sk_family ==AF_INET6 ) +if (ssk->sk_family ==AF_INET6 ) {WRITE_ONCE (inet6_sk (ssk)->tclass,READ_ONCE (inet6_sk (sk)->tclass)); +WRITE_ONCE (inet6_sk (ssk)->hop_limit,READ_ONCE (inet6_sk (sk)->hop_limit)); + } }void mptcp_sockopt_sync_locked (struct mptcp_sock *msk,struct sock *ssk)
Signature · 1 line
2.48.1
Hi Geliang,
Thank you for your modifications, that's great!
Our CI did some validations and here is its report:
- KVM Validation: normal: Unstable: 1 failed test(s): selftest_mptcp_connect_checksum 🔴
- KVM Validation: debug: Unstable: 1 failed test(s): packetdrill_mp_capable 🔴
- 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/16903540842
Initiator: Patchew Applier
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/e5177b226ee9
Patchwork: https://patchwork.kernel.org/project/mptcp/list/?series=990418
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-normalFor 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)