All of the subflows of a msk will be orphaned in mptcp_close(), which
means the subflows are in DEAD state. After then, DATA_FIN will be sent,
and the other side will response with a DATA_ACK for this DATA_FIN.
However, if the other side still has pending data, the data that received
on these subflows will not be passed to the msk, as they are DEAD and
subflow_data_ready() will not be called in tcp_data_ready(). Therefore,
these data can't be acked, and they will be retransmitted again and again,
until timeout.
Fix this by setting ssk->sk_socket and ssk->sk_wq to 'NULL', instead of
orphaning the subflows in __mptcp_close(), as Paolo suggested.
Fixes: e16163b6e2 ("mptcp: refactor shutdown and close")
Reviewed-by: Biao Jiang <benbjiang@tencent.com>
Reviewed-by: Mengen Sun <mengensun@tencent.com>
Signed-off-by: Menglong Dong <imagedong@tencent.com>
Reviewed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Use '(struct sock *)msk' to get 'sk' from 'msk' in a more direct way
instead of using '&msk->sk.icsk_inet.sk'.
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The function mptcp_subflow_process_delegated() uses the input ssk first,
while __mptcp_check_push() invokes the packet scheduler first.
So this patch adds a new parameter named 'first' for the function
__mptcp_subflow_push_pending() to deal with these two cases separately.
With this change, the code that invokes the packet scheduler in the
function __mptcp_check_push() can be removed, and replaced by invoking
__mptcp_subflow_push_pending() directly.
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Use msk instead of mptcp_sk(sk) in the functions where the variable
"msk = mptcp_sk(sk)" has been defined.
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
drivers/net/can/usb/kvaser_usb/kvaser_usb_leaf.c
2871edb32f ("can: kvaser_usb: Fix possible completions during init_completion")
abb8670938 ("can: kvaser_usb_leaf: Ignore stale bus-off after start")
8d21f5927a ("can: kvaser_usb_leaf: Fix improved state not being reported")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Our CI reported lockdep splat in the fastopen code:
======================================================
WARNING: possible circular locking dependency detected
6.0.0.mptcp_f5e8bfe9878d+ #1558 Not tainted
------------------------------------------------------
packetdrill/1071 is trying to acquire lock:
ffff8881bd198140 (sk_lock-AF_INET){+.+.}-{0:0}, at: inet_wait_for_connect+0x19c/0x310
but task is already holding lock:
ffff8881b8346540 (k-sk_lock-AF_INET){+.+.}-{0:0}, at: mptcp_sendmsg+0xfdf/0x1740
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 (k-sk_lock-AF_INET){+.+.}-{0:0}:
__lock_acquire+0xb6d/0x1860
lock_acquire+0x1d8/0x620
lock_sock_nested+0x37/0xd0
inet_stream_connect+0x3f/0xa0
mptcp_connect+0x411/0x800
__inet_stream_connect+0x3ab/0x800
mptcp_stream_connect+0xac/0x110
__sys_connect+0x101/0x130
__x64_sys_connect+0x6e/0xb0
do_syscall_64+0x59/0x90
entry_SYSCALL_64_after_hwframe+0x63/0xcd
-> #0 (sk_lock-AF_INET){+.+.}-{0:0}:
check_prev_add+0x15e/0x2110
validate_chain+0xace/0xdf0
__lock_acquire+0xb6d/0x1860
lock_acquire+0x1d8/0x620
lock_sock_nested+0x37/0xd0
inet_wait_for_connect+0x19c/0x310
__inet_stream_connect+0x26c/0x800
tcp_sendmsg_fastopen+0x341/0x650
mptcp_sendmsg+0x109d/0x1740
sock_sendmsg+0xe1/0x120
__sys_sendto+0x1c7/0x2a0
__x64_sys_sendto+0xdc/0x1b0
do_syscall_64+0x59/0x90
entry_SYSCALL_64_after_hwframe+0x63/0xcd
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(k-sk_lock-AF_INET);
lock(sk_lock-AF_INET);
lock(k-sk_lock-AF_INET);
lock(sk_lock-AF_INET);
*** DEADLOCK ***
1 lock held by packetdrill/1071:
#0: ffff8881b8346540 (k-sk_lock-AF_INET){+.+.}-{0:0}, at: mptcp_sendmsg+0xfdf/0x1740
======================================================
The problem is caused by the blocking inet_wait_for_connect() releasing
and re-acquiring the msk socket lock while the subflow socket lock is
still held and the MPTCP socket requires that the msk socket lock must
be acquired before the subflow socket lock.
Address the issue always invoking tcp_sendmsg_fastopen() in an
unblocking manner, and later eventually complete the blocking
__inet_stream_connect() as needed.
Fixes: d98a82a6af ("mptcp: handle defer connect in mptcp_sendmsg")
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The current MPTCP connect implementation duplicates a bit of inet
code and does not use nor provide a struct proto->connect callback,
which in turn will not fit the upcoming fastopen implementation.
Refactor such implementation to use the common helper, moving the
MPTCP-specific bits into mptcp_connect(). Additionally, avoid an
indirect call to the subflow connect callback.
Note that the fastopen call-path invokes mptcp_connect() while already
holding the subflow socket lock. Explicitly keep track of such path
via a new MPTCP-level flag and handle the locking accordingly.
Additionally, track the connect flags in a new msk field to allow
propagating them to the subflow inet_stream_connect call.
Fixes: d98a82a6af ("mptcp: handle defer connect in mptcp_sendmsg")
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The mptcp_pm_nl_get_local_id() code assumes that the msk local address
is available at that point. For passive sockets, we initialize such
address at accept() time.
Depending on the running configuration and the user-space timing, a
passive MPJ subflow can join the msk socket before accept() completes.
In such case, the PM assigns a wrong local id to the MPJ subflow
and later PM netlink operations will end-up touching the wrong/unexpected
subflow.
All the above causes sporadic self-tests failures, especially when
the host is heavy loaded.
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/308
Fixes: 01cacb00b3 ("mptcp: add netlink-based PM")
Fixes: d045b9eb95 ("mptcp: introduce implicit endpoints")
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We will soon introduce custom setsockopt for UDP sockets, too.
Instead of doing even more complex arbitrary checks inside
sock_use_custom_sol_socket(), add a new socket flag and set it
for the relevant socket types (currently only MPTCP).
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After commit d38afeec26 ("tcp/udp: Call inet6_destroy_sock()
in IPv6 sk->sk_destruct()."), we call inet6_destroy_sock() in
sk->sk_destruct() by setting inet6_sock_destruct() to it to make
sure we do not leak inet6-specific resources.
Now we can remove unnecessary inet6_destroy_sock() calls in
sk->sk_prot->destroy().
DCCP and SCTP have their own sk->sk_destruct() function, so we
change them separately in the following patches.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The MPTCP data path is quite complex and hard to understend even
without some foggy comments referring to modified code and/or
completely misleading from the beginning.
Update a few of them to more accurately describing the current
status.
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Daire reported a user-space application hang-up when the
peer is forcibly closed before the data transfer completion.
The relevant application expects the peer to either
do an application-level clean shutdown or a transport-level
connection reset.
We can accommodate a such user by extending the fastclose
usage: at fd close time, if the msk socket has some unread
data, and at FIN_WAIT timeout.
Note that at MPTCP close time we must ensure that the TCP
subflows will reset: set the linger socket option to a suitable
value.
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When an mptcp socket is closed due to an incoming FASTCLOSE
option, so specific sk_err is set and later syscall will
fail usually with EPIPE.
Align the current fastclose error handling with TCP reset,
properly setting the socket error according to the current
msk state and propagating such error.
Additionally sendmsg() is currently not handling properly
the sk_err, always returning EPIPE.
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The mptcp socket and its subflow sockets in accept queue can't be
released after the process exit.
While the release of a mptcp socket in listening state, the
corresponding tcp socket will be released too. Meanwhile, the tcp
socket in the unaccept queue will be released too. However, only init
subflow is in the unaccept queue, and the joined subflow is not in the
unaccept queue, which makes the joined subflow won't be released, and
therefore the corresponding unaccepted mptcp socket will not be released
to.
This can be reproduced easily with following steps:
1. create 2 namespace and veth:
$ ip netns add mptcp-client
$ ip netns add mptcp-server
$ sysctl -w net.ipv4.conf.all.rp_filter=0
$ ip netns exec mptcp-client sysctl -w net.mptcp.enabled=1
$ ip netns exec mptcp-server sysctl -w net.mptcp.enabled=1
$ ip link add red-client netns mptcp-client type veth peer red-server \
netns mptcp-server
$ ip -n mptcp-server address add 10.0.0.1/24 dev red-server
$ ip -n mptcp-server address add 192.168.0.1/24 dev red-server
$ ip -n mptcp-client address add 10.0.0.2/24 dev red-client
$ ip -n mptcp-client address add 192.168.0.2/24 dev red-client
$ ip -n mptcp-server link set red-server up
$ ip -n mptcp-client link set red-client up
2. configure the endpoint and limit for client and server:
$ ip -n mptcp-server mptcp endpoint flush
$ ip -n mptcp-server mptcp limits set subflow 2 add_addr_accepted 2
$ ip -n mptcp-client mptcp endpoint flush
$ ip -n mptcp-client mptcp limits set subflow 2 add_addr_accepted 2
$ ip -n mptcp-client mptcp endpoint add 192.168.0.2 dev red-client id \
1 subflow
3. listen and accept on a port, such as 9999. The nc command we used
here is modified, which makes it use mptcp protocol by default.
$ ip netns exec mptcp-server nc -l -k -p 9999
4. open another *two* terminal and use each of them to connect to the
server with the following command:
$ ip netns exec mptcp-client nc 10.0.0.1 9999
Input something after connect to trigger the connection of the second
subflow. So that there are two established mptcp connections, with the
second one still unaccepted.
5. exit all the nc command, and check the tcp socket in server namespace.
And you will find that there is one tcp socket in CLOSE_WAIT state
and can't release forever.
Fix this by closing all of the unaccepted mptcp socket in
mptcp_subflow_queue_clean() with __mptcp_close().
Now, we can ensure that all unaccepted mptcp sockets will be cleaned by
__mptcp_close() before they are released, so mptcp_sock_destruct(), which
is used to clean the unaccepted mptcp socket, is not needed anymore.
The selftests for mptcp is ran for this commit, and no new failures.
Fixes: f296234c98 ("mptcp: Add handling of incoming MP_JOIN requests")
Fixes: 6aeed90450 ("mptcp: fix race on unaccepted mptcp sockets")
Cc: stable@vger.kernel.org
Reviewed-by: Jiang Biao <benbjiang@tencent.com>
Reviewed-by: Mengen Sun <mengensun@tencent.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Factor out __mptcp_close() from mptcp_close(). The caller of
__mptcp_close() should hold the socket lock, and cancel mptcp work when
__mptcp_close() returns true.
This function will be used in the next commit.
Fixes: f296234c98 ("mptcp: Add handling of incoming MP_JOIN requests")
Fixes: 6aeed90450 ("mptcp: fix race on unaccepted mptcp sockets")
Cc: stable@vger.kernel.org
Reviewed-by: Jiang Biao <benbjiang@tencent.com>
Reviewed-by: Mengen Sun <mengensun@tencent.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If fastopen is used, poll must allow a first write that will trigger
the SYN+data
Similar to what is done in tcp_poll().
Acked-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Benjamin Hesmans <benjamin.hesmans@tessares.net>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When TCP_FASTOPEN_CONNECT has been set on the socket before a connect,
the defer flag is set and must be handled when sendmsg is called.
This is similar to what is done in tcp_sendmsg_locked().
Acked-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Co-developed-by: Benjamin Hesmans <benjamin.hesmans@tessares.net>
Signed-off-by: Benjamin Hesmans <benjamin.hesmans@tessares.net>
Signed-off-by: Dmytro Shytyi <dmytro@shytyi.net>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch adds a new bool variable 'do_check_data_fin' to replace the
original int variable 'copied' in __mptcp_push_pending(), check it to
determine whether to call __mptcp_check_send_data_fin().
Suggested-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Similar to mptcp_for_each_subflow(): this is clearer now that the _safe
version is used in multiple places.
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
While reading sysctl_max_skb_frags, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its readers.
Fixes: 5f74f82ea3 ("net:Add sysctl_max_skb_frags")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If the mptcp socket creation fails due to a CGROUP_INET_SOCK_CREATE
eBPF program, the MPTCP protocol ends-up leaking all the subflows:
the related cleanup happens in __mptcp_destroy_sock() that is not
invoked in such code path.
Address the issue moving the subflow sockets cleanup in the
mptcp_destroy_common() helper, which is invoked in every msk cleanup
path.
Additionally get rid of the intermediate list_splice_init step, which
is an unneeded relic from the past.
The issue is present since before the reported root cause commit, but
any attempt to backport the fix before that hash will require a complete
rewrite.
Fixes: e16163b6e2 ("mptcp: refactor shutdown and close")
Reported-by: Nguyen Dinh Phi <phind.uet@gmail.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Co-developed-by: Nguyen Dinh Phi <phind.uet@gmail.com>
Signed-off-by: Nguyen Dinh Phi <phind.uet@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading these sysctl variables, they can be changed concurrently.
Thus, we need to add READ_ONCE() to their readers.
- .sysctl_rmem
- .sysctl_rwmem
- .sysctl_rmem_offset
- .sysctl_wmem_offset
- sysctl_tcp_rmem[1, 2]
- sysctl_tcp_wmem[1, 2]
- sysctl_decnet_rmem[1]
- sysctl_decnet_wmem[1]
- sysctl_tipc_rmem[1]
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading sysctl_tcp_moderate_rcvbuf, it can be changed
concurrently. Thus, we need to add READ_ONCE() to its readers.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The in-kernel PM has a bit of duplicate code related to ack
generation. Create a new helper factoring out the PM-specific
needs and use it in a couple of places.
As a bonus, mptcp_subflow_send_ack() is not used anymore
outside its own compilation unit and can become static.
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
At disconnect time the MPTCP protocol traverse the subflows
list closing each of them. In some circumstances - MPJ subflow,
passive MPTCP socket, the latter operation can remove the
subflow from the list, invalidating the current iterator.
Address the issue using the safe list traversing helper
variant.
Reported-by: van fantasy <g1042620637@gmail.com>
Fixes: b29fcfb54c ("mptcp: full disconnect implementation")
Tested-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When setting up a subflow's flags for sending MP_PRIO MPTCP options, the
subflow socket lock was not held while reading and modifying several
struct members that are also read and modified in mptcp_write_options().
Acquire the subflow socket lock earlier and send the MP_PRIO ACK with
that lock already acquired. Add a new variant of the
mptcp_subflow_send_ack() helper to use with the subflow lock held.
Fixes: 067065422f ("mptcp: add the outgoing MP_PRIO support")
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Similar to commit 7c80b038d2 ("net: fix sk_wmem_schedule() and
sk_rmem_schedule() errors"), let the MPTCP receive path schedule
exactly the required amount of memory.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After commit 4890b686f4 ("net: keep sk->sk_forward_alloc as small as
possible"), the MPTCP protocol is the last SK_RECLAIM_CHUNK and
SK_RECLAIM_THRESHOLD users.
Update the MPTCP reclaim schema to match the core/TCP one and drop the
mentioned macros. This additionally clean the MPTCP code a bit.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The memory accounting is broken in such exceptional code
path, and after commit 4890b686f4 ("net: keep sk->sk_forward_alloc
as small as possible") we can't find much help there.
Drop the broken code.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/microchip/sparx5/sparx5_switchdev.c
9c5de246c1 ("net: sparx5: mdb add/del handle non-sparx5 devices")
fbb89d02e3 ("net: sparx5: Allow mdb entries to both CPU and ports")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When the listener socket owning the relevant request is closed,
it frees the unaccepted subflows and that causes later deletion
of the paired MPTCP sockets.
The mptcp socket's worker can run in the time interval between such delete
operations. When that happens, any access to msk->first will cause an UaF
access, as the subflow cleanup did not cleared such field in the mptcp
socket.
Address the issue explicitly traversing the listener socket accept
queue at close time and performing the needed cleanup on the pending
msk.
Note that the locking is a bit tricky, as we need to acquire the msk
socket lock, while still owning the subflow socket one.
Fixes: 86e39e0448 ("mptcp: keep track of local endpoint still available for each msk")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If the MPTCP socket shutdown happens before a fallback
to TCP, and all the pending data have been already spooled,
we never close the TCP connection.
Address the issue explicitly checking for critical condition
at fallback time.
Fixes: 1e39e5a32a ("mptcp: infinite mapping sending")
Fixes: 0348c690ed ("mptcp: add the fallback check")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
mptcp_mp_fail_no_response shouldn't be invoked on each worker run, it
should be invoked only when MP_FAIL response timeout occurs.
This patch refactors the MP_FAIL response logic.
It leverages the fact that only the MPC/first subflow can gracefully
fail to avoid unneeded subflows traversal: the failing subflow can
be only msk->first.
A new 'fail_tout' field is added to the subflow context to record the
MP_FAIL response timeout and use such field to reliably share the
timeout timer between the MP_FAIL event and the MPTCP socket close
timeout.
Finally, a new ack is generated to send out MP_FAIL notification as soon
as we hit the relevant condition, instead of waiting a possibly unbound
time for the next data packet.
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/281
Fixes: d9fb797046 ("mptcp: Do not traverse the subflow connection list without lock")
Co-developed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently, tcp_memory_allocated can hit tcp_mem[] limits quite fast.
Each TCP socket can forward allocate up to 2 MB of memory, even after
flow became less active.
10,000 sockets can have reserved 20 GB of memory,
and we have no shrinker in place to reclaim that.
Instead of trying to reclaim the extra allocations in some places,
just keep sk->sk_forward_alloc values as small as possible.
This should not impact performance too much now we have per-cpu
reserves: Changes to tcp_memory_allocated should not be too frequent.
For sockets not using SO_RESERVE_MEM:
- idle sockets (no packets in tx/rx queues) have zero forward alloc.
- non idle sockets have a forward alloc smaller than one page.
Note:
- Removal of SK_RECLAIM_CHUNK and SK_RECLAIM_THRESHOLD
is left to MPTCP maintainers as a follow up.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Each protocol having a ->memory_allocated pointer gets a corresponding
per-cpu reserve, that following patches will use.
Instead of having reserved bytes per socket,
we want to have per-cpu reserves.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Due to memcg interface, SK_MEM_QUANTUM is effectively PAGE_SIZE.
This might change in the future, but it seems better to avoid the
confusion.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The MPTCP socket's conn_list (list of subflows) requires the socket lock
to access. The MP_FAIL timeout code added such an access, where it would
check the list of subflows both in timer context and (later) in workqueue
context where the socket lock is held.
Rather than check the list twice, remove the check in the timeout
handler and only depend on the check in the workqueue. Also remove the
MPTCP_FAIL_NO_RESPONSE flag, since mptcp_mp_fail_no_response() has
insignificant overhead and can be checked on each worker run.
Fixes: 49fa1919d6 ("mptcp: reset subflow when MP_FAIL doesn't respond")
Reported-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This reverts commit 4293248c67.
Additional locks are not needed, all the touched sections
are already under mptcp socket lock protection.
Fixes: 4293248c67 ("mptcp: add data lock for sk timers")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
As per RFC, the offered MPTCP-level window should never shrink.
While we currently track the right edge, we don't enforce the
above constraint on the wire.
Additionally, concurrent xmit on different subflows can end-up in
erroneous right edge update.
Address the above explicitly updating the announced window and
protecting the update with an additional atomic operation (sic)
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Bump a counter for counter when snd_wnd is shared among subflow,
for observability's sake.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
As per RFC, mptcp subflows use a "shared" snd_wnd: the effective
window is the maximum among the current values received on all
subflows. Without such feature a data transfer using multiple
subflows could block.
Window sharing is currently implemented in the RX side:
__tcp_select_window uses the mptcp-level receive buffer to compute
the announced window.
That is not enough: the TCP stack will stick to the window size
received on the given subflow; we need to propagate the msk window
value on each subflow at xmit time.
Change the packet scheduler to ignore the subflow level window
and use instead the msk level one
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Switch net callers to the new API not requiring
the NAPI_POLL_WEIGHT argument.
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Acked-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Acked-by: Alexandra Winter <wintera@linux.ibm.com>
Link: https://lore.kernel.org/r/20220504163725.550782-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski