Future extensions with psock will override their own sk->sk_write_space
callback. This patch ensures that the overridden sk_write_space can be
invoked by MPTCP.
INDIRECT_CALL is used to keep the default path optimised.
Note that sk->sk_write_space was never called directly with MPTCP
sockets, so changing it to sk_stream_write_space in the init, and using
it from mptcp_write_space() is not supposed to change the current
behaviour.
This patch is shared early to ease discussions around future RFC and
avoid confusions with this "fix" that is needed for different future
extensions.
Suggested-by: Paolo Abeni <pabeni@redhat.com>
Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260206-net-next-mptcp-write_space-override-v2-1-e0b12be818c6@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There are two:
min_t(int, xxx, mptcp_wnd_end(msk) - msk->snd_nxt);
Both mptcp_wnd_end(msk) and msk->snd_nxt are u64, their difference
(aka the window size) might be limited to 32 bits - but that isn't
knowable from this code.
So checks being added to min_t() detect the potential discard of
significant bits.
Provided the 'avail_size' and return of mptcp_check_allowed_size()
are changed to an unsigned type (size_t matches the type the caller
uses) both min_t() can be changed to min().
Signed-off-by: David Laight <david.laight.linux@gmail.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
[ wrapped too long lines when declaring mptcp_check_allowed_size() ]
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260203-net-next-mptcp-misc-feat-6-20-v1-6-31ec8bfc56d1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
MPTCP uses several calls of the mptcp_rcv_space_init() helper to
initialize the receive space, with a catch-up call in
mptcp_rcv_space_adjust().
Drop all the other strictly not needed invocations and move constant
fields initialization at socket init/reset time.
This removes a bit of complexity from mptcp DRS code. No functional
changes intended.
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260203-net-next-mptcp-misc-feat-6-20-v1-3-31ec8bfc56d1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
MPTCP initialize the receive buffer stamp in mptcp_rcv_space_init(),
using the provided subflow stamp. Such helper is invoked in several
places; for passive sockets, space init happened at clone time.
In such scenario, MPTCP ends-up accesses the subflow stamp before
its initialization, leading to quite randomic timing for the first
receive buffer auto-tune event, as the timestamp for newly created
subflow is not refreshed there.
Fix the issue moving the stamp initialization out of the mentioned helper,
at the data transfer start, and always using a fresh timestamp.
Fixes: 013e3179db ("mptcp: fix rcv space initialization")
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260203-net-next-mptcp-misc-feat-6-20-v1-2-31ec8bfc56d1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
MPTCP-level OoOs are physiological when multiple subflows are active
concurrently and will not cause retransmissions nor are caused by
drops.
Accounting for them in mptcp_rcvbuf_grow() causes the rcvbuf slowly
drifting towards tcp_rmem[2].
Remove such accounting. Note that subflows will still account for TCP-level
OoO when the MPTCP-level rcvbuf is propagated.
This also closes a subtle and very unlikely race condition with rcvspace
init; active sockets with user-space holding the msk-level socket lock,
could complete such initialization in the receive callback, after that the
first OoO data reaches the rcvbuf and potentially triggering a divide by
zero Oops.
Fixes: e118cdc34d ("mptcp: rcvbuf auto-tuning improvement")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260203-net-next-mptcp-misc-feat-6-20-v1-1-31ec8bfc56d1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch implements .splice_read interface of mptcp struct proto_ops
as mptcp_splice_read() with reference to tcp_splice_read().
Corresponding to __tcp_splice_read(), __mptcp_splice_read() is defined,
invoking mptcp_read_sock() instead of tcp_read_sock().
mptcp_splice_read() is almost the same as tcp_splice_read(), except for
sock_rps_record_flow().
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260130-net-next-mptcp-splice-v2-4-31332ba70d7f@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Current in-kernel TCP sockets -- i.e. from nvme_tcp_try_recv() -- need
to call .read_sock interface of struct proto_ops, but it's not
implemented in MPTCP.
This patch implements it with reference to __tcp_read_sock() and
__mptcp_recvmsg_mskq().
Corresponding to tcp_recv_skb(), a new helper for MPTCP named
mptcp_recv_skb() is added to peek a skb from sk->sk_receive_queue.
Compared with __mptcp_recvmsg_mskq(), mptcp_read_sock() uses
sk->sk_rcvbuf as the max read length. The LISTEN status is checked
before the while loop, and mptcp_recv_skb() and mptcp_cleanup_rbuf()
are invoked after the loop. In the loop, all flags checks for
__mptcp_recvmsg_mskq() are removed.
Reviewed-by: Hannes Reinecke <hare@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>
Link: https://patch.msgid.link/20260130-net-next-mptcp-splice-v2-2-31332ba70d7f@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Some subflow socket errors need to be reported to the MPTCP socket: the
initial subflow connect (MP_CAPABLE), and the ones from the fallback
sockets. The others are not propagated.
The issue is that sock_error() was used to retrieve the error, which was
also resetting the sk_err field. Because of that, when notifying the
userspace about subflow close events later on from the MPTCP worker, the
ssk->sk_err field was always 0.
Now, the error (sk_err) is only reset when propagating it to the msk.
Fixes: 15cc104533 ("mptcp: deliver ssk errors to msk")
Cc: stable@vger.kernel.org
Reviewed-by: Geliang Tang <geliang@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260127-net-mptcp-dup-nl-events-v1-3-7f71e1bc4feb@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In case of subflow disconnect(), which can also happen with the first
subflow in case of errors like timeout or reset, mptcp_subflow_ctx_reset
will reset most fields from the mptcp_subflow_context structure,
including close_event_done. Then, when another subflow is closed, yet
another SUB_CLOSED event for the disconnected initial subflow is sent.
Because of the previous reset, there are no source address and
destination port.
A solution is then to also check the subflow's local id: it shouldn't be
negative anyway.
Another solution would be not to reset subflow->close_event_done at
disconnect time, but when reused. But then, probably the whole reset
could be done when being reused. Let's not change this logic, similar
to TCP with tcp_disconnect().
Fixes: d82809b6c5 ("mptcp: avoid duplicated SUB_CLOSED events")
Cc: stable@vger.kernel.org
Reported-by: Marco Angaroni <marco.angaroni@italtel.com>
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/603
Reviewed-by: Geliang Tang <geliang@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260127-net-mptcp-dup-nl-events-v1-1-7f71e1bc4feb@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The MPTCP protocol usually schedule the retransmission timer only
when there is some chances for such retransmissions to happen.
With a notable exception: __mptcp_push_pending() currently schedule
such timer unconditionally, potentially leading to unnecessary rtx
timer expiration.
The issue is present since the blamed commit below but become easily
reproducible after commit 27b0e701d3 ("mptcp: drop bogus optimization
in __mptcp_check_push()")
Fixes: 33d41c9cd7 ("mptcp: more accurate timeout")
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251205-net-mptcp-misc-fixes-6-19-rc1-v1-3-9e4781a6c1b8@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When __mptcp_retrans() kicks-in, it schedules one or more subflows for
retransmission, but such subflows could be actually left alone if there
is no more data to retransmit and/or in case of concurrent fallback.
Scheduled subflows could be processed much later in time, i.e. when new
data will be transmitted, leading to bad subflow selection.
Explicitly clear all scheduled subflows before leaving the
retransmission function.
Fixes: ee2708aeda ("mptcp: use get_retrans wrapper")
Cc: stable@vger.kernel.org
Reported-by: Filip Pokryvka <fpokryvk@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251125-net-mptcp-clear-sched-rtx-v1-1-1cea4ad2165f@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
sk->sk_timer has been used for TCP keepalives.
Keepalive timers are not in fast path, we want to use sk->sk_timer
storage for retransmit timers, for better cache locality.
Create icsk->icsk_keepalive_timer and change keepalive
code to no longer use sk->sk_timer.
Added space is reclaimed in the following patch.
This includes changes to MPTCP, which was also using sk_timer.
Alias icsk->mptcp_tout_timer and icsk->icsk_keepalive_timer
for inet_sk_diag_fill() sake.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20251124175013.1473655-4-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When the msk socket is owned or the msk receive buffer is full,
move the incoming skbs in a msk level backlog list. This avoid
traversing the joined subflows and acquiring the subflow level
socket lock at reception time, improving the RX performances.
When processing the backlog, use the fwd alloc memory borrowed from
the incoming subflow. skbs exceeding the msk receive space are
not dropped; instead they are kept into the backlog until the receive
buffer is freed. Dropping packets already acked at the TCP level is
explicitly discouraged by the RFC and would corrupt the data stream
for fallback sockets.
Special care is needed to avoid adding skbs to the backlog of a closed
msk and to avoid leaving dangling references into the backlog
at subflow closing time.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251121-net-next-mptcp-memcg-backlog-imp-v1-14-1f34b6c1e0b1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We are soon using it for incoming data processing.
MPTCP can't leverage the sk_backlog, as the latter is processed
before the release callback, and such callback for MPTCP releases
and re-acquire the socket spinlock, breaking the sk_backlog processing
assumption.
Add a skb backlog list inside the mptcp sock struct, and implement
basic helper to transfer packet to and purge such list.
Packets in the backlog are memory accounted and still use the incoming
subflow receive memory, to allow back-pressure. The backlog size is
implicitly bounded to the sum of subflows rcvbuf.
When a subflow is closed, references from the backlog to such sock
are removed.
No packet is currently added to the backlog, so no functional changes
intended here.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251121-net-next-mptcp-memcg-backlog-imp-v1-13-1f34b6c1e0b1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In the MPTCP receive path, we release the subflow allocated fwd
memory just to allocate it again shortly after for the msk.
That could increases the failures chances, especially when we will
add backlog processing, with other actions could consume the just
released memory before the msk socket has a chance to do the
rcv allocation.
Replace the skb_orphan() call with an open-coded variant that
explicitly borrows, the fwd memory from the subflow socket instead
of releasing it.
The borrowed memory does not have PAGE_SIZE granularity; rounding to
the page size will make the fwd allocated memory higher than what is
strictly required and could make the incoming subflow fwd mem
consistently negative. Instead, keep track of the accumulated frag and
borrow the full page at subflow close time.
This allow removing the last drop in the TCP to MPTCP transition and
the associated, now unused, MIB.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251121-net-next-mptcp-memcg-backlog-imp-v1-12-1f34b6c1e0b1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently, as soon as the PM closes a subflow, the msk stops accepting
data from it, even if the TCP socket could be still formally open in the
incoming direction, with the notable exception of the first subflow.
The root cause of such behavior is that code currently piggy back two
separate semantic on the subflow->disposable bit: the subflow context
must be released and that the subflow must stop accepting incoming
data.
The first subflow is never disposed, so it also never stop accepting
incoming data. Use a separate bit to mark the latter status and set such
bit in __mptcp_close_ssk() for all subflows.
Beyond making per subflow behaviour more consistent this will also
simplify the next patch.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251121-net-next-mptcp-memcg-backlog-imp-v1-11-1f34b6c1e0b1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The MPTCP protocol is not currently emitting the NL event when the first
subflow is closed before msk accept() time.
By replacing the in use close helper is such scenario, implicitly introduce
the missing notification. Note that in such scenario we want to be sure
that mptcp_close_ssk() will not trigger any PM work, move the msk state
change update earlier, so that the previous patch will offer such
guarantee.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Geliang Tang <geliang@kernel.org>
Tested-by: Geliang Tang <geliang@kernel.org>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251121-net-next-mptcp-memcg-backlog-imp-v1-8-1f34b6c1e0b1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
MPTCP currently access ack_seq outside the msk socket log scope to
generate the dummy mapping for fallback socket. Soon we are going
to introduce backlog usage and even for fallback socket the ack_seq
value will be significantly off outside of the msk socket lock scope.
Avoid relying on ack_seq for dummy mapping generation, using instead
the subflow sequence number. Note that in case of disconnect() and
(re)connect() we must ensure that any previous state is re-set.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Geliang Tang <geliang@kernel.org>
Tested-by: Geliang Tang <geliang@kernel.org>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251121-net-next-mptcp-memcg-backlog-imp-v1-6-1f34b6c1e0b1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Later patches need to ensure that all MPJ subflows are grafted to the
msk socket before accept() completion.
Currently the grafting happens under the msk socket lock: potentially
at msk release_cb time which make satisfying the above condition a bit
tricky.
Move the MPJ subflow grafting earlier, under the msk data lock, so that
we can use such lock as a synchronization point.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251121-net-next-mptcp-memcg-backlog-imp-v1-3-1f34b6c1e0b1@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Cross-merge networking fixes after downstream PR (net-6.18-rc7).
No conflicts, adjacent changes:
tools/testing/selftests/net/af_unix/Makefile
e1bb28bf13 ("selftest: af_unix: Add test for SO_PEEK_OFF.")
45a1cd8346 ("selftests: af_unix: Add tests for ECONNRESET and EOF semantics")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull networking fixes from Jakub Kicinski:
"Including fixes from IPsec and wireless.
Previous releases - regressions:
- prevent NULL deref in generic_hwtstamp_ioctl_lower(),
newer APIs don't populate all the pointers in the request
- phylink: add missing supported link modes for the fixed-link
- mptcp: fix false positive warning in mptcp_pm_nl_rm_addr
Previous releases - always broken:
- openvswitch: remove never-working support for setting NSH fields
- xfrm: number of fixes for error paths of xfrm_state creation/
modification/deletion
- xfrm: fixes for offload
- fix the determination of the protocol of the inner packet
- don't push locally generated packets directly to L2 tunnel
mode offloading, they still need processing from the standard
xfrm path
- mptcp: fix a couple of corner cases in fallback and fastclose
handling
- wifi: rtw89: hw_scan: prevent connections from getting stuck,
work around apparent bug in FW by tweaking messages we send
- af_unix: fix duplicate data if PEEK w/ peek_offset needs to wait
- veth: more robust handing of race to avoid txq getting stuck
- eth: ps3_gelic_net: handle skb allocation failures"
* tag 'net-6.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (47 commits)
vsock: Ignore signal/timeout on connect() if already established
be2net: pass wrb_params in case of OS2BMC
l2tp: reset skb control buffer on xmit
net: dsa: microchip: lan937x: Fix RGMII delay tuning
selftests: mptcp: add a check for 'add_addr_accepted'
mptcp: fix address removal logic in mptcp_pm_nl_rm_addr
selftests: mptcp: join: userspace: longer timeout
selftests: mptcp: join: endpoints: longer timeout
selftests: mptcp: join: fastclose: remove flaky marks
mptcp: fix duplicate reset on fastclose
mptcp: decouple mptcp fastclose from tcp close
mptcp: do not fallback when OoO is present
mptcp: fix premature close in case of fallback
mptcp: avoid unneeded subflow-level drops
mptcp: fix ack generation for fallback msk
wifi: rtw89: hw_scan: Don't let the operating channel be last
net: phylink: add missing supported link modes for the fixed-link
selftest: af_unix: Add test for SO_PEEK_OFF.
af_unix: Read sk_peek_offset() again after sleeping in unix_stream_read_generic().
net/mlx5: Clean up only new IRQ glue on request_irq() failure
...
The CI reports sporadic failures of the fastclose self-tests. The root
cause is a duplicate reset, not carrying the relevant MPTCP option.
In the failing scenario the bad reset is received by the peer before
the fastclose one, preventing the reception of the latter.
Indeed there is window of opportunity at fastclose time for the
following race:
mptcp_do_fastclose
__mptcp_close_ssk
__tcp_close()
tcp_set_state() [1]
tcp_send_active_reset() [2]
After [1] the stack will send reset to in-flight data reaching the now
closed port. Such reset may race with [2].
Address the issue explicitly sending a single reset on fastclose before
explicitly moving the subflow to close status.
Fixes: d21f834855 ("mptcp: use fastclose on more edge scenarios")
Cc: stable@vger.kernel.org
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/596
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Geliang Tang <geliang@kernel.org>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251118-net-mptcp-misc-fixes-6-18-rc6-v1-6-806d3781c95f@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
With the current fastclose implementation, the mptcp_do_fastclose()
helper is in charge of two distinct actions: send the fastclose reset
and cleanup the subflows.
Formally decouple the two steps, ensuring that mptcp explicitly closes
all the subflows after the mentioned helper.
This will make the upcoming fix simpler, and allows dropping the 2nd
argument from mptcp_destroy_common(). The Fixes tag is then the same as
in the next commit to help with the backports.
Fixes: d21f834855 ("mptcp: use fastclose on more edge scenarios")
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Geliang Tang <geliang@kernel.org>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251118-net-mptcp-misc-fixes-6-18-rc6-v1-5-806d3781c95f@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
I'm observing very frequent self-tests failures in case of fallback when
running on a CONFIG_PREEMPT kernel.
The root cause is that subflow_sched_work_if_closed() closes any subflow
as soon as it is half-closed and has no incoming data pending.
That works well for regular subflows - MPTCP needs bi-directional
connectivity to operate on a given subflow - but for fallback socket is
race prone.
When TCP peer closes the connection before the MPTCP one,
subflow_sched_work_if_closed() will schedule the MPTCP worker to
gracefully close the subflow, and shortly after will do another schedule
to inject and process a dummy incoming DATA_FIN.
On CONFIG_PREEMPT kernel, the MPTCP worker can kick-in and close the
fallback subflow before subflow_sched_work_if_closed() is able to create
the dummy DATA_FIN, unexpectedly interrupting the transfer.
Address the issue explicitly avoiding closing fallback subflows on when
the peer is only half-closed.
Note that, when the subflow is able to create the DATA_FIN before the
worker invocation, the worker will change the msk state before trying to
close the subflow and will skip the latter operation as the msk will not
match anymore the precondition in __mptcp_close_subflow().
Fixes: f09b0ad55a ("mptcp: close subflow when receiving TCP+FIN")
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251118-net-mptcp-misc-fixes-6-18-rc6-v1-3-806d3781c95f@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
syzbot reported use-after-free in mptcp_schedule_work() [1]
Issue here is that mptcp_schedule_work() schedules a work,
then gets a refcount on sk->sk_refcnt if the work was scheduled.
This refcount will be released by mptcp_worker().
[A] if (schedule_work(...)) {
[B] sock_hold(sk);
return true;
}
Problem is that mptcp_worker() can run immediately and complete before [B]
We need instead :
sock_hold(sk);
if (schedule_work(...))
return true;
sock_put(sk);
[1]
refcount_t: addition on 0; use-after-free.
WARNING: CPU: 1 PID: 29 at lib/refcount.c:25 refcount_warn_saturate+0xfa/0x1d0 lib/refcount.c:25
Call Trace:
<TASK>
__refcount_add include/linux/refcount.h:-1 [inline]
__refcount_inc include/linux/refcount.h:366 [inline]
refcount_inc include/linux/refcount.h:383 [inline]
sock_hold include/net/sock.h:816 [inline]
mptcp_schedule_work+0x164/0x1a0 net/mptcp/protocol.c:943
mptcp_tout_timer+0x21/0xa0 net/mptcp/protocol.c:2316
call_timer_fn+0x17e/0x5f0 kernel/time/timer.c:1747
expire_timers kernel/time/timer.c:1798 [inline]
__run_timers kernel/time/timer.c:2372 [inline]
__run_timer_base+0x648/0x970 kernel/time/timer.c:2384
run_timer_base kernel/time/timer.c:2393 [inline]
run_timer_softirq+0xb7/0x180 kernel/time/timer.c:2403
handle_softirqs+0x22f/0x710 kernel/softirq.c:622
__do_softirq kernel/softirq.c:656 [inline]
run_ktimerd+0xcf/0x190 kernel/softirq.c:1138
smpboot_thread_fn+0x542/0xa60 kernel/smpboot.c:160
kthread+0x711/0x8a0 kernel/kthread.c:463
ret_from_fork+0x4bc/0x870 arch/x86/kernel/process.c:158
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245
Cc: stable@vger.kernel.org
Fixes: 3b1d6210a9 ("mptcp: implement and use MPTCP-level retransmission")
Reported-by: syzbot+355158e7e301548a1424@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/6915b46f.050a0220.3565dc.0028.GAE@google.com/T/#u
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251113103924.3737425-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The sockmap feature allows bpf syscall from userspace, or based
on bpf sockops, replacing the sk_prot of sockets during protocol stack
processing with sockmap's custom read/write interfaces.
'''
tcp_rcv_state_process()
syn_recv_sock()/subflow_syn_recv_sock()
tcp_init_transfer(BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB)
bpf_skops_established <== sockops
bpf_sock_map_update(sk) <== call bpf helper
tcp_bpf_update_proto() <== update sk_prot
'''
When the server has MPTCP enabled but the client sends a TCP SYN
without MPTCP, subflow_syn_recv_sock() performs a fallback on the
subflow, replacing the subflow sk's sk_prot with the native sk_prot.
'''
subflow_syn_recv_sock()
subflow_ulp_fallback()
subflow_drop_ctx()
mptcp_subflow_ops_undo_override()
'''
Then, this subflow can be normally used by sockmap, which replaces the
native sk_prot with sockmap's custom sk_prot. The issue occurs when the
user executes accept::mptcp_stream_accept::mptcp_fallback_tcp_ops().
Here, it uses sk->sk_prot to compare with the native sk_prot, but this
is incorrect when sockmap is used, as we may incorrectly set
sk->sk_socket->ops.
This fix uses the more generic sk_family for the comparison instead.
Additionally, this also prevents a WARNING from occurring:
result from ./scripts/decode_stacktrace.sh:
------------[ cut here ]------------
WARNING: CPU: 0 PID: 337 at net/mptcp/protocol.c:68 mptcp_stream_accept \
(net/mptcp/protocol.c:4005)
Modules linked in:
...
PKRU: 55555554
Call Trace:
<TASK>
do_accept (net/socket.c:1989)
__sys_accept4 (net/socket.c:2028 net/socket.c:2057)
__x64_sys_accept (net/socket.c:2067)
x64_sys_call (arch/x86/entry/syscall_64.c:41)
do_syscall_64 (arch/x86/entry/syscall_64.c:63 arch/x86/entry/syscall_64.c:94)
entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
RIP: 0033:0x7f87ac92b83d
---[ end trace 0000000000000000 ]---
Fixes: 0b4f33def7 ("mptcp: fix tcp fallback crash")
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Cc: <stable@vger.kernel.org>
Link: https://patch.msgid.link/20251111060307.194196-3-jiayuan.chen@linux.dev
Convert struct proto pre_connect(), connect(), bind(), and bind_add()
callback function prototypes from struct sockaddr to struct sockaddr_unsized.
This does not change per-implementation use of sockaddr for passing around
an arbitrarily sized sockaddr struct. Those will be addressed in future
patches.
Additionally removes the no longer referenced struct sockaddr from
include/net/inet_common.h.
No binary changes expected.
Signed-off-by: Kees Cook <kees@kernel.org>
Link: https://patch.msgid.link/20251104002617.2752303-5-kees@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Update all struct proto_ops bind() callback function prototypes from
"struct sockaddr *" to "struct sockaddr_unsized *" to avoid lying to the
compiler about object sizes. Calls into struct proto handlers gain casts
that will be removed in the struct proto conversion patch.
No binary changes expected.
Signed-off-by: Kees Cook <kees@kernel.org>
Link: https://patch.msgid.link/20251104002617.2752303-2-kees@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Accessing the transmit queue without owning the msk socket lock is
inherently racy, hence __mptcp_check_push() could actually quit early
even when there is pending data.
That in turn could cause unexpected tx lock and timeout.
Dropping the early check avoids the race, implicitly relaying on later
tests under the relevant lock. With such change, all the other
mptcp_send_head() call sites are now under the msk socket lock and we
can additionally drop the now unneeded annotation on the transmit head
pointer accesses.
Fixes: 6e628cd3a8 ("mptcp: use mptcp release_cb for delayed tasks")
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Geliang Tang <geliang@kernel.org>
Tested-by: Geliang Tang <geliang@kernel.org>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20251028-net-mptcp-send-timeout-v1-1-38ffff5a9ec8@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Some protocols (e.g., TCP, UDP) implement memory accounting for socket
buffers and charge memory to per-protocol global counters pointed to by
sk->sk_proto->memory_allocated.
Sometimes, system processes do not want that limitation. For a similar
purpose, there is SO_RESERVE_MEM for sockets under memcg.
Also, by opting out of the per-protocol accounting, sockets under memcg
can avoid paying costs for two orthogonal memory accounting mechanisms.
A microbenchmark result is in the subsequent bpf patch.
Let's allow opt-out from the per-protocol memory accounting if
sk->sk_bypass_prot_mem is true.
sk->sk_bypass_prot_mem and sk->sk_prot are placed in the same cache
line, and sk_has_account() always fetches sk->sk_prot before accessing
sk->sk_bypass_prot_mem, so there is no extra cache miss for this patch.
The following patches will set sk->sk_bypass_prot_mem to true, and
then, the per-protocol memory accounting will be skipped.
Note that this does NOT disable memcg, but rather the per-protocol one.
Another option not to use the hole in struct sock_common is create
sk_prot variants like tcp_prot_bypass, but this would complicate
SOCKMAP logic, tcp_bpf_prots etc.
Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Roman Gushchin <roman.gushchin@linux.dev>
Link: https://patch.msgid.link/20251014235604.3057003-3-kuniyu@google.com
