ovpn updates dev->dstats from both process and softirq contexts. In
particular, TCP paths may run from socket callbacks, workqueues or
strparser work, while UDP receive and ovpn's ndo_start_xmit path may
update the same per-device dstats from BH context.
Add ovpn device drop-stat helpers that disable BHs around
dev_dstats_rx_dropped() and dev_dstats_tx_dropped(), and use them for
drop accounting.
The successful RX dev_dstats_rx_add() update is already covered by the
BH-disabled section around gro_cells_receive(). For the successful TCP
TX dev_dstats_tx_add() update, replace the existing preempt-disabled
section with a BH-disabled one.
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
While deleting an existing ovpn interface, there is a very
narrow window where adding a new peer via netlink may cause
the netdevice to hang and prevent its unregistration.
It may happen during ovpn_dellink(), when all existing peers are
freed and the device is queued for deregistration, but a
CMD_PEER_NEW message comes in adding a new peer that takes again
a reference to the netdev.
At this point there is no way to release the device because we are
under the assumption that all peers were already released.
Fix the race condition by releasing all peers in ndo_uninit(),
when the netdevice has already been removed from the netdev
list.
Also ovpn_peer_add() has now an extra check that forces the
function to bail out if the device reg_state is not REGISTERED.
This way any incoming CMD_PEER_NEW racing with the interface
deletion routine will simply stop before adding the peer.
Note that the above check happens while holding the netdev_lock
to prevent racing netdev state changes.
ovpn_dellink() is now empty and can be removed.
Reported-by: Hyunwoo Kim <imv4bel@gmail.com>
Closes: https://lore.kernel.org/netdev/aaVgJ16edTfQkYbx@v4bel/
Suggested-by: Sabrina Dubroca <sd@queasysnail.net>
Fixes: 80747caef3 ("ovpn: introduce the ovpn_peer object")
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
ovpn_nl_peer_new_doit()'s error path calls ovpn_peer_release() directly
rather than ovpn_peer_put(), bypassing the kref. The accompanying
comment ("peer was not yet hashed, thus it is not used in any context")
holds for UDP but not for TCP.
For UDP, the ovpn_socket union uses the .ovpn arm and never points back
at a peer; UDP encap_recv looks up peers via the not-yet-populated
hashtables, so the new peer is unreachable until ovpn_peer_add()
publishes it.
For TCP, ovpn_socket_new() sets ovpn_sock->peer and
ovpn_tcp_socket_attach() publishes ovpn_sock via rcu_assign_sk_user_data().
From that moment until ovpn_socket_release() detaches in the error path,
the TCP fd is fully wired: userspace recvmsg / sendmsg / close / poll
on the fd, as well as the strparser-driven ovpn_tcp_rcv() path, can
reach the peer through sk_user_data -> ovpn_sock->peer and bump its
refcount via ovpn_peer_hold().
ovpn_tcp_socket_wait_finish() (called inside ovpn_socket_release())
drains strparser and the tx work, but does not synchronize with
userspace syscall callers that already hold a peer reference. If
ovpn_nl_peer_modify() or ovpn_peer_add() returns an error while such
a caller is in flight - notably an ovpn_tcp_recvmsg() blocked in
__skb_recv_datagram() on peer->tcp.user_queue - the direct
ovpn_peer_release() destroys the peer while the caller still holds
the reference, and the eventual ovpn_peer_put() from that caller
operates on freed memory.
Replace the direct destructor call with ovpn_peer_put() so the kref
correctly defers destruction until the last reference is dropped.
In the common case where no concurrent user is present, behaviour is
unchanged: the kref hits zero immediately and ovpn_peer_release_kref()
runs the same destructor.
With this conversion ovpn_peer_release() has no callers outside peer.c
- ovpn_peer_release_kref() in the same translation unit is the only
remaining user - so make it static and drop its declaration from
peer.h.
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Assisted-by: Claude:claude-opus-4-7
Signed-off-by: David Carlier <devnexen@gmail.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
ovpn_tcp_close() loads the ovpn_socket via rcu_dereference_sk_user_data()
under rcu_read_lock(), takes a reference on sock->peer, caches the peer
pointer in a local, and drops the read lock. It then passes sock->peer
(rather than the cached local) to ovpn_peer_del(), re-dereferencing the
ovpn_socket after the RCU read section has ended.
Unlike ovpn_tcp_sendmsg(), which uses the same "load under RCU, use
after unlock" pattern but is protected by lock_sock() held across the
function, ovpn_tcp_close() runs without the socket lock: inet_release()
invokes sk_prot->close() without taking lock_sock first.
ovpn_socket_release() can therefore complete its kref_put -> detach ->
synchronize_rcu -> kfree(sock) sequence concurrently, in the window
after ovpn_tcp_close() drops rcu_read_lock() but before it dereferences
sock->peer. The synchronize_rcu() in ovpn_socket_release() protects
readers that use the dereferenced pointer inside the RCU read section,
not those that escape the pointer to a local and use it afterwards.
A reproducer follows the pattern of commit 94560267d6 ("ovpn: tcp -
don't deref NULL sk_socket member after tcp_close()"): trigger a peer
removal (keepalive expiration or netlink OVPN_CMD_DEL_PEER) at the same
moment userspace closes the TCP fd. That commit fixed the detach-side
of the same race window; this one fixes the close-side at a different
victim.
Tighten the entry block to read sock->peer exactly once into the cached
peer local, and route all subsequent uses (the hold check, the
ovpn_peer_del() call, and the prot->close() invocation) through that
local. sock->peer is only ever written once in ovpn_socket_new() under
lock_sock(), before rcu_assign_sk_user_data() publishes the ovpn_socket,
and is never reassigned afterwards - but the previous multi-read pattern
made that invariant implicit rather than explicit. The same multi-read
shape exists in ovpn_tcp_recvmsg(), ovpn_tcp_sendmsg(),
ovpn_tcp_data_ready() and ovpn_tcp_write_space(); those will be cleaned
up via a dedicated helper in a follow-up net-next series.
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Assisted-by: Claude:claude-opus-4-7
Signed-off-by: David Carlier <devnexen@gmail.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
ovpn injects decrypted packets into the netdev RX path through
ovpn_netdev_write() which invokes gro_cells_receive() and
dev_dstats_rx_add().
ovpn_netdev_write() is normally called in softirq context,
however, in case of TCP connections it may also be invoked
process context.
When this happens gro_cells_receive() will throw a warning:
[ 230.183747][ T12] WARNING: net/core/gro_cells.c:30 at gro_cells_receive+0x708/0xaa0, CPU#1: kworker/u16:0/12
and lockdep will also report a potential inconsistent lock state:
WARNING: inconsistent lock state
7.0.0-rc4+ #246 Tainted: G W
--------------------------------
inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage.
because attempts to acquire gro_cells->bh_lock by both
contexts may lead to a deadlock.
At the same time, dev_dstats_rx_add() does not expect to race
with a softirq (which may happen when invoked in process context),
because the latter may access its per-cpu state and corrupt
it.
Fix all this by invoking local_bh_disable/enable() around
gro_cells_receive() and dev_dstats_rx_add() to ensure that
bottom halves are always disabled before calling both of
them.
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
After decapsulating a packet, the skb->mac_header still points to the
outer transport header.
Fix this by calling skb_reset_mac_header() in ovpn_netdev_write() to
ensure the MAC header points to the beginning of
the inner IP/network packet, as expected by the rest of the stack.
Reported-by: Minqiang Chen <ptpt52@gmail.com>
Fixes: 8534731dbf ("ovpn: implement packet processing")
Signed-off-by: Qingfang Deng <qingfang.deng@linux.dev>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Currently ovpn uses three separate dynamically allocated structures to
set up cryptographic operations for both encryption and decryption. This
adds overhead to performance-critical paths and contribute to memory
fragmentation.
This commit consolidates those allocations into a single temporary blob,
similar to what esp_alloc_tmp() does.
The resulting performance gain is +7.7% and +4.3% for UDP when using AES
and ChaChaPoly respectively, and +4.3% for TCP.
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
In order to support the multipeer architecture, upon connection setup
each side of a tunnel advertises a unique ID that the other side must
include in packets sent to them. Therefore when transmitting a packet, a
peer inserts the recipient's advertised ID for that specific tunnel into
the peer ID field. When receiving a packet, a peer expects to find its
own unique receive ID for that specific tunnel in the peer ID field.
Add support for the TX peer ID and embed it into transmitting packets.
If no TX peer ID is specified, fallback to using the same peer ID both
for RX and TX in order to be compatible with the non-multipeer compliant
peers.
Cc: horms@kernel.org
Cc: donald.hunter@gmail.com
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Use bitops.h for replay window to simplify code.
Signed-off-by: Qingfang Deng <dqfext@gmail.com>
[antonio@openvpn.net: extended commit message]
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
In ovpn_nl_key_swap_doit, the attributes array used to parse the
OVPN_A_KEYCONF uses OVPN_A_PEER_MAX instead of
OVPN_A_KEYCONF_MAX. Note that this does not cause any bug, since
currently OVPN_A_KEYCONF_MAX < OVPN_A_PEER_MAX.
The wrong constant was introduced by commit 203e2bf559
("ovpn: implement key add/get/del/swap via netlink")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Pull networking fixes from Paolo Abeni:
"Including fixes from IPsec, Bluetooth and netfilter
Current release - regressions:
- wifi: fix dev_alloc_name() return value check
- rds: fix recursive lock in rds_tcp_conn_slots_available
Current release - new code bugs:
- vsock: lock down child_ns_mode as write-once
Previous releases - regressions:
- core:
- do not pass flow_id to set_rps_cpu()
- consume xmit errors of GSO frames
- netconsole: avoid OOB reads, msg is not nul-terminated
- netfilter: h323: fix OOB read in decode_choice()
- tcp: re-enable acceptance of FIN packets when RWIN is 0
- udplite: fix null-ptr-deref in __udp_enqueue_schedule_skb().
- wifi: brcmfmac: fix potential kernel oops when probe fails
- phy: register phy led_triggers during probe to avoid AB-BA deadlock
- eth:
- bnxt_en: fix deleting of Ntuple filters
- wan: farsync: fix use-after-free bugs caused by unfinished tasklets
- xscale: check for PTP support properly
Previous releases - always broken:
- tcp: fix potential race in tcp_v6_syn_recv_sock()
- kcm: fix zero-frag skb in frag_list on partial sendmsg error
- xfrm:
- fix race condition in espintcp_close()
- always flush state and policy upon NETDEV_UNREGISTER event
- bluetooth:
- purge error queues in socket destructors
- fix response to L2CAP_ECRED_CONN_REQ
- eth:
- mlx5:
- fix circular locking dependency in dump
- fix "scheduling while atomic" in IPsec MAC address query
- gve: fix incorrect buffer cleanup for QPL
- team: avoid NETDEV_CHANGEMTU event when unregistering slave
- usb: validate USB endpoints"
* tag 'net-7.0-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (72 commits)
netfilter: nf_conntrack_h323: fix OOB read in decode_choice()
dpaa2-switch: validate num_ifs to prevent out-of-bounds write
net: consume xmit errors of GSO frames
vsock: document write-once behavior of the child_ns_mode sysctl
vsock: lock down child_ns_mode as write-once
selftests/vsock: change tests to respect write-once child ns mode
net/mlx5e: Fix "scheduling while atomic" in IPsec MAC address query
net/mlx5: Fix missing devlink lock in SRIOV enable error path
net/mlx5: E-switch, Clear legacy flag when moving to switchdev
net/mlx5: LAG, disable MPESW in lag_disable_change()
net/mlx5: DR, Fix circular locking dependency in dump
selftests: team: Add a reference count leak test
team: avoid NETDEV_CHANGEMTU event when unregistering slave
net: mana: Fix double destroy_workqueue on service rescan PCI path
MAINTAINERS: Update maintainer entry for QUALCOMM ETHQOS ETHERNET DRIVER
dpll: zl3073x: Remove redundant cleanup in devm_dpll_init()
selftests/net: packetdrill: Verify acceptance of FIN packets when RWIN is 0
tcp: re-enable acceptance of FIN packets when RWIN is 0
vsock: Use container_of() to get net namespace in sysctl handlers
net: usb: kaweth: validate USB endpoints
...
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
When processing TCP stream data in ovpn_tcp_recv, we receive large
cloned skbs from __strp_rcv that may contain multiple coalesced packets.
The current implementation has two bugs:
1. Header offset overflow: Using pskb_pull with large offsets on
coalesced skbs causes skb->data - skb->head to exceed the u16 storage
of skb->network_header. This causes skb_reset_network_header to fail
on the inner decapsulated packet, resulting in packet drops.
2. Unaligned protocol headers: Extracting packets from arbitrary
positions within the coalesced TCP stream provides no alignment
guarantees for the packet data causing performance penalties on
architectures without efficient unaligned access. Additionally,
openvpn's 2-byte length prefix on TCP packets causes the subsequent
4-byte opcode and packet ID fields to be inherently misaligned.
Fix both issues by allocating a new skb for each openvpn packet and
using skb_copy_bits to extract only the packet content into the new
buffer, skipping the 2-byte length prefix. Also, check the length before
invoking the function that performs the allocation to avoid creating an
invalid skb.
If the packet has to be forwarded to userspace the 2-byte prefix can be
pushed to the head safely, without misalignment.
As a side effect, this approach also avoids the expensive linearization
that pskb_pull triggers on cloned skbs with page fragments. In testing,
this resulted in TCP throughput improvements of up to 74%.
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
When deleting a peer in case of keepalive expiration, the peer is
removed from the OpenVPN hashtable and is temporary inserted in a
"release list" for further processing.
This happens in:
ovpn_peer_keepalive_work()
unlock_ovpn(release_list)
This processing includes detaching from the socket being used to
talk to this peer, by restoring its original proto and socket
ops/callbacks.
In case of TCP it may happen that, while the peer is sitting in
the release list, userspace decides to close the socket.
This will result in a concurrent execution of:
tcp_close(sk)
__tcp_close(sk)
sock_orphan(sk)
sk_set_socket(sk, NULL)
The last function call will set sk->sk_socket to NULL.
When the releasing routine is resumed, ovpn_tcp_socket_detach()
will attempt to dereference sk->sk_socket to restore its original
ops member. This operation will crash due to sk->sk_socket being NULL.
Fix this race condition by testing-and-accessing
sk->sk_socket atomically under sk->sk_callback_lock.
Link: https://lore.kernel.org/netdev/176996279620.3109699.15382994681575380467@eldamar.lan/
Link: https://github.com/OpenVPN/ovpn-net-next/issues/29
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Link: https://patch.msgid.link/20260212213130.11497-1-antonio@openvpn.net
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Antonio Quartulli says:
====================
This batch includes the following fixes:
* set sk_user_data before installing callbacks, to avoid dropping early
packets
* fix use-after-free in ovpn_net_xmit when accessing shared skbs that
got released
* fix TX bytes stats by adding-up every positively processed GSO
segment
====================
Link: https://patch.msgid.link/20260212210340.11260-1-antonio@openvpn.net
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
In ovpn_net_xmit, after GSO segmentation and segment processing, the
first segment on the list is used to increment VPN TX statistics, which
fails to account for any subsequent segments in the chain.
Fix this by accumulating the length of every segment that successfully
passes skb_share_check into a tx_bytes variable. This ensures the peer
statistics accurately reflect the total data volume sent, regardless of
whether the original packet was segmented.
Fixes: 04ca14955f ("ovpn: store tunnel and transport statistics")
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
When building the skb_list in ovpn_net_xmit, skb_share_check will free
the original skb if it is shared. The current implementation continues
to use the stale skb pointer for subsequent operations:
- peer lookup,
- skb_dst_drop (even though all segments produced by skb_gso_segment
will have a dst attached),
- ovpn_peer_stats_increment_tx.
Fix this by moving the peer lookup and skb_dst_drop before segmentation
so that the original skb is still valid when used. Return early if all
segments fail skb_share_check and the list ends up empty.
Also switch ovpn_peer_stats_increment_tx to use skb_list.next; the next
patch fixes the stats logic.
Fixes: 08857b5ec5 ("ovpn: implement basic TX path (UDP)")
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
During initialization, we override socket callbacks and set sk_user_data
to an ovpn_socket instance. Currently, these two operations are
decoupled: callbacks are overridden before sk_user_data is set. While
existing callbacks perform safety checks for NULL or non-ovpn
sk_user_data, this condition causes a "half-formed" state where valid
packets arriving during attachment trigger error logs (e.g., "invoked on
non ovpn socket").
Set sk_user_data before overriding the callbacks so that it can be
accessed safely from them. Since we already check that the socket has no
sk_user_data before setting it, this remains safe even if an interrupt
accesses the socket after sk_user_data is set but before the callbacks
are overridden.
This also requires initializing all protocol-specific fields (such as
tcp_tx_work and peer links) before calling ovpn_socket_attach, ensuring
the ovpn_socket is fully formed before it becomes visible to any
callback.
Fixes: f6226ae7a0 ("ovpn: introduce the ovpn_socket object")
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
This patch continues the effort to refactor workqueue APIs, which has begun
with the changes introducing new workqueues and a new alloc_workqueue flag:
commit 128ea9f6cc ("workqueue: Add system_percpu_wq and system_dfl_wq")
commit 930c2ea566 ("workqueue: Add new WQ_PERCPU flag")
The point of the refactoring is to eventually alter the default behavior of
workqueues to become unbound by default so that their workload placement is
optimized by the scheduler.
Before that to happen after a careful review and conversion of each individual
case, workqueue users must be converted to the better named new workqueues with
no intended behaviour changes:
system_wq -> system_percpu_wq
system_unbound_wq -> system_dfl_wq
This way the old obsolete workqueues (system_wq, system_unbound_wq) can be
removed in the future.
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
Acked-by: Antonio Quartulli <antonio@openvpn.net>
Link: https://patch.msgid.link/20251224155006.114824-1-marco.crivellari@suse.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
openvpn TCP encapsulation uses a custom queue to deliver packets to
userspace. Currently it relies on datagram_poll, which checks
sk_receive_queue, leading to false readiness signals when that queue
contains non-userspace packets.
Switch ovpn_tcp_poll to use datagram_poll_queue with the peer's
user_queue, ensuring poll only signals readiness when userspace data is
actually available. Also refactor ovpn_tcp_poll in order to enforce the
assumption we can make on the lifetime of ovpn_sock and peer.
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Link: https://patch.msgid.link/20251021100942.195010-4-ralf@mandelbit.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
