Load balance new TCP connections across nexthops also when they
connect to the same service at a single remote address and port.
This affects only port-based multipath hashing:
fib_multipath_hash_policy 1 or 3.
Local connections must choose both a source address and port when
connecting to a remote service, in ip_route_connect. This
"chicken-and-egg problem" (commit 2d7192d6cb ("ipv4: Sanitize and
simplify ip_route_{connect,newports}()")) is resolved by first
selecting a source address, by looking up a route using the zero
wildcard source port and address.
As a result multiple connections to the same destination address and
port have no entropy in fib_multipath_hash.
This is not a problem when forwarding, as skb-based hashing has a
4-tuple. Nor when establishing UDP connections, as autobind there
selects a port before reaching ip_route_connect.
Load balance also TCP, by using a random port in fib_multipath_hash.
Port assignment in inet_hash_connect is not atomic with
ip_route_connect. Thus ports are unpredictable, effectively random.
Implementation details:
Do not actually pass a random fl4_sport, as that affects not only
hashing, but routing more broadly, and can match a source port based
policy route, which existing wildcard port 0 will not. Instead,
define a new wildcard flowi flag that is used only for hashing.
Selecting a random source is equivalent to just selecting a random
hash entirely. But for code clarity, follow the normal 4-tuple hash
process and only update this field.
fib_multipath_hash can be reached with zero sport from other code
paths, so explicitly pass this flowi flag, rather than trying to infer
this case in the function itself.
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Link: https://patch.msgid.link/20250424143549.669426-3-willemdebruijn.kernel@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
With multipath routes, try to ensure that packets leave on the device
that is associated with the source address.
Avoid the following tcpdump example:
veth0 Out IP 10.1.0.2.38640 > 10.2.0.3.8000: Flags [S]
veth1 Out IP 10.1.0.2.38648 > 10.2.0.3.8000: Flags [S]
Which can happen easily with the most straightforward setup:
ip addr add 10.0.0.1/24 dev veth0
ip addr add 10.1.0.1/24 dev veth1
ip route add 10.2.0.3 nexthop via 10.0.0.2 dev veth0 \
nexthop via 10.1.0.2 dev veth1
This is apparently considered WAI, based on the comment in
ip_route_output_key_hash_rcu:
* 2. Moreover, we are allowed to send packets with saddr
* of another iface. --ANK
It may be ok for some uses of multipath, but not all. For instance,
when using two ISPs, a router may drop packets with unknown source.
The behavior occurs because tcp_v4_connect makes three route
lookups when establishing a connection:
1. ip_route_connect calls to select a source address, with saddr zero.
2. ip_route_connect calls again now that saddr and daddr are known.
3. ip_route_newports calls again after a source port is also chosen.
With a route with multiple nexthops, each lookup may make a different
choice depending on available entropy to fib_select_multipath. So it
is possible for 1 to select the saddr from the first entry, but 3 to
select the second entry. Leading to the above situation.
Address this by preferring a match that matches the flowi4 saddr. This
will make 2 and 3 make the same choice as 1. Continue to update the
backup choice until a choice that matches saddr is found.
Do this in fib_select_multipath itself, rather than passing an fl4_oif
constraint, to avoid changing non-multipath route selection. Commit
e6b45241c5 ("ipv4: reset flowi parameters on route connect") shows
how that may cause regressions.
Also read ipv4.sysctl_fib_multipath_use_neigh only once. No need to
refresh in the loop.
This does not happen in IPv6, which performs only one lookup.
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Link: https://patch.msgid.link/20250424143549.669426-2-willemdebruijn.kernel@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The "noqueue" qdisc can either be directly attached, or get default
attached if net_device priv_flags has IFF_NO_QUEUE. In both cases, the
allocated Qdisc structure gets it's enqueue function pointer reset to
NULL by noqueue_init() via noqueue_qdisc_ops.
This is a common case for software virtual net_devices. For these devices
with no-queue, the transmission path in __dev_queue_xmit() will bypass
the qdisc layer. Directly invoking device drivers ndo_start_xmit (via
dev_hard_start_xmit). In this mode the device driver is not allowed to
ask for packets to be queued (either via returning NETDEV_TX_BUSY or
stopping the TXQ).
The simplest and most reliable way to identify this no-queue case is by
checking if enqueue == NULL.
The vrf driver currently open-codes this check (!qdisc->enqueue). While
functionally correct, this low-level detail is better encapsulated in a
dedicated helper for clarity and long-term maintainability.
To make this behavior more explicit and reusable, this patch introduce a
new helper: qdisc_txq_has_no_queue(). Helper will also be used by the
veth driver in the next patch, which introduces optional qdisc-based
backpressure.
This is a non-functional change.
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://patch.msgid.link/174559293172.827981.7583862632045264175.stgit@firesoul
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
BIG Create Sync requires the command to just generates a status so this
makes use of __hci_cmd_sync_status_sk to wait for
HCI_EVT_LE_BIG_SYNC_ESTABLISHED, also because of this chance it is not
longer necessary to use a custom method to serialize the process of
creating the BIG sync since the cmd_work_sync itself ensures only one
command would be pending which now awaits for
HCI_EVT_LE_BIG_SYNC_ESTABLISHED before proceeding to next connection.
Fixes: 42ecf19471 ("Bluetooth: ISO: Do not emit LE BIG Create Sync if previous is pending")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Broadcast Receiver requires creating PA sync but the command just
generates a status so this makes use of __hci_cmd_sync_status_sk to wait
for HCI_EV_LE_PA_SYNC_ESTABLISHED, also because of this chance it is not
longer necessary to use a custom method to serialize the process of
creating the PA sync since the cmd_work_sync itself ensures only one
command would be pending which now awaits for
HCI_EV_LE_PA_SYNC_ESTABLISHED before proceeding to next connection.
Fixes: 4a5e0ba686 ("Bluetooth: ISO: Do not emit LE PA Create Sync if previous is pending")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Move rx_lock from xsk_socket to xsk_buff_pool.
Fix synchronization for shared umem mode in
generic RX path where multiple sockets share
single xsk_buff_pool.
RX queue is exclusive to xsk_socket, while FILL
queue can be shared between multiple sockets.
This could result in race condition where two
CPU cores access RX path of two different sockets
sharing the same umem.
Protect both queues by acquiring spinlock in shared
xsk_buff_pool.
Lock contention may be minimized in the future by some
per-thread FQ buffering.
It's safe and necessary to move spin_lock_bh(rx_lock)
after xsk_rcv_check():
* xs->pool and spinlock_init is synchronized by
xsk_bind() -> xsk_is_bound() memory barriers.
* xsk_rcv_check() may return true at the moment
of xsk_release() or xsk_unbind_dev(),
however this will not cause any data races or
race conditions. xsk_unbind_dev() removes xdp
socket from all maps and waits for completion
of all outstanding rx operations. Packets in
RX path will either complete safely or drop.
Signed-off-by: Eryk Kubanski <e.kubanski@partner.samsung.com>
Fixes: bf0bdd1343 ("xdp: fix race on generic receive path")
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://patch.msgid.link/20250416101908.10919-1-e.kubanski@partner.samsung.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Remove three functions that are no longer used.
rxrpc_get_txbuf() last use was removed by 2020's
commit 5e6ef4f101 ("rxrpc: Make the I/O thread take over the call and
local processor work")
rxrpc_kernel_get_epoch() last use was removed by 2020's
commit 44746355cc ("afs: Don't get epoch from a server because it may be
ambiguous")
rxrpc_kernel_set_max_life() last use was removed by 2023's
commit db099c625b ("rxrpc: Fix timeout of a call that hasn't yet been
granted a channel")
Both of the rxrpc_kernel_* functions were documented. Remove that
documentation as well as the code.
Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org>
Acked-by: David Howells <dhowells@redhat.com>
Link: https://patch.msgid.link/20250422235147.146460-1-linux@treblig.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We will get rid of RTNL from RTM_NEWROUTE and SIOCADDRT.
Then, we may be going to add a route tied to a dying nexthop.
The nexthop itself is not freed during the RCU grace period, but
if we link a route after __remove_nexthop_fib() is called for the
nexthop, the route will be leaked.
To avoid the race between IPv6 route addition under RCU vs nexthop
deletion under RTNL, let's add a dead flag and protect it and
nh->f6i_list with a spinlock.
__remove_nexthop_fib() acquires the nexthop's spinlock and sets false
to nh->dead, then calls ip6_del_rt() for the linked route one by one
without the spinlock because fib6_purge_rt() acquires it later.
While adding an IPv6 route, fib6_add() acquires the nexthop lock and
checks the dead flag just before inserting the route.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://patch.msgid.link/20250418000443.43734-15-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The next patch adds per-nexthop spinlock which protects nh->f6i_list.
When rt->nh is not NULL, fib6_add_rt2node() will be called under the lock.
fib6_add_rt2node() could call fib6_purge_rt() for another route, which
could holds another nexthop lock.
Then, deadlock could happen between two nexthops.
Let's defer fib6_purge_rt() after fib6_add_rt2node().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Link: https://patch.msgid.link/20250418000443.43734-14-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
We will get rid of RTNL from RTM_NEWROUTE and SIOCADDRT.
If the request specifies a new table ID, fib6_new_table() is
called to create a new routing table.
Two concurrent requests could specify the same table ID, so we
need a lock to protect net->ipv6.fib_table_hash[h].
Let's add a spinlock to protect the hash bucket linkage.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Link: https://patch.msgid.link/20250418000443.43734-13-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Commit 03df156dd3 ("xdp: double protect netdev->xdp_flags with
netdev->lock") introduces the netdev lock to xdp_set_features_flag().
The change includes a _locked version of the method, as it is possible
for a driver to have already acquired the netdev lock before calling
this helper. However, the same applies to
xdp_features_(set|clear)_redirect_flags(), which ends up calling the
unlocked version of xdp_set_features_flags() leading to deadlocks in
GVE, which grabs the netdev lock as part of its suspend, reset, and
shutdown processes:
[ 833.265543] WARNING: possible recursive locking detected
[ 833.270949] 6.15.0-rc1 #6 Tainted: G E
[ 833.276271] --------------------------------------------
[ 833.281681] systemd-shutdow/1 is trying to acquire lock:
[ 833.287090] ffff949d2b148c68 (&dev->lock){+.+.}-{4:4}, at: xdp_set_features_flag+0x29/0x90
[ 833.295470]
[ 833.295470] but task is already holding lock:
[ 833.301400] ffff949d2b148c68 (&dev->lock){+.+.}-{4:4}, at: gve_shutdown+0x44/0x90 [gve]
[ 833.309508]
[ 833.309508] other info that might help us debug this:
[ 833.316130] Possible unsafe locking scenario:
[ 833.316130]
[ 833.322142] CPU0
[ 833.324681] ----
[ 833.327220] lock(&dev->lock);
[ 833.330455] lock(&dev->lock);
[ 833.333689]
[ 833.333689] *** DEADLOCK ***
[ 833.333689]
[ 833.339701] May be due to missing lock nesting notation
[ 833.339701]
[ 833.346582] 5 locks held by systemd-shutdow/1:
[ 833.351205] #0: ffffffffa9c89130 (system_transition_mutex){+.+.}-{4:4}, at: __se_sys_reboot+0xe6/0x210
[ 833.360695] #1: ffff93b399e5c1b8 (&dev->mutex){....}-{4:4}, at: device_shutdown+0xb4/0x1f0
[ 833.369144] #2: ffff949d19a471b8 (&dev->mutex){....}-{4:4}, at: device_shutdown+0xc2/0x1f0
[ 833.377603] #3: ffffffffa9eca050 (rtnl_mutex){+.+.}-{4:4}, at: gve_shutdown+0x33/0x90 [gve]
[ 833.386138] #4: ffff949d2b148c68 (&dev->lock){+.+.}-{4:4}, at: gve_shutdown+0x44/0x90 [gve]
Introduce xdp_features_(set|clear)_redirect_target_locked() versions
which assume that the netdev lock has already been acquired before
setting the XDP feature flag and update GVE to use the locked version.
Fixes: 03df156dd3 ("xdp: double protect netdev->xdp_flags with netdev->lock")
Tested-by: Mina Almasry <almasrymina@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Harshitha Ramamurthy <hramamurthy@google.com>
Signed-off-by: Joshua Washington <joshwash@google.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20250422011643.3509287-1-joshwash@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
p8022.c defines two external functions, register_8022_client()
and unregister_8022_client(), the last use of which was removed in
2018 by
commit 7a2e838d28 ("staging: ipx: delete it from the tree")
Remove the p8022.c file, it's corresponding header, and glue
surrounding it. There was one place the header was included in vlan.c
but it didn't use the functions it declared.
There was a comment in net/802/Makefile about checking
against net/core/Makefile, but that's at least 20 years old and
there's no sign of net/core/Makefile mentioning it.
Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org>
Link: https://patch.msgid.link/20250418011519.145320-1-linux@treblig.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
FDB entries are currently stored in a hash table with a fixed number of
buckets (256), resulting in performance degradation as the number of
entries grows. Solve this by converting the driver to use rhashtable
which maintains more or less constant performance regardless of the
number of entries.
Measured transmitted packets per second using a single pktgen thread
with varying number of entries when the transmitted packet always hits
the default entry (worst case):
Number of entries | Improvement
------------------|------------
1k | +1.12%
4k | +9.22%
16k | +55%
64k | +585%
256k | +2460%
In addition, the change reduces the size of the VXLAN device structure
from 2584 bytes to 672 bytes.
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Link: https://patch.msgid.link/20250415121143.345227-16-idosch@nvidia.com
Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Currently, FDB entries are stored in a hash table with a fixed number of
buckets. The table is used for both lookups and entry traversal.
Subsequent patches will convert the table to rhashtable which is not
suitable for entry traversal.
In preparation for this conversion, add FDB entries to a linked list.
Subsequent patches will convert the driver to use this list when
traversing entries during dump, flush, etc.
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Link: https://patch.msgid.link/20250415121143.345227-8-idosch@nvidia.com
Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Currently, the VXLAN driver stores FDB entries in a hash table with a
fixed number of buckets (256). Subsequent patches are going to convert
this table to rhashtable with a linked list for entry traversal, as
rhashtable is more scalable.
In preparation for this conversion, move from a per-bucket spin lock to
a single spin lock that protects the entire FDB table.
The per-bucket spin locks were introduced by commit fe1e0713bb
("vxlan: Use FDB_HASH_SIZE hash_locks to reduce contention") citing
"huge contention when inserting/deleting vxlan_fdbs into the fdb_head".
It is not clear from the commit message which code path was holding the
spin lock for long periods of time, but the obvious suspect is the FDB
cleanup routine (vxlan_cleanup()) that periodically traverses the entire
table in order to delete aged-out entries.
This will be solved by subsequent patches that will convert the FDB
cleanup routine to traverse the linked list of FDB entries using RCU,
only acquiring the spin lock when deleting an aged-out entry.
The change reduces the size of the VXLAN device structure from 3600
bytes to 2576 bytes.
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Link: https://patch.msgid.link/20250415121143.345227-7-idosch@nvidia.com
Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Currently, bnxt_en driver satisfies the requirements of the Device
memory TCP, which is HDS.
So, it implements rx-side Device memory TCP for bnxt_en driver.
It requires only converting the page API to netmem API.
`struct page` of agg rings are changed to `netmem_ref netmem` and
corresponding functions are changed to a variant of netmem API.
It also passes PP_FLAG_ALLOW_UNREADABLE_NETMEM flag to a parameter of
page_pool.
The netmem will be activated only when a user requests devmem TCP.
When netmem is activated, received data is unreadable and netmem is
disabled, received data is readable.
But drivers don't need to handle both cases because netmem core API will
handle it properly.
So, using proper netmem API is enough for drivers.
Device memory TCP can be tested with
tools/testing/selftests/drivers/net/hw/ncdevmem.
This is tested with BCM57504-N425G and firmware version 232.0.155.8/pkg
232.1.132.8.
Reviewed-by: Mina Almasry <almasrymina@google.com>
Tested-by: David Wei <dw@davidwei.uk>
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Link: https://patch.msgid.link/20250415052458.1260575-1-ap420073@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Before commit 40867d74c3 ("net: Add l3mdev index to flow struct and
avoid oif reset for port devices") it was possible to use FIB rules to
match on a L3 domain. This was done by having a FIB rule match on iif /
oif being a L3 master device. It worked because prior to the FIB rule
lookup the iif / oif fields in the flow structure were reset to the
index of the L3 master device to which the input / output device was
enslaved to.
The above scheme made it impossible to match on the original input /
output device. Therefore, cited commit stopped overwriting the iif / oif
fields in the flow structure and instead stored the index of the
enslaving L3 master device in a new field ('flowi_l3mdev') in the flow
structure.
While the change enabled new use cases, it broke the original use case
of matching on a L3 domain. Fix this by interpreting the iif / oif
matching on a L3 master device as a match against the L3 domain. In
other words, if the iif / oif in the FIB rule points to a L3 master
device, compare the provided index against 'flowi_l3mdev' rather than
'flowi_{i,o}if'.
Before cited commit, a FIB rule that matched on 'iif vrf1' would only
match incoming traffic from devices enslaved to 'vrf1'. With the
proposed change (i.e., comparing against 'flowi_l3mdev'), the rule would
also match traffic originating from a socket bound to 'vrf1'. Avoid that
by adding a new flow flag ('FLOWI_FLAG_L3MDEV_OIF') that indicates if
the L3 domain was derived from the output interface or the input
interface (when not set) and take this flag into account when evaluating
the FIB rule against the flow structure.
Avoid unnecessary checks in the data path by detecting that a rule
matches on a L3 master device when the rule is installed and marking it
as such.
Tested using the following script [1].
Output before 40867d74c3 (v5.4.291):
default dev dummy1 table 100 scope link
default dev dummy1 table 200 scope link
Output after 40867d74c3:
default dev dummy1 table 300 scope link
default dev dummy1 table 300 scope link
Output with this patch:
default dev dummy1 table 100 scope link
default dev dummy1 table 200 scope link
[1]
#!/bin/bash
ip link add name vrf1 up type vrf table 10
ip link add name dummy1 up master vrf1 type dummy
sysctl -wq net.ipv4.conf.all.forwarding=1
sysctl -wq net.ipv4.conf.all.rp_filter=0
ip route add table 100 default dev dummy1
ip route add table 200 default dev dummy1
ip route add table 300 default dev dummy1
ip rule add prio 0 oif vrf1 table 100
ip rule add prio 1 iif vrf1 table 200
ip rule add prio 2 table 300
ip route get 192.0.2.1 oif dummy1 fibmatch
ip route get 192.0.2.1 iif dummy1 from 198.51.100.1 fibmatch
Fixes: 40867d74c3 ("net: Add l3mdev index to flow struct and avoid oif reset for port devices")
Reported-by: hanhuihui <hanhuihui5@huawei.com>
Closes: https://lore.kernel.org/netdev/ec671c4f821a4d63904d0da15d604b75@huawei.com/
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Acked-by: David Ahern <dsahern@kernel.org>
Link: https://patch.msgid.link/20250414172022.242991-2-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Allow the app to request that CHALLENGEs be passed to it through an
out-of-band queue that allows recvmsg() to pick it up so that the app can
add data to it with sendmsg().
This will allow the application (AFS or userspace) to interact with the
process if it wants to and put values into user-defined fields. This will
be used by AFS when talking to a fileserver to supply that fileserver with
a crypto key by which callback RPCs can be encrypted (ie. notifications
from the fileserver to the client).
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: Simon Horman <horms@kernel.org>
cc: linux-afs@lists.infradead.org
Link: https://patch.msgid.link/20250411095303.2316168-5-dhowells@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
A number of functions separately furnish an AF_RXRPC socket with callback
function pointers into a kernel app (such as the AFS filesystem) that is
using it. Replace most of these with an ops table for the entire socket.
This makes it easier to add more callback functions.
Note that the call incoming data processing callback is retaind as that
gets set to different things, depending on the type of op.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: Simon Horman <horms@kernel.org>
cc: linux-afs@lists.infradead.org
Link: https://patch.msgid.link/20250411095303.2316168-3-dhowells@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
struct pernet_operations provides two batching hooks; ->exit_batch()
and ->exit_batch_rtnl().
The batching variant is beneficial if ->exit() meets any of the
following conditions:
1) ->exit() repeatedly acquires a global lock for each netns
2) ->exit() has a time-consuming operation that can be factored
out (e.g. synchronize_rcu(), smp_mb(), etc)
3) ->exit() does not need to repeat the same iterations for each
netns (e.g. inet_twsk_purge())
Currently, none of the ->exit_batch_rtnl() functions satisfy any of
the above conditions because RTNL is factored out and held by the
caller and all of these functions iterate over the dying netns list.
Also, we want to hold per-netns RTNL there but avoid spreading
__rtnl_net_lock() across multiple locations.
Let's add ->exit_rtnl() hook and run it under __rtnl_net_lock().
The following patches will convert all ->exit_batch_rtnl() users
to ->exit_rtnl().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Link: https://patch.msgid.link/20250411205258.63164-4-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When enabling DMA mapping in page_pool, pages are kept DMA mapped until
they are released from the pool, to avoid the overhead of re-mapping the
pages every time they are used. This causes resource leaks and/or
crashes when there are pages still outstanding while the device is torn
down, because page_pool will attempt an unmap through a non-existent DMA
device on the subsequent page return.
To fix this, implement a simple tracking of outstanding DMA-mapped pages
in page pool using an xarray. This was first suggested by Mina[0], and
turns out to be fairly straight forward: We simply store pointers to
pages directly in the xarray with xa_alloc() when they are first DMA
mapped, and remove them from the array on unmap. Then, when a page pool
is torn down, it can simply walk the xarray and unmap all pages still
present there before returning, which also allows us to get rid of the
get/put_device() calls in page_pool. Using xa_cmpxchg(), no additional
synchronisation is needed, as a page will only ever be unmapped once.
To avoid having to walk the entire xarray on unmap to find the page
reference, we stash the ID assigned by xa_alloc() into the page
structure itself, using the upper bits of the pp_magic field. This
requires a couple of defines to avoid conflicting with the
POINTER_POISON_DELTA define, but this is all evaluated at compile-time,
so does not affect run-time performance. The bitmap calculations in this
patch gives the following number of bits for different architectures:
- 23 bits on 32-bit architectures
- 21 bits on PPC64 (because of the definition of ILLEGAL_POINTER_VALUE)
- 32 bits on other 64-bit architectures
Stashing a value into the unused bits of pp_magic does have the effect
that it can make the value stored there lie outside the unmappable
range (as governed by the mmap_min_addr sysctl), for architectures that
don't define ILLEGAL_POINTER_VALUE. This means that if one of the
pointers that is aliased to the pp_magic field (such as page->lru.next)
is dereferenced while the page is owned by page_pool, that could lead to
a dereference into userspace, which is a security concern. The risk of
this is mitigated by the fact that (a) we always clear pp_magic before
releasing a page from page_pool, and (b) this would need a
use-after-free bug for struct page, which can have many other risks
since page->lru.next is used as a generic list pointer in multiple
places in the kernel. As such, with this patch we take the position that
this risk is negligible in practice. For more discussion, see[1].
Since all the tracking added in this patch is performed on DMA
map/unmap, no additional code is needed in the fast path, meaning the
performance overhead of this tracking is negligible there. A
micro-benchmark shows that the total overhead of the tracking itself is
about 400 ns (39 cycles(tsc) 395.218 ns; sum for both map and unmap[2]).
Since this cost is only paid on DMA map and unmap, it seems like an
acceptable cost to fix the late unmap issue. Further optimisation can
narrow the cases where this cost is paid (for instance by eliding the
tracking when DMA map/unmap is a no-op).
The extra memory needed to track the pages is neatly encapsulated inside
xarray, which uses the 'struct xa_node' structure to track items. This
structure is 576 bytes long, with slots for 64 items, meaning that a
full node occurs only 9 bytes of overhead per slot it tracks (in
practice, it probably won't be this efficient, but in any case it should
be an acceptable overhead).
[0] https://lore.kernel.org/all/CAHS8izPg7B5DwKfSuzz-iOop_YRbk3Sd6Y4rX7KBG9DcVJcyWg@mail.gmail.com/
[1] https://lore.kernel.org/r/20250320023202.GA25514@openwall.com
[2] https://lore.kernel.org/r/ae07144c-9295-4c9d-a400-153bb689fe9e@huawei.com
Reported-by: Yonglong Liu <liuyonglong@huawei.com>
Closes: https://lore.kernel.org/r/8743264a-9700-4227-a556-5f931c720211@huawei.com
Fixes: ff7d6b27f8 ("page_pool: refurbish version of page_pool code")
Suggested-by: Mina Almasry <almasrymina@google.com>
Reviewed-by: Mina Almasry <almasrymina@google.com>
Reviewed-by: Jesper Dangaard Brouer <hawk@kernel.org>
Tested-by: Jesper Dangaard Brouer <hawk@kernel.org>
Tested-by: Qiuling Ren <qren@redhat.com>
Tested-by: Yuying Ma <yuma@redhat.com>
Tested-by: Yonglong Liu <liuyonglong@huawei.com>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://patch.msgid.link/20250409-page-pool-track-dma-v9-2-6a9ef2e0cba8@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
DCCP was orphaned in 2021 by commit 054c4610bd ("MAINTAINERS: dccp:
move Gerrit Renker to CREDITS"), which noted that the last maintainer
had been inactive for five years.
In recent years, it has become a playground for syzbot, and most changes
to DCCP have been odd bug fixes triggered by syzbot. Apart from that,
the only changes have been driven by treewide or networking API updates
or adjustments related to TCP.
Thus, in 2023, we announced we would remove DCCP in 2025 via commit
b144fcaf46 ("dccp: Print deprecation notice.").
Since then, only one individual has contacted the netdev mailing list. [0]
There is ongoing research for Multipath DCCP. The repository is hosted
on GitHub [1], and development is not taking place through the upstream
community. While the repository is published under the GPLv2 license,
the scheduling part remains proprietary, with a LICENSE file [2] stating:
"This is not Open Source software."
The researcher mentioned a plan to address the licensing issue, upstream
the patches, and step up as a maintainer, but there has been no further
communication since then.
Maintaining DCCP for a decade without any real users has become a burden.
Therefore, it's time to remove it.
Removing DCCP will also provide significant benefits to TCP. It allows
us to freely reorganize the layout of struct inet_connection_sock, which
is currently shared with DCCP, and optimize it to reduce the number of
cachelines accessed in the TCP fast path.
Note that we keep DCCP netfilter modules as requested. [3]
Link: https://lore.kernel.org/netdev/20230710182253.81446-1-kuniyu@amazon.com/T/#u #[0]
Link: https://github.com/telekom/mp-dccp #[1]
Link: https://github.com/telekom/mp-dccp/blob/mpdccp_v03_k5.10/net/dccp/non_gpl_scheduler/LICENSE #[2]
Link: https://lore.kernel.org/netdev/Z_VQ0KlCRkqYWXa-@calendula/ #[3]
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Paul Moore <paul@paul-moore.com> (LSM and SELinux)
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Link: https://patch.msgid.link/20250410023921.11307-3-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When TCP is in TIME_WAIT state, PAWS verification uses
LINUX_PAWSESTABREJECTED, which is ambiguous and cannot be distinguished
from other PAWS verification processes.
We added a new counter, like the existing PAWS_OLD_ACK one.
Also we update the doc with previously missing PAWS_OLD_ACK.
usage:
'''
nstat -az | grep PAWSTimewait
TcpExtPAWSTimewait 1 0.0
'''
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250409112614.16153-3-jiayuan.chen@linux.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Devices in the networking path, such as firewalls, NATs, or routers, which
can perform SNAT or DNAT, use addresses from their own limited address
pools to masquerade the source address during forwarding, causing PAWS
verification to fail more easily.
Currently, packet loss statistics for PAWS can only be viewed through MIB,
which is a global metric and cannot be precisely obtained through tracing
to get the specific 4-tuple of the dropped packet. In the past, we had to
use kprobe ret to retrieve relevant skb information from
tcp_timewait_state_process().
We add a drop_reason pointer, similar to what previous commit does:
commit e34100c2ec ("tcp: add a drop_reason pointer to tcp_check_req()")
This commit addresses the PAWSESTABREJECTED case and also sets the
corresponding drop reason.
We use 'pwru' to test.
Before this commit:
''''
./pwru 'port 9999'
2025/04/07 13:40:19 Listening for events..
TUPLE FUNC
172.31.75.115:12345->172.31.75.114:9999(tcp) sk_skb_reason_drop(SKB_DROP_REASON_NOT_SPECIFIED)
'''
After this commit:
'''
./pwru 'port 9999'
2025/04/07 13:51:34 Listening for events..
TUPLE FUNC
172.31.75.115:12345->172.31.75.114:9999(tcp) sk_skb_reason_drop(SKB_DROP_REASON_TCP_RFC7323_TW_PAWS)
'''
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250409112614.16153-2-jiayuan.chen@linux.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull networking fixes from Paolo Abeni:
"Including fixes from netfilter.
Current release - regressions:
- core: hold instance lock during NETDEV_CHANGE
- rtnetlink: fix bad unlock balance in do_setlink()
- ipv6:
- fix null-ptr-deref in addrconf_add_ifaddr()
- align behavior across nexthops during path selection
Previous releases - regressions:
- sctp: prevent transport UaF in sendmsg
- mptcp: only inc MPJoinAckHMacFailure for HMAC failures
Previous releases - always broken:
- sched:
- make ->qlen_notify() idempotent
- ensure sufficient space when sending filter netlink notifications
- sch_sfq: really don't allow 1 packet limit
- netfilter: fix incorrect avx2 match of 5th field octet
- tls: explicitly disallow disconnect
- eth: octeontx2-pf: fix VF root node parent queue priority"
* tag 'net-6.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (38 commits)
ethtool: cmis_cdb: Fix incorrect read / write length extension
selftests: netfilter: add test case for recent mismatch bug
nft_set_pipapo: fix incorrect avx2 match of 5th field octet
net: ppp: Add bound checking for skb data on ppp_sync_txmung
net: Fix null-ptr-deref by sock_lock_init_class_and_name() and rmmod.
ipv6: Align behavior across nexthops during path selection
net: phy: allow MDIO bus PM ops to start/stop state machine for phylink-controlled PHY
net: phy: move phy_link_change() prior to mdio_bus_phy_may_suspend()
selftests/tc-testing: sfq: check that a derived limit of 1 is rejected
net_sched: sch_sfq: move the limit validation
net_sched: sch_sfq: use a temporary work area for validating configuration
net: libwx: handle page_pool_dev_alloc_pages error
selftests: mptcp: validate MPJoin HMacFailure counters
mptcp: only inc MPJoinAckHMacFailure for HMAC failures
rtnetlink: Fix bad unlock balance in do_setlink().
net: ethtool: Don't call .cleanup_data when prepare_data fails
tc: Ensure we have enough buffer space when sending filter netlink notifications
net: libwx: Fix the wrong Rx descriptor field
octeontx2-pf: qos: fix VF root node parent queue index
selftests: tls: check that disconnect does nothing
...
When I ran the repro [0] and waited a few seconds, I observed two
LOCKDEP splats: a warning immediately followed by a null-ptr-deref. [1]
Reproduction Steps:
1) Mount CIFS
2) Add an iptables rule to drop incoming FIN packets for CIFS
3) Unmount CIFS
4) Unload the CIFS module
5) Remove the iptables rule
At step 3), the CIFS module calls sock_release() for the underlying
TCP socket, and it returns quickly. However, the socket remains in
FIN_WAIT_1 because incoming FIN packets are dropped.
At this point, the module's refcnt is 0 while the socket is still
alive, so the following rmmod command succeeds.
# ss -tan
State Recv-Q Send-Q Local Address:Port Peer Address:Port
FIN-WAIT-1 0 477 10.0.2.15:51062 10.0.0.137:445
# lsmod | grep cifs
cifs 1159168 0
This highlights a discrepancy between the lifetime of the CIFS module
and the underlying TCP socket. Even after CIFS calls sock_release()
and it returns, the TCP socket does not die immediately in order to
close the connection gracefully.
While this is generally fine, it causes an issue with LOCKDEP because
CIFS assigns a different lock class to the TCP socket's sk->sk_lock
using sock_lock_init_class_and_name().
Once an incoming packet is processed for the socket or a timer fires,
sk->sk_lock is acquired.
Then, LOCKDEP checks the lock context in check_wait_context(), where
hlock_class() is called to retrieve the lock class. However, since
the module has already been unloaded, hlock_class() logs a warning
and returns NULL, triggering the null-ptr-deref.
If LOCKDEP is enabled, we must ensure that a module calling
sock_lock_init_class_and_name() (CIFS, NFS, etc) cannot be unloaded
while such a socket is still alive to prevent this issue.
Let's hold the module reference in sock_lock_init_class_and_name()
and release it when the socket is freed in sk_prot_free().
Note that sock_lock_init() clears sk->sk_owner for svc_create_socket()
that calls sock_lock_init_class_and_name() for a listening socket,
which clones a socket by sk_clone_lock() without GFP_ZERO.
[0]:
CIFS_SERVER="10.0.0.137"
CIFS_PATH="//${CIFS_SERVER}/Users/Administrator/Desktop/CIFS_TEST"
DEV="enp0s3"
CRED="/root/WindowsCredential.txt"
MNT=$(mktemp -d /tmp/XXXXXX)
mount -t cifs ${CIFS_PATH} ${MNT} -o vers=3.0,credentials=${CRED},cache=none,echo_interval=1
iptables -A INPUT -s ${CIFS_SERVER} -j DROP
for i in $(seq 10);
do
umount ${MNT}
rmmod cifs
sleep 1
done
rm -r ${MNT}
iptables -D INPUT -s ${CIFS_SERVER} -j DROP
[1]:
DEBUG_LOCKS_WARN_ON(1)
WARNING: CPU: 10 PID: 0 at kernel/locking/lockdep.c:234 hlock_class (kernel/locking/lockdep.c:234 kernel/locking/lockdep.c:223)
Modules linked in: cifs_arc4 nls_ucs2_utils cifs_md4 [last unloaded: cifs]
CPU: 10 UID: 0 PID: 0 Comm: swapper/10 Not tainted 6.14.0 #36
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
RIP: 0010:hlock_class (kernel/locking/lockdep.c:234 kernel/locking/lockdep.c:223)
...
Call Trace:
<IRQ>
__lock_acquire (kernel/locking/lockdep.c:4853 kernel/locking/lockdep.c:5178)
lock_acquire (kernel/locking/lockdep.c:469 kernel/locking/lockdep.c:5853 kernel/locking/lockdep.c:5816)
_raw_spin_lock_nested (kernel/locking/spinlock.c:379)
tcp_v4_rcv (./include/linux/skbuff.h:1678 ./include/net/tcp.h:2547 net/ipv4/tcp_ipv4.c:2350)
...
BUG: kernel NULL pointer dereference, address: 00000000000000c4
PF: supervisor read access in kernel mode
PF: error_code(0x0000) - not-present page
PGD 0
Oops: Oops: 0000 [#1] PREEMPT SMP NOPTI
CPU: 10 UID: 0 PID: 0 Comm: swapper/10 Tainted: G W 6.14.0 #36
Tainted: [W]=WARN
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
RIP: 0010:__lock_acquire (kernel/locking/lockdep.c:4852 kernel/locking/lockdep.c:5178)
Code: 15 41 09 c7 41 8b 44 24 20 25 ff 1f 00 00 41 09 c7 8b 84 24 a0 00 00 00 45 89 7c 24 20 41 89 44 24 24 e8 e1 bc ff ff 4c 89 e7 <44> 0f b6 b8 c4 00 00 00 e8 d1 bc ff ff 0f b6 80 c5 00 00 00 88 44
RSP: 0018:ffa0000000468a10 EFLAGS: 00010046
RAX: 0000000000000000 RBX: ff1100010091cc38 RCX: 0000000000000027
RDX: ff1100081f09ca48 RSI: 0000000000000001 RDI: ff1100010091cc88
RBP: ff1100010091c200 R08: ff1100083fe6e228 R09: 00000000ffffbfff
R10: ff1100081eca0000 R11: ff1100083fe10dc0 R12: ff1100010091cc88
R13: 0000000000000001 R14: 0000000000000000 R15: 00000000000424b1
FS: 0000000000000000(0000) GS:ff1100081f080000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000000000c4 CR3: 0000000002c4a003 CR4: 0000000000771ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
<IRQ>
lock_acquire (kernel/locking/lockdep.c:469 kernel/locking/lockdep.c:5853 kernel/locking/lockdep.c:5816)
_raw_spin_lock_nested (kernel/locking/spinlock.c:379)
tcp_v4_rcv (./include/linux/skbuff.h:1678 ./include/net/tcp.h:2547 net/ipv4/tcp_ipv4.c:2350)
ip_protocol_deliver_rcu (net/ipv4/ip_input.c:205 (discriminator 1))
ip_local_deliver_finish (./include/linux/rcupdate.h:878 net/ipv4/ip_input.c:234)
ip_sublist_rcv_finish (net/ipv4/ip_input.c:576)
ip_list_rcv_finish (net/ipv4/ip_input.c:628)
ip_list_rcv (net/ipv4/ip_input.c:670)
__netif_receive_skb_list_core (net/core/dev.c:5939 net/core/dev.c:5986)
netif_receive_skb_list_internal (net/core/dev.c:6040 net/core/dev.c:6129)
napi_complete_done (./include/linux/list.h:37 ./include/net/gro.h:519 ./include/net/gro.h:514 net/core/dev.c:6496)
e1000_clean (drivers/net/ethernet/intel/e1000/e1000_main.c:3815)
__napi_poll.constprop.0 (net/core/dev.c:7191)
net_rx_action (net/core/dev.c:7262 net/core/dev.c:7382)
handle_softirqs (kernel/softirq.c:561)
__irq_exit_rcu (kernel/softirq.c:596 kernel/softirq.c:435 kernel/softirq.c:662)
irq_exit_rcu (kernel/softirq.c:680)
common_interrupt (arch/x86/kernel/irq.c:280 (discriminator 14))
</IRQ>
<TASK>
asm_common_interrupt (./arch/x86/include/asm/idtentry.h:693)
RIP: 0010:default_idle (./arch/x86/include/asm/irqflags.h:37 ./arch/x86/include/asm/irqflags.h:92 arch/x86/kernel/process.c:744)
Code: 4c 01 c7 4c 29 c2 e9 72 ff ff ff 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 0f 1e fa eb 07 0f 00 2d c3 2b 15 00 fb f4 <fa> c3 cc cc cc cc 66 66 2e 0f 1f 84 00 00 00 00 00 90 90 90 90 90
RSP: 0018:ffa00000000ffee8 EFLAGS: 00000202
RAX: 000000000000640b RBX: ff1100010091c200 RCX: 0000000000061aa4
RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffff812f30c5
RBP: 000000000000000a R08: 0000000000000001 R09: 0000000000000000
R10: 0000000000000001 R11: 0000000000000002 R12: 0000000000000000
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
? do_idle (kernel/sched/idle.c:186 kernel/sched/idle.c:325)
default_idle_call (./include/linux/cpuidle.h:143 kernel/sched/idle.c:118)
do_idle (kernel/sched/idle.c:186 kernel/sched/idle.c:325)
cpu_startup_entry (kernel/sched/idle.c:422 (discriminator 1))
start_secondary (arch/x86/kernel/smpboot.c:315)
common_startup_64 (arch/x86/kernel/head_64.S:421)
</TASK>
Modules linked in: cifs_arc4 nls_ucs2_utils cifs_md4 [last unloaded: cifs]
CR2: 00000000000000c4
Fixes: ed07536ed6 ("[PATCH] lockdep: annotate nfs/nfsd in-kernel sockets")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20250407163313.22682-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We mostly needed rtnl_lock in qstat to make sure the queue count
is stable while we work. For "ops locked" drivers the instance
lock protects the queue count, so we don't have to take rtnl_lock.
For currently ops-locked drivers: netdevsim and bnxt need
the protection from netdev going down while we dump, which
instance lock provides. gve doesn't care.
Reviewed-by: Joe Damato <jdamato@fastly.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250408195956.412733-9-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Protect xdp_features with netdev->lock. This way pure readers
no longer have to take rtnl_lock to access the field.
This includes calling NETDEV_XDP_FEAT_CHANGE under the lock.
Looks like that's fine for bonding, the only "real" listener,
it's the same as ethtool feature change.
In terms of normal drivers - only GVE need special consideration
(other drivers don't use instance lock or don't support XDP).
It calls xdp_set_features_flag() helper from gve_init_priv() which
in turn is called from gve_reset_recovery() (locked), or prior
to netdev registration. So switch to _locked.
Reviewed-by: Joe Damato <jdamato@fastly.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Acked-by: Harshitha Ramamurthy <hramamurthy@google.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20250408195956.412733-6-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add helpers to "lock a netdev in a backward-compatible way",
which for ops-locked netdevs will mean take the instance lock.
For drivers which haven't opted into the ops locking we'll take
rtnl_lock.
The scoped foreach is dropping and re-taking the lock for each
device, even if prev and next are both under rtnl_lock.
I hope that's fine since we expect that netdev nl to be mostly
supported by modern drivers, and modern drivers should also
opt into the instance locking.
Note that these helpers are mostly needed for queue related state,
because drivers modify queue config in their ops in a non-atomic
way. Or differently put, queue changes don't have a clear-cut API
like NAPI configuration. Any state that can should just use the
instance lock directly, not the "compat" hacks.
Reviewed-by: Joe Damato <jdamato@fastly.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250408195956.412733-4-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Read accesses go via xsk_get_pool_from_qid(), the call coming
from the core and gve look safe (other "ops locked" drivers
don't support XSK).
Write accesses go via xsk_reg_pool_at_qid() and xsk_clear_pool_at_qid().
Former is already under the ops lock, latter is not (both coming from
the workqueue via xp_clear_dev() and NETDEV_UNREGISTER via xsk_notifier()).
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250408195956.412733-3-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
It's quite common to have a single UDP tunnel type active in the
whole system. In such a case we can replace the indirect call for
the UDP tunnel GRO callback with a static call.
Add the related accounting in the control path and switch to static
call when possible. To keep the code simple use a static array for
the registered tunnel types, and size such array based on the kernel
config.
Note that there are valid kernel configurations leading to
UDP_MAX_TUNNEL_TYPES == 0 even with IS_ENABLED(CONFIG_NET_UDP_TUNNEL),
Explicitly skip the accounting in such a case, to avoid compile warning
when accessing "udp_tunnel_gro_types".
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/53d156cdfddcc9678449e873cc83e68fa1582653.1744040675.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Most UDP tunnels bind a socket to a local port, with ANY address, no
peer and no interface index specified.
Additionally it's quite common to have a single tunnel device per
namespace.
Track in each namespace the UDP tunnel socket respecting the above.
When only a single one is present, store a reference in the netns.
When such reference is not NULL, UDP tunnel GRO lookup just need to
match the incoming packet destination port vs the socket local port.
The tunnel socket never sets the reuse[port] flag[s]. When bound to no
address and interface, no other socket can exist in the same netns
matching the specified local port.
Matching packets with non-local destination addresses will be
aggregated, and eventually segmented as needed - no behavior changes
intended.
Restrict the optimization to kernel sockets only: it covers all the
relevant use-cases, and user-space owned sockets could be disconnected
and rebound after setup_udp_tunnel_sock(), breaking the uniqueness
assumption
Note that the UDP tunnel socket reference is stored into struct
netns_ipv4 for both IPv4 and IPv6 tunnels. That is intentional to keep
all the fastpath-related netns fields in the same struct and allow
cacheline-based optimization. Currently both the IPv4 and IPv6 socket
pointer share the same cacheline as the `udp_table` field.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/41d16bc8d1257d567f9344c445b4ae0b4a91ede4.1744040675.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
sctp_sendmsg() re-uses associations and transports when possible by
doing a lookup based on the socket endpoint and the message destination
address, and then sctp_sendmsg_to_asoc() sets the selected transport in
all the message chunks to be sent.
There's a possible race condition if another thread triggers the removal
of that selected transport, for instance, by explicitly unbinding an
address with setsockopt(SCTP_SOCKOPT_BINDX_REM), after the chunks have
been set up and before the message is sent. This can happen if the send
buffer is full, during the period when the sender thread temporarily
releases the socket lock in sctp_wait_for_sndbuf().
This causes the access to the transport data in
sctp_outq_select_transport(), when the association outqueue is flushed,
to result in a use-after-free read.
This change avoids this scenario by having sctp_transport_free() signal
the freeing of the transport, tagging it as "dead". In order to do this,
the patch restores the "dead" bit in struct sctp_transport, which was
removed in
commit 47faa1e4c5 ("sctp: remove the dead field of sctp_transport").
Then, in the scenario where the sender thread has released the socket
lock in sctp_wait_for_sndbuf(), the bit is checked again after
re-acquiring the socket lock to detect the deletion. This is done while
holding a reference to the transport to prevent it from being freed in
the process.
If the transport was deleted while the socket lock was relinquished,
sctp_sendmsg_to_asoc() will return -EAGAIN to let userspace retry the
send.
The bug was found by a private syzbot instance (see the error report [1]
and the C reproducer that triggers it [2]).
Link: https://people.igalia.com/rcn/kernel_logs/20250402__KASAN_slab-use-after-free_Read_in_sctp_outq_select_transport.txt [1]
Link: https://people.igalia.com/rcn/kernel_logs/20250402__KASAN_slab-use-after-free_Read_in_sctp_outq_select_transport__repro.c [2]
Cc: stable@vger.kernel.org
Fixes: df132eff46 ("sctp: clear the transport of some out_chunk_list chunks in sctp_assoc_rm_peer")
Suggested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Ricardo Cañuelo Navarro <rcn@igalia.com>
Acked-by: Xin Long <lucien.xin@gmail.com>
Link: https://patch.msgid.link/20250404-kasan_slab-use-after-free_read_in_sctp_outq_select_transport__20250404-v1-1-5ce4a0b78ef2@igalia.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
timer_delete[_sync]() replaces del_timer[_sync](). Convert the whole tree
over and remove the historical wrapper inlines.
Conversion was done with coccinelle plus manual fixups where necessary.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
devmem code performs a number of safety checks to avoid having
to reimplement all of them in the drivers. Move those to
__net_mp_open_rxq() and reuse that function for binding to make
sure that io_uring ZC also benefits from them.
While at it rename the queue ID variable to rxq_idx in
__net_mp_open_rxq(), we touch most of the relevant lines.
The XArray insertion is reordered after the netdev_rx_queue_restart()
call, otherwise we'd need to duplicate the queue index check
or risk inserting an invalid pointer. The XArray allocation
failures should be extremely rare.
Reviewed-by: Mina Almasry <almasrymina@google.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Fixes: 6e18ed929d ("net: add helpers for setting a memory provider on an rx queue")
Link: https://patch.msgid.link/20250403013405.2827250-2-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
