If number of TX queues are set to 1 we get a NULL pointer
dereference during XDP_TX.

~# ethtool -L eth0 tx 1
~# ./xdp-trafficgen udp -A <ipv6-src> -a <ipv6-dst> eth0 -t 2
Transmitting on eth0 (ifindex 2)
[  241.135257] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000030

Fix this by using actual TX queues instead of max TX queues
when picking the TX channel in am65_cpsw_ndo_xdp_xmit().

Fixes: 8acacc40 ("net: ethernet: ti: am65-cpsw: Add minimal XDP support")
Signed-off-by: Roger Quadros <rogerq@kernel.org>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Julien Panis <jpanis@baylibre.com>
Reviewed-by: MD Danish Anwar <danishanwar@ti.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

The following XDP_DROP test from [1] stalls the interface after
250 packets.
~# xdb-bench drop -m native eth0
This is because new RX requests are never queued. Fix that.

The below XDP_TX test from [1] fails with a warning
[  499.947381] XDP_WARN: xdp_update_frame_from_buff(line:277): Driver BUG: missing reserved tailroom
~# xdb-bench tx -m native eth0
Fix that by using PAGE_SIZE during xdp_init_buf().

In XDP_REDIRECT case only 1 packet was processed in rx_poll.
Fix it to process up to budget packets.

Fix all XDP error cases to call trace_xdp_exception() and drop the packet
in am65_cpsw_run_xdp().

[1] xdp-tools suite https://github.com/xdp-project/xdp-tools

Fixes: 8acacc40 ("net: ethernet: ti: am65-cpsw: Add minimal XDP support")
Signed-off-by: Roger Quadros <rogerq@kernel.org>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Julien Panis <jpanis@baylibre.com>
Reviewed-by: MD Danish Anwar <danishanwar@ti.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Luiz Augusto von Dentz says:

====================
bluetooth pull request for net:

 - qca: If memdump doesn't work, re-enable IBS
 - MGMT: Fix not generating command complete for MGMT_OP_DISCONNECT
 - Revert "Bluetooth: MGMT/SMP: Fix address type when using SMP over BREDR/LE"
 - MGMT: Ignore keys being loaded with invalid type

* tag 'for-net-2024-08-30' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth:
  Bluetooth: MGMT: Ignore keys being loaded with invalid type
  Revert "Bluetooth: MGMT/SMP: Fix address type when using SMP over BREDR/LE"
  Bluetooth: MGMT: Fix not generating command complete for MGMT_OP_DISCONNECT
  Bluetooth: hci_sync: Introduce hci_cmd_sync_run/hci_cmd_sync_run_once
  Bluetooth: qca: If memdump doesn't work, re-enable IBS
====================

Link: https://patch.msgid.link/20240830220300.1316772-1-luiz.dentz@gmail.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Merge tag 'linux-can-fixes-for-6.11-20240830' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
pull-request: can 2024-08-30

The first patch is by Kuniyuki Iwashima for the CAN BCM protocol that
adds a missing proc entry removal when a device unregistered.

Simon Horman fixes the cleanup in the error cleanup path of the m_can
driver's open function.

Markus Schneider-Pargmann contributes 7 fixes for the m_can driver,
all related to the recently added IRQ coalescing support.

The next 2 patches are by me, target the mcp251xfd driver and fix ring
and coalescing configuration problems when switching from CAN-CC to
CAN-FD mode.

Simon Arlott's patch fixes a possible deadlock in the mcp251x driver.

The last patch is by Martin Jocic for the kvaser_pciefd driver and
fixes a problem with lost IRQs, which result in starvation, under high
load situations.

* tag 'linux-can-fixes-for-6.11-20240830' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can:
  can: kvaser_pciefd: Use a single write when releasing RX buffers
  can: mcp251x: fix deadlock if an interrupt occurs during mcp251x_open
  can: mcp251xfd: mcp251xfd_ring_init(): check TX-coalescing configuration
  can: mcp251xfd: fix ring configuration when switching from CAN-CC to CAN-FD mode
  can: m_can: Limit coalescing to peripheral instances
  can: m_can: Reset cached active_interrupts on start
  can: m_can: disable_all_interrupts, not clear active_interrupts
  can: m_can: Do not cancel timer from within timer
  can: m_can: Remove m_can_rx_peripheral indirection
  can: m_can: Remove coalesing disable in isr during suspend
  can: m_can: Reset coalescing during suspend/resume
  can: m_can: Release irq on error in m_can_open
  can: bcm: Remove proc entry when dev is unregistered.
====================

Link: https://patch.msgid.link/20240830215914.1610393-1-mkl@pengutronix.deSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Call rtnl_unlock() on this error path, before returning.

Fixes: bc23aa94 ("igc: Add pcie error handler support")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Gerhard Engleder <gerhard@engleder-embedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This is a clear use-after-free error. We remove it, and rely on checking
the return code of vcap_del_rule.
Reported-by: Dan Carpenter <error27@gmail.com>
Closes: https://lore.kernel.org/kernel-janitors/7bffefc6-219a-4f71-baa0-ad4526e5c198@kili.mountain/
Fixes: c956b9b3 ("net: microchip: sparx5: Adding KUNIT tests of key/action values in VCAP API")
Signed-off-by: Jens Emil Schulz Østergaard <jensemil.schulzostergaard@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Matt Johnston says:

====================
net: mctp-serial: Fix for missing tx escapes

The mctp-serial code to add escape characters was incorrect due to an
off-by-one error. This series adds a test for the chunking which splits
by escape characters, and fixes the bug.

v2: Fix kunit param const pointer
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

0x7d and 0x7e bytes are meant to be escaped in the data portion of
frames, but this didn't occur since next_chunk_len() had an off-by-one
error. That also resulted in the final byte of a payload being written
as a separate tty write op.

The chunk prior to an escaped byte would be one byte short, and the
next call would never test the txpos+1 case, which is where the escaped
byte was located. That meant it never hit the escaping case in
mctp_serial_tx_work().

Example Input: 01 00 08 c8 7e 80 02

Previous incorrect chunks from next_chunk_len():

01 00 08
c8 7e 80
02

With this fix:

01 00 08 c8
7e
80 02

Cc: stable@vger.kernel.org
Fixes: a0c2ccd9 ("mctp: Add MCTP-over-serial transport binding")
Signed-off-by: Matt Johnston <matt@codeconstruct.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Test various edge cases of inputs that contain characters
that need escaping.

This adds a new kunit suite for mctp-serial.
Signed-off-by: Matt Johnston <matt@codeconstruct.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Due to 59b047bc there could be keys stored
with the wrong address type so this attempt to detect it and ignore them
instead of just failing to load all keys.

Cc: stable@vger.kernel.org
Link: https://github.com/bluez/bluez/issues/875
Fixes: 59b047bc ("Bluetooth: MGMT/SMP: Fix address type when using SMP over BREDR/LE")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>

This reverts commit 59b047bc which
breaks compatibility with commands like:

bluetoothd[46328]: @ MGMT Command: Load.. (0x0013) plen 74  {0x0001} [hci0]
        Keys: 2
        BR/EDR Address: C0:DC:DA:A5:E5:47 (Samsung Electronics Co.,Ltd)
        Key type: Authenticated key from P-256 (0x03)
        Central: 0x00
        Encryption size: 16
        Diversifier[2]: 0000
        Randomizer[8]: 0000000000000000
        Key[16]: 6ed96089bd9765be2f2c971b0b95f624
        LE Address: D7:2A:DE:1E:73:A2 (Static)
        Key type: Unauthenticated key from P-256 (0x02)
        Central: 0x00
        Encryption size: 16
        Diversifier[2]: 0000
        Randomizer[8]: 0000000000000000
        Key[16]: 87dd2546ededda380ffcdc0a8faa4597
@ MGMT Event: Command Status (0x0002) plen 3                {0x0001} [hci0]
      Load Long Term Keys (0x0013)
        Status: Invalid Parameters (0x0d)

Cc: stable@vger.kernel.org
Link: https://github.com/bluez/bluez/issues/875
Fixes: 59b047bc ("Bluetooth: MGMT/SMP: Fix address type when using SMP over BREDR/LE")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>

MGMT_OP_DISCONNECT can be called while mgmt_device_connected has not
been called yet, which will cause the connection procedure to be
aborted, so mgmt_device_disconnected shall still respond with command
complete to MGMT_OP_DISCONNECT and just not emit
MGMT_EV_DEVICE_DISCONNECTED since MGMT_EV_DEVICE_CONNECTED was never
sent.

To fix this MGMT_OP_DISCONNECT is changed to work similarly to other
command which do use hci_cmd_sync_queue and then use hci_conn_abort to
disconnect and returns the result, in order for hci_conn_abort to be
used from hci_cmd_sync context it now uses hci_cmd_sync_run_once.

Link: https://github.com/bluez/bluez/issues/932
Fixes: 12d4a3b2 ("Bluetooth: Move check for MGMT_CONNECTED flag into mgmt.c")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>

This introduces hci_cmd_sync_run/hci_cmd_sync_run_once which acts like
hci_cmd_sync_queue/hci_cmd_sync_queue_once but runs immediately when
already on hdev->cmd_sync_work context.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>

On systems in the field, we are seeing this sometimes in the kernel logs:
  Bluetooth: qca_controller_memdump() hci0: hci_devcd_init Return:-95

This means that _something_ decided that it wanted to get a memdump
but then hci_devcd_init() returned -EOPNOTSUPP (AKA -95).

The cleanup code in qca_controller_memdump() when we get back an error
from hci_devcd_init() undoes most things but forgets to clear
QCA_IBS_DISABLED. One side effect of this is that, during the next
suspend, qca_suspend() will always get a timeout.

Let's fix it so that we clear the bit.

Fixes: 06d3fdfc ("Bluetooth: hci_qca: Add qcom devcoredump support")
Reviewed-by: Guenter Roeck <groeck@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>

Kvaser's PCIe cards uses the KCAN FPGA IP block which has dual 4K
buffers for incoming messages shared by all (currently up to eight)
channels. While the driver processes messages in one buffer, new
incoming messages are stored in the other and so on.

The design of KCAN is such that a buffer must be fully read and then
released. Releasing a buffer will make the FPGA switch buffers. If the
other buffer contains at least one incoming message the FPGA will also
instantly issue a new interrupt, if not the interrupt will be issued
after receiving the first new message.

With IRQx interrupts, it takes a little time for the interrupt to
happen, enough for any previous ISR call to do it's business and
return, but MSI interrupts are way faster so this time is reduced to
almost nothing.

So with MSI, releasing the buffer HAS to be the very last action of
the ISR before returning, otherwise the new interrupt might be
"masked" by the kernel because the previous ISR call hasn't returned.
And the interrupts are edge-triggered so we cannot loose one, or the
ping-pong reading process will stop.

This is why this patch modifies the driver to use a single write to
the SRB_CMD register before returning.
Signed-off-by: Martin Jocic <martin.jocic@kvaser.com>
Reviewed-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://patch.msgid.link/20240830153113.2081440-1-martin.jocic@kvaser.com
Fixes: 26ad340e ("can: kvaser_pciefd: Add driver for Kvaser PCIEcan devices")
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

When we cork messages in psock->cork, the last message triggers the
flushing will result in sending a sk_msg larger than the current
message size. In this case, in tcp_bpf_send_verdict(), 'copied' becomes
negative at least in the following case:

468         case __SK_DROP:
469         default:
470                 sk_msg_free_partial(sk, msg, tosend);
471                 sk_msg_apply_bytes(psock, tosend);
472                 *copied -= (tosend + delta); // <==== HERE
473                 return -EACCES;

Therefore, it could lead to the following BUG with a proper value of
'copied' (thanks to syzbot). We should not use negative 'copied' as a
return value here.

  ------------[ cut here ]------------
  kernel BUG at net/socket.c:733!
  Internal error: Oops - BUG: 00000000f2000800 [#1] PREEMPT SMP
  Modules linked in:
  CPU: 0 UID: 0 PID: 3265 Comm: syz-executor510 Not tainted 6.11.0-rc3-syzkaller-00060-gd07b4328 #0
  Hardware name: linux,dummy-virt (DT)
  pstate: 61400009 (nZCv daif +PAN -UAO -TCO +DIT -SSBS BTYPE=--)
  pc : sock_sendmsg_nosec net/socket.c:733 [inline]
  pc : sock_sendmsg_nosec net/socket.c:728 [inline]
  pc : __sock_sendmsg+0x5c/0x60 net/socket.c:745
  lr : sock_sendmsg_nosec net/socket.c:730 [inline]
  lr : __sock_sendmsg+0x54/0x60 net/socket.c:745
  sp : ffff800088ea3b30
  x29: ffff800088ea3b30 x28: fbf00000062bc900 x27: 0000000000000000
  x26: ffff800088ea3bc0 x25: ffff800088ea3bc0 x24: 0000000000000000
  x23: f9f00000048dc000 x22: 0000000000000000 x21: ffff800088ea3d90
  x20: f9f00000048dc000 x19: ffff800088ea3d90 x18: 0000000000000001
  x17: 0000000000000000 x16: 0000000000000000 x15: 000000002002ffaf
  x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
  x11: 0000000000000000 x10: ffff8000815849c0 x9 : ffff8000815b49c0
  x8 : 0000000000000000 x7 : 000000000000003f x6 : 0000000000000000
  x5 : 00000000000007e0 x4 : fff07ffffd239000 x3 : fbf00000062bc900
  x2 : 0000000000000000 x1 : 0000000000000000 x0 : 00000000fffffdef
  Call trace:
   sock_sendmsg_nosec net/socket.c:733 [inline]
   __sock_sendmsg+0x5c/0x60 net/socket.c:745
   ____sys_sendmsg+0x274/0x2ac net/socket.c:2597
   ___sys_sendmsg+0xac/0x100 net/socket.c:2651
   __sys_sendmsg+0x84/0xe0 net/socket.c:2680
   __do_sys_sendmsg net/socket.c:2689 [inline]
   __se_sys_sendmsg net/socket.c:2687 [inline]
   __arm64_sys_sendmsg+0x24/0x30 net/socket.c:2687
   __invoke_syscall arch/arm64/kernel/syscall.c:35 [inline]
   invoke_syscall+0x48/0x110 arch/arm64/kernel/syscall.c:49
   el0_svc_common.constprop.0+0x40/0xe0 arch/arm64/kernel/syscall.c:132
   do_el0_svc+0x1c/0x28 arch/arm64/kernel/syscall.c:151
   el0_svc+0x34/0xec arch/arm64/kernel/entry-common.c:712
   el0t_64_sync_handler+0x100/0x12c arch/arm64/kernel/entry-common.c:730
   el0t_64_sync+0x19c/0x1a0 arch/arm64/kernel/entry.S:598
  Code: f9404463 d63f0060 3108441f 54fffe81 (d4210000)
  ---[ end trace 0000000000000000 ]---

Fixes: 4f738adb ("bpf: create tcp_bpf_ulp allowing BPF to monitor socket TX/RX data")
Reported-by: syzbot+58c03971700330ce14d8@syzkaller.appspotmail.com
Cc: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20240821030744.320934-1-xiyou.wangcong@gmail.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Since smc_inet6_prot does not initialize ipv6_pinfo_offset, inet6_create()
copies an incorrect address value, sk + 0 (offset), to inet_sk(sk)->pinet6.

In addition, since inet_sk(sk)->pinet6 and smc_sk(sk)->clcsock practically
point to the same address, when smc_create_clcsk() stores the newly
created clcsock in smc_sk(sk)->clcsock, inet_sk(sk)->pinet6 is corrupted
into clcsock. This causes NULL pointer dereference and various other
memory corruptions.

To solve this problem, you need to initialize ipv6_pinfo_offset, add a
smc6_sock structure, and then add ipv6_pinfo as the second member of
the smc_sock structure.
Reported-by: syzkaller <syzkaller@googlegroups.com>
Fixes: d25a92cc ("net/smc: Introduce IPPROTO_SMC")
Signed-off-by: Jeongjun Park <aha310510@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2024-08-28 (igb, ice)

This series contains updates to igb and ice drivers.

Daiwei Li restores writing the TSICR (TimeSync Interrupt Cause)
register on 82850 devices to workaround a hardware issue for igb.

Dawid detaches netdev device for reset to avoid ethtool accesses during
reset causing NULL pointer dereferences on ice.

* '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  ice: Add netif_device_attach/detach into PF reset flow
  igb: Fix not clearing TimeSync interrupts for 82580
====================

Link: https://patch.msgid.link/20240828225444.645154-1-anthony.l.nguyen@intel.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

We exclude wireless drivers from the netdev@ traffic, to delegate
it to linux-wireless@, and avoid overwhelming netdev@.
Bluetooth drivers are implicitly excluded because they live under
drivers/bluetooth, not drivers/net.

In both cases DT bindings sit under Documentation/devicetree/bindings/net/
and aren't excluded. So if a patch series touches DT bindings
netdev@ ends up getting CCed, and these are usually fairly boring
series.
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20240828175821.2960423-1-kuba@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Pull networking fixes from Paolo Abeni:
 "Including fixes from bluetooth, wireless and netfilter.

  No known outstanding regressions.

  Current release - regressions:

   - wifi: iwlwifi: fix hibernation

   - eth: ionic: prevent tx_timeout due to frequent doorbell ringing

  Previous releases - regressions:

   - sched: fix sch_fq incorrect behavior for small weights

   - wifi:
      - iwlwifi: take the mutex before running link selection
      - wfx: repair open network AP mode

   - netfilter: restore IP sanity checks for netdev/egress

   - tcp: fix forever orphan socket caused by tcp_abort

   - mptcp: close subflow when receiving TCP+FIN

   - bluetooth: fix random crash seen while removing btnxpuart driver

  Previous releases - always broken:

   - mptcp: more fixes for the in-kernel PM

   - eth: bonding: change ipsec_lock from spin lock to mutex

   - eth: mana: fix race of mana_hwc_post_rx_wqe and new hwc response

  Misc:

   - documentation: drop special comment style for net code"

* tag 'net-6.11-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (57 commits)
  nfc: pn533: Add poll mod list filling check
  mailmap: update entry for Sriram Yagnaraman
  selftests: mptcp: join: check re-re-adding ID 0 signal
  mptcp: pm: ADD_ADDR 0 is not a new address
  selftests: mptcp: join: validate event numbers
  mptcp: avoid duplicated SUB_CLOSED events
  selftests: mptcp: join: check re-re-adding ID 0 endp
  mptcp: pm: fix ID 0 endp usage after multiple re-creations
  mptcp: pm: do not remove already closed subflows
  selftests: mptcp: join: no extra msg if no counter
  selftests: mptcp: join: check re-adding init endp with != id
  mptcp: pm: reset MPC endp ID when re-added
  mptcp: pm: skip connecting to already established sf
  mptcp: pm: send ACK on an active subflow
  selftests: mptcp: join: check removing ID 0 endpoint
  mptcp: pm: fix RM_ADDR ID for the initial subflow
  mptcp: pm: reuse ID 0 after delete and re-add
  net: busy-poll: use ktime_get_ns() instead of local_clock()
  sctp: fix association labeling in the duplicate COOKIE-ECHO case
  mptcp: pr_debug: add missing \n at the end
  ...

In case of im_protocols value is 1 and tm_protocols value is 0 this
combination successfully passes the check
'if (!im_protocols && !tm_protocols)' in the nfc_start_poll().
But then after pn533_poll_create_mod_list() call in pn533_start_poll()
poll mod list will remain empty and dev->poll_mod_count will remain 0
which lead to division by zero.

Normally no im protocol has value 1 in the mask, so this combination is
not expected by driver. But these protocol values actually come from
userspace via Netlink interface (NFC_CMD_START_POLL operation). So a
broken or malicious program may pass a message containing a "bad"
combination of protocol parameter values so that dev->poll_mod_count
is not incremented inside pn533_poll_create_mod_list(), thus leading
to division by zero.
Call trace looks like:
nfc_genl_start_poll()
  nfc_start_poll()
    ->start_poll()
    pn533_start_poll()

Add poll mod list filling check.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: dfccd0f5 ("NFC: pn533: Add some polling entropy")
Signed-off-by: Aleksandr Mishin <amishin@t-argos.ru>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://patch.msgid.link/20240827084822.18785-1-amishin@t-argos.ruSigned-off-by: Paolo Abeni <pabeni@redhat.com>

Pablo Neira Ayuso says:

====================
Netfilter fixes for net

The following patchset contains Netfilter fixes for net:

Patch #1 sets on NFT_PKTINFO_L4PROTO for UDP packets less than 4 bytes
payload from netdev/egress by subtracting skb_network_offset() when
validating IPv4 packet length, otherwise 'meta l4proto udp' never
matches.

Patch #2 subtracts skb_network_offset() when validating IPv6 packet
length for netdev/egress.

netfilter pull request 24-08-28

* tag 'nf-24-08-28' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  netfilter: nf_tables_ipv6: consider network offset in netdev/egress validation
  netfilter: nf_tables: restore IP sanity checks for netdev/egress
====================

Link: https://patch.msgid.link/20240828214708.619261-1-pablo@netfilter.orgSigned-off-by: Paolo Abeni <pabeni@redhat.com>

Link my old est.tech address to my active mail address
Signed-off-by: Sriram Yagnaraman <sriram.yagnaraman@ericsson.com>
Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de>
Link: https://patch.msgid.link/20240828072417.4111996-1-sriram.yagnaraman@ericsson.comSigned-off-by: Paolo Abeni <pabeni@redhat.com>

Matthieu Baerts says:

====================
mptcp: more fixes for the in-kernel PM

Here is a new batch of fixes for the MPTCP in-kernel path-manager:

Patch 1 ensures the address ID is set to 0 when the path-manager sends
an ADD_ADDR for the address of the initial subflow. The same fix is
applied when a new subflow is created re-using this special address. A
fix for v6.0.

Patch 2 is similar, but for the case where an endpoint is removed: if
this endpoint was used for the initial address, it is important to send
a RM_ADDR with this ID set to 0, and look for existing subflows with the
ID set to 0. A fix for v6.0 as well.

Patch 3 validates the two previous patches.

Patch 4 makes the PM selecting an "active" path to send an address
notification in an ACK, instead of taking the first path in the list. A
fix for v5.11.

Patch 5 fixes skipping the establishment of a new subflow if a previous
subflow using the same pair of addresses is being closed. A fix for
v5.13.

Patch 6 resets the ID linked to the initial subflow when the linked
endpoint is re-added, possibly with a different ID. A fix for v6.0.

Patch 7 validates the three previous patches.

Patch 8 is a small fix for the MPTCP Join selftest, when being used with
older subflows not supporting all MIB counters. A fix for a commit
introduced in v6.4, but backported up to v5.10.

Patch 9 avoids the PM to try to close the initial subflow multiple
times, and increment counters while nothing happened. A fix for v5.10.

Patch 10 stops incrementing local_addr_used and add_addr_accepted
counters when dealing with the address ID 0, because these counters are
not taking into account the initial subflow, and are then not
decremented when the linked addresses are removed. A fix for v6.0.

Patch 11 validates the previous patch.

Patch 12 avoids the PM to send multiple SUB_CLOSED events for the
initial subflow. A fix for v5.12.

Patch 13 validates the previous patch.

Patch 14 stops treating the ADD_ADDR 0 as a new address, and accepts it
in order to re-create the initial subflow if it has been closed, even if
the limit for *new* addresses -- not taking into account the address of
the initial subflow -- has been reached. A fix for v5.10.

Patch 15 validates the previous patch.
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
---
Matthieu Baerts (NGI0) (15):
      mptcp: pm: reuse ID 0 after delete and re-add
      mptcp: pm: fix RM_ADDR ID for the initial subflow
      selftests: mptcp: join: check removing ID 0 endpoint
      mptcp: pm: send ACK on an active subflow
      mptcp: pm: skip connecting to already established sf
      mptcp: pm: reset MPC endp ID when re-added
      selftests: mptcp: join: check re-adding init endp with != id
      selftests: mptcp: join: no extra msg if no counter
      mptcp: pm: do not remove already closed subflows
      mptcp: pm: fix ID 0 endp usage after multiple re-creations
      selftests: mptcp: join: check re-re-adding ID 0 endp
      mptcp: avoid duplicated SUB_CLOSED events
      selftests: mptcp: join: validate event numbers
      mptcp: pm: ADD_ADDR 0 is not a new address
      selftests: mptcp: join: check re-re-adding ID 0 signal

 net/mptcp/pm.c                                  |   4 +-
 net/mptcp/pm_netlink.c                          |  87 ++++++++++----
 net/mptcp/protocol.c                            |   6 +
 net/mptcp/protocol.h                            |   5 +-
 tools/testing/selftests/net/mptcp/mptcp_join.sh | 153 ++++++++++++++++++++----
 tools/testing/selftests/net/mptcp/mptcp_lib.sh  |   4 +
 6 files changed, 209 insertions(+), 50 deletions(-)
---
base-commit: 3a0504d5
change-id: 20240826-net-mptcp-more-pm-fix-ffa61a36f817

Best regards,
====================

Link: https://patch.msgid.link/20240828-net-mptcp-more-pm-fix-v2-0-7f11b283fff7@kernel.orgSigned-off-by: Paolo Abeni <pabeni@redhat.com>

This test extends "delete re-add signal" to validate the previous
commit: when the 'signal' endpoint linked to the initial subflow (ID 0)
is re-added multiple times, it will re-send the ADD_ADDR with id 0. The
client should still be able to re-create this subflow, even if the
add_addr_accepted limit has been reached as this special address is not
considered as a new address.

The 'Fixes' tag here below is the same as the one from the previous
commit: this patch here is not fixing anything wrong in the selftests,
but it validates the previous fix for an issue introduced by this commit
ID.

Fixes: d0876b22 ("mptcp: add the incoming RM_ADDR support")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

The ADD_ADDR 0 with the address from the initial subflow should not be
considered as a new address: this is not something new. If the host
receives it, it simply means that the address is available again.

When receiving an ADD_ADDR for the ID 0, the PM already doesn't consider
it as new by not incrementing the 'add_addr_accepted' counter. But the
'accept_addr' might not be set if the limit has already been reached:
this can be bypassed in this case. But before, it is important to check
that this ADD_ADDR for the ID 0 is for the same address as the initial
subflow. If not, it is not something that should happen, and the
ADD_ADDR can be ignored.

Note that if an ADD_ADDR is received while there is already a subflow
opened using the same address, this ADD_ADDR is ignored as well. It
means that if multiple ADD_ADDR for ID 0 are received, there will not be
any duplicated subflows created by the client.

Fixes: d0876b22 ("mptcp: add the incoming RM_ADDR support")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

This test extends "delete and re-add" and "delete re-add signal" to
validate the previous commit: the number of MPTCP events are checked to
make sure there are no duplicated or unexpected ones.

A new helper has been introduced to easily check these events. The
missing events have been added to the lib.

The 'Fixes' tag here below is the same as the one from the previous
commit: this patch here is not fixing anything wrong in the selftests,
but it validates the previous fix for an issue introduced by this commit
ID.

Fixes: b911c97c ("mptcp: add netlink event support")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

The initial subflow might have already been closed, but still in the
connection list. When the worker is instructed to close the subflows
that have been marked as closed, it might then try to close the initial
subflow again.

 A consequence of that is that the SUB_CLOSED event can be seen twice:

  # ip mptcp endpoint
  1.1.1.1 id 1 subflow dev eth0
  2.2.2.2 id 2 subflow dev eth1

  # ip mptcp monitor &
  [         CREATED] remid=0 locid=0 saddr4=1.1.1.1 daddr4=9.9.9.9
  [     ESTABLISHED] remid=0 locid=0 saddr4=1.1.1.1 daddr4=9.9.9.9
  [  SF_ESTABLISHED] remid=0 locid=2 saddr4=2.2.2.2 daddr4=9.9.9.9

  # ip mptcp endpoint delete id 1
  [       SF_CLOSED] remid=0 locid=0 saddr4=1.1.1.1 daddr4=9.9.9.9
  [       SF_CLOSED] remid=0 locid=0 saddr4=1.1.1.1 daddr4=9.9.9.9

The first one is coming from mptcp_pm_nl_rm_subflow_received(), and the
second one from __mptcp_close_subflow().

To avoid doing the post-closed processing twice, the subflow is now
marked as closed the first time.

Note that it is not enough to check if we are dealing with the first
subflow and check its sk_state: the subflow might have been reset or
closed before calling mptcp_close_ssk().

Fixes: b911c97c ("mptcp: add netlink event support")
Cc: stable@vger.kernel.org
Tested-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

This test extends "delete and re-add" to validate the previous commit:
when the endpoint linked to the initial subflow (ID 0) is re-added
multiple times, it was no longer being used, because the internal linked
counters are not decremented for this special endpoint: it is not an
additional endpoint.

Here, the "del/add id 0" steps are done 3 times to unsure this case is
validated.

The 'Fixes' tag here below is the same as the one from the previous
commit: this patch here is not fixing anything wrong in the selftests,
but it validates the previous fix for an issue introduced by this commit
ID.

Fixes: 3ad14f54 ("mptcp: more accurate MPC endpoint tracking")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

'local_addr_used' and 'add_addr_accepted' are decremented for addresses
not related to the initial subflow (ID0), because the source and
destination addresses of the initial subflows are known from the
beginning: they don't count as "additional local address being used" or
"ADD_ADDR being accepted".

It is then required not to increment them when the entrypoint used by
the initial subflow is removed and re-added during a connection. Without
this modification, this entrypoint cannot be removed and re-added more
than once.
Reported-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/512
Fixes: 3ad14f54 ("mptcp: more accurate MPC endpoint tracking")
Reported-by: syzbot+455d38ecd5f655fc45cf@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/00000000000049861306209237f4@google.com
Cc: stable@vger.kernel.org
Tested-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

It is possible to have in the list already closed subflows, e.g. the
initial subflow has been already closed, but still in the list. No need
to try to close it again, and increments the related counters again.

Fixes: 0ee4261a ("mptcp: implement mptcp_pm_remove_subflow")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

The checksum and fail counters might not be available. Then no need to
display an extra message with missing info.

While at it, fix the indentation around, which is wrong since the same
commit.

Fixes: 47867f0a ("selftests: mptcp: join: skip check if MIB counter not supported")
Cc: stable@vger.kernel.org
Reviewed-by: Geliang Tang <geliang@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

The initial subflow has a special local ID: 0. It is specific per
connection.

When a global endpoint is deleted and re-added later, it can have a
different ID, but the kernel should still use the ID 0 if it corresponds
to the initial address.

This test validates this behaviour: the endpoint linked to the initial
subflow is removed, and re-added with a different ID.

Note that removing the initial subflow will not decrement the 'subflows'
counters, which corresponds to the *additional* subflows. On the other
hand, when the same endpoint is re-added, it will increment this
counter, as it will be seen as an additional subflow this time.

The 'Fixes' tag here below is the same as the one from the previous
commit: this patch here is not fixing anything wrong in the selftests,
but it validates the previous fix for an issue introduced by this commit
ID.

Fixes: 3ad14f54 ("mptcp: more accurate MPC endpoint tracking")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

The initial subflow has a special local ID: 0. It is specific per
connection.

When a global endpoint is deleted and re-added later, it can have a
different ID -- most services managing the endpoints automatically don't
force the ID to be the same as before. It is then important to track
these modifications to be consistent with the ID being used for the
address used by the initial subflow, not to confuse the other peer or to
send the ID 0 for the wrong address.

Now when removing an endpoint, msk->mpc_endpoint_id is reset if it
corresponds to this endpoint. When adding a new endpoint, the same
variable is updated if the address match the one of the initial subflow.

Fixes: 3ad14f54 ("mptcp: more accurate MPC endpoint tracking")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

The lookup_subflow_by_daddr() helper checks if there is already a
subflow connected to this address. But there could be a subflow that is
closing, but taking time due to some reasons: latency, losses, data to
process, etc.

If an ADD_ADDR is received while the endpoint is being closed, it is
better to try connecting to it, instead of rejecting it: the peer which
has sent the ADD_ADDR will not be notified that the ADD_ADDR has been
rejected for this reason, and the expected subflow will not be created
at the end.

This helper should then only look for subflows that are established, or
going to be, but not the ones being closed.

Fixes: d84ad049 ("mptcp: skip connecting the connected address")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Taking the first one on the list doesn't work in some cases, e.g. if the
initial subflow is being removed. Pick another one instead of not
sending anything.

Fixes: 84dfe367 ("mptcp: send out dedicated ADD_ADDR packet")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Removing the endpoint linked to the initial subflow should trigger a
RM_ADDR for the right ID, and the removal of the subflow. That's what is
now being verified in the "delete and re-add" test.

Note that removing the initial subflow will not decrement the 'subflows'
counters, which corresponds to the *additional* subflows. On the other
hand, when the same endpoint is re-added, it will increment this
counter, as it will be seen as an additional subflow this time.

The 'Fixes' tag here below is the same as the one from the previous
commit: this patch here is not fixing anything wrong in the selftests,
but it validates the previous fix for an issue introduced by this commit
ID.

Fixes: 3ad14f54 ("mptcp: more accurate MPC endpoint tracking")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

The initial subflow has a special local ID: 0. When an endpoint is being
deleted, it is then important to check if its address is not linked to
the initial subflow to send the right ID.

If there was an endpoint linked to the initial subflow, msk's
mpc_endpoint_id field will be set. We can then use this info when an
endpoint is being removed to see if it is linked to the initial subflow.

So now, the correct IDs are passed to mptcp_pm_nl_rm_addr_or_subflow(),
it is no longer needed to use mptcp_local_id_match().

Fixes: 3ad14f54 ("mptcp: more accurate MPC endpoint tracking")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

When the endpoint used by the initial subflow is removed and re-added
later, the PM has to force the ID 0, it is a special case imposed by the
MPTCP specs.

Note that the endpoint should then need to be re-added reusing the same
ID.

Fixes: 3ad14f54 ("mptcp: more accurate MPC endpoint tracking")
Cc: stable@vger.kernel.org
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Pull random number generator fix from Jason Donenfeld:
 "Reject invalid flags passed to vgetrandom() in the same way that
  getrandom() does, so that the behavior is the same, from Yann.

  The flags argument to getrandom() only has a behavioral effect on the
  function if the RNG isn't initialized yet, so vgetrandom() falls back
  to the syscall in that case. But if the RNG is initialized, all of the
  flags behave the same way, so vgetrandom() didn't bother checking
  them, and just ignored them entirely.

  But that doesn't account for invalid flags passed in, which need to be
  rejected so we can use them later"

* tag 'random-6.11-rc6-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random:
  random: vDSO: reject unknown getrandom() flags