selftests: mptcp: declare var as local

Just to avoid classical Bash pitfall where variables are accidentally
overridden by other functions because the proper scope has not been
defined.

That's also what is done in other MPTCP selftests scripts where all non
local variables are defined at the beginning of the script and the
others are defined with the "local" keyword.

Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

selftests: mptcp: clearly declare global ns vars

It is clearer to declare these global variables at the beginning of the
file as it is done in other MPTCP selftests rather than in functions in
the middle of the script.

So for uniformity reason, we can do the same here in mptcp_sockopt.sh.

Suggested-by: Geliang Tang <geliang.tang@suse.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

selftests: mptcp: uniform 'rndh' variable

The definition of 'rndh' was probably copied from one script to another
but some times, 'sec' was not defined, not used and/or not spelled
properly.

Here all the 'rndh' are now defined the same way.

Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

selftests: mptcp: removed defined but unused vars

Some variables were set but never used.

This was not causing any issues except adding some confusion and having
shellcheck complaining about them.

Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

selftests: mptcp: run mptcp_inq from a clean netns

A new "sandbox" net namespace is available where no other netfilter
rules have been added.

Use this new netns instead of re-using "ns1" and clean it.

Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt: report FEC block stats via standard interface

I must have missed that these stats are only exposed
via the unstructured ethtool -S when they got merged.
Plumb in the structured form.

Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20221130013108.90062-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'remove-label-cpu-from-dsa-dt-binding'

Arınç ÜNAL says:

====================
remove label = "cpu" from DSA dt-binding

With this patch series, we're completely getting rid of 'label = "cpu";'
which is not used by the DSA dt-binding at all.

Information for taking the patches for maintainers:
Patch 1: netdev maintainers (based off netdev/net-next.git main)
Patch 2-3: SoC maintainers (based off soc/soc.git soc/dt)
Patch 4: MIPS maintainers (based off mips/linux.git mips-next)
Patch 5: PowerPC maintainers (based off powerpc/linux.git next-test)

I've been meaning to submit this for a few months. Find the relevant
conversation here:
https://lore.kernel.org/netdev/20220913155408.GA3802998-robh@kernel.org/

Here's how I did it, for the interested (or suggestions):

Find the platforms which have got 'label = "cpu";' defined.
grep -rnw . -e 'label = "cpu";'

Remove the line where 'label = "cpu";' is included.
sed -i /'label = "cpu";'/,+d arch/arm/boot/dts/*
sed -i /'label = "cpu";'/,+d arch/arm64/boot/dts/freescale/*
sed -i /'label = "cpu";'/,+d arch/arm64/boot/dts/marvell/*
sed -i /'label = "cpu";'/,+d arch/arm64/boot/dts/mediatek/*
sed -i /'label = "cpu";'/,+d arch/arm64/boot/dts/rockchip/*
sed -i /'label = "cpu";'/,+d arch/mips/boot/dts/qca/*
sed -i /'label = "cpu";'/,+d arch/mips/boot/dts/ralink/*
sed -i /'label = "cpu";'/,+d arch/powerpc/boot/dts/turris1x.dts
sed -i /'label = "cpu";'/,+d Documentation/devicetree/bindings/net/qca,ar71xx.yaml

Restore the symlink files which typechange after running sed.
====================

Link: https://lore.kernel.org/r/20221130141040.32447-1-arinc.unal@arinc9.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: net: qca,ar71xx: remove label = "cpu" from examples

This is not used by the DSA dt-binding, so remove it from the examples.

Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Reviewed-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-tcp-dynamically-disable-tcp-md5-static-key'

Dmitry Safonov says:

====================
net/tcp: Dynamically disable TCP-MD5 static key

The static key introduced by commit 6015c71e656b ("tcp: md5: add
tcp_md5_needed jump label") is a fast-path optimization aimed at
avoiding a cache line miss.
Once an MD5 key is introduced in the system the static key is enabled
and never disabled. Address this by disabling the static key when
the last tcp_md5sig_info in system is destroyed.

Previously it was submitted as a part of TCP-AO patches set [1].
Now in attempt to split 36 patches submission, I send this independently.

Version 5:
https://lore.kernel.org/all/20221122185534.308643-1-dima@arista.com/T/#u
Version 4:
https://lore.kernel.org/all/20221115211905.1685426-1-dima@arista.com/T/#u
Version 3:
https://lore.kernel.org/all/20221111212320.1386566-1-dima@arista.com/T/#u
Version 2:
https://lore.kernel.org/all/20221103212524.865762-1-dima@arista.com/T/#u
Version 1:
https://lore.kernel.org/all/20221102211350.625011-1-dima@arista.com/T/#u

[1]: https://lore.kernel.org/all/20221027204347.529913-1-dima@arista.com/T/#u
====================

Link: https://lore.kernel.org/r/20221123173859.473629-1-dima@arista.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/tcp: Separate initialization of twsk

Convert BUG_ON() to WARN_ON_ONCE() and warn as well for unlikely
static key int overflow error-path.

Signed-off-by: Dmitry Safonov <dima@arista.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/tcp: Do cleanup on tcp_md5_key_copy() failure

If the kernel was short on (atomic) memory and failed to allocate it -
don't proceed to creation of request socket. Otherwise the socket would
be unsigned and userspace likely doesn't expect that the TCP is not
MD5-signed anymore.

Signed-off-by: Dmitry Safonov <dima@arista.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/tcp: Disable TCP-MD5 static key on tcp_md5sig_info destruction

To do that, separate two scenarios:
- where it's the first MD5 key on the system, which means that enabling
  of the static key may need to sleep;
- copying of an existing key from a listening socket to the request
  socket upon receiving a signed TCP segment, where static key was
  already enabled (when the key was added to the listening socket).

Now the life-time of the static branch for TCP-MD5 is until:
- last tcp_md5sig_info is destroyed
- last socket in time-wait state with MD5 key is closed.

Which means that after all sockets with TCP-MD5 keys are gone, the
system gets back the performance of disabled md5-key static branch.

While at here, provide static_key_fast_inc() helper that does ref
counter increment in atomic fashion (without grabbing cpus_read_lock()
on CONFIG_JUMP_LABEL=y). This is needed to add a new user for
a static_key when the caller controls the lifetime of another user.

Signed-off-by: Dmitry Safonov <dima@arista.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/tcp: Separate tcp_md5sig_info allocation into tcp_md5sig_info_add()

Add a helper to allocate tcp_md5sig_info, that will help later to
do/allocate things when info allocated, once per socket.

Signed-off-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

jump_label: Prevent key->enabled int overflow

1. With CONFIG_JUMP_LABEL=n static_key_slow_inc() doesn't have any
   protection against key->enabled refcounter overflow.
2. With CONFIG_JUMP_LABEL=y static_key_slow_inc_cpuslocked()
   still may turn the refcounter negative as (v + 1) may overflow.

key->enabled is indeed a ref-counter as it's documented in multiple
places: top comment in jump_label.h, Documentation/staging/static-keys.rst,
etc.

As -1 is reserved for static key that's in process of being enabled,
functions would break with negative key->enabled refcount:
- for CONFIG_JUMP_LABEL=n negative return of static_key_count()
  breaks static_key_false(), static_key_true()
- the ref counter may become 0 from negative side by too many
  static_key_slow_inc() calls and lead to use-after-free issues.

These flaws result in that some users have to introduce an additional
mutex and prevent the reference counter from overflowing themselves,
see bpf_enable_runtime_stats() checking the counter against INT_MAX / 2.

Prevent the reference counter overflow by checking if (v + 1) > 0.
Change functions API to return whether the increment was successful.

Signed-off-by: Dmitry Safonov <dima@arista.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'locking/core' of git://git./linux/kernel/git/tip/tip

Pull in locking/core from tip (just a single patch) to avoid a conflict
with a jump_label change needed by a TCP cleanup.

Link: https://lore.kernel.org/all/Y4B17nBArWS1Iywo@hirez.programming.kicks-ass.net/
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

r8169: use tp_to_dev instead of open code

The open code is defined as a helper function(tp_to_dev) on r8169_main.c,
which the open code is &tp->pci_dev->dev. The helper function was added
in commit 1e1205b7d3e9 ("r8169: add helper tp_to_dev"). And then later,
commit f1e911d5d0df ("r8169: add basic phylib support") added
r8169_phylink_handler function but it didn't use the helper function.
Thus, tp_to_dev() replaces the open code. This patch doesn't change logic.

Signed-off-by: Juhee Kang <claudiajkang@gmail.com>
Reviewed-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/20221129161244.5356-1-claudiajkang@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'fix-rtnl_mutex-deadlock-with-dpaa2-and-sfp-modules'

Vladimir Oltean says:

====================
Fix rtnl_mutex deadlock with DPAA2 and SFP modules

This patch set deliberately targets net-next and lacks Fixes: tags due
to caution on my part.

While testing some SFP modules on the Solidrun Honeycomb LX2K platform,
I noticed that rebooting causes a deadlock:

============================================
WARNING: possible recursive locking detected
6.1.0-rc5-07010-ga9b9500ffaac-dirty #656 Not tainted
--------------------------------------------
systemd-shutdow/1 is trying to acquire lock:
ffffa62db6cf42f0 (rtnl_mutex){+.+.}-{4:4}, at: rtnl_lock+0x1c/0x30

but task is already holding lock:
ffffa62db6cf42f0 (rtnl_mutex){+.+.}-{4:4}, at: rtnl_lock+0x1c/0x30

other info that might help us debug this:
 Possible unsafe locking scenario:

       CPU0
       ----
  lock(rtnl_mutex);
  lock(rtnl_mutex);

 *** DEADLOCK ***

 May be due to missing lock nesting notation

6 locks held by systemd-shutdow/1:
 #0: ffffa62db6863c70 (system_transition_mutex){+.+.}-{4:4}, at: __do_sys_reboot+0xd4/0x260
 #1: ffff2f2b0176f100 (&dev->mutex){....}-{4:4}, at: device_shutdown+0xf4/0x260
 #2: ffff2f2b017be900 (&dev->mutex){....}-{4:4}, at: device_shutdown+0x104/0x260
 #3: ffff2f2b017680f0 (&dev->mutex){....}-{4:4}, at: device_release_driver_internal+0x40/0x260
 #4: ffff2f2b0e1608f0 (&dev->mutex){....}-{4:4}, at: device_release_driver_internal+0x40/0x260
 #5: ffffa62db6cf42f0 (rtnl_mutex){+.+.}-{4:4}, at: rtnl_lock+0x1c/0x30

stack backtrace:
CPU: 6 PID: 1 Comm: systemd-shutdow Not tainted 6.1.0-rc5-07010-ga9b9500ffaac-dirty #656
Hardware name: SolidRun LX2160A Honeycomb (DT)
Call trace:
 lock_acquire+0x68/0x84
 __mutex_lock+0x98/0x460
 mutex_lock_nested+0x2c/0x40
 rtnl_lock+0x1c/0x30
 sfp_bus_del_upstream+0x1c/0xac
 phylink_destroy+0x1c/0x50
 dpaa2_mac_disconnect+0x28/0x70
 dpaa2_eth_remove+0x1dc/0x1f0
 fsl_mc_driver_remove+0x24/0x60
 device_remove+0x70/0x80
 device_release_driver_internal+0x1f0/0x260
 device_links_unbind_consumers+0xe0/0x110
 device_release_driver_internal+0x138/0x260
 device_release_driver+0x18/0x24
 bus_remove_device+0x12c/0x13c
 device_del+0x16c/0x424
 fsl_mc_device_remove+0x28/0x40
 __fsl_mc_device_remove+0x10/0x20
 device_for_each_child+0x5c/0xac
 dprc_remove+0x94/0xb4
 fsl_mc_driver_remove+0x24/0x60
 device_remove+0x70/0x80
 device_release_driver_internal+0x1f0/0x260
 device_release_driver+0x18/0x24
 bus_remove_device+0x12c/0x13c
 device_del+0x16c/0x424
 fsl_mc_bus_remove+0x8c/0x10c
 fsl_mc_bus_shutdown+0x10/0x20
 platform_shutdown+0x24/0x3c
 device_shutdown+0x15c/0x260
 kernel_restart+0x40/0xa4
 __do_sys_reboot+0x1e4/0x260
 __arm64_sys_reboot+0x24/0x30

But fixing this appears to be not so simple. The patch set represents my
attempt to address it.

In short, the problem is that dpaa2_mac_connect() and dpaa2_mac_disconnect()
call 2 phylink functions in a row, one takes rtnl_lock() itself -
phylink_create(), and one which requires rtnl_lock() to be held by the
caller - phylink_fwnode_phy_connect(). The existing approach in the
drivers is too simple. We take rtnl_lock() when calling dpaa2_mac_connect(),
which is what results in the deadlock.

Fixing just that creates another problem. The drivers make use of
rtnl_lock() for serializing with other code paths too. I think I've
found all those code paths, and established other mechanisms for
serializing with them.
====================

Link: https://lore.kernel.org/r/20221129141221.872653-1-vladimir.oltean@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-mac: move rtnl_lock() only around phylink_{,dis}connect_phy()

After the introduction of a private mac_lock that serializes access to
priv->mac (and port_priv->mac in the switch), the only remaining purpose
of rtnl_lock() is to satisfy the locking requirements of
phylink_fwnode_phy_connect() and phylink_disconnect_phy().

But the functions these live in, dpaa2_mac_connect() and
dpaa2_mac_disconnect(), have contradictory locking requirements.
While phylink_fwnode_phy_connect() wants rtnl_lock() to be held,
phylink_create() wants it to not be held.

Move the rtnl_lock() from top-level (in the dpaa2-eth and dpaa2-switch
drivers) to only surround the phylink calls that require it, in the
dpaa2-mac library code.

This is possible because dpaa2_mac_connect() and dpaa2_mac_disconnect()
run unlocked, and there isn't any danger of an AB/BA deadlock between
the rtnl_mutex and other private locks.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-switch: serialize changes to priv->mac with a mutex

The dpaa2-switch driver uses a DPMAC in the same way as the dpaa2-eth
driver, so we need to duplicate the locking solution established by the
previous change to the switch driver as well.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-eth: serialize changes to priv->mac with a mutex

The dpaa2 architecture permits dynamic connections between objects on
the fsl-mc bus, specifically between a DPNI object (represented by a
struct net_device) and a DPMAC object (represented by a struct phylink).

The DPNI driver is notified when those connections are created/broken
through the dpni_irq0_handler_thread() method. To ensure that ethtool
operations, as well as netdev up/down operations serialize with the
connection/disconnection of the DPNI with a DPMAC,
dpni_irq0_handler_thread() takes the rtnl_lock() to block those other
operations from taking place.

There is code called by dpaa2_mac_connect() which wants to acquire the
rtnl_mutex once again, see phylink_create() -> phylink_register_sfp() ->
sfp_bus_add_upstream() -> rtnl_lock(). So the strategy doesn't quite
work out, even though it's fairly simple.

Create a different strategy, where all code paths in the dpaa2-eth
driver access priv->mac only while they are holding priv->mac_lock.
The phylink instance is not created or connected to the PHY under the
priv->mac_lock, but only assigned to priv->mac then. This will eliminate
the reliance on the rtnl_mutex.

Add lockdep annotations and put comments where holding the lock is not
necessary, and priv->mac can be dereferenced freely.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-eth: connect to MAC before requesting the "endpoint changed" IRQ

dpaa2_eth_connect_mac() is called both from dpaa2_eth_probe() and from
dpni_irq0_handler_thread().

It could happen that the DPNI gets connected to a DPMAC on the fsl-mc
bus exactly during probe, as soon as the "endpoint change" interrupt is
requested in dpaa2_eth_setup_irqs(). This will cause the
dpni_irq0_handler_thread() to register a phylink instance for that DPMAC.

Then, the probing function will also try to register a phylink instance
for the same DPMAC, operation which should fail (and this will fail the
probing of the driver).

Reorder dpaa2_eth_setup_irqs() and dpaa2_eth_connect_mac(), such that
dpni_irq0_handler_thread() never races with the DPMAC-related portion of
the probing path.

Also reorder dpaa2_eth_disconnect_mac() to be in the mirror position of
dpaa2_eth_connect_mac() in the teardown path.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-switch replace direct MAC access with dpaa2_switch_port_has_mac()

The helper function will gain a lockdep annotation in a future patch.
Make sure to benefit from it.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2: publish MAC stringset to ethtool -S even if MAC is missing

DPNIs and DPSW objects can connect and disconnect at runtime from DPMAC
objects on the same fsl-mc bus. The DPMAC object also holds "ethtool -S"
unstructured counters. Those counters are only shown for the entity
owning the netdev (DPNI, DPSW) if it's connected to a DPMAC.

The ethtool stringset code path is split into multiple callbacks, but
currently, connecting and disconnecting the DPMAC takes the rtnl_lock().
This blocks the entire ethtool code path from running, see
ethnl_default_doit() -> rtnl_lock() -> ops->prepare_data() ->
strset_prepare_data().

This is going to be a problem if we are going to no longer require
rtnl_lock() when connecting/disconnecting the DPMAC, because the DPMAC
could appear between ops->get_sset_count() and ops->get_strings().
If it appears out of the blue, we will provide a stringset into an array
that was dimensioned thinking the DPMAC wouldn't be there => array
accessed out of bounds.

There isn't really a good way to work around that, and I don't want to
put too much pressure on the ethtool framework by playing locking games.
Just make the DPMAC counters be always available. They'll be zeroes if
the DPNI or DPSW isn't connected to a DPMAC.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-switch: assign port_priv->mac after dpaa2_mac_connect() call

The dpaa2-switch has the exact same locking requirements when connected
to a DPMAC, so it needs port_priv->mac to always point either to NULL,
or to a DPMAC with a fully initialized phylink instance.

Make the same preparatory change in the dpaa2-switch driver as in the
dpaa2-eth one.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-eth: assign priv->mac after dpaa2_mac_connect() call

There are 2 requirements for correct code:

- Any time the driver accesses the priv->mac pointer at runtime, it
  either holds NULL to indicate a DPNI-DPNI connection (or unconnected
  DPNI), or a struct dpaa2_mac whose phylink instance was fully
  initialized (created and connected to the PHY). No changes are made to
  priv->mac while it is being used. Currently, rtnl_lock() watches over
  the call to dpaa2_eth_connect_mac(), so it serves the purpose of
  serializing this with all readers of priv->mac.

- dpaa2_mac_connect() should run unlocked, because inside it are 2
  phylink calls with incompatible locking requirements: phylink_create()
  requires that the rtnl_mutex isn't held, and phylink_fwnode_phy_connect()
  requires that the rtnl_mutex is held. The only way to solve those
  contradictory requirements is to let dpaa2_mac_connect() take
  rtnl_lock() when it needs to.

To solve both requirements, we need to identify the writer side of the
priv->mac pointer, which can be wrapped in a mutex private to the driver
in a future patch. The dpaa2_mac_connect() cannot be part of the writer
side critical section, because of an AB/BA deadlock with rtnl_lock().

So the strategy needs to be that where we prepare the DPMAC by calling
dpaa2_mac_connect(), and only make priv->mac point to it once it's fully
prepared. This ensures that the writer side critical section has the
absolute minimum surface it can.

The reverse strategy is adopted in the dpaa2_eth_disconnect_mac() code
path. This makes sure that priv->mac is NULL when we start tearing down
the DPMAC that we disconnected from, and concurrent code will simply not
see it.

No locking changes in this patch (concurrent code is still blocked by
the rtnl_mutex).

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-mac: remove defensive check in dpaa2_mac_disconnect()

dpaa2_mac_disconnect() will only be called with a NULL mac->phylink if
dpaa2_mac_connect() failed, or was never called.

The callers are these:

dpaa2_eth_disconnect_mac():

if (dpaa2_eth_is_type_phy(priv))
dpaa2_mac_disconnect(priv->mac);

dpaa2_switch_port_disconnect_mac():

if (dpaa2_switch_port_is_type_phy(port_priv))
dpaa2_mac_disconnect(port_priv->mac);

priv->mac can be NULL, but in that case, dpaa2_eth_is_type_phy() returns
false, and dpaa2_mac_disconnect() is never called. Similar for
dpaa2-switch.

When priv->mac is non-NULL, it means that dpaa2_mac_connect() returned
zero (success), and therefore, priv->mac->phylink is also a valid
pointer.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-mac: absorb phylink_start() call into dpaa2_mac_start()

The phylink handling is intended to be hidden inside the dpaa2_mac
object. Move the phylink_start() call into dpaa2_mac_start(), and
phylink_stop() into dpaa2_mac_stop().

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2: replace dpaa2_mac_is_type_fixed() with dpaa2_mac_is_type_phy()

dpaa2_mac_is_type_fixed() is a header with no implementation and no
callers, which is referenced from the documentation though. It can be
deleted.

On the other hand, it would be useful to reuse the code between
dpaa2_eth_is_type_phy() and dpaa2_switch_port_is_type_phy(). That common
code should be called dpaa2_mac_is_type_phy(), so let's create that.

The removal and the addition are merged into the same patch because,
in fact, is_type_phy() is the logical opposite of is_type_fixed().

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dpaa2-eth: don't use -ENOTSUPP error code

dpaa2_eth_setup_dpni() is called from the probe path and
dpaa2_eth_set_link_ksettings() is propagated to user space.

include/linux/errno.h says that ENOTSUPP is "Defined for the NFSv3
protocol". Conventional wisdom has it to not use it in networking
drivers. Replace it with -EOPNOTSUPP.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: microchip: sparx5: Fix error handling in vcap_show_admin()

If vcap_dup_rule() fails that leads to an error pointer dereference
side the call to vcap_free_rule().  Also it only returns an error if the
very last call to vcap_read_rule() fails and it returns success for
other errors.

I've changed it to just stop printing after the first error and return
an error code.

Fixes: 3a7921560d2f ("net: microchip: sparx5: Add VCAP rule debugFS support for the VCAP API")
Signed-off-by: Dan Carpenter <error27@gmail.com>
Reviewed-by: Steen Hegelund <Steen.Hegelund@microchip.com>
Link: https://lore.kernel.org/r/Y4XUUx9kzurBN+BV@kili
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

bonding: uninitialized variable in bond_miimon_inspect()

The "ignore_updelay" variable needs to be initialized to false.

Fixes: f8a65ab2f3ff ("bonding: fix link recovery in mode 2 when updelay is nonzero")
Signed-off-by: Dan Carpenter <error27@gmail.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Link: https://lore.kernel.org/r/Y4SWJlh3ohJ6EPTL@kili
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge tag 'mlx5-updates-2022-11-29' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2022-11-29

Misc update for mlx5 driver

1) Various trivial cleanups

2) Maor Dickman, Adds support for trap offload with additional actions

3) From Tariq, UMR (device memory registrations) cleanups,
   UMR WQE must be aligned to 64B per device spec, (not a bug fix).

* tag 'mlx5-updates-2022-11-29' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5e: Support devlink reload of IPsec core
  net/mlx5e: TC, Add offload support for trap with additional actions
  net/mlx5e: Do early return when setup vports dests for slow path flow
  net/mlx5: Remove redundant check
  net/mlx5e: Delete always true DMA check
  net/mlx5e: Don't access directly DMA device pointer
  net/mlx5e: Don't use termination table when redundant
  net/mlx5: Fix orthography errors in documentation
  net/mlx5: Use generic definition for UMR KLM alignment
  net/mlx5: Generalize name of UMR alignment definition
  net/mlx5: Remove unused UMR MTT definitions
  net/mlx5e: Add padding when needed in UMR WQEs
  net/mlx5: Remove unused ctx variables
  net/mlx5e: Replace zero-length arrays with DECLARE_FLEX_ARRAY() helper
  net/mlx5e: Remove unneeded io-mapping.h #include
====================

Link: https://lore.kernel.org/r/20221130051152.479480-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: phy: Add link between phy dev and mac dev

If the external phy used by current mac interface is
managed by another mac interface, it means that this
network port cannot work independently, especially
when the system suspends and resumes, the following
trace may appear, so we should create a device link
between phy dev and mac dev.

  WARNING: CPU: 0 PID: 24 at drivers/net/phy/phy.c:983 phy_error+0x20/0x68
  Modules linked in:
  CPU: 0 PID: 24 Comm: kworker/0:2 Not tainted 6.1.0-rc3-00011-g5aaef24b5c6d-dirty #34
  Hardware name: Freescale i.MX6 SoloX (Device Tree)
  Workqueue: events_power_efficient phy_state_machine
  unwind_backtrace from show_stack+0x10/0x14
  show_stack from dump_stack_lvl+0x68/0x90
  dump_stack_lvl from __warn+0xb4/0x24c
  __warn from warn_slowpath_fmt+0x5c/0xd8
  warn_slowpath_fmt from phy_error+0x20/0x68
  phy_error from phy_state_machine+0x22c/0x23c
  phy_state_machine from process_one_work+0x288/0x744
  process_one_work from worker_thread+0x3c/0x500
  worker_thread from kthread+0xf0/0x114
  kthread from ret_from_fork+0x14/0x28
  Exception stack(0xf0951fb0 to 0xf0951ff8)

Signed-off-by: Xiaolei Wang <xiaolei.wang@windriver.com>
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20221130021216.1052230-1-xiaolei.wang@windriver.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-devlink-return-the-driver-name-in-devlink_nl_info_fill'

Vincent Mailhol says:

====================
net: devlink: return the driver name in devlink_nl_info_fill

The driver name is available in device_driver::name. Right now,
drivers still have to report this piece of information themselves in
their devlink_ops::info_get callback function.

The goal of this series is to have the devlink core to report this
information instead of the drivers.

The first patch fulfills the actual goal of this series: modify
devlink core to report the driver name and clean-up all drivers. Both
have to be done in an atomic change to avoid attribute
duplication. This same patch also removes the
devlink_info_driver_name_put() function to prevent future drivers from
reporting the driver name themselves.

The second patch allows the core to call devlink_nl_info_fill() even
if the devlink_ops::info_get() callback is NULL. This leads to the
third and final patch which cleans up the drivers which have an empty
info_get().
====================

Link: https://lore.kernel.org/r/20221129095140.3913303-1-mailhol.vincent@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: devlink: clean-up empty devlink_ops::info_get()

devlink_ops::info_get() is now optional and devlink will continue to
report information even if that callback gets removed.

Remove all the empty devlink_ops::info_get() callbacks from the
drivers.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: devlink: make the devlink_ops::info_get() callback optional

Some drivers only reported the driver name in their
devlink_ops::info_get() callback. Now that the core provides this
information, the callback became empty. For such drivers, just
removing the callback would prevent the core from executing
devlink_nl_info_fill() meaning that "devlink dev info" would not
return anything.

Make the callback function optional by executing
devlink_nl_info_fill() even if devlink_ops::info_get() is NULL.

N.B.: the drivers with devlink support which previously did not
implement devlink_ops::info_get() will now also be able to report
the driver name.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: devlink: let the core report the driver name instead of the drivers

The driver name is available in device_driver::name. Right now,
drivers still have to report this piece of information themselves in
their devlink_ops::info_get callback function.

In order to factorize code, make devlink_nl_info_fill() add the driver
name attribute.

Now that the core sets the driver name attribute, drivers are not
supposed to call devlink_info_driver_name_put() anymore. Remove
devlink_info_driver_name_put() and clean-up all the drivers using this
function in their callback.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Tested-by: Ido Schimmel <idosch@nvidia.com> # mlxsw
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: nfc: nxp,nci: Document NQ310 compatible

The NQ310 is another NFC chip from NXP, document the compatible in the
bindings.

Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20221128173744.833018-1-luca@z3ntu.xyz
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'support-direct-read-from-region'

Jacob Keller says:

====================
support direct read from region

A long time ago when initially implementing devlink regions in ice I
proposed the ability to allow reading from a region without taking a
snapshot [1]. I eventually dropped this work from the original series due to
size. Then I eventually lost track of submitting this follow up.

This can be useful when interacting with some region that has some
definitive "contents" from which snapshots are made. For example the ice
driver has regions representing the contents of the device flash.

If userspace wants to read the contents today, it must first take a snapshot
and then read from that snapshot. This makes sense if you want to read a
large portion of data or you want to be sure reads are consistently from the
same recording of the flash.

However if user space only wants to read a small chunk, it must first
generate a snapshot of the entire contents, perform a read from the
snapshot, and then delete the snapshot after reading.

For such a use case, a direct read from the region makes more sense. This
can be achieved by allowing the devlink region read command to work without
a snapshot. Instead the portion to be read can be forwarded directly to the
driver via a new .read callback.

This avoids the need to read the entire region contents into memory first
and avoids the software overhead of creating a snapshot and then deleting
it.

This series implements such behavior and hooks up the ice NVM and shadow RAM
regions to allow it.

[1] https://lore.kernel.org/netdev/20200130225913.1671982-1-jacob.e.keller@intel.com/
====================

Link: https://lore.kernel.org/r/20221128203647.1198669-1-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ice: implement direct read for NVM and Shadow RAM regions

Implement the .read handler for the NVM and Shadow RAM regions. This
enables user space to read a small chunk of the flash without needing the
overhead of creating a full snapshot.

Update the documentation for ice to detail which regions have direct read
support.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ice: document 'shadow-ram' devlink region

78ad87da9978 ("ice: devlink: add shadow-ram region to snapshot Shadow RAM")
added support for the 'shadow-ram' devlink region, but did not document it
in the ice devlink documentation. Fix this.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ice: use same function to snapshot both NVM and Shadow RAM

The ice driver supports a region for both the flat NVM contents as well as
the Shadow RAM which is a layer built on top of the flash during device
initialization.

These regions use an almost identical read function, except that the NVM
needs to set the direct flag when reading, while Shadow RAM needs to read
without the direct flag set. They each call ice_read_flat_nvm with the only
difference being whether to set the direct flash flag.

The NVM region read function also was fixed to read the NVM in blocks to
avoid a situation where the firmware reclaims the lock due to taking too
long.

Note that the region snapshot function takes the ops pointer so the
function can easily determine which region to read. Make use of this and
re-use the NVM snapshot function for both the NVM and Shadow RAM regions.
This makes Shadow RAM benefit from the same block approach as the NVM
region. It also reduces code in the ice driver.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

devlink: support directly reading from region memory

To read from a region, user space must currently request a new snapshot of
the region and then read from that snapshot. This can sometimes be overkill
if user space only reads a tiny portion. They first create the snapshot,
then request a read, then destroy the snapshot.

For regions which have a single underlying "contents", it makes sense to
allow supporting direct reading of the region data.

Extend the DEVLINK_CMD_REGION_READ to allow direct reading from a region if
requested via the new DEVLINK_ATTR_REGION_DIRECT. If this attribute is set,
then perform a direct read instead of using a snapshot. Direct read is
mutually exclusive with DEVLINK_ATTR_REGION_SNAPSHOT_ID, and care is taken
to ensure that we reject commands which provide incorrect attributes.

Regions must enable support for direct read by implementing the .read()
callback function. If a region does not support such direct reads, a
suitable extended error message is reported.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

devlink: refactor region_read_snapshot_fill to use a callback function

The devlink_nl_region_read_snapshot_fill is used to copy the contents of
a snapshot into a message for reporting to userspace via the
DEVLINK_CMG_REGION_READ netlink message.

A future change is going to add support for directly reading from
a region. Almost all of the logic for this new capability is identical.

To help reduce code duplication and make this logic more generic,
refactor the function to take a cb and cb_priv pointer for doing the
actual copy.

Add a devlink_region_snapshot_fill implementation that will simply copy
the relevant chunk of the region. This does require allocating some
storage for the chunk as opposed to simply passing the correct address
forward to the devlink_nl_cmg_region_read_chunk_fill function.

A future change to implement support for directly reading from a region
without a snapshot will provide a separate implementation that calls the
newly added devlink region operation.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

devlink: remove unnecessary parameter from chunk_fill function

The devlink parameter of the devlink_nl_cmd_region_read_chunk_fill
function is not used. Remove it, to simplify the function signature.

Once removed, it is also obvious that the devlink parameter is not
necessary for the devlink_nl_region_read_snapshot_fill either.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

devlink: find snapshot in devlink_nl_cmd_region_read_dumpit

The snapshot pointer is obtained inside of the function
devlink_nl_region_read_snapshot_fill. Simplify this function by locating
the snapshot upfront in devlink_nl_cmd_region_read_dumpit instead. This
aligns with how other netlink attributes are handled, and allows us to
exit slightly earlier if an invalid snapshot ID is provided.

It also allows us to pass the snapshot pointer directly to the
devlink_nl_region_read_snapshot_fill, and remove the now unused attrs
parameter.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

devlink: report extended error message in region_read_dumpit()

Report extended error details in the devlink_nl_cmd_region_read_dumpit()
function, by using the extack structure from the netlink_callback.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

devlink: use min_t to calculate data_size

The calculation for the data_size in the devlink_nl_read_snapshot_fill
function uses an if statement that is better expressed using the min_t
macro.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: microchip: vcap: Change how the rule id is generated

Currently whenever a new rule id is generated, it picks up the next
number bigger than previous id. So it would always be 1, 2, 3, etc.
When the rule with id 1 will be deleted and a new rule will be added,
it will have the id 4 and not id 1.
In theory this can be a problem if at some point a rule will be added
and removed ~0 times. Then no more rules can be added because there
are no more ids.

Change this such that when a new rule is added, search for an empty
rule id starting with value of 1 as value 0 is reserved.

Fixes: c9da1ac1c212 ("net: microchip: sparx5: Adding initial tc flower support for VCAP API")
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Link: https://lore.kernel.org/r/20221128142959.8325-1-horatiu.vultur@microchip.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeontx2-af: Simplify a size computation in rvu_npc_exact_init()

We know that table_size = table->mem_table.depth * table->mem_table.ways,
so use it instead, it is less verbose.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/5230dabe27f48948a9fd0f50a62e2437b65e6a6e.1669378798.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeontx2-af: Fix the size of memory allocated for the 'id_bmap' bitmap

This allocation is really spurious.
The size of the bitmap is 'tot_ids' and it is used as such in the driver.

So we could expect something like:
   table->id_bmap = devm_kcalloc(rvu->dev, BITS_TO_LONGS(table->tot_ids),
         sizeof(long), GFP_KERNEL);

However, when the bitmap is allocated, we allocate:
   BITS_TO_LONGS(table->tot_ids) * table->tot_ids ~=
   table->tot_ids / 32 * table->tot_ids ~=
   table->tot_ids^2 / 32

It is proportional to the square of 'table->tot_ids' which seems to
potentially be big.

Allocate the expected amount of memory, and switch to the bitmap API to
have it more straightforward.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/ce2710771939065d68f95d86a27cf7cea7966365.1669378798.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeontx2-af: Use the bitmap API to allocate bitmaps

Use devm_bitmap_zalloc() instead of hand-writing it.

This also makes the comment "Allocate bitmap for 32 entry mcam" more
explicit because now 32 is really used in the allocation function, instead
of an obscure 'sizeof(long)'.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/24177a9ee7043259448b735263d9cfd6a70e89a4.1669378798.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeontx2-af: Slightly simplify rvu_npc_exact_init()

Use kzalloc() instead of kmalloc()/memset().

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/60ea220ccf3b61963f7d5a97e3df2c76a5feb837.1669378798.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeontx2-af: Fix a potentially spurious error message

When this error message is displayed, we know that the all the bits in the
bitmap are set.

So, bitmap_weight() will return the number of bits of the bitmap, which is
'table->tot_ids'.

It is unlikely that a bit will be cleared between mutex_unlock() and
dev_err(), but, in order to simplify the code and avoid this possibility,
just take 'table->tot_ids'.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/5ce01c402f86412dc57884ff0994b63f0c5b3871.1669378798.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

selftests/net: add csum offload test

Test NIC hardware checksum offload:

- Rx + Tx
- IPv4 + IPv6
- TCP + UDP

Optional features:

- zero checksum 0xFFFF
- checksum disable 0x0000
- transport encap headers
- randomization

See file header for detailed comments.

Expected results differ depending on NIC features:

- CHECKSUM_UNNECESSARY vs CHECKSUM_COMPLETE
- NETIF_F_HW_CSUM (csum_start/csum_off) vs NETIF_F_IP(V6)_CSUM

Signed-off-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20221128140210.553391-1-willemdebruijn.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/mlx5e: Support devlink reload of IPsec core

Change IPsec initialization flow to allow future creation of hardware
resources that should be released and allocated during devlink reload
operation. As part of that change, update function signature to be
void as no callers are actually interested in it.

Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: TC, Add offload support for trap with additional actions

TC trap action offload is currently supported only when trap is the sole action
in the flow.

This patch remove this limitation by changing trap action offload to not use
MLX5_ATTR_FLAG_SLOW_PATH flag and instead set the flow destination table
explicitly to be the slow table. This will allow offload of the additional
actions.

TC flow example:
tc filter add dev $REP2 protocol ip prio 2 root \
flower skip_sw dst_mac $mac0 \
action mirred egress redirect dev $REP3 \
action pedit ex munge eth dst set $mac2  pipe \
action trap

Signed-off-by: Maor Dickman <maord@nvidia.com>
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Do early return when setup vports dests for slow path flow

Adding flow flag cases in setup vport dests before the slow path
case is incorrect as the slow path should take precedence.
Current code doesn't show this importance so make the slow path
case return early and separate from the other cases and remove
the redundant comparison of it in the sample case.

Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5: Remove redundant check

If ASO failed in creation, it won't be called to destroy either.
The kernel coding pattern is to make sure that callers are calling
to destroy only for valid objects.

Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Delete always true DMA check

DMA address always exists for MACsec ASO object.

Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Don't access directly DMA device pointer

Use specialized helper to fetch DMA device pointer.

Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Don't use termination table when redundant

Current code used termination table for each vport destination
while it's only required for hairpin, i.e. uplink to uplink, or
when vlan push on rx action being used.
Fix to skip using termination table for vport destinations that
do not require it.

Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Maor Dickman <maord@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5: Fix orthography errors in documentation

Improve general readability of the device driver documentation.

Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5: Use generic definition for UMR KLM alignment

MLX5_UMR_KLM_ALIGNMENT is in units of number of entries, while
MLX5_UMR_MTT_ALIGNMENT (generalized and renamed to
MLX5_UMR_FLEX_ALIGNMENT) is in byte units. This is misleading and
confusing.
Replace this KLM definition with one based on the generic definition.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5: Generalize name of UMR alignment definition

Per the device spec, MLX5_UMR_MTT_ALIGNMENT is good not only for UMR MTT
entries, but for all other entries as well, like KLMs and KSMs.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5: Remove unused UMR MTT definitions

Defines MLX5_UMR_MTT_MASK and MLX5_UMR_MTT_MIN_CHUNK_SIZE are not in
use. Remove them.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Add padding when needed in UMR WQEs

Per the device spec, MTTs/KLMs list in a UMR WQE must be aligned to 64B.
Per our SW design, the MTT/KLMs list would need alignment only if it's
too small, for example on PPC when PAGE_SIZE is 64KB, and only 4 pages
are needed to cover a MPWQE of size 256KB.

Padding, if needed, is taken into account when calculating the UMR WQE
fields (ds_cnt and xlt_octowords), however no entries are provided,
instead garbage is passed.

No real harm though, as these parts act as gaps between the RX MPWQEs
and not used by any of them. Hence, in practice, device does not try to
write any incoming packet to them. Still, prefer providing clean padding
marking the end of the list, and do not map garbage into the RQ memory
region.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5: Remove unused ctx variables

Remove mlx5_priv.ctx_list and ctx_lock which are no longer used after
commit 601c10c89cbb ("net/mlx5: Delete custom device management logic").

Signed-off-by: Petr Pavlu <petr.pavlu@suse.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Replace zero-length arrays with DECLARE_FLEX_ARRAY() helper

Zero-length arrays are deprecated and we are moving towards adopting
C99 flexible-array members, instead. So, replace zero-length arrays
declarations in anonymous union with the new DECLARE_FLEX_ARRAY()
helper macro.

This helper allows for flexible-array members in unions.

Link: https://github.com/KSPP/linux/issues/193
Link: https://github.com/KSPP/linux/issues/222
Link: https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Remove unneeded io-mapping.h #include

The mlx5 net files don't use io_mapping functionalities. So there is no
point in including <linux/io-mapping.h>.
Remove it.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Merge branch 'master' of git://git./linux/kernel/git/klassert/ipsec-next

Steffen Klassert says:

====================
ipsec-next 2022-11-26

1) Remove redundant variable in esp6.
   From Colin Ian King.

2) Update x->lastused for every packet. It was used only for
   outgoing mobile IPv6 packets, but showed to be usefull
   to check if the a SA is still in use in general.
   From Antony Antony.

3) Remove unused variable in xfrm_byidx_resize.
   From Leon Romanovsky.

4) Finalize extack support for xfrm.
   From Sabrina Dubroca.

* 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next:
  xfrm: add extack to xfrm_set_spdinfo
  xfrm: add extack to xfrm_alloc_userspi
  xfrm: add extack to xfrm_do_migrate
  xfrm: add extack to xfrm_new_ae and xfrm_replay_verify_len
  xfrm: add extack to xfrm_del_sa
  xfrm: add extack to xfrm_add_sa_expire
  xfrm: a few coding style clean ups
  xfrm: Remove not-used total variable
  xfrm: update x->lastused for every packet
  esp6: remove redundant variable err
====================

Link: https://lore.kernel.org/r/20221126110303.1859238-1-steffen.klassert@secunet.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-pcs-altera-tse-simplify-and-clean-up-the-driver'

Maxime Chevallier says:

====================
net: pcs: altera-tse: simplify and clean-up the driver

This small series does a bit of code cleanup in the altera TSE pcs
driver, removing unused register definitions, handling 1000BaseX speed
configuration correctly according to the datasheet, and making use of
proper poll_timeout helpers.

No functional change is introduced.
====================

Link: https://lore.kernel.org/r/20221125131801.64234-1-maxime.chevallier@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: pcs: altera-tse: remove unnecessary register definitions

remove unused register definitions, left from the split with the
altera-tse mac driver.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: pcs: altera-tse: don't set the speed for 1000BaseX

When disabling the SGMII mode bit, the PCS defaults to 1000BaseX mode.
In that mode, we don't need to set the speed since it's always 1000Mbps.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: pcs: altera-tse: use read_poll_timeout to wait for reset

Software resets on the TSE PCS don't clear registers, but rather reset
all internal state machines regarding AN, comma detection and
encoding/decoding. Use read_poll_timeout to wait for the reset to clear
instead of manually polling the register.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'mptcp-msg_fastopen-and-tfo-listener-side-support'

Matthieu Baerts says:

====================
mptcp: MSG_FASTOPEN and TFO listener side support

Before this series, only the initiator of a connection was able to combine
both TCP FastOpen and MPTCP when using TCP_FASTOPEN_CONNECT socket option.

These new patches here add (in theory) the full support of TFO with MPTCP,
which means:

 - MSG_FASTOPEN sendmsg flag support (patch 1/8)
 - TFO support for the listener side (patches 2-5/8)
 - TCP_FASTOPEN socket option (patch 6/8)
 - TCP_FASTOPEN_KEY socket option (patch 7/8)

To support TFO for the server side, a few preparation patches are needed
(patches 2 to 5/8). Some of them were inspired by a previous work from
Benjamin Hesmans.

Note that TFO support with MPTCP has been validated with selftests
(patch 8/8) but also with Packetdrill tests running with a modified
but still very WIP version supporting MPTCP. Both the modified tool
and the tests are available online:

  https://github.com/multipath-tcp/packetdrill/
====================

Link: https://lore.kernel.org/r/20221125222958.958636-1-matthieu.baerts@tessares.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

selftests: mptcp: mptfo Initiator/Listener

This patch first adds TFO support in mptcp_connect.c.

This can be enabled via a new option: -o MPTFO.

Once enabled, the TCP_FASTOPEN socket option is enabled for the server
side and a sendto() with MSG_FASTOPEN is used instead of a connect() for
the client side.

Note that the first SYN has a limit of bytes it can carry. In other
words, it is allowed to send less data than the provided one. We then
need to track more status info to properly allow the next sendmsg()
starting from the next part of the data to send the rest.

Also in TFO scenarios, we need to completely spool the partially xmitted
buffer -- and account for that -- before starting sendfile/mmap xmit,
otherwise the relevant tests will fail.

Co-developed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Dmytro Shytyi <dmytro@shytyi.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

mptcp: add support for TCP_FASTOPEN_KEY sockopt

The goal of this socket option is to set different keys per listener,
see commit 1fba70e5b6be ("tcp: socket option to set TCP fast open key")
for more details about this socket option.

The only thing to do here with MPTCP is to relay the request to the
first subflow like it is already done for the other TCP_FASTOPEN* socket
options.

Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

mptcp: add TCP_FASTOPEN sock option

The TCP_FASTOPEN socket option is one way for the application to tell
the kernel TFO support has to be enabled for the listener socket.

The only thing to do here with MPTCP is to relay the request to the
first subflow like it is already done for the other TCP_FASTOPEN* socket
options.

Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Dmytro Shytyi <dmytro@shytyi.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

mptcp: add subflow_v(4,6)_send_synack()

The send_synack() needs to be overridden for MPTCP to support TFO for
two reasons:

- There is not be enough space in the TCP options if the TFO cookie has
  to be added in the SYN+ACK with other options: MSS (4), SACK OK (2),
  Timestamps (10), Window Scale (3+1), TFO (10+2), MP_CAPABLE (12).
  MPTCPv1 specs -- RFC 8684, section B.1 [1] -- suggest to drop the TCP
  timestamps option in this case.

- The data received in the SYN has to be handled: the SKB can be
  dequeued from the subflow sk and transferred to the MPTCP sk. Counters
  need to be updated accordingly and the application can be notified at
  the end because some bytes have been received.

[1] https://www.rfc-editor.org/rfc/rfc8684.html#section-b.1

Co-developed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Co-developed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Dmytro Shytyi <dmytro@shytyi.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

mptcp: implement delayed seq generation for passive fastopen

With fastopen in place, the first subflow socket is created before the
MPC handshake completes, and we need to properly initialize the sequence
numbers at MPC ACK reception.

Co-developed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Co-developed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Dmytro Shytyi <dmytro@shytyi.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

mptcp: consolidate initial ack seq generation

Currently the initial ack sequence is generated on demand whenever
it's requested and the remote key is handy. The relevant code is
scattered in different places and can lead to multiple, unneeded,
crypto operations.

This change consolidates the ack sequence generation code in a single
helper, storing the sequence number at the subflow level.

The above additionally saves a few conditional in fast-path and will
simplify the upcoming fast-open implementation.

Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

mptcp: track accurately the incoming MPC suboption type

Currently in the receive path we don't need to discriminate
between MPC SYN, MPC SYN-ACK and MPC ACK, but soon the fastopen
code will need that info to properly track the fully established
status.

Track the exact MPC suboption type into the receive opt bitmap.
No functional change intended.

Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

mptcp: add MSG_FASTOPEN sendmsg flag support

Since commit 54f1944ed6d2 ("mptcp: factor out mptcp_connect()"), all the
infrastructure is now in place to support the MSG_FASTOPEN flag, we
just need to call into the fastopen path in mptcp_sendmsg().

Co-developed-by: Benjamin Hesmans <benjamin.hesmans@tessares.net>
Signed-off-by: Benjamin Hesmans <benjamin.hesmans@tessares.net>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Dmytro Shytyi <dmytro@shytyi.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge git://git./linux/kernel/git/netdev/net

tools/lib/bpf/ringbuf.c
  927cbb478adf ("libbpf: Handle size overflow for ringbuf mmap")
  b486d19a0ab0 ("libbpf: checkpatch: Fixed code alignments in ringbuf.c")
https://lore.kernel.org/all/20221121122707.44d1446a@canb.auug.org.au/

Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge tag 'net-6.1-rc8-2' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from bpf, can and wifi.

  Current release - new code bugs:

   - eth: mlx5e:
      - use kvfree() in mlx5e_accel_fs_tcp_create()
      - MACsec, fix RX data path 16 RX security channel limit
      - MACsec, fix memory leak when MACsec device is deleted
      - MACsec, fix update Rx secure channel active field
      - MACsec, fix add Rx security association (SA) rule memory leak

  Previous releases - regressions:

   - wifi: cfg80211: don't allow multi-BSSID in S1G

   - stmmac: set MAC's flow control register to reflect current settings

   - eth: mlx5:
      - E-switch, fix duplicate lag creation
      - fix use-after-free when reverting termination table

  Previous releases - always broken:

   - ipv4: fix route deletion when nexthop info is not specified

   - bpf: fix a local storage BPF map bug where the value's spin lock
     field can get initialized incorrectly

   - tipc: re-fetch skb cb after tipc_msg_validate

   - wifi: wilc1000: fix Information Element parsing

   - packet: do not set TP_STATUS_CSUM_VALID on CHECKSUM_COMPLETE

   - sctp: fix memory leak in sctp_stream_outq_migrate()

   - can: can327: fix potential skb leak when netdev is down

   - can: add number of missing netdev freeing on error paths

   - aquantia: do not purge addresses when setting the number of rings

   - wwan: iosm:
      - fix incorrect skb length leading to truncated packet
      - fix crash in peek throughput test due to skb UAF"

* tag 'net-6.1-rc8-2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (79 commits)
  net: ethernet: renesas: ravb: Fix promiscuous mode after system resumed
  MAINTAINERS: Update maintainer list for chelsio drivers
  ionic: update MAINTAINERS entry
  sctp: fix memory leak in sctp_stream_outq_migrate()
  packet: do not set TP_STATUS_CSUM_VALID on CHECKSUM_COMPLETE
  net/mlx5: Lag, Fix for loop when checking lag
  Revert "net/mlx5e: MACsec, remove replay window size limitation in offload path"
  net: marvell: prestera: Fix a NULL vs IS_ERR() check in some functions
  net: tun: Fix use-after-free in tun_detach()
  net: mdiobus: fix unbalanced node reference count
  net: hsr: Fix potential use-after-free
  tipc: re-fetch skb cb after tipc_msg_validate
  mptcp: fix sleep in atomic at close time
  mptcp: don't orphan ssk in mptcp_close()
  dsa: lan9303: Correct stat name
  ipv4: Fix route deletion when nexthop info is not specified
  net: wwan: iosm: fix incorrect skb length
  net: wwan: iosm: fix crash in peek throughput test
  net: wwan: iosm: fix dma_alloc_coherent incompatible pointer type
  net: wwan: iosm: fix kernel test robot reported error
  ...

udp_tunnel: Add checks for nla_nest_start() in __udp_tunnel_nic_dump_write()

As the nla_nest_start() may fail with NULL returned, the return value
should be checked.

Note that this is not a real bug, nothing will break here.
The next nla_put() will fail as well and we'll bail (and
nla_nest_cancel() can handle NULL). But we keep getting
those "fixes" so whatever.

Signed-off-by: Yuan Can <yuancan@huawei.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20221129013934.55184-1-yuancan@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: renesas: ravb: Fix promiscuous mode after system resumed

After system resumed on some environment board, the promiscuous mode
is disabled because the SoC turned off. So, call ravb_set_rx_mode() in
the ravb_resume() to fix the issue.

Reported-by: Tho Vu <tho.vu.wh@renesas.com>
Fixes: 0184165b2f42 ("ravb: add sleep PM suspend/resume support")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Link: https://lore.kernel.org/r/20221128065604.1864391-1-yoshihiro.shimoda.uh@renesas.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

MAINTAINERS: Update maintainer list for chelsio drivers

This updates the maintainers for chelsio inline crypto drivers and
chelsio crypto drivers.

Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Link: https://lore.kernel.org/r/20221128231348.8225-1-ayush.sawal@chelsio.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ionic: update MAINTAINERS entry

Now that Pensando is a part of AMD we need to update
a couple of addresses.  We're keeping the mailing list
address for the moment, but that will likely change in
the near future.

Signed-off-by: Shannon Nelson <shannon.nelson@amd.com>
Signed-off-by: Brett Creeley <brett.creeley@amd.com>
Link: https://lore.kernel.org/r/20221129011734.20849-1-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

sctp: fix memory leak in sctp_stream_outq_migrate()

When sctp_stream_outq_migrate() is called to release stream out resources,
the memory pointed to by prio_head in stream out is not released.

The memory leak information is as follows:
 unreferenced object 0xffff88801fe79f80 (size 64):
   comm "sctp_repo", pid 7957, jiffies 4294951704 (age 36.480s)
   hex dump (first 32 bytes):
     80 9f e7 1f 80 88 ff ff 80 9f e7 1f 80 88 ff ff  ................
     90 9f e7 1f 80 88 ff ff 90 9f e7 1f 80 88 ff ff  ................
   backtrace:
     [<ffffffff81b215c6>] kmalloc_trace+0x26/0x60
     [<ffffffff88ae517c>] sctp_sched_prio_set+0x4cc/0x770
     [<ffffffff88ad64f2>] sctp_stream_init_ext+0xd2/0x1b0
     [<ffffffff88aa2604>] sctp_sendmsg_to_asoc+0x1614/0x1a30
     [<ffffffff88ab7ff1>] sctp_sendmsg+0xda1/0x1ef0
     [<ffffffff87f765ed>] inet_sendmsg+0x9d/0xe0
     [<ffffffff8754b5b3>] sock_sendmsg+0xd3/0x120
     [<ffffffff8755446a>] __sys_sendto+0x23a/0x340
     [<ffffffff87554651>] __x64_sys_sendto+0xe1/0x1b0
     [<ffffffff89978b49>] do_syscall_64+0x39/0xb0
     [<ffffffff89a0008b>] entry_SYSCALL_64_after_hwframe+0x63/0xcd

Link: https://syzkaller.appspot.com/bug?exrid=29c402e56c4760763cc0
Fixes: 637784ade221 ("sctp: introduce priority based stream scheduler")
Reported-by: syzbot+29c402e56c4760763cc0@syzkaller.appspotmail.com
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Reviewed-by: Xin Long <lucien.xin@gmail.com>
Link: https://lore.kernel.org/r/20221126031720.378562-1-shaozhengchao@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

packet: do not set TP_STATUS_CSUM_VALID on CHECKSUM_COMPLETE

CHECKSUM_COMPLETE signals that skb->csum stores the sum over the
entire packet. It does not imply that an embedded l4 checksum
field has been validated.

Fixes: 682f048bd494 ("af_packet: pass checksum validation status to the user")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20221128161812.640098-1-willemdebruijn.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/mlx5: Lag, Fix for loop when checking lag

The cited commit adds a for loop to check if each port supports lag
or not. But dev is not initialized correctly. Fix it by initializing
dev for each iteration.

Fixes: e87c6a832f88 ("net/mlx5: E-switch, Fix duplicate lag creation")
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reported-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Link: https://lore.kernel.org/r/20221129093006.378840-2-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Revert "net/mlx5e: MACsec, remove replay window size limitation in offload path"

This reverts commit c0071be0e16c461680d87b763ba1ee5e46548fde.

The cited commit removed the validity checks which initialized the
window_sz and never removed the use of the now uninitialized variable,
so now we are left with wrong value in the window size and the following
clang warning: [-Wuninitialized]
drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c:232:45:
       warning: variable 'window_sz' is uninitialized when used here
       MLX5_SET(macsec_aso, aso_ctx, window_size, window_sz);

Revet at this time to address the clang issue due to lack of time to
test the proper solution.

Fixes: c0071be0e16c ("net/mlx5e: MACsec, remove replay window size limitation in offload path")
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reported-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20221129093006.378840-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfp: ethtool: support reporting link modes

Add support for reporting link modes,
including `Supported link modes` and `Advertised link modes`,
via ethtool $DEV.

A new command `SPCODE_READ_MEDIA` is added to read info from
management firmware. Also, the mapping table `nfp_eth_media_table`
associates the link modes between NFP and kernel. Both of them
help to support this ability.

Signed-off-by: Yu Xiao <yu.xiao@corigine.com>
Reviewed-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Link: https://lore.kernel.org/r/20221125113030.141642-1-simon.horman@corigine.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: devlink: add WARN_ON_ONCE to check return value of unregister_netdevice_notifier_net() call

As the return value is not 0 only in case there is no such notifier
block registered, add a WARN_ON_ONCE() to yell about it.

Suggested-by: Ido Schimmel <idosch@idosch.org>
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Link: https://lore.kernel.org/r/20221125100255.1786741-1-jiri@resnulli.us
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'add-support-for-lan966x-is2-vcap'

Horatiu Vultur says:

====================
Add support for lan966x IS2 VCAP

This provides initial support for lan966x for 'tc' traffic control
userspace tool and its flower filter. For this is required to use
the VCAP library.

Currently supported flower filter keys and actions are:
- source and destination MAC address keys
- trap action
====================

Link: https://lore.kernel.org/r/20221125095010.124458-1-horatiu.vultur@microchip.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: microchip: vcap: Implement w32be

On lan966x the layout of the vcap memory is different than on sparx5.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: lan966x: Add port keyset config and callback interface

Implement vcap_operations and enable default port keyset configuration
for each port. Now it is possible actually write/read/move entries in
the VCAP.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: lan966x: add tc matchall goto action

Extend matchall with action goto. This is needed to enable the lookup in
the VCAP. It is needed to connect chain 0 to a chain that is recognized
by the HW.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>