Allow mdb entries to be forwarded to CPU and be switched at the same
time. Only remove entry when no port and the CPU isn't part of the group
anymore.
Signed-off-by: Casper Andersson <casper.casan@gmail.com>
Acked-by: Steen Hegelund <Steen.Hegelund@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

napi_build_skb() reuses NAPI skbuff_head cache in order to save some
cycles on freeing/allocating skbuff_heads on every new Rx or completed
Tx.
Use napi_consume_skb() to feed the cache with skbuff_heads of completed
Tx so it's never empty.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

NAPI cache skb_count is being checked twice without condition. Change to
checking the second time only if the first check is run.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Adding mdb entries on disabled ports allows you to do setup before
accepting any traffic, avoiding any time where the port is not in the
multicast group.
Signed-off-by: Casper Andersson <casper.casan@gmail.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add the capability to map non-linear xdp frames in XDP_TX and ndo_xdp_xmit
callback.
Tested-by: Sarkar Tirthendu <tirthendu.sarkar@intel.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

phydev->port was not set and always reported as PORT_TP.
Set phydev->port according to inserted SFP module.
Signed-off-by: Ivan Bornyakov <i.bornyakov@metrotek.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add basic description for the xilinx emaclite driver DT bindings.
Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@amd.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Alex Elder says:

====================
net: ipa: simplify completion statistics

The first patch in this series makes the name used for variables
representing a TRE ring be consistent everywhere.  The second
renames two structure fields to better represent their purpose.

The last four rework a little code that manages some tranaction and
byte transfer statistics maintained mainly for TX endpoints.  For
the most part this series is refactoring.  The last one also
includes the first step toward no longer assuming an event ring is
dedicated to a single channel.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Rename gsi_channel_tx_update() to be gsi_trans_tx_completed(), and
pass it just the transaction pointer, deriving the channel from the
transaction.  Update the comments above the function to provide a
more concise description of how statistics for TX endpoints are
maintained and used.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

In gsi_evt_ring_rx_update(), we update each transaction so its len
field reflects the actual number of bytes received.  In the process,
the total number of transactions and bytes processed on the channel
are summed, and added to a running total for the channel.

But we don't actually use those running totals for RX endpoints.
They're maintained for TX channels to support CoDel when they are
associated with a "real" network device.

So stop maintaining these totals for RX endpoints, and update the
comment where the fields are defined to make it clear they're only
valid for TX channels.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

When a TX request is issued, its channel's accumulated byte and
transaction counts are recorded.  This currently does *not* take
into account the transaction being committed.

Later, when the transaction completes, the number of bytes and
transactions that have completed since the transaction was committed
are reported to the network stack.  The transaction and its byte
count are accounted for at that time.

Instead, record the transaction and its bytes in the counts recorded
at commit time.  This avoids the need to do so when the transaction
completes, and provides a (small) simplification of that code.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Create a new function that encapsulates recording information needed
for TX channel statistics when a transaction is committed.

Record the accumulated length in the transaction before the call
(for both RX and TX), so it can be used when updating TX statistics.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

There are two fields in a GSI transaction that keep track of TRE
counts.  The first represents the number of TREs reserved for the
transaction in the TRE ring; that's currently named "tre_count".
The second is the number of TREs that are actually *used* by the
transaction at the time it is committed.

Rename the "tre_count" field to be "rsvd_count", to make its meaning
a little more specific.  The "_count" is present in the name mainly
to avoid interpreting it as a reserved (not-to-be-used) field.  This
name also distinguishes it from the "tre_count" field associated
with a channel.

Rename the "used" field to be "used_count", to match the convention
used for reserved TREs.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

All local variables that represent event rings are named "ring".

All but two functions that represent a channel's TRE ring with a
local variable use the name "tre_ring".  For consistency, use that
name in the two functions that don't fit the pattern.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Frank Wunderlich says:

====================
Support mt7531 on BPI-R2 Pro

This Series add Support for the mt7531 switch on Bananapi R2 Pro board.

This board uses port5 of the switch to conect to the gmac0 of the
rk3568 SoC.

Currently CPU-Port is hardcoded in the mt7530 driver to port 6.

Compared to v1 the reset-Patch was dropped as it was not needed and
CPU-Port-changes are completely rewriten based on suggestions/code from
Vladimir Oltean (many thanks to this).
In DTS Patch i only dropped the status-property that was not
needed/ignored by driver.

Due to the Changes i also made a regression test on mt7623 bpi-r2
(mt7623 soc + mt7530) and bpi-r64 (mt7622 soc + mt7531) with cpu-
port 6. Tests were done directly (ipv4 config on dsa user port)
and with vlan-aware bridge including vlan that was tagged outgoing
on dsa user port.
====================

Link: https://lore.kernel.org/r/20220610170541.8643-1-linux@fw-web.deSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Add Device Tree node for mt7531 switch connected to gmac0.
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

A board may have no independent reset-line, so reset cannot be used
inside switch driver.

E.g. on Bananapi-R2 Pro switch and gmac are connected to same reset-line.

Resets should be acquired only to 1 device/driver. This prevents reset to
be bound to switch-driver if reset is already used for gmac. If reset is
only used by switch driver it resets the switch *and* the gmac after the
mdio bus comes up resulting in mdio bus goes down. It takes some time
until all is up again, switch driver tries to read from mdio, will fail
and defer the probe. On next try the reset does the same again.

Make reset optional for such boards.

Allow port 5 as cpu-port and phy-mode rgmii for mt7531.

- MT7530 supports RGMII on port 5 and RGMII/TRGMII on port 6.
- MT7531 supports on port 5 RGMII and SGMII (dual-sgmii) and
  SGMII on port 6.
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Replace last occurences of hardcoded cpu-port by cpu_dp member of
dsa_port struct.

Now the constant can be dropped.
Suggested-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Enumerate available cpu-ports instead of using hardcoded constant.
Suggested-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Rework vlan_add/vlan_del functions in preparation for dynamic cpu port.

Currently BIT(MT7530_CPU_PORT) is added to new_members, even though
mt7530_port_vlan_add() will be called on the CPU port too.

Let DSA core decide when to call port_vlan_add for the CPU port, rather
than doing it implicitly.

We can do autonomous forwarding in a certain VLAN, but not add br0 to that
VLAN and avoid flooding the CPU with those packets, if software knows it
doesn't need to process them.
Suggested-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Convert txt binding to yaml binding for Mediatek switches.
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Ido Schimmel says:

====================
mlxsw: Remove XM support

The XM was supposed to be an external device connected to the
Spectrum-{2,3} ASICs using dedicated Ethernet ports. Its purpose was to
increase the number of routes that can be offloaded to hardware. This was
achieved by having the ASIC act as a cache that refers cache misses to the
XM where the FIB is stored and LPM lookup is performed.

Testing was done over an emulator and dedicated setups in the lab, but
the product was discontinued before shipping to customers.

Therefore, in order to remove dead code and reduce complexity of the
code base, revert the three patchsets that added XM support.
====================

Link: https://lore.kernel.org/r/20220613132116.2021055-1-idosch@nvidia.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

This reverts commit 923ba95e ("Merge branch
'mlxsw-spectrum-prepare-for-xm-implementation-lpm-trees'").
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

This reverts commit e7086213 ("Merge branch
'mlxsw-spectrum-prepare-for-xm-implementation-prefix-insertion-and-removal'").
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

This reverts commit 75c2a8fe ("Merge branch
'mlxsw-introduce-initial-xm-router-support'").
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Saeed Mahameed says:

====================
mlx5-next: updates 2022-06-14

1) Updated HW bits and definitions for upcoming features
 1.1) vport debug counters
 1.2) flow meter
 1.3) Execute ASO action for flow entry
 1.4) enhanced CQE compression

2) Add ICM header-modify-pattern RDMA API

Leon Says
=========

SW steering manipulates packet's header using "modifying header" actions.
Many of these actions do the same operation, but use different data each time.
Currently we create and keep every one of these actions, which use expensive
and limited resources.

Now we introduce a new mechanism - pattern and argument, which splits
a modifying action into two parts:
1. action pattern: contains the operations to be applied on packet's header,
mainly set/add/copy of fields in the packet
2. action data/argument: contains the data to be used by each operation
in the pattern.

This way we reuse same patterns with different arguments to create new
modifying actions, and since many actions share the same operations, we end
up creating a small number of patterns that we keep in a dedicated cache.

These modify header patterns are implemented as new type of ICM memory,
so the following kernel patch series add the support for this new ICM type.
==========

* 'mlx5-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux:
  net/mlx5: Add bits and fields to support enhanced CQE compression
  net/mlx5: Remove not used MLX5_CAP_BITS_RW_MASK
  net/mlx5: group fdb cleanup to single function
  net/mlx5: Add support EXECUTE_ASO action for flow entry
  net/mlx5: Add HW definitions of vport debug counters
  net/mlx5: Add IFC bits and enums for flow meter
  RDMA/mlx5: Support handling of modify-header pattern ICM area
  net/mlx5: Manage ICM of type modify-header pattern
  net/mlx5: Introduce header-modify-pattern ICM properties
====================

Link: https://lore.kernel.org/r/20220614184028.51548-1-saeed@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Add missing documentation for the TLS_TX_ZEROCOPY_RO opt-in.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Link: https://lore.kernel.org/r/20220610180212.110590-1-kuba@kernel.orgSigned-off-by: Paolo Abeni <pabeni@redhat.com>

Fix the implementation of ethtool_convert_link_mode_to_legacy_u32(), which
is supposed to return false if src has bits higher than 31 set. The current
implementation uses the complement of bitmap_fill(ext, 32) to test high
bits of src, which is wrong as bitmap_fill() fills _with long granularity_,
and sizeof(long) can be > 4. No users of this function currently check the
return value, so the bug was dormant.

Also remove the check for __ETHTOOL_LINK_MODE_MASK_NBITS > 32, as the enum
ethtool_link_mode_bit_indices contains far beyond 32 values. Using
find_next_bit() to test the src bitmask works regardless of this anyway.
Signed-off-by: Marco Bonelli <marco@mebeim.net>
Link: https://lore.kernel.org/r/20220609134900.11201-1-marco@mebeim.netSigned-off-by: Jakub Kicinski <kuba@kernel.org>

There's no point probing for phys on this artificial bus, so we can
save a little bit of boot time by telling mdiobus_register() not to do
that.

This doesn't have any functional change, since, at this point,
fixed_mdio_read() returns 0xffff for all addresses/registers, so

  mdiobus_scan() -> get_phy_device() -> get_phy_c22_id()

will return -ENODEV, which is just ignored.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Link: https://lore.kernel.org/r/20220606200208.1665417-1-linux@rasmusvillemoes.dkSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Expose ifc bits and add needed structure fields and methods to
support enhanced CQE compression feature.
The enhanced CQE compression feature improves cpu utiliziation with
better packet latency from nic to host.
Signed-off-by: Ofer Levi <oferle@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Remove not used MLX5_CAP_BITS_RW_MASK.
While at it, remove CAP_MASK, MLX5_CAP_OFF_CMDIF_CSUM
and MLX5_DEV_CAP_FLAG_*, since MLX5_CAP_BITS_RW_MASK
was their only user.
Signed-off-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Currently, the allocation of fdb software objects are done is single
function, oppose to the cleanup of them.
Group the cleanup of fdb software objects to single function.
Signed-off-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Attach flow meter to FTE with object id and index.
Use metadata register C5 to store the packet color meter result.
Signed-off-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

total_q_under_processor_handle - number of queues in error state due to an
async error or errored command.

send_queue_priority_update_flow - number of QP/SQ priority/SL update
events.

cq_overrun - number of times CQ entered an error state due to an
overflow.

async_eq_overrun -number of time an EQ mapped to async events was
overrun.

comp_eq_overrun - number of time an EQ mapped to completion events was
overrun.

quota_exceeded_command - number of commands issued and failed due to quota
exceeded.

invalid_command - number of commands issued and failed dues to any reason
other than quota exceeded.
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Michael Guralnik <michaelgur@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Add/extend structure layouts and defines for flow meter.
Signed-off-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Add support for allocate/deallocate and registering MR of the new type
of ICM area. Support exists only for devices that support sw_owner_v2.
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Added support for managing new type of ICM for devices that
support sw_owner_v2.
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Acked-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Added new fields for device memory capabilities, in order to
support creation of ICM memory for modify header patterns.
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Acked-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

__sys_accept4_file() isn't used outside of the file, make it static.

As the same time, move file_flags and nofile parameters into
__sys_accept4_file().
Signed-off-by: Yajun Deng <yajun.deng@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

sk_memory_allocated_add() has three callers, and returns
to them @memory_allocated.

sk_forced_mem_schedule() is one of them, and ignores
the returned value.

Change sk_memory_allocated_add() to return void.

Change sock_reserve_memory() and __sk_mem_raise_allocated()
to call sk_memory_allocated().

This removes one cache line miss [1] for RPC workloads,
as first skbs in TCP write queue and receive queue go through
sk_forced_mem_schedule().

[1] Cache line holding tcp_memory_allocated.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>