As the code to update ring stats is alike for different ring stats
type, this patch extract macro to simplify ring stats update code.
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This reverts commit aeeecb88.

The new SNMP variable (TCPSmallQueueFailure) can be incremented
for good reasons, even on a 100Gbit single TCP_STREAM flow.

If we really wanted to ease driver debugging [1], this would
require something more sophisticated.

[1] Usually, if a driver is delaying TX completions too much,
this can lead to stalls in TCP output. Various work arounds
have been used in the past, like skb_orphan() in ndo_start_xmit().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Menglong Dong <imagedong@tencent.com>
Link: https://lore.kernel.org/r/20211201033246.2826224-1-eric.dumazet@gmail.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Program the 1ms autonegotiation clock divisor according to the clocking
rate of neta - without this, the 1ms clock ticks at about 660us on
Armada 38x configured for 250MHz. Bring this into correct specification.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Marek Behún <kabel@kernel.org>
Link: https://lore.kernel.org/r/E1ms4WD-00EKLK-Ld@rmk-PC.armlinux.org.ukSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Tidy up the disable bit clearing where we clear a bit
and then run the link resolver.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/E1ms4Rx-00EKEc-En@rmk-PC.armlinux.org.ukSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Russell King says:

====================
net: dsa: convert two drivers to phylink_generic_validate()

Patches 1 to 3 update core DSA code to allow drivers to be converted,
and patches 4 and 5 convert hellcreek and lantiq to use this (both of
which received reviewed-by from their maintainers.) As the rest have
yet to be reviewed by their maintainers, they are not included here.

Patch 1 had a request to change the formatting of it; I have not done
so as I believe a patch should do one change and one change only -
reformatting it is a separate change that should be in its own patch.
However, as patch 2 gets rid of the reason for reformatting it, it
would be pointless, and pure noise to include such an intermediary
patch.

Instead, I have swapped the order of patches 2 and 3 from the RFC
series.
====================

Link: https://lore.kernel.org/r/YaYiiU9nvmVugqnJ@shell.armlinux.org.ukSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Populate the supported interfaces and MAC capabilities for the Lantiq
DSA switches and remove the old validate implementation to allow DSA to
use phylink_generic_validate() for this switch driver.

The exclusion of Gigabit linkmodes for MII, Reverse MII and Reduced MII
links is handled within phylink_generic_validate() in phylink, so there
is no need to make them conditional on the interface mode in the driver.
Reviewed-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Populate the supported interfaces and MAC capabilities for the
hellcreek DSA switch and remove the old validate implementation to
allow DSA to use phylink_generic_validate() for this switch driver.

The switch actually only supports MII and RGMII, but as phylib defaults
to GMII, we need to include this interface mode to keep existing DT
working.
Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de>
Tested-by: Kurt Kanzenbach <kurt@linutronix.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Support the use of phylink_generic_validate() when there is no
phylink_validate method given in the DSA switch operations and
mac_capabilities have been set in the phylink_config structure by the
DSA switch driver.

This gives DSA switch drivers the option to use this if they provide
the supported_interfaces and mac_capabilities, while still giving them
an option to override the default implementation if necessary.
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Phylink needs slightly more information than phylink_get_interfaces()
allows us to get from the DSA drivers - we need the MAC capabilities.
Replace the phylink_get_interfaces() method with phylink_get_caps() to
allow DSA drivers to fill in the phylink_config MAC capabilities field
as well.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

The code in port.c and slave.c creating the phylink instance is very
similar - let's consolidate this into a single function.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

There's build error while CONFIG_REGMAP_MMIO is not set
and CONFIG_MDIO_MSCC_MIIM=m.

ERROR: modpost: "__devm_regmap_init_mmio_clk"
[drivers/net/mdio/mdio-mscc-miim.ko] undefined!

Add the depend of REGMAP_MMIO to fix it.

Fixes: a27a7628 ("net: mdio: mscc-miim: convert to a regmap implementation")
Signed-off-by: Tianchen Ding <dtcccc@linux.alibaba.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20211130110209.804536-1-dtcccc@linux.alibaba.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

No concurrent access is possible when a bitmap is local to a function.
So prefer the non-atomic functions to save a few cycles.
   - replace a 'for' loop by an equivalent non-atomic 'bitmap_fill()' call
   - use '__set_bit()'

While at it, clear the 'bitmask' bitmap only when needed.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/3de0792f5088f00d135c835df6c19e63ae95f5d2.1638026251.git.christophe.jaillet@wanadoo.frSigned-off-by: Jakub Kicinski <kuba@kernel.org>

From: Ido Schimmel <idosch@nvidia.com>

====================
mlxsw: Spectrum-4 preparations

The Spectrum-4 ASIC will support more than 256 ports, unlike existing
ASICs. As such, various device registers were extended with two
additional bits to encode a 10 bit local port. In some cases, new
(Version 2) registers were introduced.

This patchset prepares mlxsw for Spectrum-4 support by encoding local
port in a 'u16' instead of a 'u8' and by extending the definitions of
the various registers to support the extended local port form.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The SMID-V2 register maps Multicast ID (MID) into a list of local ports.
It is a new version of SMID in order to support 1024 bits of local_port.

Add SMID-V2 register and use it instead of SMID.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The SFTR-V2 register is used for flooding packet replication.
It is a new version of SFTR in order to support 1024 bits of local_port.

Add SFTR-V2 register and use it instead of SFTR.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add 'port_page' field in SBSR to be able to query occupancy of more than
256 ports. The field determines the range of the ports specified in the
'ingress_port_mask' and 'egress_port_mask' bit masks:
>From '256 * port_page' to '256 * port_page + 255'.

For each local port, the appropriate port page is used. A query is never
performed for a port range that spans multiple port pages.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, local_port field is saved as u8, which means that maximum 256
ports can be used.

As preparation for Spectrum-4, which will support more than 256 ports,
local_port field should be extended.

Save local_port as u16 to allow use of additional ports.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Local port 255 has a special meaning in PPCNT register, it is used to
refer to all local ports. This wild card ability is not currently used
by the driver.

Special casing local port 255 in Spectrum-4 systems where it is a valid
port is going to be a problem.

Work around this issue by adding and always setting the 'lp_gl' bit
which instructs the device's firmware to treat this local port like an
ordinary port.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

'port_num' field is used to indicate the local port value which can be
assigned to a module.

Increase the field from 8 bits to 10 bits in order to support more than
255 ports.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add support for 10-bit local ports in device registers by making use of the
MLXSW_ITEM32_LP() macro that was added in the previous patch.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, local_port field uses 8 bits, which means that maximum 256
ports can be used.

As preparation for the next ASIC, which will support more than 256
ports, local_port field should be extended to 10 bits.

It is not possible to use 10 consecutive bits in all registers, and
therefore, the field is split into 2 fields:
1. local_port - the existing 8 bits, represent LSB of the extended
   field.
2. lp_msb - extra 2 bits, represent MSB of the extended field.

To avoid complex programming when reading/writing local_port, add a
dedicated macro which creates get and set functions which handle both parts
of local_port.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The functions mlxsw_reg_sfd_uc_unpack() and
mlxsw_reg_sfd_uc_lag_unpack() are not used. Remove them.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add latest verified version of Nvidia Spectrum-family switch firmware,
for Spectrum (13.2010.1006), Spectrum-2 (29.2010.1006) and
Spectrum-3 (30.2010.1006).
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Tony Nguyen says:

====================
40GbE Intel Wired LAN Driver Updates 2021-11-30

This series contains updates to iavf driver only.

Patryk adds a debug message when MTU is changed.

Grzegorz adds messaging when transitioning in and out of multicast
promiscuous mode.

Jake returns correct error codes for iavf_parse_cls_flower().

Jedrzej adds messaging for when the driver is removed and refactors
struct usage to take less memory. He also adjusts ethtool statistics to
only display information on active queues.

Tony allows for user to specify the RSS hash.

Karen resolves some static analysis warnings, corrects format specifiers,
and rewords a message to come across as informational.

v2:
- Dropped patch 1 (for net) and 5
- Change MTU message from info to debug
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Tony Nguyen says:

====================
100GbE Intel Wired LAN Driver Updates 2021-11-30

This series contains updates to ice driver only.

Shiraz corrects assignment of boolean variable and removes an unused
enum.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Tony Nguyen says:

====================
1GbE Intel Wired LAN Driver Updates 2021-11-30

Jesper Dangaard Brouer says:

Changes to fix and enable XDP metadata to a specific Intel driver igc.
Tested with hardware i225 that uses driver igc, while testing AF_XDP
access to metadata area.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Use eth_hw_addr_set function instead of writing the address directly to
net_device::dev_addr.

Fixes: adeef3e3 ("net: constify netdev->dev_addr")
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Link: https://lore.kernel.org/r/20211130143600.31970-1-jcmvbkbc@gmail.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

The skb will be checked inside kfree_skb(), so remove the
outside check.
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20211130031243.768823-1-yangyingliang@huawei.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Driver provided too many lines as an output to ethtool -S command.
Return actual length of string set of ethtool stats. Instead of predefined
maximal value use the actual value on netdev, iterate over active queues.
Without this patch, ethtool -S report would produce additional
erroneous lines of queues that are not configured.
Signed-off-by: Witold Fijalkowski <witoldx.fijalkowski@intel.com>
Signed-off-by: Przemyslaw Patynowski <przemyslawx.patynowski@intel.com>
Signed-off-by: Mateusz Palczewski <mateusz.palczewski@intel.com>
Signed-off-by: Jedrzej Jagielski <jedrzej.jagielski@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Change format to match variable type that is used in string.

Use %u format for unsigned variable and %d format for signed variable
to remove static analysis warnings.
Signed-off-by: Michal Swiatkowski <michal.swiatkowski@intel.com>
Signed-off-by: Karen Sornek <karen.sornek@intel.com>
Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

This message is intended to be informational to indicate a reset is about
to happen, but the use of "warning" in the message text can cause concern
with users.  Reword the message to make it less alarming.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Karen Sornek <karen.sornek@intel.com>
Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Change min() to min_t() to fix static code analysis warning of possible
overflow.
Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Signed-off-by: Karen Sornek <karen.sornek@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

iavf_mac_filter struct contained couple boolean
flags using up more memory than is necessary.
Change the flags to be bitfields in an anonymous struct
so all the flags now fit in one byte.
Signed-off-by: Sylwester Dziedziuch <sylwesterx.dziedziuch@intel.com>
Signed-off-by: Jedrzej Jagielski <jedrzej.jagielski@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Driver support for changing the RSS hash key exists, however, checks
have caused it to be reported as unsupported. Remove the check and
allow the hash key to be specified.
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Tony Brelinski <tony.brelinski@intel.com>

Add kernel trace that device was removed.
Currently there is no such information.
I.e. Host admin removes a PCI device from a VM,
than on VM shall be info about the event.

This patch adds info log to iavf_remove function.
Signed-off-by: Arkadiusz Kubalewski <arkadiusz.kubalewski@intel.com>
Signed-off-by: Jedrzej Jagielski <jedrzej.jagielski@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

The iavf_parse_cls_flower function returns an integer error code, and
not an iavf_status enumeration.

Fix the function to use the standard errno value EINVAL as its return
instead of using IAVF_ERR_CONFIG.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Add log when VF is entering and leaving Allmulti mode.
The change of VF state is visible in dmesg now.
Without this commit, entering and leaving Allmulti mode
is not logged in dmesg.
Signed-off-by: Grzegorz Szczurek <grzegorzx.szczurek@intel.com>
Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Add a netdev_dbg log entry in case of a change of MTU so that user is
notified about this change in the same manner as in case of pf driver.
Signed-off-by: Patryk Małek <patryk.malek@intel.com>
Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Enabling the XDP bpf_prog access to data_meta area is a very small
change. Hint passing 'true' to xdp_prepare_buff().

The SKB layers can also access data_meta area, which required more
driver changes to support. Reviewers, notice the igc driver have two
different functions that can create SKBs, depending on driver config.

Hint for testers, ethtool priv-flags legacy-rx enables
the function igc_construct_skb()

 ethtool --set-priv-flags DEV legacy-rx on
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Tested-by: Nechama Kraus <nechamax.kraus@linux.intel.com>
Reviewed-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Driver already implicitly supports XDP metadata access in AF_XDP
zero-copy mode, as xsk_buff_pool's xp_alloc() naturally set xdp_buff
data_meta equal data.

This works fine for XDP and AF_XDP, but if a BPF-prog adjust via
bpf_xdp_adjust_meta() and choose to call XDP_PASS, then igc function
igc_construct_skb_zc() will construct an invalid SKB packet. The
function correctly include the xdp->data_meta area in the memcpy, but
forgot to pull header to take metasize into account.

Fixes: fc9df2a0 ("igc: Enable RX via AF_XDP zero-copy")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Tested-by: Nechama Kraus <nechamax.kraus@linux.intel.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>