ionic: set ionic ptr before setting up ethtool ops

Set the lif->ionic value that is used in some ethtool callbacks
before setting ethtool ops.  There really shouldn't be any
race issues before this change since the netdev hasn't been
registered yet, but this seems more correct.

Signed-off-by: Shannon Nelson <shannon.nelson@amd.com>
Reviewed-by: Brett Creeley <brett.creeley@amd.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Link: https://lore.kernel.org/r/20231204210936.16587-3-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ionic: Use cached VF attributes

Each time a VF attribute is set via iproute a call to get the VF
configuration is also made. This is currently problematic because for
each VF configuration call there are multiple commands sent to the
device. Unfortunately, this doesn't scale well. Fix this by reporting
the cached VF attributes.

The original change to query the device for getting the VF attributes
    f16f5be31009 ("ionic: Query FW when getting VF info via ndo_get_vf_config")
was made to remain consistent with device set VF attributes. However,
after further investigation there is no need to query the device.

Signed-off-by: Brett Creeley <brett.creeley@amd.com>
Signed-off-by: Shannon Nelson <shannon.nelson@amd.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Link: https://lore.kernel.org/r/20231204210936.16587-2-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

packet: add a generic drop reason for receive

Commit da37845fdce2 ("packet: uses kfree_skb() for errors.") switches
from consume_skb to kfree_skb to improve error handling. However, this
could bring a lot of noises when we monitor real packet drops in
kfree_skb[1], because in tpacket_rcv or packet_rcv only packet clones
can be freed, not actual packets.

Adding a generic drop reason to allow distinguish these "clone drops".

[1]: https://lore.kernel.org/netdev/CABWYdi00L+O30Q=Zah28QwZ_5RU-xcxLFUK2Zj08A8MrLk9jzg@mail.gmail.com/
Fixes: da37845fdce2 ("packet: uses kfree_skb() for errors.")
Suggested-by: Eric Dumazet <edumazet@google.com>
Suggested-by: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
Signed-off-by: Yan Zhai <yan@cloudflare.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/ZW4piNbx3IenYnuw@debian.debian
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Documentations: fix net_cachelines documentation build warning

Original errors:
Documentation/networking/net_cachelines/index.rst:3: WARNING: Explicit markup ends without a blank line; unexpected unindent.
Documentation/networking/net_cachelines/inet_connection_sock.rst:3: WARNING: Explicit markup ends without a blank line; unexpected unindent.
Documentation/networking/net_cachelines/inet_sock.rst:3: WARNING: Explicit markup ends without a blank line; unexpected unindent.
Documentation/networking/net_cachelines/net_device.rst:3: WARNING: Explicit markup ends without a blank line; unexpected unindent.
Documentation/networking/net_cachelines/netns_ipv4_sysctl.rst:3: WARNING: Explicit markup ends without a blank line; unexpected unindent.
Documentation/networking/net_cachelines/snmp.rst:3: WARNING: Explicit markup ends without a blank line; unexpected unindent.
Documentation/networking/net_cachelines/tcp_sock.rst:3: WARNING: Explicit markup ends without a blank line; unexpected unindent.

Fixes: 14006f1d8fa2 ("Documentations: Analyze heavily used Networking related structs")
Signed-off-by: Coco Li <lixiaoyan@google.com>
Link: https://lore.kernel.org/r/20231204220728.746134-1-lixiaoyan@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: core: synchronize link-watch when carrier is queried

There are multiple ways to query for the carrier state: through
rtnetlink, sysfs, and (possibly) ethtool. Synchronize linkwatch
work before these operations so that we don't have a situation
where userspace queries the carrier state between the driver's
carrier off->on transition and linkwatch running and expects it
to work, when really (at least) TX cannot work until linkwatch
has run.

I previously posted a longer explanation of how this applies to
wireless [1] but with this wireless can simply query the state
before sending data, to ensure the kernel is ready for it.

[1] https://lore.kernel.org/all/346b21d87c69f817ea3c37caceb34f1f56255884.camel@sipsolutions.net/

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20231204214706.303c62768415.I1caedccae72ee5a45c9085c5eb49c145ce1c0dd5@changeid
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'reorganize-remaining-patch-of-networking-struct-cachelines'

Coco Li says:

====================
Reorganize remaining patch of networking struct cachelines

Rebase patches to top-of-head in https://lwn.net/Articles/951321/ to
ensure the results of the cacheline savings are still accurate.
====================

Link: https://lore.kernel.org/r/20231204201232.520025-1-lixiaoyan@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tcp: reorganize tcp_sock fast path variables

The variables are organized according in the following way:

- TX read-mostly hotpath cache lines
- TXRX read-mostly hotpath cache lines
- RX read-mostly hotpath cache lines
- TX read-write hotpath cache line
- TXRX read-write hotpath cache line
- RX read-write hotpath cache line

Fastpath cachelines end after rcvq_space.

Cache line boundaries are enforced only between read-mostly and
read-write. That is, if read-mostly tx cachelines bleed into
read-mostly txrx cachelines, we do not care. We care about the
boundaries between read and write cachelines because we want
to prevent false sharing.

Fast path variables span cache lines before change: 12
Fast path variables span cache lines after change: 8

Suggested-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Wei Wang <weiwan@google.com>
Signed-off-by: Coco Li <lixiaoyan@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20231204201232.520025-3-lixiaoyan@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net-device: reorganize net_device fast path variables

Reorganize fast path variables on tx-txrx-rx order
Fastpath variables end after npinfo.

Below data generated with pahole on x86 architecture.

Fast path variables span cache lines before change: 12
Fast path variables span cache lines after change: 4

Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Coco Li <lixiaoyan@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20231204201232.520025-2-lixiaoyan@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeon_ep: control net API framework to support offloads

Inquire firmware on supported offloads, as well as convey offloads
enabled dynamically to firmware. New control net API functionality is
required for the above. Implement control net API framework for
offloads.

Additionally, fetch/insert offload metadata from hardware RX/TX
buffer respectively during receive/transmit.

Currently supported offloads include checksum and TSO.

Signed-off-by: Shinas Rasheed <srasheed@marvell.com>
Link: https://lore.kernel.org/r/20231204154940.2583140-1-srasheed@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-mvmdio-performance-related-improvements'

Tobias Waldekranz says:

====================
net: mvmdio: Performance related improvements

Observations of the XMDIO bus on a CN9130-based system during a
firmware download showed a very low bus utilization, which stemmed
from the 150us (10x the average access time) sleep which would take
place when the first poll did not succeed.

With this series in place, bus throughput increases by about 10x,
multiplied by whatever gain you are able to extract from running the
MDC at a higher frequency (hardware dependent).
====================

Link: https://lore.kernel.org/r/20231204100811.2708884-1-tobias@waldekranz.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: mvmdio: Support setting the MDC frequency on XSMI controllers

Support the standard "clock-frequency" attribute to set the generated
MDC frequency. If not specified, the driver will leave the divisor
untouched.

Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20231204100811.2708884-4-tobias@waldekranz.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: mvmdio: Avoid excessive sleeps in polled mode

Before this change, when operating in polled mode, i.e. no IRQ is
available, every individual C45 access would be hit with a 150us sleep
after the bus access.

For example, on a board with a CN9130 SoC connected to an MV88X3310
PHY, a single C45 read would take around 165us:

    root@infix:~$ mdio f212a600.mdio-mii mmd 4:1 bench 0xc003
    Performed 1000 reads in 165ms

By replacing the long sleep with a tighter poll loop, we observe a 10x
increase in bus throughput:

    root@infix:~$ mdio f212a600.mdio-mii mmd 4:1 bench 0xc003
    Performed 1000 reads in 15ms

Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20231204100811.2708884-3-tobias@waldekranz.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tools: ynl: move private definitions to a separate header

ynl.h has a growing amount of "internal" stuff, which may confuse
users who try to take a look at the external API. Currently the
internals are at the bottom of the file with a banner in between,
but this arrangement makes it hard to add external APIs / inline
helpers which need internal definitions.

Move internals to a separate header.

Link: https://lore.kernel.org/r/20231202211225.342466-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tools: ynl: use strerror() if no extack of note provided

If kernel didn't give use any meaningful error - print
a strerror() to the ynl error message.

Reviewed-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Link: https://lore.kernel.org/r/20231202211310.342716-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tools: pynl: make flags argument optional for do()

Commit 1768d8a767f8 ("tools/net/ynl: Add support for create flags")
added support for setting legacy netlink CRUD flags on netlink
messages (NLM_F_REPLACE, _EXCL, _CREATE etc.).

Most of genetlink won't need these, don't force callers to pass
in an empty argument to each do() call.

Reviewed-by: Donald Hunter <donald.hunter@gmail.com>
Link: https://lore.kernel.org/r/20231202211005.341613-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-convert-to-platform-remove-callback-returning-void'

Uwe Kleine-König says:

====================
net*: Convert to platform remove callback returning void

(implicit) v1 of this series can be found at
https://lore.kernel.org/netdev/20231117095922.876489-1-u.kleine-koenig@pengutronix.de.

Changes since then:

 - Dropped patch #1 as Alex objected. Patch #1 (was #2 before) now
   converts ipa to remove_new() and introduces an error message in the
   error path that failed before.
====================

Link: https://lore.kernel.org/r/cover.1701713943.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: wwan: qcom_bam_dmux: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.

To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Reviewed-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Link: https://lore.kernel.org/r/20231117095922.876489-9-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/49795ee930be6a9a24565e5e7133e6f8383ab532.1701713943.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: wan/ixp4xx_hss: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.

To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Link: https://lore.kernel.org/r/20231117095922.876489-8-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/b0488fa6181a47668e5737905ae7adc8d7cd055e.1701713943.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: wan/fsl_ucc_hdlc: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.

To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Link: https://lore.kernel.org/r/20231117095922.876489-7-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/8c9ffca75ea24810f9ba05a514d5ad59847cc4fe.1701713943.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sfp: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.

To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Link: https://lore.kernel.org/r/20231117095922.876489-6-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/7c1d50d559c0e0e36a20eb3e410f6e9d3f884b6f.1701713943.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: pcs: rzn1-miic: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.

To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20231117095922.876489-5-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/82b728e14a68c421e269eff3b8083d9d6e62d956.1701713943.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: fjes: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.

To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Link: https://lore.kernel.org/r/20231117095922.876489-4-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/4889ac6a7ffa9b02fa5cdd2d3212e739741f80b8.1701713943.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.

To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Link: https://lore.kernel.org/r/20231117095922.876489-3-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/c43193b9a002e88da36b111bb44ce2973ecde722.1701713943.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: reduce stack usage in hclge_dbg_dump_tm_pri()

This function exceeds the stack frame warning limit:

drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c: In function 'hclge_dbg_dump_tm_pri':
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c:1039:1: error: the frame size of 1408 bytes is larger than 1024 bytes [-Werror=frame-larger-than=]

Use dynamic allocation for the largest stack object instead. It
would be nice to rewrite this file to completely avoid the extra
buffer and just use the one that was already allocated by debugfs,
but that is a much larger change.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20231204085735.4112882-1-arnd@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tools: ynl: remove generated user space code from git

The ynl-generated user space C code is already above 25kLoC
and is growing.

The initial reason to commit these files was to make reviewing changes
to the generator easier. Unfortunately, it has the opposite effect on
reviewing changes to specs, and we get far more changes to specs
than to the generator.

Uncommit those fails, as they are generated on the fly as needed.
netdev patchwork now runs a script on each series to create a diff
of generated code on the fly, for the rare cases when looking at
it is helpful:
https://github.com/kuba-moo/nipa/blob/master/tests/series/ynl/ynl.sh

Suggested-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'sfc-implement-ndo_hwtstamp_-get-set'

Alex Austin says:

====================
sfc: Implement ndo_hwtstamp_(get|set)

Implement ndo_hwtstamp_get and ndo_hwtstamp_set for sfc and sfc-siena.
====================

Link: https://lore.kernel.org/r/20231130135826.19018-1-alex.austin@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

sfc-siena: Implement ndo_hwtstamp_(get|set)

Update efx->ptp_data to use kernel_hwtstamp_config and implement
ndo_hwtstamp_(get|set). Remove SIOCGHWTSTAMP and SIOCSHWTSTAMP from
efx_ioctl.

Signed-off-by: Alex Austin <alex.austin@amd.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Reviewed-by: Edward Cree <ecree.xilinx@gmail.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20231130135826.19018-3-alex.austin@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

sfc: Implement ndo_hwtstamp_(get|set)

Update efx->ptp_data to use kernel_hwtstamp_config and implement
ndo_hwtstamp_(get|set). Remove SIOCGHWTSTAMP and SIOCSHWTSTAMP from
efx_ioctl.

Signed-off-by: Alex Austin <alex.austin@amd.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Reviewed-by: Edward Cree <ecree.xilinx@gmail.com>
Link: https://lore.kernel.org/r/20231130135826.19018-2-alex.austin@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ipvlan: implement .parse_protocol hook function in ipvlan_header_ops

The .parse_protocol hook function in the ipvlan_header_ops structure is
not implemented. As a result, when the AF_PACKET family is used to send
packets, skb->protocol will be set to 0.
Ipvlan is a device of type ARPHRD_ETHER (ether_setup). Therefore, use
eth_header_parse_protocol function to obtain the protocol.

Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20231202130438.2266343-1-shaozhengchao@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

macvlan: implement .parse_protocol hook function in macvlan_hard_header_ops

The .parse_protocol hook function in the macvlan_header_ops structure is
not implemented. As a result, when the AF_PACKET family is used to send
packets, skb->protocol will be set to 0.
Macvlan is a device of type ARPHRD_ETHER (ether_setup). Therefore, use
eth_header_parse_protocol function to obtain the protocol.

Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20231202130658.2266526-1-shaozhengchao@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'conver-net-selftests-to-run-in-unique-namespace-part-1'

Hangbin Liu says:

====================
Conver net selftests to run in unique namespace (Part 1)

As Guillaume pointed, many selftests create namespaces with very common
names (like "client" or "server") or even (partially) run directly in init_net.
This makes these tests prone to failure if another namespace with the same
name already exists. It also makes it impossible to run several instances
of these tests in parallel.

This patch set intend to conver all the net selftests to run in unique namespace,
so we can update the selftest freamwork to run all tests in it's own namespace
in parallel. After update, we only need to wait for the test which need
longest time.

As the total patch set is too large. I break it to severl parts. This is
the first part.

v2 -> v3:
- Convert all ip netns del to cleanup_ns (Justin Iurman)

v1 -> v2:
- Split the large patch set to small parts for easy review (Paolo Abeni)
- Move busywait from forwarding/lib.sh to net/lib.sh directly (Petr Machata)
- Update setup_ns/cleanup_ns struct (Petr Machata)
- Remove default trap in lib.sh (Petr Machata)
====================

Link: https://lore.kernel.org/r/20231202020110.362433-1-liuhangbin@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert unicast_extensions.sh to run it in unique namespace

Here is the test result after conversion.

 # ./unicast_extensions.sh
 /usr/bin/which: no nettest in (/root/.local/bin:/root/bin:/usr/share/Modules/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin)
 ###########################################################################
 Unicast address extensions tests (behavior of reserved IPv4 addresses)
 ###########################################################################
 TEST: assign and ping within 240/4 (1 of 2) (is allowed)            [ OK ]
 TEST: assign and ping within 240/4 (2 of 2) (is allowed)            [ OK ]
 TEST: assign and ping within 0/8 (1 of 2) (is allowed)              [ OK ]

 ...

 TEST: assign and ping class D address (is forbidden)                [ OK ]
 TEST: routing using class D (is forbidden)                          [ OK ]
 TEST: routing using 127/8 (is forbidden)                            [ OK ]

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert sctp_vrf.sh to run it in unique namespace

Here is the test result after conversion.

]# ./sctp_vrf.sh
Testing For SCTP VRF:
TEST 01: nobind, connect from client 1, l3mdev_accept=1, Y [PASS]
...
TEST 12: bind vrf-2 & 1 in server, connect from client 1 & 2, N [PASS]
***v6 Tests Done***

Acked-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert ndisc_unsolicited_na_test.sh to run it in unique namespace

Here is the test result after conversion.

]# ./ndisc_unsolicited_na_test.sh
    TEST: test_unsolicited_na:  drop_unsolicited_na=0  accept_untracked_na=1  forwarding=1  [ OK ]
    TEST: test_unsolicited_na:  drop_unsolicited_na=0  accept_untracked_na=0  forwarding=0  [ OK ]
    TEST: test_unsolicited_na:  drop_unsolicited_na=0  accept_untracked_na=0  forwarding=1  [ OK ]
    TEST: test_unsolicited_na:  drop_unsolicited_na=0  accept_untracked_na=1  forwarding=0  [ OK ]
    TEST: test_unsolicited_na:  drop_unsolicited_na=1  accept_untracked_na=0  forwarding=0  [ OK ]
    TEST: test_unsolicited_na:  drop_unsolicited_na=1  accept_untracked_na=0  forwarding=1  [ OK ]
    TEST: test_unsolicited_na:  drop_unsolicited_na=1  accept_untracked_na=1  forwarding=0  [ OK ]
    TEST: test_unsolicited_na:  drop_unsolicited_na=1  accept_untracked_na=1  forwarding=1  [ OK ]

Tests passed:   8
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert l2tp.sh to run it in unique namespace

Here is the test result after conversion.

]# ./l2tp.sh
TEST: IPv4 basic L2TP tunnel                                        [ OK ]
TEST: IPv4 route through L2TP tunnel                                [ OK ]
TEST: IPv6 basic L2TP tunnel                                        [ OK ]
TEST: IPv6 route through L2TP tunnel                                [ OK ]
TEST: IPv4 basic L2TP tunnel - with IPsec                           [ OK ]
TEST: IPv4 route through L2TP tunnel - with IPsec                   [ OK ]
TEST: IPv6 basic L2TP tunnel - with IPsec                           [ OK ]
TEST: IPv6 route through L2TP tunnel - with IPsec                   [ OK ]
TEST: IPv4 basic L2TP tunnel                                        [ OK ]
TEST: IPv4 route through L2TP tunnel                                [ OK ]
TEST: IPv6 basic L2TP tunnel - with IPsec                           [ OK ]
TEST: IPv6 route through L2TP tunnel - with IPsec                   [ OK ]
TEST: IPv4 basic L2TP tunnel - after IPsec teardown                 [ OK ]
TEST: IPv4 route through L2TP tunnel - after IPsec teardown         [ OK ]
TEST: IPv6 basic L2TP tunnel - after IPsec teardown                 [ OK ]
TEST: IPv6 route through L2TP tunnel - after IPsec teardown         [ OK ]

Tests passed:  16
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Reviewed-by: James Chapman <jchapman@katalix.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert ioam6.sh to run it in unique namespace

Here is the test result after conversion.

]# ./ioam6.sh

--------------------------------------------------------------------------
OUTPUT tests
--------------------------------------------------------------------------
TEST: Unknown IOAM namespace (inline mode)                          [ OK ]
TEST: Unknown IOAM namespace (encap mode)                           [ OK ]
TEST: Missing trace room (inline mode)                              [ OK ]
TEST: Missing trace room (encap mode)                               [ OK ]
TEST: Trace type with bit 0 only (inline mode)                      [ OK ]
...
TEST: Full supported trace (encap mode)                             [ OK ]

--------------------------------------------------------------------------
GLOBAL tests
--------------------------------------------------------------------------
TEST: Forward - Full supported trace (inline mode)                  [ OK ]
TEST: Forward - Full supported trace (encap mode)                   [ OK ]

- Tests passed: 88
- Tests failed: 0

Acked-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Justin Iurman <justin.iurman@uliege.be>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

sleftests/net: convert icmp.sh to run it in unique namespace

Here is the test result after conversion.

]# ./icmp.sh
OK

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert icmp_redirect.sh to run it in unique namespace

Here is the test result after conversion.

 # ./icmp_redirect.sh

 ###########################################################################
 Legacy routing
 ###########################################################################

 TEST: IPv4: redirect exception                                      [ OK ]

 ...

 TEST: IPv4: mtu exception plus redirect                             [ OK ]
 TEST: IPv6: mtu exception plus redirect                             [ OK ]

 Tests passed:  40
 Tests failed:   0
 Tests xfailed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert traceroute.sh to run it in unique namespace

Here is the test result after conversion.

]# ./traceroute.sh
TEST: IPV6 traceroute                                               [ OK ]
TEST: IPV4 traceroute                                               [ OK ]

Tests passed:   2
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert drop_monitor_tests.sh to run it in unique namespace

Here is the test result after conversion.

]# ./drop_monitor_tests.sh

Software drops test
    TEST: Capturing active software drops                               [ OK ]
    TEST: Capturing inactive software drops                             [ OK ]

Hardware drops test
    TEST: Capturing active hardware drops                               [ OK ]
    TEST: Capturing inactive hardware drops                             [ OK ]

Tests passed:   4
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert cmsg tests to make them run in unique namespace

Here is the test result after conversion.

]# ./cmsg_ipv6.sh
OK
]# ./cmsg_so_mark.sh
OK
]# ./cmsg_time.sh
OK

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert arp_ndisc_untracked_subnets.sh to run it in unique namespace

Here is the test result after conversion.

2 tests also failed without this patch

]# ./arp_ndisc_untracked_subnets.sh
    TEST: test_arp:  accept_arp=0                                       [ OK ]
    TEST: test_arp:  accept_arp=1                                       [ OK ]
    TEST: test_arp:  accept_arp=2  same_subnet=0                        [ OK ]
    TEST: test_arp:  accept_arp=2  same_subnet=1                        [ OK ]
    TEST: test_ndisc:  accept_untracked_na=0                            [ OK ]
    TEST: test_ndisc:  accept_untracked_na=1                            [ OK ]
    TEST: test_ndisc:  accept_untracked_na=2  same_subnet=0             [ OK ]
    TEST: test_ndisc:  accept_untracked_na=2  same_subnet=1             [ OK ]

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: specify the interface when do arping

When do arping, the interface need to be specified. Or we will
get error: Interface "lo" is not ARPable. And the test failed.
]# ./arp_ndisc_untracked_subnets.sh
    TEST: test_arp:  accept_arp=0                                       [ OK ]
    TEST: test_arp:  accept_arp=1                                       [FAIL]
    TEST: test_arp:  accept_arp=2  same_subnet=0                        [ OK ]
    TEST: test_arp:  accept_arp=2  same_subnet=1                        [FAIL]

After fix:
]# ./arp_ndisc_untracked_subnets.sh
    TEST: test_arp:  accept_arp=0                                       [ OK ]
    TEST: test_arp:  accept_arp=1                                       [ OK ]
    TEST: test_arp:  accept_arp=2  same_subnet=0                        [ OK ]
    TEST: test_arp:  accept_arp=2  same_subnet=1                        [ OK ]

Fixes: 0ea7b0a454ca ("selftests: net: arp_ndisc_untracked_subnets: test for arp_accept and accept_untracked_na")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: convert arp_ndisc_evict_nocarrier.sh to run it in unique namespace

Here is the test result after conversion.

]# ./arp_ndisc_evict_nocarrier.sh
run arp_evict_nocarrier=1 test
ok
run arp_evict_nocarrier=0 test
ok
run all.arp_evict_nocarrier=0 test
ok
run ndisc_evict_nocarrier=1 test
ok
run ndisc_evict_nocarrier=0 test
ok
run all.ndisc_evict_nocarrier=0 test
ok

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests/net: add lib.sh

Add a lib.sh for net selftests. This file can be used to define commonly
used variables and functions. Some commonly used functions can be moved
from forwarding/lib.sh to this lib file. e.g. busywait().

Add function setup_ns() for user to create unique namespaces with given
prefix name.

Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'intel-wired-lan-driver-updates-2023-12-01-ice'

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-12-01 (ice)

This series contains updates to ice driver only.

Konrad provides temperature reporting via hwmon.

Arkadiusz adds reporting of Clock Generation Unit (CGU) information via
devlink info.

Pawel adjusts error messaging for ntuple filters to account for additional
possibility for encountering an error.

Karol ensures that all timestamps occurring around reset are processed and
renames some E822 functions to convey additional usage for E823 devices.

Jake provides mechanism to ensure that all timestamps on E822 devices
are processed.

The following are changes since commit 15bc81212f593fbd7bda787598418b931842dc14:
  octeon_ep: set backpressure watermark for RX queues
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue 100GbE
====================

Link: https://lore.kernel.org/r/20231201180845.219494-1-anthony.l.nguyen@intel.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

ice: Rename E822 to E82X

When code is applicable for both E822 and E823 devices, rename it from
E822 to E82X.
ICE_PHY_PER_NAC_E822 was unused, so just remove it.

Signed-off-by: Karol Kolacinski <karol.kolacinski@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

ice: periodically kick Tx timestamp interrupt

The E822 hardware for Tx timestamping keeps track of how many
outstanding timestamps are still in the PHY memory block. It will not
generate a new interrupt to the MAC until all of the timestamps in the
region have been read.

If somehow all the available data is not read, but the driver has exited
its interrupt routine already, the PHY will not generate a new interrupt
even if new timestamp data is captured. Because no interrupt is
generated, the driver never processes the timestamp data. This state
results in a permanent failure for all future Tx timestamps.

It is not clear how the driver and hardware could enter this state.
However, if it does, there is currently no recovery mechanism.

Add a recovery mechanism via the periodic PTP work thread which invokes
ice_ptp_periodic_work(). Introduce a new check,
ice_ptp_maybe_trigger_tx_interrupt() which checks the PHY timestamp
ready bitmask. If any bits are set, trigger a software interrupt by
writing to PFINT_OICR.

Once triggered, the main timestamp processing thread will read through
the PHY data and clear the outstanding timestamp data. Once cleared, new
data should trigger interrupts as expected.

This should allow recovery from such a state rather than leaving the
device in a state where we cannot process Tx timestamps.

It is possible that this function checks for timestamp data
simultaneously with the interrupt, and it might trigger additional
unnecessary interrupts. This will cause a small amount of additional
processing.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Karol Kolacinski <karol.kolacinski@intel.com>
Reviewed-by: Andrii Staikov <andrii.staikov@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

ice: Re-enable timestamping correctly after reset

During reset, TX_TSYN interrupt should be processed as it may process
timestamps in brief moments before and after reset.
Timestamping should be enabled on VSIs at the end of reset procedure.
On ice_get_phy_tx_tstamp_ready error, interrupt should not be rearmed
because error only happens on resets.

Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Karol Kolacinski <karol.kolacinski@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

ice: Improve logs for max ntuple errors

Supported number of ntuple filters affect also maximum location value that
can be provided to ethtool command. Update error message to provide info
about max supported value.

Fix double spaces in the error messages.

Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Pawel Kaminski <pawel.kaminski@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

ice: add CGU info to devlink info callback

If Clock Generation Unit is present on NIC board user shall know its
details.
Provide the devlink info callback with a new:
- fixed type object (cgu.id) indicating hardware variant of onboard CGU,
- running type object (fw.cgu) consisting of CGU id, config and firmware
versions.
These information shall be known for debugging purposes.

Test (on NIC board with CGU)
$ devlink dev info <bus_name>/<dev_name> | grep cgu
        cgu.id 36
        fw.cgu 8032.16973825.6021

Test (on NIC board without CGU)
$ devlink dev info <bus_name>/<dev_name> | grep cgu -c
0

Reviewed-by: Larysa Zaremba <larysa.zaremba@intel.com>
Signed-off-by: Arkadiusz Kubalewski <arkadiusz.kubalewski@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

ice: read internal temperature sensor

Since 4.30 firmware exposes internal thermal sensor reading via admin
queue commands. Expose those readouts via hwmon API when supported.

Datasheet:

Get Sensor Reading Command (Opcode: 0x0632)

+--------------------+--------+--------------------+-------------------------+
| Name               | Bytes  | Value              |          Remarks        |
+--------------------+--------+--------------------+-------------------------+
| Flags              | 1-0    |                    |                         |
| Opcode             | 2-3    | 0x0632             | Command opcode          |
| Datalen            | 4-5    | 0                  | No external buffer.     |
| Return value       | 6-7    |                    | Return value.           |
| Cookie High        | 8-11   | Cookie             |                         |
| Cookie Low         | 12-15  | Cookie             |                         |
| Sensor             | 16     |                    | 0x00: Internal temp     |
|                    |        |                    | 0x01-0xFF: Reserved.    |
| Format             | 17     | Requested response | Only 0x00 is supported. |
|                    |        | format             | 0x01-0xFF: Reserved.    |
| Reserved           | 18-23  |                    |                         |
| Data Address high  | 24-27  | Response buffer    |                         |
|                    |        | address            |                         |
| Data Address low   | 28-31  | Response buffer    |                         |
|                    |        | address            |                         |
+--------------------+--------+--------------------+-------------------------+

Get Sensor Reading Response (Opcode: 0x0632)

+--------------------+--------+--------------------+-------------------------+
| Name               | Bytes  | Value              |          Remarks        |
+--------------------+--------+--------------------+-------------------------+
| Flags              | 1-0    |                    |                         |
| Opcode             | 2-3    | 0x0632             | Command opcode          |
| Datalen            | 4-5    | 0                  | No external buffer      |
| Return value       | 6-7    |                    | Return value.           |
|                    |        |                    | EINVAL: Invalid         |
|                    |        |                    | parameters              |
|                    |        |                    | ENOENT: Unsupported     |
|                    |        |                    | sensor                  |
|                    |        |                    | EIO: Sensor access      |
|                    |        |                    | error                   |
| Cookie High        | 8-11   | Cookie             |                         |
| Cookie Low         | 12-15  | Cookie             |                         |
| Sensor Reading     | 16-23  |                    | Format of the reading   |
|                    |        |                    | is dependent on request |
| Data Address high  | 24-27  | Response buffer    |                         |
|                    |        | address            |                         |
| Data Address low   | 28-31  | Response buffer    |                         |
|                    |        | address            |                         |
+--------------------+--------+--------------------+-------------------------+

Sensor Reading for Sensor 0x00 (Internal Chip Temperature):

+--------------------+--------+--------------------+-------------------------+
| Name               | Bytes  | Value              |          Remarks        |
+--------------------+--------+--------------------+-------------------------+
| Thermal Sensor     | 0      |                    | Reading in degrees      |
| reading            |        |                    | Celsius. Signed int8    |
| Warning High       | 1      |                    | Warning High threshold  |
| threshold          |        |                    | in degrees Celsius.     |
|                    |        |                    | Unsigned int8.          |
|                    |        |                    | 0xFF when unsupported   |
| Critical High      | 2      |                    | Critical High threshold |
| threshold          |        |                    | in degrees Celsius.     |
|                    |        |                    | Unsigned int8.          |
|                    |        |                    | 0xFF when unsupported   |
| Fatal High         | 3      |                    | Fatal High threshold    |
| threshold          |        |                    | in degrees Celsius.     |
|                    |        |                    | Unsigned int8.          |
|                    |        |                    | 0xFF when unsupported   |
| Reserved           | 4-7    |                    |                         |
+--------------------+--------+--------------------+-------------------------+

Driver provides current reading from HW as well as device specific
thresholds for thermal alarm (Warning, Critical, Fatal) events.

$ sensors

Output
=========================================================
ice-pci-b100
Adapter: PCI adapter
temp1:        +62.0°C  (high = +95.0°C, crit = +105.0°C)
                       (emerg = +115.0°C)

Tested on Intel Corporation Ethernet Controller E810-C for SFP

Co-developed-by: Marcin Domagala <marcinx.domagala@intel.com>
Signed-off-by: Marcin Domagala <marcinx.domagala@intel.com>
Co-developed-by: Eric Joyner <eric.joyner@intel.com>
Signed-off-by: Eric Joyner <eric.joyner@intel.com>
Reviewed-by: Marcin Szycik <marcin.szycik@linux.intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Konrad Knitter <konrad.knitter@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'net-sched-act_api-contiguous-action-arrays'

Pedro Tammela says:

====================
net/sched: act_api: contiguous action arrays

When dealing with action arrays in act_api it's natural to ask if they
are always contiguous (no NULL pointers in between). Yes, they are in
all cases so far, so make use of the already present tcf_act_for_each_action
macro to explicitly document this assumption.

There was an instance where it was not, but it was refactorable (patch 2)
to make the array contiguous.

v1->v2:
- Respin
- Added Jamal's acked-by
====================

Link: https://lore.kernel.org/r/20231201175015.214214-1-pctammela@mojatatu.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net/sched: act_api: use tcf_act_for_each_action in tcf_idr_insert_many

The actions array is contiguous, so stop processing whenever a NULL
is found. This is already the assumption for tcf_action_destroy[1],
which is called from tcf_actions_init.

[1] https://elixir.bootlin.com/linux/v6.7-rc3/source/net/sched/act_api.c#L1115

Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net/sched: act_api: stop loop over ops array on NULL in tcf_action_init

The ops array is contiguous, so stop processing whenever a NULL is found

Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net/sched: act_api: avoid non-contiguous action array

In tcf_action_add, when putting the reference for the bound actions
it assigns NULLs to just created actions passing a non contiguous
array to tcf_action_put_many.
Refactor the code so the actions array is always contiguous.

Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net/sched: act_api: use tcf_act_for_each_action

Use the auxiliary macro tcf_act_for_each_action in all the
functions that expect a contiguous action array

Suggested-by: Marcelo Ricardo Leitner <mleitner@redhat.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'doc-update-bridge-doc'

Hangbin Liu says:

====================
Doc: update bridge doc

The current bridge kernel doc is too old. It only pointed to the
linuxfoundation wiki page which lacks of the new features.

Here let's start the new bridge document and put all the bridge info
so new developers and users could catch up the last bridge status soon.

v3 -> v4:
- Patch01: Reference and borrow definitions from the IEEE 802.1Q-2022 standard
           for bridge (Stephen Hemminger)
- Patch04: Remind that kAPI is unstable. Add back sysfs part, but only note
           that sysfs is deprecated. (Stephen Hemminger, Florian Fainelli)
- Patch05: Mention the RSTP and IEEE 802.1D developing info. (Stephen Hemminger)
- Some other grammar fixes.

v2 -> v3:
- Split the bridge doc update and adding kAPI/uAPI field to 2 part (Nikolay Aleksandrov)
- Update bridge and bridge enum descriptions (Nikolay Aleksandrov)
- Add user space stp help for STP doc (Vladimir Oltean)

v1 -> v2:
- Update bridge and bridge port enum descriptions (Vladimir Oltean)

RFCv3 -> v1:
- Fix up various typos, grammar and technical issues (Nikolay Aleksandrov)

RFCv2 -> RFCv3:
- Update netfilter part (Florian Westphal)
- Break the one large patch in to multiparts for easy reviewing. Please tell
  me if I break it too much.. (Nikolay Aleksandrov)
- Update the description of each enum and doc (Nikolay Aleksandrov)
- Add more descriptions for STP/Multicast/VLAN.

RFCv1 -> RFCv2:
- Drop the python tool that generate iproute man page from kernel doc
====================

Link: https://lore.kernel.org/r/20231201081951.1623069-1-liuhangbin@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

docs: bridge: add other features

Add some features that are not appropriate for the existing section to
the "Others" part of the bridge document.

Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

docs: bridge: add netfilter doc

Add netfilter part for bridge document.

Reviewed-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

docs: bridge: add switchdev doc

Add switchdev part for bridge document.

Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

docs: bridge: add multicast doc

Add multicast part for bridge document.

Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

docs: bridge: add VLAN doc

Add VLAN part for bridge document.

Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

docs: bridge: add STP doc

Add STP part for bridge document.

Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

docs: bridge: Add kAPI/uAPI fields

Add kAPI/uAPI field for bridge doc. Update struct net_bridge_vlan
comments to fix doc build warning.

Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: bridge: add document for IFLA_BRPORT enum

Add document for IFLA_BRPORT enum so we can use it in
Documentation/networking/bridge.rst.

Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: bridge: add document for IFLA_BR enum

Add document for IFLA_BR enum so we can use it in
Documentation/networking/bridge.rst.

Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

docs: bridge: update doc format to rst

The current bridge kernel doc is too old. It only pointed to the
linuxfoundation wiki page which lacks of the new features.

Here let's start the new bridge document and put all the bridge info
so new developers and users could catch up the last bridge status soon.

In this patch, Convert the doc to rst format. Add bridge brief introduction,
FAQ and contact info.

Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'net-stmmac-est-implementation'

Rohan G Thomas says:

====================
net: stmmac: EST implementation

This patchset extends EST interrupt handling support to DWXGMAC IP
followed by refactoring of EST implementation. Added a separate
module for EST and moved all EST related functions to the new module.

Also added support for EST cycle-time-extension.
====================

Link: https://lore.kernel.org/r/20231201055252.1302-1-rohan.g.thomas@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: stmmac: Add support for EST cycle-time-extension

Add support for cycle-time-extension. TER GCL-register needs to be
updated with the cycle-time-extension. Width of TER register is EST
time interval width + 7 bits.

Signed-off-by: Rohan G Thomas <rohan.g.thomas@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20231201055252.1302-4-rohan.g.thomas@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: stmmac: Refactor EST implementation

Refactor EST implementation by moving common code for DWMAC4 and
DWXGMAC IPs into a separate EST module. EST implementation for DWMAC4
and DWXGMAC differs only for CSR base address, PTOV field offset
width, and PTOV clock multiplier value.

Thanks, Serge Semin and Jakub Kicinski for the suggestions on
refactoring EST implementation into a separate EST module.

Signed-off-by: Rohan G Thomas <rohan.g.thomas@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20231201055252.1302-3-rohan.g.thomas@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: stmmac: xgmac: EST interrupts handling

Enabled the following EST related interrupts:
  1) Constant Gate Control Error (CGCE)
  2) Head-of-Line Blocking due to Scheduling (HLBS)
  3) Head-of-Line Blocking due to Frame Size (HLBF)
  4) Base Time Register error (BTRE)
  5) Switch to S/W owned list Complete (SWLC)
Also, add EST errors into the ethtool statistic.

The commit e49aa315cb01 ("net: stmmac: EST interrupts handling and
error reporting") and commit 9f298959191b ("net: stmmac: Add EST
errors into ethtool statistic") add EST interrupts handling and error
reporting support to DWMAC4 core. This patch enables the same support
for XGMAC.

Signed-off-by: Rohan G Thomas <rohan.g.thomas@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20231201055252.1302-2-rohan.g.thomas@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: ti: davinci_mdio: Update K3 SoCs list for errata i2329

The errata i2329 affects all the currently available silicon revisions of
AM62x, AM64x, AM65x, J7200, J721E and J721S2. So remove the revision
string from the SoC list.

The silicon revisions affected by the errata i2329 can be found under
the MDIO module in the "Advisories by Modules" section of each
SoC errata document listed below

AM62x: https://www.ti.com/lit/er/sprz487c/sprz487c.pdf
AM64X: https://www.ti.com/lit/er/sprz457g/sprz457g.pdf
AM65X: https://www.ti.com/lit/er/sprz452i/sprz452i.pdf
J7200: https://www.ti.com/lit/er/sprz491d/sprz491d.pdf
J721E: https://www.ti.com/lit/er/sprz455d/sprz455d.pdf
J721S2: https://www.ti.com/lit/er/sprz530b/sprz530b.pdf

Signed-off-by: Ravi Gunasekaran <r-gunasekaran@ti.com>
Reviewed-by: Nishanth Menon <nm@ti.com>
Link: https://lore.kernel.org/r/20231201132033.29576-1-r-gunasekaran@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'introduce-queue-and-napi-support-in-netdev-genl-was-introduce-napi-queues-support'

Amritha Nambiar says:

====================
Introduce queue and NAPI support in netdev-genl (Was: Introduce NAPI queues support)

Add the capability to export the following via netdev-genl interface:
- queue information supported by the device
- NAPI information supported by the device

Introduce support for associating queue and NAPI instance.
Extend the netdev_genl generic netlink family for netdev
with queue and NAPI data.

The queue parameters exposed are:
- queue index
- queue type
- ifindex
- NAPI id associated with the queue

Additional rx and tx queue parameters can be exposed in follow up
patches by stashing them in netdev queue structures. XDP queue type
can also be supported in future.

The NAPI fields exposed are:
- NAPI id
- NAPI device ifindex
- Interrupt number associated with the NAPI instance
- PID for the NAPI thread

This series only supports 'get' ability for retrieving
certain queue and NAPI attributes. The 'set' ability for
configuring queue and associated NAPI instance via netdev-genl
will be submitted as a separate patch series.

Previous discussion at:
https://lore.kernel.org/netdev/c8476530638a5f4381d64db0e024ed49c2db3b02.camel@gmail.com/T/#m00999652a8b4731fbdb7bf698d2e3666c65a60e7

$ ./cli.py --spec netdev.yaml --do queue-get  --json='{"ifindex": 12, "id": 0, "type": 0}'
{'id': 0, 'ifindex': 12, 'napi-id': 593, 'type': 'rx'}

$ ./cli.py --spec netdev.yaml  --do queue-get --json='{"ifindex": 12, "id": 0, "type": 1}'
{'id': 0, 'ifindex': 12, 'napi-id': 593, 'type': 'tx'}

$ ./cli.py --spec netdev.yaml  --dump queue-get --json='{"ifindex": 12}'
[{'id': 0, 'ifindex': 12, 'napi-id': 593, 'type': 'rx'},
 {'id': 1, 'ifindex': 12, 'napi-id': 594, 'type': 'rx'},
 {'id': 2, 'ifindex': 12, 'napi-id': 595, 'type': 'rx'},
 {'id': 3, 'ifindex': 12, 'napi-id': 596, 'type': 'rx'},
 {'id': 0, 'ifindex': 12, 'napi-id': 593, 'type': 'tx'},
 {'id': 1, 'ifindex': 12, 'napi-id': 594, 'type': 'tx'},
 {'id': 2, 'ifindex': 12, 'napi-id': 595, 'type': 'tx'},
 {'id': 3, 'ifindex': 12, 'napi-id': 596, 'type': 'tx'}]

$ ./cli.py --spec netdev.yaml --do napi-get --json='{"id": 593}'
{'id': 593, 'ifindex': 12, 'irq': 291, 'pid': 3727}

$ ./cli.py --spec netdev.yaml --dump napi-get --json='{"ifindex": 12}'
[{'id': 596, 'ifindex': 12, 'irq': 294, 'pid': 3724},
 {'id': 595, 'ifindex': 12, 'irq': 293, 'pid': 3725},
 {'id': 594, 'ifindex': 12, 'irq': 292, 'pid': 3726},
 {'id': 593, 'ifindex': 12, 'irq': 291, 'pid': 3727}]
====================

Link: https://lore.kernel.org/r/170147307026.5260.9300080745237900261.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

eth: bnxt: link NAPI instances to queues and IRQs

Make bnxt compatible with the newly added netlink queue GET APIs.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/170147336340.5260.6773000274196548907.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev-genl: Add PID for the NAPI thread

In the threaded NAPI mode, expose the PID of the NAPI thread.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147335818.5260.10253384006102593087.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev-genl: spec: Add PID in netdev netlink YAML spec

Add support in netlink spec(netdev.yaml) for PID of the
NAPI thread. Add code generated from the spec.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147335301.5260.11872351477120434501.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: Add NAPI IRQ support

Add support to associate the interrupt vector number for a
NAPI instance.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147334728.5260.13221803396905901904.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev-genl: spec: Add irq in netdev netlink YAML spec

Add support in netlink spec(netdev.yaml) for interrupt number
among the NAPI attributes. Add code generated from the spec.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147334210.5260.18178387869057516983.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev-genl: Add netlink framework functions for napi

Implement the netdev netlink framework functions for
napi support. The netdev structure tracks all the napi
instances and napi fields. The napi instances and associated
parameters can be retrieved this way.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147333637.5260.14807433239805550815.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev-genl: spec: Extend netdev netlink spec in YAML for NAPI

Add support in netlink spec(netdev.yaml) for napi related information.
Add code generated from the spec.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147333119.5260.7050639053080529108.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev-genl: Add netlink framework functions for queue

Implement the netdev netlink framework functions for
exposing queue information.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147332603.5260.7982559672617639065.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ice: Add support in the driver for associating queue with napi

After the napi context is initialized, map the napi instance
with the queue/queue-set on the corresponding irq line.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147332060.5260.13310934657151560599.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: Add queue and napi association

Add the napi pointer in netdev queue for tracking the napi
instance for each queue. This achieves the queue<->napi mapping.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147331483.5260.15723438819994285695.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev-genl: spec: Extend netdev netlink spec in YAML for queue

Add support in netlink spec(netdev.yaml) for queue information.
Add code generated from the spec.

Note: The "queue-type" attribute takes values 0 and 1 for rx
and tx queue type respectively.

Signed-off-by: Amritha Nambiar <amritha.nambiar@intel.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Link: https://lore.kernel.org/r/170147330963.5260.2576294626647300472.stgit@anambiarhost.jf.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'bnxt_en-support-new-5760x-p7-devices'

Michael Chan says:

====================
bnxt_en: Support new 5760X P7 devices

This series completes the basic support for the new 5760X P7 devices
with new PCI IDs added in the last patch.

Thie first patch fixes a backing store issue introduced in the last
patchset last week.  The 2nd patch is the new firmware interface
required to support the new chips.  The next few patches are doorbell
changes, refactoring, and new hardware interface structures.  New
changes to support packet reception including TPA are added in patch 10.
The next 4 patches are ethernet link related changes to support the
new chip.
====================

Link: https://lore.kernel.org/r/20231201223924.26955-1-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Add 5760X (P7) PCI IDs

Now with basic support for the new chip family, add the PCI IDs of the
new devices.

Reviewed-by: Andy Gospodarek <gospo@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-16-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Report the new ethtool link modes in the new firmware interface

Add new look up entries to convert the new supported speeds, advertised
speeds, etc to ethtool link modes.

Reviewed-by: Damodharam Ammepalli <damodharam.ammepalli@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-15-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Support force speed using the new HWRM fields

Modify bnxt_force_link_speed() to support the new speeds stored in
link_info->support_speeds2, including the new 400G speed.

Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-14-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Support new firmware link parameters

Newer firmware supporting PAM4 112Gbps speeds use new parameters in
firmware message structures.  Detect the new firmware capability and
add basic logic to report and store these new fields.

Reviewed-by: Hongguang Gao <hongguang.gao@broadcom.com>
Reviewed-by: Damodharam Ammepalli <damodharam.ammepalli@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-13-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Refactor ethtool speeds logic

Add helper functions to refactor the logic that converts firmware
speed masks to ethtool speeds.  Pass the phy_flags to
bnxt_get_ethtool_speeds() and the call chain.  The refactoring and the
phy_flags will be needed when adding support for the new speeds in the
next patches.

Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-12-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Add support for new RX and TPA_START completion types for P7

These new completion types are supported on the new P7 chips.
These new types have commonalities with the legacy types.  After
the refactoring, we mainly have to add new functions to handle the
the new meta data formats and the RX hash information in the new
types.

Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-11-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Refactor and refine bnxt_tpa_start() and bnxt_tpa_end().

Refactor bnxt_tpa_start() by adding bnxt_tpa_metadata() to gather the
metadata from the TPA_START completion.  This makes it easier to
support the new P7 chip which has a modified TPA_START completion
structure with different metadata formats.  We also add vlan_valid
and cfa_code_valid fields to the bnxt_tpa_info structure so that the
VLAN and VF rep logic can be common for all chips.  The VLAN metadata
is now collected in bnxt_tpa_start() only when it is valid and the
vlan_valid field will be set.  bnxt_tpa_end() can now use common VLAN
logic for all chips.

Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-10-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Refactor RX VLAN acceleration logic.

Refactor the logic in the RX path that checks for the accelerated VLAN
tag by adding a new function.  This will make it easier to support
the new receive logic on P7 chips.

Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-9-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Add new P7 hardware interface definitions

Add new RX, TX, and TPA hardware interface structures and macros for the
P7 chips.

Reviewed-by: Andy Gospodarek <gospo@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-8-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Refactor RSS capability fields

Add a new rss_cap field in the per device struct bnxt and move all
the RSS capability fields there.  It will be easier to add new RSS
capabilities for the new P7 chips.

Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-7-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Implement the new toggle bit doorbell mechanism on P7 chips

The new chip family passes the Toggle bits to the driver in the NQE
notification.  The driver now stores this value and sends it back to
hardware when it re-arms the RX and TX CQs.  Together with the earlier
patch that guarantees the driver will only re-arm the CQ at the end of
NAPI polling if it has seen a new NQE, this method allows the hardware
to detect any dropped doorbells.

Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-6-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Consolidate DB offset calculation

The doorbell offset on P5 chips is hard coded.  On the new P7 chips,
it is returned by the firmware.  Simplify the logic that determines
this offset and store it in a new db_offset field in struct bnxt.
Also, provide this offset to the RoCE driver in struct bnxt_en_dev.

Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Hongguang Gao <hongguang.gao@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-5-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Define basic P7 macros

Repurpose the BNXT_FLAG_CHIP_SR2 flag by renaming it to
BNXT_FLAG_CHIP_P7 since the SR2 chip never went to production.  The SR2
statictics structure is also renamed for the P7 chip.  Define the basic
P7 doorbell bits (Epoch. Toggle, etc) and implement the Epoch bit
logic.  The next higher bit beyond the legal doorbell mask is the
Epoch bit used for doorbells on P7 chips.  This bit is used by the
chip to detect dropped doorbells.

The 57608 chip ID belonging to the P7 family is also defined.  Note
that the PCI ID is not added until the last patch in the series.

Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-4-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Update firmware interface to 1.10.3.15

This updated interface supports the new 5760X P7 chip family.  It has
the changes to support the new link speeds/modes and other changes
for the basic L2 features.

Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20231201223924.26955-3-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>