octeontx2-pf: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Subbaraya Sundeep <sbhatta@marvell.com>
Link: https://patch.msgid.link/20240901112803.212753-15-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

gianfar: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Link: https://patch.msgid.link/20240901112803.212753-14-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: enetc: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Wei Fang <wei.fang@nxp.com>
Link: https://patch.msgid.link/20240901112803.212753-13-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: fec: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Wei Fang <wei.fang@nxp.com>
Link: https://patch.msgid.link/20240901112803.212753-12-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Jijie Shao <shaojijie@huawei.com>
Link: https://patch.msgid.link/20240901112803.212753-11-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: rtsn: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Link: https://patch.msgid.link/20240901112803.212753-10-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: renesas: rswitch: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Link: https://patch.msgid.link/20240901112803.212753-9-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ravb: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Link: https://patch.msgid.link/20240901112803.212753-8-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ionic: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Shannon Nelson <shannon.nelson@amd.com>
Reviewed-by: Brett Creeley <brett.creeley@amd.com>
Link: https://patch.msgid.link/20240901112803.212753-7-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tsnep: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Gerhard Engleder <gerhard@engleder-embedded.com>
Link: https://patch.msgid.link/20240901112803.212753-6-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

can: peak_usb: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Marc Kleine-Budde <mkl@pengutronix.de>
Link: https://patch.msgid.link/20240901112803.212753-5-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

can: peak_canfd: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Marc Kleine-Budde <mkl@pengutronix.de>
Link: https://patch.msgid.link/20240901112803.212753-4-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

can: dev: Remove setting of RX software timestamp

The responsibility for reporting of RX software timestamp has moved to
the core layer (see __ethtool_get_ts_info()), remove usage from the
device drivers.

Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Reviewed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Marc Kleine-Budde <mkl@pengutronix.de>
Link: https://patch.msgid.link/20240901112803.212753-3-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ethtool: RX software timestamp for all

All devices support SOF_TIMESTAMPING_RX_SOFTWARE by virtue of
net_timestamp_check() being called in the device independent code.

Move the responsibility of reporting SOF_TIMESTAMPING_RX_SOFTWARE and
SOF_TIMESTAMPING_SOFTWARE, and setting PHC index to -1 to the core.
Device drivers no longer need to use them.

Suggested-by: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
Link: https://lore.kernel.org/netdev/661550e348224_23a2b2294f7@willemb.c.googlers.com.notmuch/
Co-developed-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Marc Kleine-Budde <mkl@pengutronix.de>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/20240901112803.212753-2-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch '1GbE' of git://git./linux/kernel/git/tnguy/next-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2024-08-30 (igc, e1000e, i40e)

This series contains updates to igc, e1000e, and i40 drivers.

Kurt Kanzenbach adds support for MQPRIO offloads and stops unintended,
excess interrupts on igc.

Sasha adds reporting of EEE (Energy Efficient Ethernet) ability and
moves a register define to a better suited file for igc.

Vitaly stops reporting errors on shutdown and suspend as they are not
fatal for e1000e.

Alex adds reporting of EEE to i40e.

* '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue:
  i40e: Add Energy Efficient Ethernet ability for X710 Base-T/KR/KX cards
  e1000e: avoid failing the system during pm_suspend
  igc: Move the MULTI GBT AN Control Register to _regs file
  igc: Add Energy Efficient Ethernet ability
  igc: Get rid of spurious interrupts
  igc: Add MQPRIO offload support
====================

Link: https://patch.msgid.link/20240830210451.2375215-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge tag 'ieee802154-for-net-2024-09-01' of git://git./linux/kernel/git/wpan/wpan

Stefan Schmidt says:

====================
pull-request: ieee802154 for net 2024-09-01

Simon Horman catched two typos in our headers. No functional change.

* tag 'ieee802154-for-net-2024-09-01' of git://git.kernel.org/pub/scm/linux/kernel/git/wpan/wpan:
  ieee802154: Correct spelling in nl802154.h
  mac802154: Correct spelling in mac802154.h
====================

Link: https://patch.msgid.link/20240901184213.2303047-1-stefan@datenfreihafen.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ioam6: improve checks on user data

This patch improves two checks on user data.

The first one prevents bit 23 from being set, as specified by RFC 9197
(Sec 4.4.1):

  Bit 23    Reserved; MUST be set to zero upon transmission and be
            ignored upon receipt.  This bit is reserved to allow for
            future extensions of the IOAM Trace-Type bit field.

The second one checks that the tunnel destination address !=
IPV6_ADDR_ANY, just like we already do for the tunnel source address.

Signed-off-by: Justin Iurman <justin.iurman@uliege.be>
Link: https://patch.msgid.link/20240830191919.51439-1-justin.iurman@uliege.be
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'cleanup-chelsio-driver-declarations'

Yue Haibing says:

====================
cleanup chelsio driver declarations

v2: correct patch 3 commit citation
====================

Link: https://patch.msgid.link/20240830093338.3742315-1-yuehaibing@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

cxgb: Remove unused declarations

These functions were never implenmented since introduction in
commit 8199d3a79c22 ("[PATCH] A new 10GB Ethernet Driver by Chelsio
Communications")

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

cxgb4: Remove unused declarations

Commit e2d14b42c25c ("cxgb4: Remove WOL get/set ethtool support") removed
t4_wol_magic_enable() and t4_wol_pat_enable() but leave declarations.

Commit 02d805dc5fe3 ("cxgb4: use new fw interface to get the VIN and smt
index") leave behind cxgb4_tp_smt_idx().

cxgb4_dcb_set_caps() is never implemented and used since introduction in
commit 76bcb31efc06 ("cxgb4 : Add DCBx support codebase and dcbnl_ops").

Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

cxgb3: Remove unused declarations

Commit 4d22de3e6cc4 ("Add support for the latest 1G/10G Chelsio adapter,
T3.") declared but never implemented these.

Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

selftests: netfilter: nft_queue.sh: fix spurious timeout on debug kernel

The sctp selftest is very slow on debug kernels.

Its possible that the nf_queue listener program exits due to timeout
before first sctp packet is processed.

In this case socat hangs until script times out.
Fix this by removing the -t option where possible and kill the test
program once the file transfer/socat has exited.

-t sets SO_RCVTIMEO, its inteded for the 'ping' part of the selftest
where we want to make sure that packets get reinjected properly without
skipping a second queue request.

While at it, add a helper to compare the (binary) files instead of diff.
The 'diff' part was copied from a another sub-test that compares text.

Let helper dump file sizes on error so we can see the progress made.

Tested on an old 2010-ish box with a debug kernel and 100 iterations.

This is a followup to the earlier filesize reduction change.

Reported-by: Jakub Kicinski <kuba@kernel.org>
Closes: https://lore.kernel.org/netdev/20240829080109.GB30766@breakpoint.cc/
Fixes: 0a8b08c554da ("selftests: netfilter: nft_queue.sh: reduce test file size for debug build")
Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://patch.msgid.link/20240830092254.8029-1-fw@strlen.de
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'net-simplified-with-scoped-function'

Jinjie Ruan says:

====================
net: Simplified with scoped function

Simplify with scoped for each OF child loop, as well as dev_err_probe().

Changes in v4:
- Drop the fix patch and __free() patch.
- Rebased on the fix patch has been stripped out.
- Remove the extra parentheses.
- Ensure Signed-off-by: should always be last.
- Add Reviewed-by.
- Update the cover letter commit message.

Changes in v3:
- Sort the variables, longest first, shortest last.
- Add Reviewed-by.

Changes in v2:
- Subject prefix: next -> net-next.
- Split __free() from scoped for each OF child loop clean.
- Fix use of_node_put() instead of __free() for the 5th patch.
====================

Link: https://patch.msgid.link/20240830031325.2406672-1-ruanjinjie@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: bcmasp: Simplify with scoped for each OF child loop

Use scoped for_each_available_child_of_node_scoped() when
iterating over device nodes to make code a bit simpler.

Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Justin Chen <justin.chen@broadcom.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dsa: microchip: Use scoped function to simplfy code

Avoids the need for manual cleanup of_node_put() in early exits
from the loop by using for_each_available_child_of_node_scoped().

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: mv643xx_eth: Simplify with scoped for each OF child loop

Use scoped for_each_available_child_of_node_scoped() when iterating
over device nodes to make code a bit simpler.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: mdio: mux-mmioreg: Simplified with dev_err_probe()

Use the dev_err_probe() helper to simplify code.

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: mdio: mux-mmioreg: Simplified with scoped function

Avoids the need for manual cleanup of_node_put() in early exits
from the loop by using for_each_available_child_of_node_scoped().

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: phy: Use for_each_available_child_of_node_scoped()

Avoid need to manually handle of_node_put() by using
for_each_available_child_of_node_scoped(), which can simplfy code.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dsa: realtek: Use for_each_child_of_node_scoped()

Avoid need to manually handle of_node_put() by using
for_each_child_of_node_scoped(), which can simplfy code.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: stmmac: dwmac-sun8i: Use for_each_child_of_node_scoped()

Avoid need to manually handle of_node_put() by using
for_each_child_of_node_scoped(), which can simplfy code.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'netdev_features-start-cleaning-netdev_features_t-up'

Alexander Lobakin says:

====================
netdev_features: start cleaning netdev_features_t up

NETDEV_FEATURE_COUNT is currently 64, which means we can't add any new
features as netdev_features_t is u64.
As per several discussions, instead of converting netdev_features_t to
a bitmap, which would mean A LOT of changes, we can try cleaning up
netdev feature bits.
There's a bunch of bits which don't really mean features, rather device
attributes/properties that can't be changed via Ethtool in any of the
drivers. Such attributes can be moved to netdev private flags without
losing any functionality.

Start converting some read-only netdev features to private flags from
the ones that are most obvious, like lockless Tx, inability to change
network namespace etc. I was able to reduce NETDEV_FEATURE_COUNT from
64 to 60, which mean 4 free slots for new features. There are obviously
more read-only features to convert, such as highDMA, "challenged VLAN",
HSR (4 bits) - this will be done in subsequent series.

Please note that currently netdev features are not uAPI/ABI by any means.
Ethtool passes their names and bits to the userspace separately and there
are no hardcoded names/bits in the userspace, so that new Ethtool could
work on older kernels and vice versa.
This, however, isn't true for Ethtools < 3.4. I haven't changed the bit
positions of the already existing features and instead replaced the freed
bits with stubs. But it's anyway theoretically possible that Ethtools
older than 2011 will break. I hope no currently supported distros supply
such an ancient version.
Shell scripts also most likely won't break since the removed bits were
always read-only, meaning nobody would try touching them from a script.
====================

Link: https://patch.msgid.link/20240829123340.789395-1-aleksander.lobakin@intel.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

netdev_features: remove NETIF_F_ALL_FCOE

NETIF_F_ALL_FCOE is used only in vlan_dev.c, 2 times. Now that it's only
2 bits, open-code it and remove the definition from netdev_features.h.

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

netdev_features: convert NETIF_F_FCOE_MTU to dev->fcoe_mtu

Ability to handle maximum FCoE frames of 2158 bytes can never be changed
and thus more of an attribute, not a toggleable feature.
Move it from netdev_features_t to "cold" priv flags (bitfield bool) and
free yet another feature bit.

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

netdev_features: convert NETIF_F_NETNS_LOCAL to dev->netns_local

"Interface can't change network namespaces" is rather an attribute,
not a feature, and it can't be changed via Ethtool.
Make it a "cold" private flag instead of a netdev_feature and free
one more bit.

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

netdev_features: convert NETIF_F_LLTX to dev->lltx

NETIF_F_LLTX can't be changed via Ethtool and is not a feature,
rather an attribute, very similar to IFF_NO_QUEUE (and hot).
Free one netdev_features_t bit and make it a "hot" private flag.

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

netdevice: convert private flags > BIT(31) to bitfields

Make dev->priv_flags `u32` back and define bits higher than 31 as
bitfield booleans as per Jakub's suggestion. This simplifies code
which accesses these bits with no optimization loss (testb both
before/after), allows to not extend &netdev_priv_flags each time,
but also scales better as bits > 63 in the future would only add
a new u64 to the structure with no complications, comparing to
that extending ::priv_flags would require converting it to a bitmap.
Note that I picked `unsigned long :1` to not lose any potential
optimizations comparing to `bool :1` etc.

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dsa: vsc73xx: implement FDB operations

This commit introduces implementations of three functions:
.port_fdb_dump
.port_fdb_add
.port_fdb_del

The FDB database organization is the same as in other old Vitesse chips:
It has 2048 rows and 4 columns (buckets). The row index is calculated by
the hash function 'vsc73xx_calc_hash' and the FDB entry must be placed
exactly into row[hash]. The chip selects the bucket number by itself.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
Link: https://patch.msgid.link/20240827123938.582789-1-paweldembicki@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge tag 'linux-can-next-for-6.12-20240830' of git://git./linux/kernel/git/mkl/linux-can-next

Marc Kleine-Budde says:

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

The first patch is by Duy Nguyen and document the R-Car V4M support in
the rcar-canfd DT bindings.

Frank Li's patch converts the microchip,mcp251x.txt DT bindings
documentation to yaml.

A patch by Zhang Changzhong update a comment in the j1939 CAN
networking stack.

Stefan Mätje's patch updates the CAN configuration netlink code, so
that the bit timing calculation doesn't work on stale
can_priv::ctrlmode data.

Martin Jocic contributes a patch for the kvaser_pciefd driver to
convert some ifdefs into if (IS_ENABLED()).

The last patch is by Yan Zhen and simplifies the probe() function of
the kvaser USB driver by using dev_err_probe().

* tag 'linux-can-next-for-6.12-20240830' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next:
  can: kvaser_usb: Simplify with dev_err_probe()
  can: kvaser_pciefd: Use IS_ENABLED() instead of #ifdef
  can: netlink: avoid call to do_set_data_bittiming callback with stale can_priv::ctrlmode
  can: j1939: use correct function name in comment
  dt-bindings: can: convert microchip,mcp251x.txt to yaml
  dt-bindings: can: renesas,rcar-canfd: Document R-Car V4M support
====================

Link: https://patch.msgid.link/20240830214406.1605786-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

r8169: add support for RTL8126A rev.b

Add support for RTL8126A rev.b. Its XID is 0x64a. It is basically
based on the one with XID 0x649, but with different firmware file.

Signed-off-by: ChunHao Lin <hau@realtek.com>
Reviewed-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://patch.msgid.link/20240830021810.11993-1-hau@realtek.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev-genl: Set extack and fix error on napi-get

In commit 27f91aaf49b3 ("netdev-genl: Add netlink framework functions
for napi"), when an invalid NAPI ID is specified the return value
-EINVAL is used and no extack is set.

Change the return value to -ENOENT and set the extack.

Before this commit:

$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
                          --do napi-get --json='{"id": 451}'
Netlink error: Invalid argument
nl_len = 36 (20) nl_flags = 0x100 nl_type = 2
error: -22

After this commit:

$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
                         --do napi-get --json='{"id": 451}'
Netlink error: No such file or directory
nl_len = 44 (28) nl_flags = 0x300 nl_type = 2
error: -2
extack: {'bad-attr': '.id'}

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Joe Damato <jdamato@fastly.com>
Link: https://patch.msgid.link/20240831121707.17562-1-jdamato@fastly.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: stmmac: drop the ethtool begin() callback

This callback doesn't seem to serve much purpose, and prevents things
like:

    - systemd.link files from disabling autonegotiation
    - carrier detection in NetworkManager
    - any ethtool setting

prior to userspace bringing the link up.

The only fear I can think of is accessing unclocked resources due to
pm_runtime, but ethtool ioctls handle that as of commit
f32a21376573 ("ethtool: runtime-resume netdev parent before ethtool ioctl ops")

Reviewed-by: Dmitry Dolenko <d.dolenko@metrotek.ru>
Tested-by: Dmitry Dolenko <d.dolenko@metrotek.ru>
Signed-off-by: Andrew Halaney <ahalaney@redhat.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'octeontx2-af-cpt-update'

Srujana Challa says:

====================
octeontx2-af: update CPT block for CN10KB and CN10KA B0

This commit addresses two key updates for the CN10KB and CN10KA B0:

1. The number of FLT interrupt vectors has been reduced to 2 on CN10KB.
The code is updated to reflect this change across the CN10K series.
2. The maximum CPT credits that RX can use are now configurable through
a hardware CSR on CN10KA B0. This patch sets the default value to optimize
peak performance, aligning it with other chip versions.

v2:
- Addressed the review comments.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

octeontx2-af: configure default CPT credits for CN10KA B0

The maximum CPT credits that RXC can use are now configurable on CN10KA B0
through a hardware CSR. This patch sets the default value to optimize peak
performance, aligning it with other chip versions.

Signed-off-by: Srujana Challa <schalla@marvell.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

octeontx2-af: avoid RXC register access for CN10KB

This patch modifies the driver to prevent access to RXC hardware
registers on the CN10KB, as RXC is not available on this chip.

Signed-off-by: Srujana Challa <schalla@marvell.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

octeontx2-af: use dynamic interrupt vectors for CN10K

This patch updates the driver to use a dynamic number of vectors instead
of a hard-coded value. This change accommodates the CN10KB, which has 2
vectors, unlike the previously supported chips that have 3 vectors.

Signed-off-by: Srujana Challa <schalla@marvell.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'unmask-dscp-bits'

Ido Schimmel says:

====================
Unmask upper DSCP bits - part 2

tl;dr - This patchset continues to unmask the upper DSCP bits in the
IPv4 flow key in preparation for allowing IPv4 FIB rules to match on
DSCP. No functional changes are expected. Part 1 was merged in commit
("Merge branch 'unmask-upper-dscp-bits-part-1'").

The TOS field in the IPv4 flow key ('flowi4_tos') is used during FIB
lookup to match against the TOS selector in FIB rules and routes.

It is currently impossible for user space to configure FIB rules that
match on the DSCP value as the upper DSCP bits are either masked in the
various call sites that initialize the IPv4 flow key or along the path
to the FIB core.

In preparation for adding a DSCP selector to IPv4 and IPv6 FIB rules, we
need to make sure the entire DSCP value is present in the IPv4 flow key.
This patchset continues to unmask the upper DSCP bits, but this time in
the output route path.

Patches #1-#3 unmask the upper DSCP bits in the various places that
invoke the core output route lookup functions directly.

Patches #4-#6 do the same in three helpers that are widely used in the
output path to initialize the TOS field in the IPv4 flow key.

The rest of the patches continue to unmask these bits in call sites that
invoke the following wrappers around the core lookup functions:

Patch #7 - __ip_route_output_key()
Patches #8-#12 - ip_route_output_flow()

The next patchset will handle the callers of ip_route_output_ports() and
ip_route_output_key().

No functional changes are expected as commit 1fa3314c14c6 ("ipv4:
Centralize TOS matching") moved the masking of the upper DSCP bits to
the core where 'flowi4_tos' is matched against the TOS selector.

Changes since v1 [1]:

* Remove IPTOS_RT_MASK in patch #7 instead of in patch #6

[1] https://lore.kernel.org/netdev/20240827111813.2115285-1-idosch@nvidia.com/
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

bpf: Unmask upper DSCP bits in __bpf_redirect_neigh_v4()

Unmask the upper DSCP bits when calling ip_route_output_flow() so that
in the future it could perform the FIB lookup according to the full DSCP
value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

vrf: Unmask upper DSCP bits in vrf_process_v4_outbound()

Unmask the upper DSCP bits when calling ip_route_output_flow() so that
in the future it could perform the FIB lookup according to the full DSCP
value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipvlan: Unmask upper DSCP bits in ipvlan_process_v4_outbound()

Unmask the upper DSCP bits when calling ip_route_output_flow() so that
in the future it could perform the FIB lookup according to the full DSCP
value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: sit: Unmask upper DSCP bits in ipip6_tunnel_xmit()

The function calls flowi4_init_output() to initialize an IPv4 flow key
with which it then performs a FIB lookup using ip_route_output_flow().

The 'tos' variable with which the TOS value in the IPv4 flow key
(flowi4_tos) is initialized contains the full DS field. Unmask the upper
DSCP bits so that in the future the FIB lookup could be performed
according to the full DSCP value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: Unmask upper DSCP bits in ip_send_unicast_reply()

The function calls flowi4_init_output() to initialize an IPv4 flow key
with which it then performs a FIB lookup using ip_route_output_flow().

'arg->tos' with which the TOS value in the IPv4 flow key (flowi4_tos) is
initialized contains the full DS field. Unmask the upper DSCP bits so
that in the future the FIB lookup could be performed according to the
full DSCP value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

xfrm: Unmask upper DSCP bits in xfrm_get_tos()

The function returns a value that is used to initialize 'flowi4_tos'
before being passed to the FIB lookup API in the following call chain:

xfrm_bundle_create()
tos = xfrm_get_tos(fl, family)
xfrm_dst_lookup(..., tos, ...)
__xfrm_dst_lookup(..., tos, ...)
xfrm4_dst_lookup(..., tos, ...)
__xfrm4_dst_lookup(..., tos, ...)
fl4->flowi4_tos = tos
__ip_route_output_key(net, fl4)

Unmask the upper DSCP bits so that in the future the output route lookup
could be performed according to the full DSCP value.

Remove IPTOS_RT_MASK since it is no longer used.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: Unmask upper DSCP bits when building flow key

build_sk_flow_key() and __build_flow_key() are used to build an IPv4
flow key before calling one of the FIB lookup APIs.

Unmask the upper DSCP bits so that in the future the lookup could be
performed according to the full DSCP value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: Unmask upper DSCP bits in get_rttos()

The function is used by a few socket types to retrieve the TOS value
with which to perform the FIB lookup for packets sent through the socket
(flowi4_tos). If a DS field was passed using the IP_TOS control message,
then it is used. Otherwise the one specified via the IP_TOS socket
option.

Unmask the upper DSCP bits so that in the future the lookup could be
performed according to the full DSCP value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: Unmask upper DSCP bits in ip_sock_rt_tos()

The function is used to read the DS field that was stored in IPv4
sockets via the IP_TOS socket option so that it could be used to
initialize the flowi4_tos field before resolving an output route.

Unmask the upper DSCP bits so that in the future the output route lookup
could be performed according to the full DSCP value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: icmp: Unmask upper DSCP bits in icmp_route_lookup()

The function is called to resolve a route for an ICMP message that is
sent in response to a situation. Based on the type of the generated ICMP
message, the function is either passed the DS field of the packet that
generated the ICMP message or a DS field that is derived from it.

Unmask the upper DSCP bits before resolving and output route via
ip_route_output_key_hash() so that in the future the lookup could be
performed according to the full DSCP value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: Unmask upper DSCP bits in ip_route_output_key_hash()

Unmask the upper DSCP bits so that in the future output routes could be
looked up according to the full DSCP value.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: Unmask upper DSCP bits in RTM_GETROUTE output route lookup

Unmask the upper DSCP bits when looking up an output route via the
RTM_GETROUTE netlink message so that in the future the lookup could be
performed according to the full DSCP value.

No functional changes intended since the upper DSCP bits are masked when
comparing against the TOS selectors in FIB rules and routes.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

can: kvaser_usb: Simplify with dev_err_probe()

dev_err_probe() is used to log an error message during the probe process
of a device.

It can simplify the error path and unify a message template.

Using this helper is totally fine even if err is known to never
be -EPROBE_DEFER.

The benefit compared to a normal dev_err() is the standardized format
of the error code, it being emitted symbolically and the fact that
the error code is returned which allows more compact error paths.

Signed-off-by: Yan Zhen <yanzhen@vivo.com>
Link: https://patch.msgid.link/20240830110651.519119-1-yanzhen@vivo.com
mkl: fix indention
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

can: kvaser_pciefd: Use IS_ENABLED() instead of #ifdef

Use the IS_ENABLED() macro to check kernel config defines instead of
ifdef. Use upper_32_bits() to avoid warnings about "right shift count
>= width of type" on systems with CONFIG_ARCH_DMA_ADDR_T_64BIT not
set. In kvaser_pciefd_write_dma_map_altera() use lower_32_bits() for
symmetry.

Signed-off-by: Martin Jocic <martin.jocic@kvaser.com>
Link: https://patch.msgid.link/20240830141038.1402217-1-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

can: netlink: avoid call to do_set_data_bittiming callback with stale can_priv::ctrlmode

This patch moves the evaluation of data[IFLA_CAN_CTRLMODE] in function
can_changelink in front of the evaluation of data[IFLA_CAN_BITTIMING].

This avoids a call to do_set_data_bittiming providing a stale
can_priv::ctrlmode with a CAN_CTRLMODE_FD flag not matching the
requested state when switching between a CAN Classic and CAN-FD bitrate.

In the same manner the evaluation of data[IFLA_CAN_CTRLMODE] in function
can_validate is also moved in front of the evaluation of
data[IFLA_CAN_BITTIMING].

This is a preparation for patches where the nominal and data bittiming
may have interdependencies on the driver side depending on the
CAN_CTRLMODE_FD flag state.

Signed-off-by: Stefan Mätje <stefan.maetje@esd.eu>
Link: https://patch.msgid.link/20240808164224.213522-1-stefan.maetje@esd.eu
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

can: j1939: use correct function name in comment

The function j1939_cancel_all_active_sessions() was renamed to
j1939_cancel_active_session() but name in comment wasn't updated.

Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Acked-by: Oleksij Rempel <o.rempel@pengutronix.de>
Fixes: 9d71dd0c7009 ("can: add support of SAE J1939 protocol")
Link: https://patch.msgid.link/1724935703-44621-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

dt-bindings: can: convert microchip,mcp251x.txt to yaml

Convert binding doc microchip,mcp251x.txt to yaml.
Additional change:
- add ref to spi-peripheral-props.yaml

Fix below warning:
arch/arm64/boot/dts/freescale/imx8dx-colibri-eval-v3.dtb: /bus@5a000000/spi@5a020000/can@0:
failed to match any schema with compatible: ['microchip,mcp2515']

Signed-off-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://patch.msgid.link/20240814164407.4022211-1-Frank.Li@nxp.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

dt-bindings: can: renesas,rcar-canfd: Document R-Car V4M support

Document support for the CAN-FD Interface on the Renesas R-Car V4M
(R8A779H0) SoC, which supports up to four channels.

The CAN-FD module on R-Car V4M is very similar to the one on R-Car V4H,
but differs in some hardware parameters, as reflected by the Parameter
Status Information part of the Global IP Version Register.  However,
none of this parameterization should have any impact on the driver, as
the driver does not access any register that is impacted by the
parameterization (except for the number of channels).

Signed-off-by: Duy Nguyen <duy.nguyen.rh@renesas.com>
[geert: Clarify R-Car V4M differences]
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Simon Horman <horms@kernel.org>
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://patch.msgid.link/68b5f910bef89508e3455c768844ebe859d6ff1d.1722520779.git.geert+renesas@glider.be
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

ieee802154: Correct spelling in nl802154.h

Correct spelling in nl802154.h.
As reported by codespell.

Signed-off-by: Simon Horman <horms@kernel.org>
Message-ID: <20240829-wpan-spell-v1-2-799d840e02c4@kernel.org>
Signed-off-by: Stefan Schmidt <stefan@datenfreihafen.org>

mac802154: Correct spelling in mac802154.h

Correct spelling in mac802154.h.
As reported by codespell.

Signed-off-by: Simon Horman <horms@kernel.org>
Message-ID: <20240829-wpan-spell-v1-1-799d840e02c4@kernel.org>
Signed-off-by: Stefan Schmidt <stefan@datenfreihafen.org>

i40e: Add Energy Efficient Ethernet ability for X710 Base-T/KR/KX cards

Add "EEE: Enabled/Disabled" to dmesg for supported X710 Base-T/KR/KX
cards. According to the IEEE standard report the EEE ability and the
EEE Link Partner ability. Use the kernel's 'ethtool_keee' structure
and report EEE link modes.

Example:
dmesg | grep 'NIC Link is'
ethtool --show-eee <device>

Before:
NIC Link is Up, 10 Gbps Full Duplex, Flow Control: None

        Supported EEE link modes:  Not reported
        Advertised EEE link modes:  Not reported
        Link partner advertised EEE link modes:  Not reported

After:
NIC Link is Up, 10 Gbps Full Duplex, Flow Control: None, EEE: Enabled

        Supported EEE link modes:  100baseT/Full
                                   1000baseT/Full
                                   10000baseT/Full
        Advertised EEE link modes:  100baseT/Full
                                    1000baseT/Full
                                    10000baseT/Full
        Link partner advertised EEE link modes:  100baseT/Full
                                                 1000baseT/Full
                                                 10000baseT/Full

Reviewed-by: Arkadiusz Kubalewski <arkadiusz.kubalewski@intel.com>
Signed-off-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@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>

e1000e: avoid failing the system during pm_suspend

Occasionally when the system goes into pm_suspend, the suspend might fail
due to a PHY access error on the network adapter. Previously, this would
have caused the whole system to fail to go to a low power state.
An example of this was reported in the following Bugzilla:
https://bugzilla.kernel.org/show_bug.cgi?id=205015

[ 1663.694828] e1000e 0000:00:19.0 eth0: Failed to disable ULP
[ 1664.731040] asix 2-3:1.0 eth1: link up, 100Mbps, full-duplex, lpa 0xC1E1
[ 1665.093513] e1000e 0000:00:19.0 eth0: Hardware Error
[ 1665.596760] e1000e 0000:00:19.0: pci_pm_resume+0x0/0x80 returned 0 after 2975399 usecs

and then the system never recovers from it, and all the following suspend failed due to this
[22909.393854] PM: pci_pm_suspend(): e1000e_pm_suspend+0x0/0x760 [e1000e] returns -2
[22909.393858] PM: dpm_run_callback(): pci_pm_suspend+0x0/0x160 returns -2
[22909.393861] PM: Device 0000:00:1f.6 failed to suspend async: error -2

This can be avoided by changing the return values of __e1000_shutdown and
e1000e_pm_suspend functions so that they always return 0 (success). This
is consistent with what other drivers do.

If the e1000e driver encounters a hardware error during suspend, potential
side effects include slightly higher power draw or non-working wake on
LAN. This is preferred to a system-level suspend failure, and a warning
message is written to the system log, so that the user can be aware that
the LAN controller experienced a problem during suspend.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=205015
Suggested-by: Dima Ruinskiy <dima.ruinskiy@intel.com>
Signed-off-by: Vitaly Lifshits <vitaly.lifshits@intel.com>
Tested-by: Mor Bar-Gabay <morx.bar.gabay@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

ethtool: pse-pd: move pse validation into set

Move validation into set, removing .set_validate operation as its current
implementation holds the rtnl lock for acquiring the PHY device, defeating
the intended purpose of checking before grabbing the lock.

Reported-by: syzbot+ec369e6d58e210135f71@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=ec369e6d58e210135f71
Fixes: 31748765bed3 ("net: ethtool: pse-pd: Target the command to the requested PHY")
Signed-off-by: Diogo Jahchan Koike <djahchankoike@gmail.com>
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Link: https://patch.msgid.link/20240829184830.5861-1-djahchankoike@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Documentation: Add missing fields to net_cachelines

Two fields, page_pools and *irq_moder, were added to struct net_device
in commit 083772c9f972 ("net: page_pool: record pools per netdev") and
commit f750dfe825b9 ("ethtool: provide customized dim profile
management"), respectively.

Add both to the net_cachelines documentation, as well.

Signed-off-by: Joe Damato <jdamato@fastly.com>
Link: https://patch.msgid.link/20240829155742.366584-1-jdamato@fastly.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'icmp-avoid-possible-side-channels-attacks'

Eric Dumazet says:

====================
icmp: avoid possible side-channels attacks

Keyu Man reminded us that linux ICMP rate limiting was still allowing
side-channels attacks.

Quoting the fine document [1]:

4.4 Private Source Port Scan Method
...
 We can then use the same global ICMP rate limit as a side
 channel to infer if such an ICMP message has been triggered. At
 first glance, this method can work but at a low speed of one port
 per second, due to the per-IP rate limit on ICMP messages.
 Surprisingly, after we analyze the source code of the ICMP rate
 limit implementation, we find that the global rate limit is checked
 prior to the per-IP rate limit. This means that even if the per-IP
 rate limit may eventually determine that no ICMP reply should be
 sent, a packet is still subjected to the global rate limit check and one
 token is deducted. Ironically, such a decision is consciously made
 by Linux developers to avoid invoking the expensive check of the
 per-IP rate limit [ 22], involving a search process to locate the per-IP
 data structure.
 This effectively means that the per-IP rate limit can be disre-
 garded for the purpose of our side channel based scan, as it only
 determines if the final ICMP reply is generated but has nothing to
 do with the global rate limit counter decrement. As a result, we can
 continue to use roughly the same scan method as efficient as before,
 achieving 1,000 ports per second
...

This series :

1) Changes the order of the two rate limiters to fix the issue.

2-3) Make the 'host-wide' rate limiter a per-netns one.

[1]
Link: https://dl.acm.org/doi/pdf/10.1145/3372297.3417280
====================

Link: https://patch.msgid.link/20240829144641.3880376-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

icmp: icmp_msgs_per_sec and icmp_msgs_burst sysctls become per netns

Previous patch made ICMP rate limits per netns, it makes sense
to allow each netns to change the associated sysctl.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://patch.msgid.link/20240829144641.3880376-4-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

icmp: move icmp_global.credit and icmp_global.stamp to per netns storage

Host wide ICMP ratelimiter should be per netns, to provide better isolation.

Following patch in this series makes the sysctl per netns.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://patch.msgid.link/20240829144641.3880376-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

icmp: change the order of rate limits

ICMP messages are ratelimited :

After the blamed commits, the two rate limiters are applied in this order:

1) host wide ratelimit (icmp_global_allow())

2) Per destination ratelimit (inetpeer based)

In order to avoid side-channels attacks, we need to apply
the per destination check first.

This patch makes the following change :

1) icmp_global_allow() checks if the host wide limit is reached.
   But credits are not yet consumed. This is deferred to 3)

2) The per destination limit is checked/updated.
   This might add a new node in inetpeer tree.

3) icmp_global_consume() consumes tokens if prior operations succeeded.

This means that host wide ratelimit is still effective
in keeping inetpeer tree small even under DDOS.

As a bonus, I removed icmp_global.lock as the fast path
can use a lock-free operation.

Fixes: c0303efeab73 ("net: reduce cycles spend on ICMP replies that gets rate limited")
Fixes: 4cdf507d5452 ("icmp: add a global rate limitation")
Reported-by: Keyu Man <keyu.man@email.ucr.edu>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Cc: Jesper Dangaard Brouer <hawk@kernel.org>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20240829144641.3880376-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: openvswitch: Use ERR_CAST() to return

Using ERR_CAST() is more reasonable and safer, When it is necessary
to convert the type of an error pointer and return it.

Signed-off-by: Yan Zhen <yanzhen@vivo.com>
Acked-by: Eelco Chaudron <echaudro@redhat.com>
Reviewed-by: Aaron Conole <aconole@redhat.com>
Link: https://patch.msgid.link/20240829095509.3151987-1-yanzhen@vivo.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfp: Convert to use ERR_CAST()

Use ERR_CAST() as it is designed for casting an error pointer to
another type.

Signed-off-by: Shen Lichuan <shenlichuan@vivo.com>
Link: https://patch.msgid.link/20240829072538.33195-1-shenlichuan@vivo.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

sfc: Convert to use ERR_CAST()

As opposed to open-code, using the ERR_CAST macro clearly indicates that
this is a pointer to an error value and a type conversion was performed.

Signed-off-by: Shen Lichuan <shenlichuan@vivo.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Martin Habets <habetsm.xilinx@gmail.com>
Reviewed-by: Edward Cree <ecree.xilinx@gmail.com>
Link: https://patch.msgid.link/20240829021253.3066-1-shenlichuan@vivo.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

igc: Move the MULTI GBT AN Control Register to _regs file

MULTI GBT AN Control Register is IEEE Standard Register 7.32 (not a mask).
The right place should be in igc_reg.h file. In accordance with the
registers naming convention added IGC_' prefix.

Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Tested-by: Avigail Dahan <avigailx.dahan@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

igc: Add Energy Efficient Ethernet ability

According to the IEEE standard report the EEE ability (registers 7.60 and
7.62) and the EEE Link Partner ability (registers 7.61 and 7.63). Use the
kernel's 'ethtool_keee' structure and report EEE link modes.

Example:
ethtool --show-eee <device>

Before:
Advertised EEE link modes:  Not reported
Link partner advertised EEE link modes:  Not reported

After:
Advertised EEE link modes:  100baseT/Full
                            1000baseT/Full
                            2500baseT/Full
Link partner advertised EEE link modes:  100baseT/Full
                                         1000baseT/Full
                                         2500baseT/Full

Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
Tested-by: Avigail Dahan <avigailx.dahan@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

igc: Get rid of spurious interrupts

When running the igc with XDP/ZC in busy polling mode with deferral of hard
interrupts, interrupts still happen from time to time. That is caused by
the igc task watchdog which triggers Rx interrupts periodically.

That mechanism has been introduced to overcome skb/memory allocation
failures [1]. So the Rx clean functions stop processing the Rx ring in case
of such failure. The task watchdog triggers Rx interrupts periodically in
the hope that memory became available in the mean time.

The current behavior is undesirable for real time applications, because the
driver induced Rx interrupts trigger also the softirq processing. However,
all real time packets should be processed by the application which uses the
busy polling method.

Therefore, only trigger the Rx interrupts in case of real allocation
failures. Introduce a new flag for signaling that condition.

[1] - https://git.kernel.org/pub/scm/linux/kernel/git/tglx/history.git/commit/?id=3be507547e6177e5c808544bd6a2efa2c7f1d436

Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Mor Bar-Gabay <morx.bar.gabay@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

igc: Add MQPRIO offload support

Add support for offloading MQPRIO. The hardware has four priorities as well
as four queues. Each queue must be a assigned with a unique priority.

However, the priorities are only considered in TSN Tx mode. There are two
TSN Tx modes. In case of MQPRIO the Qbv capability is not required.
Therefore, use the legacy TSN Tx mode, which performs strict priority
arbitration.

Example for mqprio with hardware offload:

|tc qdisc replace dev ${INTERFACE} handle 100 parent root mqprio num_tc 4 \
|   map 0 0 0 0 0 1 2 3 0 0 0 0 0 0 0 0 \
|   queues 1@0 1@1 1@2 1@3 \
|   hw 1

The mqprio Qdisc also allows to configure the `preemptible_tcs'. However,
frame preemption is not supported yet.

Tested on Intel i225 and implemented by following data sheet section 7.5.2,
Transmit Scheduling.

Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Mor Bar-Gabay <morx.bar.gabay@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Merge branch 'am-qt2025-phy-rust'

FUJITA Tomonori says:

====================
net: phy: add Applied Micro QT2025 PHY driver

This patchset adds a PHY driver for Applied Micro Circuits Corporation
QT2025.

The first patch adds Rust equivalent to include/linux/sizes.h, makes
code more readable. The 2-5th patches update the PHYLIB Rust bindings.
The 4th and 5th patches have been reviewed previously in a different
thread [1].

QT2025 PHY support was implemented as a part of an Ethernet driver for
Tehuti Networks TN40xx chips. Multiple vendors (DLink, Asus, Edimax,
QNAP, etc) developed adapters based on TN40xx chips. Tehuti Networks
went out of business and the driver wasn't merged into mainline. But
it's still distributed with some of the hardware (and also available
on some vendor sites).

The original driver handles multiple PHY hardware (AMCC QT2025, TI
TLK10232, Aqrate AQR105, and Marvell MV88X3120, MV88X3310, and
MV88E2010). I divided the original driver into MAC and PHY drivers and
implemented a QT2025 PHY driver in Rust.

The MAC driver for Tehuti Networks TN40xx chips was already merged in
6.11-rc1. The MAC and this PHY drivers have been tested with Edimax
EN-9320SFP+ 10G network adapter.

[1] https://lore.kernel.org/rust-for-linux/20240607052113.69026-1-fujita.tomonori@gmail.com/

v7:
- add Trevor as Reviewer to MAINTAINERS file entry
- add Trevor Reviewed-by
- add/fix comments
- replace uppercase hex with lowercase
- remove unnecessary code
- update the commit message (1st patch)
v6: https://lore.kernel.org/netdev/20240820225719.91410-1-fujita.tomonori@gmail.com/
- improve comments
- make the logic to load firmware more readable
- add Copy trait to reg::{C22 and C45}
- add Trevor Reviewed-by
v5: https://lore.kernel.org/netdev/20240819005345.84255-1-fujita.tomonori@gmail.com/
- fix the comments (3th patch)
- add RUST_FW_LOADER_ABSTRACTIONS dependency
- add Andrew and Benno Reviewed-by
v4: https://lore.kernel.org/netdev/20240817051939.77735-1-fujita.tomonori@gmail.com/
- fix the comments
- add Andrew's Reviewed-by
- fix the order of tags
- remove wrong endianness conversion
v3: https://lore.kernel.org/netdev/20240804233835.223460-1-fujita.tomonori@gmail.com/
- use addr_of_mut!` to avoid intermediate mutable reference
- update probe callback's Safety comment
- add MODULE_FIRMWARE equivalent
- add Alice's Reviewed-by
v2: https://lore.kernel.org/netdev/20240731042136.201327-1-fujita.tomonori@gmail.com/
- add comments in accordance with the hw datasheet
- unify C22 and C45 APIs
- load firmware in probe callback instead of config_init
- use firmware API
- handle firmware endian
- check firmware size
- use SZ_*K constants
- avoid confusing phy_id variable
v1: https://lore.kernel.org/netdev/20240415104701.4772-1-fujita.tomonori@gmail.com/
====================

rom: FUJITA Tomonori <fujita.tomonori@gmail.com>
To: netdev@vger.kernel.org
Cc: rust-for-linux@vger.kernel.org, andrew@lunn.ch,
tmgross@umich.edu, miguel.ojeda.sandonis@gmail.com,
benno.lossin@proton.me, aliceryhl@google.com
Subject: [PATCH net-next v7 0/6] net: phy: add Applied Micro QT2025 PHY driver
Date: Sat, 24 Aug 2024 02:06:10 +0000 [thread overview]
Message-ID: <20240824020617.113828-1-fujita.tomonori@gmail.com> (raw)

This patchset adds a PHY driver for Applied Micro Circuits Corporation
QT2025.

The first patch adds Rust equivalent to include/linux/sizes.h, makes
code more readable. The 2-5th patches update the PHYLIB Rust bindings.
The 4th and 5th patches have been reviewed previously in a different
thread [1].

QT2025 PHY support was implemented as a part of an Ethernet driver for
Tehuti Networks TN40xx chips. Multiple vendors (DLink, Asus, Edimax,
QNAP, etc) developed adapters based on TN40xx chips. Tehuti Networks
went out of business and the driver wasn't merged into mainline. But
it's still distributed with some of the hardware (and also available
on some vendor sites).

The original driver handles multiple PHY hardware (AMCC QT2025, TI
TLK10232, Aqrate AQR105, and Marvell MV88X3120, MV88X3310, and
MV88E2010). I divided the original driver into MAC and PHY drivers and
implemented a QT2025 PHY driver in Rust.

The MAC driver for Tehuti Networks TN40xx chips was already merged in
6.11-rc1. The MAC and this PHY drivers have been tested with Edimax
EN-9320SFP+ 10G network adapter.

[1] https://lore.kernel.org/rust-for-linux/20240607052113.69026-1-fujita.tomonori@gmail.com/

v7:
- add Trevor as Reviewer to MAINTAINERS file entry
- add Trevor Reviewed-by
- add/fix comments
- replace uppercase hex with lowercase
- remove unnecessary code
- update the commit message (1st patch)
v6: https://lore.kernel.org/netdev/20240820225719.91410-1-fujita.tomonori@gmail.com/
- improve comments
- make the logic to load firmware more readable
- add Copy trait to reg::{C22 and C45}
- add Trevor Reviewed-by
v5: https://lore.kernel.org/netdev/20240819005345.84255-1-fujita.tomonori@gmail.com/
- fix the comments (3th patch)
- add RUST_FW_LOADER_ABSTRACTIONS dependency
- add Andrew and Benno Reviewed-by
v4: https://lore.kernel.org/netdev/20240817051939.77735-1-fujita.tomonori@gmail.com/
- fix the comments
- add Andrew's Reviewed-by
- fix the order of tags
- remove wrong endianness conversion
v3: https://lore.kernel.org/netdev/20240804233835.223460-1-fujita.tomonori@gmail.com/
- use addr_of_mut!` to avoid intermediate mutable reference
- update probe callback's Safety comment
- add MODULE_FIRMWARE equivalent
- add Alice's Reviewed-by
v2: https://lore.kernel.org/netdev/20240731042136.201327-1-fujita.tomonori@gmail.com/
- add comments in accordance with the hw datasheet
- unify C22 and C45 APIs
- load firmware in probe callback instead of config_init
- use firmware API
- handle firmware endian
- check firmware size
- use SZ_*K constants
- avoid confusing phy_id variable
v1: https://lore.kernel.org/netdev/20240415104701.4772-1-fujita.tomonori@gmail.com/
Signed-off-by: David S. Miller <davem@davemloft.net>
rom: FUJITA Tomonori <fujita.tomonori@gmail.com>
To: netdev@vger.kernel.org
Cc: rust-for-linux@vger.kernel.org, andrew@lunn.ch,
tmgross@umich.edu, miguel.ojeda.sandonis@gmail.com,
benno.lossin@proton.me, aliceryhl@google.com
Subject: [PATCH net-next v7 0/6] net: phy: add Applied Micro QT2025 PHY driver
Date: Sat, 24 Aug 2024 02:06:10 +0000 [thread overview]
Message-ID: <20240824020617.113828-1-fujita.tomonori@gmail.com> (raw)

This patchset adds a PHY driver for Applied Micro Circuits Corporation
QT2025.

The first patch adds Rust equivalent to include/linux/sizes.h, makes
code more readable. The 2-5th patches update the PHYLIB Rust bindings.
The 4th and 5th patches have been reviewed previously in a different
thread [1].

QT2025 PHY support was implemented as a part of an Ethernet driver for
Tehuti Networks TN40xx chips. Multiple vendors (DLink, Asus, Edimax,
QNAP, etc) developed adapters based on TN40xx chips. Tehuti Networks
went out of business and the driver wasn't merged into mainline. But
it's still distributed with some of the hardware (and also available
on some vendor sites).

The original driver handles multiple PHY hardware (AMCC QT2025, TI
TLK10232, Aqrate AQR105, and Marvell MV88X3120, MV88X3310, and
MV88E2010). I divided the original driver into MAC and PHY drivers and
implemented a QT2025 PHY driver in Rust.

The MAC driver for Tehuti Networks TN40xx chips was already merged in
6.11-rc1. The MAC and this PHY drivers have been tested with Edimax
EN-9320SFP+ 10G network adapter.

[1] https://lore.kernel.org/rust-for-linux/20240607052113.69026-1-fujita.tomonori@gmail.com/

v7:
- add Trevor as Reviewer to MAINTAINERS file entry
- add Trevor Reviewed-by
- add/fix comments
- replace uppercase hex with lowercase
- remove unnecessary code
- update the commit message (1st patch)
v6: https://lore.kernel.org/netdev/20240820225719.91410-1-fujita.tomonori@gmail.com/
- improve comments
- make the logic to load firmware more readable
- add Copy trait to reg::{C22 and C45}
- add Trevor Reviewed-by
v5: https://lore.kernel.org/netdev/20240819005345.84255-1-fujita.tomonori@gmail.com/
- fix the comments (3th patch)
- add RUST_FW_LOADER_ABSTRACTIONS dependency
- add Andrew and Benno Reviewed-by
v4: https://lore.kernel.org/netdev/20240817051939.77735-1-fujita.tomonori@gmail.com/
- fix the comments
- add Andrew's Reviewed-by
- fix the order of tags
- remove wrong endianness conversion
v3: https://lore.kernel.org/netdev/20240804233835.223460-1-fujita.tomonori@gmail.com/
- use addr_of_mut!` to avoid intermediate mutable reference
- update probe callback's Safety comment
- add MODULE_FIRMWARE equivalent
- add Alice's Reviewed-by
v2: https://lore.kernel.org/netdev/20240731042136.201327-1-fujita.tomonori@gmail.com/
- add comments in accordance with the hw datasheet
- unify C22 and C45 APIs
- load firmware in probe callback instead of config_init
- use firmware API
- handle firmware endian
- check firmware size
- use SZ_*K constants
- avoid confusing phy_id variable
v1: https://lore.kernel.org/netdev/20240415104701.4772-1-fujita.tomonori@gmail.com/
Signed-off-by: David S. Miller <davem@davemloft.net>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>

net: phy: add Applied Micro QT2025 PHY driver

This driver supports Applied Micro Circuits Corporation QT2025 PHY,
based on a driver for Tehuti Networks TN40xx chips.

The original driver for TN40xx chips supports multiple PHY hardware
(AMCC QT2025, TI TLK10232, Aqrate AQR105, and Marvell 88X3120,
88X3310, and MV88E2010). This driver is extracted from the original
driver and modified to a PHY driver in Rust.

This has been tested with Edimax EN-9320SFP+ 10G network adapter.

Reviewed-by: Trevor Gross <tmgross@umich.edu>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

rust: net::phy unified genphy_read_status function for C22 and C45 registers

Add unified genphy_read_status function for C22 and C45
registers. Instead of having genphy_c22 and genphy_c45 methods, this
unifies genphy_read_status functions for C22 and C45.

Reviewed-by: Trevor Gross <tmgross@umich.edu>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

rust: net::phy unified read/write API for C22 and C45 registers

Add the unified read/write API for C22 and C45 registers. The
abstractions support access to only C22 registers now. Instead of
adding read/write_c45 methods specifically for C45, a new reg module
supports the unified API to access C22 and C45 registers with trait,
by calling an appropriate phylib functions.

Reviewed-by: Trevor Gross <tmgross@umich.edu>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

rust: net::phy implement AsRef<kernel::device::Device> trait

Implement AsRef<kernel::device::Device> trait for Device. A PHY driver
needs a reference to device::Device to call the firmware API.

Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Reviewed-by: Trevor Gross <tmgross@umich.edu>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

rust: net::phy support probe callback

Support phy_driver probe callback, used to set up device-specific
structures.

Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Reviewed-by: Trevor Gross <tmgross@umich.edu>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

rust: sizes: add commonly used constants

Add rust equivalent to include/linux/sizes.h, makes code more
readable. Only SZ_*K that QT2025 PHY driver uses are added.

Make generated constants accessible with a proper type.

Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Reviewed-by: Trevor Gross <tmgross@umich.edu>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'bnxt_en-update-for-net-next'

Michael Chan says:

====================
bnxt_en: Update for net-next

This series starts with 2 patches to support firmware crash dump.  The
driver allocates the required DMA memory ahead of time for firmware to
store the crash dump if and when it crashes.  Patch 3 adds priority and
TPID for the .ndo_set_vf_vlan() callback.  Note that this was rejected
and reverted last year and it is being re-submitted after recent changes
in the guidelines.  The remaining patches are MSIX related.  Legacy
interrupt is no longer supported by firmware so we remove the support
in the driver.  We then convert to use the newer kernel APIs to
allocate and enable MSIX vectors.  The last patch adds support for
dynamic MSIX.

v3: https://lore.kernel.org/20240823195657.31588-1-michael.chan@broadcom.com
v2: https://lore.kernel.org/20240816212832.185379-1-michael.chan@broadcom.com
v1: https://lore.kernel.org/20240713234339.70293-1-michael.chan@broadcom.com
====================

Link: https://patch.msgid.link/20240828183235.128948-1-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Support dynamic MSIX

A range of MSIX vectors are allocated at initialization for the number
needed for RocE and L2.  During run-time, if the user increases or
decreases the number of L2 rings, all the MSIX vectors have to be
freed and a new range has to be allocated.  This is not optimal and
causes disruptions to RoCE traffic every time there is a change in L2
MSIX.

If the system supports dynamic MSIX allocations, use dynamic
allocation to add new L2 MSIX vectors or free unneeded L2 MSIX
vectors.  RoCE traffic is not affected using this scheme.

Reviewed-by: Hongguang Gao <hongguang.gao@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Link: https://patch.msgid.link/20240828183235.128948-10-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Allocate the max bp->irq_tbl size for dynamic msix allocation

If dynamic MSIX allocation is supported, additional MSIX can be
allocated at run-time without reinitializing the existing MSIX entries.
The first step to support this dynamic scheme is to allocate a large
enough bp->irq_tbl if dynamic allocation is supported.

Reviewed-by: Hongguang Gao <hongguang.gao@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20240828183235.128948-9-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Replace deprecated PCI MSIX APIs

Use the new pci_alloc_irq_vectors() and pci_free_irq_vectors() to
replace the deprecated pci_enable_msix_range() and pci_disable_msix().

Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20240828183235.128948-8-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Remove register mapping to support INTX

In legacy INTX mode, a register is mapped so that the INTX handler can
read it to determine if the NIC is the source of the interrupt.  This
and all the related macros are no longer needed now that INTX is no
longer supported.

Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Hongguang Gao <hongguang.gao@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20240828183235.128948-7-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Remove BNXT_FLAG_USING_MSIX flag

Now that we only support MSIX, the BNXT_FLAG_USING_MSIX is always true.
Remove it and any if conditions checking for it.  Remove the INTX
handler and associated logic.

Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Hongguang Gao <hongguang.gao@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20240828183235.128948-6-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Deprecate support for legacy INTX mode

Firmware has deprecated support for legacy INTX in 2022 (since v2.27)
and INTX hasn't been tested for many years before that.  INTX was
only used as a fallback mechansim in case MSIX wasn't available.  MSIX
is always supported by all firmware.  If MSIX capability in PCI config
space is not found during probe, abort.

Reviewed-by: Hongguang Gao <hongguang.gao@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20240828183235.128948-5-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: Support QOS and TPID settings for the SRIOV VLAN

With recent changes in the .ndo_set_vf_*() guidelines, resubmitting
this patch that was reverted eariler in 2023:

c27153682eac ("Revert "bnxt_en: Support QOS and TPID settings for the SRIOV VLAN")

Add these missing settings in the .ndo_set_vf_vlan() method.
Older firmware does not support the TPID setting so check for
proper support.

Remove the unused BNXT_VF_QOS flag.

Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20240828183235.128948-4-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: add support for retrieving crash dump using ethtool

Add support for retrieving crash dump using ethtool -w on the
supported interface.

Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Vikas Gupta <vikas.gupta@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20240828183235.128948-3-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt_en: add support for storing crash dump into host memory

Newer firmware supports automatic DMA of crash dump to host memory
when it crashes.  If the feature is supported, allocate the required
memory using the existing context memory infrastructure.  Communicate
the page table containing the DMA addresses to the firmware.

Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Vikas Gupta <vikas.gupta@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20240828183235.128948-2-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'adding-so_peek_off-for-tcpv6'

Jon Maloy says:

====================
Adding SO_PEEK_OFF for TCPv6

Adding SO_PEEK_OFF for TCPv6 and selftest for both TCPv4 and TCPv6
====================

Link: https://patch.msgid.link/20240828183752.660267-1-jmaloy@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>