net: phy: Improved PHY error reporting in state machine

When the PHY library calls phy_error() something bad has happened, and
we halt the PHY state machine. Calling phy_error() from the main state
machine however is not precise enough to know whether the issue is
reading the link status or starting auto-negotiation.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/ism: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

isdn: mISDN: netjet: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: phy: micrel: Add support for PTP_PF_EXTTS for lan8841

Extend the PTP programmable gpios to implement also PTP_PF_EXTTS
function. The pins can be configured to capture both of rising
and falling edge. Once the event is seen, then an interrupt is
generated and the LTC is saved in the registers.

This was tested using:
ts2phc -m -l 7 -s generic -f ts2phc.cfg

Where the configuration was the following:
[global]
ts2phc.pin_index  6

[eth2]

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ethernet: ec_bhf: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

ionic: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Acked-by: Shannon Nelson <shannon.nelson@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: mana: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx5: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/fungible: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Acked-by: Dimitris Michailidis <dmichail@fungible.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: cxgb4vf: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: hisilicon: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: liquidio: Remove redundant pci_clear_master

Remove pci_clear_master to simplify the code,
the bus-mastering is also cleared in do_pci_disable_device,
like this:
./drivers/pci/pci.c:2197
static void do_pci_disable_device(struct pci_dev *dev)
{
u16 pci_command;

pci_read_config_word(dev, PCI_COMMAND, &pci_command);
if (pci_command & PCI_COMMAND_MASTER) {
pci_command &= ~PCI_COMMAND_MASTER;
pci_write_config_word(dev, PCI_COMMAND, pci_command);
}

pcibios_disable_device(dev);
}.
And dev->is_busmaster is set to 0 in pci_disable_device.

Signed-off-by: Cai Huoqing <cai.huoqing@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>

fix typos in net/sched/* files

This patch fixes typos in net/sched/* files.

Signed-off-by: Taichi Nishimura <awkrail01@gmail.com>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

ethernet: broadcom/sb1250-mac: clean up after SIBYTE_BCM1x55 removal

With commit b984d7b56dfc ("MIPS: sibyte: Remove unused config option
SIBYTE_BCM1x55"), some #if's in the Broadcom SiByte SOC built-in Ethernet
driver can be simplified.

Simplify prepreprocessor conditions after config SIBYTE_BCM1x55 removal.

Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Reviewed-by: Steen Hegelund <Steen.Hegelund@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: prevent router_solicitations for team port

The issue fixed for bonding in commit c2edacf80e15 ("bonding / ipv6: no
addrconf for slaves separately from master") also exists in team driver.
However, we can't just disable ipv6 addrconf for team ports, as 'teamd'
will need it when nsns_ping watch is used in the user space.

Instead of preventing ipv6 addrconf, this patch only prevents RS packets
for team ports, as it did in commit b52e1cce31ca ("ipv6: Don't send rs
packets to the interface of ARPHRD_TUNNEL").

Note that we do not prevent DAD packets, to avoid the changes getting
intricate / hacky. Also, usually sysctl dad_transmits is set to 1 and
only 1 DAD packet will be sent, and by now no libteam user complains
about DAD packets on team ports, unlike RS packets.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'main' of ssh://gitolite./linux/kernel/git/netfilter/nf-next

Florian Westphal says:

====================
netfilter updates for net-next

This pull request contains changes for the *net-next* tree.

1. Change IPv6 stack to keep conntrack references until ipsec policy
   checks are done, like ipv4, from Madhu Koriginja.
   This update was missed when IPv6 NAT support was added 10 years ago.

2. get rid of old 'compat' structure layout in nf_nat_redirect
   core and move the conversion to the only user that needs the
   old layout for abi reasons. From Jeremy Sowden.

3. Compact some common code paths in nft_redir, also from Jeremy.

4. Time to remove the 'default y' knob so iptables 32bit compat interface
   isn't compiled in by default anymore, from myself.

5. Move ip(6)tables builtin icmp matches to the udptcp one.
   This has the advantage that icmp/icmpv6 match doesn't load the
   iptables/ip6tables modules anymore when iptables-nft is used.
   Also from myself.

* 'main' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next:
  netfilter: keep conntrack reference until IPsecv6 policy checks are done
  xtables: move icmp/icmpv6 logic to xt_tcpudp
  netfilter: xtables: disable 32bit compat interface by default
  netfilter: nft_masq: deduplicate eval call-backs
  netfilter: nft_redir: use `struct nf_nat_range2` throughout and deduplicate eval call-backs
====================

Link: https://lore.kernel.org/r/20230322210802.6743-1-fw@strlen.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

docs: netdev: add note about Changes Requested and revising commit messages

One of the most commonly asked questions is "I answered all questions
and don't need to make any code changes, why was the patch not applied".
Document our time honored tradition of asking people to repost with
improved commit messages, to record the answers to reviewer questions.

Take this opportunity to also recommend a change log format.

Link: https://lore.kernel.org/r/20230322231202.265835-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnx2: remove deadcode in bnx2_init_cpus()

The load_cpu_fw function has no error return code
and always returns zero. Checking the value returned by
this function does not make sense.
Now checking the value of the return value is misleading when reading
the code. Path with error handling was deleted in 57579f7629a3
("bnx2: Use request_firmware()").
As a result, bnx2_init_cpus() will also return only zero
Therefore, it will be safe to change the type of functions
to void and remove checking to improving readability.

Found by Security Code and Linux Verification
Center (linuxtesting.org) with SVACE

Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Maxim Korotkov <korotkov.maxim.s@gmail.com>
Link: https://lore.kernel.org/r/20230322162843.3452-1-korotkov.maxim.s@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: add IPA v5.0 to ipa_version_string()

In the IPA device sysfs directory, the "version" file can be read to
find out what IPA version is implemented.  The content of this file
is supplied by ipa_version_string(), which needs to be updated to
properly handle IPA v5.0.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230322144742.2203947-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ynl: allow to encode u8 attr

Playing with dpll netlink, I came across following issue:
$ sudo ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/dpll.yaml --do pin-set --json '{"id": 0, "pin-idx": 1, "pin-state": 1}'
Traceback (most recent call last):
  File "tools/net/ynl/cli.py", line 52, in <module>
    main()
  File "tools/net/ynl/cli.py", line 40, in main
    reply = ynl.do(args.do, attrs)
  File "tools/net/ynl/lib/ynl.py", line 520, in do
    return self._op(method, vals)
  File "tools/net/ynl/lib/ynl.py", line 476, in _op
    msg += self._add_attr(op.attr_set.name, name, value)
  File "tools/net/ynl/lib/ynl.py", line 344, in _add_attr
    raise Exception(f'Unknown type at {space} {name} {value} {attr["type"]}')
Exception: Unknown type at dpll pin-state 1 u8

I'm not that familiar with ynl code, but from a quick peek, I suspect
that couple other types are missing for both encoding and decoding.
Ignoring those here as I'm scratching my local itch only.

Fix the issue by adding u8 attr packing.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20230322154242.1739136-1-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

docs: networking: document NAPI

Add basic documentation about NAPI. We can stop linking to the ancient
doc on the LF wiki.

Link: https://lore.kernel.org/all/20230315223044.471002-1-kuba@kernel.org/
Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Pavel Pisa <pisa@cmp.felk.cvut.cz> # for ctucanfd-driver.rst
Reviewed-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Stephen Hemminger <stephen@networkplumber.org>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Link: https://lore.kernel.org/r/20230322053848.198452-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/sched: act_api: use the correct TCA_ACT attributes in dump

4 places in the act api code are using 'TCA_' definitions where they
should be using 'TCA_ACT_', which is confusing for the reader, although
functionally they are equivalent.

Cc: Hangbin Liu <haliu@redhat.com>
Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Acked-by: Hangbin Liu <haliu@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

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

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-03-21 (igb, igbvf, igc)

This series contains updates to igb, igbvf, and igc drivers.

Andrii changes igb driver to utilize diff_by_scaled_ppm() implementation
over an open-coded version.

Dawid adds pci_error_handlers for reset_prepare and reset_done for
igbvf.

Sasha removes unnecessary code in igc.
====================

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

Merge branch 'ipv4-address-protocol'

Petr Machata says:

====================
net: Allow changing IPv4 address protocol

IPv4 and IPv6 addresses can be assigned a protocol value that indicates the
provenance of the IP address. The attribute is modeled after ip route
protocols, and essentially allows the administrator or userspace stack to
tag addresses in some way that makes sense to the actor in question.

When IP address protocol field was added in commit 47f0bd503210 ("net: Add
new protocol attribute to IP addresses"), the semantics included the
ability to change the protocol for IPv6 addresses, but not for IPv4
addresses. It seems this was not deliberate, but rather by accident.

One particular use case is tagging the addresses differently depending on
whether the routing stack should advertise them or not. Without support for
protocol replacement, this can not be done.

In this patchset, extend IPv4 to allow changing the protocol defined at an
address (in patch #1). Then in patches #2 and #3 add selftest coverage for
ip address protocols.

Currently the kernel simply ignores the new value. Thus allowing the
replacement changes the observable behavior. However, since IPv6 already
behaves like this anyway, and since the feature as such is relatively new,
it seems like the change is safe to make.

An example session with the feature in action:

bash-5.2# ip address add dev d 192.0.2.1/28 proto 0xab
bash-5.2# ip address show dev d
4: d: <BROADCAST,NOARP> mtu 1500 qdisc noop state DOWN group default qlen 1000
    link/ether 06:29:74:fd:1f:eb brd ff:ff:ff:ff:ff:ff
    inet 192.0.2.1/28 scope global proto 0xab d
       valid_lft forever preferred_lft forever

bash-5.2# ip address replace dev d 192.0.2.1/28 proto 0x11
bash-5.2# ip address show dev d
4: d: <BROADCAST,NOARP> mtu 1500 qdisc noop state DOWN group default qlen 1000
    link/ether 06:29:74:fd:1f:eb brd ff:ff:ff:ff:ff:ff
    inet 192.0.2.1/28 scope global proto 0x11 d
       valid_lft forever preferred_lft forever
====================

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

selftests: rtnetlink: Add an address proto test

Add coverage of "ip address {add,replace} ... proto" support.

Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests: rtnetlink: Make the set of tests to run configurable

Extract the list of all tests into a variable, ALL_TESTS. Then assume the
environment variable TESTS holds the list of tests to actually run, falling
back to ALL_TESTS if TESTS is empty. This is the same interface that
forwarding selftests use to make the set of tests to run configurable.
In addition to this, allow setting the value explicitly through a command
line option "-t" along the lines of what fib_nexthops.sh does.

Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipv4: Allow changing IPv4 address protocol

When IP address protocol field was added in commit 47f0bd503210 ("net: Add
new protocol attribute to IP addresses"), the semantics included the
ability to change the protocol for IPv6 addresses, but not for IPv4
addresses. It seems this was not deliberate, but rather by accident.

A userspace that wants to change the protocol of an address might drop and
recreate the address, but that disrupts routing and is just impractical.

So in this patch, when an IPv4 address is replaced (through RTM_NEWADDR
request with NLM_F_REPLACE flag), update the proto at the address to the
one given in the request, or zero if none is given. This matches the
behavior of IPv6. Previously, any new value given was simply ignored.

Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'net-ipa-fully-support-ipa-v5-0'

Alex Elder says:

====================
net: ipa: fully support IPA v5.0

At long last, add the IPA and GSI register definitions, and the
configuration data required to support IPA v5.0.  This enables IPA
support for the Qualcomm SDX65 SoC.

The first version of this series had build errors due to a
non-existent source file being required.  This version addresses
that by changing how required files are specified in the Makefile.

Note that the second patch has some warnings about lines starting
with spaces; those spaces align text with the open parenthesis on
the previous line.
====================

Link: https://lore.kernel.org/r/20230321182644.2143990-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: add IPA v5.0 configuration data

Add the configuration data required for IPA v5.0, which is used in
the SDX65 SoC.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: add IPA v5.0 GSI register definitions

Add the definitions of GSI register offsets and fields for IPA v5.0.
These are used for the SDX65 SoC.  Increase the maximum channel and
event ring counts supported by the driver, so those implemented by
the SDX65 are supported.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: add IPA v5.0 register definitions

Add the definitions of IPA register offsets and fields for IPA v5.0.
These are used for the SDX65 SoC.

In the Makefile, split IPA_VERSIONS to use IPA_REG_VERSIONS and
IPA_DATA_VERSIONS instead, to allow IPA register definitions for a
new version to be added separate from the IPA data.

Rename GSI_IPA_VERSIONS to be GSI_REG_VERSIONS for consistency.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'quirk-for-oem-sfp-2-5g-t-copper-module'

Russell King says:

====================
Quirk for OEM SFP-2.5G-T copper module

Frank Wunderlich reports that this copper module requires a quirk in
order to function - in that the module needs to use 2500base-X.
Moreover, negotiation must be disabled.

An example of this device would be:

  https://www.optcore.net/product/sfp-2g-t-gen
====================

Link: https://lore.kernel.org/r/ZBniMlTDZJQ242DP@shell.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sfp: add quirk for 2.5G copper SFP

Add a quirk for a copper SFP that identifies itself as "OEM"
"SFP-2.5G-T". This module's PHY is inaccessible, and can only run
at 2500base-X with the host without negotiation. Add a quirk to
enable the 2500base-X interface mode with 2500base-T support, and
disable autonegotiation.

Reported-by: Frank Wunderlich <frank-w@public-files.de>
Tested-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sfp-bus: allow SFP quirks to override Autoneg and pause bits

Allow SFP quirks to override the Autoneg, Pause and Asym_Pause bits in
the support mask.

Some modules have an inaccessible PHY on which is only accessible via
2500base-X without Autonegotiation. We therefore want to be able to
clear the Autoneg bit. Rearrange sfp_parse_support() to allow a SFP
modes quirk to override this bit.

Tested-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-remove-some-skb_mac_header-assumptions'

Eric Dumazet says:

====================
net: remove some skb_mac_header assumptions

Historically, we tried o maintain skb_mac_header available in most of
networking paths.

When reaching ndo_start_xmit() handlers, skb_mac_header() should always
be skb->data.

With recent additions of skb_mac_header_was_set() and
DEBUG_NET_WARN_ON_ONCE() in skb_mac_header(), we can attempt
to remove our reliance on skb_mac_header in TX paths.

When this effort completes we will remove skb_reset_mac_header()
from __dev_queue_xmit() and replace it by
skb_unset_mac_header() on DEBUG_NET builds.
====================

Link: https://lore.kernel.org/r/20230321164519.1286357-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/sched: remove two skb_mac_header() uses

tcf_mirred_act() and tcf_mpls_act() can use skb_network_offset()
instead of relying on skb_mac_header().

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

sch_cake: do not use skb_mac_header() in cake_overhead()

We want to remove our use of skb_mac_header() in tx paths,
eg remove skb_reset_mac_header() from __dev_queue_xmit().

Idea is that ndo_start_xmit() can get the mac header
simply looking at skb->data.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: do not use skb_mac_header() in qdisc_pkt_len_init()

We want to remove our use of skb_mac_header() in tx paths,
eg remove skb_reset_mac_header() from __dev_queue_xmit().

Idea is that ndo_start_xmit() can get the mac header
simply looking at skb->data.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'remove-phylink_state-s-an_enabled-member'

Russell King says:

====================
Remove phylink_state's an_enabled member

Now that all the fixes and correctness patches have been merged, it is
time to switch the two users that make use of .an_enabled to check the
Autoneg bit in the advertising mask, and finally remove the
.an_enabled member.

The first two patches remove the last uses of .an_enabled, which are
in DPAA2 and XPCS. The final patch removes the member.
====================

Link: https://lore.kernel.org/r/ZBnT6yW9UY1sAsiy@shell.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: phylink: remove an_enabled

The Autoneg bit in the advertising bitmap and state->an_enabled are
always identical. state->an_enabled is now no longer used by any
drivers, so lets kill this duplication.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: pcs: xpcs: use Autoneg bit rather than an_enabled

The Autoneg bit in the advertising bitmap and state->an_enabled are
always identical. Thus, we will be removing state->an_enabled.

Use the Autoneg bit in the advertising bitmap to indicate whether
autonegotiation should be used, rather than using the an_enabled
member which will be going away.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dpaa2-mac: use Autoneg bit rather than an_enabled

The Autoneg bit in the advertising bitmap and state->an_enabled are
always identical. Thus, we will be removing state->an_enabled.

Use the Autoneg bit in the advertising bitmap to indicate whether
autonegotiation should be used, rather than using the an_enabled
member which will be going away. This means we use the same condition
as phylink_mii_c22_pcs_config().

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netdev: Enforce index cap in netdev_get_tx_queue

When requesting a TX queue at a given index, warn on out-of-bounds
referencing if the index is greater than the allocated number of
queues.

Specifically, since this function is used heavily in the networking
stack use DEBUG_NET_WARN_ON_ONCE to avoid executing a new branch on
every packet.

Signed-off-by: Nick Child <nnac123@linux.ibm.com>
Link: https://lore.kernel.org/r/20230321150725.127229-2-nnac123@linux.ibm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: Catch invalid index in XPS mapping

When setting the XPS value of a TX queue, warn the user once if the
index of the queue is greater than the number of allocated TX queues.

Previously, this scenario went uncaught. In the best case, it resulted
in unnecessary allocations. In the worst case, it resulted in
out-of-bounds memory references through calls to `netdev_get_tx_queue(
dev, index)`. Therefore, it is important to inform the user but not
worth returning an error and risk downing the netdevice.

Signed-off-by: Nick Child <nnac123@linux.ibm.com>
Reviewed-by: Piotr Raczynski <piotr.raczynski@intel.com>
Link: https://lore.kernel.org/r/20230321150725.127229-1-nnac123@linux.ibm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'add-cpswxg-sgmii-support-for-j7200-and-j721e'

Siddharth Vadapalli says:

====================
Add CPSWxG SGMII support for J7200 and J721E

This series adds support to configure the CPSW Ethernet Switch in SGMII
mode, using the am65-cpsw-nuss driver. SGMII mode is supported by the
CPSWxG instances on TI's J7200 and J721E SoCs. Thus, SGMII mode is added
in the list of extra_modes for the appropriate compatibles corresponding
to the aforementioned SoCs.

Additionally, the method of setting the supported interface via struct
"phylink_config" is simplified by converting the IF/ELSE statements to
SWITCH statements.
====================

Link: https://lore.kernel.org/r/20230321111958.2800005-1-s-vadapalli@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: ti: am65-cpsw: Enable SGMII mode for J721E

TI's J721E SoC supports SGMII mode with the CPSW9G instance of the CPSW
Ethernet Switch. Thus, enable it by adding SGMII mode to the extra_modes
member of the "j721e_cpswxg_pdata" SoC data.

Signed-off-by: Siddharth Vadapalli <s-vadapalli@ti.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: ti: am65-cpsw: Enable SGMII mode for J7200

TI's J7200 SoC supports SGMII mode with the CPSW5G instance of the CPSW
Ethernet Switch. Thus, enable it by adding SGMII mode to the extra_modes
member of the "j7200_cpswxg_pdata" SoC data.

Signed-off-by: Siddharth Vadapalli <s-vadapalli@ti.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: ti: am65-cpsw: Add support for SGMII mode

Add support for configuring the CPSW Ethernet Switch in SGMII mode.

Depending on the SoC, allow selecting SGMII mode as a supported interface,
based on the compatible used.

Signed-off-by: Siddharth Vadapalli <s-vadapalli@ti.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: ti: am65-cpsw: Simplify setting supported interface

Convert the existing IF/ELSE statement based approach of setting the
supported_interfaces member of struct "phylink_config", to SWITCH
statements. This will help scale to newer PHY-MODES as well as newer
compatibles.

Signed-off-by: Siddharth Vadapalli <s-vadapalli@ti.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

liquidio: remove unused IQ_INSTR_MODE_64B function

clang with W=1 reports
drivers/net/ethernet/cavium/liquidio/request_manager.c:43:19: error:
  unused function 'IQ_INSTR_MODE_64B' [-Werror,-Wunused-function]
static inline int IQ_INSTR_MODE_64B(struct octeon_device *oct, int iq_no)
                  ^
This function and its macro wrapper are not used, so remove them.

Signed-off-by: Tom Rix <trix@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230321184811.1827306-1-trix@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge tag 'ipsec-libreswan-mlx5' of https://git./linux/kernel/git/mellanox/linux

Leon Romanovsky says:

====================
Extend packet offload to fully support libreswan

The following patches are an outcome of Raed's work to add packet
offload support to libreswan [1].

The series includes:
 * Priority support to IPsec policies
 * Statistics per-SA (visible through "ip -s xfrm state ..." command)
 * Support to IKE policy holes
 * Fine tuning to acquire logic.

[1] https://github.com/libreswan/libreswan/pull/986
Link: https://lore.kernel.org/all/cover.1678714336.git.leon@kernel.org
* tag 'ipsec-libreswan-mlx5' of https://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux:
  net/mlx5e: Update IPsec per SA packets/bytes count
  net/mlx5e: Use one rule to count all IPsec Tx offloaded traffic
  net/mlx5e: Support IPsec acquire default SA
  net/mlx5e: Allow policies with reqid 0, to support IKE policy holes
  xfrm: copy_to_user_state fetch offloaded SA packets/bytes statistics
  xfrm: add new device offload acquire flag
  net/mlx5e: Use chains for IPsec policy priority offload
  net/mlx5: fs_core: Allow ignore_flow_level on TX dest
  net/mlx5: fs_chains: Refactor to detach chains from tc usage
====================

Link: https://lore.kernel.org/r/20230320094722.1009304-1-leon@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: net: Drop unneeded quotes

Cleanup bindings dropping unneeded quotes. Once all these are fixed,
checking for this can be enabled in yamllint.

Acked-by: Marc Kleine-Budde <mkl@pengutronix.de> # for bindings/net/can
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Steen Hegelund <Steen.Hegelund@microchip.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Acked-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp> # for bindings/net/toshiba,visconti-dwmac.yaml
Reviewed-by: Heiko Stuebner <heiko@sntech.de> #rockchip
Link: https://lore.kernel.org/r/20230320233758.2918972-1-robh@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-dsa-b53-configure-6318-and-63268-rgmii-ports'

Álvaro Fernández Rojas says:

====================
net: dsa: b53: configure 6318 and 63268 RGMII ports

BCM6318 and BCM63268 need special configuration for their RGMII ports, so we
need to be able to identify them as a special BCM63xx switch.
In the meantime, let's add some missing BCM63xx SoCs to B53 MMAP device table.

This should be applied after "net: dsa: b53: add support for BCM63xx RGMIIs":
https://patchwork.kernel.org/project/netdevbpf/patch/20230319220805.124024-1-noltari@gmail.com/

Álvaro Fernández Rojas (4):
  dt-bindings: net: dsa: b53: add more 63xx SoCs
  net: dsa: b53: mmap: add more 63xx SoCs
  net: dsa: b53: mmap: allow passing a chip ID
  net: dsa: b53: add BCM63268 RGMII configuration

 .../devicetree/bindings/net/dsa/brcm,b53.yaml |  3 ++
 drivers/net/dsa/b53/b53_common.c              | 19 +++++++++++-
 drivers/net/dsa/b53/b53_mmap.c                | 29 +++++++++++++++----
 drivers/net/dsa/b53/b53_priv.h                |  9 +++++-
 drivers/net/dsa/b53/b53_regs.h                |  1 +
 5 files changed, 53 insertions(+), 8 deletions(-)
====================

Link: https://lore.kernel.org/r/20230321173359.251778-1-noltari@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: b53: add BCM63268 RGMII configuration

BCM63268 requires special RGMII configuration to work.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: b53: mmap: allow passing a chip ID

BCM6318 and BCM63268 SoCs require a special handling for their RGMIIs, so we
should be able to identify them as a special BCM63xx switch.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: b53: mmap: add more 63xx SoCs

BCM6318, BCM6362 and BCM63268 are SoCs with a B53 MMAP switch.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: net: dsa: b53: add more 63xx SoCs

BCM6318, BCM6362 and BCM63268 are SoCs with a B53 MMAP switch.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ethernet: remove superfluous clearing of phydev

phy_disconnect() calls phy_detach() which already clears 'phydev' if it
is attached to a struct net_device.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20230321131745.27688-1-wsa+renesas@sang-engineering.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'bnxt-ptp-optimizations'

Pavan Chebbi says:

====================
bnxt PTP optimizations

Patches to
1. Enforce software based freq adjustments only on shared PHC NIC

2. A prerequisite change to expand capability storage field to
   accommodate more Firmware reported capabilities
====================

Link: https://lore.kernel.org/r/20230321144449.15289-1-pavan.chebbi@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt: Enforce PTP software freq adjustments only when in non-RTC mode

Currently driver performs software based frequency adjustments
when RTC capability is not discovered or when in shared PHC mode.
But there may be some old firmware versions that still support
hardware freq adjustments without RTC capability being exposed.
In this situation driver will use non-realtime mode even on single
host NICs.

Hence enforce software frequency adjustments only when running in
shared PHC mode. Make suitable changes for cyclecounter for the
same.

Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Acked-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt: Defer PTP initialization to after querying function caps

Driver uses the flag BNXT_FLAG_MULTI_HOST to determine whether
to use non-realtime mode PHC when running on a multi host NIC.
However when ptp initializes on a NIC with shared PHC, we still
don't have this flag set yet because HWRM_FUNC_QCFG is issued
much later.

Move the ptp initialization code after we have issued func_qcfg.
The next patch will use the BNXT_FLAG_MULTI_HOST flag during PTP
initialization.

Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Acked-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnxt: Change fw_cap to u64 to accommodate more capability bits

The current fw_cap field (u32) has run out of bits to save any
new capability.

Change the field to u64.

Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Acked-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netfilter: keep conntrack reference until IPsecv6 policy checks are done

Keep the conntrack reference until policy checks have been performed for
IPsec V6 NAT support, just like ipv4.

The reference needs to be dropped before a packet is
queued to avoid having the conntrack module unloadable.

Fixes: 58a317f1061c ("netfilter: ipv6: add IPv6 NAT support")
Signed-off-by: Madhu Koriginja <madhu.koriginja@nxp.com>
Signed-off-by: Florian Westphal <fw@strlen.de>

xtables: move icmp/icmpv6 logic to xt_tcpudp

icmp/icmp6 matches are baked into ip(6)_tables.ko.

This means that even if iptables-nft is used, a rule like
"-p icmp --icmp-type 1" will load the ip(6)tables modules.

Move them to xt_tcpdudp.ko instead to avoid this.

This will also allow to eventually add kconfig knobs to build kernels
that support iptables-nft but not iptables-legacy (old set/getsockopt
interface).

Signed-off-by: Florian Westphal <fw@strlen.de>

netfilter: xtables: disable 32bit compat interface by default

This defaulted to 'y' because before this knob existed the 32bit
compat layer was always compiled in if CONFIG_COMPAT was set.

32bit iptables on 64bit kernel isn't common anymore, so remove
the default-y now.

Signed-off-by: Florian Westphal <fw@strlen.de>

netfilter: nft_masq: deduplicate eval call-backs

nft_masq has separate ipv4 and ipv6 call-backs which share much of their
code, and an inet one switch containing a switch that calls one of the
others based on the family of the packet.  Merge the ipv4 and ipv6 ones
into the inet one in order to get rid of the duplicate code.

Const-qualify the `priv` pointer since we don't need to write through it.

Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Florian Westphal <fw@strlen.de>

netfilter: nft_redir: use `struct nf_nat_range2` throughout and deduplicate eval call-backs

`nf_nat_redirect_ipv4` takes a `struct nf_nat_ipv4_multi_range_compat`,
but converts it internally to a `struct nf_nat_range2`.  Change the
function to take the latter, factor out the code now shared with
`nf_nat_redirect_ipv6`, move the conversion to the xt_REDIRECT module,
and update the ipv4 range initialization in the nft_redir module.

Replace a bare hex constant for 127.0.0.1 with a macro.

Remove `WARN_ON`.  `nf_nat_setup_info` calls `nf_ct_is_confirmed`:

/* Can't setup nat info for confirmed ct. */
if (nf_ct_is_confirmed(ct))
return NF_ACCEPT;

This means that `ct` cannot be null or the kernel will crash, and
implies that `ctinfo` is `IP_CT_NEW` or `IP_CT_RELATED`.

nft_redir has separate ipv4 and ipv6 call-backs which share much of
their code, and an inet one switch containing a switch that calls one of
the others based on the family of the packet.  Merge the ipv4 and ipv6
ones into the inet one in order to get rid of the duplicate code.

Const-qualify the `priv` pointer since we don't need to write through
it.

Assign `priv->flags` to the range instead of OR-ing it in.

Set the `NF_NAT_RANGE_PROTO_SPECIFIED` flag once during init, rather
than on every eval.

Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: Florian Westphal <fw@strlen.de>

selftests/net: Add SHA256 computation over data sent in tcp_mmap

Add option to compute and send SHA256 over data sent (-i).

This is to ensure the correctness of data received.
Data is randomly populated from /dev/urandom.

Tested:
./tcp_mmap -s -z -i
./tcp_mmap -z -H $ADDR -i
SHA256 is correct

./tcp_mmap -s -i
./tcp_mmap -H $ADDR -i
SHA256 is correct

Signed-off-by: Coco Li <lixiaoyan@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230321081202.2370275-2-lixiaoyan@google.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net-zerocopy: Reduce compound page head access

When compound pages are enabled, although the mm layer still
returns an array of page pointers, a subset (or all) of them
may have the same page head since a max 180kb skb can span 2
hugepages if it is on the boundary, be a mix of pages and 1 hugepage,
or fit completely in a hugepage. Instead of referencing page head
on all page pointers, use page length arithmetic to only call page
head when referencing a known different page head to avoid touching
a cold cacheline.

Tested:
See next patch with changes to tcp_mmap

Correntess:
On a pair of separate hosts as send with MSG_ZEROCOPY will
force a copy on tx if using loopback alone, check that the SHA
on the message sent is equivalent to checksum on the message received,
since the current program already checks for the length.

echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages
./tcp_mmap -s -z
./tcp_mmap -H $DADDR -z

SHA256 is correct
received 2 MB (100 % mmap'ed) in 0.005914 s, 2.83686 Gbit
  cpu usage user:0.001984 sys:0.000963, 1473.5 usec per MB, 10 c-switches

Performance:
Run neper between adjacent hosts with the same config
tcp_stream -Z --skip-rx-copy -6 -T 20 -F 1000 --stime-use-proc --test-length=30

Before patch: stime_end=37.670000
After patch: stime_end=30.310000

Signed-off-by: Coco Li <lixiaoyan@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230321081202.2370275-1-lixiaoyan@google.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

sh_eth: remove open coded netif_running()

It had a purpose back in the days, but today we have a handy helper.

Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20230321065826.2044-1-wsa+renesas@sang-engineering.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: ethernet: ti: am65-cpts: adjust estf following ptp changes

When the CPTS clock is synced/adjusted by running linuxptp (ptp4l/phc2sys),
it will cause the TSN EST schedule to drift away over time. This is because
the schedule is driven by the EstF periodic counter whose pulse length is
defined in ref_clk cycles and it does not automatically sync to CPTS clock.
   _______
 _|
  ^
  expected cycle start time boundary
   _______________
 _|_|___|_|
  ^
  EstF drifted away -> direction

To fix it, the same PPM adjustment has to be applied to EstF as done to the
PHC CPTS clock, in order to correct the TSN EST cycle length and keep them
in sync.

Drifted cycle:
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230373377017
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230373877017
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230374377017
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230374877017
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230375377017
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230375877023
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230376377018
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230376877018
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635968230377377018

Stable cycle:
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635966863193375473
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635966863193875473
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635966863194375473
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635966863194875473
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635966863195375473
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635966863195875473
AM65_CPTS_EVT: 7 e1:01770001 e2:000000ff t:1635966863196375473

Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
Signed-off-by: Siddharth Vadapalli <s-vadapalli@ti.com>
Reviewed-by: Roger Quadros <rogerq@kernel.org>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Link: https://lore.kernel.org/r/20230321062600.2539544-1-s-vadapalli@ti.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net-sysfs: display two backlog queue len separately

Sometimes we need to know which one of backlog queue can be exactly
long enough to cause some latency when debugging this part is needed.
Thus, we can then separate the display of both.

Signed-off-by: Jason Xing <kernelxing@tencent.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230321015746.96994-1-kerneljasonxing@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

virtio/vsock: check transport before skb allocation

Pointer to transport could be checked before allocation of skbuff, thus
there is no need to free skbuff when this pointer is NULL.

Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
Reviewed-by: Bobby Eshleman <bobby.eshleman@bytedance.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Link: https://lore.kernel.org/r/08d61bef-0c11-c7f9-9266-cb2109070314@sberdevices.ru
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

tools: ynl: skip the explicit op array size when not needed

Jiri suggests it reads more naturally to skip the explicit
array size when possible. When we export the symbol we want
to make sure that the size is right but for statics its
not needed.

Link: https://lore.kernel.org/r/20230321044159.1031040-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-remove-some-rcu_bh-cruft'

Eric Dumazet says:

====================
net: remove some rcu_bh cruft

There is no point using rcu_bh variant hoping to free objects faster,
especially hen using call_rcu() or kfree_rcu().

Disabling/enabling BH has a non-zero cost, and adds distracting
hot spots in kernel profiles eg. in ip6_xmit().
====================

Link: https://lore.kernel.org/r/20230321040115.787497-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: remove rcu_dereference_bh_rtnl()

This helper is no longer used in the tree.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

neighbour: switch to standard rcu, instead of rcu_bh

rcu_bh is no longer a win, especially for objects freed
with standard call_rcu().

Switch neighbour code to no longer disable BH when not necessary.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ipv6: flowlabel: do not disable BH where not needed

struct ip6_flowlabel are rcu managed, and call_rcu() is used
to delay fl_free_rcu() after RCU grace period.

There is no point disabling BH for pure RCU lookups.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: atheros: atl1c: remove unused atl1c_irq_reset function

clang with W=1 reports
drivers/net/ethernet/atheros/atl1c/atl1c_main.c:214:20: error:
  unused function 'atl1c_irq_reset' [-Werror,-Wunused-function]
static inline void atl1c_irq_reset(struct atl1c_adapter *adapter)
                   ^
This function is not used, so remove it.

Signed-off-by: Tom Rix <trix@redhat.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Link: https://lore.kernel.org/r/20230320232317.1729464-1-trix@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

igc: Remove obsolete DMA coalescing code

DMA coalescing is not applicable for i225 parts. This patch comes to tidy
up the driver code.

Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
Tested-by: Naama Meir <naamax.meir@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

igbvf: add PCI reset handler functions

There was a problem with resuming ping after conducting a PCI reset.
This commit adds two functions, igbvf_io_prepare and igbvf_io_done,
which, after being added to the pci_error_handlers struct,
will prepare the drivers for a PCI reset and then bring the interface up
and reset it after. This will prevent the driver from ending up in
incorrect state. Test_and_set_bit is highly reliable in this context,
so we are not including a timeout in this commit

This introduces 900ms - 1100ms of overhead to this operation but it's in
non-time-critical flow. And also allows the driver to continue
functioning after the reset.

Functionality documented in ethernet-controller-i350-datasheet 4.2.1.3
https://www.intel.com/content/www/us/en/products/details/ethernet/gigabit-controllers/i350-controllers/docs.html

Signed-off-by: Dawid Wesierski <dawidx.wesierski@intel.com>
Signed-off-by: Kamil Maziarz <kamil.maziarz@intel.com>
Reviewed-by: Michal Kubiak <michal.kubiak@intel.com>
Tested-by: Marek Szlosek <marek.szlosek@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

igb: refactor igb_ptp_adjfine_82580 to use diff_by_scaled_ppm

Driver's .adjfine interface functions use adjust_by_scaled_ppm and
diff_by_scaled_ppm introduced in commit 1060707e3809
("ptp: introduce helpers to adjust by scaled parts per million")
to calculate the required adjustment in a concise manner,
but not igb_ptp_adjfine_82580.
Fix it by introducing IGB_82580_BASE_PERIOD and changing function logic
to use diff_by_scaled_ppm.

Signed-off-by: Andrii Staikov <andrii.staikov@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>

net: pasemi: Fix return type of pasemi_mac_start_tx()

With clang's kernel control flow integrity (kCFI, CONFIG_CFI_CLANG),
indirect call targets are validated against the expected function
pointer prototype to make sure the call target is valid to help mitigate
ROP attacks. If they are not identical, there is a failure at run time,
which manifests as either a kernel panic or thread getting killed. A
warning in clang aims to catch these at compile time, which reveals:

  drivers/net/ethernet/pasemi/pasemi_mac.c:1665:21: error: incompatible function pointer types initializing 'netdev_tx_t (*)(struct sk_buff *, struct net_device *)' (aka 'enum netdev_tx (*)(struct sk_buff *, struct net_device *)') with an expression of type 'int (struct sk_buff *, struct net_device *)' [-Werror,-Wincompatible-function-pointer-types-strict]
          .ndo_start_xmit         = pasemi_mac_start_tx,
                                    ^~~~~~~~~~~~~~~~~~~
  1 error generated.

->ndo_start_xmit() in 'struct net_device_ops' expects a return type of
'netdev_tx_t', not 'int'. Adjust the return type of
pasemi_mac_start_tx() to match the prototype's to resolve the warning.
While PowerPC does not currently implement support for kCFI, it could in
the future, which means this warning becomes a fatal CFI failure at run
time.

Link: https://github.com/ClangBuiltLinux/linux/issues/1750
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Link: https://lore.kernel.org/r/20230319-pasemi-incompatible-pointer-types-strict-v1-1-1b9459d8aef0@kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: geneve: accept every ethertype

The Geneve encapsulation, as defined in RFC 8926, has a Protocol Type
field, which states the Ethertype of the payload appearing after the
Geneve header.

Commit 435fe1c0c1f7 ("net: geneve: support IPv4/IPv6 as inner protocol")
introduced a new IFLA_GENEVE_INNER_PROTO_INHERIT flag that allowed the
use of other Ethertypes than Ethernet. However, it did not get rid of a
restriction that prohibits receiving payloads other than Ethernet,
instead the commit white-listed additional Ethertypes, IPv4 and IPv6.

This patch removes this restriction, making it possible to receive any
Ethertype as a payload, if the IFLA_GENEVE_INNER_PROTO_INHERIT flag is
set.

The restriction was set in place back in commit 0b5e8b8eeae4
("net: Add Geneve tunneling protocol driver"), which implemented a
protocol layer driver for Geneve to be used with Open vSwitch. The
relevant discussion about introducing the Ethertype white-list can be
found here:
https://lore.kernel.org/netdev/CAEP_g=_1q3ACX5NTHxLDnysL+dTMUVzdLpgw1apLKEdDSWPztw@mail.gmail.com/

<quote>
>> +       if (unlikely(geneveh->proto_type != htons(ETH_P_TEB)))
>
> Why? I thought the point of geneve carrying protocol field was to
> allow protocols other than Ethernet... is this temporary maybe?

Yes, it is temporary. Currently OVS only handles Ethernet packets but
this restriction can be lifted once we have a consumer that is capable
of handling other protocols.
</quote>

This white-list was then ported to a generic Geneve netdevice in commit
371bd1061d29 ("geneve: Consolidate Geneve functionality in single
module."). Preserving the Ethertype white-list at this point made sense,
as the Geneve device could send out only Ethernet payloads anyways.

However, now that the Geneve netdevice supports encapsulating other
payloads with IFLA_GENEVE_INNER_PROTO_INHERIT and we have a consumer
capable of other protocols, it seems appropriate to lift the restriction
and allow any Geneve payload to be received.

Signed-off-by: Josef Miegl <josef@miegl.cz>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Eyal Birger <eyal.birger@gmail.com>
Link: https://lore.kernel.org/r/20230319220954.21834-1-josef@miegl.cz
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dsa: b53: add support for BCM63xx RGMIIs

BCM63xx RGMII ports require additional configuration in order to work.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20230319220805.124024-1-noltari@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dsa: qca8k: remove assignment of an_enabled in pcs_get_state()

pcs_get_state() implementations are not supposed to alter an_enabled.
Remove this assignment.

Fixes: b3591c2a3661 ("net: dsa: qca8k: Switch to PHYLINK instead of PHYLIB")
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/E1pdsE5-00Dl2l-8F@rmk-PC.armlinux.org.uk
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: dsa: mv88e6xxx: fix mdio bus' phy_mask member

Commit 2c7e46edbd03 ("net: dsa: mv88e6xxx: mask apparently non-existing
phys during probing") added non-trivial bus->phy_mask in
mv88e6xxx_mdio_register() in order to avoid excessive mdio bus
transactions during probing.

But the mask is incorrect for switches with non-zero phy_base_addr (such
as 88E6341).

Fix this.

Fixes: 2c7e46edbd03 ("net: dsa: mv88e6xxx: mask apparently non-existing phys during probing")
Signed-off-by: Marek Behún <kabel@kernel.org>
Tested-by: Klaus Kudielka <klaus.kudielka@gmail.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Link: https://lore.kernel.org/r/20230319140238.9470-1-kabel@kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: cxgb3: remove unused fl_to_qset function

clang with W=1 reports
drivers/net/ethernet/chelsio/cxgb3/sge.c:169:32: error: unused function
  'fl_to_qset' [-Werror,-Wunused-function]
static inline struct sge_qset *fl_to_qset(const struct sge_fl *q, int qidx)
                               ^
This function is not used, so remove it.

Signed-off-by: Tom Rix <trix@redhat.com>
Reviewed-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Link: https://lore.kernel.org/r/20230319172433.1708161-1-trix@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-ethernet-mtk_eth_soc-various-enhancements'

Daniel Golle says:

====================
net: ethernet: mtk_eth_soc: various enhancements

This series brings a variety of fixes and enhancements for mtk_eth_soc,
adds support for the MT7981 SoC and facilitates sharing the SGMII PCS
code between mtk_eth_soc and mt7530.

The whole series has been tested on MT7622+MT7531 (BPi-R64),
MT7623+MT7530 (BPi-R2), MT7981+GPY211 (GL.iNet GL-MT3000) and
MT7986+MT7531 (BPi-R3). On the BananaPi R3 a variete of SFP modules
have been tested, all of them (some SGMII with PHY, others 2500Base-X
or 1000Base-X without PHY) are working well now, however, some of them
need manually disabling of autonegotiation for the link to come up.
====================

Link: https://lore.kernel.org/r/cover.1679230025.git.daniel@makrotopia.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: mt7530: use external PCS driver

Implement regmap access wrappers, for now only to be used by the
pcs-mtk-lynxi driver.
Make use of this external PCS driver and drop the now reduntant
implementation in mt7530.c.
As a nice side effect the SGMII registers can now also more easily be
inspected for debugging via /sys/kernel/debug/regmap.

Tested-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: mtk_eth_soc: switch to external PCS driver

Now that we got a PCS driver, use it and remove the now redundant
PCS code and it's header macros from the Ethernet driver.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: pcs: add driver for MediaTek SGMII PCS

The SGMII core found in several MediaTek SoCs is identical to what can
also be found in MediaTek's MT7531 Ethernet switch IC.
As this has not always been clear, both drivers developed different
implementations to deal with the PCS.
Recently Alexander Couzens pointed out this fact which lead to the
development of this shared driver.

Add a dedicated driver, mostly by copying the code now found in the
Ethernet driver. The now redundant code will be removed by a follow-up
commit.

Suggested-by: Alexander Couzens <lynxis@fe80.eu>
Suggested-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Frank Wunderlich <frank-w@public-files.de>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: mtk_eth_soc: ppe: add support for flow accounting

The PPE units found in MT7622 and newer support packet and byte
accounting of hw-offloaded flows. Add support for reading those counters
as found in MediaTek's SDK[1].

[1]: https://git01.mediatek.com/plugins/gitiles/openwrt/feeds/mtk-openwrt-feeds/+/bc6a6a375c800dc2b80e1a325a2c732d1737df92
Tested-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: mtk_eth_soc: set MDIO bus clock frequency

Set MDIO bus clock frequency and allow setting a custom maximum
frequency from device tree.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: net: pcs: mediatek,sgmiisys: add MT7981 SoC

Add mediatek,pnswap boolean property needed on many boards using the
MediaTek MT7981 SoC.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: arm: mediatek: sgmiisys: Convert to DT schema

Convert mediatek,sgmiiisys bindings to DT schema format.
Add maintainer Matthias Brugger, no maintainers were listed in the
original documentation.
As this node is also referenced by the Ethernet controller and used
as SGMII PCS add this fact to the description.
Move the file to Documentation/devicetree/bindings/net/pcs/ which seems
more appropriate given that the great majority of registers are related
to SGMII PCS functionality and only one register represents clock bits.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: net: mediatek,net: add mt7981-eth binding

Introduce DT bindings for the MT7981 SoC to mediatek,net.yaml.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ethernet: mtk_eth_soc: add support for MT7981 SoC

The MediaTek MT7981 SoC comes with two 1G/2.5G SGMII ports, just like
MT7986.

In addition MT7981 is equipped with a built-in 1000Base-T PHY which can
be used with GMAC1.

As many MT7981 boards make use of inverting SGMII signal polarity, add
new device-tree attribute 'mediatek,pn_swap' to support them.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

MAINTAINERS: remove file entry in NFC SUBSYSTEM after platform_data movement

Commit 053fdaa841bd ("nfc: mrvl: Move platform_data struct into driver")
moves the nfcmrvl.h header file from include/linux/platform_data to the
driver's directory, but misses to adjust MAINTAINERS.

Hence, ./scripts/get_maintainer.pl --self-test=patterns complains about a
broken reference.

Just remove the file entry in NFC SUBSYSTEM, as the new location of the
code is already covered by another pattern in that section.

Fixes: 053fdaa841bd ("nfc: mrvl: Move platform_data struct into driver")
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

r8169: consolidate disabling ASPM before EPHY access

Now that rtl_hw_aspm_clkreq_enable() is a no-op for chip versions < 32,
we can consolidate disabling ASPM before EPHY access in rtl_hw_start().

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>