net: ethernet: rmnet: Support for ingress MAPv5 checksum offload

Adding support for processing of MAPv5 downlink packets.
It involves parsing the Mapv5 packet and checking the csum header
to know whether the hardware has validated the checksum and is
valid or not.

Based on the checksum valid bit the corresponding stats are
incremented and skb->ip_summed is marked either CHECKSUM_UNNECESSARY
or left as CHEKSUM_NONE to let network stack revalidate the checksum
and update the respective snmp stats.

Current MAPV1 header has been modified, the reserved field in the
Mapv1 header is now used for next header indication.

Signed-off-by: Sharath Chandra Vurukala <sharathv@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

docs: networking: Add documentation for MAPv5

Adding documentation explaining the new MAPv4/v5 packet formats
and the corresponding checksum offload headers.

Signed-off-by: Sharath Chandra Vurukala <sharathv@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'iwl-next' of git://git./linux/kernel/git/tnguy/linux

Tony Nguyen says:

====================
iwl-next Intel Wired LAN Driver Updates 2021-06-01

This pull request is targeting net-next and rdma-next branches.
These patches have been reviewed by netdev and rdma mailing lists[1].

This series adds RDMA support to the ice driver for E810 devices and
converts the i40e driver to use the auxiliary bus infrastructure
for X722 devices. The PCI netdev drivers register auxiliary RDMA devices
that will bind to auxiliary drivers registered by the new irdma module.

[1] https://lore.kernel.org/netdev/20210520143809.819-1-shiraz.saleem@intel.com/
---
v3:
- ice_aq_add_rdma_qsets(), ice_cfg_vsi_rdma(), ice_[ena|dis]_vsi_rdma_qset(),
and ice_cfg_rdma_fltr() no longer return ice_status
- Remove null check from ice_aq_add_rdma_qsets()

v2:
- Added patch 'i40e: Replace one-element array with flexible-array
member'

Changes since linked review (v6):
- Removed unnecessary checks in i40e_client_device_register() and
i40e_client_device_unregister()
- Simplified the i40e_client_device_register() API
====================

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

net: mdio: Fix spelling mistakes

informations  ==> information
typicaly  ==> typically
derrive  ==> derive
eventhough  ==> even though

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: usb: Fix spelling mistakes

wierdness  ==> weirdness
multicat  ==> multicast
limite  ==> limit
adddress  ==> address
operater  ==> operator
intial  ==> initial
smaler  ==> smaller
Communcation  ==> Communication
funcitons  ==> functions
everytime  ==> every time
Neigbor  ==> Neighbor
performace  ==> performance

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

vrf: Fix a typo

possibile  ==> possible

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

gtp: Fix a typo

Suppport  ==> Support

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

macvlan: Fix a typo

underlaying  ==> underlying

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: Return the correct errno code

When kalloc or kmemdup failed, should return ENOMEM rather than ENOBUF.

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: dcb: Return the correct errno code

When kalloc or kmemdup failed, should return ENOMEM rather than ENOBUF.

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

igb: Fix -Wunused-const-variable warning

If CONFIG_IGB_HWMON is n, gcc warns:

drivers/net/ethernet/intel/igb/e1000_82575.c:2765:17:
 warning: ‘e1000_emc_therm_limit’ defined but not used [-Wunused-const-variable=]
 static const u8 e1000_emc_therm_limit[4] = {
                 ^~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/intel/igb/e1000_82575.c:2759:17:
 warning: ‘e1000_emc_temp_data’ defined but not used [-Wunused-const-variable=]
 static const u8 e1000_emc_temp_data[4] = {
                 ^~~~~~~~~~~~~~~~~~~

Move it into #ifdef block to fix this.

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

cxgb4: Fix -Wunused-const-variable warning

If CONFIG_PCI_IOV is n, make W=1 warns:

drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c:3909:33:
 warning: ‘cxgb4_mgmt_ethtool_ops’ defined but not used [-Wunused-const-variable=]
 static const struct ethtool_ops cxgb4_mgmt_ethtool_ops = {
                                 ^~~~~~~~~~~~~~~~~~~~~~

Move it into #ifdef block to fix this.

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

hamradio: bpqether: Fix -Wunused-const-variable warning

If CONFIG_PROC_FS is n, gcc warns:

drivers/net/hamradio/bpqether.c:437:36:
 warning: ‘bpq_seqops’ defined but not used [-Wunused-const-variable=]
 static const struct seq_operations bpq_seqops = {
                                    ^~~~~~~~~~
Use #ifdef macro to gurad this.

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: stmmac: enable platform specific safety features

On Intel platforms, not all safety features are enabled on the hardware.
The current implementation enable all safety features by default. This
will cause mass error and warning printouts after the module is loaded.

Introduce platform specific safety features flag to enable or disable
each safety features.

Signed-off-by: Wong Vee Khee <vee.khee.wong@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

NFC: microread: Remove redundant assignment to variable err

In the case MICROREAD_CB_TYPE_READER_ALL clang reports a dead code
warning. The error code assigned to variable err is already passed
to async_cb(). The assignment is redundant and can be removed.

Addresses-Coverity: ("Unused value")
Signed-off-by: Nigel Christian <nigel.l.christian@gmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'hdlc-cleanups'

Peng Li says:

====================
net: hdlc: clean up some code style issues

This patchset clean up some code style issues.
====================

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

net: hdlc: add braces {} to all arms of the statement

Braces {} should be used on all arms of this statement.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: hdlc: move out assignment in if condition

Should not use assignment in if condition.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: hdlc: replace comparison to NULL with "!param"

According to the chackpatch.pl, comparison to NULL could
be written "!param".

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: hdlc: fix an code style issue about EXPORT_SYMBOL(foo)

According to the chackpatch.pl,
EXPORT_SYMBOL(foo); should immediately follow its function/variable.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: hdlc: fix an code style issue about "foo* bar"

Fix the checkpatch error as "foo* bar" and should be "foo *bar".

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: hdlc: add blank line after declarations

This patch fixes the checkpatch error about missing a blank line
after declarations.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: hdlc: remove redundant blank lines

This patch removes some redundant blank lines.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'act_vlan-allow-modify-zero'

Boris Sukholitko says:

====================
net/sched: act_vlan: Fix modify to allow 0

Currently vlan modification action checks existence of vlan priority by
comparing it to 0. Therefore it is impossible to modify existing vlan
tag to have priority 0.

For example, the following tc command will change the vlan id but will
not affect vlan priority:

tc filter add dev eth1 ingress matchall action vlan modify id 300 \
        priority 0 pipe mirred egress redirect dev eth2

The incoming packet on eth1:

ethertype 802.1Q (0x8100), vlan 200, p 4, ethertype IPv4

will be changed to:

ethertype 802.1Q (0x8100), vlan 300, p 4, ethertype IPv4

although the user has intended to have p == 0.

The fix is to add tcfv_push_prio_exists flag to struct tcf_vlan_params
and rely on it when deciding to set the priority.

The same flag is used to avoid dumping unset vlan priority.

Change Log:
v3 -> v4:
- revert tcf_vlan_get_fill_size change: total size calculation may race vs dump

v2 -> v3:
- Push assumes that the priority is being set
- tcf_vlan_get_fill_size accounts for priority existence

v1 -> v2:
- Do not dump unset priority and fix tests accordingly
- Test for priority 0 modification
====================

Reviewed-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/sched: act_vlan: Test priority 0 modification

Because explicitly being set, the priority 0 should appear
in the output.

Signed-off-by: Boris Sukholitko <boris.sukholitko@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/sched: act_vlan: No dump for unset priority

Dump vlan priority only if it has been previously set.

Fix the tests accordingly.

Signed-off-by: Boris Sukholitko <boris.sukholitko@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/sched: act_vlan: Fix modify to allow 0

Currently vlan modification action checks existence of vlan priority by
comparing it to 0. Therefore it is impossible to modify existing vlan
tag to have priority 0.

For example, the following tc command will change the vlan id but will
not affect vlan priority:

tc filter add dev eth1 ingress matchall action vlan modify id 300 \
        priority 0 pipe mirred egress redirect dev eth2

The incoming packet on eth1:

ethertype 802.1Q (0x8100), vlan 200, p 4, ethertype IPv4

will be changed to:

ethertype 802.1Q (0x8100), vlan 300, p 4, ethertype IPv4

although the user has intended to have p == 0.

The fix is to add tcfv_push_prio_exists flag to struct tcf_vlan_params
and rely on it when deciding to set the priority.

Fixes: 45a497f2d149a4a8061c (net/sched: act_vlan: Introduce TCA_VLAN_ACT_MODIFY vlan action)
Signed-off-by: Boris Sukholitko <boris.sukholitko@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

octeontx2-af: Fix spelling mistake "vesion" -> "version"

There is a spelling mistake in a dev_warning message. Fix it.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

NFC: nci: Remove redundant assignment to len

Variable 'len' is set to conn_info->max_pkt_payload_len but this
value is never read as it is overwritten with a new value later on,
hence it is a redundant assignment and can be removed.

Clean up the following clang-analyzer warning:

net/nfc/nci/hci.c:164:3: warning: Value stored to 'len' is never read
[clang-analyzer-deadcode.DeadStores]

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: vxge: Declare the function vxge_reset_all_vpaths as void

variable 'status' is unneeded and it's noneed to check the
return value of function vxge_reset_all_vpaths,so declare
it as void.

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

r8152: support pauseparam of ethtool_ops

Support get_pauseparam and set_pauseparam of ethtool_ops.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qlcnic: Remove the repeated declaration

Function 'qlcnic_82xx_hw_write_wx_2M' is declared twice, so remove the
repeated declaration.

Cc: Shahed Shaikh <shshaikh@marvell.com>
Cc: Manish Chopra <manishc@marvell.com>
Cc: GR-Linux-NIC-Dev@marvell.com
Signed-off-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

fjes: Use DEFINE_RES_MEM() and DEFINE_RES_IRQ() to simplify code

No functional change.

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: enetc: catch negative return code from enetc_pf_to_port()

After the refactoring introduced in commit 87614b931c24 ("net: enetc:
create a common enetc_pf_to_port helper"), enetc_pf_to_port was coded up
to return -1 in case the passed PCIe device does not have a recognized
BDF.

Make sure the -1 value is checked by the callers, to appease static
checkers.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

netpoll: don't require irqs disabled in rt kernels

write_msg(netconsole.c:836) calls netpoll_send_udp after a call to
spin_lock_irqsave, which normally disables interrupts; but in PREEMPT_RT
this call just locks an rt_mutex without disabling irqs. In this case,
netpoll_send_udp is called with interrupts enabled.

Signed-off-by: Wander Lairson Costa <wander@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'marvell-prestera-firmware-3-0'

Vadym Kochan says:

====================
Marvell Prestera Switchdev initial updates for firmware version 3.0

This series adds minimal support for firmware version 3.0 which
has such changes like:

    - initial routing support

    - LAG support

    - events interrupt handling changes

Changes just make able to work with new firmware version but
supported features in driver will be added later.

New firmware version was recently merged into linux-firmware tree.

Added ability of loading previous fw major version if the latest one
is missing, also add support for previous FW ABI.

PATCH -> RFC:
    1) Load previous fw version if the latest one is missing (suggested by Andrew Lunn)

    2) Add support for previous FW ABI version (suggested by Andrew Lunn)

RFC v2:
    1) Get rid of automatic decrementing of
       major version but hard code it.

    2) Print error message with file path if
       previous FW could not be loaded.
====================

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

net: marvell: prestera: try to load previous fw version

Lets try to load previous fw version in case the latest one is missing on
existing system.

Signed-off-by: Vadym Kochan <vkochan@marvell.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: marvell: prestera: bump supported firmware version to 3.0

New firmware version has some ABI and feature changes like:

    - LAG support
    - initial L3 support
    - changed events handling logic

Signed-off-by: Vadym Kochan <vkochan@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: marvell: prestera: align flood setting according to latest firmware version

Latest FW IPC flood message format was changed to configure uc/mc
flooding separately, so change code according to this.

Signed-off-by: Vadym Kochan <vkochan@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: marvell: prestera: disable events interrupt while handling

There are change in firmware which requires that receiver will
disable event interrupts before handling them and enable them
after finish with handling. Events still may come into the queue
but without receiver interruption.

Signed-off-by: Vadym Kochan <vkochan@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: neterion: fix doc warnings in s2io.c

Add description for may_sleep to fix the W=1 warnings:

  drivers/net/ethernet/neterion/s2io.c:1110: warning: Function parameter or member 'may_sleep' not described in 'init_tti'
  drivers/net/ethernet/neterion/s2io.c:3335: warning: Function parameter or member 'may_sleep' not described in 'wait_for_cmd_complete'
  drivers/net/ethernet/neterion/s2io.c:4881: warning: Function parameter or member 'may_sleep' not described in 's2io_set_multicast'

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'net-hns3-add-vlan-filter-control-support'

Huazhong Tan says:

====================
net: hns3: add VLAN filter control support

This patchset add VLAN filter control support for HNS3 driver.
====================

Link: https://lore.kernel.org/r/1622428725-30049-1-git-send-email-tanhuazhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: add debugfs support for vlan configuration

Add debugfs support for vlan configuraion. create a single file
"vlan_config" for it, and query it by command "cat vlan_config",
return the result to userspace.

The new display style is below:
$ cat vlan_config
I_PORT_VLAN_FILTER: on
E_PORT_VLAN_FILTER: off
FUNC_ID  I_VF_VLAN_FILTER  E_VF_VLAN_FILTER  PORT_VLAN_FILTER_BYPASS
pf       off               on                off
vf0      off               on                off
FUNC_ID  PVID    ACCEPT_TAG1  ACCEPT_TAG2  ACCEPT_UNTAG1  ACCEPT_UNTAG2
pf       0       on           on           on             on
vf0      0       on           on           on             on

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: add support for VF modify VLAN filter state

Previously, there is hardware limitation for VF to modify
the VLAN filter state, and the VLAN filter state is default
enabled. Now the limitation has been removed in some device,
so add capability flag to check whether the device supports
modify VLAN filter state. If flag on, user will be able to
modify the VLAN filter state by ethtool -K.
VF needs to send mailbox to request the PF to modify the VLAN
filter state for it.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: add query basic info support for VF

There are some features of VF depend on PF, so it's necessary
for VF to know whether PF supports. For compatibility, modify
the mailbox HCLGE_MBX_GET_TCINFO, extend its function, use to
get the basic information of PF, including mailbox api version
and PF capabilities.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: add support for modify VLAN filter state

Previously, with hardware limitation, the port VLAN filter are
effective for both PF and its VFs simultaneously, so a single
function is not able to enable/disable separately, and the VLAN
filter state is default enabled. Now some device supports each
function to bypass port VLAN filter, then each function can
switch VLAN filter separately. Add capability flag to check
whether the device supports modify VLAN filter state. If flag
on, user will be able to modify the VLAN filter state by ethtool
-K.

Furtherly, the default VLAN filter state is also changed
according to whether non-zero VLAN used. Then the device can
receive packet with any VLAN tag if only VLAN 0 used.

The function hclge_need_enable_vport_vlan_filter() is used to
help implement above changes. And the VLAN filter handle for
promisc mode can also be simplified by this function.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: refine function hclge_set_vf_vlan_cfg()

The struct hclge_vf_vlan_cfg is firstly designed for setting
VLAN filter tag. And it's reused for enable RX VLAN offload
later. It's strange to use member "is_kill" to indicate "enable".
So redefine the struct hclge_vf_vlan_cfg to adapt it.

For there are already 3 subcodes being used in function
hclge_set_vf_vlan_cfg(), use "switch-case" style for each
branch, rather than "if-else". Also simplify the assignment for
each branch to make it more clearly.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: remove unnecessary updating port based VLAN

For the PF have called hclge_update_port_base_vlan_cfg() already
before notify VF, it's unnecessary to update port based VLAN again
when received mailbox request from VF.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: refine for hclge_push_vf_port_base_vlan_info()

Use struct "hclge_vlan_info" instead of separately parameters
for function hclge_push_vf_port_base_vlan_info(), to make it
more concise.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: hns3: add 'QoS' support for port based VLAN configuration

Currently, option "qos" is igored by HNS3 driver for command
"ip link set ethx vf <vf id> vlan <vlan id> qos <qos value>".
Add support for it.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

sctp: sm_statefuns: Fix spelling mistakes

Fix some spelling mistakes in comments:
genereate ==> generate
correclty ==> correctly
boundries ==> boundaries
failes ==> fails
isses ==> issues
assocition ==> association
signe ==> sign
assocaition ==> association
managemement ==> management
restransmissions ==> retransmission
sideffect ==> sideeffect
bomming ==> booming
chukns ==> chunks
SHUDOWN ==> SHUTDOWN
violationg ==> violating
explcitly ==> explicitly
CHunk ==> Chunk

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Link: https://lore.kernel.org/r/20210601020801.3625358-1-zhengyongjun3@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

rds: Fix spelling mistakes

Fix some spelling mistakes in comments:
alloced  ==> allocated

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Reviewed-by: Håkon Bugge <haakon.bugge@oracle.com>
Link: https://lore.kernel.org/r/20210531063617.3018637-1-zhengyongjun3@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sched: Fix spelling mistakes

Fix some spelling mistakes in comments:
sevaral  ==> several
sugestion  ==> suggestion
unregster  ==> unregister
suplied  ==> supplied
cirsumstances  ==> circumstances

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Link: https://lore.kernel.org/r/20210531020048.2920054-1-zhengyongjun3@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: hci: Fix spelling mistakes

Fix some spelling mistakes in comments:
occured  ==> occurred
negociate  ==> negotiate

Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531020019.2919799-1-zhengyongjun3@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'part-2-of-sja1105-dsa-driver-preparation-for-new-switch-introduction-sja1110'

Vladimir Oltean says:

====================
Part 2 of SJA1105 DSA driver preparation for new switch introduction (SJA1110)

This series is a continuation of:
https://patchwork.kernel.org/project/netdevbpf/cover/20210524131421.1030789-1-olteanv@gmail.com/

even though it isn't the first time these patches are submitted (they
were part of the group previously called "Add NXP SJA1110 support to the
sja1105 DSA driver"):
https://patchwork.kernel.org/project/netdevbpf/cover/20210526135535.2515123-1-vladimir.oltean@nxp.com/

but I broke that up again since these patches are already reviewed, for
the most part. There are no changes compared to v2 and v1.

This series of patches contains:

- an adaptation of the driver to the new "ethernet-ports" OF node name
- an adaptation of the driver to support more than 1 SGMII port
- a generalization of the supported phy_interface_t values per port
- an adaptation to encode SPEED_10, SPEED_100, SPEED_1000 into the
  hardware registers differently depending on switch revision
- a consolidation of the PHY interface type used for RGMII and another
  one for the API exposed for sja1105_dynamic_config_read()
====================

Link: https://lore.kernel.org/r/20210530225939.772553-1-olteanv@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: sja1105: some table entries are always present when read dynamically

The SJA1105 has a static configuration comprised of a number of tables
with entries. Some of these can be read and modified at runtime as well,
through the dynamic configuration interface.

As a careful reader can notice from the comments in this file, the
software interface for accessing a table entry through the dynamic
reconfiguration is a bit of a no man's land, and varies wildly across
switch generations and even from one kind of table to another.

I have tried my best to come up with a software representation of a
'common denominator' SPI command to access a table entry through the
dynamic configuration interface:

struct sja1105_dyn_cmd {
bool search;
u64 valid; /* must be set to 1 */
u64 rdwrset; /* 0 to read, 1 to write */
u64 errors;
u64 valident; /* 0 if entry is invalid, 1 if valid */
u64 index;
};

Relevant to this patch is the VALIDENT bit, which for READ commands is
populated by the switch and lets us know if we're looking at junk or at
a real table entry.

In SJA1105, the dynamic reconfiguration interface for management routes
has notably not implemented the VALIDENT bit, leading to a workaround to
ignore this field in sja1105_dynamic_config_read(), as it will be set to
zero, but the data is valid nonetheless.

In SJA1110, this pattern has sadly been abused to death, and while there
are many more tables which can be read back over the dynamic config
interface compared to SJA1105, their handling isn't in any way more
uniform. Generally speaking, if there is a single possible entry in a
given table, and loading that table in the static config is mandatory as
per the documentation, then the VALIDENT bit is deemed as redundant and
more than likely not implemented.

So it is time to make the workaround more official, and add a bit to the
flags implemented by dynamic config tables. It will be used by more
tables when SJA1110 support arrives.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: sja1105: always keep RGMII ports in the MAC role

In SJA1105, the xMII Mode Parameters Table field called PHY_MAC denotes
the 'role' of the port, be it a PHY or a MAC. This makes a difference in
the MII and RMII protocols, but RGMII is symmetric, so either PHY or MAC
settings result in the same hardware behavior.

The SJA1110 is different, and the RGMII ports only work when configured
in MAC mode, so keep the port roles in MAC mode unconditionally.

Why we had an RGMII port in the PHY role in the first place was because
we wanted to have a way in the driver to denote whether RGMII delays
should be applied based on the phy-mode property or not. This is already
done in sja1105_parse_rgmii_delays() based on an intermediary
struct sja1105_dt_port (which contains the port role). So it is a
logical fallacy to use the hardware configuration as a scratchpad for
driver data, it isn't necessary.

We can also remove the gating condition for applying RGMII delays only
for ports in the PHY role. The .setup_rgmii_delay() method looks at
the priv->rgmii_rx_delay[port] and priv->rgmii_tx_delay[port] properties
which are already populated properly (in the case of a port in the MAC
role they are false). Removing this condition generates a few more SPI
writes for these ports (clearing the RGMII delays) which are perhaps
useless for SJA1105P/Q/R/S, where we know that the delays are disabled
by default. But for SJA1110, the firmware on the embedded microcontroller
might have done something funny, so it's always a good idea to clear the
RGMII delays if that's what Linux expects.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: sja1105: add a translation table for port speeds

In order to support the new speed of 2500Mbps, the SJA1110 has achieved
the great performance of changing the encoding in the MAC Configuration
Table for the port speeds of 10, 100, 1000 compared to SJA1105.

Because this is a common driver, we need a layer of indirection in order
to program the hardware with the right values irrespective of switch
generation.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: sja1105: add a PHY interface type compatibility matrix

On the SJA1105, all ports support the parallel "xMII" protocols (MII,
RMII, RGMII) except for port 4 on SJA1105R/S which supports only SGMII.
This was relatively easy to model, by special-casing the SGMII port.

On the SJA1110, certain ports can be pinmuxed between SGMII and xMII, or
between SGMII and an internal 100base-TX PHY. This creates problems,
because the driver's assumption so far was that if a port supports
SGMII, it uses SGMII.

We allow the device tree to tell us how the port pinmuxing is done, and
check that against a PHY interface type compatibility matrix for
plausibility.

The other big change is that instead of doing SGMII configuration based
on what the port supports, we do it based on what is the configured
phy_mode of the port.

The 2500base-x support added in this patch is not complete.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: sja1105: cache the phy-mode port property

So far we've succeeded in operating without keeping a copy of the
phy-mode in the driver, since we already have the static config and we
can look at the xMII Mode Parameters Table which already holds that
information.

But with the SJA1110, we cannot make the distinction between sgmii and
2500base-x, because to the hardware's static config, it's all SGMII.
So add a phy_mode property per port inside struct sja1105_private.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: sja1105: the 0x1F0000 SGMII "base address" is actually MDIO_MMD_VEND2

Looking at the SGMII PCS from SJA1110, which is accessed indirectly
through a different base address as can be seen in the next patch, it
appears odd that the address accessed through indirection still
references the base address from the SJA1105S register map (first MDIO
register is at 0x1f0000), when it could index the SGMII registers
starting from zero.

Except that the 0x1f0000 is not a base address at all, it seems. It is
0x1f << 16 | 0x0000, and 0x1f is coding for the vendor-specific MMD2.
So, it turns out, the Synopsys PCS implements all its registers inside
the vendor-specific MMDs 1 and 2 (0x1e and 0x1f). This explains why the
PCS has no overlaps (for the other MMDs) with other register regions of
the switch (because no other MMDs are implemented).

Change the code to remove the SGMII "base address" and explicitly encode
the MMD for reads/writes. This will become necessary for SJA1110 support.

Cc: Russell King <linux@armlinux.org.uk>
Cc: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: sja1105: allow SGMII PCS configuration to be per port

The SJA1105 R and S switches have 1 SGMII port (port 4). Because there
is only one such port, there is no "port" parameter in the configuration
code for the SGMII PCS.

However, the SJA1110 can have up to 4 SGMII ports, each with its own
SGMII register map. So we need to generalize the logic.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: sja1105: be compatible with "ethernet-ports" OF node name

Since commit f2f3e09396be ("net: dsa: sja1105: be compatible with
"ethernet-ports" OF node name"), DSA supports the "ethernet-ports" name
for the container node of the ports, but the sja1105 driver doesn't,
because it handles some device tree parsing of its own.

Add the second node name as a fallback.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bnx2x: Remove the repeated declaration

Function 'bnx2x_vfpf_release' is declared twice, so remove the
repeated declaration.

Cc: Ariel Elior <aelior@marvell.com>
Cc: GR-everest-linux-l2@marvell.com
Signed-off-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
Acked-by: Sudarsana Reddy Kalluru <skalluru@marvell.com>
Link: https://lore.kernel.org/r/1622449756-2627-1-git-send-email-zhangshaokun@hisilicon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ipv6: align code with context

The Tab key is used three times, causing the code block to
be out of alignment with the context.

Signed-off-by: Rocco Yue <rocco.yue@mediatek.com>
Link: https://lore.kernel.org/r/20210530113811.8817-1-rocco.yue@mediatek.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-sealevel-clean-up-some-code-style-issues'

Guangbin Huang says:

====================
net: sealevel: clean up some code style issues

This patchset clean up some code style issues.
====================

Link: https://lore.kernel.org/r/1622355874-18933-1-git-send-email-huangguangbin2@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: fix the alignment issue

Alignment should match open parenthesis.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: fix the comments style issue

Networking block comments don't use an empty /* line,
use /* Comment...

Block comments use * on subsequent lines.
Block comments use a trailing */ on a separate line.

This patch fixes the comments style issues.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: remove meaningless comments

Remove the meaningless stylistically wrong comment.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: fix a code style issue about switch and case

According to the chackpatch.pl, switch and case should be
at the same indent.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: remove redundant initialization for statics

Should not initialise statics to false.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: add some required spaces

Add spaces required around that '='.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: open brace '{' following struct go on the same line

Fix the checkpatch error as open brace '{' following struct should
go on the same line.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: fix the code style issue about "foo* bar"

Fix the checkpatch error as "(foo*)" should be "(foo *)".

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: add blank line after declarations

This patch fixes the checkpatch error about missing a blank line
after declarations.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sealevel: remove redundant blank lines

This patch removes some redundant blank lines.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ipv6: use prandom_u32() for ID generation

This is a complement to commit aa6dd211e4b1 ("inet: use bigger hash
table for IP ID generation"), but focusing on some specific aspects
of IPv6.

Contary to IPv4, IPv6 only uses packet IDs with fragments, and with a
minimum MTU of 1280, it's much less easy to force a remote peer to
produce many fragments to explore its ID sequence. In addition packet
IDs are 32-bit in IPv6, which further complicates their analysis. On
the other hand, it is often easier to choose among plenty of possible
source addresses and partially work around the bigger hash table the
commit above permits, which leaves IPv6 partially exposed to some
possibilities of remote analysis at the risk of weakening some
protocols like DNS if some IDs can be predicted with a good enough
probability.

Given the wide range of permitted IDs, the risk of collision is extremely
low so there's no need to rely on the positive increment algorithm that
is shared with the IPv4 code via ip_idents_reserve(). We have a fast
PRNG, so let's simply call prandom_u32() and be done with it.

Performance measurements at 10 Gbps couldn't show any difference with
the previous code, even when using a single core, because due to the
large fragments, we're limited to only ~930 kpps at 10 Gbps and the cost
of the random generation is completely offset by other operations and by
the network transfer time. In addition, this change removes the need to
update a shared entry in the idents table so it may even end up being
slightly faster on large scale systems where this matters.

The risk of at least one collision here is about 1/80 million among
10 IDs, 1/850k among 100 IDs, and still only 1/8.5k among 1000 IDs,
which remains very low compared to IPv4 where all IDs are reused
every 4 to 80ms on a 10 Gbps flow depending on packet sizes.

Reported-by: Amit Klein <aksecurity@gmail.com>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20210529110746.6796-1-w@1wt.eu
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

virtio-net: Add validation for used length

This adds validation for used length (might come
from an untrusted device) to avoid data corruption
or loss.

Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Link: https://lore.kernel.org/r/20210531135852.113-1-xieyongji@bytedance.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: st95hf: fix indentation to tabs

Use tabs to indent instead of spaces. No functional change.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073902.7111-7-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: st-nci: drop ftrace-like debugging messages

Now that the kernel has ftrace, any debugging calls that just do "made
it to this function!" and "leaving this function!" can be removed.
Better to use standard debugging tools.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073902.7111-6-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: st21nfca: drop ftrace-like debugging messages

Now that the kernel has ftrace, any debugging calls that just do "made
it to this function!" and "leaving this function!" can be removed.
Better to use standard debugging tools.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073902.7111-5-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: pn544: drop ftrace-like debugging messages

Now that the kernel has ftrace, any debugging calls that just do "made
it to this function!" and "leaving this function!" can be removed.
Better to use standard debugging tools.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073902.7111-4-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: pn533: drop unneeded braces {} in if

{} braces are not needed over single if-statement.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073902.7111-3-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: pn533: drop ftrace-like debugging messages

Now that the kernel has ftrace, any debugging calls that just do "made
it to this function!" and "leaving this function!" can be removed.
Better to use standard debugging tools.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073902.7111-2-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: mrvl: simplify with module_driver

Remove standard module init/exit boilerplate with module_driver() which
also annotates the functions with __init.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073902.7111-1-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: mrvl: correct minor coding style violations

Correct block comments and usage of tab in function definition.  No
functional change.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073522.6720-4-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: mrvl: use SPDX-License-Identifier

Use SPDX-License-Identifier: GPL-2.0-only, instead of hand writing it.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073522.6720-3-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: mei_phy: drop ftrace-like debugging messages

Now that the kernel has ftrace, any debugging calls that just do "made
it to this function!" and "leaving this function!" can be removed.
Better to use standard debugging tools.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073522.6720-2-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfc: fdp: drop ftrace-like debugging messages

Now that the kernel has ftrace, any debugging calls that just do "made
it to this function!" and "leaving this function!" can be removed.
Better to use standard debugging tools.

This allows also to remove several local variables and entire
fdp_nci_recv_frame() function (whose purpose was only to log).

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210531073522.6720-1-krzysztof.kozlowski@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'fixes-for-yt8511-phy-driver'

Peter Geis says:

====================
fixes for yt8511 phy driver

The Intel clang bot caught a few uninitialized variables in the new
Motorcomm driver. While investigating the issue, it was found that the
driver would have unintended effects when used in an unsupported mode.

Fixed the uninitialized ret variable and abort loading the driver in
unsupported modes.
====================

Link: https://lore.kernel.org/r/20210529110556.202531-1-pgwipeout@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: phy: abort loading yt8511 driver in unsupported modes

While investigating the clang `ge` uninitialized variable report, it was
discovered the default switch would have unintended consequences. Due to
the switch to __phy_modify, the driver would modify the ID values in the
default scenario.

Fix this by promoting the interface mode switch and aborting when the
mode is not a supported RGMII mode.

This prevents the `ge` and `fe` variables from ever being used
uninitialized.

Fixes: 48e8c6f1612b ("net: phy: add driver for Motorcomm yt8511 phy")
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Peter Geis <pgwipeout@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: phy: fix yt8511 clang uninitialized variable warning

clang doesn't preinitialize variables. If phy_select_page failed and
returned an error, phy_restore_page would be called with `ret` being
uninitialized.
Even though phy_restore_page won't use `ret` in this scenario,
initialize `ret` to silence the warning.

Fixes: 48e8c6f1612b ("net: phy: add driver for Motorcomm yt8511 phy")
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Peter Geis <pgwipeout@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-dsa-qca8k-check-return-value-of-read-functions-correctly'

Yang Yingliang says:

====================
net: dsa: qca8k: check return value of read functions correctly

patch #1 - Change return type and add output parameter to make check
return value of read  functions correctly.

patch #2 - Add missing check return value in qca8k_phylink_mac_config().
====================

Link: https://lore.kernel.org/r/20210529030439.1723306-1-yangyingliang@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: qca8k: add missing check return value in qca8k_phylink_mac_config()

Now we can check qca8k_read() return value correctly, so if
it fails, we need return directly.

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: qca8k: check return value of read functions correctly

Current return type of qca8k_mii_read32() and qca8k_read() are
unsigned, it can't be negative, so the return value check is
unuseful. For check the return value correctly, change return
type of the read functions and add a output parameter to store
the read value.

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeontx2-pf: Fix fall-through warning for Clang

In preparation to enable -Wimplicit-fallthrough for Clang, fix a warning
by explicitly adding a break statement instead of letting the code fall
through to the next case.

Link: https://github.com/KSPP/linux/issues/115
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20210528202225.GA39855@embeddedor
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: axienet: Fix fall-through warning for Clang

In preparation to enable -Wimplicit-fallthrough for Clang, fix multiple
warnings by explicitly adding a fallthrough statement instead of just
letting the code fall through to the next case.

Link: https://github.com/KSPP/linux/issues/115
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20210528195831.GA39131@embeddedor
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'npc-kpu-updates'

George Cherian says:

====================
NPC KPU updates

Add support for
 - Loading Custom KPU profile entries
 - Add NPC profile Load from System Firmware DB
 - Add Support fo Coalescing KPU profiles
 - General Updates/Fixes to default KPU profile
====================

Link: https://lore.kernel.org/r/20210527094439.1910013-1-george.cherian@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeontx2-af: Update the default KPU profile and fixes

Add support for parsing following
 1. NGIO
 2. PPPOE
 3. 24 byte custom L2 header
 4. CPT Header
 5. Fragmented CPT packets
 6. VLAN EXDSA

Fix for
 1. EDSA VLAN parsing
 2. Enhance FDSA
 3. CPT Header parsing

Remove ITAG support

Signed-off-by: Sunil Kovvuri Goutham <Sunil.Goutham@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: George Cherian <george.cherian@marvell.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

octeontx2-af: support for coalescing KPU profiles

Adding support to load a new type of KPU image, known as coalesced/
consolidated KPU image via firmware database. This image is a
consolidation of multiple KPU profiles into a single image.

During kernel bootup this coalesced image will be read via
firmware database and only the relevant KPU profile will be loaded.
Existing functionality of loading single KPU/MKEX profile
is intact as the images are differentiated based on the image signature.

Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: Sunil Kovvuri Goutham <Sunil.Goutham@marvell.com>
Signed-off-by: George Cherian <george.cherian@marvell.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>