net/ixgbe/base: fix getting PHY type for some x550 devices

Return correct physical layer for some x550 devices.

Fixes: 76d5b807ff74 ("ixgbe/base: new X557 phy")
Fixes: d2e72774e58c ("ixgbe/base: support X550")

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: limit iXFI setup to X552 devices

The MAC register NW_MNG_IF_SEL fields have been redefined for X553.
These changes impact the iXFI driver code flow. Since iXFI is only
supported in X552, this patch adds X552 MAC check for iXFI flows.

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: enable LASI only for X552 devices

Enable the Link Alarm Status Interrupt (LASI) only for X552 devices
to receive notifications of the link configurations of the external
PHY and correspondingly support the configuration of the internal
iXFI link, since iXFI does not support auto-negotiation.

This is not required for X553 devices having KR support, which
performs auto-negotiations and which is used as the internal
link to the external PHY.

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: cleanup X540 checksum calculation

The variable checksum_last_word is used only for bounds check.
So remove this variable and use IXGBE_EEPROM_CHECKSUM directly.

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: store link active LED

Add support to get the link active LED index via the LEDCTL register.
If the LEDCTL register does not have link active LED set then
use MAC default LED index.
Link active LED is used for adapter identify/blink support.

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: add driver version to firmware

Send the driver version string to firmware through
the host interface command on x550 devices.

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: clean up X557 link status check

This patch cleans up the code and clarifies the comment around
the X557 PHY link status check in ixgbe_check_link_t_x550em().

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: fix PHY identification for x550a

Method to identify the CS4223/CS4227 is incorrect and unreliable.
Provide a new register to differentiate between these PHY SKUs.

Fixes: fc0559bdb5e3 ("net/ixgbe/base: add link MAC setup for X550a SFP+")

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: use fast MDIO for non-10G devices

Devices that cannot go 10G speeds can safely select a faster
Management Data Input/Output (MDIO) speed.
Select fast MDIO clock speed for for those devices.

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: fix clearing SAN MAC address

Receive Address Register (RAR) entries, including SAN MAC address,
are cleared when VMDq pool bits are cleared.
Prevent SAN MAC address to be cleared.

Fixes: af75078fece3 ("first public release")

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/ixgbe/base: fix PHY reset check for x550em-ext

PHY type ixgbe_phy_x550em_ext_t requires different check
to verify reset status.

Fixes: af75078fece3 ("first public release")

Signed-off-by: Wei Dai <wei.dai@intel.com>

net/nfp: extend speed capabilities advertised

NFP supports more speeds than just 40 and 100GB, which were
what was advertised before.

Signed-off-by: Alejandro Lucero <alejandro.lucero@netronome.com>

net/nfp: report link speed using hardware info

Previous reported speed was hardcoded because there was not firmware
support for getting this information. This change needs also to support
old firmware versions, but instead of the previous hardcoded report, no
speed is reported to the user avoiding to give the wrong speed when link
is not configured to 40G.

Signed-off-by: Alejandro Lucero <alejandro.lucero@netronome.com>

net/nfp: fix typo in Tx offload capabilities

Because macros for TCP and UDP related to offload cksums have
same values, this was not a main problem. But better to use the
right ones.

Fixes: d4a27a3b092a ("nfp: add basic features")

Signed-off-by: Alejandro Lucero <alejandro.lucero@netronome.com>

net/nfp: remove Rx port metadata

This was required for middlebox-like firmware which NFP does
not support anymore.

Signed-off-by: Alejandro Lucero <alejandro.lucero@netronome.com>

net/vmxnet3: fix Rx deadlock

Our use case is that we have an app that needs to keep mbufs around
for a while. We've seen cases when calling vmxnet3_post_rx_bufs() from
vmxet3_recv_pkts(), it might not succeed to add any mbufs to any RX
descriptors (where it returns -err). Since there are no mbufs that the
virtual hardware can use, no packets will be received after this; the
driver won't refill the mbuf after this so it gets stuck in this
state. I call this a deadlock for lack of a better term - the virtual
HW waits for free mbufs, while the app waits for the hardware to
notify it for data (by flipping the generation bit on the used Rx
descriptors). Note that after this, the app can't recover.

This fix is a rework of this patch by Marco Lee:
http://dpdk.org/dev/patchwork/patch/6575/. I had to forward port
it, address review comments and also reverted the allocation
failure handling to the first version of the patch
(http://dpdk.org/ml/archives/dev/2015-July/022079.html), since
that's the only approach that seems to work, and seems to be what
other drivers are doing (I checked ixgbe and em). Reusing the mbuf
that's getting passed to the application doesn't seem to make
sense, and it was causing weird issues in our app. Also, reusing
rxm without checking if it's NULL could cause the code to crash.

Fixes: 14680e3747d5 ("vmxnet3: improve Rx performance")

Signed-off-by: Stefan Puiu <stefan.puiu@gmail.com>
Acked-by: Yong Wang <yongwang@vmware.com>

net/af_packet: guard against buffer overruns in Tx path

Signed-off-by: Michał Mirosław <michal.miroslaw@atendesoftware.pl>
Acked-by: John W. Linville <linville@tuxdriver.com>

net/af_packet: guard against buffer overruns in Rx path

Signed-off-by: Michał Mirosław <michal.miroslaw@atendesoftware.pl>
Acked-by: John W. Linville <linville@tuxdriver.com>

net/pcap: fix timestamps in output pcap file

Fixes: 4c173302c307 ("pcap: add new driver")

Signed-off-by: Michał Mirosław <michal.miroslaw@atendesoftware.pl>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/sfc: support firmware-assisted TSO

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Mark Spender <mspender@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: update RSS redirection table

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: query RSS redirection table

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: set RSS key and hash types config

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: query RSS key and hash types config

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support RSS hash offload

Extract RSS hash provided by the HW in the prefix and put it to mbuf.

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: add basic stubs for RSS support on driver attach

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc/base: do not use enum type when values are bitmask

ICC complains that enumerated type mixed with another type.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support VLAN offload on transmit path

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support deferred start of transmit queues

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support Tx free threshold

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: get transmit queue information

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support deferred start of receive queues

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support scattered Rx DMA

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: get RxQ descriptor done

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: get RxQ pending descriptors count

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support Rx free threshold

Rx free threshold defines minimum number of free Rx descriptors
when Rx ring refill should be done.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: get receive queue information

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: handle received packet type info from HW

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support checksum offloads on receive

IPv4 header and TCP/UDP checksums for both IPv4 and IPv6 are supported.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support multicast addresses list controls

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support main MAC address change

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support promiscuous and all-multicast control

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support link up/down

Signed-off-by: Artem Andreev <artem.andreev@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support link speed and duplex settings

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support MTU change

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support link status change interrupt

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support flow control settings

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support extended statistics

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: add basic statistics

Does not implement any deprecated statistics.
No per-queue statistics yet.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: support parameter to choose performance profile

Supported options are auto (based on NIC firmware variant and
installed licences), throughput, low-latency.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: implement MCDI logging callback

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/tap: add TUN/TAP device PMD

The PMD allows for DPDK and the host to communicate using a raw
device interface on the host and in the DPDK application. The device
created is a Tap device with a L2 packet header.

Signed-off-by: Keith Wiles <keith.wiles@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Tested-by: Aws Ismail <aismail@ciena.com>
Tested-by: Vasily Philipov <vasilyf@mellanox.com>

ethdev: remove some VF functions

remove the following API's:

rte_eth_dev_set_vf_rxmode
rte_eth_dev_set_vf_rx
rte_eth_dev_set_vf_tx
rte_eth_dev_set_vf_vlan_filter
rte_eth_dev_set_vf_rate_limit

Increment LIBABIVER in Makefile
Remove deprecation notice for removing rte_eth_dev_set_vf_* API's.

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>

net/ixgbe: remove static set VF functions

remove the following static functions:

ixgbe_set_pool_rx_mode
ixgbe_set_pool_rx
ixgbe_set_pool_tx
ixgbe_set_pool_vlan_filter
ixgbe_set_vf_rate_limit

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>

examples/ethtool: use ixgbe public function

Replace rte_eth_dev_set_vf_rxmode with rte_pmd_ixgbe_set_vf_rx_mode.

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>

app/testpmd: cleanup parameter checking

Parameter checking is done in the rte_pmd_ixgbe_* functions.
Remove parameter checking from before calls to the rte_pmd_ixgbe_*
functions.

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>

app/testpmd: use ixgbe public functions

Use the the following ixgbe public functions:

rte_pmd_ixgbe_set_vf_rate_limit
rte_pmd_ixgbe_set_vf_rx
rte_pmd_ixgbe_set_vf_rxmode
rte_pmd_ixgbe_set_vf_tx
rte_pmd_ixgbe_set_vf_vlan_filter

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>

net/ixgbe: move set VF functions from the ethdev

Move the following functions from eth_dev_ops to the ixgbe PMD and rename:

ixgbe_set_pool_rx_mode
ixgbe_set_pool_rx
ixgbe_set_pool_tx
ixgbe_set_pool_vlan_filter
ixgbe_set_vf_rate_limit

Rename the functions to the following:

rte_pmd_ixgbe_set_vf_rxmode
rte_pmd_ixgbe_set_vf_rx
rte_pmd_ixgbe_set_vf_tx
rte_pmd_ixgbe_set_vf_vlan_filter
rte_pmd_ixgbe_set_vf_rate_limit

Use public function internally

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>

net/qede: fix resource leak

Current code does not close 'fd' on function exit, leaking resources.

Fixes: 2ea6f76aff40 ("qede: add core driver")

Signed-off-by: Yong Wang <wang.yong19@zte.com.cn>
Acked-by: Harish Patil <harish.patil@qlogic.com>

net/i40e: remove unused macro

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: remove unused macro

remove X722_SUPPORT and I40E_NDIS_SUPPORT MACROs

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: fix byte order

Big Endian platform will accidentally send the wrong
data to the firmware command. This patch fixes the issue.

Fixes: 788fc17b2dec ("i40e/base: support proxy config for X722")
Fixes: 3c89193a36fd ("i40e/base: support WOL config for X722")

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: fix division by zero

For some cases when reading from device are incorrect or image is
incorrect, this part of code causes crash due to division by zero.

Fixes: 8db9e2a1b232 ("i40e: base driver")

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: add broadcast promiscuous control per VLAN

Add a new adminq function that allows driver to configure per-VLAN
broadcast promiscuous mode, similar to how we handle unicast and
multicast promiscuous modes.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: add error state for NVM update state machine

This patch adds I40E_NVMUPD_STATE_ERROR state for NVM update.
Without this patch driver has no possibility to return NVM image write
failure.This state is being set when ARQ rises error.
arq_last_status is also updated every time when ARQ event comes,
not only on error cases.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: remove duplicate definitions

We already define I40E_AQ_PHY_TYPE_EXT_25G* flags in the response adminq
structure above, and do not need to re-define these.

While we are here, replace 0X with 0x as normal style.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: comment that UDP port must be in host order

The firmware expects the Port number to be in Little Endian format, and
the i40e_aq_add_udp_tunnel command clearly expects the udp_port variable
to be in Host order, as it uses CPU_TO_LE16(). It was recently
discovered in the Linux driver that we were passing a Big Endian port
number, which was therefor not enabling the UDP tunnel correctly.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: convert shift values to hex

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: fix NVM access interfering

Acquire NVM lock before reads on all devices.  Previously, locks were
only used for X722 and later.  Fixes an issue where simultaneous X710
NVM accesses were interfering with each other.

Fixes: 8db9e2a1b232 ("i40e: base driver")

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: save link FEC info from link up event

Store the FEC status bits from the link up event into the
hw_link_info structure.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: fix WoL failure on PF reset

By default the device clears all MAC filter information on PF Reset.
However, this will cause Wake-On-LAN to fail because the wake filters
are deleted on transition to D3 power state. To get around this,
firmware is adding functionality to preserve certain MAC filters during
PFR. These bits allow the driver tell the FW which filters to preserve.

Set the datalen field and add I40E_AQ_FLAG_BUF/I40E_AQ_FLAG_RD flags in the
desc struct for the WoL/Proxy AQ descriptors. The WoL/Proxy AQ commands
were failing because these were missing.

Fixes: 3c89193a36fd ("i40e/base: support WOL config for X722")

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: implement VSI full promiscuous mode

This patch implements a function to set a VSI to broadcast, multicast, and
unicast promiscuous mode all at once. This is specifically needed to set
the WoL/Proxy VSI created by FW to full promiscuous mode during power down
for WoL patterns and protocol offloads to function properly.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: implement clear all WoL filters

This patch implements the clear Wake on LAN (WoL) filters admin queue
function which clears out ALL WoL patterns programmed into
the flex filters.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: adjust 25G PHY type values

Define the values for the 25G PHY type bit-fields that match
reported values from firmware. There was a gap in the bit
fields but no corresponding gap i40e_aq_phy_type enum.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: use BIT macro instead of bit fields

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: add FEC bits to PHY capabilities

Add FEC bits to the PHY capabilities AQ command struct. This is required
for 25GbE support. Change the name of the generic mod_type_ext field to
indicate that it is now used for handling FEC.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: remove FPK HyperV VF device ID

Microsoft recently removed the requirement for VFs to use the VMBus.
The Fort Park Windows VF has been changed to use only the hardware
mailbox, so the Hyper-V VF device ID can be removed.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: deprecate unused macro

I40E_MAC_X710 was supposed to be for 10G and I40E_MAC_XL710
was supposed to be for 40G. But i40e_set_mac_type() sets
I40E_MAC_XL710 for all device IDS. I40E_MAC_X710 is not
used at all. Thus deprecating this extra macro.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: replace memcpy

To align with current memcpy use, replace existing legacy memcpy() calls
with i40e_memcpy() calls.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: fix unknown PHYs incorrect identification

The PHY type value for unrecognized PHYs and cables was changed
based on firmware version number. Newer hardware use lower firmware
version numbers and this was causing some PHYs to be identified
as type 0x16 instead of 0xe (unknown).

Without this patch, newer card will incorrectly identify unknown
PHYs and cables.

This change adds hardware type to the check for firmware version
so the PHY type is reported correctly.

Fixes: 8db9e2a1b232 ("i40e: base driver")

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: add protocols when discover capabilities

Add logical_id to I40E_AQ_CAP_ID_MNG_MODE capability starting from major
version 2.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: add bus number info

Currently i40e_bus_info has PCI device and function info only. However
in log messages slot number (i.e hw->bus.device) is being printed
as bus number. Another field should be added to provide bus number
info and preserve existing information.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: add clause22 and clause45 implementation

Some external PHYs require Clause22 and Clause45 method for
accessing registers. Mostly used for X722 support.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: add media type detection for 25G link

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: fix long link down notification time

This patch fixes a problem where it could take a very
long time (>100 msec) to print the link down notification.
This problem is fixed by changing how often we update link
info from fw, when link is down. Without this patch, it can
take over 100msec to notify user link is down.

Fixes: e6691b428eb1 ("i40e/base: fix PHY NVM interaction")

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: group base mode VF offload flags

Group together the minimum set of offload capabilities that are always
supported by VF in base mode. This define would be used by PF to make
sure VF in base mode gets minimum of base capabilities.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: fix bit test mask

Incorrect bit mask was used for testing "get link status" response.
Instead of I40E_AQ_LSE_ENABLE (which is actually 0x03) it should
be I40E_AQ_LSE_IS_ENABLED (which is defined as 0x01).

Fixes: 8db9e2a1b232 ("i40e: base driver")

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: remove unnecessary code

This patch changes some assignments and removing the unnecessary
code to avoid error reported by static analysis tools.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: fix flow control set for 25G

Add phy_type_ext copied from old setting to rpevents 25G PHY
types from being disabled when setting the flow control modes.

Fixes: 51131ae119 ("net/i40e/base: get PHY abilities for 25G")

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e/base: add encap checksum VF offload flag

Add ENCAP_CSUM offload negotiation flag. Currently VF assumes checksum
offload for encapsulated packets is supported by default. Going forward,
this feature needs to be negotiated with PF before advertising to the
stack. Hence, we need a flag to control it.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>

net/nfp: document supported features

Signed-off-by: Alejandro Lucero <alejandro.lucero@netronome.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

net/ixgbe: fix typo in comment

Fixes: c03fcee9abbd ("ixgbe: remove CRC size from byte counters")

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>

net/sfc: restart TxQ in case of exception on event queue

Examples of recoverable exceptions because of Tx error are:
 - Tx descriptor PCI read error
 - invalid Tx option descriptor
 - Tx option descriptor not supported by the firmware variant
 - unexpected Tx option descriptor (e.g. missing FATSO2A before
   FATSO2B, missing FATSO2B after FATSO2A)
 - incomplete packet push (CONT bit set in the latest pushed DMA
   descriptor)

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: restart RxQ in case of exception on event queue

Examples of recoverable exceptions are:
 - unexpected Rx event (Rx scatter abort with non-zero size,
   too big Rx descriptors batch completed)
 - Rx error due to invalid Rx descriptors push
 - Rx error due to Rx descriptor read error (e.g. unmapped Rx ring
   and denied by IOMMU)

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andrew Lee <alee@solarflare.com>
Reviewed-by: Robert Stonehouse <rstonehouse@solarflare.com>

net/sfc: send bursts of packets

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andy Moreton <amoreton@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/sfc: implement transmit path start / stop

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andy Moreton <amoreton@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/sfc: set up and release Tx queues

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andy Moreton <amoreton@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/sfc: add function to check configured Tx mode

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andy Moreton <amoreton@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/sfc: provide basic stubs for Tx subsystem

Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andy Moreton <amoreton@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/sfc: discard scattered packet on Rx correctly

Since Rx scatter is not supported, all scattered packets are discarded.
It is not always possible to disable scatter on Huntington, so we
should handle scattered packets correctly in any case.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Andy Moreton <amoreton@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>