net/ixgbe/base: support X550em_a 10G PHY

This patch uses the shared Management Data Input/output (MDIO)
functions ixgbe_read_phy_reg_x550a and ixgbe_write_phy_reg_x550a
for X550em_a 10G PHY which supports flow control auto-negotiation.

Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>

net/ixgbe/base: change endianness of PHY data

The x550a related firmware reverses the endianness of the PHY data read
and written via host interface command, so make corresponding changes
to that.

Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>

net/ixgbe/base: remove X550em SFP iXFI setup

This patch removes X550em SFP iXFI setup for the drivers since there
is no released hardware production with SFP iXFI.

Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>

net/ixgbe/base: fix pointer check

The vfta_delta value rather than the pointer should be checked.

Fixes: b978f7b38c14 ("net/ixgbe/base: simplify VLAN management")

Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>

net/i40e: enable 25G device

Add support for 25G link speed to enable 25G device.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e: use PHY type to check PHY capability

Using device ID to check PHY capability is not extensible.

Now we are using PHY type to detect PHY capability.

All link speeds supported by the device are encoded into PHY type,
and PHY type value can be read by admin queue "get_phy_capability"
command at initialization stage.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>

net/enic: support scatter Rx in MTU update

Re-initialize Rq's when MTU is changed. This allows for more
efficient use of mbufs when moving from an MTU that is greater
than the mbuf size to one that is less. Also move to using Rx
scatter mode when moving from an MTU less than the mbuf size
to one that is greater.

Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Signed-off-by: John Daley <johndale@cisco.com>

net/enic: fix freeing memory for descriptor ring

The function vnic_dev_free_desc_ring() didn't actually free memory. Fix
this by first changing vnic_dev_alloc_desc_ring() to use the common
allocation function, then in vnic_dev_free_desc_ring call the common
free function.

Fixes: fefed3d1e62c ("enic: new driver")

Signed-off-by: Nelson Escobar <neescoba@cisco.com>

net/enic: enable link check interrupt

Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>

net/enic: extract code for checking link status

Move link check code to a new function so that it can be reused
by the interrupt handler.

Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>

net/enic: unregister interrupt handler when stopping

enic_disable() wasn't calling rte_intr_disable() or
rte_intr_callback_unregister().  stopping/starting a port would
result in the same interrupt callback being registered multiple
times, which would then cause it to be called multiple times on
every interrupt.

Fixes: fefed3d1e62c ("enic: new driver")

Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>

net/mlx5: support Mellanox OFED 3.4

Some macros are renamed by Mellanox OFED 3.4.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>

maintainers: add new bnxt driver maintainer

Add Ajit Khaparde as the maintainer of the bnxt PMD

Reviewed-by: David Christensen <david.christensen@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Stephen Hurd <stephen.hurd@broadcom.com>

net/i40e: support ARM platform

And add read memory barrier to avoid status inconsistency
between two Rx descriptors readings.

Signed-off-by: Jianbo Liu <jianbo.liu@linaro.org>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/enic: get Rx queue count

Add support to enic for rte_eth_rx_queue_count().

Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>

net/ixgbe: use macro for PCI log format

Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Acked-by: Remy Horton <remy.horton@intel.com>

net/i40e/base: support FCoE

Add Fibre Channel Over Ethernet (FCoE) in PHY capabilities.
Add data structure for FCoE statistics collection.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: get SAN MAC address

Add function that reads the adapter's Storage Area Network (SAN) MAC
address from NVM.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: minor clean up

Remove unused device ID.
Remove unused compile option.
Adjust code alignment.
Standardize macros.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: enable proxy command for X722

Add definitions to enable proxy cmd.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: add multicast magic packet enable flag

Add a definition for multicast magic packet enable flag

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: enable set/get HMC resource profile

Enable admin queue command that can get or set host memory cache (HMC)
resource profile.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: enable wake on LAN for X722

Add macros and data structure to support wake on
LAN feature for X722.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: get PHY abilities for 25G

Add extended PHY types to get PHY abilities for 25G, then put
PHY types for 25G into the proper field when the PHY capabilities
are queried.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: add 25G PHY capability

Add PHY type macros for 25G PHY capabilities.
Change data type and bit setting defines for 25G.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: define PHY type capability constants

Define macros for PHY type capabilities to replace exist enum type.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: fix UDP packet header

Correct Rx ptype payload layer for non_tunneled ipv6 with UDP.

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

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e/base: add function to clear default VSI

Add a function to return the VSI to default settings.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e: fix parsing QinQ packets type

Previously, PTYPE filed in the RX descriptors is not set properly
for QinQ packets. The wrong PTYPE is generated because outer Tag did
not have ORT/PIT configured, so fix this issue by configuring ORT/PIT.
This patch also changes bitmask of outer VLAN tag in L2 header
to support RSS and flow director for QinQ.

Fixes: 4861cde46116 ("i40e: new poll mode driver")
Fixes: 4072d503aaa5 ("i40e: fix VLAN bitmasks for input set")

Signed-off-by: Beilei Xing <beilei.xing@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e: fix null pointer dereferences when using VMDq+RSS

When using VMDQ+RSS, the queue ids used by the application are not
contiguous (see i40e_pf_config_rss). Most of the driver already handled
this, but there were a few cases where it assumed all configured queues
had been setup.

Fixes: 4861cde46116 ("i40e: new poll mode driver")
Fixes: 6b4537128394 ("i40e: free queue memory when closing")
Fixes: 8e109464c022 ("i40e: allow vector Rx and Tx usage")

Signed-off-by: Rich Lane <rich.lane@bigswitch.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>

net/ixgbe: fix VF reset to apply to correct VF

In SR-IOV configuration, queues 0 - nb_rx_queues belong to VF0,
which means that with the current implementation when a reset mbox
message comes from any VF, it affects the settings of VF0.

Fix this by using PF queue index to update the correct queue.

Fixes: dbb0b8737f64 ("ixgbe: add vlan offload support")

Signed-off-by: Alex Zelezniak <alexz@att.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>

net/ixgbe: start Rx/Tx after all config done

Starting rxtx before flow director config will cause driver not to
receive packets from NIC.

Signed-off-by: Wang Wei <lnykww@gmail.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>

net/i40e: fix dropping packets with ethertype 0x88A8

In FW default settings, Ethertype 0x88A8 is treated as S-TAG,
and packets with S-TAG should be received in Port Virtualizer mode.
However, Port Virtualizer mode is not initialized in DPDK, so X710 will
drop packets with Ethertype 0x88A8.
This patch fixes this issue by turning off S-TAG identification.

Fixes: 4861cde46116 ("i40e: new poll mode driver")

Signed-off-by: Beilei Xing <beilei.xing@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>

net/enic: fix bad L4 checksum flag on ICMP packets

The bad L4 checksum flag was set on IP packets which were not
also TCP or UDP packets. This includes ICMP, IGMP and OSPF packets.

L4 ptypes were being treated as bits instead of values within the
L4 mask causing the code to check L4 checksum in the completion
queue and incorrectly set the L4 bad checksum flag.

Fixes: 947d860c821f ("enic: improve Rx performance")

Reviewed-by: Nelson Escobar <neescoba@cisco.com>
Signed-off-by: John Daley <johndale@cisco.com>

net/fm10k: fix MAC address removal from switch

When testpmd quits with two ports, the second port's MAC address
remains in the MAC table of switch manager.

There needs to be some time for HW to quiesce when closing a port,
otherwise a subsequent port close won't be handled correctly.

This patch adds a delay after turning off a logic port, just as
the kernel driver does.

Fixes: 8b5c9ec20b7b ("fm10k: support VMDQ in MAC/VLAN filter")

Reported-by: Xueqin Lin <xueqin.lin@intel.com>
Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
Acked-by: Jing Chen <jing.d.chen@intel.com>

net/enic: move link checking init to probe time

The enic DMAs link status information to the host and this requires a
little setup. This setup was being done as a result of calling
rte_eth_dev_start(). But applications expect to be able to check link
status before calling rte_eth_dev_start().

This patch moves the link status setup to enic_init() which is called
at device probe time so that link status can be checked anytime.

Fixes: fefed3d1e62c ("enic: new driver")

Signed-off-by: Nelson Escobar <neescoba@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>

net/pcap: fix memory leak in jumbo frames

If rte_pktmbuf_alloc() fails on any segment that is not the initial
segment, previously allocated mbufs are not freed.

Fixes: 6db141c91e1f ("pcap: support jumbo frames")

Signed-off-by: Dror Birkman <dror.birkman@lightcyber.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/mlx5: remove gather loop on segments

Tx function was handling a double loop to send segmented packets, it can be
done in a single one.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Signed-off-by: Vasily Philipov <vasilyf@mellanox.com>

net/mlx5: reduce memory overhead for WQE handling

PMD uses only power of two number of Work Queue Elements (aka WQE), storing
the number of elements in log2 helps to reduce the size of the container to
store it.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/mlx5: reduce memory overhead for BF handling

Blue Flame (aka BF) is a buffer allocated with a power of two value, its
size is returned by Verbs in log2.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/mlx5: reduce memory overhead for CQE handling

PMD uses only power of two number of Completion Queue Elements (aka CQE),
storing the number of elements in log2 helps to reduce the size of the
container to store it.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/mlx5: reduce memory overhead of Rx/Tx descriptors

PMD uses only power of two number of descriptors, storing the number of
elements in log2 helps to reduce the size of the container to store it.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>

net/mlx5: rework hardware structures

Rework Work Queue Element (aka WQE) structures to fit PMD needs.
A WQE is an aggregation of 16 bytes elements known as "data segments"
(aka dseg).

The only common part is the first two elements i.e. the control one to
define the job type, and the Ethernet segment which embed offload requests
with other information, after that, it can have:
  - a raw data packet,
  - a data pointer to the packet itself,
  - both.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/nfp: unregister interrupt callback when closing

With an app using hotplug feature, when a device is unplugged without
unregistering makes the interrupt handling unstable.

Fixes: 6c53f87b3497 ("nfp: add link status interrupt")

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

net/nfp: fix copying MAC address

Fixes: defb9a5dd156 ("nfp: introduce driver initialization")

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

net/nfp: use random MAC address if not configured

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

net/thunderx: support 81xx SoC

81xx NIC subsystem differs in new PCI subsystem_device_id and
NICVF_CAP_CQE_RX2 capability.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

net/thunderx: add tunneling extension info capability flag

Certain thunderx SoC pass has additional optional word
in Rx descriptor to hold tunneling extension info.
Based on this capability, the location where packet pointer
address stored in Rx descriptor will vary.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

net/thunderx: remove generic passX references

thunderx pmd driver needs to support multiple SoC
variants in ThunderX family.
Remove generic pass references from driver as each SoC
can have same pass number.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

net/mlx: fix debug build with gcc 6.1

With recent gcc versions, e.g. gcc 6.1, compilation of mlx drivers with
debug enabled produces lots of errors complaining that "pedantic" is
not a warning level that can be ignored.

error: ‘-pedantic’ is not an option that controls warnings [-Werror=pragmas]
 #pragma GCC diagnostic ignored "-pedantic"
                                 ^~~~~~~~~~~

These errors can be removed by changing the "-pedantic" to "-Wpedantic".

Fixes: 7fae69eeff13 ("mlx4: new poll mode driver")
Fixes: 771fa900b73a ("mlx5: introduce new driver for Mellanox ConnectX-4 adapters")

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>

net/pcap: fix checkpatch warnings

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: remove rte prefix from static functions

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: coding convention updates

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: fix missing Tx interface assignment

Missing pcap assignment may cause pcap file/interface to be opened
again, and previous one not closed.

Fixes: 1e38a7c66923 ("pcap: fix storage of name and type in queues")

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: simplify function

simplify function rte_eth_from_pcaps_common by using interim variables.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: remove redundant assignment

data->name assigned twice.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: remove unnecessary check

Both fields are fields of same type of struct, one's size can't be bigger
than others.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: update single interface handling

Remove hardcoded single interface values, make it more obvious.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: reorder functions

Reorder functions to be able to remove function declarations in .c file.
Function definitions not modified.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: reorder includes

Remove unused ones and sort remaining.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: make const array static

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: group stats related fields into a struct

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: use single interface flag instead of hardcoding

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: remove duplicated max queue number check

Remove duplicated check by reorganizing the code, no functional change.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: move comment

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: do not carry kvlist argument

Don't carry kvlist argument into sub function and used it, use kvlist
argument in upper level of call stack.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: do not carry numa node argument

Instead of defining numa_node variable upper level of call stack and
carry into sub function, set it where needs to be used.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: update function to allow reuse

rte_eth_from_pcaps and rte_eth_from_pcaps_n_dumpers functions are very
close, updated rte_eth_from_pcaps function and reused.

No functional update.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: check max queue number

Number of queues is defined by devargs, a check added to be sure this
number is not bigger than configured max number of queue.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: reorganize private structs

struct rx_pcaps and tx_pcaps used to point parsed devargs, but it is not
clear with current names.

Merged both into single struct and modified struct name and field names.

Functionality not changed, only struct names.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: use macros for parameter string

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/pcap: convert config option to a macro

pcap PMD is using ring PMD configuration parameters to set max number of
queues. This creates an unnecessary dependency and confusion.

Stop using configuration parameter to set max number of queues and
convert this variable into a macro within source code, to simplify
configuration file.

Default value of macro is same as ring parameter's default.

pcap pmd doesn't need to be configured in a detail to set rx and tx max
queue numbers separately, so using same macro for both queues.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/e1000: fix returned number of available Rx descriptors

Fixes: 0f6b7c7f7a37 ("igb: use DD bit to count RX available descriptors")

Signed-off-by: Ali Volkan Atli <volkan.atli@argela.com.tr>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/mlx5: fix inline logic

To improve performance the NIC expects for large packets to have a pointer
to a cache aligned address, old inline code could break this assumption
which hurts performance.

Fixes: 2a66cf378954 ("net/mlx5: support inline send")

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Signed-off-by: Vasily Philipov <vasilyf@mellanox.com>

net/mlx5: re-factorize functions

Rework logic of wqe_write() and wqe_write_vlan() which are pretty similar
to keep a single one.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/mlx5: force inline for completion function

This function was supposed to be inlined, but was not because several
functions calls it.  This function should always be inline avoid
external function calls and to optimize code in data-path.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/mlx5: fix flow director drop mode

Packet rejection was routed to a polled queue.  This patch route them to a
dummy queue which is not polled.

Fixes: 76f5c99e6840 ("mlx5: support flow director")

Signed-off-by: Yaacov Hazan <yaacovh@mellanox.com>
Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/mlx5: refactor allocation of flow director queues

This is done to prepare support for drop queues, which are not related to
existing Rx queues and need to be managed separately.

Signed-off-by: Yaacov Hazan <yaacovh@mellanox.com>
Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>

net/mlx5: fix removing VLAN filter

memmove was moving bytes as the number of elements next to i, while it
should move the number of elements multiplied by the size of each element.

Fixes: e9086978 ("mlx5: support VLAN filtering")

Signed-off-by: Raslan Darawsheh <rdarawsheh@asaltech.com>

net/mlx5: fix Rx VLAN offload capability report

This capability is implemented but not reported.

Fixes: f3db9489188a ("mlx5: support Rx VLAN stripping")

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>

net/mlx5: fix inconsistent return value in flow director

The return value in DPDK is negative errno on failure.
Since internal functions in mlx driver return positive
values need to negate this value when it returned to
dpdk layer.

Fixes: 76f5c99 ("mlx5: support flow director")

Signed-off-by: Yaacov Hazan <yaacovh@mellanox.com>

doc: fix typo in VF guide

This patch fixes a typo in doc/guides/nics/intel_vf.rst.

Fixes: fc1f2750a3ec ("doc: programmers guide")

Signed-off-by: Rami Rosen <rami.rosen@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/mlx5: fix possible NULL dereference in Rx path

The user is allowed to call ->rx_pkt_burst() even without free
mbufs in the pool. In this scenario we'll fail allocating a rep mbuf
on the first iteration (where pkt is still NULL). This would cause us
to deref a NULL pkt (reset refcount and free).

Fix this by checking the pkt before freeing it.

Fixes: a1bdb71a32da ("net/mlx5: fix crash in Rx")

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>

net/i40e: add packet type translation for X722

To make the PCTYPE in x722 compatible with original PCTYPE in
flow director (FD) filters, the PCTYPE in the FD programming
descriptor needs to be translated into a different PCTYPE using
GLQF_FD_PCTYPE table.
Translation needs to be done before the FD filter is programmed.

Signed-off-by: Jeff Guo <jia.guo@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>

net/i40e: add new packet types for device X722

There are 6 new PCTYPEs enabled in the device x722.

The 6 new PCTYPEs As below:
* NonF Unicast IPv4, UDP
* NonF Multicast IPv4, UDP
* NonF IPv4, TCP, SYN, no ACK
* NonF Unicast IPv6, UDP
* NonF Multicast IPv6, UDP

Signed-off-by: Jeff Guo <jia.guo@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>

net/vmxnet3: enable LRO

This change enables device LRO if requested.

The current implementation of jumbo frame Rx can be used for LRO
directly without changes.

Note that since jumbo frame uses both ring0 and ring1, it cannot
be enabled in UPT (VMDirectPath) mode.

Signed-off-by: Yong Wang <yongwang@vmware.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/vmxnet3: update NIC documentation

Signed-off-by: Yong Wang <yongwang@vmware.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/vmxnet3: update feature doc

Signed-off-by: Yong Wang <yongwang@vmware.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/vmxnet3: reallocate shared memzone on re-config

When adding a DPDK port to ovs-vswitchd with DPDK, the vmxnet3 device
fails to activate due to mismatched magic number.  This failure causes
following operations to run: start the port, stop the port,
reconfigure and re-start the port.

During reconfigure, if there is an existing memzone, driver will reuse
it. But reconfigure may request different number of Tx/Rx queues.
This results in a memzone with wrong size and potential invalid memory
access.

To fix this, free the memzone if found and reserve a new one.

Signed-off-by: Yong Wang <yongwang@vmware.com>
Reviewed-by: Guolin Yang <gyang@vmware.com>
Reviewed-by: Daniele Di Proietto <ddiproietto@vmware.com>
Tested-by: Daniele Di Proietto <ddiproietto@vmware.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/vmxnet3: coding style changes

Signed-off-by: Yong Wang <yongwang@vmware.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/vmxnet3: improve error checks and return values

Signed-off-by: Yong Wang <yongwang@vmware.com>
Reviewed-by: Juho Snellman <jsnell@iki.fi>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

net/i40e: fix mbuf leak during Rx queue release

For the vector PMD, release all mbufs from the Rx queue if no packets
are received after device start.

Fixes: 9ed94e5bb04e ("i40e: add vector Rx")

Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>

net/ixgbe: fix mbuf leak during Rx queue release

For the vector PMD, release all mbufs from the Rx queue if no packets
are received after device start.

Fixes: 11b220c6498d ("ixgbe: fix release queue mbufs")

Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>

ip_frag: fix missing dependency on hash library

Not sure what exactly changed and where, but I've started getting
build failures on Fedora rawhide i386:
    lib/librte_ip_frag/ip_frag_internal.c:36:23: fatal error:
    rte_jhash.h: No such file or directory
     #include <rte_jhash.h>
                       ^
Looking at librte_ip_frag, it clearly depends on librte_hash so
its probably more a question of something commonly masking the issue.

Signed-off-by: Panu Matilainen <pmatilai@redhat.com>

kni: remove unnecessary ethtool files

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Remy Horton <remy.horton@intel.com>

kni: remove unused ethtool files

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Remy Horton <remy.horton@intel.com>

app/testpmd: reset headroom after txonly packet allocation

This patch fixes txonly raw packets allocations by resetting the
available headroom.

Indeed, some PMDs such as Virtio might prepend some data to the
packet, resulting in mbuf's data_off field to be decremented each
time the mbuf gets re-allocated.

For Virtio PMD, it means that we use only single descriptors for the
first times mbufs get allocated, as at some point there is not
enough headroom to store the header.

Other alternative would be use standard API to allocate the packets,
which does reset the headroom, but the impact on performance is too
big to consider this an option.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

mbuf: add function to reset headroom

Some application use rte_mbuf_raw_alloc() function to improve
performance by not resetting mbuf's fields to their default state.

This can be however problematic for mbuf consumers that need some
headroom, meaning that data_off field gets decremented after
allocation. When the mbuf is re-used afterwards, there might not
be enough room for the consumer to prepend anything, if the data_off
field is not reset to its default value.

This patch adds a new rte_pktmbuf_reset_headroom() function that
applications can call to reset the data_off field.
This patch also replaces current data_off affectations in the mbuf
lib with a call to this function.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

mbuf: fix error handling on pool creation

On error, the mempool object has to be freed, and rte_errno should be a
positive value.

Fixes: 152ca517900b ("mbuf: use default mempool handler from config")

Signed-off-by: Olivier Matz <olivier.matz@6wind.com>

hash: modify lookup bulk pipeline

This patch replaces the pipelined rte_hash lookup mechanism with a
loop-and-jump model, which performs significantly better,
especially for smaller table sizes and smaller table occupancies.

Signed-off-by: Byron Marohn <byron.marohn@intel.com>
Signed-off-by: Saikrishna Edupuganti <saikrishna.edupuganti@intel.com>
Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Sameh Gobriel <sameh.gobriel@intel.com>