net/hns3: report Rx free threshold

This patch reports .rx_free_thresh value in the .dev_infos_get ops
implementation function named hns3_dev_infos_get and
hns3vf_dev_infos_get.
In addition, the name of the member variable of struct hns3_rx_queue is
modified and comments are added to improve code readability.

Signed-off-by: Chengwen Feng <fengchengwen@huawei.com>
Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>

examples: use new link status print format

Add usage of rte_eth_link_to_str function to example
applications.

Signed-off-by: Ivan Dyukov <i.dyukov@samsung.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

app: use new link status print format

Add usage of rte_eth_link_to_str function to applications and docs.

Signed-off-by: Ivan Dyukov <i.dyukov@samsung.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

ethdev: format link status text

There is new link_speed value introduced. It's INT_MAX value which
means that speed is unknown. To simplify processing of the value
in application, new function is added which convert link_speed to
string. Also dpdk examples have many duplicated code which format
entire link status structure to text.

This commit adds two functions:
  * rte_eth_link_speed_to_str - format link_speed to string
  * rte_eth_link_to_str - convert link status structure to string

Signed-off-by: Ivan Dyukov <i.dyukov@samsung.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/af_xdp: fix umem size

The kernel expects the start address of the UMEM to be page size
aligned.
Since the mempool is not guaranteed to have such alignment, we have been
aligning the address to the start of the page the mempool is on. However
when passing the 'size' of the UMEM during it's creation we did not take
this into account.

This commit adds the amount by which the address was aligned to the size
of the UMEM.

Bugzilla ID: 532
Fixes: d8a210774e1d ("net/af_xdp: support unaligned umem chunks")
Cc: stable@dpdk.org
Signed-off-by: Ciara Loftus <ciara.loftus@intel.com>

net/virtio: fix variable assignment in helper macro

Inside Macro ASSIGN_UNLESS_EQUAL(var, val), assignment to var is always
failing as assignment done using var_ having local scope only.
This leads to TX packets not going out and found broken due to cleanup
malfunctioning. This patch fixes the wrong variable assignment.

Fixes: 57f90f894588 ("net/virtio: reuse packed ring functions")
Cc: stable@dpdk.org
Signed-off-by: Vipul Ashri <vipul.ashri@oracle.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

vdpa/mlx5: fix completion queue polling

The CQ polling is done in order to notify the guest about new traffic
bursts and to release FW resources for the next bursts management.

When HW is faster than SW, it may be that all the FW resources are busy
in SW due to late polling.
In this case, due to wrong WQE counter masking, the fullness
calculation of the completions number is 0 while the queue is full.

Change the WQE counter masking to 16-bit wideness instead of the CQ
size mask as defined by the CQE format.

Fixes: c5f714e50b0e ("vdpa/mlx5: optimize completion queue poll")
Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@nvidia.com>
Acked-by: Xueming Li <xuemingl@nvidia.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

vdpa/mlx5: fix completion queue assertion

The CQ configuration enables the collapse feature in HW what cause HW to
write all the completions in the first CQE.
When this feature is enabled the HW doesn't switch the owner bit when it
starts a new cycle of the CQ, not like working without the collapse
feature.

The current SW CQ polling wrongly added an assertion to validate the
owner bit switch what causes a panic in debug mode.

Remove the aforementioned assertion.

Fixes: c5f714e50b0e ("vdpa/mlx5: optimize completion queue poll")
Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@nvidia.com>
Acked-by: Xueming Li <xuemingl@nvidia.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

vhost: fix IOTLB mempool single-consumer flag

Control thread (which handles iotlb msg) and forwarding thread
both use iotlb to translate address. The former may modify the
same entry of mempool and may cause a loop in iotlb_pending_entries
list.

Bugzilla ID: 523
Fixes: d012d1f293f4 ("vhost: add IOTLB helper functions")
Cc: stable@dpdk.org
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

vdpa/mlx5: fix event channel setup

During vDPA device setup, if some error happens, event channel
release stucks at polling event channel.

Event channel fd is set to non-blocking in cqe setup, so if any
error happens before this function and after event channel created,
the pooling before releasing resources will stuck.

This patch moves event channel to non-blocking mode right after
creation.

Fixes: 8395927cdfaf ("vdpa/mlx5: prepare HW queues")
Cc: stable@dpdk.org
Signed-off-by: Xueming Li <xuemingl@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

vhost: add device reset status

vhost lib now does not have definition of reset status. This patch
adds the reset status definition and changes related log.

Signed-off-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

net/virtio-user: enable feature checking

virtio 1.0 introduced a mechanism for the driver to verify that the
feature bits it sets are accepted by the device. This mechanism consists
in setting the VIRTIO_STATUS_FEATURE_OK status bit and re-reading it,
which gives a chance for the device to clear it if the features
were not accepted.

This is currently being done only in modern virtio-pci devices but since
the appropriate vhost-user messages have been added, it can also be done
in virtio-user (vhost-user only).

This patch activates this mechanism on virtio-user.

Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Signed-off-by: Adrian Moreno <amorenoz@redhat.com>

net/virtio-user: support vhost status getting

This patch adds support for VHOST_USER_GET_STATUS request.

Only vhost-user backed is supported for now

Signed-off-by: Adrian Moreno <amorenoz@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>

net/virtio-user: support vhost status setting

This patch adds support for VHOST_USER_SET_STATUS
request. It is used to make the backend aware of
Virtio devices status update.

It is useful for the backend to know when the Virtio
driver is done with the Virtio device configuration.

Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Signed-off-by: Adrian Moreno <amorenoz@redhat.com>

net/virtio: add device reset status bit

For the sake of completeness, add the definition of the missing status
bit in accordance with the virtio spec

Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Signed-off-by: Adrian Moreno <amorenoz@redhat.com>

net/cxgbe: support RSS redirection table update

Implement eth_dev_ops to manipulate RSS redirection table.

Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>

net/cxgbe: improve Rx congestion control

Chelsio T6 NIC can support up to 8 priority channels to manage
congestion. So, increase to 8 congestion channels for T6. Also,
add Rxq state to avoid unnecessarily ringing doorbell and polling
the hardware for more traffic when the Rxq is stopped.

Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>

net/cxgbe: rework queue allocation between ports

Firmware returns the max queues that can be allocated on the entire
PF. The driver evenly distributes them across all the ports belonging
to the PF. However, some ports may need more queues than others and
this equal distribution scheme prevents accessing these other ports
unused queues. So, remove the equal distribution scheme and allow the
ports to allocate as many queues as they need.

Also remove the hardcoded 64 max limit on queue allocation. Instead,
use the max limit given by firmware.

Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>

net/cxgbe: release port resources during port close

Enable RTE_ETH_DEV_CLOSE_REMOVE during PCI probe for all ports
enumerated under the PF. Free up the underlying port Virtual
Identifier (VI) and associated resources during port close.
Once all the ports under the PF are closed, free up the PF-wide
shared resources. Invoke port close function of all ports under
the PF, in PCI remove too.

Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>

net/cxgbe: fix queue DMA ring leaks during port close

Free up the DMA memzones properly for all the port's queues during
port close. So, rework DMA ring allocation/free logic to use
rte_eth_dma_zone_reserve()/rte_eth_dma_zone_free() helper functions
for allocating/freeing the memzones.

The firmware event queue doesn't have an associated freelist queue.
So, remove check that tries to give memzone name for a non-existent
freelist queue.

Also, add a missing free for the control queue mempools.

Fixes: 0462d115441d ("cxgbe: add device related operations")
Cc: stable@dpdk.org
Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>

net/sfc/base: fix tunnel configuration

Tunnel configuration may fail because of insufficient access rights
on a virtual function. Ignore the failure if a tunnel configuration
with empty UDP ports is requested.

Fixes: 17551f6dffcc ("net/sfc/base: add API to control UDP tunnel ports")
Cc: stable@dpdk.org
Signed-off-by: Igor Romanov <igor.romanov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>

net/octeontx2: support RSS hash level

Add support to choose rss hash level from ethdev rss config.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>

app/testpmd: support RSS level configuration

Adding support to set RSS level from ethdev config.
level-default will requests the default behavior.
level-outer will requests RSS to be performed on the outermost packet
encapsulation level.
level-inner will request RSS to be performed on the specified inner
packet encapsulation level, from outermost to innermost.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

ethdev: support encapsulation level for RSS offload

This patch reserves 2 bits as input selection to select inner and outer
encapsulation level for RSS computation. It is combined with existing
ETH_RSS_* to choose inner or outer layers.
This functionality already exists in rte_flow through level parameter in
RSS action configuration rte_flow_action_rss.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net: adjust header length parse size

Enlarge the L3 and tunnel header length from 8-bit to 16-bit to handle
the bigger headers. And reorder the fields to avoid creating a structure
hole.

Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>

doc: improve multiport PF in nfp guide

The Agilio CX family of smartNIC's generally have a 1:many mapping of PF
to physical ports. Elaborate on this mapping in the PF multiport section
of the NFP PMD documentation.

Fixes: d625beafc8be ("doc: update NFP with PF support information")
Cc: stable@dpdk.org
Signed-off-by: Heinrich Kuhn <heinrich.kuhn@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>

net/cxgbe: fix crash when accessing empty Tx mbuf list

Ensure packets are available before accessing the mbuf list in Tx
burst function. Otherwise, just reclaim completed Tx descriptors and
exit.

Fixes: b1df19e43e1d ("net/cxgbe: fix prefetch for non-coalesced Tx packets")
Cc: stable@dpdk.org
Reported-by: Brian Poole <brian90013@gmail.com>
Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>

gso: fix payload unit size for UDP

Fragment offset of IPv4 header is measured in units of
8 bytes. Fragment offset of UDP fragments will be wrong
after GSO if pyld_unit_size isn't multiple of 8. Say
pyld_unit_size is 1500, fragment offset of the second
UDP fragment will be 187 (i.e. 1500 / 8), which means 1496,
and it will result in 4-byte data loss (1500 - 1496 = 4).
So UDP GRO will reassemble out a wrong packet.

Fixes: b166d4f30b66 ("gso: support UDP/IPv4 fragmentation")
Cc: stable@dpdk.org
Signed-off-by: Yi Yang <yangyi01@inspur.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>

net/ice: support new devices

Added support for below new devices:

ICE_DEV_ID_E822L_BACKPLANE      0x1897
ICE_DEV_ID_E822L_SFP            0x1898
ICE_DEV_ID_E822L_10G_BASE_T     0x1899
ICE_DEV_ID_E822L_SGMII          0x189A

The patch also reordered items in pci_id_ice_map to align with
ice_devids.h

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Qiming Yang <qiming.yang@intel.com>

net/iavf: reject floating RSS attribute

For RSS attribute don't have an associated RSS type, we need
to reject it.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Junfeng Guo <junfeng.guo@intel.com>

net/bnxt: add separate mutex for FW health check

def_cp_lock was added to sync race between dev_configure and
int_handler. It should not be used to synchronize scheduling of FW
health check between dev_start and async event handler as well,
use a separate mutex for the same.

Fixes: a73b8e939f10 ("net/bnxt: fix race between start and interrupt handler")
Cc: stable@dpdk.org
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: fix checking VNIC in shutdown path

Add a couple of NULL pointer checks in bnxt_free_all_filters()
and bnxt_free_vnics() respectively to guard against certain error
injection/recovery scenarios where it was found that the application
was crashing with the bp->vnic_info pointer being NULL.

Fixes: 51fafb89a9a0 ("net/bnxt: get rid of ff pools and use VNIC info array")
Cc: stable@dpdk.org
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: add locks in flow database

Added support for mutex protection for the flow database to prevent
simultaneous access to flow database and protect flow creation and
deletion.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Shahaji Bhosle <sbhosle@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: fix representor data path

1.Representor Rx ring producer index was not getting reset in
the ring full case. Fix it by incrementing only in
success case.
2.Instead of calling the mbuf specific routine to free the mbuf when
representor ring is full rte_free was being called leading to
'invalid memory' errors being logged.
3. Do not account the pkt meant for the representor in the parent
Rx ring's array that is returned to the application.

Fixes: 6dc83230b43b ("net/bnxt: support port representor data path")
Cc: stable@dpdk.org
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com>
Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: provide switch info if VFR are configured

Some applications need switch_info of the device to be returned
as a part of eth_dev_info_get(). The offload logic in such
applications could use this info. Pass this info to the when VF
representors are configured.

Fixes: 322bd6e70272 ("net/bnxt: add port representor infrastructure")
Cc: stable@dpdk.org
Signed-off-by: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: fix out of bound access in bit handling

Fix out of bounds access in action bit handling.
The act_val is changed to be array to resolve out of bound access issue.

Fixes: 52799debdf1c ("net/bnxt: support action bitmap opcode")
Cc: stable@dpdk.org
Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Shahaji Bhosle <sbhosle@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: enable NAT action with tagged traffic

Added support for performing L3 or L4 rewrite for VLAN tagged
flows. The outer most DMAC, SMAC and VLAN are used to overwrite
when NAT operations are performed.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: enable VXLAN IPv6 encapsulation

Add code to support vxlan ipv6 tunnel encapsulation. The
ipv6 flow traffic class and flow label wild card match
can be ignored to support offload on some applications.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: check and set initial counter ID

Instead of relying on value of Flow counter ID to determine validity
have an explicit boolean flag for the same to check and set.

Fixes: 306c2d28e247 ("net/bnxt: support count action in flow query")
Fixes: 9cf9c8385df7 ("net/bnxt: add ULP flow counter manager")
Cc: stable@dpdk.org
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: increase counter support from 8K to 16K

The number of internal stats counter is increased to 16k
in both egress and ingress direction.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: remove VLAN pop action for egress flows

Whitney platform does not support VLAN pop action in the egress
direction. Hence the VLAN pop action is removed from the egress
action templates.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: use direct HWRM message for interface table

Change interface tables to use direct or non-tunneled HWRM messaging
instead of tunneled messaging. Update HWRM API to a new version to
allow this change.

Signed-off-by: Randy Schacher <stuart.schacher@broadcom.com>
Reviewed-by: Shahaji Bhosle <sbhosle@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: update resource settings

Update default resource configuration.
Resources include ENCAP records, TCAM, wild card, source property
functions and such.

Signed-off-by: Shahaji Bhosle <sbhosle@broadcom.com>
Reviewed-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: fix VFR cleanup during init failure

If VF-rep port add fails for some reason, code was rolling back
all ports added so far. With some applications, there is no need
to do that. Just log failure message for the VF rep port add and
continue.
Also include RTE_MAX_ETH_PORTS value in the bounds check as one port
will be taken by the uplink port anyway

Fixes: 6dc83230b43b ("net/bnxt: support port representor data path")
Cc: stable@dpdk.org
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: fix crash in VFR queue select

Instead of bounds checking against max possible rings while selecting
queue index for the VF representor, do it against the number of rings
configured.

Fixes: 6dc83230b43b ("net/bnxt: support port representor data path")
Cc: stable@dpdk.org
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>

net/bnxt: refactor VFR port clean up

When parent VF or PF ports are cleaned up, the child VF representor
ports also need to be cleaned up. If not cleaned up, then deleting
the parent VF shall result in not cleaning up the hardware rules and
updating the firmware of VFR removal. The issue can occur even when
application is exited without deleting VFR ports.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Shahaji Bhosle <sbhosle@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: fix function id used in flow flush

The function id being used in the flush is incorrect, fixed the
flush of the flows.

Fixes: 74bcfc062489 ("net/bnxt: add session and function flow flush")
Cc: stable@dpdk.org
Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Shahaji Bhosle <sbhosle@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: modify default flow rule creation

Change default flow rule to use 8-byte encap.
The VFR conduit uses VLAN encap to send packets. So the encap record
is changed from 16B to 8B. That frees up 8B of encap records.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Shahaji Bhosle <sbhosle@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: add null check for resource manager

Verify the resource manager has been allocated prior to using it.
This can avoid potential segmentation faults.

Signed-off-by: Shahaji Bhosle <sbhosle@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: free EM index on failure

When a Exact Match entry fails insertion, the allocated index needs to
be pushed back to the allocation stack. This patch takes care of that.

Signed-off-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: fix coexistence of IPv4 and IPv6 ingress rules

The ingress rule to match on ipv4 and ipv6 is now two rules to
make sure both rules can coexist at the same time. Added count
action only for ingress flows.

Fixes: fe82f3e02701 ("net/bnxt: support exact match templates")
Cc: stable@dpdk.org
Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: reduce debug log messages

Removed the mark id log message since it is in the data path.
Also optimized the link status debug message.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: reject flow offload with invalid MAC

Reject offload flows that have broadcast or multicast
Ethernet addresses.

Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

net/bnxt: fix flow drop action to support count

Changed the action template to support count action in addition
to a flow that does drop action.

Fixes: fe82f3e02701 ("net/bnxt: support exact match templates")
Cc: stable@dpdk.org
Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>

net/bnxt: fix port stop process and cleanup resources

The port deinitialization now cleans up all the resources
properly. If all the ports are stopped then ULP context is
freed.
Added fix to update the correct tfp pointer in the ULP context
with the changes to support multi control channels.

Fixes: 70e64b27af5b ("net/bnxt: support ULP session manager cleanup")
Cc: stable@dpdk.org
Signed-off-by: Kishore Padmanabha <kishore.padmanabha@broadcom.com>
Reviewed-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Shahaji Bhosle <sbhosle@broadcom.com>

bus/dpaa: fix fd check before close

The fd is possibly a negative value while it is passed as an
argument to function "close". Fix the check to the fd.

Fixes: b9c94167904f ("bus/dpaa: decouple FQ portal alloc and init")
Cc: stable@dpdk.org
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>

net/ice: fix ptype parsing

The ptype mask for flexible descriptor in Rx function ice_recv_pkts_vec
has a reversed order, which leads to an incorrect value of the final
ptype. This patch fix the mask to parse the correct ptype of RX packets.

Fixes: c68a52b8b38c ("net/ice: support vector SSE in Rx")
Cc: stable@dpdk.org
Signed-off-by: Ting Xu <ting.xu@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/i40e: fix recreating flexible flow director rule

This patch fixes the failure of recreate flexible fdir rule.
The root cause is that the flex_mask_flag is not reset during
flow destroy and flow flush.

Fixes: 6ced3dd72f5f ("net/i40e: support flexible payload parsing for FDIR")
Cc: stable@dpdk.org
Signed-off-by: Guinan Sun <guinanx.sun@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/ice: remove devargs for flow mark

Currently, all data paths already support flow mark, so remove devargs
"flow-mark-support". FDIR matched ID will display in verbose
when packets match the created rule.

Signed-off-by: Guinan Sun <guinanx.sun@intel.com>
Acked-by: Leyi Rong <leyi.rong@intel.com>

net/ice: support flow mark in SSE path

Support flow director mark ID parsing from flexible
Rx descriptor in SSE path.

Signed-off-by: Guinan Sun <guinanx.sun@intel.com>
Acked-by: Leyi Rong <leyi.rong@intel.com>

net/ice: support flow mark in AVX path

Support flow director mark ID parsing from flexible
Rx descriptor in AVX path.

Signed-off-by: Guinan Sun <guinanx.sun@intel.com>
Acked-by: Leyi Rong <leyi.rong@intel.com>

net/ice: add flow director enabled switch

The patch adds fdir_enabled flag to identify if parse flow director mark
ID from flexible Rx descriptor.

Signed-off-by: Guinan Sun <guinanx.sun@intel.com>
Acked-by: Leyi Rong <leyi.rong@intel.com>

net/ice: support flex Rx descriptor RxDID22

This patch supports RxDID #22 by the following changes:
- add structure and macro definition for RxDID #22.
- support RxDID #22 format in normal path.
- change RSS hash parsing from RxDID #22 in AVX/SSE data path.

Signed-off-by: Junyu Jiang <junyux.jiang@intel.com>
Acked-by: Leyi Rong <leyi.rong@intel.com>

net/mlx5: fix hairpin dependency on destination DevX TIR

The PMD supports hairpin only if DevX is supported and DV flow is
enabled.

When destination DevX TIR is not supported, the PMD tries to create TIR
action, and fails.

Avoid supporting hairpin when destination DevX TIR is not supported.

Fixes: b6b3bf86bd1a ("net/mlx5: get hairpin capabilities")
Cc: stable@dpdk.org
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>

net/mlx5: fix Rx objects creator selection

There are 2 creators for Rx objects, DevX and Verbs.
There are supported DR versions when a DevX destination TIR flow action
creation cannot be supported, using this versions the TIR object should
be created by Verbs, what forces all the Rx objects to be created by
Verbs.

The selection of the Rx objects creator, wrongly, didn't take into
account the destination TIR action support what caused a failure in the
Rx flows creation.

Select Verbs creator when destination TIR action creation is not
supported by the DR version.

Fixes: 6deb19e1b2d2 ("net/mlx5: separate Rx queue object creations")

Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>

net/hns3: fix queue offload capability

Currently, offload capabilities are only enabled for all Rx/Tx queues in
hns3 PF/VF PMD driver, and offload capability only applied in a Rx/Tx
queue is not supported.
So this patch moves 'DEV_TX_OFFLOAD_MBUF_FAST_FREE' from
tx_queue_offload_capa to tx_offload_capa.

Fixes: 1f5ca0b460cd ("net/hns3: support some device operations")
Fixes: a5475d61fa34 ("net/hns3: support VF")
Cc: stable@dpdk.org
Signed-off-by: Huisong Li <lihuisong@huawei.com>
Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>

app/testpmd: fix name of bitrate library in meson build

The bitrate library in DPDK is actually in a "bitratestats" directory,
so that is used by meson for the macro and library name.
Therefore, we need to update references to RTE_LIBRTE_BITRATE to
RTE_LIBRTE_BITRATESTATS in testpmd to have it found. Rather than
supporting both defines, since make is being removed, we can just
replace all instances of the former define with the latter.

To ensure testpmd links ok when this is done, we also need to add
bitratestats to the list of library dependencies.

Fixes: 5b9656b157d3 ("lib: build with meson")
Cc: stable@dpdk.org
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Tested-by: Wei Ling <weix.ling@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/nfp: expand device info get

Report Rx and Tx descriptor related limitations in the nfp dev_info_get
callback function. This commit also adds NFP_ALIGN_RING_DESC to replace
a static integer value used during rx/tx queue setups to validate
descriptor alignment.

Cc: stable@dpdk.org
Signed-off-by: Heinrich Kuhn <heinrich.kuhn@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>

common/mlx5: fix aligned malloc

Before this commit system call memalign was used for aligned
allocations, however memalign is deprecated.

Based on (1) - POSIX requires that memory aligned allocations can be
freed using free. Some systems provide no way to reclaim memory
allocated with memalign (because one can only pass to free a pointer
gotten from malloc, while, memalign would call malloc and then align the
obtained value).
Another issue is that 64/32 bits architectures use a minimal alignment
size. So any requested alignment below the minimal system size can be
simplified by calling malloc.

The glibc implementation allows memory obtained from posix_memalign to
be reclaimed with free.  This commit replaces system call memalign with
system call posix_memalign. It also calls malloc in case the requested
alignment is below the minimal system size.

(1) https://linux.die.net/man/3/memalign

Fixes: d38e3d526657 ("common/mlx5: add memory management functions")
Cc: stable@dpdk.org
Signed-off-by: Ophir Munk <ophirmu@nvidia.com>
Acked-by: Matan Azrad <matan@mellanox.com>

net/mlx5: fix RSS RETA reset on start

The following sequences was working fine on mlx5:
   rte_eth_dev_configure(portid, ...);

   for (queueid = 0; queueid < nb_txq; queueid++)
      rte_eth_tx_queue_setup(portid, queueid, ...);

   for (queueid = 0; queueid < nb_rxq; queueid++)
      rte_eth_rx_queue_setup(portid, queueid, ...);

  // use a custom reta configuration
  rte_eth_dev_rss_reta_update(portid, reta_conf, reta_size);
  rte_eth_dev_start(portid);

We were able to configure a custom reta before starting the port.

The commit "net/mlx5: support RSS on hairpin" breaks this logic by
moving the code initializing the RSS reta from rte_eth_dev_configure
into rte_eth_dev_start.

To fix the issue, the skip_default_rss_reta is always set to 1 in
rte_eth_dev_rss_reta to avoid reconfigure the rss reta when the device
is started.

Fixes: 63bd16292c3a ("net/mlx5: support RSS on hairpin")
Cc: stable@dpdk.org
Signed-off-by: Maxime Leroy <maxime.leroy@6wind.com>
Acked-by: Ori Kam <orika@nvidia.com>

net/iavf: support RSS for IPv6 64-bit prefix

RSS for IPv6 prefix 64bit fields are supported in this patch, so that
we can use prefix instead of full IPv6 address for RSS. The prefix
here only includes the first 64 bits of both SRC and DST IPv6 address.

Signed-off-by: Junfeng Guo <junfeng.guo@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/iavf: replace function name with macro

Replace some function name with macro to shrink coding characters.
VIRTCHNL_DEL_PROTO_HDR_FIELD, VIRTCHNL_ADD_PROTO_HDR_FIELD
--> REFINE_PROTO_FLD.

Signed-off-by: Junfeng Guo <junfeng.guo@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/ice: merge inner/outer seg info for flow director

For tunnel and non-tunnel packets, it can share the same seg_tun info.
seg_tun[1] can be used for supporting inner fields with tunnel flow rule
or for non-tunnel packets, seg_tun[0] only used for tunnel outer part.
Add outer_input_set to distinguish inner/outer input set. So we can
identify different fields in outer or inner part.

Signed-off-by: Zhirun Yan <zhirun.yan@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/ice: refactor flow director set config

The original set conf function in FDIR was very long. Refactor to
increase readability to make it clearer and allow for more convenient
further changes.

No functional change here.

Signed-off-by: Zhirun Yan <zhirun.yan@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/iavf: support outer IP hash for no inner GTPU

Outer IP hash can be configured as input sets for no inner GTPU packets.

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

app/flow-perf: allow fixed values for actions

Sometime the user want to have fixed values of
encap/decap or header modify for all flows.

This will introduce the ability to choose from
fixed or dynamic values by setting the flag in
config.h

To use different value for each flow:
config.h: #define FIXED_VALUES 0

To use single value for all flows:
config.h: #define FIXED_VALUES 1

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: support ICMP matching

Start support matching on icmpv4 and icmpv6.

Usage:
--icmpv4: add icmp item to match on.
--icmpv6: add icmpv6 item to match on.

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: add port mask option

Sometimes you need to check flow performance for
certain port and not all ports. Thus a portmask
option is needed.

Usage:
--portmask=N

Where N represent the hexadecimal bitmask of ports
used.

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: add random mark values

Instead of having single id value, use up to 256
values, thus we make sure that all flows will not
use same mark action.

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: fix IPv4 source matching

All value must be converted into intended endianness.

Fixes: bf3688f1e816 ("app/flow-perf: add insertion rate calculation")
Cc: stable@dpdk.org
Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: support VXLAN encap/decap actions

Introduce vxlan-encap and vxlan-decap actions.

vxlan-encap have fixed pattern and values for
encap data.

Usage example:
--vxlan-encap
--vxlan-decap

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: support raw encap/decap actions

Introduce raw-encap and raw-decap actions.
The two actions are added in command line
options, and for the data to encap or decap
the user need to parse it within the command
line.

All values of raw-encap data is set to be fixed
values.

Usage example:

--raw-encap=ether,ipv4,udp,vxlan

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: fix memory leak from RSS action

Currently, each call for add_rss_action will allocate
extra memory for rss_data, which will reflect bad results
on memory consumption for all flows, and will leads into
memory leak.

In this fix, it will check if it's allocated before
reallocating it.

Fixes: bf3688f1e816 ("app/flow-perf: add insertion rate calculation")
Cc: stable@dpdk.org
Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: support flag action

Introduce flag action support to flow perf
application.

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: support header modify actions

Introduce headers modify actions in the app.
All header modify actions will add different value
for each flow, to make sure each flow will create
and use it's own actions.

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: support user order

The old design was using the bit mask to identify
items, action and attributes.

So it was all based on the order of the code itself,
to place the order of the actions, items & attributes
inside the flows. Such design will lead into many failures
when some PMD support order different than other PMD,
in the end the rules will fail to create. Also sometimes
the user needs to have one action before other actions
and vice versa, so using new design of arrays that
take user order into consideration make more sense.

After this patch, we start supporting inner items
and more than one instance of same action.

Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

doc: fix section layout of the flow perf app guide

Currently all the sections are considered as main title under
DPDK Tools User Guides.

This fix will collect all flow perf sections under one title
which is Flow Performance Tool

Fixes: 3344cf2e3001 ("app/flow-perf: add flow performance skeleton")
Cc: stable@dpdk.org
Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

app/flow-perf: fix actions mask

Actions have it's own macro which is FLOW_ACTION_MASK

Fixes: bf3688f1e816 ("app/flow-perf: add insertion rate calculation")
Cc: stable@dpdk.org
Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Alexander Kozyrev <akozyrev@nvidia.com>

net/axgbe: support IEEE 1588 PTP

Add ethdev APIs to support PTP timestamping

Signed-off-by: Selwin Sebastian <selwin.sebastian@amd.com>
Acked-by: Amaranath Somalapuram <asomalap@amd.com>

net/cxgbe: fix duplicate MAC addresses in MPS TCAM

During MAC address insertion to MPS TCAM, add a default mask when
the mask is not explicitly specified. Otherwise, driver misses the
mask comparison and ends up inserting duplicate entries in the
MPS TCAM.

Fixes: 6fda3f0ddda9 ("net/cxgbe: add API to program hardware MPS table")
Cc: stable@dpdk.org
Signed-off-by: Karra Satwik <kaara.satwik@chelsio.com>
Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>

net/dpaa: fix port ID type in API

Updating the type for 'port' variable from 'uint8_t' to 'uint16_t'.

Fixes: 8c3495f5d2dd ("net/dpaa: support loopback API")
Cc: stable@dpdk.org
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>

net/pfe: fix misuse of interface index

Pfe pmd has no need to bound host interface
for which we require if_index field.
Setting it to 0 as unused.

Fixes: fe38ad9ba73e ("net/pfe: add device start/stop")
Cc: stable@dpdk.org
Reported-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Sachin Saxena <sachin.saxena@nxp.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/dpaa2: fix misuse of interface index

Dpaa2 pmd has no need to bound host interface
for which we require if_index field.
Setting it to 0 as unused.

Fixes: 3e5a335d3f88 ("net/dpaa2: add basic operations")
Cc: stable@dpdk.org
Reported-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Sachin Saxena <sachin.saxena@nxp.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

doc: update RSS flow action with best effort

Using the rte_flow action RSS types field,
may result in undefined outcome.

For example selecting both UDP and TCP,
selecting TCP RSS type but the pattern is targeting UDP traffic.
another option is that the PMD doesn't support all requested types.

Until now, it wasn't clear what will happen in such cases.
This commit clarify this issue by stating that the PMD
will work in the best-effort mode, and will fail
in case the requested type is not supported.

Signed-off-by: Ori Kam <orika@nvidia.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>

net/tap: free mempool when closing

When setup tx queues, we will create a mempool for the 'gso_ctx'.
The mempool is not freed when closing tap device. If free the tap
device and create it with different name, it will create a new
mempool. This maybe cause an OOM.

The snprintf function return value is not checked and the mempool
name may be truncated. This patch also fix it.

Fixes: 050316a88313 ("net/tap: support TSO (TCP Segment Offload)")
Cc: stable@dpdk.org
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

maintainers: update for igb/igc/ixgbe

Co-work with Jeff, setting me as new maintainer for igb, igc and ixgbe.

Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Acked-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Jeff Guo <jia.guo@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

net: check first segment length on SW VLAN insertion

SW VLAN insertion relies on Ethernet addresses location in contiguous
memory (do not split across mbuf segments). There is no any formal
requirements on data location and mbuf structure which guarantee it.
So, check it explicitly to avoid corrupted packets if the condition
is violated. Typically software VLAN insertion is done on Tx prepare
stage and application will get indication that the packet is invalid
and cannot be transmitted.

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

net/octeontx2: support TM length adjust and packet mode

This patch adds support to packet length adjust TM feature
for private shaper. It also adds support to packet mode
feature that applies both to private shaper and node DWRR
scheduling of SP children.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

app/testpmd: add TM command for non-leaf and packet mode

Add TM command to enable packet mode for all SP children
in non leaf node. This is a new command as
"add tm nonleaf node pktmode".

Also add support to shaper profile add command to take
packet mode parameter used to setup shaper in packet mode.
This adds an extra argument "packet_mode" to shaper profile add command
"add port tm node shaper profile" as last argument.

This patch also dumps new tm port/level/node capabilities
sched_wfq_packet_mode_supported, sched_wfq_byte_mode_supported,
shaper_private_packet_mode_supported, shaper_private_byte_mode_supported,
shaper_shared_packet_mode_supported, shaper_shared_byte_mode_supported.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

drivers/net: update TM capability

Since existing PMD's support shaper byte mode and scheduler
wfq byte mode, update the same in their port/level/node capabilities
that are added. SoftNIC PMD is already upto date with new capabilities.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>