Increase the maximum supported burst size for the bnxt vector
mode PMD from 32 to 64.
With larger burst sizes, per-burst overhead is amortized over more
packets, improving overall performance. For small packets this has
been measured to provide a 4-10% increase in single-core throughput
with testpmd iofwd.
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Don't allocate extra completion queue entries for aggregation
ring when aggregation ring will not be used.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Improve support for small ring sizes:
- Ensure that transmit free threshold is no more than 1/4 ring size.
- Ensure that receive free threshold is no more than 1/4 ring size.
- Validate requested ring sizes against minimum supported size.
- Use rxq receive free threshold instead of fixed maximum burst
size to trigger bulk receive buffer allocation.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Disable support for vector mode when async completions can be placed
in a receive completion ring and change the default for all platforms
to use a dedicated async completion ring.
Simplify completion handling in vector mode receive paths now that
it no longer needs to handle async completions.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Change the type of the software receive mbuf ring from an array
of structures containing an mbuf pointer to an array of pointers
to struct rte_mbuf for consistency with how this ring is currently
used by the vector mode receive function.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
We should use rte_pktmbuf_free() instead of rte_free() to free the mbuf.
Fixes: 6dc83230b43b ("net/bnxt: support port representor data path")
Cc: stable@dpdk.org
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
This patch removes logically dead code reported by coverity.
Coverity issue: 360824
Fixes: 6dc83230b43b ("net/bnxt: support port representor data path")
Cc: stable@dpdk.org
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
If PF reset is finished but VF reset is pending, VF should no need to
send any invalid cmd to PF. That would avoid mass unexpected behaviors
affecting the robust.
Fixes: 22b123a36d07 ("net/avf: initialize PMD")
Fixes: 9e03acd726cf ("net/iavf: fix flow access")
Cc: stable@dpdk.org
Signed-off-by: Jeff Guo <jia.guo@intel.com>
Tested-by: Hailin Xu <hailinx.xu@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Some IPv6 prefix related protocol header fields are defined in this
patch, so that we can use prefix instead of full IPv6 address for RSS.
Ref https://tools.ietf.org/html/rfc6052.
Signed-off-by: Junfeng Guo <junfeng.guo@intel.com>
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>
Most values from the VIRTCHNL_OP_GET_VF_RESOURCES are stored in the
iavf_hw_capabilities structure. Unfortunately, it seems that
max_mtu was missed. Add this member to the structure and save it when
parsing hw config.
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>
1. use BIT to replace <<
2. move VIRTCHNL_VF_CAP_DCF to keep order
3. align the vc msg validate
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>
Currently, there is no way for a VF driver to specify that it wants to
change its device/primary unicast MAC address. This makes it
difficult/impossible for the PF driver to track the VF's device/primary
unicast MAC address, which is used for VM/VF reboot and displaying on
the host. Fix this by using 2 bits of a pad byte in the
virtchnl_ether_addr structure so the VF can specify what type of MAC
it's adding/deleting.
Below are the values that should be used by all VF drivers going
forward.
VIRTCHNL_ETHER_ADDR_LEGACY(0):
- The type should only ever be 0 for legacy AVF drivers (i.e.
drivers that don't support the new type bits). The PF drivers
will track VF's device/primary unicast MAC using with best
effort.
VIRTCHNL_ETHER_ADDR_PRIMARY(1):
- This type should only be used when the VF is changing their
device/primary unicast MAC. It should be used for both delete
and add cases related to the device/primary unicast MAC.
VIRTCHNL_ETHER_ADDR_EXTRA(2):
- This type should be used when the VF is adding and/or deleting
MAC addresses that are not the device/primary unicast MAC. For
example, extra unicast addresses and multicast addresses
assuming the PF supports "extra" addresses at all.
If a PF is parsing the type field of the virtchnl_ether_addr, then it
should use the VIRTCHNL_ETHER_ADDR_TYPE_MASK to mask the first two bits
of the type field since 0, 1, and 2 are the only valid values.
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>
Kernel driver shows firmware version as hex but ice PMD shows
as decimal. This patch fixes the issue to make consistent with
kernel driver.
Fixes: f9204d8a23c3 ("net/ice: fix firmware version result of ethtool")
Cc: stable@dpdk.org
Signed-off-by: Shougang Wang <shougangx.wang@intel.com>
Acked-by: Qiming Yang <qiming.yang@intel.com>
'__builtin_add_overflow' added to gcc in version 5, earlier versions
causing build error, like gcc 4.8.5 in RHEL7.
Replaced compiler builtin check with arithmetic check.
Fixes: 7838d3a6ae7a ("net/netvsc: check for overflow on packet info from host")
Reported-by: Raslan Darawsheh <rasland@mellanox.com>
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Tested-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
'_rte_eth_dev_callback_process()' & '_rte_eth_dev_reset()' internal APIs
has unconventional underscore ('_') prefix.
Although this is not documented most probably this is to mark them as
internal. Since we have '__rte_internal' flag to mark this, removing '_'
from API names.
For '_rte_eth_dev_reset()', there is already a public API named
'rte_eth_dev_reset()', so renaming '_rte_eth_dev_reset()' to
'rte_eth_dev_internal_reset'.
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: David Marchand <david.marchand@redhat.com>
Acked-by: Sachin Saxena <sachin.saxena@nxp.com>
Hairpin helper functions were not used by drivers, but it was used only
local to ethdev. They are:
'rte_eth_dev_is_rx_hairpin_queue()'
'rte_eth_dev_is_tx_hairpin_queue()'
Exposing them as internal APIs and update mlx5 driver (only user of
hairpin) to use them.
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: David Marchand <david.marchand@redhat.com>
This patch is a preparation to hide the 'struct eth_dev_ops' from
applications by moving some device operations from 'struct eth_dev_ops'
to 'struct rte_eth_dev'.
Mentioned ethdev APIs are in the data path and implemented as inline
because of performance reasons.
Exposing 'struct eth_dev_ops' to applications is bad because it is a
contract between ethdev and PMDs, not really needs to be known by
applications, also changes in the struct causing ABI breakages which
shouldn't.
To be able to both keep APIs inline and hide the 'struct eth_dev_ops',
moving device operations used in ethdev inline APIs to 'struct
rte_eth_dev' to the same level with Rx/Tx burst functions.
The list of dev_ops moved:
eth_rx_queue_count_t rx_queue_count;
eth_rx_descriptor_done_t rx_descriptor_done;
eth_rx_descriptor_status_t rx_descriptor_status;
eth_tx_descriptor_status_t tx_descriptor_status;
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Reviewed-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: David Marchand <david.marchand@redhat.com>
Acked-by: Sachin Saxena <sachin.saxena@nxp.com>
Move Rx queue drop action similar resources allocations from Verbs
module to a shared location.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Separate Rx queue drop creation into both Verbs and DevX modules.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Move Rx hash queue object similar resources allocations from DevX and
Verbs modules to a shared location.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Move Rx indirection table object similar resources allocations from DevX
and Verbs modules to a shared location.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Once the separation between Verbs and DevX is done using function
pointers, the type field of the indirection table structure becomes
redundant and no more code is used.
Remove the unnecessary field from the structure.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Separate Rx hash queue creation into both Verbs and DevX modules.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Separate Rx indirection table object creation into both Verbs and DevX
modules.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Use new modify_wq functions for Rx object creation in DevX and Verbs
modules.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Separate Rx object modification to the Verbs and DevX modules.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Rearrangement of WQ and CQ creation for Verbs Rx queue:
1. Rename the allocation function.
2. Reduce the number of arguments that the creation functions receive.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Rearrangement of RQ and CQ resource handling for DevX Rx queue:
1. Rename the allocation function so that it is understood that it
allocates all resources and not just the CQ or RQ.
2. Move the allocation and release of the doorbell into creation and
release functions.
3. Reduce the number of arguments that the creation functions receive.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Move Rx object similar resources allocations and debug logs from DevX
and Verbs modules to a shared location.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Separate interrupt event handler into both Verbs and DevX modules.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
As an arrangement to Windows OS support, the Verbs operations should be
separated to another file.
By this way, the build can easily cut the unsupported Verbs APIs from
the compilation process.
Define operation structure and DevX module in addition to the existing
linux Verbs module.
Separate Rx object creation into the Verbs/DevX modules and update the
operation structure according to the OS support and the user
configuration.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
The Rx queue structures manage 2 different reference counter per queue:
rxq_ctrl reference counter and rxq_obj reference counter.
There is no real need to use two different counters, it just complicates
the release functions.
Remove the rxq_obj counter and use only the rxq_ctrl counter.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Rx HW objects can be created by both Verbs and DevX operations.
The management of the 2 types of operations are done directly in the
main flow of the object’s creations.
Some arrangements and validations were wrongly done to the irrelevant
type:
1. LRO related validations were done for Verbs type where LRO is not
supported at all.
2. Verbs allocation arrangements were done for DevX operations where it
is not needed.
3. Doorbell destroy was considered for Verbs types where it is
irrelevant.
Adjust the aforementioned points only for the relevant types.
Fixes: e79c9be91515 ("net/mlx5: support Rx hairpin queues")
Fixes: 08d1838f645a ("net/mlx5: implement CQ for Rx using DevX API")
Fixes: 17ed314c6c0b ("net/mlx5: allow LRO per Rx queue")
Fixes: dc9ceff73c99 ("net/mlx5: create advanced RxQ via DevX")
Cc: stable@dpdk.org
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
In order to support DevX Rx queue stop and start operations, the state
of the queue should be updated in FW.
The state update PRM command requires to set both the current state and
the new requested state.
The current state and the new requested state fields setting were
wrongly switched.
Switch them back to the correct setting.
Fixes: 161d103b231c ("net/mlx5: add queue start and stop")
Cc: stable@dpdk.org
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
The mlx5_hrxq_new function allocates several resources and if one of the
allocations fails, the function jumps to an error label where it
releases all the allocated resources.
When the TIR action creation fails, the hrxq memory is not released what
can cause a resource leak.
Add an appropriate release to the hrxq pointer in the error flow.
Fixes: 772dc0eb83d3 ("net/mlx5: convert hrxq to indexed")
Fixes: dc9ceff73c99 ("net/mlx5: create advanced RxQ via DevX")
Cc: stable@dpdk.org
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
Replace behavior with RTE_LIBRTE_ARK_MIN_TX_PKTLEN
with a default value of 0.
Update documentation as needed.
Signed-off-by: Ed Czeck <ed.czeck@atomicrules.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Use ARK_PMD_LOG in place of PMD_DRV_LOG, PMD_DEBUG_LOG, PMD_FUNC_LOG,
PMD_STATS_LOG, PMD_RX_LOG, and PMD_TX_LOG.
Review and adjust log levels and messages as needed.
Signed-off-by: Ed Czeck <ed.czeck@atomicrules.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
When loading the OS default package and the pipeline mode is enabled
by the "pipeline-mode-support=1" operation. In this case, the wrong
parser is selected for processing and it will cause the unsupported
patterns(pppoes/pfcp/l2tpv3/esp/ah) to be validated successfully.
This patch corrects the parser selection issue.
Fixes: 47d460d63233 ("net/ice: rework switch filter")
Cc: stable@dpdk.org
Signed-off-by: Guinan Sun <guinanx.sun@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Current RSS implementation is not easy to scale and maintain.
The patch refactor the code base on below design:
1. iavf_pattern_match_item->input_set_mask is the superset of
ETH_RSS_xxx.
2. iavf_pattern_match_item->meta is the virtchnl_proto_hdrs template.
3. iavf_hash_parse_pattern will generate pattern hint.
4. iavf_hash_parse_action will refine the virtchnl_proto_hdrs base on
pattern hint and ETH_RSS_xxx.
5. The refine process include
1) refine field selector of l2, l3, l4.
2) insert gtpu proto_hdr at the beginning base on pattern hint.
3) refine field selector for gtpu header.
The patch reduce the code from 4000+ line to less than 1000.
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Jeff Guo <jia.guo@intel.com>
ptype 264, 265, 266, 267, 275 should not be set
in ice_ptypes_ipv6_ofos_all.
Fixes: 88824213be8a ("net/ice/base: enable RSS for PFCP/L2TP/ESP/AH")
Cc: stable@dpdk.org
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Jeff Guo <jia.guo@intel.com>
With fmlib (FMCLESS) mode now RSS can be modified on runtime.
This patch add support for RSS update functions
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Signed-off-by: Sachin Saxena <sachin.saxena@nxp.com>
FMC tool generates and saves the setup in a file.
This patch help Parse the /tmp/fmc.bin generated by FMC to
setup RXQs for each port on FMC mode.
The parser gets the fqids and vspids from fmc.bin
Signed-off-by: Jun Yang <jun.yang@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
This patch adds support for Virtual Storage profile (VSP) feature.
With VSP support when memory pool is created, the hw buffer pool id
i.e. bpid is not allocated; the bpid is identified by dpaa flow
create API.
The memory pool of RX queue is attached to specific BMan pool
according to the VSP ID when RX queue is setup.
For fmlib based hash queue, VSP base ID is assigned to each queue.
Signed-off-by: Jun Yang <jun.yang@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
A shared MAC interface is an interface which can be used
by both kernel and userspace based on classification configuration
It is defined in dts with the compatible string
"fsl,dpa-ethernet-shared" which bpool will be seeded by the dpdk
partition and configured as a netdev by the dpaa Linux eth driver.
User space buffers from the bpool will be kmapped by the kernel.
Signed-off-by: Radu Bulie <radu-andrei.bulie@nxp.com>
Signed-off-by: Jun Yang <jun.yang@nxp.com>
Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
This patch uses fmlib to configure the FMAN HW for flow
and distribution configuration, thus avoiding the need
for static FMC tool execution optionally.
Signed-off-by: Sachin Saxena <sachin.saxena@nxp.com>
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>
