dpdk.git
5 years agonet/sfc/base: implement proxy auth MCDI event handling
Gautam Dawar [Mon, 10 Jun 2019 07:38:40 +0000 (08:38 +0100)]
net/sfc/base: implement proxy auth MCDI event handling

Add the capability to receive MCDI proxy event from firmware and
invoke the client driver registered function to handle it.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: support proxy auth operations for SR-IOV
Gautam Dawar [Mon, 10 Jun 2019 07:38:39 +0000 (08:38 +0100)]
net/sfc/base: support proxy auth operations for SR-IOV

VMware expects that certain kind of configurations made on the VFs are
authorized by the ESXi before these are applied e.g. assigning a MAC
address to the VF, setting MTU etc. Firmware supports a feature called
MCDI proxy which will be used to implement this authorization check.
The proxy auth module is governed by EFSYS_OPT_MCDI_PROXY_AUTH switch.
This patch adds the framework for proxy auth module along with the APIs
required for SR-IOV initialization.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: support data path with EVB module
Gautam Dawar [Mon, 10 Jun 2019 07:38:38 +0000 (08:38 +0100)]
net/sfc/base: support data path with EVB module

ef10_nic_init() allocates a vAdaptor for the physical port in current
flow. In case of SR-IOV, this vAdaptor must be created for the PF as the
vSwitch is allocated on the physical port. So, the call to
efx_mcdi_vadaptor_alloc() should be avoided in ef10_nic_init() in SR-IOV
flow. To achieve this, for SR-IOV use case, the vSwitch is created
before NIC initialization and its handle is used to prevent vAdaptor
allocation in ef10_nic_init(). This approach has been taken to minimize
the changes in NIC initialization flow.

This is also the case with Linux driver where vSwitch creation happens
before NIC initialization.

Also, when DMA queues need to be allocated for Tx/Rx functionality
(MC_CMD_INIT_RXQ / MC_CMD_INIT_TXQ), the correct vPort is selected
based on efx_vswitch_t property of efx_nic_t structure - vport
corresponding to PF in case of SR-IOV use case and EVB_PORT_ID_ASSIGNED
for physical port.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: factor out upstream port vAdaptor allocation
Gautam Dawar [Mon, 10 Jun 2019 07:38:37 +0000 (08:38 +0100)]
net/sfc/base: factor out upstream port vAdaptor allocation

Separate out vAdaptor allocation from ef10_nic_init() as it is not
required for SR-IOV use case. In case of SR-IOV, vAdaptor is allocated
early along with vSwitch creation and vPort configuration.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: implement vSwitch create/destroy
Gautam Dawar [Mon, 10 Jun 2019 07:38:36 +0000 (08:38 +0100)]
net/sfc/base: implement vSwitch create/destroy

The vSwitch create API takes an array of num_vports client driver
allocated vPort config entries where entry at index 0 contains the PF
configuration and rest num_vports-1 entries refer to vPort configuration
for VFs 0 to (num_vports-2). The required hierarchy
(vswitch/vport/vadaptor) is created within this API. The destroy API
tears down this hierarchy and releases memory for the vSwitch object.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add EVB module vSwitch/vPort/vAdaptor ops
Gautam Dawar [Mon, 10 Jun 2019 07:38:35 +0000 (08:38 +0100)]
net/sfc/base: add EVB module vSwitch/vPort/vAdaptor ops

Implement functions to allocate and free vSwitch, vPort and vAdaptor.
Also, implement functions to add and delete vPort MAC address and EVB
port assign.

Most of the efx_evb_ops_t functions take vSwitch ID as a parameter for
future enhancements. Currently, firmware doesn't implement vSwitch
identifier and hence this parameter is unused for EF10 architecture
implementation.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add MCDI wrappers for vPort and vSwitch in EVB
Gautam Dawar [Mon, 10 Jun 2019 07:38:34 +0000 (08:38 +0100)]
net/sfc/base: add MCDI wrappers for vPort and vSwitch in EVB

Add the functions to implement the vPort and vSwitch MCDI calls.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: introduce EVB module for SR-IOV
Gautam Dawar [Mon, 10 Jun 2019 07:38:33 +0000 (08:38 +0100)]
net/sfc/base: introduce EVB module for SR-IOV

Implement the framework for Enterprise Virtual Briding (EVB) module.
SR-IOV augments the software virtual switch with NIC capabilities
supported from EVB module.
Further patches will add APIs to create and destroy EVB switching
hierarchy required for SR-IOV and APIs to set vPort properties like MAC,
VLAN etc.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: export the zero-based MCDI port number
Gautam Dawar [Mon, 10 Jun 2019 07:38:32 +0000 (08:38 +0100)]
net/sfc/base: export the zero-based MCDI port number

Proxy authorization module for SR-IOV requires one instance of proxy
data structures per card. In order to achieve this, proxy data
structures will be allocated only for primary port (port id 0) and other
secondary ports in the card will access those data structures through
reference to primary port. Accordingly, the port number obtained from
efx_mcdi_get_port_assignment is stored in NIC configuration as
enc_mcdi_port.

Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add definition of OEM TLV
Kevin Lampis [Mon, 10 Jun 2019 07:38:31 +0000 (08:38 +0100)]
net/sfc/base: add definition of OEM TLV

Allows to enable additional functionality related to this OEM
(e.g. vendor extensions to VPD, NC-SI etc.)

Signed-off-by: Kevin Lampis <klampis@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add definition of bundle metadata partition
Paul Fox [Mon, 10 Jun 2019 07:38:30 +0000 (08:38 +0100)]
net/sfc/base: add definition of bundle metadata partition

Signed-off-by: Paul Fox <pfox@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add background mode firmware updating
Richard Houldsworth [Mon, 10 Jun 2019 07:38:29 +0000 (08:38 +0100)]
net/sfc/base: add background mode firmware updating

Request firmware updates be performed in background mode.
In this mode MCDI to the function processing the update
remains accessible and the client polls for completion.
This is supported for lengthy partition updates such as
MCFW and bundles. The MC ignores the flags used for this
mode for other partition updates.

Signed-off-by: Richard Houldsworth <rhouldsworth@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: transition to the extensible NVRAM info API
Richard Houldsworth [Mon, 10 Jun 2019 07:38:28 +0000 (08:38 +0100)]
net/sfc/base: transition to the extensible NVRAM info API

Old NVRAM info API required function prototype too often.

Signed-off-by: Richard Houldsworth <rhouldsworth@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: support direct FW update for bundle partitions
Richard Houldsworth [Mon, 10 Jun 2019 07:38:27 +0000 (08:38 +0100)]
net/sfc/base: support direct FW update for bundle partitions

All signed images other than for the MCFW partition should
be written fully to the partition with no rearrangement.

Signed-off-by: Richard Houldsworth <rhouldsworth@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add firmware ID header
Richard Houldsworth [Mon, 10 Jun 2019 07:38:26 +0000 (08:38 +0100)]
net/sfc/base: add firmware ID header

Add definitions header for reflash header IDs. This is required to
support different handling modes for signed firmware images.

Signed-off-by: Richard Houldsworth <rhouldsworth@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: update MCDI headers
Andrew Rybchenko [Mon, 10 Jun 2019 07:38:25 +0000 (08:38 +0100)]
net/sfc/base: update MCDI headers

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add NVRAM info
Richard Houldsworth [Mon, 10 Jun 2019 07:38:24 +0000 (08:38 +0100)]
net/sfc/base: add NVRAM info

Add function to query partition characteristics.
Refactor efx_nvram_size to share implementation.

Signed-off-by: Richard Houldsworth <rhouldsworth@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add extensible NVRAM info function
Richard Houldsworth [Mon, 10 Jun 2019 07:38:23 +0000 (08:38 +0100)]
net/sfc/base: add extensible NVRAM info function

Includes the partition read-only flag, to allow for
checks before opening the partition.

Signed-off-by: Richard Houldsworth <rhouldsworth@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add capabilities for bundle partition
Richard Houldsworth [Mon, 10 Jun 2019 07:38:22 +0000 (08:38 +0100)]
net/sfc/base: add capabilities for bundle partition

Signed-off-by: Richard Houldsworth <rhouldsworth@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: add driver version string registration
Richard Houldsworth [Mon, 10 Jun 2019 07:38:21 +0000 (08:38 +0100)]
net/sfc/base: add driver version string registration

Support the registration of a version string by the libefx
client. The string is passed on to the MC in efx_nic_probe
only to allow the MC to advertise the OS driver version in
NC-SI, and the content is considered opaque for libefx.

Signed-off-by: Richard Houldsworth <rhouldsworth@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: do not rely on indirect header inclusion
Andrew Rybchenko [Mon, 10 Jun 2019 07:38:20 +0000 (08:38 +0100)]
net/sfc/base: do not rely on indirect header inclusion

Types defined in efx_types.h are used in efx.h and it is better do
not rely on the header inclusion from somewhere else (typically from
efsys.h).

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: improve code style in sensors decoding
Mark Spender [Mon, 10 Jun 2019 07:38:19 +0000 (08:38 +0100)]
net/sfc/base: improve code style in sensors decoding

Add more comments to simplify code reading and understanding.

Signed-off-by: Mark Spender <mspender@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: fix shift by more bits than field width
Mark Spender [Mon, 10 Jun 2019 07:38:18 +0000 (08:38 +0100)]
net/sfc/base: fix shift by more bits than field width

This was probably an oversight when support for multiple sensor pages
was added.

Despite being undefined behaviour in C, it probably worked on Intel
x32/x64 as on them bit shift operations wrap round.

Fixes: dfb3b1ce15f6 ("net/sfc/base: import monitors access via MCDI")
Cc: stable@dpdk.org
Signed-off-by: Mark Spender <mspender@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: fix signed/unsigned mismatch
Andrew Lee [Mon, 10 Jun 2019 07:38:17 +0000 (08:38 +0100)]
net/sfc/base: fix signed/unsigned mismatch

Use UINT32_MAX instead of assigning -1 to a uint32_t variable to
resolve "conversion from 'int' to 'uint32_t', signed/unsigned mismatch"
errors produced by the Visual Studio 2017 toolchain [with the default
/W4 /WX C compiler options which set warning level 4 and treat warnings
as errors].

Fixes: 107cf1d792cb ("net/sfc/base: move limits config to ef10 NIC board config")
Cc: stable@dpdk.org
Signed-off-by: Andrew Lee <alee@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc/base: enable chained multicast on all EF10 cards
Gautam Dawar [Mon, 10 Jun 2019 07:38:16 +0000 (08:38 +0100)]
net/sfc/base: enable chained multicast on all EF10 cards

Set WORKAROUND_BUG26807 which does the job.

Fix the misunderstanding in the Medford code: i.e. the workaround is
always supported by firmware, but the driver still needs to enable it.
Also, as it now applies to all EF10 controllers, the implementation is
moved to EF10 common place.

Fixes: 94190e3543bf ("net/sfc/base: import SFN8xxx family support")
Fixes: 2b38e7b7b7e1 ("net/sfc/base: add Medford2 support to NIC module")
Cc: stable@dpdk.org
Signed-off-by: Gautam Dawar <gdawar@solarflare.com>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agotelemetry: fix memory leak
Reshma Pattan [Wed, 29 May 2019 12:43:17 +0000 (13:43 +0100)]
telemetry: fix memory leak

Free the `values` pointer before returning
from rte_telemetry_command_ports_all_stat_values()
to avoid memory leak.

Fixes: c12cefa379 ("telemetry: fix mapping of statistics")
Cc: stable@dpdk.org
Signed-off-by: Reshma Pattan <reshma.pattan@intel.com>
Acked-by: Kevin Laatz <kevin.laatz@intel.com>
5 years agomaintainers: add Windows exports script
Thomas Monjalon [Thu, 13 Jun 2019 14:11:55 +0000 (23:11 +0900)]
maintainers: add Windows exports script

A script used with meson was missing in the list of files.

Fixes: 4c773788e05a ("build: generate Windows exports file")

Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Reviewed-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
5 years agokni: remove PCI related information
Igor Ryzhov [Thu, 6 Jun 2019 13:26:26 +0000 (16:26 +0300)]
kni: remove PCI related information

As there is no ethtool support in KNI anymore,
PCI related information is no longer needed.

Fixes: ea6b39b5b847 ("kni: remove ethtool support")

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agodoc: fix typo in EAL guide
Thomas Monjalon [Sun, 16 Jun 2019 14:32:06 +0000 (17:32 +0300)]
doc: fix typo in EAL guide

The EAL init diagram had a typo for "lauch"
instead of "launch".

Fixes: fc1f2750a3ec ("doc: programmers guide")
Cc: stable@dpdk.org
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Reviewed-by: David Marchand <david.marchand@redhat.com>
5 years agoapp/testpmd: fix crash
Ferruh Yigit [Fri, 14 Jun 2019 17:19:39 +0000 (18:19 +0100)]
app/testpmd: fix crash

Testpmd tries to calculate mbuf size based on "max Rx packet size" and
"max MTU segment number".
When driver set a "nb_mtu_seg_max" to zero, it causes division by zero
segmentation fault in testpmd.

If the PMD set "nb_mtu_seg_max" to zero, testpmd shouldn't try to
calculate the mbuf size.

Fixes: 33f9630fc23d ("app/testpmd: create mbuf based on max supported segments")

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agobus/pci: fix TOCTOU for sysfs access
Stephen Hemminger [Tue, 2 Apr 2019 18:51:05 +0000 (11:51 -0700)]
bus/pci: fix TOCTOU for sysfs access

Using access followed by open causes a static analysis warning
about Time of check versus Time of use. Also, access() and
open() have different UID permission checks.

This is not a serious problem; but easy to fix by using errno instead.

Coverity issue: 300870
Fixes: 4a928ef9f611 ("bus/pci: enable write combining during mapping")
Cc: stable@dpdk.org
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Reviewed-by: David Marchand <david.marchand@redhat.com>
5 years agonet/bnxt: fix build with some compilers
Ajit Khaparde [Fri, 14 Jun 2019 04:20:57 +0000 (09:50 +0530)]
net/bnxt: fix build with some compilers

The driver was defining its own version of roundup which was
conflicting with another version defined elsewhere.

Change the local definition of roundup to avoid compilation errors.

Fixes: f8168ca0e690 ("net/bnxt: support thor controller")

Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/mlx5: fix master device Netlink socket sharing
Viacheslav Ovsiienko [Thu, 6 Jun 2019 12:23:01 +0000 (12:23 +0000)]
net/mlx5: fix master device Netlink socket sharing

There is the patch [1] that uses master device Netlink socket
to retrieve master device link settings. This is not thread safe
because this resource may be in use by other call to the master
device itself. Using the same Netlink socket concurrently from
the multiple threads causes Netlink requests malfunction and
must be eliminated. The patch replaces master Netlink socket
with the socket from representor device.

[1] http://patches.dpdk.org/patch/53120/

Fixes: 0333b2f584d9 ("net/mlx5: inherit master link settings for representors")
Cc: stable@dpdk.org
Signed-off-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: recover secondary process Tx errors
Matan Azrad [Thu, 30 May 2019 10:20:39 +0000 (10:20 +0000)]
net/mlx5: recover secondary process Tx errors

The SQ errors recovery mechanism in the PMD invokes a Verbs
functions to modify the RQ states in order to reset the SQ and to
reactivate it.

These Verbs functions are not allowed to be invoked from a secondary
process, hence the PMD skips the recovery when the error is captured
by secondary processes queues.

Using the DPDK IPC mechanism the secondary process can request Verbs
queues state modifications to be done synchronically by the primary
process.

Add support for secondary process Tx errors recovery.

Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: recover secondary process Rx errors
Matan Azrad [Thu, 30 May 2019 10:20:38 +0000 (10:20 +0000)]
net/mlx5: recover secondary process Rx errors

The RQ errors recovery mechanism in the PMD invokes a Verbs functions to
modify the RQ states in order to reset the RQ and to reactivate it.

These Verbs functions are not allowed to be invoked from a secondary
process, hence the PMD skips the recovery when the error is captured by
secondary processes queues.

Using the DPDK IPC mechanism the secondary process can request Verbs
queues state modifications to be done synchronically by the primary
process.

Add support for secondary process Rx errors recovery.

Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: handle Tx completion with error
Matan Azrad [Thu, 30 May 2019 10:20:37 +0000 (10:20 +0000)]
net/mlx5: handle Tx completion with error

When WQEs are posted to the HW to send packets, the PMD may get a
completion report with error from the HW, aka error CQE which is
associated to a bad WQE.

The error reason may be bad address, wrong lkey, bad sizes, etc.
that can wrongly be configured by the PMD or by the user.

Checking all the optional mistakes to prevent error CQEs doesn't make
sense due to performance impacts and huge complexity.

The error CQEs change the SQ state to error state what causes all the
next posted WQEs to be completed with CQE flush error forever.

Currently, the PMD doesn't handle Tx error CQEs and even may crashed
when one of them appears.

Extend the Tx data-path to detect these error CQEs, to report them by
the statistics error counters, to recover the SQ by moving the state
to ready again and adjusting the management variables appropriately.

Sometimes the error CQE root cause is very hard to debug and even may
be related to some corner cases which are not reproducible easily, hence
a dump file with debug information will be created for the first number
of error CQEs, this number can be configured by the PMD probe
parameters.

Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: extend Rx completion with error handling
Matan Azrad [Thu, 30 May 2019 10:20:36 +0000 (10:20 +0000)]
net/mlx5: extend Rx completion with error handling

When WQEs are posted to the HW to receive packets, the PMD may receive
a completion report with error from the HW, aka error CQE which is
associated to a bad WQE.

The error reason may be bad address, wrong lkey, small buffer size,
etc. that can wrongly be configured by the PMD or by the user.

Checking all the optional mistakes to prevent error CQEs doesn't make
sense due to performance impacts, moreover, some error CQEs can be
triggered because of the packets coming from the wire when the DPDK
application has no any control.

Most of the error CQE types change the RQ state to error state what
causes all the next received packets to be dropped by the HW and to be
completed with CQE flush error forever.

The current solution detects these error CQEs and even reports the
errors to the user by the statistics error counters but without
recovery, so if the RQ inserted to the error state it never moves to
ready state again and all the next packets ever will be dropped.

Extend the error CQEs handling for recovery by moving the state to
ready again, and rearranging all the RQ WQEs and the management
variables appropriately.

Sometimes the error CQE root cause is very hard to debug and even may
be related to some corner cases which are not reproducible easily,
hence a dump file with debug information will be created for the first
number of error CQEs, this number can be configured by the PMD probe
parameters.

Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: separate Rx queue initialization
Matan Azrad [Thu, 30 May 2019 10:20:35 +0000 (10:20 +0000)]
net/mlx5: separate Rx queue initialization

Move the RQ WQEs initialization code to separate function as an
arrangement to CQE error recovering for code reuse.

CC: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: mitigate Rx doorbell memory barrier
Matan Azrad [Thu, 30 May 2019 10:20:34 +0000 (10:20 +0000)]
net/mlx5: mitigate Rx doorbell memory barrier

The RQ WQEs must be written in the memory before the HW gets the RQ
doorbell, hence a memory barrier should be triggered after the WQEs
writing and before the doorbell writing.

The current code used rte_wmb barrier which ensures that all the memory
stores were done while it is enough to use rte_cio_wmb barrier for the
local memory stores because the WQEs are in local memory.

CC: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: fix device arguments error detection
Matan Azrad [Thu, 30 May 2019 10:20:33 +0000 (10:20 +0000)]
net/mlx5: fix device arguments error detection

When bad device arguments are added to the DPDK command line, the PMD
ignores all the command line arguments specified by the user and uses
the default values instead.

This behavior doesn't make sense because the user intention is to force
some device parameters and expects to get an error in case of
problematic issues with the arguments.

Stop probing and report an error in case of problematic command line
arguments.

Fixes: e72dd09b614e ("net/mlx5: add support for configuration through kvargs")
Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: add log file procedure for debug data
Matan Azrad [Thu, 30 May 2019 10:20:32 +0000 (10:20 +0000)]
net/mlx5: add log file procedure for debug data

Add a global function in the PMD which dumps debug information to
specific file.

The data can be printed in hexadecimal format or as regular string.

The number of debug files per PMD entity should be limited by a new PMD
probe parameter called max_dump_files_num.

The files will be created in the /var/log directory or in the current
directory.

Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/mlx5: remove Rx queues indexes correlation
Matan Azrad [Thu, 30 May 2019 10:20:31 +0000 (10:20 +0000)]
net/mlx5: remove Rx queues indexes correlation

There is a full correlation between the CQE indexes to the WQE indexes
in the vectorized Rx queues management.

When the RQ is inserted to the reset state, the correlation may break
because the HW starts the RQ polling from index 0 while the CQ polling
continues regularly.

As an arrangement to CQE errors handling, when the RQ can be reset,
the correlation dependence should be removed from all the Rx queues
index managements.

Remove the aforementioned dependence from the vectorized Rx burst
functions.

Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
5 years agonet/ixgbevf: add full link status check option
Haiyue Wang [Fri, 7 Jun 2019 15:56:21 +0000 (23:56 +0800)]
net/ixgbevf: add full link status check option

To get the VF's link status by calling 'rte_eth_link_get_nowait()', the
VF not only check PF's physical link status, but also check the mailbox
running status. And mailbox checking will generate mailbox interrupt in
PF, it will be worse if many VFs are running in the system, the PF will
have to handle many interrrupts.

Normally, checking the PF's physical link status is enough for nowait.
For different scenarios, adding an 'pflink_fullchk' option to control
whether to check the link fully or not.

Fixes: 91546fb62e67 ("net/ixgbevf: fix link state")
Cc: stable@dpdk.org
Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Acked-by: Scott Daniels <daniels@research.att.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
5 years agonet/bnxt: fix icc build
Naresh Kumar PBS [Sat, 8 Jun 2019 19:22:08 +0000 (23:22 +0400)]
net/bnxt: fix icc build

Address build errors reported by intel compiler while compiling
on Windows. Instead of typeof() using the actual type in ALLOW_FUNC

Cc: stable@dpdk.org
Signed-off-by: Naresh Kumar PBS <nareshkumar.pbs@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
5 years agonet/bnxt: fix interrupt vector initialization
Naresh Kumar PBS [Sat, 8 Jun 2019 19:22:07 +0000 (23:22 +0400)]
net/bnxt: fix interrupt vector initialization

Initialize the vector array when it is valid, thereby
preventing a case were it may be accessed when
the array is unallocated

Fixes: 1fe427fd08ee ("net/bnxt: support enable/disable interrupt")
Cc: stable@dpdk.org
Signed-off-by: Naresh Kumar PBS <nareshkumar.pbs@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
5 years agonet/bnxt: fix xstats
Kalesh AP [Sat, 8 Jun 2019 19:22:06 +0000 (23:22 +0400)]
net/bnxt: fix xstats

If the HWRM_PORT_QSTATS_EXT fails to initialize
fw_rx_port_stats_ext_size or fw_tx_port_stats_ext_size,
the driver can end up passing junk statistics to the application.

Instead of relying on the application to initialize the xstats
buffer before calling the xstats_get dev_op, memset xstats
with zeros to avoid returning or displaying incorrect statistics.

Also fixed the buffer starting offset.

Fixes: f55e12f33416 ("net/bnxt: support extended port counters")
Cc: stable@dpdk.org
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
5 years agonet/bnxt: use configured MTU during load
Somnath Kotur [Sat, 8 Jun 2019 19:22:05 +0000 (23:22 +0400)]
net/bnxt: use configured MTU during load

The MTU value of a port can be (re)configured out-of-band.
FW will be returning this configured MTU as part of func_qcfg cmd.
Driver to use this value during load time.

Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: check for null completion ring doorbell
Somnath Kotur [Sat, 8 Jun 2019 19:22:04 +0000 (23:22 +0400)]
net/bnxt: check for null completion ring doorbell

It is observed that sometimes during init, the bnxt_int_handler() gets
invoked while the cpr->cp_db.doorbell is not yet initialized. Check for
the same and return.

Fixes: f7ecea911ec5 ("net/bnxt: fix interrupt handler")
Cc: stable@dpdk.org
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: support redirecting tunnel packets to VF
Somnath Kotur [Sat, 8 Jun 2019 19:22:03 +0000 (23:22 +0400)]
net/bnxt: support redirecting tunnel packets to VF

Add code to redirect GRE, NVGRE and VXLAN tunnel packets
to the specified VF.

Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/enic: remove flow locks
John Daley [Thu, 6 Jun 2019 15:26:58 +0000 (08:26 -0700)]
net/enic: remove flow locks

There is no requirement for thread safety in the flow PMD code and no
need for the locks.

Fixes: 6ced137607d0 ("net/enic: flow API for NICs with advanced filters enabled")
Cc: stable@dpdk.org
Signed-off-by: John Daley <johndale@cisco.com>
Reviewed-by: Hyong Youb Kim <hyonkim@cisco.com>
5 years agonet/enic: remove flow count action support
John Daley [Thu, 6 Jun 2019 15:26:57 +0000 (08:26 -0700)]
net/enic: remove flow count action support

The firmware in 1400 series VIC adapters which would support COUNT
flow action was postponed and reworked. The capability will be
re-added in a future release when the firmware is available.

This reverts the following commits.
commit 86df6c4e2fce ("net/enic: support flow counter action")
commit 1b4ce87dc5e6 ("net/enic: fix counter action")

Cc: stable@dpdk.org
Signed-off-by: John Daley <johndale@cisco.com>
Reviewed-by: Hyong Youb Kim <hyonkim@cisco.com>
5 years agonet/enic: report speed capabilities
Hyong Youb Kim [Thu, 6 Jun 2019 15:26:56 +0000 (08:26 -0700)]
net/enic: report speed capabilities

Available link speeds are based on VIC adapter model, which is encoded
in PCI subsystem device ID.

Signed-off-by: Hyong Youb Kim <hyonkim@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>
5 years agonet/enic: set min and max MTU
Hyong Youb Kim [Thu, 6 Jun 2019 15:26:55 +0000 (08:26 -0700)]
net/enic: set min and max MTU

These values correspond to those used in the MTU handler (enic_set_mtu).

Signed-off-by: Hyong Youb Kim <hyonkim@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>
5 years agonet/sfc: support Rx interrupts for ef10 datapath
Georgiy Levashov [Thu, 6 Jun 2019 17:33:25 +0000 (18:33 +0100)]
net/sfc: support Rx interrupts for ef10 datapath

Similar to support for efx datapath, Rx interrupt disabling
just avoids rearming the next time.

Signed-off-by: Georgiy Levashov <georgiy.levashov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/sfc: support Rx interrupts for efx datapath
Georgiy Levashov [Thu, 6 Jun 2019 17:33:24 +0000 (18:33 +0100)]
net/sfc: support Rx interrupts for efx datapath

When Rx interrupts are disabled, we simply disable rearm when
the interrupt fires the next time. So, the next packet will
trigger interrupt (if it is not happened yet after previous Rx
burst processing).

Signed-off-by: Georgiy Levashov <georgiy.levashov@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/vmxnet3: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:16:04 +0000 (09:16 +0100)]
net/vmxnet3: fix Tx prepare to set positive rte_errno

Fixes: baf3bbae5556 ("net/vmxnet3: add Tx preparation")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: Yong Wang <yongwang@vmware.com>
5 years agonet/qede: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:16:03 +0000 (09:16 +0100)]
net/qede: fix Tx prepare to set positive rte_errno

Fixes: 29540be7efce ("net/qede: support LRO/TSO offloads")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/ixgbe: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:16:02 +0000 (09:16 +0100)]
net/ixgbe: fix Tx prepare to set positive rte_errno

Fixes: 7829b8d52be0 ("net/ixgbe: add Tx preparation")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/ice: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:16:01 +0000 (09:16 +0100)]
net/ice: fix Tx prepare to set positive rte_errno

Fixes: 17c7d0f9d6a4 ("net/ice: support basic Rx/Tx")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/iavf: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:16:00 +0000 (09:16 +0100)]
net/iavf: fix Tx prepare to set positive rte_errno

Fixes: a2b29a7733ef ("net/avf: enable basic Rx Tx")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/i40e: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:15:59 +0000 (09:15 +0100)]
net/i40e: fix Tx prepare to set positive rte_errno

Fixes: 3f33e643e5c6 ("net/i40e: add Tx preparation")
Fixes: bfeed0262b0c ("net/i40e: check illegal packets")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/fm10k: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:15:58 +0000 (09:15 +0100)]
net/fm10k: fix Tx prepare to set positive rte_errno

Fixes: 9b134aa39716 ("net/fm10k: add Tx preparation")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: Xiao Wang <xiao.w.wang@intel.com>
5 years agonet/enic: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:15:57 +0000 (09:15 +0100)]
net/enic: fix Tx prepare to set positive rte_errno

Fixes: 1e81dbb5321b ("net/enic: add Tx prepare handler")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/e1000: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:15:56 +0000 (09:15 +0100)]
net/e1000: fix Tx prepare to set positive rte_errno

Fixes: 2b76648872c9 ("net/e1000: add Tx preparation")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/atlantic: fix Tx prepare to set positive rte_errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:15:55 +0000 (09:15 +0100)]
net/atlantic: fix Tx prepare to set positive rte_errno

Fixes: 2b1472d7150c ("net/atlantic: implement Tx path")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agoethdev: fix Tx prepare documentation to use positive errno
Andrew Rybchenko [Wed, 5 Jun 2019 08:15:54 +0000 (09:15 +0100)]
ethdev: fix Tx prepare documentation to use positive errno

Tx prepare documentation was incorrectly suggesting to use
negative rte_errno.

Fixes: 4fb7e803eb1a ("ethdev: add Tx preparation")
Fixes: 439a90b5f2a7 ("ethdev: reorder inline functions")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/bnx2x: fix supported max Rx/Tx descriptor count
Shahed Shaikh [Tue, 4 Jun 2019 18:53:52 +0000 (11:53 -0700)]
net/bnx2x: fix supported max Rx/Tx descriptor count

Driver does not provide limit on number Rx and Tx descriptors per queue,
this may result in application configuring 64k descriptors (default set
by rte_eth_dev_info_get()) and further result in issues in PMD and HW
flows due to unsupported number.

Fixes: 540a211084a7 ("bnx2x: driver core")
Cc: stable@dpdk.org
Signed-off-by: Shahed Shaikh <shshaikh@marvell.com>
Acked-by: Rasesh Mody <rmody@marvell.com>
5 years agonet/bnx2x: fix link state
Shahed Shaikh [Tue, 4 Jun 2019 18:53:51 +0000 (11:53 -0700)]
net/bnx2x: fix link state

Don't call bnx2x_link_status_update() from bnx2x_link_update().
Actual use case of bnx2x_link_status_update() is to update the link
status in shared memory between driver and MFW, and not to get the
link status from HW.

So ideally, bnx2x_link_status_update() should be called when there
is an actual link event or change in link status.

Calling bnx2x_link_status_update() from bnx2x_link_update() may
corrupt the data of link status in shared memory and result
in inconsistent state of link.

Fixes: 540a211084a7 ("bnx2x: driver core")
Cc: stable@dpdk.org
Signed-off-by: Shahed Shaikh <shshaikh@marvell.com>
Acked-by: Rasesh Mody <rmody@marvell.com>
5 years agonet/bnx2x: fix memory leak
Shahed Shaikh [Tue, 4 Jun 2019 18:53:50 +0000 (11:53 -0700)]
net/bnx2x: fix memory leak

bnx2x_free_hsi_mem() does not free DMA memory.
Fix it here.

Fixes: 540a211084a7 ("bnx2x: driver core")
Cc: stable@dpdk.org
Signed-off-by: Shahed Shaikh <shshaikh@marvell.com>
Acked-by: Rasesh Mody <rmody@marvell.com>
5 years agonet/bnx2x: fix interrupt flood
Shahed Shaikh [Tue, 4 Jun 2019 18:53:49 +0000 (11:53 -0700)]
net/bnx2x: fix interrupt flood

PMD sets up and clears the slow path interrupt status block in dev_start
and dev_stop flow and slow path interrupt status block DMA memory for
device is allocated in dev_configure flow.

This situation creates a state where, after dev_stop is called, and if
there is a slow path interrupt from device, PMD sees the old value of
status block consumer in dev_start flow, since DMA memory for status
block belongs to old configuration and dev_start will result in
new slow path interrupt status block configuration.
And since PMD fails to ack new slow path interrupt with correct status
block consumer value, device continues to trigger interrupt causing an
interrupt flood.

Fix is to create and destroy status block DMA memory in dev_start and
dev_stop flow instead of dev_configure and dev_close flow.

Fixes: 540a211084a7 ("bnx2x: driver core")
Cc: stable@dpdk.org
Signed-off-by: Shahed Shaikh <shshaikh@marvell.com>
Acked-by: Rasesh Mody <rmody@marvell.com>
5 years agonet/bnx2x: fix packet drop
Shahed Shaikh [Tue, 4 Jun 2019 18:53:48 +0000 (11:53 -0700)]
net/bnx2x: fix packet drop

Patch "8bd31421c593 ("net/bnx2x: fix ramrod timeout")"
introduced a regression where sc->scan_fp flags is
set for unexpectedly long time. So the slow path completion
handler flow is run unnecessarily which walks over receive
descriptor ring of fast path and drops the data packets while looking
for slow path completion descriptor out of fast path ring.

This issue is seen under heavy traffic with link events happening
in background.

Fixes: 8bd31421c593 ("net/bnx2x: fix ramrod timeout")
Cc: stable@dpdk.org
Signed-off-by: Shahed Shaikh <shshaikh@marvell.com>
Acked-by: Rasesh Mody <rmody@marvell.com>
5 years agonet/ena: fix assigning NUMA node to IO queue
Michal Krawczyk [Tue, 4 Jun 2019 10:59:36 +0000 (12:59 +0200)]
net/ena: fix assigning NUMA node to IO queue

Previous solution was using memzones in invalid way in hope to assign
IO queue to the appropriate NUMA zone.

The right way is to use socket_id from the rx/tx queue setup function
and then pass it to the IO queue.

Fixes: 3d3edc265fc8 ("net/ena: make coherent memory allocation NUMA-aware")
Cc: stable@dpdk.org
Signed-off-by: Michal Krawczyk <mk@semihalf.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
5 years agoapp/testpmd: create mbuf based on max supported segments
Sunil Kumar Kori [Mon, 27 May 2019 08:35:29 +0000 (14:05 +0530)]
app/testpmd: create mbuf based on max supported segments

Configuring buffer size based following parameters:
- max-pkt-len
- max supported segments per MTU

Buffer size are configured as given below:
- If platform supports infinite segments per packet then default
  buffer size is used.
- If platform supports nb_mtu_seg_max segments then buffer size
  is configured as (max-pkt-len / nb_mtu_seg_max) + headroom

Signed-off-by: Sunil Kumar Kori <skori@marvell.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agoethdev: add default value for max segment
Sunil Kumar Kori [Mon, 27 May 2019 08:35:28 +0000 (14:05 +0530)]
ethdev: add default value for max segment

rte_eth_dev_info structure exposes, nb_seg_max & nb_mtu_seg_max
to provide maximum number of supported segments for a given platform.

Defining UINT16_MAX as default value of above mentioned variables to
expose support of infinite/maximum segments.

Based on above values, application can decide best size for buffers
while creating mbuf pool.

Signed-off-by: Sunil Kumar Kori <skori@marvell.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/af_xdp: remove unused struct member
Xiaolong Ye [Thu, 30 May 2019 09:07:07 +0000 (17:07 +0800)]
net/af_xdp: remove unused struct member

Fixes: f1debd77efaf ("net/af_xdp: introduce AF_XDP PMD")
Cc: stable@dpdk.org
Signed-off-by: Xiaolong Ye <xiaolong.ye@intel.com>
5 years agonet/af_xdp: support multi-queue
Xiaolong Ye [Thu, 30 May 2019 09:07:06 +0000 (17:07 +0800)]
net/af_xdp: support multi-queue

This patch adds two parameters `start_queue` and `queue_count` to
specify the range of netdev queues used by AF_XDP pmd.

Signed-off-by: Xiaolong Ye <xiaolong.ye@intel.com>
5 years agonet/af_xdp: enable zero copy by external mbuf
Xiaolong Ye [Thu, 30 May 2019 09:07:05 +0000 (17:07 +0800)]
net/af_xdp: enable zero copy by external mbuf

Implement zero copy of af_xdp pmd through mbuf's external memory
mechanism to achieve high performance.

This patch also provides a new parameter "pmd_zero_copy" for user, so they
can choose to enable zero copy of af_xdp pmd or not.

To be clear, "zero copy" here is different from the "zero copy mode" of
AF_XDP, it is about zero copy between af_xdp umem and mbuf used in dpdk
application.

Suggested-by: Vipin Varghese <vipin.varghese@intel.com>
Suggested-by: Tummala Sivaprasad <sivaprasad.tummala@intel.com>
Suggested-by: Olivier Matz <olivier.matz@6wind.com>
Signed-off-by: Xiaolong Ye <xiaolong.ye@intel.com>
5 years agonet/failsafe: fix reported device info
Stephen Hemminger [Thu, 16 May 2019 03:53:10 +0000 (20:53 -0700)]
net/failsafe: fix reported device info

The failsafe driver device info had several issues in the
info it reported in dev_info_get:
  - it cleared dev_info->device set in rte_eth_dev_info_get
  - many fields (for example max_rx_queue) should be the minimum
    of all sub devices
  - it reported tx capa for the active transmit device, but
    the device may change.

There was enough messed up that ended up reworking the info_get
handler. There is no need to save current values or have a
template for defaults.

Fixes: 4e31ee26ed51 ("net/failsafe: report actual device capabilities")
Cc: stable@dpdk.org
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Acked-by: Gaetan Rivet <gaetan.rivet@6wind.com>
5 years agonet/i40e: fix uninitialized value
Mesut Ali Ergin [Thu, 6 Jun 2019 16:30:42 +0000 (09:30 -0700)]
net/i40e: fix uninitialized value

Initializes mark_spec pointer to NULL.

Coverity issue: 341075
Fixes: 0bbcfc706a2b ("net/i40e: support MARK and RSS flow action")

Signed-off-by: Mesut Ali Ergin <mesut.a.ergin@intel.com>
Acked-by: Kevin Traynor <ktraynor@redhat.com>
5 years agonet/bnxt: enable RSS for thor-based controllers
Lance Richardson [Sun, 2 Jun 2019 17:42:47 +0000 (13:42 -0400)]
net/bnxt: enable RSS for thor-based controllers

Make changes needed to support rss for thor-based controllers.

Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: enable completion coalescing for thor
Lance Richardson [Sun, 2 Jun 2019 17:42:45 +0000 (13:42 -0400)]
net/bnxt: enable completion coalescing for thor

Enable completion coalescing for Thor-based adapters.

Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: support thor controller
Lance Richardson [Sun, 2 Jun 2019 17:42:44 +0000 (13:42 -0400)]
net/bnxt: support thor controller

This commit adds support to the bnxt PMD for devices
based on the BCM57508 "thor" Ethernet controller.

Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
5 years agonet/bnxt: refactor ring allocation
Lance Richardson [Sun, 2 Jun 2019 17:42:43 +0000 (13:42 -0400)]
net/bnxt: refactor ring allocation

Reduce code duplication and prepare for supporting hardware with
different ring allocation requirements by refactoring ring
allocation code.

Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: refactor doorbell handling
Lance Richardson [Sun, 2 Jun 2019 17:42:42 +0000 (13:42 -0400)]
net/bnxt: refactor doorbell handling

Reduce code duplication and prepare for newer controllers that
use different doorbell protocols by refactoring doorbell handling
code.

Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: support extended HWRM request sizes
Lance Richardson [Sun, 2 Jun 2019 17:42:41 +0000 (13:42 -0400)]
net/bnxt: support extended HWRM request sizes

Enable support for extended request sizes.

Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: reset function earlier in initialization
Lance Richardson [Sun, 2 Jun 2019 17:42:40 +0000 (13:42 -0400)]
net/bnxt: reset function earlier in initialization

Move function reset to beginnng of initialization sequence.

Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: use consistent values for VNIC RSS rule
Lance Richardson [Sun, 2 Jun 2019 17:42:39 +0000 (13:42 -0400)]
net/bnxt: use consistent values for VNIC RSS rule

Use consistent values for vnic->rss_rule. No functional change,
these all equate to uint16_t 0xffff.

Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: fix variable width in endian conversion
Lance Richardson [Sun, 2 Jun 2019 17:42:38 +0000 (13:42 -0400)]
net/bnxt: fix variable width in endian conversion

Use 32-bit conversion width when converting to 32-bit values.

Fixes: 6371b91fb66d ("net/bnxt: add ring alloc/free")
Cc: stable@dpdk.org
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: fix ring type macro name
Lance Richardson [Sun, 2 Jun 2019 17:42:37 +0000 (13:42 -0400)]
net/bnxt: fix ring type macro name

Use consistent macro names for ring type values. (There is no
functional change, the "alloc" and "free" values are identical.)

Fixes: 6371b91fb66d ("net/bnxt: add ring alloc/free")
Cc: stable@dpdk.org
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/bnxt: fix endianness in ring macros
Lance Richardson [Sun, 2 Jun 2019 17:42:36 +0000 (13:42 -0400)]
net/bnxt: fix endianness in ring macros

Descriptor fields in CP ring are in little-endian form, convert
to CPU endian before performing arithmetic operations.

Also use more general comparison when checking for ring
index wrap.

Fixes: f2a768d4d186 ("net/bnxt: add completion ring")
Cc: stable@dpdk.org
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
5 years agonet/sfc: advertise offload capabilities by Tx datapaths
Andrew Rybchenko [Sat, 1 Jun 2019 08:42:44 +0000 (09:42 +0100)]
net/sfc: advertise offload capabilities by Tx datapaths

Tx datapath feature bits were useful on migration from the old offload API
to the new one. However, right now it just adds indirection which
complicates code reading and understanding. Also addition of a new
offloads requires addition of a new feature bits and makes patches longer
and harder to understand. So, remove feature bits which correspond to Tx
offloads and simply advertise device and per-queue offloads directly.
Generic code could still mask some offloads if running HW or FW does not
support it.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Igor Romanov <igor.romanov@oktetlabs.ru>
Reviewed-by: Ivan Malov <ivan.malov@oktetlabs.ru>
5 years agonet/sfc: advertise offload capabilities by Rx datapaths
Andrew Rybchenko [Sat, 1 Jun 2019 08:42:43 +0000 (09:42 +0100)]
net/sfc: advertise offload capabilities by Rx datapaths

Rx datapath feature bits were useful on migration from the old offload API
to the new one. However, right now it just adds indirection which
complicates code reading and understanding. Also addition of a new
offloads requires addition of a new feature bits and makes patches longer
and harder to understand. So, remove feature bits which correspond to Rx
offloads and simply advertise device and per-queue offloads directly.
Generic code could still mask some offloads if running HW or FW does not
support it.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Igor Romanov <igor.romanov@oktetlabs.ru>
Reviewed-by: Ivan Malov <ivan.malov@oktetlabs.ru>
5 years agoethdev: add a check on mempool during RxQ setup
David Marchand [Fri, 17 May 2019 08:15:08 +0000 (10:15 +0200)]
ethdev: add a check on mempool during RxQ setup

We currently have no check on the mempool pointer passed to
rte_eth_rx_queue_setup.
Let's avoid a plain crash when dereferencing it.

Suggested-by: Jens Freimann <jfreimann@redhat.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Jens Freimann <jfreimann@redhat.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
5 years agonet/ena: fix Rx checksum errors statistics
Michal Krawczyk [Tue, 28 May 2019 08:28:34 +0000 (10:28 +0200)]
net/ena: fix Rx checksum errors statistics

Rx checksum flags and input errors shouldn't be updated on Tx, as it
would work only for packets forwarding.

The ierrors statistic should be updated on Rx, right after checking
Rx checksum flags if the Rx checksum offload is enabled.

Fixes: 1173fca25af9 ("ena: add polling-mode driver")
Cc: stable@dpdk.org
Signed-off-by: Michal Krawczyk <mk@semihalf.com>
5 years agonet/ena: fix Tx statistics
Rafal Kozik [Tue, 14 May 2019 11:11:26 +0000 (13:11 +0200)]
net/ena: fix Tx statistics

Instead of counting number of used NIC Tx bufs just count number
of Tx packets.

Fixes: 45b6d86184fc ("net/ena: add per-queue software counters stats")
Cc: stable@dpdk.org
Signed-off-by: Rafal Kozik <rk@semihalf.com>
Acked-by: Michal Krawczyk <mk@semihalf.com>
5 years agonet/memif: introduce memory interface PMD
Jakub Grajciar [Thu, 6 Jun 2019 11:38:50 +0000 (13:38 +0200)]
net/memif: introduce memory interface PMD

Shared memory packet interface (memif) PMD allows for DPDK and any other
client using memif (DPDK, VPP, libmemif) to communicate using shared
memory. The created device transmits packets in a raw format. It can be
used with Ethernet mode, IP mode, or Punt/Inject. At this moment, only
Ethernet mode is supported in DPDK memif implementation. Memif is Linux
only.

Signed-off-by: Jakub Grajciar <jgrajcia@cisco.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
5 years agonet/virtio: remove useless pointer checks
Maxime Coquelin [Wed, 5 Jun 2019 10:00:39 +0000 (12:00 +0200)]
net/virtio: remove useless pointer checks

This patch removes useless checks on 'prev' pointer, as it
is always set before with a valid value.

Reviewed-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Tiwei Bie <tiwei.bie@intel.com>
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
5 years agonet/virtio: fix segment length in mergeable packed Rx
Maxime Coquelin [Wed, 5 Jun 2019 10:00:38 +0000 (12:00 +0200)]
net/virtio: fix segment length in mergeable packed Rx

Head segment data_len field is wrongly summed with the length
of all the segments of the chain, whereas it should be the
length of the first segment only.

Fixes: a76290c8f1cf ("net/virtio: implement Rx path for packed queues")
Cc: stable@dpdk.org
Reported-by: Yaroslav Brustinov <ybrustin@cisco.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Jens Freimann <jfreimann@redhat.com>
Reviewed-by: Tiwei Bie <tiwei.bie@intel.com>
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
5 years agonet/virtio: fix mergeable Rx with segmented packet
Maxime Coquelin [Wed, 5 Jun 2019 10:00:37 +0000 (12:00 +0200)]
net/virtio: fix mergeable Rx with segmented packet

After having dequeued a burst of descriptors, there may be a
need to dequeue a few more if the last packet was segmented
and not complete. When it happens, the extra segments were
not properly attached to the mbuf chain, and so were lost.

Also, head segment data_len field is wrongly summed with
the length of all the segments of the chain.

This patch fixes both the mbuf chaining and head segment's
data_len field

Fixes: bcac5aa207f8 ("net/virtio: improve batching in mergeable path")
Cc: stable@dpdk.org
Reported-by: Yaroslav Brustinov <ybrustin@cisco.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Tiwei Bie <tiwei.bie@intel.com>
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
5 years agonet/virtio: fix in-order Rx with segmented packet
Maxime Coquelin [Wed, 5 Jun 2019 10:00:36 +0000 (12:00 +0200)]
net/virtio: fix in-order Rx with segmented packet

After having dequeued a burst of descriptors, there may be a
need to dequeue a few more if the last packet was segmented
and not complete. When it happens, the extra segments were
not properly attached to the mbuf chain, and so were lost.

Also, head segment data_len field is wrongly summed with
the length of all the segments of the chain.

This patch fixes both the mbuf chaining and head segment's
data_len field.

Fixes: e5f456a98d3c ("net/virtio: support in-order Rx and Tx")
Cc: stable@dpdk.org
Reported-by: Yaroslav Brustinov <ybrustin@cisco.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Tiwei Bie <tiwei.bie@intel.com>
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>