doc: reduce features tables column width

The font size of the header row is reduced in order to shrink
the column size of the tables.
It is required for the NICs features table which is too large to fit
in the page width.

Cc: stable@dpdk.org
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>

maintainers: resign from GSO lib

I will not be directly working on the DPDK project anymore.

Signed-off-by: Mark Kavanagh <mark.b.kavanagh@intel.com>
Acked-by: Helin Zhang <helin.zhang@intel.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>

maintainers: resign from vhost/virtio

Signed-off-by: Yuanhan Liu <yliu@fridaylinux.org>

maintainers: update for vhost lib and PMD

Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Acked-by: Zhiyong Yang <zhiyong.yang@intel.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
Acked-by: Maxime Coquelin <maxime.coquelin@redhat.com>

drivers: rename bbdev directory to baseband

The drivers directory contains some sub-directories
for each kind of device (or bus, mem):
net, crypto, event, raw
They are not suffixed with "dev" because it is obvious.

For consistency, the sub-directory drivers/bbdev/
is renamed to drivers/baseband/.

Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Amr Mokhtar <amr.mokhtar@intel.com>

vfio: fix headers for C++ support

Fixes: 279b581c897d ("vfio: expose functions")
Cc: stable@dpdk.org
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>

eal: ignore IPC messages until init is complete

If we receive messages that don't have a callback registered for
them, and we haven't finished initialization yet, it can be reasonably
inferred that we shouldn't have gotten the message in the first
place. Therefore, send requester a special message telling them to
ignore response to this request, as if this process wasn't there.

Since it is not possible for primary process to receive any messages
during initialization, this change in practice only applies to
secondary processes.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: simplify IPC sync request timeout

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: lock IPC directory on init and send

When sending IPC messages, prevent new sockets from initializing.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: do not hardcode socket filter value in IPC

Currently, filter value is hardcoded and disconnected from actual
value returned by eal_mp_socket_path(). Fix this to generate filter
value by deriving it from eal_mp_socket_path() instead.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: abstract away IPC socket path generation

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: add internal flag of init completed

Currently, primary process initialization is finalized by setting
the RTE_MAGIC value in the shared config. However, it is not
possible to check whether secondary process initialization has
completed. Add such a value to internal config.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>

eal: fix race condition in IPC request

Unlocking the action list before sending message and locking it
again afterwards introduces a window where a response might
arrive before we have a chance to start waiting on a condition,
resulting in timeouts on valid messages.

Fixes: 783b6e54971d ("eal: add synchronous multi-process communication")
Cc: stable@dpdk.org
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: fix errno handling in IPC

Fixes: bacaa2754017 ("eal: add channel for multi-process communication")
Cc: stable@dpdk.org
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: fix IPC request socket path

Fixes: 783b6e54971d ("eal: add synchronous multi-process communication")
Cc: stable@dpdk.org
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: fix IPC socket path

Fixes: bacaa2754017 ("eal: add channel for multi-process communication")
Cc: stable@dpdk.org
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

eal: fix IPC timeout

Fixes: 783b6e54971d ("eal: add synchronous multi-process communication")
Cc: stable@dpdk.org
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>

build: add meson cross-compilation config for dpaaX

Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

build: add meson support for dpaaX platforms

Signed-off-by: Akhil Goyal <akhil.goyal@nxp.com>
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

config: add dpaaX build for all Linux platforms

Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

bus/fslmc: enable compilation for other platforms

Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

bus/dpaa: enable compilation for other platforms

Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

bus/fslmc: add 32-bit functional support for ARM

This patch adds the functional logic to make the dpaa2 drivers
work on 32bit machine.

Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

dpaa2: prepare for 32-bit build

This patch prepare the dpaa2 drivers for compilation on 32 bit machine.

Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

dpaa: prepare for 32-bit build

This patch prepares the dpaa drivers for compilation on 32 bit machine.

Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

bus/dpaa: fix big endian build

The array pointers were used without index.

Fixes: b9083ea5e084 ("net/dpaa: further push mode optimizations")
Cc: stable@dpdk.org
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Shreyansh Jain <shreyansh.jain@nxp.com>

event/dpaa: fix header include

rte_cycles.h shall be included instead of rte_cycles_64.h

dpaa_eventdev.c:32:27:
fatal error: rte_cycles_64.h: No such file or directory

Fixes: 9caac5dd1e7f ("event/dpaa: introduce PMD")
Cc: stable@dpdk.org
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

net/pcap: simplify dependency checking using meson

Rather than trying to use meson's build-in detection for libpcap, and
having to special-case cross-building, just check for the presence of
pcap.h and the pcap library.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Tested-by: Hemant Agrawal <hemant.agrawal@nxp.com>

net/virtio: add to meson build

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Maxime Coquelin <maxime.coquelin@redhat.com>

app: add all remaining apps to meson build

Add remaining subdirectories in the app folder to the meson build.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

app/procinfo: rename folder to remove underscore

All other apps in the app folder use "-" rather than "_" to separate words
in the app name, so rename proc_info to be consistent.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

app: generalize meson build

Since most apps are built in largely the same way, generalize the logic
into a foreach loop in app/meson.build file.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

mk: fix dependencies of dpaaX drivers

This  patch fixes the build dependency of various
dpaaX components, when the dpaa or fslmc bus is disabled,
or VFIO is disabled.

Fixes: 1ee9569576f6 ("config: enable dpaaX drivers for generic ARMv8")
Cc: stable@dpdk.org
Reported-by: Yongseok Koh <yskoh@mellanox.com>
Suggested-by: Thomas Monjalon <thomas@monjalon.net>
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

app/testpmd: fix DPAA shared library dependency

The dynamic link is broken for ARM platform because the dependencies
of the DPAA PMD are not declared.

Fixes: 83c82e15e1c0 ("app/testpmd: support loopback config for DPAA")
Cc: stable@dpdk.org
Reported-by: Marco Varlese <mvarlese@suse.de>
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>

eal/ppc: remove braces in SMP memory barrier macro

This patch fixes the compilation problem with rte_smp_mb,
when there is else clause following it, as in test_barrier.c.

Fixes: 05c3fd7110 ("eal/ppc: atomic operations for IBM Power")
Cc: stable@dpdk.org
Signed-off-by: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
Acked-by: Luca Boccassi <bluca@debian.org>
Acked-by: Chao Zhu <chaozhu@linux.vnet.ibm.com>

examples: fix build with meson

The ip_pipeline and qos_meter example apps now use experimental APIs so
this fact needs to be flagged in their meson.build files.

Fixes: c06ddf9698e0 ("meter: add configuration profile")

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>

meter: add configuration profile

This patch adds support for meter configuration profiles.
Benefits: simplified configuration procedure, improved performance.

Q1: What is the configuration profile and why does it make sense?
A1: The configuration profile represents the set of configuration
    parameters for a given meter object, such as the rates and sizes for
    the token buckets. The configuration profile concept makes sense when
    many meter objects share the same configuration, which is the typical
    usage model: thousands of traffic flows are each individually metered
    according to just a few service levels (i.e. profiles).

Q2: How is the configuration profile improving the performance?
A2: The performance improvement is achieved by reducing the memory
    footprint of a meter object, which results in better cache utilization
    for the typical case when large arrays of meter objects are used. The
    internal data structures stored for each meter object contain:
       a) Constant fields: Low level translation of the configuration
          parameters that does not change post-configuration. This is
          really duplicated for all meters that use the same
          configuration. This is the configuration profile data that is
          moved away from the meter object. Current size (implementation
          dependent): srTCM = 32 bytes, trTCM = 32 bytes.
       b) Variable fields: Time stamps and running counters that change
          during the on-going traffic metering process. Current size
          (implementation dependent): srTCM = 24 bytes, trTCM = 32 bytes.
          Therefore, by moving the constant fields to a separate profile
          data structure shared by all the meters with the same
          configuration, the size of the meter object is reduced by ~50%.

Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Jasvinder Singh <jasvinder.singh@intel.com>

version: 18.05-rc0

Signed-off-by: Thomas Monjalon <thomas@monjalon.net>

doc: add template release notes for 18.05

Add template release notes for DPDK 18.05 with inline
comments and explanations of the various sections.

Signed-off-by: John McNamara <john.mcnamara@intel.com>
Acked-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com>

version: 18.02.0

Signed-off-by: Thomas Monjalon <thomas@monjalon.net>

doc: update release notes for 18.02

Fix grammar, spelling and formatting of DPDK 18.02 release notes.

Signed-off-by: John McNamara <john.mcnamara@intel.com>

doc: add virtio guest announce support to release notes

Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: add tested platforms with Mellanox NICs

Signed-off-by: Raslan Darawsheh <rasland@mellanox.com>

doc: fix outdated link to IPsec white paper

Fixes: 924e84f87306 ("aesni_mb: add driver for multi buffer based crypto")
Cc: stable@dpdk.org
Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>

doc: remove reference to old distro in Linux guide

Remove reference to Fedora 18 which is EOL-ed, reword
surrounding sentences to read correctly.

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: announce ethdev API fix for port id size

Fix rte_eth_dev_get_sec_ctx() parameter port_id storage size,
from uint8_t to uint16_t.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

doc: announce ethdev ABI change for filter structs

Signed-off-by: Kirill Rybalchenko <kirill.rybalchenko@intel.com>
Acked-by: Marko Kovacevic <marko.kovacevic@intel.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

doc: announce ethdev API changes for callbacks

Ethdev APIs to add callback return the callback object as "void *",
update return type to actual object type
"struct rte_eth_rxtx_callback *"

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>

doc: announce ethdev API change for default MAC

Announce internal PMD API change in the function to set the default MAC
address. The objective is to be able to notify errors occurring in the
PMD.

Link: https://dpdk.org/dev/patchwork/patch/32284/
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

doc: announce ethdev API change for VF representors

This is following the RFC being discussed and targets 18.05

http://dpdk.org/ml/archives/dev/2018-January/085716.html

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Declan Doherty <declan.doherty@intel.com>
Acked-by: Remy Horton <remy.horton@intel.com>
Acked-by: Luca Boccassi <luca.boccassi@intl.att.com>
Acked-by: Alex Zelezniak <alexz@att.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

doc: announce ethdev API change for preferred burst size

rte_eth_rx_burst(..,nb_pkts) function has semantic that if return value
is smaller than requested, application can consider it end of packet
stream. Some hardware can only support smaller burst sizes which need
to be advertised. Similar is the case for Tx burst.

This patch adds deprecation notice for rte_eth_dev_info structure as
new members, for preferred Rx and Tx burst and ring size would be
added - impacting the size of the structure.

Signed-off-by: Shreyansh Jain <shreyansh.jain@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Zhiyong Yang <zhiyong.yang@intel.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

doc: announce ethdev API change for RSS configuration

Update deprecation notice for the new rss_level field of
rte_eth_rss_conf.

Link: http://www.dpdk.org/dev/patchwork/patch/31891
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

doc: announce mbuf sched field change

Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Acked-by: Jasvinder Singh <jasvinder.singh@intel.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
Acked-by: Kevin Laatz <kevin.laatz@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

doc: announce control mbuf removal

Link: http://dpdk.org/ml/archives/dev/2017-July/069813.html
Link: http://dpdk.org/dev/patchwork/patch/32041/
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

doc: announce mempool API changes

An API/ABI changes are planned for 18.05 [1]:

 * Allow to customize how mempool objects are stored in memory.
 * Deprecate mempool XMEM API.
 * Add mempool driver ops to get information from mempool driver and
   dequeue contiguous blocks of objects if driver supports it.

[1] http://dpdk.org/ml/archives/dev/2018-January/088698.html

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>

doc: announce EAL function removal for default mempool

Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

doc: announce memory hotplug changes

Due to coming changes outlined in memory hotplug RFC, there will
be several API/ABI changes.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Jonas Pfefferle <pepperjo@japf.ch>
Acked-by: Yongseok Koh <yskoh@mellanox.com>

doc: announce EAL ABI change for NUMA node count

There will be a new function added in v18.05 that will return
number of detected sockets, which will change the ABI.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Jonas Pfefferle <pepperjo@japf.ch>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

doc: announce EAL API change to lcore role function

This an API/ABI change notice for DPDK 18.05 announcing a change in
the meaning of the return values of the rte_lcore_has_role() function.

Signed-off-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

doc: update deprecation notice for devargs

The declaration and identification of devices will change in v18.05.

Remove the precedent deprecation notice.

Add new one reflecting the planned changes more accurately,
updated for v18.05.

Signed-off-by: Gaetan Rivet <gaetan.rivet@6wind.com>
Acked-by: Shreyansh Jain <shreyansh.jain@nxp.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

net/failsafe: fix Rx interrupt reinstallation

Fail-safe dev_start() operation can be called by both the application
and the hot-plug alarm mechanism.

The installation of Rx interrupt are triggered from dev_start() in any
time it is called while actually the Rx interrupt should be installed
only by the application calls.

So, each plug-in event causes reinstallation which causes memory leak
and spoils the fail-safe Rx interrupt mechanism.

Trigger the Rx interrupt installation only when it does not exist.

Fixes: 9e0360aebf23 ("net/failsafe: register as Rx interrupt mode")

Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Gaetan Rivet <gaetan.rivet@6wind.com>

net/tap: fix promiscuous rules double insertions

Running testpmd command "port stop all" followed by command "port start
all" may result in a TAP error:
PMD: Kernel refused TC filter rule creation (17): File exists

Root cause analysis: during the execution of "port start all" command
testpmd calls rte_eth_promiscuous_enable() while during the execution
of "port stop all" command testpmd does not call
rte_eth_promiscuous_disable().
As a result the TAP PMD is trying to add tc (traffic control command)
promiscuous rules to the remote netvsc device consecutively. From the
kernel point of view it is seen as an attempt to add the same rule more
than once. In recent kernels (e.g. version 4.13) this attempt is rejected
with a "File exists" error. In less recent kernels (e.g. version 4.4) the
same rule may have been successfully accepted twice, which is undesirable.

In the corrupted code every tc promiscuous rule included a different
handle number parameter. If instead an identical handle number is
used for all tc promiscuous rules - all kernels will reject the second
identical rule with a "File exists" error, which is easy to identify and
to silently ignore.

Fixes: 2bc06869cd94 ("net/tap: add remote netdevice traffic capture")
Cc: stable@dpdk.org
Signed-off-by: Ophir Munk <ophirmu@mellanox.com>
Acked-by: Pascal Mazon <pascal.mazon@6wind.com>

maintainers: update for stable branches

I am maintaining 16.11 until November 2018

Signed-off-by: Luca Boccassi <bluca@debian.org>

doc: add maintainers section to the contributors guide

Add a maintainers section to the contributors guide to have a low tech
location to check/link to the current maintainers. This file is included
dynamically from the MAINTAINERS file in the root directory of the DPDK
source when the docs are built. This also allows us to link to the file
from other sections of the docs.

Signed-off-by: Ciara Power <ciara.power@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: improve HTML spacing in release notes

Vertical spacing is lower before an item title than after.
So the items with paragraphs are not well separated.

A custom CSS is added to override the rtd theme.

Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: add vhost live migration features to release notes

Signed-off-by: Jiayu Hu <jiayu.hu@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: add VxLAN GRO to release notes

Signed-off-by: Jiayu Hu <jiayu.hu@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: add note for shared library in NICs guide

Add note information to intimate about use of option '-d' for shared
library in DPDK application.

Signed-off-by: Vipin Varghese <vipin.varghese@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

examples/cmdline: convert to SPDX license tag

This updates the license on files in examples to be the standard
BSD-3-Clause license used for the rest of DPDK,
bringing the files in compliance with the DPDK licensing policy.

Signed-off-by: Lee Daly <lee.daly@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

examples/ip_pipeline: convert dual-license to SDPX

This updates the Intel and Oliver Matz licenses on a file in examples
to be the standard BSD-3-Clause license used for the rest of DPDK,
bringing the files in compliance with the DPDK licensing policy.

Signed-off-by: Lee Daly <lee.daly@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

examples/performance-thread: update lthread license

This updates Hasan Alayli's license to be the standard BSD-3-Clause
license used for the rest of DPDK, bringing the files in compliance
with the DPDK licensing policy.

Signed-off-by: Lee Daly <lee.daly@intel.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

examples/ipsec-secgw: fix printed crypto name

When AES-256 was used aes-128 was printed in the console

Fixes: fa9088849e12 ("examples/ipsec-secgw: support AES 256")

Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Anoob Joseph <anoob.joseph@caviumnetworks.com>

examples/exception_path: align stats on cache line

Align stats structure to cache line to prevent bouncing per CPU stats
structure between cache lines.

Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org
Signed-off-by: Dustin Lundquist <dustin@null-ptr.net>
Acked-by: Remy Horton <remy.horton@intel.com>

net/failsafe: fix FreeBSD build

The type pthread_t is not portable because it is freely defined.
On Linux, it is an unsigned long int which can be printed with %l.
On FreeBSD, it is a pointer which can be printed with %p.

That's why there was this error:
drivers/net/failsafe/failsafe_private.h:377:53: error:
format specifies type 'unsigned long' but the argument has
type 'pthread_t' (aka 'struct pthread *')

Fixes: 655fcd68c7d2 ("net/failsafe: fix hotplug races")

Reported-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Tested-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Acked-by: Matan Azrad <matan@mellanox.com>

net/virtio-user: fix interrupts with kernel vhost

intr_handle->fd was wrongly initialized as 0 (usually as the stdio fd)
when virtio-user is used with vhost-kernel. So the interrupt thread
might wrongly treat stdin events as LSC interrupts.

Fixes: 3d4fb6fd2505 ("net/virtio-user: support Rx interrupt")
Cc: stable@dpdk.org
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

net/virtio-user: fix start with kernel vhost

After reset owner in below patch, we failed to set owner before
sending further vhost messages. It is OK with vhost user implemented
DPDK/VPP/Contrail, but it sees "Operation not permitted" error when
used with vhost kernel.

We fix this by setting owner every time the device is started.

Fixes: 0d6a8752ac9d ("net/virtio-user: fix crash as features change")
Cc: stable@dpdk.org
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

net/vhost: fix log messages on create/destroy

The original words are not accurate. For example, as destroy_device
callback gets called, it does not necessarily mean that the connection
is closed.

Fixes: ee584e9710b9 ("vhost: add driver on top of the library")
Cc: stable@dpdk.org
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

vhost: do not take lock on owner reset

A deadlock happens when handling VHOST_USER_RESET_OWNER request
for the same reason the lock is not taken for
VHOST_USER_GET_VRING_BASE.

It is safe not to take the lock, as the queues are no more used
by the application when the virtqueues and the device are reset.

Fixes: a3688046995f ("vhost: protect active rings from async ring changes")
Cc: stable@dpdk.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Tiwei Bie <tiwei.bie@intel.com>
Reviewed-by: Jianfeng Tan <jianfeng.tan@intel.com>

net/virtio: fix resuming port with Rx vector path

Since commit efc83a1e7fc3 ("net/virtio: fix queue setup consistency"),
when resuming a virtio port, the rx rings are refilled with new mbufs
until they are full (vq->vq_free_cnt == 0). This is done without
ensuring that the descriptor index remains a multiple of
RTE_VIRTIO_VPMD_RX_REARM_THRESH, which is a prerequisite when using the
vector mode. This can cause an out of bound access in the rx ring.

This commit changes the vector refill method from
virtqueue_enqueue_recv_refill_simple() to virtio_rxq_rearm_vec(), which
properly checks that the refill is done by batch of
RTE_VIRTIO_VPMD_RX_REARM_THRESH.

As virtqueue_enqueue_recv_refill_simple() is no more used, this
patch also removes the function.

Fixes: efc83a1e7fc3 ("net/virtio: fix queue setup consistency")
Cc: stable@dpdk.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Reviewed-by: Jianfeng Tan <jianfeng.tan@intel.com>

net/virtio: fix mbuf data offset for simple Rx

The mbuf->data_off was was not properly set for the first received
mbufs. Fix this by setting it in virtqueue_enqueue_recv_refill_simple(),
which is used to enqueue the first mbuf in the ring.

The function virtio_rxq_rearm_vec(), which is used to rearm the ring
with new mbufs, is valid and does not need to be updated.

Fixes: cab0461234e7 ("virtio: fill Rx avail ring with blank mbufs")
Cc: stable@dpdk.org
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

net/failsafe: fix reconfiguration

Fail-safe PMD manages the states of its sub-devices gradually:
DEV_UNDEFINED, DEV_PARSED, DEV_PROBED, DEV_ACTIVE, DEV_STARTED.

When the sub-device arguments successfully was parsed, the state is
raised from DEV_UNDEFINED to DEV_PARSED.
When the sub-device successfully was probed, the state is raised from
DEV_PARSED to DEV_PROBED.
When the sub-device successfully was configured by
rte_eth_dev_configure(), the state is raised from DEV_PROBED to
DEV_ACTIVE.
When the sub-device successfully was started by rte_eth_dev_start(), the
state is raised from DEV_ACTIVE to DEV_STARTED.

When the sub-device successfully was stopped by rte_eth_dev_stop(), the
state is degraded from DEV_STARTED to DEV_ACTIVE.
When the sub-device successfully was closed by rte_eth_dev_close(), the
state is degraded from DEV_ACTIVE to DEV_PROBED.
When the sub-device successfully was removed by
rte_eal_hotplug_remove(), the state is degraded from DEV_PROBED to
DEV_UNDEFINED.

Fail-safe dev_configure() operation calls to its sub-devices
dev_configure() operation, but only for sub-devices which are in
DEV_PROBED state, means that sub-devices which are in DEV_ACTIVE state
because the application triggered dev_stop() operation cannot be
reconfigured again by dev_configure() operation which is really
problematic when application wants to reconfigure its ports.
Actually, the application may get success report when some of the
sub-devices are not in the wanted configuration.

The current behavior of fail-safe dev_configure() is correct only for
the first time dev_configure() is triggered by the application or for
sub-device synchronization in plug-in event, but it ignores the option
for reconfiguration from application side.

Allow calling to sub-devices dev_configure() operations also in
DEV_ACTIVE state when the call was triggered by the application.

Fixes: a46f8d584eb8 ("net/failsafe: add fail-safe PMD")
Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Gaetan Rivet <gaetan.rivet@6wind.com>

net/failsafe: fix hotplug races

Fail-safe uses a periodic alarm mechanism, running from the host
thread, to manage the hot-plug events of its sub-devices. This
management requires a lot of sub-devices PMDs operations
(stop, close, start, configure, etc.).

While the hot-plug alarm runs in the host thread, the application may
call fail-safe operations, which directly trigger the sub-devices PMDs
operations as well. This call may occur from any thread decided by the
application (probably the master thread).

Thus, more than one operation can be executed to a sub-device at the
same time. This can initiate a lot of races in the sub-PMDs.

Moreover, some control operations update the fail-safe internal
databases, which can be used by the alarm mechanism at the same time.
This can also initiate races and crashes.

Fail-safe is the owner of its sub-devices and must synchronize their
use according to the ETHDEV ownership rules.

Synchronize hot-plug management by a new lock mechanism uses a mutex to
atomically defend each critical section in the fail-safe hot-plug
mechanism and control operations to prevent any races between them.

Fixes: a46f8d5 ("net/failsafe: add fail-safe PMD")
Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Gaetan Rivet <gaetan.rivet@6wind.com>

net/failsafe: fix removal scope

The fail-safe PMD uses a per sub-device flag called "remove" to
indicate the scope where the sub-device was removed physically and
whether its software resources should be released.

This flag is set when the fail-safe receives an RMV notification
about the physical removal of the sub-device, and should be unset when
all the sub-device resources are released.

The previous code wrongly unsets the flag in dev_configure(), instead
of when the software resources release is completed.

Change the remove flag unsetting to take action in the end of the
software resources release.

Fixes: a46f8d5 ("net/failsafe: add fail-safe PMD")
Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Gaetan Rivet <gaetan.rivet@6wind.com>

net/failsafe: fix hotplug alarm cancel

The hot-plug alarm mechanism of fail-safe PMD is responsible for
handling removed devices during a plug-out event and to restore them
back to activity following a plug-in event.

Fail-safe sets a flag called "pending_alarm" to validate that only one
alarm callback is pending at any time. While this flag is required to
avoid simultaneous initiations of the alarm thread - it should not be
considered during alarm thread cancellation.

So, when failsafe_hotplug_alarm_cancel() was called while the alarm
callback was being executed the alarm mechanism was not stopped.

Skip checking the "pending_alarm" flag to allow alarm thread
cancellation all the times.

Fixes: ebea83f899d8 ("net/failsafe: add plug-in support")
Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Gaetan Rivet <gaetan.rivet@6wind.com>

net/tap: add CRC stripping capability

CRC stripping is executed in the kernel outside of TAP PMD scope.
There is no prevention that the TAP PMD will report on Rx CRC
stripping capability.
In the corrupted code, TAP PMD did not report on this capability.
The fix enables TAP PMD to report that Rx CRC stripping is supported.

Fixes: 02f96a0a82d1 ("net/tap: add TUN/TAP device PMD")
Cc: stable@dpdk.org
Signed-off-by: Ophir Munk <ophirmu@mellanox.com>

doc: remove jumbo frames entry from tap features

The current implementation of the Tap PMD doesn't support packet size
grater than 1522B, so no need to set jumbo frames in the Tap features.

Fixes: 04b2c3c94346 ("doc: add tap features")
Cc: stable@dpdk.org
Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Pascal Mazon <pascal.mazon@6wind.com>

doc: add missing SFN8xxx adapters in sfc guide

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: update Intel VF guide

update for intel VF usage:
 - Add note for AVF device ID generation.
 - Add steps for vfio-pci pass through.

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: rename nfp PF features file

Other NICs with VF support do not use the _pf suffix for these files
but just for the VF description.

This change makes NFP files consistent in this regard and helps with
how the DPDK web nics features matrix shows NFP devices.

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

net/mlx5: add log on flow creation error

Add error message dump when flow create error happened.

Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/mlx5: revert multicast rule verbs flow type

This is to revert the following commits:
commit da646bd93888 ("net/mlx5: fix all multi verification code position")
commit 0a40a1363a4d ("net/mlx5: fix flow type for allmulti rules")

The last one introduced a bug in the following diff:
@ -1262,6 +1274,7 @@ struct ibv_spec_header {
                eth.val.ether_type &= eth.mask.ether_type;
        }
        mlx5_flow_create_copy(parser, &eth, eth_size);
+       parser->allmulti = eth.val.dst_mac[0] & 1;
        return 0;
 }

As broadcast rules will be considered of type allmulti as well.

The patch was originally intended to enable VF to receive all multicast
traffic by using the IBV_FLOW_ATTR_MC_DEFAULT flow type.
Since the support was removed from the kernel there is no point with
fixing this issue, hence the revert.

Fixes: da646bd93888 ("net/mlx5: fix all multi verification code position")
Fixes: 0a40a1363a4d ("net/mlx5: fix flow type for allmulti rules")
Cc: stable@dpdk.org
Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/mlx5: fix close after start failure

This patch fixed primary socket assertion error during close on a device
that failed to start.

Fixes: f8b9a3bad467 ("net/mlx5: install a socket to exchange a file descriptor")
Cc: stable@dpdk.org
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

net/dpaa2: remove unused global variable

Due to the missing __packed definition, compiler treating
it as as global variable.

grep "Allocating common symbols" -A 100 build/app/testacl.map
__packed   0x8  /build/lib/librte_pmd_dpaa2.a(dpaa2_rxtx.o)

Fixes: a5fc38d422a7 ("net/dpaa2: support Rx packet parsing")
Cc: stable@dpdk.org
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>

ethdev: increase log level of port allocation failure

DPDK API does not propagate the reason of device allocation failure
from rte_eth_dev_allocate() up to the DPDK application (e.g. Open
vSwitch).
Log level of associated log entries was changed to warning. So user
can find additional details in log files also in production systems,
where debug messages cannot be turned on.

Signed-off-by: Martin Klozik <martinx.klozik@intel.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

ethdev: fix data alignment

The struct rte_eth_dev_data is used in ethdev fastpath routines
and it not aligned to cache line size. This patch fixes the ethdev
data alignment.

The alignment was broken from the "first public release" changeset
where ethdev data address was aligned only to the first port.
Remaining ports alignment was defined by the size of the struct
(rte_eth_dev_data). This scheme is not guaranteed to be cache line
aligned all the time.

"ethdev: add port ownership" change set introduced a
rte_eth_dev_shared_data container for port ownership change,
This resulted in rte_eth_dev->data memory for the first port also
as cache unaligned.

Added a compiler alignment attribute to make sure
rte_eth_dev->data always cache aligned so that CPU/compiler
1) Avoid sharing the element with another cache line
2) Can load/store the elements in struct rte_eth_dev_data as
naturally aligned.

Some platform like thunderX could see performance regression of 1%
at "ethdev: add port ownership" change set with
1 port/1 queue l3fwd application and this patch fixes that regression.

example command:
sudo ./examples/l3fwd/build/l3fwd -c 0xff00 -- -p 0x1 --config="(0,0,9)"

Fixes: af75078fece3 ("first public release")
Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
Cc: stable@dpdk.org
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

eal: fix errno in IPC API

rte_errno values should not be negative.

Fixes: bacaa2754017 ("eal: add channel for multi-process communication")
Fixes: 783b6e54971d ("eal: add synchronous multi-process communication")

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Reviewed-by: Jianfeng Tan <jianfeng.tan@intel.com>

lib: remove unused map symbols

Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

maintainers: update for driver information tool

I wrote pmdinfogen initially, and since there isn't a maintainer for it,
I'll volunteer to take care of it

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

version: 18.02-rc4

Signed-off-by: Thomas Monjalon <thomas@monjalon.net>

doc: add tested Intel platforms with Intel NICs

Add tested Intel platforms with Intel NICs to the release note.

Signed-off-by: Yulong Pei <yulong.pei@intel.com>