latency: added new library for latency stats

Add a library designed to calculate latency statistics and report them
to the application when queried. The library measures minimum, average and
maximum latencies, and jitter in nano seconds. The current implementation
supports global latency stats, i.e. per application stats.

Signed-off-by: Reshma Pattan <reshma.pattan@intel.com>
Signed-off-by: Remy Horton <remy.horton@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>

app/testpmd: add bitrate statistics calculation

Calculate bitrate statistics using the bitrate stats library. The
resulting statistics can be viewed via proc_info.

Signed-off-by: Remy Horton <remy.horton@intel.com>

bitrate: add bitrate statistics library

This patch adds a library that calculates peak and average data-rate
statistics. For ethernet devices. These statistics are reported using
the metrics library.

Signed-off-by: Remy Horton <remy.horton@intel.com>

app/proc_info: add metrics displaying

Modify the dpdk-procinfo process to display the newly added metrics.
Added new command line option "--metrics" to display metrics.

Signed-off-by: Reshma Pattan <reshma.pattan@intel.com>
Signed-off-by: Remy Horton <remy.horton@intel.com>

metrics: add information metrics library

This patch adds a new information metrics library. This Metrics
library implements a mechanism by which producers can publish
numeric information for later querying by consumers. Metrics
themselves are statistics that are not generated by PMDs, and
hence are not reported via ethdev extended statistics.

Metric information is populated using a push model, where
producers update the values contained within the metric
library by calling an update function on the relevant metrics.
Consumers receive metric information by querying the central
metric data, which is held in shared memory.

Signed-off-by: Remy Horton <remy.horton@intel.com>

doc: add missing step in PVP reference benchmark howto

After we configure QEMU we also need to call 'make' to build the binary.
This step is currently missing in the description and this patch fixes
it.

Signed-off-by: Jens Freimann <jfreiman@redhat.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: remove old deprecation notices

The ring and distributor reworks are done.

Fixes: a6619414e0a9 ("ring: make struct and macros type agnostic")
Fixes: 775003ad2f96 ("distributor: add new burst-capable library")

Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>

app/testpmd: add command to show/clear VF stats

Add two commands to show/clear VF stats
show vf stats <port_id> <vf_id>
clear vf stats <port_id> <vf_id>

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

net/i40e: use dynamic log type for control logs

This is an example of how a dynamic log type can be used in a
PMD.

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

app/testpmd: new command to dump log types

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

app/test: new command to dump log types

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

eal: deprecate log functions

Deprecate the following functions:
- rte_set_log_level(), replaced by rte_log_set_global_level()
- rte_get_log_level(), replaced by rte_log_get_global_level()
- rte_set_log_type(), replaced by rte_log_set_level()
- rte_get_log_type(), replaced by rte_log_get_level()

The new functions provide a better control of the per-type log level,
and have a better name prefix (rte_log_).

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

eal: change specific log levels at startup

Example of use:
  ./app/test-pmd --log-level='pmd\.i40e.*,8'

  This enables debug logs for all dynamic logs whose type starts with
  'pmd.i40e'.

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

eal: change several log levels matching a regexp

Introduce a function to set the log level of several log types that
match a regular expression.

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

eal: dump registered log types

Introduce a function to dump the global level and the registered log
types.

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

eal: support dynamic log types

Introduce 2 new functions to support dynamic log types:

- rte_log_register(): register a log name, and return a log type id
- rte_log_set_level(): set the log level of a given log type

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

mbuf: bump library version

The reorganization of the mbuf structure induces an ABI breakage.
Bump the library version, and update the documentation accordingly.

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

mbuf: add a timestamp field

The field itself is not fully described yet, but this commit reserves
the room in the mbuf.

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

mbuf: move sequence number in second cache line

Move this field in the second cache line, since no driver use it
in Rx path. The freed space will be used by a timestamp in next
commit.

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

mbuf: use 2 bytes for port and nb segments

Change the size of m->port and m->nb_segs to 16 bits. It is now possible
to reference a port identifier larger than 256 and have a mbuf chain
larger than 256 segments.

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

mbuf: make rearm data address naturally aligned

To avoid multiple stores on fast path, Ethernet drivers
aggregate the writes to data_off, refcnt, nb_segs and port
to an uint64_t data and write the data in one shot
with uint64_t* at &mbuf->rearm_data address.

Some of the non-IA platforms have store operation overhead
if the store address is not naturally aligned.This patch
fixes the performance issue on those targets.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>

drivers/net: do not touch mbuf next or nb segs on Rx

Now that the m->next pointer and m->nb_segs is expected to be set (to
NULL and 1 respectively) after a mempool_get(), we can avoid to write them
in the Rx functions of drivers.

Only some drivers are patched, it's not an exhaustive patch. It gives
the idea to do the same in other drivers.

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

mbuf: set mbuf fields while in pool

Set the value of m->refcnt to 1, m->nb_segs to 1 and m->next
to NULL when the mbuf is stored inside the mempool (unused).
This is done in rte_pktmbuf_prefree_seg(), before freeing or
recycling a mbuf.

Before this patch, the value of m->refcnt was expected to be 0
while in pool.

The objectives are:

- to avoid drivers to set m->next to NULL in the early Rx path, since
  this field is in the second 64B of the mbuf and its access could
  trigger a cache miss

- rationalize the behavior of raw_alloc/raw_free: one is now the
  symmetric of the other, and refcnt is never changed in these functions.

To optimize the freeing of the segments, we try try to only update
m->refcnt, m->next, and m->nb_segs when it's required (idea from
Konstantin Ananyev <konstantin.ananyev@intel.com>).

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

mbuf: make raw free function public

Rename __rte_mbuf_raw_free() as rte_mbuf_raw_free() and make
it public. The old function is kept for compat but is marked as
deprecated.

The next commit changes the behavior of rte_mbuf_raw_free() to
make it more consistent with rte_mbuf_raw_alloc().

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

mbuf: make segment prefree function public

Document the function and make it public, since it is used at several
places in the drivers. The old one is marked as deprecated.

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

doc: add Cavium OCTEONTX eventdev PMD to release notes

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

doc: add OCTEONTX ssovf details

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>

test/eventdev: add remaining tests based on existing helpers

Add the following tests based existing helper functions
- Queue based producer-consumer ingress order test
- Run existing queue and flow based ordering test in dequeue timeout
  mode

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx producer-consumer based order

Add flow based producer-consumer based ingress order test

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx queue and flow based max stage

Add queue and flow based pipeline test with maximum number of
stages available in the device.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx queue based max stage

Add queue based pipeline test with maximum number of stages available
in the device.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx flow based max stage

Add flow based pipeline test with maximum number of stages available
in the device.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx queue based two stage sched

Add queue based two stage pipeline test with all combination
of schedule types.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx flow based two stage sched

Add flow based two stage pipeline test with all combination
of schedule types.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx multi link establishment

Add unit test case to verify queue to port multi link
establishment operation.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx single link establishment

Add test case to verify queue to port single link establishment operation.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx multi queue and multi port

Add unit test case to verify multi queue enqueue and multi core/port
dequeue operation.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx priority test

Added unit test case to verify the priority associated with
each event queue available in the device.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx multi queue enq/deq tests

Added unit test case to verify enqueue and dequeue operations
with multiple queues and a single port.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx simple enq/deq tests

Added unit test case to verify simple event enqueue and dequeue
operation with different schedule types

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

test/eventdev: add octeontx unit test infrastructure

add test setup and teardown routines.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

event/octeontx: add stop and close functions

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: add start function

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: support worker dequeue

If device is configured with RTE_EVENT_DEV_CFG_PER_DEQUEUE_TIMEOUT
configuration then use different fast path dequeue handler to wait till
requested amount of nanosecond if the event is not available.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>

event/octeontx: support worker enqueue

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: add SSO HW device operations

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: add dump function for easier debugging

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: support dequeue timeout tick conversion

SSO co-processor runs at a different frequency than core clock.
Request PF to convert the ns to SSO get_work timeout period.
On dequeue, If device is configured with
RTE_EVENT_DEV_CFG_PER_DEQUEUE_TIMEOUT configuration then
use different fast path dequeue handler to wait till requested
amount of nanosecond if the event is not available.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: support linking queues to ports

queues to port link and unlink establishment is through
setting/resetting the queue/group membership in
SSOW_VHWS_GRPMSK_CHGX

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: support event ports

Add in the data-structures for the ports used by workers
to sent events to/from the HW scheduler. Also add a
function to release the resource allocated in setup

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: support event queues

Pretty much everything done in HW. Need to configure
the priority associated with event queue aka sso group
through a mailbox request to PF

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: add configure function

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: add device capabilities function

Add the info_get function to return details on the queues, flow,
prioritization capabilities, etc. which this device has.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: add octeontx eventdev driver

This adds the minimal changes to allow a octeontx eventdev
implementation to be compiled, linked and created at run time.
The eventdev does nothing, but can be created via vdev
on command line, e.g.

sudo ./build/app/test -c 0xff00 --vdev=event_octeontx
...
Initializing event_octeontx domain=4 max_queues=64 max_ports=32
RTE>>

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: add mailbox support

ssovf VF device has mailbox mechanism to communicate
with PF device. This patch adds support for a mbox API to
send the mailbox request to PF device.
The ssovf VF device will be used as the communication channel
to talk to PF devices of all the network accelerated
co-processors in Octeontx. Exposing as shared function to
use it from pool, crypto, network devices.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>

event/octeontx: add vdev interface functions

ssovf and ssowvf PCIe VF devices are shared
between eventdev PMD and ethdev PMD. This patch
expose a set of interface API to get info
about probed ssovf and ssowvf VF resources to use
with eventdev and ethdev vdev devices latter.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: probe ssowvf pcie devices

An event device consists of event queues and event ports.
On Octeontx HW, each event queues(sso group/ssovf) and
event ports(sso hws/ssowvf) are enumerated as separate
SRIOV VF PCIe device. In order to expose as an event device,
On PCIe probe, the driver stores the information associated
with the PCIe device and later with vdev infrastructure
creates event device with earlier probed PCIe VF devices.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: probe ssovf pcie devices

An event device consists of event queues and event ports.
On Octeontx HW, each event queues(sso group/ssovf) and
event ports(sso hws/ssowvf) are enumerated as separate
SRIOV VF PCIe device. In order to expose as an event device,
On PCIe probe, the driver stores the information associated
with the PCIe device and later with vdev infrastructure
creates event device with earlier probed PCIe VF devices.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

event/octeontx: add build and log infrastructure

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Gage Eads <gage.eads@intel.com>

maintainers: add eventdev section and claim SW PMD

Add a section for the eventdev PMDs, and note the next-tree.
Claim maintainership of the software eventdev PMD.

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

doc: add SW eventdev PMD to release notes

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: David Hunt <david.hunt@intel.com>

doc: add event device and software eventdev

This commit adds a section to the docs listing the event
device PMDs available.

It then adds the software eventdev PMD to the listed event
devices.

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

test/eventdev: add SW deadlock tests

This commit adds the worker loopback test to verify
that the deadlock avoidance scheme is functioning, and
a holb (head-of-line-blocking) test to ensure the head
of line blocking avoidance is correct.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: David Hunt <david.hunt@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

test/eventdev: add SW xstats tests

This commit introduces xstats tests for statistics
and reset functionality.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: David Hunt <david.hunt@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

test/eventdev: add SW tests for load balancing

This commit adds various tests for load-balancing and
queue prioritization.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: David Hunt <david.hunt@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

test/eventdev: add basic SW tests

This commit adds basic enqueue and dequeue unit tests,
some negative invalid tests, and configuration.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: David Hunt <david.hunt@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

test/eventdev: add SW test infrastructure

Add the test infrastructure, create and destroy the test
instance.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: David Hunt <david.hunt@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

event/sw: support xstats

Add support for xstats to report out on the state of the eventdev.
Useful for debugging and for unit tests, as well as observability
at runtime and performance tuning of apps to work well with the
scheduler.

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

event/sw: add dump function for easier debugging

Segfault issue resolved when only partially configured and
rte_event_dev_dump() is called before start(),

Reported-by: Vipin Varghese <vipin.varghese@intel.com>
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: David Hunt <david.hunt@intel.com>

event/sw: add start stop and close functions

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

event/sw: add scheduling logic

Add in the scheduling function which takes the events from the
producer queues and buffers them before scheduling them to consumer
queues. The scheduling logic includes support for atomic, reordered,
and parallel scheduling of flows.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: Gage Eads <gage.eads@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: David Hunt <david.hunt@intel.com>

event/sw: add worker core functions

add the event enqueue, dequeue and release functions to the eventdev.
These also include tracking of stats for observability in the load of
the scheduler.
Internally in the enqueue function, the various types of enqueue
operations, to forward an existing event, to send a new event, to
drop a previous event, are converted to a series of flags which will
be used by the scheduler code to perform the needed actions for that
event.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: Gage Eads <gage.eads@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

event/sw: support linking queues to ports

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

event/sw: support event ports

Add in the data-structures for the ports used by workers to send
packets to/from the scheduler. Also add in the functions to
create/destroy those ports.

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

event/sw: support event queues

Add in the data structures for the event queues, and the eventdev
functions to create and destroy those queues.

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

event/sw: return default port/queue config

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

event/sw: add configure function

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

event/sw: add device capabilities function

Add in the info_get function to return details on the queues, flow,
prioritization capabilities, etc. that this device has.

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

event/sw: add new software-only eventdev driver

This adds the minimal changes to allow a SW eventdev implementation to
be compiled, linked and created at run time. The eventdev does nothing,
but can be created via vdev on commandline, e.g.

  sudo ./x86_64-native-linuxapp-gcc/app/test --vdev=event_sw0
  ...
  PMD: Creating eventdev sw device event_sw0, numa_node=0, sched_quanta=128
  RTE>>

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

test/eventdev: pass timeout ticks unsupported

This commit reworks the return value handling of the
timeout ticks test. This feature is not mandatory for
a pmd, the eventdev layer returns -ENOTSUP if the PMD
doesn't implement the function.

The test is modified to check if the return value is
-ENOTSUP, and return -ENOTSUP to the test framework,
which can handle "unsupported" tests since patch[1].

As such, this test will function correctly if the
patchset linked below is applied, it fails if the
patch is not applied and the PMD doesn't the timeout
ticks function.

Note it does not depend (as a compile time dependency)
on the patchset linked below.

[1] http://dpdk.org/dev/patchwork/patch/21979/

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

eventdev: improve docs of start function

This commit documents two error return values for the
rte_event_dev_start() function.

-ESTALE  indicates not all ports are configured
-ENOLINK indicates that not all queues are linked to ports. If an
         application enqueues to such a queue it can lead to deadlock

Suggested-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>

doc: add eventdev library to release notes

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

eventdev: add errno-style return values

This commit adds rte_errno return values to rte_event_enqueue_burst() and
rte_event_dequeue_burst().

These return values allows user software to differentiate between an
invalid argument (such as an invalid queue_id or sched_type in an enqueued
event) and backpressure from the event device.

The port and device ID checks are placed in RTE_LIBRTE_EVENTDEV_DEBUG
header guards to avoid the performance hit in non-debug execution.

Signed-off-by: Gage Eads <gage.eads@intel.com>

eventdev: add extended stats

Add in APIs for extended stats so that eventdev implementations can report
out information on their internal state. The APIs are based on, but not
identical to, the equivalent ethdev functions.

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

test/eventdev: link all queues before start

The software eventdev can lock-up if not all queues are
linked to a port. For this reason, the software evendev
fails to start if queues are not linked to anything.

This commit creates dummy links from all queues to port
0 in the eventdev setup function and start/stop test,
which would otherwise fail due to unlinked queues.

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

eventdev: remove default queue overriding

PMDs that only do a specific type of scheduling cannot provide
CFG_ALL_TYPES, so the Eventdev infrastructure should not demand
that every PMD supports CFG_ALL_TYPES.

By not overriding the default configuration of the queue as
suggested by the PMD, the eventdev_common unit tests can pass
on all PMDs, regardless of their capabilities.

RTE_EVENT_QUEUE_CFG_DEFAULT is no longer used by the eventdev layer
it can be removed now. Applications should use CFG_ALL_TYPES
if they require enqueue of all types a queue, or specify which
type of queue they require.

The CFG_DEFAULT value is changed to CFG_ALL_TYPES in event/skeleton,
to not break the compile.

A capability flag is added that indicates if the underlying PMD
supports creating queues of ALL_TYPES.

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

test/eventdev: fix reconfigure values

Minimum value of nb_event_ports and/or nb_event_queues
should be one before reconfiguring the event device.

Fixes: f8f9d233ea0e ("test/eventdev: add unit tests")

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

eventdev: return code in dequeue timeout conversion

eventdev driver may return error on dequeue timeout tick conversion.
Change the pmd callback interface to address the same.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

eventdev: fix links map initialization for SW PMD

This patch initializes the links_map array entries to
EVENT_QUEUE_SERVICE_PRIORITY_INVALID, as expected by
rte_event_port_links_get(). This is necessary for the sw eventdev PMD,
which does not initialize links_map when rte_event_port_setup() calls
rte_event_port_unlink().

Fixes: 4f0804bbdfb9 ("eventdev: implement the northbound APIs")

Signed-off-by: Gage Eads <gage.eads@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

eventdev: use generic device holder

rte_device is a generic device which is available to the applications
and EAL. This patch replaces rte_pci_device in 'struct rte_eventdev'
and in 'struct rte_event_dev_info' with common rte_device.

Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Shreyansh Jain <shreyansh.jain@nxp.com>

eventdev: improve API doc for timeout ticks

Improve the documentation of the return values of the
rte_event_dequeue_timeout_ticks() function, adding a
-ENOTSUP value for eventdevs that do not support waiting.

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

eventdev: increase size of enq/deq conf variables

Large port enqueue sizes were not supported as the value
it was stored in was a uint8_t. Using uint8_ts to save
space in config apis makes no sense - increasing the 3
instances of uint8_t enqueue / dequeue depths to more
appropriate values (based on the context around them).

Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

eventdev: amend timeout criteria comment for burst dequeue

Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

eventdev: clarify some parameter descriptions

This commit clarifies the usage of nb_links and nb_unlinks when passing
a NULL pointer as the queues argument.

Signed-off-by: Gage Eads <gage.eads@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>

eventdev: amend comments for events limit and threshold

Updated the comments on 'nb_events_limit' of 'struct rte_event_dev_config'
and 'new_event_threshold' of 'struct rte_event_port_conf'.

Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

eventdev: limit port link operation to configured queues

On port_setup, the link_map is updated only
for configured number of event queues.
Limit the port_links_get scan only to configured number
of event queues. Also, Limit the port link and unlink queue
validation to configured number of event queues.

Fixes: 4f0804bbdfb9 ("eventdev: implement the northbound APIs")

Reported-by: Nipun Gupta <nipun.gupta@nxp.com>
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Nipun Gupta <nipun.gupta@nxp.com>

event/skeleton: support vdev uninit

Removed global index based device name
generation as vdev uninit needs the exact driver
name used vdev init.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

eventdev: support vdev uninit

Added eventdev vdev uninit support to release the resources
allocated in eventdev vdev init.

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

eventdev: fix event driver name to eventdev lookup

- Removed uninitialized max_devs value
- Corrected dev assignment

Fixes: 4f0804bbdfb9 ("eventdev: implement the northbound APIs")

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>

eventdev: remove unneeded dependencies

Since eventdev uses event structures rather than working directly on
mbufs, there is no actual dependencies on the mbuf library. The
inclusion of an mbuf pointer element inside the event itself does not
require the inclusion of the mbuf header file. Similarly the pci
header is not needed, but following their removal, rte_memory.h is
needed for the definition of the __rte_cache_aligned macro.

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