Context memory is allocated from mempool. Ideally
it will get all memory from single segment, so simple offset
calculation is used for address conversion for such addresses
from context memory.
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>
Intel VT-d supports different address widths for the IOVAs, from
39 bits to 56 bits.
While recent processors support at least 48 bits, VT-d emulation
currently only supports 39 bits. It makes DMA mapping to fail in this
case when using VA as IOVA mode, as user-space virtual addresses uses
up to 47 bits (see kernel's Documentation/x86/x86_64/mm.txt).
This patch parses VT-d CAP register value available in sysfs, and
forbid VA as IOVA mode if the GAW is 39 bits or unknown.
Fixes: f37dfab21c98 ("drivers/net: enable IOVA mode for Intel PMDs")
Cc: stable@dpdk.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Tested-by: Chas Williams <chas3@att.com>
In the info get API we were returning number of event
ports as MAX number of LCORE's. After this change actual
number of event ports (i.e. number of DPIO's) is provided
in the info get API.
Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Given ethernet Rx queues can be attached with event queue in
parallel or atomic mode. Patch imlmplements Rx queue
configuration, attachment/detachment with given event queue and their
corresponding callbacks to handle events from respective queues.
Signed-off-by: Sunil Kumar Kori <sunil.kori@nxp.com>
Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
To receive events from given event port, corresponding
function needs to be added which receives events
from portal. Also added function to consume received
events based on entry index.
Signed-off-by: Sunil Kumar Kori <sunil.kori@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
With the current QBMAN multi-tx API, we need to create separate
enqueue descriptors for each of the packet which is required to
be enqueued to the hardware, once we support Atomic Queues
(with DCA) in dpaa2 drivers. Creating enqueue descriptor for
each packet is costly and have significant performance impact.
This patch introduces a flag parameter in the QBMAN multi-tx API,
so that DCA configuration (and later on ORP/ODP for ordered queues)
can be passed using flags and be updated in the EQCR using this flag.
Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
This patch provides cleaner approach to store the DQRR entries,
which are yet to be consumed in case of atomic queues.
Also, this patch changes the storage of the DQRR entry index
into the mbuf->seqn instead of ev->opaque
Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
A new API qbman_swp_dqrr_idx_consume is defined which takes
input as DQRR index to consume corresponding DQRR entry.
Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Update software eventdev ops to invoke selftest when application invokes
`rte_event_dev_selftest`.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
Modify test_eventdev_sw to be standalone selftest independent of test
framework.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
Move software eventdev specific test (test_eventdev_sw) to
driver/event/sw/.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
Add selftest as a device argument that can be enabled by suppling
'self_test=1' as a vdev parameter
--vdev="event_octeontx,self_test=1"
The selftest is run after vdev creation is successfully
complete.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Modify test_eventdev_octeontx to be standalone selftest independent of
test framework.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Move octeontx eventdev specific test (test_eventdev_octeontx.c) to
driver/event/octeontx.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
This commit fixes a bug in which, when the sw PMD is reconfigured, it would
leave stale IQ chunk pointers in each queue's IQ structure. Now, the PMD
initializes all IQs at eventdev start time and releases all IQ chunk
pointers at eventdev stop time (which has the consequence that any events
in a queue when the eventdev is stopped will be lost). This approach should
be resilient to any reconfiguration done between the stop and start, such
as adding or removing queues.
This commit also fixes two potential issues in iq_chunk.h. iq_init()
now initializes the IQ's count field to 0, and iq_dequeue_burst() sets
iq->head to the appropriate next pointer.
Fixes: dca926ca9faa ("event/sw: use dynamically-sized IQs")
Reported-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Signed-off-by: Gage Eads <gage.eads@intel.com>
Reviewed-by: Harry van Haaren <harry.van.haaren@intel.com>
Acked-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
This commit adds unit test inside the OPDL PMD. There is a PMd parameter
"self_test" can be used to triger the test when vdev bus probe opdl device
e.g.
sudo ./app/test --vdev="event_opdl0,self_test=1"
Signed-off-by: Liang Ma <liang.j.ma@intel.com>
Signed-off-by: Peter Mccarthy <peter.mccarthy@intel.com>
This commit adds a OPDL implementation of the eventdev API. The
implementation here is intended to enable the community to use
the OPDL infrastructure under eventdev API.
The main components of the implementation is three files:
- opdl_evdev.c Creation, configuration, etc
- opdl_evdev_xstats.c helper function to support stats collection
- opdl_evdev.h include the main data structure of opdl
device and all the function prototype
need to be exposed to support eventdev API.
- opdl_evdev_init.c implement all initailization helper function
This commit only adds the implementation, no existing DPDK files
are modified.
Signed-off-by: Liang Ma <liang.j.ma@intel.com>
Signed-off-by: Peter Mccarthy <peter.mccarthy@intel.com>
OPDL ring is the core infrastructure of OPDL PMD. OPDL ring library
provide the core data structure and core helper function set. The Ring
implements a single ring multi-port/stage pipelined packet distribution
mechanism. This mechanism has the following characteristics:
• No multiple queue cost, therefore, latency is significant reduced.
• Fixed dependencies between queue/ports is more suitable for complex.
fixed pipelines of stateless packet processing (static pipeline).
• Has decentralized distribution (no scheduling core).
• Packets remain in order (no reorder core(s)).
* Update build system to enable compilation.
Signed-off-by: Liang Ma <liang.j.ma@intel.com>
Signed-off-by: Peter Mccarthy <peter.mccarthy@intel.com>
Reviewed-by: Seán Harte <seanbh@gmail.com>
This commit modifies the sw PMD credit scheme such that credits are
consumed when enqueueing a NEW event and released when an event is
released -- typically, the beginning and end of a pipeline. Workers that
simply forward events do not interact with the credit pool.
Signed-off-by: Gage Eads <gage.eads@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
This commit introduces a capability for disabling the "implicit" release
functionality for a port, which prevents the eventdev PMD from issuing
outstanding releases for previously dequeued events when dequeuing a new
batch of events.
If a PMD does not support this capability, the application will receive an
error if it attempts to setup a port with implicit releases disabled.
Otherwise, if the port is configured with implicit releases disabled, the
application must release each dequeued event by invoking
rte_event_enqueue_burst() with RTE_EVENT_OP_RELEASE or
RTE_EVENT_OP_FORWARD.
Signed-off-by: Gage Eads <gage.eads@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
This commit introduces dynamically-sized IQs, by switching the underlying
data structure from a fixed-size ring to a linked list of queue 'chunks.'
This has a number of benefits:
- Certain corner cases were observed in which all of a pipeline's flows
could be pinned to one port for extended periods, effectively turning a
multi-core pipeline into single-core one. This was caused by an event
producer having a larger new_event_threshold than the IQ depth, and
injecting large numbers of packets that are ultimately backpressured in a
worker's rx_ring, causing those packets' flows to be scheduled to that
port.
The dynamically sized IQ does not have this problem because each IQ can
grow large enough to store all the system's events, such that
backpressure will not reach the worker_ring.
- Slight performance improvement (~1-2%) in high throughput scenarios,
tested with eventdev_pipeline_sw_pmd.
This implementation has a small increase in the queue storage memory
footprint (~70KB). This commit also removes the iq_size xstat, which no
longer applies to this implementation.
Signed-off-by: Gage Eads <gage.eads@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
This commit reinitializes a queue before it is reconfigured, such that
reorder buffer memory is not leaked.
This bug masked a few other problems, which this commit corrects as well:
- sw_port_link() allowed a port to link to a queue twice, such that the
port could then successfully unlink the queue twice. Now the link
function checks whether a port is already linked to the queue, and if so
returns success but doesn't assign the a port a second slot in the
queue's cq map.
- test_eventdev.c's test_eventdev_unlink() was unlinking a queue twice
from the same port, and expecting the second unlink to succeed. Now the
test unlinks, links, then unlinks again.
- test_eventdev.c's test_eventdev_link_get() was linking a single queue but
expecting the unlink function to return nb_queues (where nb_queues > 1).
The test now checks for a return value of 1.
Fixes: 5ffb2f142d95 ("event/sw: support event queues")
Fixes: 371a688fc159 ("event/sw: support linking queues to ports")
Fixes: f8f9d233ea0e ("test/eventdev: add unit tests")
Cc: stable@dpdk.org
Signed-off-by: Gage Eads <gage.eads@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
- bbdev 'turbo_sw' is the software accelerated version of 3GPP L1
Turbo coding operation using the optimized Intel FlexRAN SDK libraries.
- 'turbo_sw' pmd is disabled by default
Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
- 'bbdev_null' is a basic pmd that performs a minimalistic
bbdev operation
- useful for bbdev smoke testing and in measuring the overhead
introduced by the bbdev library
- 'bbdev_null' pmd is enabled by default
Signed-off-by: Amr Mokhtar <amr.mokhtar@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
The mbox messages are naturally aligned and in some cases compiler
optimization might disregard natural alignment.
Use volatile key word to force compiler to disable optimizing and
maintain alignment.
Fixes: aecb8e093b52 ("event/octeontx: introduce specialized mbox message copy")
Cc: stable@dpdk.org
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Acked-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Remove RTE_LOG_LEVEL config option, use existing RTE_LOG_DP_LEVEL config
option for controlling datapath log level.
RTE_LOG_LEVEL is no longer needed as dynamic logging can be used to
control global and module specific log levels.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
Reviewed-by: Olivier Matz <olivier.matz@6wind.com>
Many exported headers rely on definitions found in rte_config.h without
including it, as shown by the following command:
grep -L '^#include <rte_config.h>' -- \
$(grep -Rl \
$(sed -n '/^#define \([^ ]\+\).*$/{s//\1/;H;};${x;s/\n//;s/\n/\\|/g;p;}' \
build/include/rte_config.h) \
-- build/include/)
We cannot assume external applications will include rte_config.h on their
own, neither directly nor through a -include parameter like DPDK does
internally.
This not only causes obvious compilation failures that can be reproduced
with check-includes.sh such as:
[...]/rte_memory.h:88:43: error: ‘RTE_CACHE_LINE_SIZE’ was not declared in
this scope
#define __rte_cache_aligned __rte_aligned(RTE_CACHE_LINE_SIZE)
^
It also results in less visible issues, for instance rte_hash_crc.h relying
on RTE_ARCH_X86_64's presence to provide dedicated inline functions.
This patch partially reverts the commit below and adds missing include
lines to the remaining files.
Fixes: f1a7a5c5f404 ("remove include of generated config header")
Cc: stable@dpdk.org
Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
Reported by check-includes.sh:
[...]/rte_pmd_i40e.h:97:30: error: ISO C restricts enumerator values to
range of `int' [-Werror=pedantic]
RTE_PMD_I40E_PKG_INFO_MAX = 0xFFFFFFFF
^
Fixes: edeab742edac ("net/i40e: get information about DDP profile")
Cc: stable@dpdk.org
Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>
This patch adds dev_pause, dev_resume and inject_pkts APIs to allow
driver to pause the worker threads and inject special packets into
Tx queue. The next patch will be based on this.
Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
