Adding macro based framework to hook rx/tx burst function
pointers to the appropriate function based on rx/tx offloads.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Adding multi segment support to the octeontx PMD. Also
adding the logic to share rx/tx ofloads with the eventdev
code.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Add libatomic as a global dependency when compiling for 32-bit using
clang. As we need libatomic for 64-bit atomic ops.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Add redundant stack variable initialization to work around
false-positive warnings in older versions of GCC.
Fixes: 1f2b99e8d9 ("event/dsw: improve migration mechanism")
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Use c11 atomics with RELAXED ordering instead of rte_atomic ops which
enforce unnessary barriers on arm64.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Reviewed-by: Phil Yang <phil.yang@arm.com>
DSW keeps an internal port load estimate, used by the load balancing
mechanism. As a side effect, it keeps track of the total number of
busy cycles since startup. This metric is indirectly exposed in the
form of DSW xstats' "port_<n>_event_proc_latency", which is the total
number of busy cycles divided by the total number of events processed
on a particular port.
An external application can take (event_latency * dequeued) to go back
to busy_cycles. One reason for doing this is to measure the port's
load during a longer time period, without resorting to sampling
"port_<n>_load". However, as the number dequeued events grows, a
rounding error in event_latency renders the application-calculated
busy_cycles inaccurate.
Thus, it makes sense to directly expose the number of busy cycles as a
DSW xstats, even though it might seem redundant.
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
On dequeue, polling the control ring once is enough.
Fixes: f6257b22e7 ("event/dsw: add load balancing")
Cc: stable@dpdk.org
Suggested-by: Ola Liljedahl <ola.liljedahl@arm.com>
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
DSW limits the rate of migrations on a per-port basis. Hence, as the
number of cores grows, so does the total migration capacity.
In high core-count systems, this allows for a situation where flows
are migrated to a lightly loaded port which recently already received
a number of new flows (from other ports). The processing load
generated by these new flows may not yet be reflected in the lightly
loaded port's load estimate. The result is that the previously lightly
loaded port is now overloaded.
This patch adds a rough estimate of the size of the inbound migrations
to a particular port, which can be factored into the migration logic,
avoiding the above problem.
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Allowing moving multiple flows in one migration transaction, to
rebalance load more quickly.
Introduce a threshold to avoid migrating flows between ports with very
similar load.
Simplify logic for selecting which flow to migrate. The aim is now to
move flows in such a way that the receiving port is as lightly-loaded
as possible (after receiving the flow), while still migrating enough
flows from the source port to reduce its load. This is essentially how
legacy strategy work as well, but the code is more readable.
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
To allow visualization of migrations, track the number flow
immigrations in "port_<N>_immigrations". The "port_<N>_migrations"
retains legacy semantics, but is renamed "port_<N>_emigrations".
Expose the number of events currently undergoing processing
(i.e. pending releases) at a particular port.
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Reduce the maximum number of DSW flows from 32k to 8k, to be able
rebalance load faster.
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
In DSW, in case a port can't produce any events for the application to
consume, the port is considered idle.
To slightly reduce wall-time latency, flush the port's output buffer
in case of such an empty dequeue.
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Each workslot is always bound to a specific lcore there is no multi-core
contention to cause cache trashing as a result it is safe to remove the
WFE. Also, in dual workslot dequeue work will mostlikely be available on
the pair workslot making WFE impractical.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Reviewed-by: Gavin Hu <gavin.hu@arm.com>
Remove setting ALLOW_EXPERIMENTAL_API individually for each Makefile and
meson.build. Instead, enable ALLOW_EXPERIMENTAL_API flag across app, lib
and drivers.
This changes reduces the clutter across the project while still
maintaining the functionality of ALLOW_EXPERIMENTAL_API i.e. warning
external applications about experimental API usage.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
The EAL API (with doxygen documentation) is moved from
common/include/ to include/, which makes more clear that
it is the global API for all environments and architectures.
Note that the arch-specific and OS-specific include files are not
in this global include directory, but include/generic/ should
cover the doxygen documentation for them.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: David Marchand <david.marchand@redhat.com>
This patch sets the number of atomic ordered sequences
supported by the driver.
Fixes: dbf63bd43a ("event/dpaa2: support ordered queue")
Cc: stable@dpdk.org
Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Three warnings are commonly disabled in DPDK with make and meson:
* address-of-packed-member
always disabled
* missing-field-initializers
disabled with meson
disabled with make + clang or make + gcc < 4.7
disabled with make + gcc <= 5 for test files and event drivers
* packed-not-aligned
disabled with meson
This change is removing exceptions for missing-field-initializers.
As it is always disabled, some redundant configs are cleaned up.
Now the situation is:
* address-of-packed-member
always disabled
* missing-field-initializers
always disabled
* packed-not-aligned
disabled with meson
It could alternatively be decided to disable missing-field-initializers
only for old gcc (< 6).
The warning packed-not-aligned is not modified in this change.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Added new flag for SECURITY in compiler optimized Tx fastpath
framework. With this, compiler autogenerates functions which
have security enabled.
Signed-off-by: Ankur Dwivedi <adwivedi@marvell.com>
Signed-off-by: Anoob Joseph <anoobj@marvell.com>
Signed-off-by: Archana Muniganti <marchana@marvell.com>
Signed-off-by: Tejasree Kondoj <ktejasree@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Replace DSW's use of regular DPDK rings (and code for
packing/unpacking control messages into void pointers) with custom
size rings.
In addition to cleaner code, this change allows DSW to support up to
the eventdev API's maximum of 255 ports by tweaking DSW_MAX_PORTS.
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Use the new API to wait in low power state instead of continuous
polling to save CPU cycles and power.
Signed-off-by: Gavin Hu <gavin.hu@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Reviewed-by: Jerin Jacob <jerinj@marvell.com>
Relax memory requirement for event timers when internal mempool used is
octeontx2 mempool.
Add debug log to print the memory used.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
If an application issues rte_event_enqueue_new_burst() or
rte_event_enqueue_forward_burst() call with a burst of events longer
than the configured max enqueue burst size, DSW allocates credits not
only for events actually enqueued, but for the complete burst. If this
process is repeated, enough credits will have leaked to cause the
event device to backpressure (i.e. disallow) any new enqueue
operations.
In addition, the port-level enqueue xstats will log the wrong number
of events enqueued for oversized enqueues.
This patch makes DSW gracefully handle oversized enqueue bursts.
Fixes: 1c8e3caa3b ("event/dsw: add event scheduling and device start/stop")
Cc: stable@dpdk.org
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Fix DSW's rte_event_enqueue_burst(), so that a call with a zero-sized
event array immediately flushes the port's output buffers. Prior to
this patch, the flush operation would be deferred to the next enqueue
or dequeue call, which is inconsistent with DSW documentation.
Fixes: 1c8e3caa3b ("event/dsw: add event scheduling and device start/stop")
Cc: stable@dpdk.org
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
When net/octeontx is connected to event/octeontx as an event Rx adapter,
PKI aka 'net/octeontx' can forward packets directly to SSO aka
'event/octeontx'.
When pumping traffic to PKI if flow control is disabled internal FIFOs
might be overrun causing partial l2 packets to be enqueued.
SSO receives <31:0> TAG tag calculated by PKI, in normal cases <31:28>
is always 0 which signifies RTE_EVENT_TYPE_ETHDEV. But in case of
partial received packets PKI sets the <31:0> TAG as 0xFFFFFFFF which
is an invalid event type.
Add a check to see if TAG is 0xFFFFFFFF and free the partial receive
packet.
Fixes: d0d6549860 ("net/octeontx: support event Rx adapter")
Cc: stable@dpdk.org
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
Since TIM buckets are always aligned to 32B and our cache line size being
128B, we will always have a cache miss when reading current_chunk pointer.
Avoid the cache miss by storing the current_chunk pointer in the bucket
opposite to the current bucket.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Fix HW race condition observed when timeout resolution is low (<5us).
When HW traverses a given TIM bucket it will clear chunk_remainder,
but since SW always decreases the chunk_remainder at the start of the
arm routine it might cause a race where SW updates chunk_remainder
after HW has cleared it that lead to nasty side effects.
Fixes: 95e4e4ec74 ("event/octeontx2: add timer arm timeout burst")
Cc: stable@dpdk.org
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Sync mail box data structures to version 1.1.3.
Add mail box version verification and defer initializing octeontx
devices if mail box version mismatches.
Update OCTEON TX limitaion with max mempool size used.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Reviewed-by: Jerin Jacob <jerinj@marvell.com>
Merge all versions in linker version script files to DPDK_20.0.
This commit was generated by running the following command:
:~/DPDK$ buildtools/update-abi.sh 20.0
Signed-off-by: Pawel Modrak <pawelx.modrak@intel.com>
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
Since the library versioning for both stable and experimental ABI's is
now managed globally, the LIBABIVER and version variables no longer
serve any useful purpose, and can be removed.
The replacement in Makefiles was done using the following regex:
^(#.*\n)?LIBABIVER\s*:=\s*\d+\n(\s*\n)?
(LIBABIVER := numbers, optionally preceded by a comment and optionally
succeeded by an empty line)
The replacement for meson files was done using the following regex:
^(#.*\n)?version\s*=\s*\d+\n(\s*\n)?
(version = numbers, optionally preceded by a comment and optionally
succeeded by an empty line)
[David]: those variables are manually removed for the files:
- drivers/common/qat/Makefile
- lib/librte_eal/meson.build
[David]: the LIBABIVER is restored for the external ethtool example
library.
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
Some variables are commented out. Remove them.
Fixes: d548ef513c ("event/opdl: add unit tests")
Cc: stable@dpdk.org
Signed-off-by: Kevin Traynor <ktraynor@redhat.com>
Acked-by: Liang Ma <liang.j.ma@intel.com>
During LTO build compiler reports some 'false positive' warnings about
variables being possibly used uninitialized. This patch silences these
warnings.
Exemplary compiler warning to suppress (with LTO enabled):
error: ‘chunk’ may be used uninitialized in this function
[-Werror=maybe-uninitialized]
bkt->current_chunk = (uintptr_t)chunk;
Signed-off-by: Andrzej Ostruszka <aostruszka@marvell.com>
Use WFE to save power while waiting for tag to become head.
SSO signals EVENTI to allow cores to exit from wfe when they
are waiting for specific operations in which one of them is
setting HEAD bit in GWS_TAG.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Reviewed-by: Gavin Hu <gavin.hu@arm.com>
This code is being shared by more than 1 type of driver.
Common is most appropriate place for it.
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>
While meson always adds -Wall flag to C compiles, the make build adds extra
warning flags that are not present in the meson build. This addresses that
shortcoming by adding additional warning flags to our builds. The one
omission is the -Wcast-align flag, which though present in make gcc builds,
gives a lot of warnings/errors when used with clang.
The removed warning "-Wunused-parameter" is covered by the "-Wextra"
parameter so is unnecessary.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Luca Boccassi <bluca@debian.org>
This patch adds the support of Tx adapter for DPAA2 platform
Signed-off-by: Nipun Gupta <nipun.gupta@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
