Marko Kovacevic [Tue, 27 Mar 2018 12:14:30 +0000 (13:14 +0100)]
crypto/aesni_gcm: support IPsec Multi-buffer lib v0.49
Adds support for the v0.49 of the IPsec Multi-buffer lib,
which now gets compiled and installed as a shared object.
Therefore, there is no need to pass the AESNI_MULTI_BUFFER_LIB_PATH
Signed-off-by: Marko Kovacevic <marko.kovacevic@intel.com> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Marko Kovacevic [Tue, 27 Mar 2018 12:14:29 +0000 (13:14 +0100)]
crypto/aesni_mb: support IPsec Multi-buffer lib v0.49
Adds support for the v0.49 of the IPsec Multi-buffer lib,
which now gets compiled and installed as a shared object.
Therefore, there is no need to pass the AESNI_MULTI_BUFFER_LIB_PATH
Signed-off-by: Marko Kovacevic <marko.kovacevic@intel.com> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Tomasz Duszynski [Wed, 14 Mar 2018 13:12:54 +0000 (14:12 +0100)]
test/crypto: add MRVL to hash test cases
MRVL Crypto PMD supports most of the hash algorithms covered
by test suites thus specific bits should be set in pmd_masks.
Otherwise blockcipher authonly test returns success even though no
real tests have been executed.
Fixes: 84e0ded38ac5 ("test/crypto: add mrvl crypto unit tests") Cc: stable@dpdk.org Signed-off-by: Tomasz Duszynski <tdu@semihalf.com> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Fiona Trahe [Fri, 9 Mar 2018 17:44:38 +0000 (17:44 +0000)]
cryptodev: change argument of driver registration
Pass an rte_driver to the RTE_PMD_REGISTER_CRYPTO_DRIVER macro
rather than an unspecified container which holds an rte_driver.
All the macro actually needs is the rte_driver, not the
container holding it.
This paves the way for a later patch in which a driver
will be registered which does not naturally derive from a
container and so avoids having to create an arbitrary container
to pass in the rte_driver.
This patch changes the cryptodev lib macro and all the
PMDs which use it.
Maxime Coquelin [Wed, 14 Mar 2018 16:31:25 +0000 (17:31 +0100)]
vhost: deprecate unsafe GPA translation API
This patch marks rte_vhost_gpa_to_vva() as deprecated because
it is unsafe. Application relying on this API should move
to the new rte_vhost_va_from_guest_pa() API, and check
returned length to avoid out-of-bound accesses.
This issue has been assigned CVE-2018-1059.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Maxime Coquelin [Thu, 19 Apr 2018 08:52:33 +0000 (10:52 +0200)]
vhost/crypto: move to safe GPA translation API
This patch uses the new rte_vhost_va_from_guest_pa() API
to ensure all the descriptor buffer is mapped contiguously
in the application virtual address space.
It does not handle buffers discontiguous in host virtual
address space, but only return an error.
This issue has been assigned CVE-2018-1059.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Maxime Coquelin [Wed, 14 Mar 2018 15:46:46 +0000 (16:46 +0100)]
examples/vhost_scsi: move to safe GPA translation API
This patch uses the new rte_vhost_va_from_guest_pa() API
to ensure all the descriptor buffer is mapped contiguously
in the application virtual address space.
As the application did not checked return of previous API,
this patch just print an error if the buffer address isn't in
the vhost memory regions or if it is scattered. Ideally, it
should handle scattered buffers gracefully.
This issue has been assigned CVE-2018-1059.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Maxime Coquelin [Tue, 13 Mar 2018 18:43:48 +0000 (19:43 +0100)]
examples/vhost: move to safe GPA translation API
This patch uses the new rte_vhost_va_from_guest_pa() API
to ensure the application doesn't perform out-of-bound
accesses either because of a malicious guest providing an
incorrect descriptor length, or because the buffer is
contiguous in guest physical address space but not in the
host process virtual address space.
This issue has been assigned CVE-2018-1059.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Maxime Coquelin [Tue, 23 Jan 2018 13:37:50 +0000 (14:37 +0100)]
vhost: introduce safe API for GPA translation
This new rte_vhost_va_from_guest_pa API takes an extra len
parameter, used to specify the size of the range to be mapped.
Effective mapped range is returned via len parameter.
This issue has been assigned CVE-2018-1059.
Reported-by: Yongji Xie <xieyongji@baidu.com> Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Maxime Coquelin [Tue, 23 Jan 2018 13:26:02 +0000 (14:26 +0100)]
vhost: check all range is mapped when translating GPAs
There is currently no check done on the length when translating
guest addresses into host virtual addresses. Also, there is no
guanrantee that the guest addresses range is contiguous in
the host virtual address space.
This patch prepares vhost_iova_to_vva() and its callers to
return and check the mapped size. If the mapped size is smaller
than the requested size, the caller handle it as an error.
This issue has been assigned CVE-2018-1059.
Reported-by: Yongji Xie <xieyongji@baidu.com> Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
This patch fixes the size passed at the indirect descriptor
table translation time, which is the len field of the descriptor,
and not a single descriptor.
This issue has been assigned CVE-2018-1059.
Fixes: 62fdb8255ae7 ("vhost: use the guest IOVA to host VA helper") Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Thomas Monjalon [Thu, 19 Apr 2018 23:53:10 +0000 (01:53 +0200)]
mbuf: improve tunnel Tx offloads API doc
Add few details to remind TSO flag, checksum flags and header lengths.
The doxygen syntax for MPLS-in-UDP is fixed.
Fixes: d95188551fa1 ("mbuf: introduce new Tx offload flag for MPLS-in-UDP") Cc: stable@dpdk.org Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Olivier Matz <olivier.matz@6wind.com>
Thomas Monjalon [Thu, 19 Apr 2018 23:39:52 +0000 (01:39 +0200)]
mbuf: fix Tx checksum offload API doc
When introducing rte_eth_tx_prepare(), the constraints on checksum
pre-filling for Tx offloads were relaxed because implemented in
the PMDs with rte_net_intel_cksum_flags_prepare() helper.
As a consequence, these old requirements are removed for:
- PKT_TX_OUTER_IP_CKSUM
- PKT_TX_IP_CKSUM
- PKT_TX_[L4]_CKSUM
- PKT_TX_TCP_SEG
Not sure SCTP offload is properly implemented though.
A reference to rte_eth_tx_prepare() is added in rte_eth_tx_burst() doc.
Fixes: 609dd68ef14f ("mbuf: enhance the API documentation of offload flags") Fixes: 4fb7e803eb1a ("ethdev: add Tx preparation") Cc: stable@dpdk.org Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Olivier Matz <olivier.matz@6wind.com>
Thomas Monjalon [Mon, 23 Apr 2018 01:18:35 +0000 (03:18 +0200)]
maintainers: fix typo and ordering
Fix logical/alphabetical ordering, spacing, and syntax typo.
Fixes: 8fb3b2576025 ("maintainers: call out subtree committers") Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Tomasz Duszynski [Wed, 11 Apr 2018 11:45:07 +0000 (13:45 +0200)]
net/mvpp2: add to meson build
Add support for building MRVL MVPP2 PMD with meson. To avoid cluttering
the build environment and to keep all relevant settings local to a cross
build we get MUSDK library installation path from a meson option.
Signed-off-by: Tomasz Duszynski <tdu@semihalf.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Jerin Jacob [Mon, 9 Apr 2018 14:39:46 +0000 (20:09 +0530)]
build: fix default arm64 instruction level
The make based build system has crc+crypto instruction
support for the default arm64 build.
http://dpdk.org/browse/dpdk/tree/mk/machine/armv8a/rte.vars.mk#n31
This patch fixes the disparity with meson build flags for armv8.
As a bonus, This patch fixes the following errors with
ip_pipeline example application.
Assembler messages:
Error: selected processor does not support `crc32cx w3,w3,x0'
Fixes: c6e536e38437 ("build: add more implementers IDs and PNs for ARM") Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com> Tested-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Bruce Richardson [Thu, 29 Mar 2018 13:54:36 +0000 (14:54 +0100)]
drivers/dpaa: reduce meson dependency lists
Meson build currently tracks the dependencies between libraries, which
can often make things easier, but has the side-effect of slowing down
the initial meson run if too many duplicated dependencies are provided.
Therefore, we remove dependencies from the dpaa items where other
dependencies already depend on those. This provides a noticable speed-up
in meson configuration runs when lots of sample apps are included in the
build.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com> Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Bruce Richardson [Thu, 29 Mar 2018 13:54:35 +0000 (14:54 +0100)]
examples: improve error report for missing meson deps
When a required library is missing on a platform, rather than having
meson report an error about the missing variable, catch the problem
earlier and provide a more readable message.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
Bruce Richardson [Thu, 29 Mar 2018 13:54:34 +0000 (14:54 +0100)]
examples: allow building all as part of meson build
To test building all relevant example applications as part of a build, we
add support for the "all" keyword to be passed to the "examples" build
option. Since not all examples can actually be built on all systems,
we also add support for the "build" option inside the sub-dirs. However,
in case where "all" is not used, and a particular example is requested
to be built, we will error out if building the requested app is not
possible.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
Bruce Richardson [Thu, 29 Mar 2018 13:54:31 +0000 (14:54 +0100)]
examples: add empty meson files for unsupported ones
A number of example apps are not supported by the meson build system yet,
but to allow future testing with "-Dexamples=all" we add in a placeholder
meson.build file indicating that the apps should not be built.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Tested-by: Harry van Haaren <harry.van.haaren@intel.com>
Bruce Richardson [Thu, 29 Mar 2018 13:48:47 +0000 (14:48 +0100)]
build: remove checks for non-optional libraries
Unless a library cannot be built for a specific platform (generally
BSD), it will always be available. Therefore remove checks for IP
fragmentation and ACL libraries, since these are built for all
platforms.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
build: clean up building kernel modules using meson
The meson.build files for building the kernel modules directory could
be improved now that it is extracted from the EAL. For example, no
global processing is necessary inside the kernel folder, just need to
subdir to the appropriate bsd or linux folder to do the actual work.
To avoid potential race conditions with the BSD module builds when
the kernel build system is creating the dev_if.h and other files,
we serialize the kernel module builds (all 2 of them!) by setting
up each module to depend on all the previous.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Tested-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Vipin Varghese [Mon, 26 Feb 2018 06:22:38 +0000 (11:52 +0530)]
app/pdump: check for ports
In case of application build with shared library mode unless
option '-d' is passed, poll mode driver for devices is not
initialized. Notifying the user just after rte_eal_init is
pro active way of intimating the user.
Ferruh Yigit [Mon, 12 Mar 2018 17:53:08 +0000 (17:53 +0000)]
app/testpmd: print Rx/Tx offload values on start
Which per port offloads are enabled is not clear. Printing offloads
values at forwarding start.
CRC strip offload value was printed in more verbose manner, it is
removed since Rx/Tx offload values covers it and printing only CRC one
can cause confusion.
Hexadecimal offloads values are not very user friendly but preferred to
not create to much noise during forwarding start.
Pablo de Lara [Tue, 17 Apr 2018 13:13:42 +0000 (14:13 +0100)]
eventdev: fix build with icc
ICC complains about variable being used before its value is set.
Since the variable is only assigned in the for loop,
its declaration is moved inside and is initialized.
lib/librte_eventdev/rte_event_timer_adapter.c(708): error #592:
variable "ret" is used before its value is set
RTE_SET_USED(ret);
Fixes: 6750b21bd6af ("eventdev: add default software timer adapter") Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com> Acked-by: Ferruh Yigit <ferruh.yigit@intel.com> Acked-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com>
Sometimes gcc does not inline the function despite keyword *inline*,
we observe rte_movX is not inline when doing performance profiling,
so use *always_inline* keyword to force gcc to inline the function.
Signed-off-by: Junjie Chen <junjie.j.chen@intel.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Thomas Monjalon [Thu, 5 Apr 2018 15:33:22 +0000 (17:33 +0200)]
ethdev: deprecate port count function
Some DPDK applications wrongly assume these requirements:
- no hotplug, i.e. ports are never detached
- all allocated ports are available to the application
Such application iterates over ports by its own mean.
The most common pattern is to request the port count and
assume ports with index in the range [0..count[ can be used.
In order to fix this common mistake in all external applications,
the function rte_eth_dev_count is deprecated, while introducing
the new functions rte_eth_dev_count_avail and rte_eth_dev_count_total.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Thomas Monjalon [Thu, 5 Apr 2018 15:33:21 +0000 (17:33 +0200)]
fix ethdev port id validation
Some DPDK applications wrongly assume these requirements:
- no hotplug, i.e. ports are never detached
- all allocated ports are available to the application
Such application assume a valid port index is in the range [0..count[.
There are three consequences when using such wrong design:
- new ports having an index higher than the port count won't be valid
- old ports being detached (RTE_ETH_DEV_UNUSED) can be valid
Such mistake will be less common with growing hotplug awareness.
All applications and examples inside this repository - except testpmd -
must be fixed to use the function rte_eth_dev_is_valid_port.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Thomas Monjalon [Thu, 5 Apr 2018 15:33:20 +0000 (17:33 +0200)]
fix ethdev ports enumeration
Some DPDK applications wrongly assume these requirements:
- no hotplug, i.e. ports are never detached
- all allocated ports are available to the application
Such application iterates over ports by its own mean.
The most common pattern is to request the port count and
assume ports with index in the range [0..count[ can be used.
There are three consequences when using such wrong design:
- new ports having an index higher than the port count won't be seen
- old ports being detached (RTE_ETH_DEV_UNUSED) can be seen as ghosts
- failsafe sub-devices (RTE_ETH_DEV_DEFERRED) will be seen by the application
Such mistake will be less common with growing hotplug awareness.
All applications and examples inside this repository - except testpmd -
must be fixed to use the iterator RTE_ETH_FOREACH_DEV.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
ring: relax alignment constraint on ring structure
The initial objective of
commit d9f0d3a1ffd4 ("ring: remove split cacheline build setting")
was to add an empty cache line between the producer and consumer
data (on platform with cache line size = 64B), preventing from
having them on adjacent cache lines.
Following discussion on the mailing list, it appears that this
also imposes an alignment constraint that is not required.
This patch removes the extra alignment constraint and adds the
empty cache lines using padding fields in the structure. The
size of rte_ring structure and the offset of the fields remain
the same on platforms with cache line size = 64B:
While debugging startup issues encountered with Clang (see "eal: fix
undefined behavior in fbarray"), I noticed that fbarray stores indices,
sizes and masks on signed integers involved in bitwise operations.
Such operations almost invariably cause undefined behavior with values that
cannot be represented by the result type, as is often the case with
bit-masks and left-shifts.
This patch replaces them with unsigned integers as a safety measure and
promotes a few internal variables to larger types for consistency.
According to GCC documentation [1], the __builtin_clz() family of functions
yield undefined behavior when fed a zero value. There is one instance in
the fbarray code where this can occur.
Clang (at least version 3.8.0-2ubuntu4) seems much more sensitive to this
than GCC and yields random results when compiling optimized code, as shown
below:
#include <stdio.h>
int main(void)
{
volatile unsigned long long moo;
int x;
$ gcc -O3 -o test test.c && ./test
63
$ clang -O3 -o test test.c && ./test 1742715559
$ clang -O0 -o test test.c && ./test
63
Even 63 can be considered an unexpected result given the number of leading
zeroes should be the full width of the underlying type, i.e. 64.
In practice it causes find_next_n() to sometimes return negative values
interpreted as errors by caller functions, which prevents DPDK applications
from starting due to inability to find free memory segments:
Earlier fix for race condition introduced a bug where mutex
wasn't unlocked if message failed to be sent. Fix all of this
by moving locking out of mp_request_sync() altogether.
We are trying to notify sender that response from current process
should be ignored, but we didn't specify which request this response
was for. Fix by copying request name from the original message.
Fixes: 579a4ccc345c ("eal: ignore IPC messages until init is complete") Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com> Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>
Previously, we were removing request from the list only if we
have succeeded to send it. This resulted in leaving an invalid
pointer in the request list.
Fix this by only adding new requests to the request list if we
have succeeded in sending them.
Previously, we were adding synchronous requests to request list, we
were doing it after checking if request existed. However, we only
removed the request from the request list if we have succeeded in
sending the request. In case of failed request send, we left an
invalid pointer in the request list.
Fix this by only adding request to the list once we succeed in
sending it.
EAL did not stop processing further asynchronous requests on
encountering a request that should trigger the callback. This
resulted in erasing valid requests but not triggering them.
Fix this by stopping the loop once we have a request that
can trigger the callback. Once triggered, we go back to scanning
the request queue until there are no more callbacks to trigger.
Previously, VFIO functions were not compiled in and exported if
VFIO compilation was disabled. Fix this by actually compiling
all of the functions unconditionally, and provide missing
prototypes on Linux.
Fixes: 279b581c897d ("vfio: expose functions") Fixes: 73a639085938 ("vfio: allow to map other memory regions") Fixes: 964b2f3bfb07 ("vfio: export some internal functions") Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Bruce Richardson [Mon, 16 Apr 2018 13:39:11 +0000 (14:39 +0100)]
net/ixgbe: fix build when vector driver disabled
The new functions for Rx and Tx offloads should not be inside the
conditional block for the vector driver, otherwise compile errors occur
when vector driver is disabled. For example:
ixgbe_ethdev.c:3636:36: error:
implicit declaration of function ‘ixgbe_get_rx_queue_offloads’;
This shows up as an error when doing ARM builds using meson as the vector
driver is not (yet) enabled for those builds.
Fixes: 51215925a32f ("net/ixgbe: convert to new Tx offloads API") Fixes: ec3b1124d14d ("net/ixgbe: convert to new Rx offloads API") Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
test_perf_common.c: In function ‘perf_event_timer_producer’:
test_perf_common.c:99:3: error: missing initializer for
field ‘priority’ of ‘struct <anonymous>’
[-Werror=missing-field-initializers]
.ev.sched_type = t->opt->sched_type_list[0],
Fixes: d008f20bce23 ("app/eventdev: add event timer adapter as a producer") Reported-by: Andrew Rybchenko <arybchenko@solarflare.com> Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com> Tested-by: Andrew Rybchenko <arybchenko@solarflare.com>
dev_info->max_event_ports is uint8_t. dpaa_event_dev_info_get assigns
DPAA_EVENT_MAX_EVENT_PORT (which is RTE_MAX_LCORE, upto 256 in ppc64le)
into this variable, which breaks compile in ppc64le.
drivers/event/dpaa/dpaa_eventdev.c: In function ‘dpaa_event_dev_info_get’:
rte_config.h:23:23: error:
large integer implicitly truncated to unsigned type [-Werror=overflow]
#define RTE_MAX_LCORE 256
This patch removes the decalartion of rte_eventdev_driver from
rte_eventdev.h, as it not used anymore; pci_eventdev_skeleton_pmd
moved to use rte_pci_driver instead of rte_eventdev_driver.
event/octeontx: add option to use fpavf as chunk pool
Add compile-time configurable option to force TIMvf to use Octeontx
FPAvf pool manager as its chunk pool.
When FPAvf is used as pool manager the TIMvf automatically frees the
chunks to FPAvf through gpool-id.
When application sets `RTE_EVENT_TIMER_ADAPTER_F_ADJUST_RES` flag
while creating adapter underlying driver is free to optimize the
resolution for best possible configuration.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
event/octeontx: add single producer timer arm variant
When application creates the timer adapter by passing
`RTE_EVENT_TIMER_ADAPTER_F_SP_PUT` flag, we can optimize the arm sequence
by removing the locking overhead.
When application requests to start the timer adapter through
`rte_event_timer_adapter_start`, Octeontx TIMvf ring does the
following:
- Uses mbox to communicate TIMpf driver about,
* SCLK frequency used to convert ns<->cycles.
* program the ring control parameters and start the ring.
* get the exact cycle at which the TIMvf ring has started which can be
used to estimate the bucket position.
On `rte_event_timer_adapter_stop` i.e stop, Octeontx TIMvf ring does the
following:
- Use mbox to communicate TIMpf driver about,
* reset the ring control parameters and stop the ring.
When the application requests to create a timer device, Octeontx TIM
create does the following:
- Get the requested TIMvf ring based on adapter_id.
- Verify the config parameters supplied.
- Allocate memory required for
* Buckets based on min and max timeout supplied.
* Allocate the chunk pool based on the number of timers.
- Clear the interrupts.
On Free:
- Free the allocated bucket and chunk memory.
- Free private data used by TIMvf.
On Octeontx HW, each event timer device is enumerated as separate SRIOV VF
PCIe device.
In order to expose as a event timer device:
On PCIe probe, the driver stores the information associated with the
PCIe device and later when application requests for a event timer device
through `rte_event_timer_adapter_create` the driver infrastructure creates
the timer adapter with earlier probed PCIe VF devices.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Add options to configure expiry timeout, max number of timers and number
of event timer adapters through command line parameters.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com> Acked-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
app/eventdev: add burst mode for event timer adapter
Add burst mode for event timer adapter that can be selected by passing
--prod_type_timerdev_burst.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com> Acked-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
app/eventdev: add event timer adapter as a producer
Add event timer adapter as producer option that can be selected by
passing --prod_type_timerdev.
Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com> Acked-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com> Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com> Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
If an eventdev PMD does not wish to provide event timer adapter ops
definitions, the library will fall back to a default software
implementation whose entry points are added by this commit.
Signed-off-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com> Acked-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
The introduction of the event timer adapter library adds a dependency
on the rte_timer library from the rte_eventdev library. Update the
order so that the timer library comes after the eventdev library in the
linker command when statically linking applications.
Signed-off-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com> Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Event devices can be coupled with various components to provide
new event sources by using event adapters. The event timer adapter
is one such adapter; it bridges event devices and timer mechanisms.
This library extends the event-driven programming model by
introducing a new type of event that represents a timer expiration,
and it provides APIs with which adapters can be created or destroyed
and event timers can be armed and canceled.
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com> Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com> Signed-off-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
When an event device is stopped, it drains all event queues and ports.
These events may contain pointers, so to prevent memory leaks eventdev now
supports a user-provided flush callback that is called during the queue
drain process. This callback is stored in process memory, so the callback
must be registered by any process that may call rte_event_dev_stop().
This commit also clarifies the behavior of rte_event_dev_stop().
This follows this mailing list discussion:
http://dpdk.org/ml/archives/dev/2018-January/087484.html
Signed-off-by: Gage Eads <gage.eads@intel.com> Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Liang Ma [Tue, 27 Mar 2018 14:18:12 +0000 (15:18 +0100)]
event/opdl: fix atomic queue race condition
If application link one atomic queue to multiple ports,
and each worker core update flow_id, there will have a
chance to hit race condition issue and lead to double processing
same event. This fix solve the problem and eliminate
the race condition issue.
Fixes: 4236ce9bf5bf ("event/opdl: add OPDL ring infrastructure library") Cc: stable@dpdk.org Signed-off-by: Liang Ma <liang.j.ma@intel.com> Signed-off-by: Peter Mccarthy <peter.mccarthy@intel.com> Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
Gage Eads [Mon, 12 Mar 2018 14:55:22 +0000 (09:55 -0500)]
event/sw: perform partial burst enqueues
Previously, the sw PMD would enqueue either all or no events, depending on
if enough inflight credits were available for the new events in the burst.
If a port is enqueueing a large burst (i.e. a multiple of the credit update
quanta), this can result in suboptimal performance, and requires an
understanding of the sw PMD implementation (in particular, its credit
scheme) to tune an application's burst size.
This affects software that enqueues large bursts of new events, such as the
ethernet event adapter which uses a 128-deep event buffer, when the input
packet rate is sufficiently high.
This change makes the sw PMD enqueue as many events as it has credits, if
there are any new events in the burst.
Signed-off-by: Gage Eads <gage.eads@intel.com> Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
Vipin Varghese [Tue, 27 Feb 2018 20:08:05 +0000 (01:38 +0530)]
event/sw: add unlikely branch predict
For most run cases 'sw->started' holds true. Adding a branch prediction
suggestion to compiler helps as this is first conditional check just
after entering the function.
Signed-off-by: Vipin Varghese <vipin.varghese@intel.com> Acked-by: Kevin Laatz <kevin.laatz@intel.com>
git.droids-corp.org / dpdk.git / log