Dharmik Thakkar [Wed, 21 Oct 2020 22:50:04 +0000 (17:50 -0500)]
hash: implement RCU resources reclamation
Currently, users have to use external RCU mechanisms to free resources
when using lock free hash algorithm.
Integrate RCU QSBR process to make it easier for the applications to use
lock free algorithm.
Refer to RCU documentation to understand various aspects of
integrating RCU library into other libraries.
Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com> Signed-off-by: Dharmik Thakkar <dharmik.thakkar@arm.com> Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com> Acked-by: Ray Kinsella <mdr@ashroe.eu> Acked-by: Yipeng Wang <yipeng1.wang@intel.com>
Thomas Monjalon [Thu, 22 Oct 2020 19:22:45 +0000 (21:22 +0200)]
remove config prefix used with make
The config options CONFIG_RTE_* are simple RTE_* defines with meson.
Now that make support is dropped, update the names in logs and comments.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Ruifeng Wang <ruifeng.wang@arm.com> Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Acked-by: Haiyue Wang <haiyue.wang@intel.com> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Acked-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Thu, 22 Oct 2020 17:37:34 +0000 (19:37 +0200)]
eal: remove comment about old partition option
The main initialization function (rte_eal_init) has documentation
about a feature from another era: memory partition.
Curiously, this lost treasure is found only now,
suggesting there may be other interesting things to discover in the doc.
To all aspiring Indiana Jones: the hunt is open!
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: David Marchand <david.marchand@redhat.com>
Kevin Laatz [Wed, 21 Oct 2020 08:17:24 +0000 (09:17 +0100)]
doc: update patch cheatsheet to use meson
With 'make' being removed, the patch cheatsheet needs to be updated to
remove any references to 'make'. These references have been replaced with
meson alternatives in this patch.
Signed-off-by: Kevin Laatz <kevin.laatz@intel.com>
Ciara Power [Wed, 21 Oct 2020 08:17:21 +0000 (09:17 +0100)]
doc: remove reference to make from tools guide
Make is no longer supported for compiling DPDK, references are now
removed in the documentation.
Signed-off-by: Ciara Power <ciara.power@intel.com> Reviewed-by: Kevin Laatz <kevin.laatz@intel.com> Acked-by: Nicolas Chautru <nicolas.chautru@intel.com>
Ciara Power [Wed, 21 Oct 2020 08:17:20 +0000 (09:17 +0100)]
doc: remove references to make from apps guide
While make has been deprecated for DPDK, it's still applicable for
some example apps to be built standalone, this patch adjusts the
guides to take that into consideration.
Signed-off-by: Ciara Power <ciara.power@intel.com> Acked-by: Nicolas Chautru <nicolas.chautru@intel.com> Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
David Marchand [Thu, 22 Oct 2020 07:59:59 +0000 (09:59 +0200)]
hash: use x86 common flag for jhash
jhash has been forgotten when factorising the x86 arch check.
Fixes: dbf17d44f375 ("hash: use common x86 flag") Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
David Marchand [Thu, 22 Oct 2020 07:55:45 +0000 (09:55 +0200)]
drivers: add headers install helper
A lot of drivers export headers, reproduce the same facility than for
libraries.
Note: this change fixes an issue with the crypto scheduler headers which
were not installed properly. A separate backport will be sent to stable
branches.
Suggested-by: Bruce Richardson <bruce.richardson@intel.com> Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
David Marchand [Thu, 22 Oct 2020 07:47:04 +0000 (09:47 +0200)]
build: fix version map file references in documentation
Fixes: 63b3907833d8 ("build: remove library name from version map file name") Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Ray Kinsella <mdr@ashroe.eu> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Yunjian Wang [Sat, 23 May 2020 10:52:21 +0000 (18:52 +0800)]
eal/linux: fix memory leak in uevent handling
When the memory for uevent.devname is allocated in dev_uev_parse(). It
is not freed when parse the subsystem layer fails in dev_uev_parse().
Before return, it is also not freed in dev_uev_handler(). These cause a
memory leak.
Fixes: 0d0f478d0483 ("eal/linux: add uevent parse and process") Cc: stable@dpdk.org Signed-off-by: Yunjian Wang <wangyunjian@huawei.com> Reviewed-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Wed, 14 Oct 2020 20:18:13 +0000 (22:18 +0200)]
bus/pci: remove unused scan by address
The function pci_update_device was used to scan a device
for probing by PCI address.
This private function (and implementations) are unused
since such probing is removed.
Fixes: f3bac43b60da ("bus/pci: remove unused function to probe by address") Cc: stable@dpdk.org Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: David Marchand <david.marchand@redhat.com>
Tal Shnaiderman [Tue, 20 Oct 2020 08:48:15 +0000 (11:48 +0300)]
eal/windows: add missing stdint include
Following the addition of the in_addr/in6_addr structs
to in.h the header file must have stdint.h included
for the definitions of the uint8_t/uint32_t types used
within the new structs.
Not having it could results in the following errors
in places where in.h is included:
in.h:30:2: error: unknown type name 'uint32_t'
uint32_t s_addr;
in.h:34:2: error: unknown type name 'uint8_t'
uint8_t s6_addr[16];
Bruce Richardson [Tue, 20 Oct 2020 09:17:43 +0000 (10:17 +0100)]
config: fix kni build
When building kernel modules such as kni, the "config" directory is not
passed as a standard path in the EXTRA_CFLAGS value, meaning that the
rte_compatibility_defines.h is not found from rte_config.h. However, since
both headers are in the same directory, we can just use quotes rather than
angle-brackets to ensure the second header is always found if the first is.
Fixes: b0b672aeadaa ("build: add defines for compatibility with make build") Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: David Marchand <david.marchand@redhat.com>
Update the coding style document to include a policy against
introducing new master/slave usage. This is taken from the similar
place in the Linux kernel coding style.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
app/procinfo: provide way to request info on owned ports
There are cases where a port maybe owned by another (failsafe, netvsc,
bond); but currently proc-info has no way to look at stats of those
ports. This patch provides way for the user to explicitly ask for these
ports.
If no portmask is given the output is unchanged; it only shows the
top level ports. If portmask requests a specific port it will be
shown even if owned.
Increase the size of port mask variable to unsigned long to
allow up to 64 ports to be handled on 64 bit architecture.
The device owner is also a useful thing to show in port info.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Revise the display of port information to include more data
and be more human friendly.
* Show driver and device information
* Show MAC address
* Show flow control information
* Combine lines if possible
* Show all multicast mode
* Show queue mempool name
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Ophir Munk [Sun, 18 Oct 2020 14:21:47 +0000 (14:21 +0000)]
app/regex: fix crash in options parsing
getopt_long() parses command-line arguments. One of its arguments
'longopts' is a pointer to the first element of an array of struct
option. The last element of the array has to be filled with zeros
to mark the end of options. For example:
This commit adds the last element. Prior to this commit getopt_long()
continued parsing beyond the longopts[] array which occasionally caused
segmentation faults.
Fixes: de06137cb295 ("app/regex: add RegEx test application") Cc: stable@dpdk.org Signed-off-by: Ophir Munk <ophirmu@nvidia.com> Acked-by: Ori Kam <orika@nvidia.com> Acked-by: Lukasz Wojciechowski <l.wojciechow@partner.samsung.com>
Bruce Richardson [Thu, 15 Oct 2020 15:05:53 +0000 (16:05 +0100)]
build: standardize component names and defines
As discussed on the dpdk-dev mailing list[1], we can make some easy
improvements in standardizing the naming of the various components in DPDK,
and their associated feature-enabled macros.
Following this patch, each library will have the name in format,
'librte_<name>.so', and the macro indicating that library is enabled in the
build will have the form 'RTE_LIB_<NAME>'.
Similarly, for libraries, the equivalent name formats and macros are:
'librte_<class>_<name>.so' and 'RTE_<CLASS>_<NAME>', where class is the
device type taken from the relevant driver subdirectory name, i.e. 'net',
'crypto' etc.
To avoid too many changes at once for end applications, the old macro names
will still be provided in the build in this release, but will be removed
subsequently.
Bruce Richardson [Thu, 15 Oct 2020 15:05:52 +0000 (16:05 +0100)]
devtools: load all drivers in null test
Rather than specifying specific drivers in the driver directory to load, we
can just pass in the whole driver directory to the "-d" EAL flag, causing
all drivers to load. This makes the load of driver independent of any
specific driver names.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Bruce Richardson [Thu, 15 Oct 2020 15:05:51 +0000 (16:05 +0100)]
build: remove library name from version map file name
Since each version map file is contained in the subdirectory of the library
it refers to, there is no need to include the library name in the filename.
This makes things simpler in case of library renaming.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Luca Boccassi <bluca@debian.org> Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru> Acked-by: Rosen Xu <rosen.xu@intel.com>
Bruce Richardson [Thu, 15 Oct 2020 15:05:50 +0000 (16:05 +0100)]
common/qat: build drivers from common folder
Since the drivers in the common directory can be processed out of order, in
this case following the "bus" directory, we can simplify somewhat the build
of the QAT driver to be done entirely from the "common/qat" folder rather
than having it's build distributed across 3 folders.
This also opens up the possibility of building the QAT driver with crypto
only and the compression part disabled. It further allows more sensible
naming of the resulting shared library in case of standardizing library
names based on device class; i.e. common_qat is more descriptive for a
combined crypto/compression driver than either of the other two prefixes
individually.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Luca Boccassi <bluca@debian.org>
Bruce Richardson [Thu, 15 Oct 2020 15:05:49 +0000 (16:05 +0100)]
build: add defines for compatibility with make build
The defines used to indicate what crypto, compression and eventdev drivers
were being built were different to those used in the make build, with meson
defining them with "_PMD" at the end, while make defined them with "_PMD"
in the middle and the specific driver name at the end. This might cause
compatibility issues for applications which used the older defines, which
switching to build against new DPDK releases.
As well as changing the default to match that of make, meson also
special-cases the crypto/compression/event drivers to have both defines
provided. This ensures compatibility for these macros with both meson and
make from older versions.
For a selection of other libraries and drivers, there were other
incompatibilities between the meson and make-defined macros which were not
previously highlighted in a deprecation notice, so we add per-macro
compatibility defines for these to ease the transition from make to meson.
Fixes: 5b9656b157d3 ("lib: build with meson") Fixes: 9314afb68a53 ("drivers: add infrastructure for meson build") Fixes: dcadbbde8e61 ("crypto/null: build with meson") Fixes: 3c32e89f68e1 ("compress/isal: add skeleton ISA-L compression PMD") Fixes: eca504f318db ("drivers/event: build skeleton and SW drivers with meson") Cc: stable@dpdk.org Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Luca Boccassi <bluca@debian.org>
Bruce Richardson [Thu, 15 Oct 2020 15:05:47 +0000 (16:05 +0100)]
app: fix missing dependencies
A number of lib and driver dependencies for various apps were missed on
build because the proper macro names for their use were mismatched between
meson and make build systems. Before adding in equivalent compatibility
macros we need to ensure to add the proper dependencies to ensure a valid
build.
Fixes: 16ade738fd0d ("app/testpmd: build with meson") Fixes: b5dc795a8a55 ("test: build app with meson as dpdk-test") Fixes: 996ef1176111 ("app: add all remaining apps to meson build") Cc: stable@dpdk.org Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Luca Boccassi <bluca@debian.org>
Bruce Richardson [Thu, 10 Sep 2020 11:32:05 +0000 (12:32 +0100)]
doc: add SPDX license tag header to meson guide
The build-sdk-meson.rst file originates from the short plain-text meson
instructions added in 2018. Add SPDX tag and copyright notice based on the
original commit.
Fixes: 9c3adc289c5e ("doc: add instructions on build using meson") Cc: stable@dpdk.org Reported-by: Stephen Hemminger <stephen@networkplumber.org> Signed-off-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Ciara Power [Tue, 13 Oct 2020 15:56:48 +0000 (16:56 +0100)]
test/telemetry: fix socket resource leak
If an error occurred when reading from the socket, the function
returned without closing the socket. This is now fixed to avoid the
resource leak of the sock variable going out of scope.
Coverity issue: 363043 Fixes: bd78cf693ebd ("test/telemetry: add unit tests for data to JSON") Signed-off-by: Ciara Power <ciara.power@intel.com> Acked-by: Kevin Laatz <kevin.laatz@intel.com>
Ciara Power [Tue, 13 Oct 2020 16:28:20 +0000 (17:28 +0100)]
test/telemetry: fix error handling for socket
When the socket connection failed, an error was printed to screen but
the function did not return an error, and continued to try read from the
socket. This is now corrected to close the socket and return -1 when the
connection fails.
Fixes: bd78cf693ebd ("test/telemetry: add unit tests for data to JSON") Signed-off-by: Ciara Power <ciara.power@intel.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Ciara Power [Mon, 19 Oct 2020 13:48:58 +0000 (15:48 +0200)]
acl: check max SIMD bitwidth
When choosing a vector path to take, an extra condition must be
satisfied to ensure the max SIMD bitwidth allows for the CPU enabled
path. These checks are added in the check alg helper functions.
Signed-off-by: Ciara Power <ciara.power@intel.com> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com> Tested-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Ciara Power [Mon, 19 Oct 2020 13:48:57 +0000 (15:48 +0200)]
node: choose vector path at runtime
When choosing the vector path, max SIMD bitwidth is now checked to
ensure the vector path is suitable. To do this, the scalar function is
chosen by default in the struct, but at node initialisation time, this
function pointer is updated to the vector version if supported, and
if it is within the max SIMD bitwidth limit.
Signed-off-by: Ciara Power <ciara.power@intel.com> Acked-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Ciara Power [Mon, 19 Oct 2020 13:48:56 +0000 (15:48 +0200)]
net: check max SIMD bitwidth
When choosing a vector path to take, an extra condition must be
satisfied to ensure the max SIMD bitwidth allows for the CPU enabled
path.
The vector path was initially chosen in RTE_INIT, however this is no
longer suitable as we cannot check the max SIMD bitwidth at that time.
Default handlers are now chosen on initialisation, these default
handlers are used the first time the crc calc is called, and they set
the suitable handlers to be used going forward.
Ciara Power [Mon, 19 Oct 2020 13:48:50 +0000 (15:48 +0200)]
net/ixgbe: check max SIMD bitwidth
When choosing a vector path to take, an extra condition must be
satisfied to ensure the max SIMD bitwidth allows for the CPU enabled
path.
Signed-off-by: Ciara Power <ciara.power@intel.com> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com> Reviewed-by: Haiyue Wang <haiyue.wang@intel.com>
Ciara Power [Mon, 19 Oct 2020 13:48:41 +0000 (15:48 +0200)]
eal: control max SIMD bitwidth
This patch adds a max SIMD bitwidth EAL configuration. The API allows
for an app to set this value. It can also be set using EAL argument
--force-max-simd-bitwidth, which will lock the value and override any
modifications made by the app.
Each arch has a define for the default SIMD bitwidth value, this is used
on EAL init to set the config max SIMD bitwidth.
Signed-off-by: Ciara Power <ciara.power@intel.com> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com> Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com> Acked-by: Ray Kinsella <mdr@ashroe.eu>
Gagandeep Singh [Mon, 19 Oct 2020 09:48:01 +0000 (17:48 +0800)]
crypto/caam_jr: fix device tree parsing for SEC_ERA
Previously, SEC_ERA was hardcoded and it was removed in [1].
Now when that hardcoded was removed, it is supposed to be
read from the device tree but it is not done correctly.
This patch calls a necessary API of_init() before using any
of_* APIs to retrieve information from the device tree and
if reading integer value that must be converted to cpu endianness
before using it.
[1] eef9e0412a84 ("drivers/crypto: fix build with -fno-common")
Jun Yang [Thu, 15 Oct 2020 09:47:06 +0000 (17:47 +0800)]
raw/dpaa2_qdma: support scatter gather in enqueue
This patch add support to add Scatter Gather support
for different jobs for qdma queues.
It also supports gathering multiple enqueue jobs into SG enqueue job(s).
Jun Yang [Thu, 15 Oct 2020 09:47:04 +0000 (17:47 +0800)]
raw/dpaa2_qdma: refactor the code
This patch moves qdma queue specific configurations from driver
global configuration to per-queue setup. This is required
as each queue can be configured differently.
eal/linux: add function to allow interruptible epoll
The existing definition of rte_epoll_wait retries if interrupted
by a signal. This behavior makes it hard to use rte_epoll_wait
for applications that want to use signals do do things like
exit polling loop and shutdown.
Since changing existing semantic might break applications, add
a new rte_epoll_wait_interruptible() function that does the
same thing as rte_epoll_wait but will return -1 and errno of EINTR
if it receives a signal.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Reviewed-by: Harman Kalra <hkalra@marvell.com>
Sarosh Arif [Tue, 8 Sep 2020 10:22:04 +0000 (15:22 +0500)]
test/distributor: fix mbuf leak on failure
rte_mempool_get_bulk is used to get bufs/many_bufs from the pool,
but at some locations when test fails the bufs/many_bufs are
not returned back to the pool.
Due to this, multiple executions of distributor_autotest gives the
following error message: Error getting mbufs from pool.
To resolve this issue rte_mempool_put_bulk is used whenever the test
fails and returns.
Fixes: c3eabff124e6 ("distributor: add unit tests") Cc: stable@dpdk.org Signed-off-by: Sarosh Arif <sarosh.arif@emumba.com> Acked-by: Lukasz Wojciechowski <l.wojciechow@partner.samsung.com> Reviewed-by: David Hunt <david.hunt@intel.com>
test/distributor: fix quitting workers in burst mode
Sending number of packets equal to number of workers isn't enough
to stop all workers in burst version of distributor as more than
one packet can be matched and consumed by a single worker. This way
some of workers might not be awaken from rte_distributor_get_pkt().
This patch fixes it by sending packets one by one. Each sent packet
causes exactly one worker to quit.
test/distributor: ensure all packets are delivered
In all distributor tests there is a chance that tests
will send packets to distributor with rte_distributor_process()
before workers are started and requested for packets.
This patch ensures that all packets are delivered to workers
by calling rte_distributor_process() in loop until number
of successfully processed packets reaches required by test.
Change is applied to every first call in test case.
Cc: stable@dpdk.org Signed-off-by: Lukasz Wojciechowski <l.wojciechow@partner.samsung.com> Acked-by: David Hunt <david.hunt@intel.com>
rte_distributor_flush() is using total_outstanding()
function to calculate if it should still wait
for processing packets. However in burst mode
only backlog packets were counted.
This patch fixes that issue by counting also in flight
packets. There are also sum fixes to properly keep
count of in flight packets for each worker in bufs[].count.
Fixes: 775003ad2f96 ("distributor: add new burst-capable library") Cc: stable@dpdk.org Signed-off-by: Lukasz Wojciechowski <l.wojciechow@partner.samsung.com> Acked-by: David Hunt <david.hunt@intel.com>
git.droids-corp.org / dpdk.git / log