net/sfc: ensure that device is closed on removal

PCI device may be unplugged when ethdev is not closed yet.

Fixes: aaa3f5f0f79d ("net/sfc: add configure and close stubs")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ivan Malov <ivan.malov@oktetlabs.ru>

ethdev: avoid error on PCI unplug of closed port

If PCI Ethernet device driver removes it on close
(RTE_ETH_DEV_CLOSE_REMOVE) and later PCI device itself is unplugged,
it should not fail because of Ethernet device is already removed.

Fixes: 23ea57a2a0ce ("ethdev: complete closing of port")
Cc: stable@dpdk.org
Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Reported-by: David Marchand <david.marchand@redhat.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

doc: fix build with latest meson

The latest versions of meson don't build targets when build_by_default is
false but install is true, unlike older versions. We can fix this by having
both build_by_default and install settings come from the build-time option.

Bugzilla ID: 303
Fixes: d02a2dab2dfb ("doc: support building HTML guides with meson")
Fixes: 720b14db3ae2 ("build: generate API documentation with meson")
Cc: stable@dpdk.org
Reported-by: Ali Alnubani <alialnu@mellanox.com>
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Luca Boccassi <bluca@debian.org>

timer: fix resource leak in finalize

Currently, whenever timer library is initialized, the memory
is leaked because there is no telling when primary or secondary
processes get to use the state, and there is no way to
initialize/deinitialize timer library state without race
conditions [1] because the data itself must live in shared memory.

Add a spinlock to the shared mem config to have a way to
exclusively initialize/deinitialize the timer library without
any races, and implement the synchronization mechanism based
on this lock in the timer library.

Also, update the API doc. Note that the behavior of the API
itself did not change - the requirement to call init in every
process was simply not documented explicitly.

[1] See the following email thread:
https://mails.dpdk.org/archives/dev/2019-May/131498.html

Fixes: c0749f7096c7 ("timer: allow management in shared memory")
Cc: stable@dpdk.org
Signed-off-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com>
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Erik Gabriel Carrillo <erik.g.carrillo@intel.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>

eal: prevent different primary/secondary process versions

Currently, nothing stops DPDK to attempt to run primary and
secondary processes while having different versions. This
can lead to all sorts of weird behavior and makes it harder
to maintain compatibility without breaking ABI every once
in a while.

Fix it by explicitly disallowing running different DPDK
versions as primary and secondary processes.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: David Marchand <david.marchand@redhat.com>

eal: unify internal config init

Currently, each EAL will update internal/shared config in their
own way at init, resulting in needless duplication of code and
OS-dependent behavior. Move the functions to a common file and
add missing FreeBSD steps.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>

eal: unify wait for complete init

Currently, mcfg completion function exists in two independent
implementations doing the same thing, which is bug prone.
Unify the two functions and move them into one place.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>

eal: uninline wait for complete init

Currently, the function to wait until config completion is
static inline for no reason. Move its implementation to
an EAL common file.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>

eal: remove packed attribute from mcfg structure

There is no reason to pack the memconfig structure, and doing so
gives out warnings in some static analyzers. Fix it by removing
the packed attributed.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>

eal: hide shared memory config

Now that everything that has ever accessed the shared memory
config is doing so through the public API's, we can make it
internal. Since we're removing quite a few headers from
rte_eal_memconfig.h, we need to add them back in places
where this header is used.

This bumps the ABI, so also change all build files and make
update documentation.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>

eal: add API to lock/unlock mempool list

Currently, in order to lock access to the mempool list, a direct
access to the shared memory structure is needed. Add an API to do
the same, and search-and-replace all usages.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>

eal: add API to lock/unlock tailq list

Currently, locking/unlocking the TAILQ list requires direct
access to the shared memory config. Add an API to do the same,
and search-and-replace all usages.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>

mem: add API to lock/unlock memory hotplug

Currently, the memory hotplug is locked automatically by all
memory-related _walk() functions, but sometimes locking the
memory subsystem outside of them is needed. There is no
public API to do that, so it creates a dependency on shared
memory config to be public. Fix this by introducing a new
API to lock/unlock the memory hotplug subsystem.

Create a new common file for all things mem config, and a
new API namespace rte_mcfg_*, and search-and-replace all
usages of the locks with the new API.

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>

bus/pci: consider only usable devices for IOVA mode

When selecting the preferred IOVA mode of the pci bus, the current
heuristic ("are devices bound?", "are devices bound to UIO?", "are pmd
drivers supporting IOVA as VA?" etc..) should honor the device
white/blacklist so that an unwanted device does not impact the decision.

There is no reason to consider a device which has no driver available.

This applies to all OS, so implements this in common code then call a
OS specific callback.

On Linux side:
- the VFIO special considerations should be evaluated only if VFIO
  support is built,
- there is no strong requirement on using VA rather than PA if a driver
  supports VA, so defaulting to DC in such a case.

Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>

eal: compute IOVA mode based on PA availability

Currently, if the bus selects IOVA as PA, the memory init can fail when
lacking access to physical addresses.
This can be quite hard for normal users to understand what is wrong
since this is the default behavior.

Catch this situation earlier in eal init by validating physical addresses
availability, or select IOVA when no clear preferrence had been expressed.

The bus code is changed so that it reports when it does not care about
the IOVA mode and let the eal init decide.

In Linux implementation, rework rte_eal_using_phys_addrs() so that it can
be called earlier but still avoid a circular dependency with
rte_mem_virt2phys().
In FreeBSD implementation, rte_eal_using_phys_addrs() always returns
false, so the detection part is left as is.

If librte_kni is compiled in and the KNI kmod is loaded,
- if the buses requested VA, force to PA if physical addresses are
  available as it was done before,
- else, keep iova as VA, KNI init will fail later.

Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

kni: abort when IOVA is not PA

If a forced iova-mode has been passed at init, kni is not supposed to
work.

Fixes: 075b182b54ce ("eal: force IOVA to a particular mode")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>

mem: fix typo in API description

Fixes: 552afc420a67 ("mem: add contig walk function")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

raw/ioat: fix icc build

When using _mm_set_epi64() rather than _mm_set_epi64x() intrinsic, ICC
tries to use the x87 floating point registers, leading to warnings about
not properly clearing value when switching between x87 and SSE/AVX modes.

error #13203: No EMMS instruction before call to function

Fix this by using the set64x() intrinsic.

Fixes: 0a92e63fc4cd ("raw/ioat: add local API to perform copies")

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

examples/ntb: add example for NTB

Enable an example for rawdev ntb. Support interactive mode to send
file on one host and receive file from another host. The command line
would be 'send [filepath]' and 'receive [filepath]'.

But since the FIFO is not enabled right now, use rte_memcpy as the enqueue
and dequeue functions and only support transmitting file no more than 4M.

Signed-off-by: Xiaoyun Li <xiaoyun.li@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>

raw/ntb: add handshake process

Add handshake process using doorbell so that two hosts can
communicate to start and stop.

Signed-off-by: Xiaoyun Li <xiaoyun.li@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>

raw/ntb: support Intel NTB

Add in the list of registers for the device.
And enable NTB device ops for Intel Skylake platform.

Signed-off-by: Xiaoyun Li <xiaoyun.li@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>

raw/ntb: introduce NTB raw device driver

Introduce rawdev driver support for NTB (Non-transparent Bridge) which
can help to connect two separate hosts with each other.

Signed-off-by: Xiaoyun Li <xiaoyun.li@intel.com>
Acked-by: Jingjing Wu <jingjing.wu@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>

usertools: add octeontx2 DMA device binding

Update the devbind script with new section of DMA devices, also
added OCTEONTX2 DMA device ID to DMA device list

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Satha Rao <skoteshwar@marvell.com>

raw/octeontx2_dma: add driver self test

Sample test to verify DMA functionality, this test covers
internal transfer mode.

Signed-off-by: Satha Rao <skoteshwar@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

raw/octeontx2_dma: add dequeue and device control operations

Add dequeue, device start, stop, close and reset operations.

Signed-off-by: Satha Rao <skoteshwar@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

raw/octeontx2_dma: add enqueue operation

Add enqueue operation.

Signed-off-by: Satha Rao <skoteshwar@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

raw/octeontx2_dma: add device close operation

Send message to PF to stop DMA queue when device close is
called from application.
Defined the required data structures to support enqueue and
dequeue APIs.

Signed-off-by: Satha Rao <skoteshwar@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

raw/octeontx2_dma: add device configuration

Register dev_configure API to configure DPI PCI devices.
After successful initialization send message to PF to open
corresponding DPI DMA queue. At present hardware doesn't
support mail box for DPI, so PMD to PF communication uses
pre build kernel devfs.

Signed-off-by: Satha Rao <skoteshwar@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

raw/octeontx2_dma: update probe function

Probe function enhanced to allocate and initialize PMD private data.

Signed-off-by: Satha Rao <skoteshwar@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

raw/octeontx2_dma: add build infra and device probe

Add the make and meson based build infrastructure along
with the DMA device probe with documentation infrastructure.

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Satha Rao <skoteshwar@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

mempool/octeontx2: fix clang build for arm64

The ARMv8.1 CASP instruction works with even register pairs and since
there no register constraint in older versions of GCC/Clang, use
explicit register allocation to satisfy CASP requirements.
Remove function level optimization specification as Clang doesn't have
support for it and explicit register allocation removes the need for it.

Fixes build issue with arm64-armv8a-linux-clang.

Fixes: ee338015e7a9 ("mempool/octeontx2: add optimized dequeue operation for arm64")

Reported-by: Gavin Hu <gavin.hu@arm.com>
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

test: fix build without timer library

The action for timer_secondary_spawn_wait should be enabled only when
CONFIG_RTE_LIBRTE_TIMER is enabled.

Fixes: 50247fe03fe0 ("test/timer: exercise new APIs in secondary process")
Cc: stable@dpdk.org
Signed-off-by: Timothy Redaelli <tredaelli@redhat.com>

test/hash: use array for small amount of memory

Variables of size 128B can make use of stack instead of dynamically
allocated memory.

Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>

test/hash: free allocated memory

Free allocated memory.

Fixes: 3f9aab961ed3 ("test/hash: check lock-free extendable bucket")
Cc: stable@dpdk.org
Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Dharmik Thakkar <dharmik.thakkar@arm.com>

test/hash: init parameters in the correct function

Each test case initializes its hash parameters in the test case
function. To be consistent, generate keys function should initialize
hash parameters similarly.

Fixes: c7eb0972e74b ("test/hash: add lock-free r/w concurrency")
Cc: stable@dpdk.org
Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Dharmik Thakkar <dharmik.thakkar@arm.com>

test/hash: fix data reset on new run

Reset tbl_rwc_test_param to discard data from previous run
of the test.

Fixes: c7eb0972e74b ("test/hash: add lock-free r/w concurrency")
Cc: stable@dpdk.org
Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Dharmik Thakkar <dharmik.thakkar@arm.com>

test/power: add cases for turbo feature

Add UT check_power_turbo.

Signed-off-by: Lukasz Krakowiak <lukaszx.krakowiak@intel.com>
Acked-by: David Hunt <david.hunt@intel.com>
Tested-by: Marcin Hajkowski <marcinx.hajkowski@intel.com>

test/power: remove prefix ACPI

This patch removes prefix _acpi from power UT function/test names,
and renames file test_power_acpi_cpufreq.c ->
 app/test/test_power_cpufreq.c.

Signed-off-by: Lukasz Krakowiak <lukaszx.krakowiak@intel.com>

test/eal: add cases for options proc-type and uio

Added unit test cases for EAL flags --proc-type=auto and
--create-uio-dev in order to cover the below functions
eal_proc_type_detect()
rte_eal_create_uio_dev()

Signed-off-by: Lavanya Govindarajan <lavanyax.govindarajan@intel.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

test/eal: add cases for in-memory and single-file-segments

Added unit test case for eal command line '--in-memory' option
which will cover below functions.
get_seg_memfd()
test_memfd_create()
pagesz_flags()

Added unit test case for eal command line '--single-file-segments' option
which will cover resize_hugefile().

Signed-off-by: Pallantla Poornima <pallantlax.poornima@intel.com>
Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>

test: fix autotest crash

On some systems when dpdk test is executed with make test command
autotest_runner crashes in first_cpu_on_node. This happens when list
of available cpus contains something that is not a cpu as first element.
Fixed by removing all non-cpu values from list of available cpus.

Bugzilla ID: 253
Fixes: 22dcd9a4d90f ("test: parallelize unit tests")
Cc: stable@dpdk.org
Signed-off-by: Herakliusz Lipiec <herakliusz.lipiec@intel.com>
Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>

table: fix crash in LPM IPv6

Unit test table_autotest results in segmentation fault.
Crash occurs in test_table_lpm_ipv6_combined().

Variable 'nht_pos0' used as array subscript is not initialized
in rte_table_lpm_ipv6_entry_add(). It will not be assigned,
if a rule does not exist.

In such case a junk number or invalid array index might result in
segmentation fault due to array out of bounds when
lpm->nht_users is used with such invalid array index.

Fix is to initialize the variables used for array subscript.

Bugzilla ID: 285
Fixes: d89a5bce1d ("lpm6: extend next hop field")
Cc: stable@dpdk.org
Signed-off-by: Jananee Parthasarathy <jananeex.m.parthasarathy@intel.com>
Tested-by: David Marchand <david.marchand@redhat.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>

usertools: add baseband device binding

Allows binding of baseband devices

Signed-off-by: Nicolas Chautru <nicolas.chautru@intel.com>
Acked-by: Amr Mokhtar <amr.mokhtar@intel.com>

doc: add a note for multi-process in mempool guide

The mempool library assigns handler ops indexes based on the dynamic load
order of mempool handlers. Indexes are used so a mempool can be used by
multiple processes, but this only works if all processes agree on the
mapping from index to mempool handler.

When using the '-d' argument, it's possible for different processes to load
mempool handlers in different orders, and thus have different
index->handler mappings. Using a mempool in multiple of such processes will
result in undefined behavior.

This commit adds a note to the mempool library programmer's guide warning
users against this.

Fixes: 449c49b93a6b ("mempool: support handler operations")
Cc: stable@dpdk.org
Signed-off-by: Gage Eads <gage.eads@intel.com>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>

usertools: fix refresh binding infos

Currently clear_data (dpdk-devbind.py) doesn't work as expected
since "global devices" is missing and so "devices" is considered
a local variable.

This commit changes "clear_data" function in order to really clear
devices by adding "global devices".

Fixes: ea9f00f7289a ("usertools: refactor NIC and crypto binding details")
Cc: stable@dpdk.org
Signed-off-by: Timothy Redaelli <tredaelli@redhat.com>
Acked-by: Luca Boccassi <bluca@debian.org>
Reviewed-by: David Marchand <david.marchand@redhat.com>

devtools: pass custom options to checkpatch

Add the ability to pass custom options to checkpatch script. An example
of use is to change the output format so it can run in emacs compilation
mode:

  DPDK_CHECKPATCH_PATH=/path/to/linux/scripts/checkpatch.pl \
    DPDK_CHECKPATCH_OPTIONS='--emacs --showfile --no-color' \
    /path/to/dpdk.org/devtools/checkpatches.sh

Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

examples/power: remove double copy of FIFO path

Removed doubled created fifo path string for channel info.

Signed-off-by: Lukasz Krakowiak <lukaszx.krakowiak@intel.com>
Signed-off-by: Lukasz Gosiewski <lukaszx.gosiewski@intel.com>
Acked-by: David Hunt <david.hunt@intel.com>

telemetry: fix build with gcc 9

Suppress the unaligned packed member address warnings by extending
the telemetry library build flags with -Wno-address-of-packed-member
option, through the WERROR_FLAGS makefile variable.

With this change additional warnings are turned on to be treated as errors,
which causes the following build issues to be seen:
- no previous prototype [-Werror=missing-prototypes]
- initialization discards ‘const’ qualifier from pointer target type
  [-Werror=discarded-qualifiers]
- old-style function definition [-Werror=old-style-definition]
- variable may be used before its value is set (when using icc compiler).

Fixes: 0fe3a37924d4 ("telemetry: format json response when sending stats")
Fixes: ee5ff0d3297e ("telemetry: add client feature and sockets")
Fixes: 8877ac688b52 ("telemetry: introduce infrastructure")
Fixes: 1b756087db93 ("telemetry: add parser for client socket messages")
Fixes: fff6df7bf58e ("telemetry: fix using ports of different types")
Fixes: 4080e46c8078 ("telemetry: support global metrics")
Cc: stable@dpdk.org
Signed-off-by: Reshma Pattan <reshma.pattan@intel.com>
Signed-off-by: Flavia Musatescu <flavia.musatescu@intel.com>
Acked-by: Kevin Laatz <kevin.laatz@intel.com>

config: increase maximum lcores for ppc

Setting RTE_MAX_LCORE to reflect the largest available configuration.

Signed-off-by: David Wilder <dwilder@us.ibm.com>
Acked-by: David Christensen <drc@linux.vnet.ibm.com>

config: update for ppc build with meson

Adding defines for missing RTE_MACHINE_CPUFLAGs.

Signed-off-by: David Wilder <dwilder@us.ibm.com>
Acked-by: David Christensen <drc@linux.vnet.ibm.com>

mempool/dpaa: fix leak in pool creation failure

When 'rte_zmalloc' failed dpaa_mbuf_create_pool() forgets freeing
'bp' thus leading resource leak. This patch avoids this.

Coverity issue: 337679

Signed-off-by: Li Qiang <liq3ea@163.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>

event/octeontx2: add libatomic dependency for 32-bit clang

When compiling with clang on 32-bit platforms, we are missing
copies of 64-bit atomic functions. We can solve this by linking
against libatomic for the drivers and libs which need those
atomic ops.

Fixes: f0b9982cb3a7 ("event/octeontx2: add TIM bucket operations")

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

drivers/octeontx2: fix icc build for i686

Fix the following ICC specific compilation issue with i686 build.

dpdk/drivers/common/octeontx2/otx2_mbox.c(47): error #2259:
non-pointer conversion from "unsigned long long" to "struct mbox_hdr *"
may lose significant bits
(struct mbox_hdr *)((uintptr_t)mdev->mbase + mbox->tx_start);

Fixes: 732377a6792e ("mempool/octeontx2: add build infra and device probe")
Fixes: 371d3212cbed ("common/octeontx2: add build infrastructure and HW definition")

Signed-off-by: Amit Gupta <agupta3@marvell.com>

net: forbid VLAN insert in shared mbuf

The vlan_insert() is buggy when it tries to handle the shared mbufs,
instead don't support inserting VLAN tag into shared mbufs and return
an error for that case.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

raw/ifpga: fix unintentional integer overflow

cast unsigned int ports_per_retimer, unsigned int nums_retimer,
unsigned int nums_fvl and unsigned int ports_per_fvl to uint64_t
before multiply operation, to avoid Unintentional integer overflow.

Coverity issue: 337924, 337926
Fixes: d1cd4eb2d48e ("raw/ifpga: support ipn3ke")
Cc: stable@dpdk.org
Signed-off-by: Andy Pei <andy.pei@intel.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>

raw/ifpga/base: fix retimer link status

Fix the readout retimer link status incorrectly when we
remove the linux intel-fpga-driver and run the DPDK application.
The linux driver will stop the retimer when remove the kernel
modules.

Fixes: 8a256bef ("raw/ifpga/base: add eth group driver")
Cc: stable@dpdk.org
Reported-by: Amrutha Sampath <amrutha.sampath@intel.com>
Signed-off-by: Tianfei Zhang <tianfei.zhang@intel.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>

raw/ifpga/base: fix physical address info

Fix miss phy_addr on ifpga_acc_get_region_info() function.

Fixes: 56bb54ea1bd ("raw/ifpga/base: add Intel FPGA OPAE share code")
Cc: stable@dpdk.org
Signed-off-by: Tianfei Zhang <tianfei.zhang@intel.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>

raw/ifpga/base: fix bit fields definition

Fix CTRL_DEV_SELECT bit fields definition about eth_group devices.

Fixes: 8a256bef32 ("raw/ifpga/base: add eth group driver")
Cc: stable@dpdk.org
Signed-off-by: Tianfei Zhang <tianfei.zhang@intel.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>

raw/ifpga/base: fix logically dead code

add temporary variable in max10_reg_write().

Coverity issue: 337927
Fixes: 96ebfcf ("raw/ifpga/base: add SPI and MAX10 device driver")
Cc: stable@dpdk.org
Signed-off-by: Tianfei Zhang <tianfei.zhang@intel.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>

raw/ifpga/base: fix use of untrusted scalar value

Add checking the buffer size and use
const char * for buffer declaration.

Coverity issue: 279449
Fixes: ef1e8ede ("raw/ifpga: add Intel FPGA bus rawdev driver")
Cc: stable@dpdk.org
Signed-off-by: Tianfei Zhang <tianfei.zhang@intel.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>

test/bpf: add test-case for function return value

New test-case to cover situation when external function returns a
pointer the data.

Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

bpf: fix validate for function return value

eval_call() blindly calls eval_max_bound() for external function
return value for all return types.
That causes wrong estimation for returned pointer min and max boundaries.
So any attempt to dereference that pointer value causes verifier to fail
with error message: "memory boundary violation at pc: ...".
To fix - estimate min/max boundaries based on the return value type.

Bugzilla ID: 298

Fixes: 8021917293d0 ("bpf: add extra validation for input BPF program")
Cc: stable@dpdk.org
Reported-by: Michel Machado <michel@digirati.com.br>
Suggested-by: Michel Machado <michel@digirati.com.br>
Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

raw/ioat: add local API to perform copies

Add local APIs to trigger data copies, and retrieve handle values once
those copies are completed. Included are unit tests to validate the data
is copies correctly.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>
Tested-by: Harry van Haaren <harry.van.haaren@intel.com>

raw/ioat: add statistics functions

Add stats functions to track what is happening in the driver, and put
unit tests to check those.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>
Tested-by: Harry van Haaren <harry.van.haaren@intel.com>

raw/ioat: add configure, start and stop functions

Allow initializing a driver instance. Include selftest to validate these
functions.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>
Tested-by: Harry van Haaren <harry.van.haaren@intel.com>

raw/ioat: add device info function

Add in the "info_get" function to the driver, to allow us to query the
device. This allows us to have the unit test pick up the presence of
supported hardware or not.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>
Tested-by: Harry van Haaren <harry.van.haaren@intel.com>

raw/ioat: create device on probe and destroy on release

Add the create/destroy driver functions so that we can actually allocate
a rawdev and destroy it when done. No rawdev API functions are actually
implemented at this point.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>
Tested-by: Harry van Haaren <harry.van.haaren@intel.com>

raw/ioat: add register definition file

Add in the list of registers for the device. File is taken from the SPDK
project:

  https://github.com/spdk/spdk/blob/master/include/spdk/ioat_spec.h

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

usertools: support IOAT device binding

In order to allow binding/unbinding of devices for use by the
ioat_rawdev, we need to update the devbind script to add a new class
of device, and add device ids for the specific HW instances.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>
Tested-by: Harry van Haaren <harry.van.haaren@intel.com>

raw/ioat: introduce IOAT driver

Add stubs for ioat rawdev driver support in DPDK, specifically:

  * makefile and meson build hooks
  * initial public header file
  * rawdev main C file, with probe and release functions
  * release note update announcing the driver
  * initial documentation for the new section in the rawdev doc
  * unit test stubs for device unit tests

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Jiayu Hu <jiayu.hu@intel.com>
Tested-by: Harry van Haaren <harry.van.haaren@intel.com>

rawdev: allow devices to skip extra memory allocation

Some device drivers want to allocate their own private memory, and should
be allowed to do so. Therefore skip memory allocation and associated error
checks if zero-length private memory is requested.

While adjusting the code for new indent level, fix incorrect error
message.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Shreyansh Jain <shreyansh.jain@nxp.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>

devtools: fix lib directory in pkg-config test

With Debian and Ubuntu, the default installation path for the 64-bit
libraries is set to e.g. /usr/local/lib/x86_64-linux-gnu/, compared to
/usr/local/lib64 on Fedora and Redhat distributions. This causes issues
when using "pkg-config --define-prefix" since pkg-config assumes the prefix
to be the grandparent of where the .pc file is. On Ubuntu we then get the
cflags include path as being "/path/to/install-root/usr/local/lib/include"
i.e. with an extra "lib" in the path.

This issue only applies for test installs on Ubuntu and similar distros,
and is not a problem for regular installs since the --define-prefix
parameter would not be passed to pkg-config in those cases.

The workaround for this in our test build script is to explicitly make
"lib" the "libdir" setting for the install, overriding the distro-provided
default.

Fixes: 7f80a2102bbb ("devtools: test pkg-config file")

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Luca Boccassi <bluca@debian.org>

examples: fix pkg-config detection with older make

Make versions before 4.2 did not have support for the .SHELLSTATUS
variable, so use another method to detect shell success.

Fixes: 22119c4591a0 ("examples: use pkg-config in makefiles")
Cc: stable@dpdk.org
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Luca Boccassi <bluca@debian.org>

net: add definition for DSCP and ECN masks

Added definitions of DSCP and ECN masks.

Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

net: fix definition of IPv6 traffic class mask

IPv6 header TC field is 8 bits, not 4.

Fixes: 9b20c6e9028e ("net: add IPv6 header fields macros")
Cc: stable@dpdk.org
Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

net: move TCP flags from ethdev header

TCP flags were moved to the TCP header file from the Ethernet control
header file, and the RTE prefix was added to their names.

Missing TCP ECN flags were added.

The ALL mask did not include TCP ECN flags, so it was renamed to reflect
that it applies to N-tuple filtering only.

Updated other files affected by the renaming accordingly.

Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>

app/testpmd: fix offloads config

There is no need to use default offloads configuration
if offloads configuration has been pass down from upper layer.
The default offloads are overwritten if not zero.

Fixes: 5e91aeef218c ("app/testpmd: fix offload flags after port config")
Cc: stable@dpdk.org
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

net/mvneta: optimize checksum generation offload

For tx checksum offload it is not mandatory to provide
a valid packet_type in addition to a valid ol_flags.
This patch only use ol_flags information for this feature
as a performance improvement.

Signed-off-by: Liron Himi <lironh@marvell.com>
Reviewed-by: Yuri Chipchev <yuric@marvell.com>
Tested-by: Liron Himi <lironh@marvell.com>

net/mvneta: reset stats during device init

reset device statistics on device initialization

Signed-off-by: Yuri Chipchev <yuric@marvell.com>
Reviewed-by: Liron Himi <lironh@marvell.com>
Tested-by: Yuri Chipchev <yuric@marvell.com>

net/mvneta: fix ierror statistics

Error packets were counted twice due to use of redundant counters.

Fixes: e9d5faffd10c ("net/mvneta: support basic stats")
Cc: stable@dpdk.org
Signed-off-by: Liron Himi <lironh@marvell.com>
Reviewed-by: Yuri Chipchev <yuric@marvell.com>
Tested-by: Liron Himi <lironh@marvell.com>

net/ice: support UDP tunnelling port

Enabled UDP tunnel port add and delete functions.

Signed-off-by: Qiming Yang <qiming.yang@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>

net/ice: add generic flow API

This patch adds ice_flow_create, ice_flow_destroy,
ice_flow_flush and ice_flow_validate support,
these are used to handle all the generic filters.

Signed-off-by: Qiming Yang <qiming.yang@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>

net/ice: enable switch filter

The patch enables the backend of rte_flow. It transfers
rte_flow_xxx to device specific data structure and
configures packet process engine's binary classifier
(switch) properly.

Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>

net/ice/base: fix inner L2 offset in GRE dummy packet

The offset for the inner L2 header in the dummy GRE packet
was off by 2 bytes so updated the offset.

Fixes: 839c0a4b77e6 ("net/ice/base: enable additional switch rules")

Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com>
Signed-off-by: Leyi Rong <leyi.rong@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/ice/base: support ethertype switch filters

Add protocol definitions for ethertype.
Add ice_ethtype_hdr structure definition into the ice_prot_hdr union.
Add ethtype offsets into the training packet maps.

Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com>
Signed-off-by: Leyi Rong <leyi.rong@intel.com>
Tested-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/ice/base: support IPv6 based switch filters

- Add IPv6 switch rule support.
- Add IPv6 training packet.
- Correct name of IPv6 header variable.
- Fix enum values so that they point to the proper
  ice_prot_ext_tbl_entry field.

Signed-off-by: Kevin Scott <kevin.c.scott@intel.com>
Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com>
Signed-off-by: Leyi Rong <leyi.rong@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>

net/ipn3ke: modify AFU configurations

Modify AFU configurations for new BBS (Blue Bitstream) of A10 on N3000
card:
- AFU register access: RTL changes the UPL (User Programmable Logic
which is the container of vBNG IP) base address and the read/write
commands of register indirect access.
- Poll the INIT_STS register to wait for the vBNG IP and DDR reset
completion.
- Refine log for debug: print UPL_version not only for vBNG bit stream,
but also for other bit streams.

Fixes: c01c748e4ae6 ("net/ipn3ke: add new driver")
Cc: stable@dpdk.org
Signed-off-by: Dan Wei <dan.wei@intel.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>

net/nfb: support Silicom Mango card

Add support for Silicom FB2CGG3 smart NIC

Signed-off-by: Rastislav Cernay <cernay@netcope.com>

net/szedata2: support Silicom Mango card

Add support for Silicom FB2CGG3 smart NIC

Signed-off-by: Rastislav Cernay <cernay@netcope.com>
Acked-by: Jan Remes <remes@netcope.com>

event/octeontx2: add devargs to control timer adapters

Add devargs to control each event timer adapter i.e. TIM rings internal
parameters uniquely. The following dict format is expected
[ring-chnk_slots-disable_npa-stats_ena]. 0 represents default values.

Example:
--dev "0002:0e:00.0,tim_ring_ctl=[2-1023-1-0]"

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add devargs to limit timer adapters

Add devargs to limit the max number of TIM rings reserved on probe.
Since, TIM rings are HW resources we can avoid starving other
applications by not grabbing all the rings.

Example:
--dev "0002:0e:00.0,tim_rings_lmt=2"

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add timer adapter start and stop

Add event timer adapter start and stop functions.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add timer stats get and reset

Add event timer adapter statistics get and reset functions.
Stats are disabled by default and can be enabled through devargs.

Example:
--dev "0002:0e:00.0,tim_stats_ena=1"

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add timer cancel function

Add function to cancel event timer that has been armed.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add timer arm timeout burst

Add event timer arm timeout burst function.
All the timers requested to be armed have the same timeout.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add timer arm routine

Add event timer arm routine.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add TIM bucket operations

Add TIM bucket operations used for event timer arm and cancel.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add timer adapter info function

Add TIM event timer adapter info get function.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: allow adapters to resize inflight buffers

Add internal SSO functions to allow event adapters to resize SSO buffers
that are used to hold in-flight events in DRAM.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>

event/octeontx2: add TIM IRQ handlers

Register and implement TIM IRQ handlers for error interrupts

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>