cryptodev: document usage of digest-appended operations

This patch explains what are the conditions
and how to use digest appended for auth-cipher
operations.

Signed-off-by: Damian Nowak <damianx.nowak@intel.com>
Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>

crypto/openssl: fix free of asymmetric crypto keys

In case big number need to be freed, data it contains should
also be cleared before especially if it is critical data like
private keys.

Fixes: 3e9d6bd447fb ("crypto/openssl: add RSA and mod asym operations")
Cc: stable@dpdk.org
Signed-off-by: Arek Kusztal <arkadiuszx.kusztal@intel.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>

cryptodev: free memzone when releasing device

When a cryptodev is created in a primary process,
rte_cryptodev_data_alloc reserves a memzone.
However, this memzone was not released when the cryptodev
is uninitialized. After that, new cryptodev cannot be
created due to memzone name conflict.

This commit frees the memzone when a cryptodev is
uninitialized, fixing this bug. This approach is chosen
instead of keeping and reusing the old memzone, because
the new cryptodev could belong to a different NUMA socket.

Also, rte_cryptodev_data pointer is now properly recorded
in cryptodev_globals.data array.

Bugzilla ID: 105

Signed-off-by: Junxiao Shi <git@mail1.yoursunny.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

test/ipsec: fix destroy SA

Call rte_cryptodev_sym_session_clear() function.

Fixes: 05fe65eb66b2 ("test/ipsec: introduce functional test")
Cc: stable@dpdk.org
Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

test/ipsec: fix tear down

Set freed cop pointers to NULL
Set other freed pointers to NULL instead of 0

Fixes: 05fe65eb66b2 ("test/ipsec: introduce functional test")
Cc: stable@dpdk.org
Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

examples/ipsec-secgw: add bypass test case

Add simple test-case with all traffic in BYPASS mode.
Useful for some basic test of your network environment.

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

examples/ipsec-secgw: add multi-segment test cases

Enhance test scripts to support fragmentation/reassemble functionality.

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

examples/ipsec-secgw: support fragmentation and reassembly

Add optional ability to fragment packet bigger then mtu,
and reassemble fragmented packet.
To minimize possible performance effect, reassembly is
implemented as RX callback.
To support these features ipsec-secgw relies on librte_ipsec ability
to handle multi-segment packets.
Also when reassemble/fragmentation support is enabled, attached
crypto devices have to support 'In Place SGL' offload capability.
To enable/disable this functionality, two new optional command-line
options are introduced:
  --reassemble <val> - number of entries in reassemble table
  --mtu <val> - MTU value for all attached ports
As separate '--mtu' option is introduced, '-j <val>' option is now used
to specify mbuf data buffer size only.

Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

examples/ipsec-secgw: fix packet length

for packets smaller then 64B some NICs reports pkt_len=64B.
As ipsec-secgw (and librte_ipsec) relies on pkt_len value to determine
payload length, that causes problems for small packets.
To fix the issue, check that pkt_len matches values in IPv4/IPv6 header
and re-adjust pkt_len if necessary.

Fixes: 906257e965b7 ("examples/ipsec-secgw: support IPv6")
Fixes: d299106e8e31 ("examples/ipsec-secgw: add IPsec sample application")
Cc: stable@dpdk.org
Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

examples/ipsec-secgw: fix inline modes

Application ipsec-secgw is not working for IPv4 transport mode and for
IPv6 both transport and tunnel mode.

IPv6 tunnel mode is not working due to wrongly assigned fields of
security association patterns, as it was IPv4, during creation of
inline crypto session.

IPv6 and IPv4 transport mode is iterating through security capabilities
until it reaches tunnel, which causes session to be created as tunnel,
instead of transport. Another issue, is that config file does not
provide source and destination ip addresses for transport mode, which
are required by NIC to perform inline crypto. It uses default addresses
stored in security association (all zeroes), which causes dropped
packages.

To fix that, reorganization of code in create_session() is needed,
to behave appropriately to given protocol (IPv6/IPv4). Change in
iteration through security capabilities is also required, to check
for expected mode (not only tunnel).

For lack of addresses issue, some resolving mechanism is needed.
Approach is to store addresses in security association, as it is
for tunnel mode. Difference is that they are obtained from sp rules,
instead of config file. To do that, sp[4/6]_spi_present() function
is used to find addresses based on spi value, and then stored in
corresponding sa rule. This approach assumes, that every sp rule
for inline crypto have valid addresses, as well as range of addresses
is not supported.

New flags for ipsec_sa structure are required to distinguish between
IPv4 and IPv6 transport modes. Because of that, there is need to
change all checks done on these flags, so they work as expected.

Fixes: ec17993a145a ("examples/ipsec-secgw: support security offload")
Fixes: 9a0752f498d2 ("net/ixgbe: enable inline IPsec")
Cc: stable@dpdk.org
Signed-off-by: Mariusz Drost <mariuszx.drost@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>
Tested-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

net/ixgbe: fix IP type for crypto session

When ixgbe_crypto_add_sa() is called, it checks whether the ip type is
IPv6 or IPv4 to write correct addresses to the registers. Type itself
is never specified, and act as IPv4, which is the default value.
It causes lack of support for IPv6.

To fix that, ip type needs to be stored in device private data, based on
crypto session ip type field, before the checking is done.

Fixes: ec17993a145a ("examples/ipsec-secgw: support security offload")
Fixes: 9a0752f498d2 ("net/ixgbe: enable inline IPsec")
Cc: stable@dpdk.org
Signed-off-by: Mariusz Drost <mariuszx.drost@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Tested-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

app/compress-perf: fix reliance on integer endianness

Reliance on integer endianness (INCOMPATIBLE_CAST) in function
*parse_window_sz*.

Coverity issue: 328524
Fixes: e0b6287c035d ("app/compress-perf: add parser")
Cc: stable@dpdk.org
Signed-off-by: Artur Trybula <arturx.trybula@intel.com>
Acked-by: Shally Verma <shallyv@marvell.com>

app/crypto-perf: check lcore job failure

Currently, there's no return status check from lcore's jobs.
In case of fail - crypto-perf tool returns success anyway.
This patch adds such a detection and returns proper status at the end.

Fixes: ce8af1a4398d ("app/crypto-perf: wait for cores launched by app")
Cc: stable@dpdk.org
Signed-off-by: Tomasz Jozwiak <tjozwiakgm@gmail.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

app/crypto-perf: fix display once detection

This patch changes 'only_once' variable to 'display_once',
which should be atomic type due to fact, that all runner functions
can be executed in paraller way on different lcores.

Fixes: df52cb3b6e13 ("app/crypto-perf: move verify as single test type")
Cc: stable@dpdk.org
Signed-off-by: Tomasz Jozwiak <tjozwiakgm@gmail.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

crypto/octeontx: remove unsupported capabilities

Removing capabilities not supported with latest firmware.

Signed-off-by: Anoob Joseph <anoobj@marvell.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

baseband/fpga_lte_fec: add driver for FEC on FPGA

Supports for FEC 4G PMD Driver on FPGA card PAC N3000

Signed-off-by: Nicolas Chautru <nicolas.chautru@intel.com>
Acked-by: Kamil Chalupnik <kamilx.chalupnik@intel.com>

baseband/turbo_sw: support meson build

Turbo_sw PMD driver now building with meson/ninja
with or without SDK libraries.

Signed-off-by: Nicolas Chautru <nicolas.chautru@intel.com>
Acked-by: Kamil Chalupnik <kamilx.chalupnik@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

doc: update turbo_sw build steps in bbdev guide

The documentation is clarified to point to steps on building the
SDK libraries which are now publicly available:
https://software.intel.com/en-us/articles/flexran-lte-and-5g-nr-fec-software-development-kit-modules

Signed-off-by: Nicolas Chautru <nicolas.chautru@intel.com>
Acked-by: Kamil Chalupnik <kamilx.chalupnik@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

baseband/turbo_sw: allow to build without SDK dependency

Adding compile flag to allow to build the turbo_sw PMD
without dependency to have the SDK libraries installed.

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

test/ipsec: improve debug in group tests

Add dump_grp_pkt() function.
Dump invalid mbuf data in case of error in group tests 12 and 13.

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

test/ipsec: add extra test configuration

Add extra test configuration for BURST_SIZE packets

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

test/ipsec: fix log messages

Error log messages after the create_sa() function calls
were using the wrong function name.
The function name is replaced with create_sa.

Fixes: 05fe65eb66b2 ("test/ipsec: introduce functional test")
Cc: stable@dpdk.org
Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

ipsec: include high order bytes of ESN in packet length

When esn is used then high-order 32 bits are included in ICV
calculation however are not transmitted. Update packet length
to be consistent with auth data offset and length before crypto
operation. High-order 32 bits of esn will be removed from packet
length in crypto post processing.

Signed-off-by: Lukasz Bartosik <lbartosik@marvell.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

ipsec: support multi-segment packets

Add support for packets that consist of multiple segments.
Take into account that trailer bytes (padding, ESP tail, ICV)
can spawn across multiple segments.

Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

doc: clarify xform zeroing in crypto guide

Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

test/crypto: add NULL algo to loop test mechanism

Added NULL algo tests into loop test mechanism used
by block cipher tests as easier to extend there.
Included chain, cipher-only and auth-only use-cases.
Extended to cover out-of-place use-cases and use-cases
where data length is not an 8-byte multiple.

Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

examples/ipsec-secgw: add scapy based tests

Add new unittest-like mechanism which uses scapy to craft custom
packets and a set of assertions to check how ipsec-secgw example
application is processing them. Python3 with scapy module is
required by pkttest.sh to run test scripts.

A new mechanism is used to test IPv6 transport mode traffic with
header extensions (trs_ipv6opts.py).

Fix incomplete test log problem by disabling buffering of ipsec-secgw
standard output with stdbuf application.

Signed-off-by: Marcin Smoczynski <marcinx.smoczynski@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>
Tested-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

examples/ipsec-secgw: support IPv6 options

Using transport with IPv6 and header extensions requires calculating
total header length including extensions up to ESP header which is
achieved with iteratively parsing extensions when preparing traffic
for processing. Calculated l3_len is later used to determine SPI
field offset for an inbound traffic and to reconstruct L3 header by
librte_ipsec.

Signed-off-by: Marcin Smoczynski <marcinx.smoczynski@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>
Tested-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

ipsec: fix transport mode for IPv6 with extensions

Reconstructing IPv6 header after encryption or decryption requires
updating 'next header' value in the preceding protocol header, which
is determined by parsing IPv6 header and iteratively looking for
next IPv6 header extension.

It is required that 'l3_len' in the mbuf metadata contains a total
length of the IPv6 header with header extensions up to ESP header.

Fixes: 4d7ea3e1459b ("ipsec: implement SA data-path API")
Cc: stable@dpdk.org
Signed-off-by: Marcin Smoczynski <marcinx.smoczynski@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>
Tested-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

net: add IPv6 header extension parsing function

Introduce new function for IPv6 header extension parsing able to
determine extension length and next protocol number.

This function is helpful when implementing IPv6 header traversing.

Signed-off-by: Marcin Smoczynski <marcinx.smoczynski@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>
Tested-by: Konstantin Ananyev <konstantin.ananyev@intel.com>

crypto/aesni_mb: fix out-of-bounds access

This patch fixes the out-of-bounds coverity issue by adding
missed algorithms to the array.

Coverity issue: 337683
Fixes: c68d7aa354f6 ("crypto/aesni_mb: use architecture independent macros")
Cc: stable@dpdk.org
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>

cryptodev: add feature flags to disable

Adding a new field, ff_disable, to allow applications to control the
features enabled on the crypto device. This would allow for efficient
usage of HW/SW offloads.

Signed-off-by: Anoob Joseph <anoobj@marvell.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

maintainers: claim maintainership of bbdev

Reorg within Intel teams causing transfer of maintainership

Signed-off-by: Nicolas Chautru <nicolas.chautru@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>

cryptodev: add option to support both IV and J0 for GCM

This patch adds an option to support both IV (of all supported sizes)
and J0 when using Galois Counter Mode of crypto operation.

Signed-off-by: Arek Kusztal <arkadiuszx.kusztal@intel.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>
Acked-by: Anoob Joseph <anoobj@marvell.com>

crypto/virtio: check PCI config read

Fix unchecked return value issue for rte_pci_read_config.

Coverity issue: 302861
Fixes: 25500d4b8076 ("crypto/virtio: support device init")
Cc: stable@dpdk.org
Signed-off-by: Chenbo Xia <chenbo.xia@intel.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

app/compress-perf: print socket id

Added info about socket id used by compression perf
test process and by the driver

Signed-off-by: Tomasz Cel <tomaszx.cel@intel.com>
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Acked-by: Artur Trybula <arturx.trybula@intel.com>

doc: clarify data plane error handling in compressdev

Fixed some typos and clarified how errors on ops which
fail to get submitted on the enqueue API should be handled.

Fixes: a584d3bea902 ("doc: add compressdev library guide")
Cc: stable@dpdk.org
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Acked-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Acked-by: Shally Verma <shallyv@marvell.com>

examples/fips_validation: fix logically dead code

Coverity issue: 336866, 336841, 336838
Fixes: 41d561cbdd24 ("examples/fips_validation: add power on self test")
Cc: stable@dpdk.org
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
Acked-by: Marko Kovacevic <marko.kovacevic@intel.com>

crypto/openssl: fix usage of non constant time memcmp

ANSI C memcmp is not constant time function per spec so it should
be avoided in cryptography usage.

Fixes: d61f70b4c918 ("crypto/libcrypto: add driver for OpenSSL library")
Cc: stable@dpdk.org
Signed-off-by: Arek Kusztal <arkadiuszx.kusztal@intel.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>

crypto/octeontx: sync mbox changes from kernel driver

Synchronize mbox with latest changes in kernel driver

Signed-off-by: Lukasz Bartosik <lbartosik@marvell.com>
Signed-off-by: Anoob Joseph <anoobj@marvell.com>

app/crypto-perf: fix CSV format

The format for printing float is incorrect for the following fields
in pmd-cyclecount test: Cycles/Op,Cycles/Enq,Cycles/Deq.

Currently, the format is %.f3. This format will round off the number to
the nearest integer and append a 3 after that.

This patch changes the format to %.3f. This will print the number as a
floating point with a precision of 3 fractional digits.

Fixes: 96dfeb609be1 ("app/crypto-perf: add new PMD benchmarking mode")
Cc: stable@dpdk.org
Signed-off-by: Ankur Dwivedi <adwivedi@marvell.com>
Signed-off-by: Anoob Joseph <anoobj@marvell.com>

crypto/qat: set message field to zero in sym SGL case

Fixes: e09231eaa2af ("crypto/qat: add SGL capability")
Cc: stable@dpdk.org
Signed-off-by: Gosia Bakota <gosia.bakota@intel.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>

drivers: fix typo in NXP comments

Fix another copy'n'paste typo.

Fixes: 531b17a780dc ("bus/fslmc: add QBMAN driver to bus")
Fixes: 6ef6beca8844 ("crypto/caam_jr: add enqueue/dequeue operations")
Fixes: c3e85bdcc6e6 ("crypto/dpaa_sec: add crypto driver for NXP DPAA platform")
Cc: stable@dpdk.org
Signed-off-by: Thierry Herbelot <thierry.herbelot@6wind.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

doc: fix grammar in prog guides

A copy'n'paste typo was present in three guides.

Fixes: a584d3bea902 ("doc: add compressdev library guide")
Fixes: 4935e1e9f76e ("bbdev: introduce wireless base band device lib")
Fixes: 0318c02b57cf ("doc: add cryptodev chapter in prog guide")
Cc: stable@dpdk.org
Signed-off-by: Thierry Herbelot <thierry.herbelot@6wind.com>
Acked-by: Rami Rosen <ramirose@gmail.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>

doc: fix triplicated typo in prog guides

A bad formula was copied and pasted.

Fixes: 4935e1e9f76e ("bbdev: introduce wireless base band device lib")
Fixes: 0318c02b57cf ("doc: add cryptodev chapter in prog guide")
Fixes: a9bb0c44c775 ("doc: add rawdev library guide and doxygen page")
Cc: stable@dpdk.org
Signed-off-by: Thierry Herbelot <thierry.herbelot@6wind.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>

crypto/mvsam: fix typo in comment

Inverted letters.

Fixes: 8a61c83af2fa ("crypto/mrvl: add mrvl crypto driver")
Cc: stable@dpdk.org
Signed-off-by: Thierry Herbelot <thierry.herbelot@6wind.com>
Acked-by: Tomasz Duszynski <tdu@semihalf.com>

doc: cleanup test removal in armv8 and openssl guides

The perf tests were removed in a previous commit.

Fixes: 2ac67c32837a ("test/crypto: remove crypto perf tests")
Cc: stable@dpdk.org
Signed-off-by: Thierry Herbelot <thierry.herbelot@6wind.com>
Acked-by: Akhil Goyal <akhil.goyal@nxp.com>

app/compress-perf: fix improper use of negative value

test_data->input_data_sz is passed to a parameter that
cannot be negative.

Coverity issue: 328504
Fixes: b68a82425da4 ("app/compress-perf: add performance measurement")
Cc: stable@dpdk.org
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>

compress/isal: fix use after free

The release function was using qp->stream after already
releasing it and the null pointer checking was missing.

Also since rte_free(NULL) is a no-op, remove unnecessary
checks for NULL.

Coverity issure: 340860
Fixes: dc49e6aa4879 ("compress/isal: add ISA-L compression functionality")
Cc: stable@dpdk.org
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Lee Daly <lee.daly@intel.com>
Acked-by: Fiona Trahe <fiona.trahe@intel.com>

doc: update bnxt guide

- Include list of supported adapters.
- Include list of supported features.
- Document requirements for vector mode PMD.
- Remove obsolete limitation (scatter rx has been supported for
  some time).
- Fixed broken links.

Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>

app/testpmd: fix queue offload configuration

When adding offloads from commandline, not only port
related configuration bits should be set, but also queue
related offloads configuration bits, or it will cause error.
For example, test in this process for ixgbe:
(1)./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4
-- -i --portmask=0x1 --port-topology=loop --disable-crc-strip
(2)port stop all
(3)port config all crc-strip on
(4)port start all
we will see "Fail to configure port 0 rx queues" of warning info.

Fixes: 0074d02fca21 ("app/testpmd: convert to new Rx offloads API")
Cc: stable@dpdk.org
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>

app/testpmd: fix parsing RSS queue rule

Declare a local variable 'arg' to hold the pointer to ARGS_ENTRY_ARB,
otherwise, the pop 'arg' in 'parse_int' holds wrong data like 'arg->
offset = 5912737, arg->size = 0', this caused the parse failure.

If calling like 'push_args(ctx, ARGS_ENTRY_ARB..)', the below code for
assigning 'offset & size' will be missed for some gcc optimization.

parse_vc_action_rss_queue
    29c5: c5 f9 ef c0           vpxor  %xmm0,%xmm0,%xmm0
    29c9: 43 8d 44 24 68        lea    0x68(%r12,%r12,1),%eax
    29ce: 48 c7 44 24 20 00 00  movq   $0x0,0x20(%rsp)
    29d5: 00 00
    29d7: c5 f8 29 44 24 10     vmovaps %xmm0,0x10(%rsp)
    29dd: 89 44 24 18           mov    %eax,0x18(%rsp)
    29e1: 48 63 83 0c 01 00 00  movslq 0x10c(%rbx),%rax
    29e8: c7 44 24 1c 02 00 00  movl   $0x2,0x1c(%rsp)
    29ef: 00
    29f0: c5 f8 29 04 24        vmovaps %xmm0,(%rsp)

The above assembly code is from this modification with:
    gcc version 9.1.1 20190503 (Red Hat 9.1.1-1) (GCC)
    Fedora release 30 (Thirty)
    5.1.11-300.fc30.x86_64

Fixes: d0ad8648b1c5 ("app/testpmd: fix RSS flow action configuration")
Cc: stable@dpdk.org
Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>

net/mlx5: check memory allocation in flow creation

rte_calloc functions returns a non-null pointer in case of
success and null pointer in case of failure.

The return value should be checked and the function flow
should take that into consideration.

This patch adds a check for rte_calloc return value in function
flow_list_create.

Fixes: 84c406e74524 ("net/mlx5: add flow translate function")
Cc: stable@dpdk.org
Signed-off-by: Asaf Penso <asafp@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>

net/mlx5: fix condition for link update fallback

mlx5_link_update uses the newer ethtool command
ETHTOOL_GLINKSETTINGS to determine interface capabilities but falls
back to the older (deprecated) ETHTOOL_GSET command if the new
method fails for any reason.
The older method only supports reporting of capabilities up to 40G.

However, mlx5_link_update_unlocked_gs can return a failure for a
number of reasons (including the link being down).
Using the older method in cases of transient failure of the method
can result in reporting of reduced capabilities to the application.

The older method (mlx5_link_update_unlocked_gset) should only be
invoked if the newer method returns EOPNOTSUPP.

Fixes: 7d2e32f76cfc ("net/mlx5: fix ethtool link setting call order")
Cc: stable@dpdk.org
Reported-by: Srinivas Narayan <srinivas.narayan@att.com>
Signed-off-by: Asaf Penso <asafp@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>

net/mlx5: remove TCF support

This commit removes the support of configuring the device E-switch
using TCF since it is now possible to configure it via DR (direct
verbs rules), and by that to also remove the PMD dependency in libmnl.

Signed-off-by: Moti Haimovsky <motih@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>

net/mlx5: fix crash on null operation

mlx5 implements mlx5_flow_null_drv_ops to be used when a specific
flow typei/driver is not available or invalid.
This routines return error without modifying the rte_flow_error
parameter passed to them which causes testpmd, for example, to crash.
This commit addresses the issue by modifying the rte_flow_error
parameter in theses routines.

Fixes: 0c76d1c9a18d ("net/mlx5: add abstraction for multiple flow drivers")
Fixes: 684dafe795d0 ("net/mlx5: add flow query abstraction interface")
Cc: stable@dpdk.org
Signed-off-by: Moti Haimovsky <motih@mellanox.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>

net/octeontx2: add link status operations

Add support for setting the link up and down.

Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

net/octeontx2: support Rx interrupts

This patch implements rx interrupts feature required for power
saving. These interrupts can be enabled/disabled on demand.

Signed-off-by: Harman Kalra <hkalra@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: add MTU set operation

Add MTU set operation and MTU update feature.

Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Signed-off-by: Sunil Kumar Kori <skori@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: add device stop and close operations

Add device stop, close and reset operations.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

net/octeontx2: add device start operation

Add device start operation and update the correct
function pointers for Rx and Tx burst functions.

This patch also update the octeontx2 NIC specific documentation.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Signed-off-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: add Tx vector version

Add vector version of packet transmit function.

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

net/octeontx2: add Tx multi segment version

Add multi segment version of packet Transmit function.

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

net/octeontx2: support Tx

Add Tx burst support.

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>

net/octeontx2: add Rx vector version

Add vector version of packet Receive function.

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

net/octeontx2: add Rx multi segment version

Add multi segment version of packet Receive function.

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

net/octeontx2: support Rx

Add Rx burst support.

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>

net/octeontx2: add FW version get operation

Add firmware version get operation.

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

net/octeontx2: support VLAN TPID and PVID for Tx

Implement support for setting VLAN TPID and PVID for Tx packets.

Signed-off-by: Vivek Sharma <viveksharma@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: support VLAN filters

Support setting up VLAN filters so as to allow tagged
packet's reception after VLAN HW Filter offload is enabled.

Signed-off-by: Vivek Sharma <viveksharma@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: support VLAN offloads

Support configuring VLAN offloads for an ethernet device and
dynamic promiscuous mode configuration for VLAN filters where
filters are updated according to promiscuous mode of the device.

Signed-off-by: Vivek Sharma <viveksharma@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: implement VLAN utility functions

Implement accessory functions needed for VLAN functionality.
Introduce VLAN related structures as well.

Signed-off-by: Vivek Sharma <viveksharma@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: connect flow API to ethdev ops

Connect rte_flow driver ops to ethdev via .filter_ctrl op.

Signed-off-by: Vivek Sharma <viveksharma@marvell.com>
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>

net/octeontx2: add flow init and fini

Adding the flow init and fini functionality. These will be called from
dev init and will initialize and de-initialize the flow related memory.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: support flow destroy ops

Adding few more flow operations like flow_destroy, flow_isolate
and flow_flush.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: add flow operations

Adding the initial flow ops like flow_create and flow_validate.
These will be used to alloc and write flow rule to device and
validate the flow rule.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: support flow parse actions

Adding support to parse flow actions like drop, count, mark, rss, queue.
On egress side, only drop and count actions were supported.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: add flow parsing for inner layers

Adding functionality to parse inner layers from la to lc.
These will be used to parse inner layers L2, L3, L4 types.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: add flow parsing for outer layers

Adding functionality to parse outer layers from ld to lh.
These will be used parse outer layers L2, L3, L4 and tunnel types.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: add flow MCAM utility functions

Adding MCAM utility functions to alloc and write the entries.
These will be used to arrange the flow rules based on priority.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: add flow mbox utility functions

Adding mailbox utility functions for rte_flow. These will be used
to alloc, reserve and write the entries to the device on request.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: add flow utility functions

First pass rte_flow utility functions for octeontx2.
These will be used to communicate with AF driver.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: introduce flow driver

Introducing flow infra for octeontx2.
This will be used to maintain rte_flow rules.

Create, destroy, validate, query, flush, isolate flow operations
will be supported.

Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Vivek Sharma <viveksharma@marvell.com>

net/octeontx2: add remaining PTP operations

Add remaining PTP configuration/slowpath operations.
Timesync feature is available only for PF devices.

Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: Zyta Szpak <zyta@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: support base PTP

Add PTP enable and disable operations.

Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: Zyta Szpak <zyta@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

net/octeontx2: support flow control

Add flow control operations and exposed
otx2_nix_update_flow_ctrl_mode() to enable on the
configured mode in dev_start().

Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

net/octeontx2: add module EEPROM dump

Add module EEPROM dump operation.

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

net/octeontx2: add Rx and Tx descriptor operations

Add Rx and Tx queue descriptor related operations.

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>

net/octeontx2: add queue info and pool supported ops

Add Rx and Tx queue info get and pool ops supported operations.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>

net/octeontx2: support packet type

The fields from CQE needs to be converted to
ptype and rx ol flags in mbuf. This patch adds
create lookup memory for those items to be
used in Fastpath.

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>

net/octeontx2: enable Tx through traffic manager

This patch enables pkt transmit through traffic manager
hierarchy by clearing software XOFF on the nodes and linking
tx queues to corresponding leaf nodes.
It also adds support to start and stop tx queue using
traffic manager.

Signed-off-by: Krzysztof Kanas <kkanas@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

net/octeontx2: configure TM HW resources

This patch sets up and configure hierarchy in hw
nodes. Since all the registers are with RVU AF,
register configuration is also done using mbox
communication.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Krzysztof Kanas <kkanas@marvell.com>

net/octeontx2: alloc and free TM HW resources

Allocate and free shaper/scheduler hardware resources for
nodes of hierarchy levels in sw.

Signed-off-by: Krzysztof Kanas <kkanas@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

net/octeontx2: introduce traffic manager

Introduce traffic manager infra and default hierarchy
creation.

Upon ethdev configure, a default hierarchy is
created with one-to-one mapped tm nodes. This topology
will be overridden when user explicitly creates and commits
a new hierarchy using rte_tm interface.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Krzysztof Kanas <kkanas@marvell.com>

net/octeontx2: add queue start and stop operations

Add queue start and stop operations. Tx queue needs
to update the flow control value, Which will be
added in sub subsequent patch.

Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

net/octeontx2: handle port reconfigure

setup tx & rx queues with the previous configuration during
port reconfig, it handles cases like port reconfigure without
reconfiguring tx & rx queues.

Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

net/octeontx2: add Tx queue setup and release

Add Tx queue setup and release.

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

net/octeontx2: add Rx queue setup and release

Add Rx queue setup and release.

Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>

net/octeontx2: support RSS

Add RSS support and expose RSS related functions
to implement RSS action for rte_flow driver.

Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>

net/octeontx2: add unicast MAC filter

Add unicast MAC filter for PF device and
update the respective feature list.

Signed-off-by: Sunil Kumar Kori <skori@marvell.com>
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>