Tunnel type format should be translated to ixgbe required format
before register set in FDIR cloud mode, Ans also some register
not useful in cloud mode but only useful in IP mode should be set
to zero as datasheet request.
Fixes: 82fb702077 ("ixgbe: support new flow director modes for X550")
Fixes: 11777435c7 ("net/ixgbe: parse flow director filter")
Cc: stable@dpdk.org
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
In cloud mode for FDIR, tunnel id should be set as protocol
request, the lower 8 bits should be set as reserved.
Fixes: 82fb702077 ("ixgbe: support new flow director modes for X550")
Fixes: 11777435c7 ("net/ixgbe: parse flow director filter")
Cc: stable@dpdk.org
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
In IP mode FDIR, X550 can support not only 4 tuple parameters
but also vlan tci in protocol, so add this feature to flow parser.
Fixes: 11777435c7 ("net/ixgbe: parse flow director filter")
Cc: stable@dpdk.org
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
There are many registeres in x550 support stats of
flow director filters, for example the number of added
or removed rules and the number match or miss match packet
count for this for port, all these important information
can be read form registeres in x550 and display with command
xstats.
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
When detaching a port bound to ixgbe PMD, if the port
does not have any VFs, *vfinfo is not set and there is
a NULL dereference attempt, when calling
rte_eth_switch_domain_free(), which expects VFs to be used,
causing a segmentation fault.
Steps to reproduce:
./testpmd -- -i
testpmd> port stop all
testpmd> port close all
testpmd> port detach 0
Bugzilla ID: 57
Fixes: cf80ba6e20 ("net/ixgbe: add support for representor ports")
Cc: stable@dpdk.org
Reported-by: Anatoly Burakov <anatoly.burakov@intel.com>
Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Tested-by: Anatoly Burakov <anatoly.burakov@intel.com>
Acked-by: Remy Horton <remy.horton@intel.com>
rte_errno should be saved only if error has occurred because rte_errno
could have garbage value.
Fixes: a6d83b6a92 ("net/mlx5: standardize on negative errno values")
Cc: stable@dpdk.org
Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Split maintainers logs from user logs.
A lot of debug logs are present providing internal information on how
the PMD works to users. Such logs should not be available for them and
thus should remain available only when the PMD is compiled in debug
mode.
This commits removes some useless debug logs, move the Maintainers ones
under DEBUG and also move dump into debug mode only.
Cc: stable@dpdk.org
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Write combining (WC) increases NIC performance by making better
utilization of PCI bus. ENA PMD may make usage of this feature.
To enable it load igb_uio driver with wc_activate set to 1.
Signed-off-by: Rafal Kozik <rk@semihalf.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Write combining (WC) increases NIC performance by making better
utilization of PCI bus, but cannot be used by all PMDs.
It will be enabled only if RTE_PCI_DRV_WC_ACTIVATE will be set in
drivers flags. For proper work also igb_uio driver must be loaded with
wc_activate set to 1.
When mapping PCI resources, firstly check if it support WC
and then try to use it.
In case of failure, it will fallback to normal mode.
Signed-off-by: Rafal Kozik <rk@semihalf.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Add pointer to driver structure before calling rte_pci_map_device.
It allows to use driver flags for adjusting configuration.
Signed-off-by: Rafal Kozik <rk@semihalf.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Write combining (WC) increases NIC performance by making better
utilization of PCI bus, but cannot be use by all PMD.
To get internal_addr memory need to be mapped. But as memory could not be
mapped twice: with and without WC it should be skipped for WC. [1]
To do not spoil other drivers that potentially could use internal_addr,
parameter wc_activate adds possibility to skip it for those PMDs,
that do not use it.
[1] https://www.kernel.org/doc/ols/2008/ols2008v2-pages-135-144.pdf
section 5.3 and 5.4
Signed-off-by: Rafal Kozik <rk@semihalf.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Ferruh and Andrew are doing excellent reviews and contributions
to ethdev API.
They become official maintainers and responsibles of this major lib.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
Previously, we were putting an exclusive lock to prevent secondary
processes spinning up while we are sending our messages. However,
using exclusive locks had an effect of disallowing multiple
simultaenous unrelated messages/requests being sent, which was
not the intention behind locking.
Fix it to put a shared lock on the directory. That way, we still
prevent secondary process initializations while sending data over
IPC, but allow multiple unrelated transmissions to proceed.
Fixes: 89f1fe7e6d ("eal: lock IPC directory on init and send")
Cc: stable@dpdk.org
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Tested-by: Qi Zhang <qi.z.zhang@intel.com>
In 17.08, the crypto operation was restructured,
and some reserved bytes (5) were added to have the mempool
pointer aligned to 64 bits, since the structure is expected
to be aligned to 64 bits, allowing future additions with no
ABI breakage needed.
In 18.05, a new 2-byte field was added, so the reserved bytes
were reduced to 3. However, this field was added after the first 3 bytes
of the structure, causing it to be placed in an offset of 4 bytes,
and therefore, forcing the mempool pointer to be placed after 16 bytes,
instead of a 8 bytes, causing unintentionally the ABI breakage.
This commit fixes the breakage, by swapping the reserved bytes
and the private_data_offset field, so the latter is aligned to 2 bytes
and the offset of the mempool pointer returns to its original offset,
8 bytes.
Fixes: 54c8368466 ("cryptodev: set private data for session-less mode")
Cc: stable@dpdk.org
Reported-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Abhinandan Gujjar <abhinandan.gujjar@intel.com>
Converted the license header of the files
that still have the old full header.
Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Acked-by: Kevin Laatz <kevin.laatz@intel.com>
For all trace on the data-path move to dynamic logging.
Add new QAT_DP_LOG macro to separate from non-dp trace.
Use the overall RTE_LOG_DP_LEVEL to override and compile out
debug data-path trace.
Add "unlikely" in two appropriate error checks.
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
For all trace not on the data-path move to dynamic logging.
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Removed INIT_FUNC trace and other unused macros.
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Document the driver and device naming formats.
Changed the underscores alignment.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Other services, apart from symmetric crypto,
such as compression, will be added in future patches.
Therefore, the assumption that only symmetric crypto
operations are processed will not be valid anymore,
and service type needs to be checked.
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Optimize the dequeue function by inlining the response
processing function, assuming only symmetric
operations are supported.
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Process response function is only implemented
for crypto symmetric operations, which do not require
some of the arguments.
Therefore, these arguments can be removed from the
function prototype.
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
For every QAT PCI device probed, populate a local rte_device
containing an rte_driver. The rte_driver was created in a previous
patch to provide a crypto-specific driver name: "crypto_qat".
This was previously only used for driver registration, now it's also
used in device creation.
This allows applications to find devices driven by "crypto_qat".
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
QAT_64_BTYE_ALIGN_MASK macro is not specific to
symmetric crypto, but it is common for other
future services.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Removed crypto reference in common debug message.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
Renamed sgl_cookie to cookie and qp_gen_config
to qat_gen_config, as it is intended to hold
more than just queue pair data.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Signed-off-by: Tomasz Jozwiak <tomaszx.jozwiak@intel.com>
As bundle_num is included in qat_gen1_qps static array
there shouldn't be a multiplier used in qat_qps_per_service()
Then removed ADF_NUM_BUNDLES_PER_DEV as no longer used.
Also renamed ADF_MAX_QPS_PER_BUNDLE to ADF_MAX_QPS_ON_ANY_SERVICE
and reduced from 4 to 2 which is enough for all current devices.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Add lock around accesses to the arbiter CSR
and use & instead of ^ as ^ not safe if
arb_disable called when already disabled.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Move all code into appropriate files, no actual code changes. Specifically:
- Rename rte_qat_cryptodev.c to qat_sym_pmd.c
- Create qat_sym_pmd.h and populate with fn prototypes for qat_sym_pmd.c
- Create qat_comp_pmd.c/.h and populate with placeholder functions
- Create qat_asym_pmd.c/.h and populate with placeholder functions
- Rename qat_crypto_capabilities.h to qat_sym_capabilities.h
- Move CRYPTODEV_NAME_QAT_SYM_PMD from qat_common.h to qat_sym_pmd.h
- Move qat_sym_dev_private from qat_device.h to qat_sym_pmd.h
- Move prototype for qat_sym_dev_info_get frm qat_device.h 2 qat_sym_pmd.h
- Move all qat_device.c sym dev_ops fns to qat_sym_pmd.c file
- Move all qat_sym.c dev_ops fns to qat_sym_pmd.c file
- Remove unused header file #includes from all files.
- Move pci_id_qat_map, probe/release/register from
rte_qat_cryptodev.c to qat_device.c
- Moved stray comment for bpi_cipher_ctx_init() from qat_sym.c
to qat_sym_session.c
- Changed all files to use SPDX license header
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Split qat_sym_stats_get/reset into 2 functions, a wrapper function calling
a new qat_stats_get/reset function which can be called per service.
Remove cryptodev stats struct from qat_qp, replace with qat_common_stats.
Add links for qat_qp into qat_pci_device using an array per service to
avoid need for a lock and so qp_id for the service can be used as index.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Struct qat_pmd_private held the data needed by cryptodev, common code
now gets most data from struct qat_pci_device instead. qat_pmd_private
is trimmed to hold only sym crypto data and renamed qat_sym_private
to reflect its usage.
Specifically
- remove max_nb_queue_pairs from qat_pmd_private, get from qp_hw_data
- remove max_nb_sesssions from qat_pmd_private as not needed.
- remove qat_gen from qat_pmd_private, get from qat_pci_device instead.
- use qat_pci_device throughout common code instead of qat_pmd_private
- rename qat_pmd_private to qat_sym_dev_private - this now holds only
sym-specific data for the cryptodev API
- extend pci device name to <bdf>_qat for clarity, was just <bdf>
- update qp mem and cookiepool names to reflect the appropriate device,
service and qp.
- rename qat_dev_info_get() to qat_sym_dev_info_get() as mostly sym,
not enough common info to warrant a generic fn.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
The QAT PMD used to register with PCI using the name "crypto_qat".
Keep this name for the driver registered with cryptodev
and use a more generic name "qat" for the PCI registration.
This paves the way for the PCI device to host other services.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
- Added struct qat_pci_device to use internally in QAT PMD
to avoid dependencies on rte_cryptodev or rte_compressdev
- Added a global array of these
- Restructured probe/release to separate QAT common init/clear
from crypto pmd create/destroy.
- In QAT common part allocated a qat_pci_device and populated it
- Removed meaningless check in probe for RTE_PROC_PRIMARY
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Create data structures to support different generations
of qat hardware supplying services through different queue pairs.
- Add two new structs qat_gen_hw_data and qat_qp_hw_dat
- Add a qat_service_type enum
An array of qat_qp_hw_data elements is
initialised with constants, these are arranged so that the qp_id used
on the API can be used as an index to pick up the qp data to use.
The constants are common to current generations,
new arrays will be added for future generations.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Move defines related to coalescing from sym header file to qp header
file as these will be common for all services.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Remove dependence on rte_cryptodev from common qp code
to facilitate being used by other device types in future.
Transferred required data into qat-specific structures.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Extracted all sym-specific code from qp setup fns, leaving
generic qat_qp_setup fn and helper fns. Created a new
meta-data struct qat_qp_config to hold all the data needed
to create a qp, filled this out in the sym-specific code
and passed to the generic qp_setup fn.
No need now for rx and tx queue_create fns, one generic
queue_create fn replaces these.
Included the service name (e.g. "sym") in the qp memzone
and cookie pool names.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Move the generic enqueue and dequeue fns from
the qat_sym.c file to the qat_qp.c file
Move generic qp structs to a new qat_qp.h file
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
Queue-handling code in dequeue is made generic, so it can
be used by other services in future. This is done by
- Removing all sym-specific refs in input params - replace with void ptrs.
- Wrapping this generic dequeue with the sym-specific dequeue
called through the API.
- extracting the sym-specific response processing into a new fn.
- Setting a fn ptr for process_response in qp on qp creation
- Passing void * params to this, in the service-specific implementation
qat_sym_process_response cast back to sym structs.
Signed-off-by: Fiona Trahe <fiona.trahe@intel.com>
