When strings are added to an array variable, we need to properly escape
the invalid json characters in the strings.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Ciara Power <ciara.power@intel.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
For string values returned from telemetry, escape any values that cannot
normally appear in a json string. According to the json spec[1], the
characters than need to be handled are control chars (char value < 0x20)
and '"' and '\' characters.
To handle this, we replace the snprintf call with a separate string
copying and encapsulation routine which checks each character as it
copies it to the final array.
[1] https://www.rfc-editor.org/rfc/rfc8259.txt
Bugzilla ID: 1037
Fixes: 6dd571fd07 ("telemetry: introduce new functionality")
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Ciara Power <ciara.power@intel.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
To save issues with encoding the names of values in dicts, we limit the
allowed names to a subset of character values. This list of allowed
characters can be expanded as necessary in future.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Ciara Power <ciara.power@intel.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Remove deprecated fdir_conf from device configuration.
Assume that mode is equal to RTE_FDIR_MODE_NONE.
Add internal Flow Director configuration copy in ixgbe and txgbe device
private data since flow API supports requires it. Initialize mode to
the first flow rule mode on the rule validation or creation.
Since Flow Director configuration data types are still used by some
drivers internally, move it from public API to ethdev driver internal
API.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Dongdong Liu <liudongdong3@huawei.com>
Remove deprecated ``ETH_VLAN_*`` and ``ETH_QINQ_`` defines.
Use corresponding defines with ``RTE_`` prefix instead.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>
Acked-by: Ferruh Yigit <ferruh.yigit@xilinx.com>
Remove deprecated ``DEV_RX_OFFLOAD_*`` and ``DEV_TX_OFFLOAD_`` defines.
Use corresponding defines with ``RTE_ETH_RX_OFFLOAD_`` and
``RTE_ETH_TX_OFFLOAD_`` prefix instead.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>
Acked-by: Ferruh Yigit <ferruh.yigit@xilinx.com>
Remove deprecated ``ETH_RSS_*`` defines used for hash function and RETA
size specification. Use corresponding defines with ``RTE_`` prefix
instead.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>
Acked-by: Ferruh Yigit <ferruh.yigit@xilinx.com>
Remove deprecated ``ETH_MQ_RX_*`` and ``ETH_MQ_TX_*`` defines.
Use corresponding defines with ``RTE_`` prefix instead.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>
Acked-by: Ferruh Yigit <ferruh.yigit@xilinx.com>
Remove deprecated ``ETH_LINK_SPEED_``, ``ETH_SPEED_NUM_`` and
``ETH_LINK_`` defines. Use corresponding defines with ``RTE_`` prefix
instead.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: David Marchand <david.marchand@redhat.com>
Acked-by: Ferruh Yigit <ferruh.yigit@xilinx.com>
Add pipeline control path API to manage direct meters. These meters
are identified by a table key, whose entry ID is used as the index
into the meter array.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Add pipeline control path API to read/write direct registers. These
registers are identified by a table key, whose entry ID is used as the
index into the register array.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Add the entry ID instruction that reads the entry ID of the latest
table lookup operation from the pipeline into the meta-data. The entry
ID is then used by the register and meter instructions as the index
into the register or meter array.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Add support for unique ID for each learner table entry. The entry ID
is retrieved as part of the learner table lookup operation and is
saved by the pipeline for later use.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Add support for unique ID for each table entry. The entry ID is
retrieved as part of the table lookup operation and is saved by the
pipeline for later use.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Make the hash function configurable for the learner pipeline tables.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Make the hash function configurable. The internal hash function that
was not configurable, mask-based and limited to 64 bytes is removed.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Make the hash function configurable for the regular pipeline tables.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Make the hash function configurable. The internal hash function that
was not configurable, mask-based and limited to 64 bytes is removed.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Add key comparison functions to be used by the exact match and the
learner table types as part of the performance critical lookup
operation. Since the key size is fixed, it is possible to select a
specialized memory copy function as opposed to using the variable size
version, resulting in a performance improvement of around 5%.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Add hash function prototype to be used by the exact match and the
learner table types. The hash function is not mask-based, so the table
key fields have to be contiguous in memory.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Make rte_device opaque for non internal users.
This will make extending this object possible without breaking the ABI.
Some applications may have been dereferencing rte_device objects, mark
this object's accessors as stable.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
For diagnostic, it may be useful to provide a description of the device
with bus specific information.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Prepare for making the device object opaque by adding accessors.
Update existing "external" users.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Make rte_driver opaque for non internal users.
This will make extending this object possible without breaking the ABI.
Introduce a new driver header and move rte_driver definition.
Update drivers and library to use the internal header.
Some applications may have been dereferencing rte_driver objects, mark
this object's accessors as stable.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Akhil Goyal <gakhil@marvell.com>
Acked-by: Abhinandan Gujjar <abhinandan.gujjar@intel.com>
Prepare for making the driver object opaque by adding accessors.
Update existing "external" users.
Internal users may still dereference a rte_driver object.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Make rte_bus opaque for non internal users.
This will make extending this object possible without breaking the ABI.
Introduce a new driver header and move rte_bus definition and helpers.
Update drivers and library to use the internal header.
Some applications may have been dereferencing rte_bus objects, mark
this object's accessors as stable.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Add helpers to get a rte_bus object details.
This will be used externally.
Internal users may still dereference a rte_bus object.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
iova enum definition does not need to be defined as part of the bus API.
Move it to rte_eal.h.
With this step, rte_eal.h does not depend on rte_bus.h and rte_dev.h.
Fix existing code that was relying on these implicit inclusions.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
The vdev bus interface is for drivers only.
Mark as internal and move the header in the driver headers list.
While at it, cleanup the code:
- fix indentation,
- remove unneeded reference to bus specific singleton object,
- remove unneeded list head structure type,
- reorder the definitions and macro manipulating the bus singleton object,
- remove inclusion of rte_bus.h and fix the code that relied on implicit
inclusion,
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
The pci bus interface is for drivers only.
Mark as internal and move the header in the driver headers list.
While at it, cleanup the code:
- fix indentation,
- remove unneeded reference to bus specific singleton object,
- remove unneeded list head structure type,
- reorder the definitions and macro manipulating the bus singleton object,
- remove inclusion of rte_bus.h and fix the code that relied on implicit
inclusion,
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>
We don't need to include rte_bus.h in rte_devargs.h.
Only a forward declaration of rte_bus and an inclusion of rte_dev.h are
needed.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Those macros have no real value and are easily replaced with a simple
if() block.
Existing users have been converted using a new cocci script.
Deprecate them.
Signed-off-by: David Marchand <david.marchand@redhat.com>
For any bus that does not support device iteration, rte_dev_iterator_init
both returned an error code and logged an error message.
An application (like testpmd) that only wants to list devices, would have
no choice but to inspect a bus object to avoid spewing error logs.
Make those log messages debug level, and remove the check in testpmd.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Add support to the move instruction for operands bigger than 64 bits.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Harshad Narayane <harshad.suresh.narayane@intel.com>
Support reading structure fields that are bigger than 64 bits on the
control path for the table update operations.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Harshad Narayane <harshad.suresh.narayane@intel.com>
Remove the generic limitation of structure fields to 64 bits or less
and push this restriction to the instructions that require it.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Harshad Narayane <harshad.suresh.narayane@intel.com>
DLB2 has a need to parse a user supplied coremask as part
of an optimization that associates optimal core/resource
pairs. Therefore eal_parse_coremask has been renamed
to rte_eal_parse_coremask and exported but kept internal.
Signed-off-by: Abdullah Sevincer <abdullah.sevincer@intel.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Do not flush the buffered packets unnecessarily when a burst was sent
since the last flush call.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Free the buffered packets as opposed to retrying to send them when the
output port is freed.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Drop packets that cannot be sent instead of retry sending the same
packets potentially forever when the ring consumer that is down.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Do not flush the buffered packets unnecessarily when a burst was sent
since the last flush call.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Free the buffered packets as opposed to retrying to send them when the
output port is freed.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Drop packets that cannot be sent instead of retry sending the same
packets potentially forever when the Ethernet device that is down.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
The sink port is tasked to drop all packets, hence the packet and byte
counters should be named to reflect the drop operation.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Add packet drop statistics counters for the output ports. Required by
the non-blocking output port behavior where the packets that cannot
be sent at the time of the operation are dropped as opposed to the
send operation being retried potentially forever for the same packets.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Do not include <ctype.h>, <errno.h>, and <stdlib.h> from <rte_common.h>,
because they are not used by this file.
Include the needed headers directly from the files that need them.
Signed-off-by: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
There is no reason for rte_str_to_size() to be inline.
Move the implementation out of <rte_common.h>.
Export it as a stable ABI because it always has been public.
Signed-off-by: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
RTE_CACHE_LINE_ROUNDUP() implementation repeated RTE_ALIGN_MUL_CEIL().
In other places RTE_CACHE_LINE_SIZE is assumed to be a power-of-2,
so define RTE_CACHE_LINE_ROUNDUP() using RTE_ALIGN_CEIL().
Signed-off-by: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>
Reviewed-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
There is no point in such a call and UBSan complains about a call to
memcpy with a null pointer as second arg.
When building with -Db_sanitize=undefined, Clang gives the following
warning
../lib/bpf/bpf_load.c:37:20: runtime error: null pointer passed as
argument 2, which is declared to never be null
A check of the sz before calling memcpy fixes that.
Signed-off-by: Henning Schild <henning.schild@siemens.com>
__rte_raw_cksum() (used by rte_raw_cksum() among others) accessed its
data through an uint16_t pointer, which allowed the compiler to assume
the data was 16-bit aligned. This in turn would, with certain
architectures and compiler flag combinations, result in code with SIMD
load or store instructions with restrictions on data alignment.
This patch keeps the old algorithm, but data is read using memcpy()
instead of direct pointer access, forcing the compiler to always
generate code that handles unaligned input. The __may_alias__ GCC
attribute is no longer needed.
The data on which the Internet checksum functions operates are almost
always 16-bit aligned, but there are exceptions. In particular, the
PDCP protocol header may (literally) have an odd size.
Performance impact seems to range from none to a very slight
regression.
Bugzilla ID: 1035
Fixes: 6006818cfb ("net: new checksum functions")
Cc: stable@dpdk.org
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
A mempool consumes 3 memzones (with the default ring mempool driver).
The default DPDK configuration allows RTE_MAX_MEMZONE (2560) memzones.
Assuming there is no other memzones that means that we can have a
maximum of 853 mempools.
In the vhost library, the IOTLB cache code so far was requesting a
mempool per vq, which means that at the maximum, the vhost library
could request mempools for 426 qps.
This limit was recently reached on big systems with a lot of virtio
ports (and multiqueue in use).
While the limit on mempool count could be something we fix at the DPDK
project level, there is no reason to use mempools for the IOTLB cache:
- the IOTLB cache entries do not need to be DMA-able and are only used
by the current process (in multiprocess context),
- getting/putting objects from/in the mempool is always associated with
some other locks, so some level of lock contention is already present,
We can convert to a malloc'd pool with objects put in a free list
protected by a spinlock.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Having a back reference to the index of the vq in the dev->virtqueue[]
array makes it possible to unify the internal API, with only passing dev
and vq.
It also allows displaying the vq index in log messages.
Remove virtqueue index checks where unneeded (like in static helpers
called from a loop on all available virtqueue).
Move virtqueue index validity checks the sooner possible.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
translate_ring_addresses and numa_realloc may change a virtio device and
virtio queue. Callers of those helpers must be extra careful and refresh
any reference to old data.
Change those functions prototype as a way to hint about this issue and
always ask for an indirect pointer.
Besides, when reallocating the device and queue, the code already made
sure it will return a pointer to a valid device. The checks on such
returned pointer can be removed.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
translate_ring_addresses (via numa_realloc) may change a virtio device and
virtio queue.
The virtqueue object must be refreshed before accessing the lock.
Fixes: 04c27cb673 ("vhost: fix unsafe vring addresses modifications")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Previously, the pipeline build operation was done based on the
specification file (typically produced by the P4 compiler), then the C
code with optimized functions for the pipeline actions and
instructions was generated, built into a shared object library, loaded
and installed into the pipeline in a completely hardcoded and
non-customizable way.
Now, this process is split into three explicit stages:
i) code generation (specification file -> C file);
ii) code build (C file -> shared object library);
iii) code installation (library load into the pipeline).
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Previously, the C code generation for the pipeline was hidden under
the hood; now, we make this an explicit API operation. Besides the
functions for the pipeline actions and the pipeline instructions,
the generated C source code now includes the pipeline specification
structure required for the pipeline configuration operations.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Add specification data structure and API for the pipeline I/O ports
and related pipeline configuration such as packet mirroring.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Add support to export the pipeline specification data structure to a C
source code file.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Rework the specification file-based pipeline build operation to first
parse the specification file into the previously introduced pipeline
specification data structure, then use this structure to configure
and build the pipeline.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Add specification data structure for the entire pipeline.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Move all the pipeline object specification data structures to an
internal header file.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Add an unique name to every pipeline. This enables the library to
maintain a list of the existing pipeline objects, which can be
queried by the application.
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Signed-off-by: Kamalakannan R <kamalakannan.r@intel.com>
Some NIC drivers support MBUF_FAST_FREE (device supports optimization
for fast release of mbufs. When set, application must guarantee that
per-queue all mbufs comes from the same mempool, has refcnt = 1, direct
and non-segmented.) offload.
In order to adapt to this offload function, add this API.
Add some test data for this API.
Signed-off-by: Huichao Cai <chcchc88@163.com>
Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
This patch aims at supporting the unlikely case where a
Virtio-net header is spanned across more than two
descriptors.
CVE-2022-2132
Fixes: fd68b4739d ("vhost: use buffer vectors in dequeue path")
Cc: stable@dpdk.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
This patch discards descriptor chains which are smaller
than the Virtio-net header size, and ones that are equal.
Indeed, such descriptor chains sizes mean there is no
packet data.
This patch also has the advantage of requesting the exact
packets sizes for the mbufs.
CVE-2022-2132
Fixes: 62250c1d09 ("vhost: extract split ring handling from Rx and Tx functions")
Fixes: c3ff0ac70a ("vhost: improve performance by supporting large buffer")
Fixes: 84d5204310 ("vhost: support async dequeue for split ring")
Cc: stable@dpdk.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
The "offset" and "n_bits" fields were generated incorrectly, hence the
output C file was producing compilation errors when the "recircid"
instruction was used.
Fixes: 5ec76d29dc ("pipeline: support packet recirculation")
Cc: stable@dpdk.org
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Added changes to enable CMAN (RED or PIE) at init
from profile configuration file.
By default CMAN code is enabled but not in use, when
there is no RED or PIE profile configured.
Signed-off-by: Marcin Danilewicz <marcinx.danilewicz@intel.com>
Acked-by: Jasvinder Singh <jasvinder.singh@intel.com>
Start a new release cycle with empty release notes.
The ABI version becomes 23.0.
The map files are updated to the new ABI major number (23).
The ABI exceptions are dropped and CI ABI checks are disabled because
compatibility is not preserved.
Special handling of removed drivers is also dropped in check-abi.sh and
a note has been added in libabigail.abignore as a reminder.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
The dirty page logging is only required in vhost enqueue direction for
live migration. This patch removes the unnecessary dirty page logging
in vhost dequeue direction. Otherwise, it will result in a performance
drop. Some if-else judgements are also optimized to improve performance.
Fixes: 6d823bb302 ("vhost: prepare sync for descriptor to mbuf refactoring")
Fixes: b6eee3e834 ("vhost: fix sync dequeue offload")
Signed-off-by: Xuan Ding <xuan.ding@intel.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Tested-by: Xingguang He <xingguang.he@intel.com>
A packet with RTE_PTYPE_L4_FRAG(0x300) contains both RTE_PTYPE_L4_TCP
(0x100) & RTE_PTYPE_L4_UDP (0x200). A fragmented packet as defined in
rte_mbuf_ptype.h cannot be recognized as other L4 types and hence the
GRO layer should not use IS_IPV4_TCP_PKT or IS_IPV4_UDP_PKT for
RTE_PTYPE_L4_FRAG. Hence, if the packet type is RTE_PTYPE_L4_FRAG the
IP header should be parsed to recognize the appropriate IP type and
invoke the respective gro handler.
Fixes: 1ca5e67408 ("gro: support UDP/IPv4")
Cc: stable@dpdk.org
Signed-off-by: Kumara Parameshwaran <kumaraparamesh92@gmail.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
This commit fixes an issue where calling rte_service_lcore_stop()
would result in a service's "active on lcore" status becoming stale.
The stale status would result in rte_service_may_be_active() always
returning "1", indicating that the service is not certainly stopped.
This is fixed by ensuring the "active on lcore" status of each service
is set to 0 when an lcore is stopped.
Fixes: e30dd31847 ("service: add mechanism for quiescing")
Fixes: 8929de043e ("service: retrieve lcore active state")
Reported-by: Naga Harish K S V <s.v.naga.harish.k@intel.com>
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
We recently improved the log messages in the vhost library, adding some
context that helps filtering for a given vhost-user device.
However, some parts of the code were missed, and some later code changes
broke this new convention (fixes were sent previous to this patch).
Change the VHOST_LOG_CONFIG/DATA helpers and always ask for a string
used as context. This should help limit regressions on this topic.
Most of the time, the context is the vhost-user device socket path.
For the rest when a vhost-user device can not be related, generic
names were chosen:
- "dma", for vhost-user async DMA operations,
- "device", for vhost-user device creation and lookup,
- "thread", for threads management,
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Those messages were missed when adding socket context.
Fix this.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
device information in the log messages was dropped.
Fixes: 52ade97e36 ("vhost: fix physical address mapping")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
This patch checks the return value of rte_dma_info_get()
called in rte_vhost_async_dma_configure().
Coverity issue: 379066
Fixes: 53d3f4778c ("vhost: integrate dmadev in asynchronous data-path")
Cc: stable@dpdk.org
Signed-off-by: Jiayu Hu <jiayu.hu@intel.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
This patch fixes the missing virtio net header copy in sync
dequeue path caused by refactoring, which affects dequeue
offloading.
Fixes: 6d823bb302 ("vhost: prepare sync for descriptor to mbuf refactoring")
Signed-off-by: Xuan Ding <xuan.ding@intel.com>
Tested-by: Wei Ling <weix.ling@intel.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
In the virtio blk vDPA live migration use case, before the live
migration process, QEMU will set call fd to vDPA back-end. QEMU
and vDPA back-end stand by until live migration starts.
During live migration process, QEMU sets kick fd and a new call
fd. However, after the kick fd is set to the vDPA back-end, the
vDPA back-end configures device and data path starts. The new
call fd will cause some kind of "re-configuration", this kind
of "re-configuration" cause IO drop.
After this patch, vDPA back-end configures device after kick fd
and call fd are well set and make sure no IO drops.
This patch only impact virtio blk vDPA device and does not impact
net device.
Fixes: 7015b65771 ("vdpa/ifc: add block device SW live-migration")
Signed-off-by: Andy Pei <andy.pei@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Add spinlock protection in queue delete function.
This protects the data path while the queue delete operation
is in progress.
Fixes: a3bbf2e097 ("eventdev: add eth Tx adapter implementation")
Cc: stable@dpdk.org
Signed-off-by: Naga Harish K S V <s.v.naga.harish.k@intel.com>
Acked-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
These are functions related to interrupts that have been
in since 20.02 release or earlier.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Ray Kinsella <mdr@ashroe.eu>
These API's have been around for a long time and by now are fixed.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Ray Kinsella <mdr@ashroe.eu>
The RTE_LOG_REGISTER is not experimental, and the experimental
tag was never enforced on these.
Make rte_log_can_log a fully supported function.
It was introduced nearly 2yrs ago.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Ray Kinsella <mdr@ashroe.eu>
The comments in rte_rib6 were cut-and-pasted from rte_rib
and because of that some references to rte_rib_node were
not updated.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Found by nullfree.cocci.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
[David: for lpm parts:]
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Acked-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
[David: for vdpa/mlx5 parts:]
Acked-by: Matan Azrad <matan@nvidia.com>
[David: for dma/dpaa2, raw/ifpga, vdpa/mlx5:]
Acked-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
Reviewed-by: Chengwen Feng <fengchengwen@huawei.com>
[David: reran cocci.sh and updated common/mlx5 and cryptodev asym test]
Signed-off-by: David Marchand <david.marchand@redhat.com>
Make sure all functions which use the convention that XXX_free(NULL)
is a nop are all documented.
The wording is chosen to match the documentation of free(3).
"If ptr is NULL, no operation is performed."
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
[David: squashed with other series updates, unified wording]
Remove extraneous phrase "This API is used to" and use
active instead of passive voice when describing a function.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
[David: for raw/ioat and dmadev parts:]
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Conor Walsh <conor.walsh@intel.com>
Add support for using hugepages for worker lcore stack memory. The
intent is to improve performance by reducing stack memory related TLB
misses and also by using memory local to the NUMA node of each lcore.
EAL option '--huge-worker-stack[=stack-size-in-kbytes]' is added to allow
the feature to be enabled at runtime. If the size is not specified,
the system pthread stack size will be used.
Signed-off-by: Don Wallwork <donw@xsightlabs.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
GCC 12 raises warnings on usage of rte_memcpy with IPv4 options handling
in fragments for both the ip_frag library and unit tests.
For example in the library:
In function ‘_mm256_storeu_si256’,
inlined from ‘rte_mov32’ at
../lib/eal/x86/include/rte_memcpy.h:347:2,
inlined from ‘rte_mov128’ at
../lib/eal/x86/include/rte_memcpy.h:369:2,
inlined from ‘rte_memcpy_generic’
at ../lib/eal/x86/include/rte_memcpy.h:445:4,
inlined from ‘rte_memcpy’
at ../lib/eal/x86/include/rte_memcpy.h:851:10,
inlined from ‘__create_ipopt_frag_hdr’
at ../lib/ip_frag/rte_ipv4_fragmentation.c:68:4,
inlined from ‘rte_ipv4_fragment_packet’
at ../lib/ip_frag/rte_ipv4_fragmentation.c:242:16:
/usr/lib/gcc/x86_64-redhat-linux/12/include/avxintrin.h:935:8: error:
array subscript ‘__m256i_u[1]’ is partly outside array bounds of
‘uint8_t[60]’ {aka ‘unsigned char[60]’} [-Werror=array-bounds]
935 | *__P = __A;
| ~~~~~^~~~~
../lib/ip_frag/rte_ipv4_fragmentation.c: In function
‘rte_ipv4_fragment_packet’:
../lib/ip_frag/rte_ipv4_fragmentation.c:122:17: note: at offset [52, 60]
into object ‘ipopt_frag_hdr’ of size 60
122 | uint8_t ipopt_frag_hdr[IPV4_HDR_MAX_LEN];
| ^~~~~~~~~~~~~~
To resolve the compilation warning, replace the rte_memcpy with memcpy.
Fixes: b50a14a853 ("ip_frag: add IPv4 options fragment")
Signed-off-by: Huichao Cai <chcchc88@163.com>
If called to allocate memory of size is between multiple of hugepage
size minus malloc_header_len and hugepage size, rte_malloc fails.
This fix replaces malloc_elem_trailer_len with malloc_elem_overhead in
try_expand_heap() to include malloc_elem_header_len when calculating
n_seg.
Bugzilla ID: 800
Fixes: 07dcbfe010 ("malloc: support multiprocess memory hotplug")
Cc: stable@dpdk.org
Signed-off-by: Fidaullah Noonari <fidaullah.noonari@emumba.com>
Acked-by: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>
rte_dump_stack() needs to be usable in situations when a bug is
encountered and from signal handlers (such as SEGV).
Glibc backtrace_symbols() calls malloc which makes it
dangerous in a signal handler that is handling errors that maybe
due to memory corruption. Additionally, rte_log() is unsafe because
syslog() is not signal safe; printf() is also documented as
not being safe.
This version formats message and uses writev for each line in a manner
similar to what glibc version of backtrace_symbols_fd() does. The
FreeBSD version of backtrace_symbols_fd() is not signal safe.
Sample output:
0: ./build/app/dpdk-testpmd (rte_dump_stack+0x2b) [560a6e9c002b]
1: ./build/app/dpdk-testpmd (main+0xad) [560a6decd5ad]
2: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xcd) [7fd43d3e27fd]
3: ./build/app/dpdk-testpmd (_start+0x2a) [560a6e83628a]
Bugzilla ID: 929
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Reviewed-by: David Marchand <david.marchand@redhat.com>
copy_data was returning a pointer to an increased (off by one) descriptor.
Subsequent calls to copy_data in the library were then failing.
Fix this by incrementing the descriptor only if there is some left data
to copy.
Fixes: 4414bb6701 ("vhost/crypto: fix build with GCC 12")
Cc: stable@dpdk.org
Reported-by: Jakub Poczatek <jakub.poczatek@intel.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Tested-by: Jakub Poczatek <jakub.poczatek@intel.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
GCC 12 raises the following warning:
In file included from ../lib/mempool/rte_mempool.h:46,
from ../lib/mbuf/rte_mbuf.h:38,
from ../lib/vhost/vhost_crypto.c:7:
../lib/vhost/vhost_crypto.c: In function ‘rte_vhost_crypto_fetch_requests’:
../lib/eal/x86/include/rte_memcpy.h:371:9: warning: array subscript 1 is
outside array bounds of ‘struct virtio_crypto_op_data_req[1]’
[-Warray-bounds]
371 | rte_mov32((uint8_t *)dst + 3 * 32, (const uint8_t *)src + 3 * 32);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../lib/vhost/vhost_crypto.c:1178:42: note: while referencing ‘req’
1178 | struct virtio_crypto_op_data_req req;
| ^~~
Split this function and separate the per descriptor copy.
This makes the code clearer, and the compiler happier.
Note: logs for errors have been moved to callers to avoid duplicates.
Fixes: 3c79609fda ("vhost/crypto: handle virtually non-contiguous buffers")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Since the commit 02798b0735 ("vhost: improve virtio-net layer logs"),
vhost logs contain the socket path as a prefix.
Async dequeue path was copied from the sync dequeue path but a log
was incorrect.
Fixes: 84d5204310 ("vhost: support async dequeue for split ring")
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
This patchs renames the local variables free_entries to
avail_entries in the dequeue path.
Indeed, this variable represents the number of new packets
available in the Virtio transmit queue, so these entries
are actually used, not free.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
This patch implements packed ring dequeue data path
for asynchronous vhost.
Signed-off-by: Cheng Jiang <cheng1.jiang@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
rte_vhost_clear_queue_thread_unsafe() supports to clear
in-flight packets for async enqueue only. But after
supporting async dequeue, this API should support async dequeue too.
This patch also adds the thread-safe version of this API,
the difference between the two API is that thread safety uses lock.
These APIs maybe used to clean up packets in the async channel
to prevent packet loss when the device state changes or
when the device is destroyed.
Signed-off-by: Yuan Wang <yuanx.wang@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
The Virtio specification requires that in case of checksum
offloading, the pseudo-header checksum must be set in the
L4 header.
When received from another Vhost-user port, the packet
checksum might already contain the pseudo-header checksum
but we have no way to know it. So we have no other choice
than doing the pseudo-header checksum systematically.
This patch handles this using the rte_net_intel_cksum_prepare()
helper.
Fixes: 859b480d5a ("vhost: add guest offload setting")
Cc: stable@dpdk.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
During adapter create, memory is allocated for storing event port
configuration which is freed during adapter free. The following
error is seen during free "EAL: Error: Invalid memory"
The service data pointer storage for txa_service_data_array is
allocated during adapter create with incorrect size which is less
than the required size.
Initialization of this memory causes buffer overflow and result in
metadata overwrite of event port config memory allocated above
and results in the above error message during free.
Allocating the correct size of memory for txa_service_data_array
prevents overwriting other memory areas like event port config
memory.
Fixes: a3bbf2e097 ("eventdev: add eth Tx adapter implementation")
Cc: stable@dpdk.org
Signed-off-by: Naga Harish K S V <s.v.naga.harish.k@intel.com>
Acked-by: Jay Jayatheerthan <jay.jayatheerthan@intel.com>
Fix the UDP header fields, wrong byte order used for src and dst port
and wrong offset used when updating UDP datagram length.
Fixes: 01eef5907f ("ipsec: support NAT-T")
Cc: stable@dpdk.org
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
ALIGNMENT_MASK is only used internally.
Besides it lacks a DPDK-related prefix.
Hide it from external eyes.
Fixes: f5472703c0 ("eal: optimize aligned memcpy on x86")
Cc: stable@dpdk.org
Reported-by: Morten Brørup <mb@smartsharesystems.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
The function rte_pie_drop was attempting to do a random probability
drop, but because of incorrect usage of fixed point divide
it would always return 1.
Change to use new rte_drand() instead.
Fixes: 44c730b0e3 ("sched: add PIE based congestion management")
Cc: stable@dpdk.org
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Jasvinder Singh <jasvinder.singh@intel.com>
The qdelay variable is derived from and compared to 64 bit
value so it doesn't have to be floating point.
Fixes: 44c730b0e3 ("sched: add PIE based congestion management")
Cc: stable@dpdk.org
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Jasvinder Singh <jasvinder.singh@intel.com>
The PIE code and other applications can benefit from having a
fast way to get a random floating point value. This new function
is equivalent to drand() in the standard library.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Ray Kinsella <mdr@ashroe.eu>
These header includes have been flagged by the iwyu_tool
and removed.
Signed-off-by: Sean Morrissey <sean.morrissey@intel.com>
Acked-by: Chenbo Xia <chenbo.xia@intel.com>
These header includes have been flagged by the iwyu_tool
and removed.
Signed-off-by: Sean Morrissey <sean.morrissey@intel.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Bug scenario:
1. start testpmd:
$ dpdk-testpmd -l 4-6 -a 0000:7d:00.0 --trace=.* \
--file-prefix=trace_autotest -- -i
2. then observed:
EAL: eal_trace_init():93 failed to initialize trace [File exists]
EAL: FATAL: Cannot init trace
EAL: Cannot init trace
EAL: Error - exiting with code: 1
The root cause it that the offset set wrong with long file-prefix and
then lead the strftime return failed.
At the same time, trace_session_name_generate() uses errno as the return
value, but the errno was not set if strftime returned zero.
A previously set errno (EEXIST or ENOENT from call to mkdir for creating
the runtime configuration directory) was returned in this case.
This is fragile and may lead to incorrect logic if errno was set
to 0 previously.
This also resulted in inaccurate prompting.
Set errno to ENOSPC if strftime return zero.
Fixes: 321dd5f8fa ("trace: add internal init and fini interface")
Cc: stable@dpdk.org
Signed-off-by: Chengwen Feng <fengchengwen@huawei.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
Bug scenario:
1. start testpmd:
$ dpdk-testpmd -l 4-6 -a 0000:7d:00.0 --trace=.* -- -i
2. quit testpmd and then observed segment fault:
Bye...
Segmentation fault (core dumped)
The root cause is that rte_trace_save() and eal_trace_fini() access
the huge pages which were cleanup by rte_eal_memory_detach().
This patch moves rte_trace_save() and eal_trace_fini() before
rte_eal_memory_detach() to fix the bug.
Fixes: dfbc61a2f9 ("mem: detach memsegs on cleanup")
Cc: stable@dpdk.org
Signed-off-by: Chengwen Feng <fengchengwen@huawei.com>
Tested-by: Jerin Jacob <jerinj@marvell.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
The P4 language requires marking a header as valid before any of the
header fields are written as opposed to after the writes are done.
Hence, the optimization of replacing the sequence of instructions to
generate a header by reading it from the table action data with a
single DMA internal instruction are reworked from "mov all + validate
-> dma" to "validate + mov all -> dma".
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Fix comparison used to check against the maximum number of learner
table timeouts.
Fixes: e2ecc53582 ("pipeline: improve learner table timers")
Signed-off-by: Harshad Narayane <harshad.suresh.narayane@intel.com>
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Fix segmentation fault due to null pointer dereferencing inside the
"mirror" instruction when number of mirroring slots is set to 0. This
was taking place when the "mirror" instruction was used without the
mirror feature being properly configured, i.e. the API function
rte_swx_pipeline_mirroring_config was not called at initialization.
Fixes: dac0ecd909 ("pipeline: support packet mirroring")
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
rte_xmm_t is a union type which wraps around xmm_t and maps its contents
to scalar structures. Since C++ has stricter type conversion rules than
C, the rte_xmm_t::x has to be used instead of C-casting.
The generated assembly is identical to the code without the fix (checked
both on x86 and RISC-V).
Fixes: 406937f89f ("lpm: add scalar version of lookupx4")
Signed-off-by: Stanislaw Kardach <kda@semihalf.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
rte_xmm_t is a union type which wraps around xmm_t and maps its contents
to scalar structures. Since C++ has stricter type conversion rules than
C, the rte_xmm_t::x has to be used instead of C-casting.
Fixes: f22e705ebf ("eal/riscv: support RISC-V architecture")
Signed-off-by: Stanislaw Kardach <kda@semihalf.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
The cmdline library cmdline_parse() function parses a command and
executes the action automatically too. The cmdline_valid_buffer function
also uses this function to validate commands, meaning that there is no
function to validate a command as ok without executing it.
To fix this omission, we extract the body of cmdline_parse into a new
static inline function with an extra parameter to indicate whether the
action should be performed or not. Then we create two wrappers around
that - a replacement for the existing cmdline_parse function where the
extra parameter is "true" to execute the command, and a new function
"cmdline_parse_check" which passes the parameter as "false" to perform
cmdline validation only.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Tested-by: Weiyuan Li <weiyuanx.li@intel.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
A new event RTE_ETH_EVENT_RX_AVAIL_THRESH should be generated by HW
when number of available descriptors in Rx queue goes below the
threshold.
The threshold is defined as a percentage of an Rx queue size with valid
values from 0 to 99 (inclusive). Zero (default) value disables it.
There is no capability reporting for the feature. Application should
simply try to set required threshold value and handle result.
Add testpmd commands to control the threshold:
set port <port_id> rxq <rxq_id> avail_thresh <avail_thresh_num>
Signed-off-by: Spike Du <spiked@nvidia.com>
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
MCS lock, PF lock and Ticket lock have no arch specific implementation,
there is no need for the extra redirection in headers.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Stanislaw Kardach <kda@semihalf.com>
Add all necessary elements for DPDK to compile and run EAL on SiFive
Freedom U740 SoC which is based on SiFive U74-MC (ISA: rv64imafdc)
core complex.
This includes:
- EAL library implementation for rv64imafdc ISA.
- meson build structure for 'riscv' architecture. RTE_ARCH_RISCV define
is added for architecture identification.
- xmm_t structure operation stubs as there is no vector support in the
U74 core.
Compilation was tested on Ubuntu and Arch Linux using riscv64 toolchain.
Clang compilation currently not supported due to issues with missing
relocation relaxation.
Two rte_rdtsc() schemes are provided: stable low-resolution using rdtime
(default) and unstable high-resolution using rdcycle. User can override
the scheme by defining RTE_RISCV_RDTSC_USE_HPM=1 during compile time of
both DPDK and the application. The reasoning for this is as follows.
The RISC-V ISA mandates that clock read by rdtime has to be of constant
period and synchronized between all hardware threads within 1 tick
(chapter 10.1 in version 20191213 of RISC-V spec).
However this clock may not be of high-enough frequency for dataplane
uses. I.e. on HiFive Unmatched (FU740) it is 1MHz.
There is a high-resolution alternative in form of rdcycle which is
clocked at the core clock frequency. The drawbacks are that it may be
disabled during sleep (WFI), its frequency might change due to DVFS and
it is core-local and therefore cannot be used as a wall-clock. It can
however be used for micro-benchmarking user applications, similarly to
Aarch64's PMCCNTR PMU counter.
The platform is currently marked as linux-only because rte_cycles
implementation uses the timebase-frequency device-tree node read through
the proc file system. Such approach was chosen because Linux kernel
depends on the presence of this device-tree node.
The i40e PMD driver is disabled on RISC-V as the rv64gc ISA has no vector
operations.
The compilation of following modules has been disabled by this commit
and will be re-enabled in later commits as fixes are introduced:
net/ixgbe, net/memif, net/tap, example/l3fwd.
Sponsored-by: Frank Zhao <frank.zhao@starfivetech.com>
Sponsored-by: Sam Grove <sam.grove@sifive.com>
Signed-off-by: Michal Mazurek <maz@semihalf.com>
Signed-off-by: Stanislaw Kardach <kda@semihalf.com>
To allow other projects to easily use DPDK as a subproject, add in the
necessary dependency definitions. Slightly different definitions are
necessary for static and shared builds, since for shared builds the
drivers should not be linked in, and the internal meson dependency
objects are more complete.
To use DPDK as a subproject fallback i.e. use installed DPDK if present,
otherwise the shipped one, the following meson statement can be used:
libdpdk = dependency('libdpdk', fallback: ['dpdk', 'dpdk_dep'])
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Reviewed-by: Ben Magistro <koncept1@gmail.com>
Tested-by: Ben Magistro <koncept1@gmail.com>
This patch replaces instances of zero-sized arrays i.e. those at the end
of structures with "[0]" with the more standard syntax of "[]".
Replacement was done using coccinelle script, with some revert and
cleanup of whitespace afterwards.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Add functions for setting and getting the priority of a thread.
Priorities on multiple platforms are similarly determined by a priority
value and a priority class/policy.
Currently in DPDK most threads operate at the OS-default priority level
but there are cases when increasing the priority is useful. For
example, high performance applications may require elevated priority
levels.
For these reasons, EAL will expose two priority levels which are named
suggestively "normal" and "realtime_critical" and are computed as
follows:
On Linux, the following mapping is created:
RTE_THREAD_PRIORITY_NORMAL corresponds to
* policy SCHED_OTHER
* priority value: (sched_get_priority_min(SCHED_OTHER) +
sched_get_priority_max(SCHED_OTHER))/2;
RTE_THREAD_PRIORITY_REALTIME_CRITICAL corresponds to
* policy SCHED_RR
* priority value: sched_get_priority_max(SCHED_RR);
On Windows, the following mapping is created:
RTE_THREAD_PRIORITY_NORMAL corresponds to
* class NORMAL_PRIORITY_CLASS
* priority THREAD_PRIORITY_NORMAL
RTE_THREAD_PRIORITY_REALTIME_CRITICAL corresponds to
* class REALTIME_PRIORITY_CLASS (when running with privileges)
* class HIGH_PRIORITY_CLASS (when running without privileges)
* priority THREAD_PRIORITY_TIME_CRITICAL
Note that on Linux the resulting priority value will be 0, in
accordance to the documentation that mention the value should be 0 for
SCHED_OTHER policy.
Signed-off-by: Narcisa Vasile <navasile@linux.microsoft.com>
Signed-off-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
A sequence lock (seqlock) is a synchronization primitive which allows
for data-race free, low-overhead, high-frequency reads, suitable for
data structures shared across many cores and which are updated
relatively infrequently.
A seqlock permits multiple parallel readers. A spinlock is used to
serialize writers. In cases where there is only a single writer, or
writer-writer synchronization is done by some external means, the
"raw" sequence counter type (and accompanying rte_seqcount_*()
functions) may be used instead.
To avoid resource reclamation and other issues, the data protected by
a seqlock is best off being self-contained (i.e., no pointers [except
to constant data]).
One way to think about seqlocks is that they provide means to perform
atomic operations on data objects larger than what the native atomic
machine instructions allow for.
DPDK seqlocks (and the underlying sequence counters) are not
preemption safe on the writer side. A thread preemption affects
performance, not correctness.
A seqlock contains a sequence number, which can be thought of as the
generation of the data it protects.
A reader will
1. Load the sequence number (sn).
2. Load, in arbitrary order, the seqlock-protected data.
3. Load the sn again.
4. Check if the first and second sn are equal, and even numbered.
If they are not, discard the loaded data, and restart from 1.
The first three steps need to be ordered using suitable memory fences.
A writer will
1. Take the spinlock, to serialize writer access.
2. Load the sn.
3. Store the original sn + 1 as the new sn.
4. Perform load and stores to the seqlock-protected data.
5. Store the original sn + 2 as the new sn.
6. Release the spinlock.
Proper memory fencing is required to make sure the first sn store, the
data stores, and the second sn store appear to the reader in the
mentioned order.
The sn loads and stores must be atomic, but the data loads and stores
need not be.
The original seqlock design and implementation was done by Stephen
Hemminger. This is an independent implementation, using C11 atomics.
For more information on seqlocks, see
https://en.wikipedia.org/wiki/Seqlock
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Reviewed-by: Ola Liljedahl <ola.liljedahl@arm.com>
Reviewed-by: Chengwen Feng <fengchengwen@huawei.com>
Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
musl lacks __ppc_get_timebase() but has __builtin_ppc_get_timebase()
Signed-off-by: Duncan Bellamy <dunk@denkimushi.com>
Reviewed-by: David Christensen <drc@linux.vnet.ibm.com>
Telemetry commands are now registered through the dmadev library
for the gathering of DSA stats. The corresponding callback
functions for listing dmadevs and providing info and stats for a
specific dmadev are implemented in the dmadev library.
An example usage can be seen below:
Connecting to /var/run/dpdk/rte/dpdk_telemetry.v2
{"version": "DPDK 22.03.0-rc2", "pid": 2956551, "max_output_len": 16384}
Connected to application: "dpdk-dma"
--> /
{"/": ["/", "/dmadev/info", "/dmadev/list", "/dmadev/stats", ...]}
--> /dmadev/list
{"/dmadev/list": [0, 1]}
--> /dmadev/info,0
{"/dmadev/info": {"name": "0000:00:01.0", "nb_vchans": 1, "numa_node": 0,
"max_vchans": 1, "max_desc": 4096, "min_desc": 32, "max_sges": 0,
"capabilities": {"mem2mem": 1, "mem2dev": 0, "dev2mem": 0, ...}}}
--> /dmadev/stats,0,0
{"/dmadev/stats": {"submitted": 0, "completed": 0, "errors": 0}}
Signed-off-by: Sean Morrissey <sean.morrissey@intel.com>
Reviewed-by: Bruce Richardson <bruce.richardson@intel.com>
Reviewed-by: Conor Walsh <conor.walsh@intel.com>
Tested-by: Sunil Pai G <sunil.pai.g@intel.com>
Tested-by: Kevin Laatz <kevin.laatz@intel.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Clarify that once an operation has completed, the output of that
operation is visible to all cores.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Vhost backend of different devices have different features.
Add an API to get vDPA device type, net device or blk device
currently, so users can set different features for different
kinds of devices.
Signed-off-by: Andy Pei <andy.pei@intel.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Add support for VHOST_USER_GET_CONFIG and VHOST_USER_SET_CONFIG.
VHOST_USER_GET_CONFIG and VHOST_USER_SET_CONFIG message is only
supported by virtio blk VDPA device.
Signed-off-by: Andy Pei <andy.pei@intel.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Get_config and set_config are necessary ops for blk device.
Add get_config and set_config ops to vDPA ops.
Signed-off-by: Andy Pei <andy.pei@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
