Prefetching packet was missing when do_macswap() was optimized.
Fixes: 62b52877ad ("app/testpmd: batch MAC swap for performance on x86")
Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
The pointer is misused and could set wrong MAC address. As a result, some
of packets can be dropped in receiver side due to MAC address mismatch.
This can be shown as performance degradation.
Bugzilla ID: 188
Fixes: 62b52877ad ("app/testpmd: batch MAC swap for performance on x86")
Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
This patch proposes a slightly different test-pmd quit operation: stop
all devices before starting to close any device. Basically, stop all
moving parts before beginning to remove them. The current test-pmd quit
is stopping and closing each device before moving to the next device.
If all devices in the system are independent of each other, this
difference is usually not important. In case of Soft NIC devices, any
such virtual device typically depends on one or more physical devices
being alive, as it accesses their queues, so this difference becomes
important.
Without this straightforward fix, all the Soft NIC devices need to be
manually stopped before the quit command is issued, otherwise the quit
command can sometimes crash the test-pmd application.
Fixes: d3a274ce9d ("app/testpmd: handle SIGINT and SIGTERM")
Cc: stable@dpdk.org
Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
There's very commonly that more than 4G DDR memory in NIC for HQoS,
so right now the queue threshold size of RED needs to expand to
uint64_t. For struct rte_tm_red_params, it has been fixed, but for
test-pmd TM configuration, it hasn't been fixed. This patch fixes it.
Fixes: bddc2f40b5 ("app/testpmd: add commands for shaper and wred profiles")
Cc: stable@dpdk.org
Signed-off-by: Rosen Xu <rosen.xu@intel.com>
Reviewed-by: Jasvinder Singh <jasvinder.singh@intel.com>
There is need to add boundary for input number from commandline,
If it beyond the definition, code will return error.
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Do four packets macswap in same loop iterate to squeeze more
CPU cycles.
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
The patch optimizes the mac swap operation by taking advantage
of SSE instructions, it only impacts x86 platform.
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Move macswap workload to dedicate function, so we can further enable
platform specific optimized version.
Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Add missing implementation for 64-bit log2 function, and extend
the unit test to test this new function. Also, remove duplicate
reimplementation of this function from testpmd and memalloc.
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Add an rte_bsf64 function that follows the convention of existing
rte_bsf32 function. Also, add missing implementation for safe
version of rte_bsf32, and implement unit tests for all recently
added bsf varieties.
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
We already changed to use generic IPC in pdump since below commit:
commit 660098d61f ("pdump: use generic multi-process channel")
The `rte_pdump_set_socket_dir()`, the `path` parameter of
`rte_pdump_init()` and the `enum rte_pdump_socktype` have been
deprecated since then. This commit removes these deprecated
APIs and also bumps the pdump ABI.
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Acked-by: Reshma Pattan <reshma.pattan@intel.com>
In function cmd_set_mplsogre_encap_parsed(), MPLS label value was
set in mplsogre_encap_conf struct without the required offset.
As a result the value was copied incorrectly into
rte_flow_item_mpls struct.
This patch sets MPLS label value in appropriate location at
mplsogre_encap_conf struct, so it is correctly copied to
rte_flow_item_mpls struct.
Fixes: 3e77031be8 ("app/testpmd: add MPLSoGRE encapsulation")
Cc: stable@dpdk.org
Signed-off-by: Dekel Peled <dekelp@mellanox.com>
Acked-by: Ori Kam <orika@mellanox.com>
In function cmd_set_mplsoudp_encap_parsed(), MPLS label value was
set in mplsoudp_encap_conf struct without the required offset.
As a result the value was copied incorrectly into
rte_flow_item_mpls struct.
This patch sets MPLS label value in appropriate location at
mplsoudp_encap_conf struct, so it is correctly copied to
rte_flow_item_mpls struct.
Fixes: a1191d39cb ("app/testpmd: add MPLSoUDP encapsulation")
Cc: stable@dpdk.org
Signed-off-by: Dekel Peled <dekelp@mellanox.com>
Acked-by: Ori Kam <orika@mellanox.com>
The Bottom-of-Stack (bos) bit of MPLS indicates
whether its the last MPLS layer (1) or not (0).
Indicating that the encapsulating MPLS is the
last MPLS layer in the packet as the default
behavior is more appropriate since multiple
encapsulation actions is not supported.
Fixes: 3e77031be8 ("app/testpmd: add MPLSoGRE encapsulation")
Fixes: 3e77031be8 ("app/testpmd: add MPLSoGRE encapsulation")
Fixes: a1191d39cb ("app/testpmd: add MPLSoUDP encapsulation")
cc: stable@dpdk.org
Signed-off-by: Saleh Alsouqi <salehals@mellanox.com>
Acked-by: Ori Kam <orika@mellanox.com>
Because the user's callback is invoked in eal interrupt callback, the
interrupt callback need to be finished before it can be unregistered
when detaching device. So finish callback soon and use a deferred
removal to detach device is need.
It is a workaround, once the device detaching be moved into the eal in
the future, the deferred removal could be deleted. This patch aim to
add this workaround and refine the function name and the description to
be more explicit and comment the limitation.
Fixes: 2049c5113f ("app/testpmd: use hotplug failure handler")
Signed-off-by: Jeff Guo <jia.guo@intel.com>
The patch fixes the memory allocation for the meter DSCP table.
Fixes: e63b50162a ("app/testpmd: clean metering and policing commands")
Cc: stable@dpdk.org
Signed-off-by: Jasvinder Singh <jasvinder.singh@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
This reverts the patch that enables default RSS action by setting
key=NULL and key_len=0.
In current testpmd implementation a key pointer must exist if
key_len!=0. For example, the following flow rule will cause a
segmentation fault:
flow create 0 <pattern> actions rss queues 0 1 end key_len 40 / end
Fixes: a4391f8bae ("app/testpmd: set default RSS key as null")
Cc: stable@dpdk.org
Signed-off-by: Ophir Munk <ophirmu@mellanox.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
testpmd only sets the L4 len in case of TCP packets.
some PMD's like tap rely on mbuf meta data to calculate csum
This will set the L4 len for UDP packets same as TCP
Fixes: 160c3dc945 ("app/testpmd: introduce IP parsing functions in csum fwd engine")
CC: stable@dpdk.org
Signed-off-by: Raslan Darawsheh <rasland@mellanox.com>
Signed-off-by: Ophir Munk <ophirmu@mellanox.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
There is an error in function search_rx_offload(),
it will break when get unexpected return value from function
rte_eth_dev_rx_offload_name(), but rte_eth_dev_rx_offload_name()
will return some unexpected value indeed.
Fixes: c73a907187 ("app/testpmd: add commands to test new offload API")
Cc: stable@dpdk.org
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Tested-by: Yuan Peng <yuan.peng@intel.com>
When creating an RSS rule without specifying a key (see [1]) it is
expected that the device will use the default key.
A NULL key is used to indicate to a PMD it should use
its default key, however testpmd assigns a non-NULL dummy key
(see [2]) instead.
This does not enable testing any PMD behavior when the RSS key is not
specified. This commit fixes this limitation by setting key to NULL.
[1]
RSS rule example without specifying a key:
flow create 0 ingress <pattern> / end actions rss queues 0 1 end / end
[2]
Testpmd default key assignment:
.key= "testpmd's default RSS hash key, "
"override it for better balancing"
Signed-off-by: Ophir Munk <ophirmu@mellanox.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Set port status to stopped for newly added devices.
Fixes: 2950a76931 ("bond: testpmd support")
Cc: stable@dpdk.org
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
ol_flags can be wrong if DEV_TX_OFFLOAD_VLAN_INSERT is not set in
tx_offloads
Fixes: 3eecba267c ("app/testpmd: cleanup internal Tx offloads flags field")
Cc: stable@dpdk.org
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Y stands for Yellow, R stands for Red.
Fixes: 30ffb4e67e ("app/testpmd: add commands traffic metering and policing")
Cc: stable@dpdk.org
Signed-off-by: Leah Tekoa <leah@ethernitynet.com>
As struct rte_tm_shaper_params defined, the command line of
testpmd should include committed and peak parameters, but
right now the command line doesn't identify whether it's
committed or peak parameter. This patch identifies and
adds the clarify definition
Fixes: bddc2f40b5 ("app/testpmd: add commands for shaper and wred profiles")
Cc: stable@dpdk.org
Signed-off-by: Rosen Xu <rosen.xu@intel.com>
This patch adds telemetry as a dependecy to all applications. Without these
changes, the --telemetry flag will not be recognised and applications will
fail to run if they want to enable telemetry.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: Kevin Laatz <kevin.laatz@intel.com>
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Harry van Haaren <harry.van.haaren@intel.com>
Fix a typo on DEV_RX_OFFLOAD_QINQ_STRIP selection.
Fixes: 0074d02fca ("app/testpmd: convert to new Rx offloads API")
Cc: stable@dpdk.org
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
After probing is done, each new port must be setup.
The new ports are currently guessed by iterating on ports
matching the devargs string used for probing.
When probing a port, it is possible that one more port probing
get triggered (e.g. PF is automatically probed when probing
a VF representor). Such automatic probing will be caught only on event.
The iterator loop may be replaced by a call from the event callback.
In order to be able to test both modes, a command is added
to choose between iterator and event modes.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
The callback for ethdev events was registered on port start,
so it was missing some events.
It is now registered at the beginning of the main function.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
It is possible to request probing of a device twice,
and possibly get new ports for this device.
However, the ports which were already probed and setup
must not be setup again. That's why it is checked whether
the port is already part of fwd_ports_ids array at the beginning
of the function setup_attached_port().
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
The arrays ports_ids and fwd_ports_ids require the same kind
of update when some ports are removed or added.
The functions update_fwd_ports() and remove_unused_fwd_ports()
are merged in the new function remove_invalid_ports().
The part for adding new port is moved into setup_attached_port().
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
The command "port detach" is removing the EAL rte_device
of the ethdev port specified as parameter.
The function name and some comments are updated to make clear
that we are detaching the whole device.
After detaching, the pointer, which maps a port to its device,
is reset. This way, it is possible to check whether a port
is still associated to a (not removed) device.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Previous patch introduces the Tx metadata feature, with unnecessary
restrictions on data entry.
It also used the metadata in txonly fwd engine only.
This fix removes the data entry restrictions on metadata item.
It also implements callback function to add the metadata in every
Tx packet, sent by any fwd engine.
Fixes: c18feafa19 ("app/testpmd: support metadata as flow rule item")
Signed-off-by: Dekel Peled <dekelp@mellanox.com>
Acked-by: Ori Kam <orika@mellanox.com>
Some user and tester require flow RSS to support more types,
so add "all" and "none" to make configuration more easy for users.
Tested-by: Yuan Peng <yuan.peng@intel.com>
Signed-off-by: Wei Zhao <wei.zhao1@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
The hotplug attach/detach features are implemented in EAL layer.
There is a new ethdev iterator to retrieve ports from ethdev layer.
As announced earlier, the (buggy) ethdev functions are now removed.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Reviewed-by: Andrew Rybchenko <arybchenko@solarflare.com>
Example for MPLSoGRE tunnel:
ETH / IPV4 / GRE / MPLS / IP / L4..L7
In order to encapsulate such a tunnel there is a need to remove L2 of
the inner packet and encap the remaining tunnel, this is done by
applying 2 rte flow commands l2_decap followed by mplsogre_encap.
Both commands must appear in the same flow, and from the point of the
packet it both actions are applied at the same time. (There is no part
where a packet doesn't have L2 header).
Decapsulating such a tunnel works the other way, first we need to decap
the outer tunnel header and then apply the new L2.
So the commands will be mplsogre_decap / l2_encap
Due to the complex encapsulation of MPLSoGRE flow action and
based on the fact testpmd does not allocate memory, this patch adds a
new command in testpmd to initialise a global structure containing the
necessary information to make the outer layer of the packet. This same
global structures will then be used by the flow commands in testpmd when
the action mplsogre_encap, mplsogre_decap, will be parsed, at this
point, the conversion into such action becomes trivial.
Signed-off-by: Ori Kam <orika@mellanox.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
MPLSoUDP is an example for L3 tunnel encapsulation.
L3 tunnel type is a tunnel that is missing the layer 2 header of the
inner packet.
Example for MPLSoUDP tunnel:
ETH / IPV4 / UDP / MPLS / IP / L4..L7
In order to encapsulate such a tunnel there is a need to remove L2 of
the inner packet and encap the remaining tunnel, this is done by
applying 2 rte flow commands l2_decap followed by mplsoudp_encap.
Both commands must appear in the same flow, and from the point of the
packet it both actions are applied at the same time. (There is no part
where a packet doesn't have L2 header).
Decapsulating such a tunnel works the other way, first we need to decap
the outer tunnel header and then apply the new L2.
So the commands will be mplsoudp_decap / l2_encap
Due to the complex encapsulation of MPLSoUDP and L2 flow actions and
based on the fact testpmd does not allocate memory, this patch adds a
new command in testpmd to initialise a global structures containing the
necessary information to make the outer layer of the packet. This same
global structures will then be used by the flow commands in testpmd when
the action mplsoudp_encap, mplsoudp_decap, l2_encap, l2_decap, will be
parsed, at this point, the conversion into such action becomes trivial.
The l2_encap and l2_decap actions can also be used for other L3 tunnel
types.
Signed-off-by: Ori Kam <orika@mellanox.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
As described in [1], this series adds option to set metadata value
as match pattern when creating a new flow rule.
This patch introduces additional options in testpmd commands:
- New item type "meta" "data"
- New per-port offload flag "match_metadata".
It also adds commands to configure the tx_metadata value to use:
- New 'config' command takes a 32 bit value and stores it per port:
port config <port_id> tx_metadata <value>
testpmd will add to any Tx packet sent from this port the metadata
value, and set ol_flags accordingly.
- A matching 'show' command is added to read the configured value:
port config <port_id> tx_metadata <value>
[1] "ethdev: support metadata as flow rule criteria"
Signed-off-by: Dekel Peled <dekelp@mellanox.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
After closing a port, it cannot be restarted.
So there is no reason to not free all associated resources.
The last step was done with rte_eth_dev_detach() which is deprecated.
Instead of blindly removing the associated rte_device, the driver should
check if no more port (ethdev, cryptodev, etc) is open for the device.
The last ethdev freeing which were done by rte_eth_dev_detach(),
are now done at the end of rte_eth_dev_close() if the driver supports
the flag RTE_ETH_DEV_CLOSE_REMOVE.
There will be a transition period for PMDs to enable this new flag
and migrate to the new behaviour.
When enabling RTE_ETH_DEV_CLOSE_REMOVE, the PMD must free all its
private resources for the port, in its dev_close function.
It is advised to call the dev_close function in the remove function
in order to support removing a device without closing its ports.
Some drivers does not allocate MAC addresses dynamically or separately.
In those cases, the pointer is set to NULL, in order to avoid wrongly
freeing them in rte_eth_dev_release_port().
A closed port will have the state RTE_ETH_DEV_UNUSED which is
considered as invalid by rte_eth_dev_is_valid_port().
So validity is not checked anymore for closed ports in testpmd.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Reviewed-by: Andrew Rybchenko <arybchenko@solarflare.com>
The testpmd application aim is for testing;
so order of operations should not be enforced.
There was a test to forbid detaching before closing a port.
However, it may interesting to test what happens in such case.
It is possible for a PMD to automatically close the port when detaching.
in order to avoid a crash, it is checked that the port must be stopped
before detaching (as for closing).
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
When detaching a port, the full rte_device is removed.
If the rte_device was hosting several ports,
the testpmd list of ports must be updated for multiple removals.
Signed-off-by: Wisam Jaddo <wisamm@mellanox.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
If user explicitly requested memory to be allocated from a socket via
`port-numa-config` and `ring-numa-config`, and if that socket is
valid, add that socket into socket_ids[] so that mempool allocated for
that socket.
Fixes: dbfb8ec709 ("app/testpmd: optimize mbuf pool allocation")
Cc: stable@dpdk.org
Suggested-by: Ferruh Yigit <ferruh.yigit@intel.com>
Signed-off-by: Phil Yang <phil.yang@arm.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
We use _GNU_SOURCE all over the place, but often times we miss
defining it, resulting in broken builds on musl. Rather than
fixing every library's and driver's and application's makefile,
fix it by simply defining _GNU_SOURCE by default for all
builds.
Remove all usages of _GNU_SOURCE in source files and makefiles,
and also fixup a couple of instances of using __USE_GNU instead
of _GNU_SOURCE.
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
when changing verbosity level it will configure rx/tx callbacks to dump
packets based on the verbosity value as following:
1- dump only received packets:
testpmd> set verbose 1
2- dump only sent packets:
testpmd> set verbose 2
3- dump sent and received packets:
testpmd> set verbose (any number > 2)
4- disable dump
testpmd> set verbose 0
Signed-off-by: Raslan Darawsheh <rasland@mellanox.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
add new rx/tx callback functions to be used for dumping the packets.
Signed-off-by: Raslan Darawsheh <rasland@mellanox.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
verbosity for the received/sent packets is needed in all of the
forwarding engines so moving it to be in a separate function
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Signed-off-by: Raslan Darawsheh <rasland@mellanox.com>
Once the lcore list setting excluded the socket which physical device
attached, it will cause failure. Meanwhile, it will disable Testpmd
cross NUMA scenario.
Fixes: dbfb8ec709 ("app/testpmd: optimize mbuf pool allocation")
Cc: stable@dpdk.org
Signed-off-by: Phil Yang <phil.yang@arm.com>
Reviewed-by: Gavin Hu <gavin.hu@arm.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
add commands to support following actions:
- RTE_FLOW_ACTION_TYPE_SET_MAC_SRC
- RTE_FLOW_ACTION_TYPE_SET_MAC_DST
Signed-off-by: Xiaoyu Min <jackmin@mellanox.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
