| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
In the port section of the PROX configuration file, a new parameter can
be added: all_rx_queues. The default is all_rx_queues=no.
When the rx port variable of a task is listing a port name, all rx
queues will be used when the all_rx_queues is set for that port.
If not set, only one queue of that port will be "consumed", each time
the port is referenced in the rx port variable.
When the all_rx_queues is set for a port, you should NOT reference that
port more than once in the rx port variables.
Change-Id: If9662b1ac07adeec9db88d2a25ca68aed0b9e213
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
As described in https://patches.dpdk.org/patch/60354/, we cannot set
mq_mode for a vertio device to ETH_MQ_RX_RSS. Therefore adding this
extra test before setting mq_mode.
Change-Id: I8a33cf73afaa2cef8b9b012bf8ebe7296e1e692d
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
Change-Id: I7ac1bb78fe00407ff66945fcdf7b5d54c9025214
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
UDP and IP packet length and source UDP port and destination UDP port
were wrongly set for IPv6 packets.
In the *.test files, there is now an optional boolean field (ipv6) in
the [TestParameters] section. When set to true, the generator will use
the IPv6 packet layout to place the right values at the proper offset
in the packet. If not present or set to false, the IPv4 layout will be
used. Note also that packet size needs to be at least 84 bytes for IPV6
traffic since we need space for latency and other information in each
packet.
Packet loss ratio is now printed as a percentage of the total packets
sent.
Change-Id: I2136e87f3032348ae95b69052be7da8461c6303c
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
The IRQ test has been reworked to increase the accuracy of the IRQ
measurements. Results can now also be pushed to a Prometheus push
gateway or to OPNFV's Xtesting. In order to do so, a new file
format.yaml has been introduced. Please use this file now to specify the
details of the PushGateway or the Xtesting server.
Added new test: increment_till_fail.test
Change-Id: I111aae3e099bc03e3d2ddd1014a0301bac356e0b
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
Due to the recent changes in the rapid scripts, Kubernetes environment need to be adjusted.
Ready for test flag file for the runrapid script created during the container start.
TUN device configuration added.
Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
Change-Id: I097aafbd1d1d102aa61b8b86842afcb2d6a51338
|
|
Change-Id: I187ed4f282c6332a7e906bc03b51d26295a475ac
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: Iae0a0d95b5a216463ec7511337be4019fdd4359f
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: I7eeaaa3a0dfa1c0a831863e772b50697c5ff5054
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Using vfio in stead of igb_uio. DPDK 2020.5 is NOT compiling the igb_uio
driver by default any more.
For the l3 mode, we are now using the new PROX vdev feature. This is
achieved by adding the vdev parameter in the port section of the PROX
config files. One of the advantages is that we can now ping a port
managed by PROX: the requests will be forwarded by PROX to the LINUX
stack, who will reply to the ping. A tap device is being created by PROX
to support this feature.
Installing old version of nasm since latest causes some issues.
If not pushgateway is being used, it should be commented out in the
config_file.
vfio is now loaded by specifying the module in /etc/modules-load.d
There is also an after_boot.sh script that will be run by
check_prox_system_setup.sh, which can be used for further configuration
of the instance after boot.
There is now also a file created /opt/rapid/system_ready_for_rapid by
the check_prox_system_setup.sh script. This file is created once all
configuration is done. The runrapid.py script will wait till this file
is created before executing any tests.
Change-Id: Ic5c41af82642066af42134c3323297f5a06f6f72
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
Option for Kubernetes environment clean up (createrapidk8s.py -c) from the test pods added.
Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
Change-Id: I1c6c7090462b55fbadd1d8f1a277470a6eeb2dee
|
|
Fix scripts and build procedure to use /opt/rapid as a main directory for
binaries and config files.
Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
Change-Id: Icb38c37a85a78a6f2de2e751df5fafd624b4f1b3
|
|
All config files are now copied into /opt/rapid. We are not using the
home directory any more. The prox executable is now also copied into the
same /opt/rapid.
Change-Id: I838636167f370bf8a855a3cfb0796c659e27a3ad
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
New files created in 2020 have now the right date in the copyright
notice.
Change-Id: I8eb244beec3d793e819cce1f63c39671f18286a1
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
Change-Id: Iad6882cc28ac2039984b12f6b7a897b81d975d9b
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Also fixed minor issues such as
- prevent using vdev on DPDK <= 17.05 as not supported
- potential segmentation fault after panic (if panic thrown early
in prox initialization)
Change-Id: I49a1962e95a4a44dddadd1fdd906e1b3b232465c
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
- vdev devices now support VLAN.
- kernel tap device can be configured with a netmask (/24 was
always used in previous version).
- when sending a (fake) packet to the kernel, this packet will now
not be routed by the kernel (i.e. it will leave through the
interface configured by PROX). This might change in the futture
when PROX supports multiple VLANs per port. But today it prevents
ARP being sent on management interfaces.
- Log error in case kernel unable to send packet.
- Added support for comments (';') in lua sections.
- Prevent duplication of local_ipv4 - should now be configured in port section
local_ipv4 in core section still supported
Change-Id: I8f9a40fe6ad6f3013ff91b58b44627c7f34081e6
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Before this patch latency measurements were only
supported if signature was disabled.
Change-Id: I565b00f76818ab42e9cca4ddc32f9d857648929d
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: Idd72b593504139ac9d42c33767f8a18395b2cc87
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: I8cbbe1b08af9b5472c304f2fba3635f0a084fdd8
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: I45ea9741be1c552d8f13a7de027994c23995df69
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Following messages are now handled by PROX
- router_solicitation
- neighbour_solicitation
- router_advertisement
- neighbour_advertisement
The following parameters are supported (through the PROX config file)
- sub mode=ndp
This will enable handling of router and neighbour solicitation
and advertisement.
- local ipv6=xxxx:xxxx:xxxxx:xxxx:xxxx:xxxx:xxxx:xxxx
This will configure the local IPv6 address of the port.
This parameter is optional. If not specified, the local IPv6
will be calculated from the EUI.
- global ipv6=xxxx:xxxx:xxxxx:xxxx:xxxx:xxxx:xxxx:xxxx
This will configure the global IPv6 address of the port.
This parameter is optional. If not specified, the global IPv6
will be calculated from the EUI and the router prefix received
from the router.
- ipv6 router=yes
This will cause the core to behave as an IPv6 router
i.e. it will generate Router Advertisement messages
This is only useful in back to back cases, when no real
IPv6 router is present in the setup.
- router prefix=xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx
The router prefix usedin the router advertisement
The prefix will be used by the node to build an IPv6 global
address in cases none were configured.
"Unsollicited NA" parameter has been added within the core/task section.
If set to yes (Unsollicited NA=yes), then an unsollicited neighbour
Advertisement is sent at startup
A same core/task cannot support both l3 and ndp mode.
Those messages will be generated or handled when submode
is set to "ndp":
- neighbour sollicitation
- neighbour advertisement
- router sollicitation
- router advertisement
An example configuration is provided: config/ipv6.cfg in which
port 0 / core 1 plays the role of the generator and port 1 /
core 2 plays the role of the swap.
Change-Id: Id0ab32d384448b4cf767fb4a1c486fc023f4f395
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
In Prox configuration IMIX can be specified using the following syntax:
imix=pkt_size,pkt_size... Up to 127 different pkt sizes can be specified.
Through command line, the following syntax can be used
imix core_id task_id pkt_size,pkt_size... Up to 127 packet sizes can be specified.
When IMIX is enabled, PROX will loop through the list of packet sizes.
PROX supports two different submodes to handle pcap files.
In the 1st submode (default, no submode specified), timestamps from
the pcap file are ignored. In this submode IMIX is supported as well and each
packets in the pcap file will be generated with all IMIX sizes.
In the second submode, called pcap in gen mode, packets are generated at
the timestamp specified in the pcap file. This submode does not support IMIX.
Change-Id: I53cbf1378a5364254285b81e6848350d98561184
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
A virtual device can be created adding the following option on the eal field within
the eal options section:
--vdev=net_ring0
This will cause a new (virtual) dpdk port to be created.
Such a device can then be used in a primary process as other DPDK devices i.e.
a [port x] section must be defined in the configuration file
For a secondary process, there is no need to add --vdev, but, of course
--proc-type=secondary (or --proc-type=auto) must be set within the eal field.
There are two ways to use such a virtual device in a secondary process
1) Using the DPDK port number. Note however that the DPDK port number of virtual devices
in the secondary process might not start from 0 and might hence have to be guessed
e.g. if using 2 virtual devices in primary process (--vdev=net_ring0 --vdev=net_ring1)
the port_id will be 0 and 1 in primary process and 2 & 3 in secondary process.
2) Using the port name, as defined in --vdev in the primary process. In this way,
no [port] section must be defined, and the virtual port names are directly used within
the "rx port" and "tx port" configuration e.g.:
rx port=net_ring0
tx port=net_ring1
Limitations
===========
There seems to be a PROX leak causing DPDK port id to increase when restarting the
secondary process (causing the secondary process configuration through port id more difficult).
As the primary process configures the port (including the number of rx and tx queues)
based on its config file, the secondary process can't setup its own queues.
Simple configurations (such as using gen or lat in primary, and swap in secondary) work as PROX
allocates a RX queue even in gen mode and a TX queue even in lat mode. Better configuration/support
for secondary process should be designed.
The use of dpdk port names is only supported by PROX within "rx port" and "tx port". It is not
supported by other configuration fields such as "tx ports from routing table".
The use of dpdk port names in "rx port" and "tx port" is limited to only 1 port.
Change-Id: Iaa606625da471403713a21df79d3ded4bb91b91e
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
In addition to a mmbuf leak when receiving netlink routes packets
this commit prevents sending the PROX pseudo packet i.e.
the packet PROX sends to the kernel socket to have it generate
an ARP request.
Change-Id: Iabbdecbe412e4b90ac0df7e30fa36d096c5326f0
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
For kernel supported devices, add for vlan tag support
This can be configured through port parameter:
vlan tag=<vlan tag>
If this parameter is set, a vlan tagged interface is created
on top of the tap device
This is only supported for vdev tap devices
When sending (untagged) packet to the tap device (through socket)
the tap should react in sending tagged packet
Note that receiving in L3 mode (w/o tap support) a tagged packet
is not yet supported.
Change-Id: I363fa2f8d2341ac41ef23620222ece1d944bf336
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Two cases where performance has been improved
- When using a gateway from a routing table (l3 submode),
store the mac within the next hop table, to avoid a hash_lookup.
This gives ~10% improvement.
- Read tsc only once per bulk (of up to 64 packets).
This gives ~10% improvement to swap, 4% to gen.
In addition a small fix has been added, preventig "No route" Error
to be written too aften.
Change-Id: I8a7ab74a32f09c8ff47f751ee91e84afee1b2147
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Through this commit BGP messages are forwarded to tap device
Netlink messages are enabled to receive route Updates.
In addition, generating tasks can also specify a routing table
which will be used when sending packets
The routes initialized by the routing table can be changed through
the reception of BGP messages
Change-Id: I187ba9a921885cbc9b209aae5fb654309e3388b8
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: I8f570940d2c4e1fc09ebae03105e436f5caebbd3
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
The crlplane performance has been improved. This is necessary when
handling many IP in L3 mode (i.e. many ARP Requests/Replies).
Before this change, arp timer was updated as soon as an
arp request was tentativaly sent to the master. This means that,
if the request failed to be sent (e.g. ring full) we had to wait
arp_update_time (default 1sec) before trying again.
Now arp_update_time is only used when an arp reply is received.
If a request has been successfully sent, then the arp timer
is updated by 1 second (i.e. we will not send any arp reuest for this
IP within this second)
If we failed to send the request, then the timer is updated by 100 msec
which means we will wait 100 msec before trying again.
A too high value here would have meant that we have to wait a long time
before trying again. A too short value overload the master ring with
request to transmit ARP requests, so that master is unable to handle
arp replies.
arp_update_time is now also partly randomized to avoid all IPs to request
arp request at the beginning of the test. So if arp_update_time is 10 sec
then the configuration sets it to 5 to 15 sec.
As a final note the following parameter (already implemented before)
can be changed to improve ctrlplane performance (at the cost of dataplane)
ctrl path polling frequency=10000
Change-Id: I9660458a22e7442eaa0c83aaa0c9893a15069f98
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Through this commit ARP and ICMP messages are forwarded to the kernel
when vdev tap devices are enabled, as well as PROX l3 mode.
ICMP support has also been added to master (i.e. PROX L3 mode) and to
swap (so when L3 submode is not enabled).
Change-Id: Ie6bf52cbae7171bfca041ff18651d4ec866f44cd
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
This commit adds support for the creation of tap devices.
To create a tap device, and associate it with a DPDK port,
add within the port section:
vdev=<tap_name>
local ipv4=<ip_assigned_to_port>
<tap_name> will appear as a kernel network device, with an IP <ip_assigned_to_port>
ARP packets received from the network on the DPDK interface will be forwarded to the kernel.
1st packet which needs to be generated to a new IP will cause an IP packet
to be sent to the kernel (using socket sendto) forcing the kernel to send an ARP_REQUEST.
In this commit, there is no notification from the kernel about ARP table changes.
This is subject to further commits.
Two config files, gen_tap.cfg and swap_tap.cfg have been created. They work one against each other.
*** Note however that they do not work if running within the same host
(with the port being connected back to back) ***
This is due to the fact that the kernel would bypass the ports and use local interface...
Change-Id: Iadeec0d99e3c693472ea44bdb9163a3bf97df2fa
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Since runrapid.py is now supporting more than 1 dataplane interface, we
need to use a port index in the rapid.env file for the dataplane ports.
In k8sdeployment.py, we only support 1 dataplane port, but we are
writing dp_ip1 and dp_mac1 in the environment file. The copyright dates
are also updated. Some references to VM have also been changed to instance.
Change-Id: Ie1d036c13ef5d6f0aa1f35aa068a9760a1cfc2a9
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
IPV6 test has been added. The generator will now generate IPV6 packets
using NDP to resolve MAC addresses. The hex IP addresses in the inline
definition of the packets in the cfg files is now calculated by a lua
function called in parameters.lua. The hard-coded path /home/centos has
been replaced by ~. An error in the speed calculations has been fixed.
See the RapidTest class for details. The actual number of generated
flows is now reported. If the requested number of flows is not a power
of 2, the lowest power of 2 that is larger than the requested number of
flows will be used and reported.
Change-Id: I36b8d345b4a03c2cf358c19ec353ca8ad2831293
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
All packet size in the *.test files have been replaces with imix sizes.
imix is now a list of packet sizes. If you want to run with only a
pakcet size of 64, you need to specify a list containing only 1 size:
[64].
You can also specify multiple imix values by specifying a list of list
of sizes. The keyworkds in the test files are now imix & imixs.
Change-Id: Iea47c0266b022133b7bbfc9a6811e3199ae41521
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
yaml files have been added to the repo as an example. Please check the
README explaining the output section reqs for this yaml file. There is
also a new file (config_file): it also specifies which yaml files to
use. multiple dataplane interfaces per VM can now be specified and will
appear in the <STACK>.env file. An error in setting the packet size has
been fixed (see set_udp_packet_size for packet size setting details)
Change-Id: Ie89a4940521dac7dd3652acca477739abb9f5497
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
The code has been rewritten in multiple classes and files for better
readability and maintainability. You should now also use Python3 to run
the scripts. Some of the variables in the config files have moved to
other sections. The pushgateway parameter moved to the [varia] section
of the rapid environment file.
Change-Id: Ic61a1bc3baa39447d42b21dfcc0f16c28fcdfa6e
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
Change-Id: I9e90c9bdd4d6e83ce0533fd38e1980dca1841bdb
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
dump packets (e.g. dump 1 0 1) was in some cases dumping more
packets than expected.
Proper count of packets is now dumped.
Change-Id: I0974465d617eda00e77ee87148bf96991ecd3ed7
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: I58fa7a0ab7aaa27e9f9383f7995bb71bf0952618
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Variables were initiales based on the file name of the PROX
configuration files. These decisions are now based on the content of
these files and other variables, rather than just the name of a file.
This commit also fixes some bugs: irq test was not executing any more,
some issues with the push gateway reporting.
Change-Id: I39308effad3c80cf52d96e4d85d92d4b5f0be9a5
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
In case of kubernetes based deployments procedure for test environment
creation is different. Please refer to README.k8s file for more
details.
Dockerfile and dockerimage.sh files can be used to create a docker
image with prox.
There is separate createrapidk8s.py script which is used to create
appropriate number of PODs for testing based on the rapid.pods
configuration.
port_info is a helper application to identify port MAC address.
Change-Id: I73611f066ac54260aa11dd17c173379df06f46a5
Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
* changes:
New PROX commit ID and bug fix
code review comments and PROX commit ID
Code review feedback & back-ground traffic support
Fix run_iteration when not using the sut_sock
Specify latency bucket size & correct calculation
Initial support for percentile latency measurements
Combine code for fixed rate and variable rate tests
Better Latency accuray warning message
Support for ETSI GS NFV-TST 009
|
|
Updated deploycentos.sh with latest PROX commit ID and took another code
review comment into account.
Change-Id: Iedfa695be7496459d990f36e2dc00c8890b19fb3
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
|
|
Implemented the code review comments I got on previous commits and using
commit ID 2cd03b7 to select the PROX version that will be installed in
the VMs through deploycentostools.sh
Change-Id: Ifc4c18c657a8bd57c68a7fc9d885c75ffdcb7b5d
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
Change-Id: I04e0bdc6679a74a8ace26eb30d6f8d2857d6c1c7
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
When PROX detects an error condition at startup, it calls PROX_PANIC
which prints an message before calling rte_panic to end the application.
If ncurses was already started, this results in the message being printed
in ncurses (and prox.log) then leaving ncurses; hence the message is usually
not visible to the user (until it open prox.log), giving the impression of an
unexpected crash.
With this fix, the error message is repeated after closing ncurses, hence the message
will be printed on stdout.
Note that it might also be printed twice on stdout if ncurses was not already started.
As part of this fix, prox.log is now properly closed.
Change-Id: If41875843f1a39bc715f4264b3992c3fa018394e
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
lat all stats <core_id> <task_id> combines "lat stats" and "lat packets" within one command.
This command returns:
- One line in case the core_task parser fails. In that case the line will start with: "error: invalid syntax"
- One set of lines for each core task:
- if core/task is invalid, or not running lat mode, the line will contain (start with) "error...". this will be the only line for that core/task.
- Otherwise the set of lines will be:
- One line as returned by "lat stats". This line will be
- One line starting with "error" in case of error, or
- One line containing lat_min_usec, lat_max_usec, lat_avg_usec, tot_lat_min_usec, tot_lat_max_usec, last_tsc, rte_get_tsc_hz(), lcore_id, task_id
- The return from lat packets i.e.
- One line starting with "error" in case of error, or
- 128 lines - one for each bucket - as returned by "lat packets" command
This command is more useful then lat packets, as it contains the tsc - needed to see whether the sample is new or not.
Change-Id: I04231aa7c5bd2d352ed6f94d40e88d3b411ce744
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
- Latency Distribution (i.e. histograms) has been added when latency is enabled.
- Histograms are now compiled by default.
- A bug has also been fixed, which was causing strange behavior in the histograms
(due to bucket size being overwritten).
- The default bucket size has been changed to "11" which means that the size of each bucket
is now (1 cycle << 11) = 2048 cycles = 1 microsecond at 2GHz. It was (1 << (10 - 1)).
As there are 128 latency buckets, it is not possible to show all of them on the display.
Hence the latency buckets are displayed based on on the assumption that the minimum latency is usually relatively
stable and that the maximum latency change more often.
- The first empty buckets are not shown (i.e. buckets empty for all tasks)
- If more than 15 buckets are non empty, then the buckets are combined, so that a maximum of
15 (bigger) buckets are shown
- If less than 15 buckets are non empty, the following (empty) buckets are shown
(this avoid seeing every x seconds some columns being added and removed).
Change-Id: I27fe6ac0e513a5558e42ff2e74255c55ba79516d
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
Luc Provoost
Yury Kylulin
Xavier Simonart