Age | Commit message (Collapse) | Author | Files | Lines |
|
scheduler priority"
|
|
Add warning if gateway_ip is configured and l3 is not configured.
If l3 sub mode is not configured, then gateway_ip parameter is ignored
(an error is not returned to make it easy to switch from an l3 sub
mode to default submode).
Change-Id: Ica4a522b037a024dd54bf616d32d69e29d5b8b92
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
priority
Set "realtime scheduling=yes" option to change Linux scheduler policy
to SCHED_RR and increase the priority to the maximum possible for the policy.
Change-Id: I3ecef5cbc3816cf2b56364bb4e806ae5ac093c23
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
In addition to resetting all values set by the user, reset value was also
resetting internal configuration. It resulted that the src mac address
in the outgoing packet was set to whatever was specified in pkt inline,
which is wrong (by default src mac address should be set the the port mac address).
Note that reset value reset any "set values" but does not reset
the packet size.
Change-Id: I9d1cd92be9f5814bd7b104932dd6ffe5df3b6edd
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
* changes:
Some fixes after code review
Cosmetic change when printing warnings
Multiple changes for June release
|
|
Taking into account comments from Patrice and Xavier
Change-Id: Ifdabd1945e074c9ee97b059956f107901392c020
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
On Ubuntu 18.04, the keyboard was not properly handled in PROX
For instance, many keys pressed were silently discarded.
The issue is related to a change in histedit/libedit version.
Recent libedit have (silently) changed the prototype of the get_char function
passed in el_set(el, EL_GETCFN, get_char), with some input parameters
changing from char to wchar_t.
As PROX used the old prototype (char based), this resulted in some
uninitialized field (= garbage) in the wchar_t, causing libedit to discard
the character.
PROX now uses different get_char prototypes, depending of the libedit version
being used.
If PROX was already compiled and a the OS is updated (e.g. from Ubuntu 16.04
to Ubuntu 18.04), this will require a 'make clean'.
Change-Id: Icb0e555a21e13cdaf98172bad17f2f838fb7bc3a
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Warnings can be printed in the following cases:
- When not enough packets are taken into account for measuring packet
latency accuracy.
- When there us a potential network throughput issue, meaning the generator
is generating more packets than the the NIC can handle
- When the generator cannot generate the requested load
If there are no warnings for a certain measurement, nothing gets printed,
suppressing an empty line
Change-Id: Iee07c12142e28dcc0ac406bfed7626731ab08f98
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
- Change inittest in warmuptest. This has now a warmuptime, warmupspeed,
packet size and flow size.
- Change in centos.json. This will now also copy deploycentostools.sh
during image building so that it can be used to re-compile prox in the VM
by typing "./deploycentostools.sh compile"
- runrapid.py parameters needing a file name, need to be complete file
names: no extensions will be added any more by the scripts.
- Changes in createrapid.py to handel the openstack CLI output in a
simpler way.
- The management interface of the VMs can now also be an SRIOV interface.
We have now an extra optional parameter in the VM sections of the
rapiVMs.vms file: SRIOV_mgmt_port
- Changed the name of some sections and keys in the environment file since
runrapid.py will not always communicate to OpenStack VMs. This could be
containers or any other (virtual) machine.
- The previous MachineMap.cfg has been renamed to machine.map
- A new test has been added: monitorswap is just showing the statistics of
a swap (virtual) machine, without generating any packets. This is useful
in situations where an external tester is used.
- Latency and core statistics can now be measured even if there are
multiple PROX cores and tasks running. A new parameter has been added in
the test files with following default value: tasks=[0]. During the
collection of statistics, all tasks in this list will be queried. It is
ok to have a non-existing task in such a query since it will be ignored.
- A --screenlog parameter was added for runrapid.py allowing for more
detailed output on the screen during debugging. No need to check the log
file.
- Previous tests to run multiple packet size with a given flow size and to
run multiple flow size, with a given packet size are now combined by
specifiying 2 lists: packetsizes & flows
- The screen output of this test has also been reworked with more
meaningful column names and the test result is now in the field "core
received". This allows also for faster termination of the test: When the
all packets sent by Gen NIC are received by the cores within the
thresholds for packet loss and latency, the test is now stopped, even we
were requesting more packets to be sent.
Change-Id: I3307e7a972f2140e739f376f146fe875df0303e6
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
The scripts are orrignally written to support an OpenStack deployment.
We are now making changes to support container environments, hence
renaming the directory to rapid.
Change-Id: Icaa2fad83c6da451eee733934151e98e9cd07de4
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
Change-Id: Ifb55091feedb28ffc5947ffe61e6f6a5ce034459
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Accuracy of latency is obtained by writing TX accuracy for packet N
in a later packet i.e. packet N + WINDOW.
If this window is too small, packet N + WINDOW (conveying accuracy
for packet N) might arrive before packet N, resulting in no accuracy
for packet N.
This change increases this window from 64 to 8K packets.
This change should result in a higher number of packets used for
latency measurements when packets are reordered.
This change might have a performance impact as it uses more memory.
Change-Id: I1016fddb66af86605c73a24050238da41cf54152
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
Change-Id: Iaae603a7558666b5794ec4c76f447eb3260a1df9
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
Similar support and comments as for "dp core stats" command
Change-Id: I215fc57c292fbc1f12a3fe95d40a0a80c29b714a
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
The "dp core stats" command supported, as other commands, multiple cores
but not multiple tasks. It now supports both (e.g. dp core stats 1-2 0-1).
This is necessary so that users can collect all statistics from PROX in one
command line.
It will return similar information on the screen or through the socket:
- if the syntax is wrong (e.g. missing task, too many cores or tasks, ...)
=> an error message is printed on the screen, and the single error line
"error: invalid syntax" is returned through the socket
- if the syntax is correct
=> one line is printed or returned for each core/task pair
=> if the core/task pair is invalid, the line reports an error
=> otherwise, the line provides dataplane statistics for the core/task pair,
with the core_id and task_id added, either at the end of the line returned
through the socket, or at the beginning of the line printed on the screen.
This means a change in behaviour when using the socket: before, when an error
happened (e.g. bad task), then nothing was returned which made it difficult for
the script to catch the syntax error.
Similar behaviour could/should be implemented for other commands.
However, care must be taken as some scripts like NSB might count the number
of lines within the output, or the number of items within the line.
This should not be an issue for dp core stats as NSB does not use it, and other
PROX scripts do not count the number of items and will be adjusted to count the
number of lines.
Change-Id: I582a671ae3d2f6493f791e80fc28e70f8e3a38d1
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
and rename parse_core_task() to parse_cores_task()
Change-Id: I256dfb8d02520a70d65d6623feddcc9037938501
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
Warning:
This change breaks backward-compatibility of PROX socket scripting API,
because the "total TX fail" count is reported before the "last TSC"
field.
As a result, PROX scripts that make use of the "dp core stats" command,
such as the core_stats() method in helper-scripts/openstackrapid/prox_ctrl.py,
MUST be reworked to correctly parse the returned set of fields.
Change-Id: I3fe7f37dccee19f5f0b2719f262c6cfa7bad850f
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
by means of calling core_task_is_valid()
Change-Id: Ice64aabc84066633288ff43d4dc9af3f5b378e33
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
Soft CRC was wrong in some cases, such as when compiled with gcc -O2
This might have caused the checksum to be set to 0 in those cases.
Change-Id: Idf01df92876bd2a152ad028293b67a31861a0dfc
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
|
|
PROX version was not udpated since a long time, resulting in
difficulty to easily understand which version is being run.
Improved solution should be to support sha1.
Change-Id: I9d68bd64f52f32544d31d2b8c6bcd0c55c7c228c
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Different PMD have different requirements for minimum rx buffer size.
vmxnet3 has (in DPDK 18.08) requires minimum 1518 + 128 (RTE_PKTMBUF_HEADROOM).
Change-Id: I327ae5b8dcf7eef8dd71fb4e2f0603b9bab6e1d3
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Bug fixes for the script for building an image via the packer tool. This
also installs a newer version of PROX with some bug fixes in PROX. The
*.test files have been simplified. Also introduced a new file
MachineMap.cfg to map test machines on actual VMs.
Change-Id: I69445327ef0a907bc8c1566aaa60c733418c541e
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
When PROX fails to transmit packets (e.g. because switch or NIC
is overloaded) some cleanup needs to be done related to packet id
and latency.
In addition, in some casesi, the clock estimate is slighly over-estimated
which might result, in rare cases, in negative (hence very high) latencies,
due to the extrapolation of timestamps.
This has been workarounded bydecreasing the extrapolation, tolerating up to 1%
clock estimate error
Change-Id: I1ba17dfe0d5e2f9f9167f4f087ed0d96da1293c2
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: Icac622092c35dc197a0f7cb1dcee8afce9f06fa1
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: I73209cfb0534223e24f1f688391062a36e8fc20a
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
The Signature is now written in the pkt template
instead of being written in each packet
Change-Id: I58ca99ff47260b82ce9471b7a47a11edbd14c582
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Improve PROX generator performance by pre-calculating bytes_to_tsc.
This improvement is only implemented for non-pcap
generator, where only few different packet sizes are usually generated.
This change might have a negative performance impact in some cases, if
many different packet sizes are generated, resulting in higher memory usage.
This is the case for instance if random is applied to packet size.
In addition, simplified the rx path, receiving now only MAX_PKT_BURST packets
per handle loop.
Before we were trying to empty the NIC looping on RX packets, ending up
with many rx packets per handle loop. This was used to determine an lower bound
for the time the packet was received.
We now set the lower bound when less than MAX_PKT_BURST has been received.
Change-Id: I1ce813d7e4ac1577ea412c590add5d6f94b36ec7
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Improve PROX generator performance by
- Cleanup some old code checking for unset flags
Change-Id: Icd7ff0a8d17f11f1a1bd09bef4fb40574257c093
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
|
|
|
|
|
|
JIRA: SAMPLEVNF-151
link speed is used in gen and lat latency extrapolations.
Using a link_speed value lower than the actual link speed
might result in errors (e.g. negative latencies).
Negotiated link speed might be reported slowly (as reported through irq)
Hence it is better to use the device capability link speed.
In addition, this remove the check for link speed changes in fastpath.
Change-Id: I0f475fe5e139b046012de6cd0b710e4390735078
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: I5b550a42b27d697f77c4bafab890ba57279f8d13
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Number of mbufs must at least cover for number of rx descriptors,
number of tx descriptors, number of mbuf cached and number of
mbufs handled by the application.
If this is not the case, a warning is returned.
This ony check for the more basic cases. This will not check for
instance for cases with multiples rings where more mbufs might be
cached.
Change-Id: If2c0c9fc76ed4500849d92cf7586bb0b25d8ab22
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
With this use case, the generator is generating at maximum speed
We measure the forwarding rate.
Throughput, latency, number of packets sent and received are printed
every seconds.
Change-Id: I05eb874d6ecefd58fb0c876bccf6a32872306076
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
|
|
* changes:
Prevent dropping ARP packets
Do not add count of non dataplane packets to dropped count
|
|
|
|
In L3 submodes, there were two memory leaks
- when a L3 core was restarted, causing around 2MB leak and a
potential issue after 256 start/stop
- a potential mbuf leak when handling arp replies
Those have been fixed
Change-Id: I348478fa5967936297850432e93667e12b0adac4
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
L3 mode supports two timers:
- arp_update_time, defaulted to 1 second, which makes PROX to send
arp request every second for active flows
- arp_timeout, previously defaulted to 30 seconds, which makes PROX
consider a MAC address as invalid if no arp_reply was received
within those 30 seconds.
Those timers values were hardcoded. They can now be configured through
the configuration file (within the core section), using resp.
"arp update time" and "arp timeout" keywords. Unit is milli seconds.
The default becomes respectively 1 second and 2 weeks.
Change-Id: I35e46e97df32ca44c2cdfae85a20ee015de5d6e1
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
RSS was enabled when rss=yes is set in the configuration or
when multiqueue is being used.
But when RSS was enabled the types of packets on which RSS
hashing was applied might not be supported by the PMD, as
almost every PMD supported a different set of packets.
Change-Id: I1d9516da61df7ef18c1d6953879ee15e33fca077
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Change-Id: Ia9df816d522e5db460d9129109525d5d345d997f
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
Reworked deploycentos1.sh and deploycentos2.sh which is installing
a PROX version that deals with non data plane packet counting for
improved packet loss measurements. Fixed some issues with the packer
scripts.
Change-Id: I64aab09d364ab5abbbc5fc75c73afcbe11fed681
Signed-off-by: Luc Provoost <luc.provoost@intel.com>
|
|
JIRA: SAMPLEVNF-152
When system is overloaded, ARP packets were sometimes dropped, as any
other packets.
This was causing two issues:
- The count of TX non dataplane packets was wrong
- If many consecutive ARP packets were dropped, the underlying switch might
see its ARP timer expiring, causing performance degradation (packets being
broadcasted).
ARP packets are now always sent as no-drop.
Change-Id: I9a86cbf8c4b56a178f86bc789153f1fa49ddf73f
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
RX packets count = TX count + dropped (handled + discarded) count +
non dataplane packets count.
Hence non dataplane packets such as unexpected packets (e.g. dhcp request)
should not b considerded as handled.
Change-Id: I45cef19fed09bb4f86644d56d689a0959a9038db
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
|
|
Gen mode requires a tx ring or a tx port. A configuration without those
should gracefully report an error. This was causing a crash.
Change-Id: Ia30d4d4c87a694da1dfed86587f90b9199c493ab
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
When master core was on socket 1, the rings setup to exchange
messages related to arp between dataplane cores and master core
was not properly setup, resulting in arp not working.
Change-Id: Icc08df81f2e39e53de7c212248cdc7b78d7e2618
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|
|
In L3 mode, prox is extracting IP destination address in the packets.
It uses this destination address to find the MAC address.
If the MAC address is not found, it sends a ARP request.
It also sends ARP request when it realizes that some timeout expired.
However, PROX was using the mbuf of the existing packets (to be forwarded)
to send the ARP. This resulted in packet loss.
Now PROX is generating ARP requests using mbuf from an ARP mempool.
Some clean up was also done.
Change-Id: Icb6083a8cdf88789553ad23c32ca12d6b7ba7f08
Signed-off-by: Xavier Simonart <xavier.simonart@intel.com>
|