Age | Commit message (Collapse) | Author | Files | Lines |
|
In order to compare performance number with existed routing based ipip
interface mode 40k tunnels ipsec test.
Signed-off-by: xinfeng zhao <xinfengx.zhao@intel.com>
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
Change-Id: I94816a8c18f08b539114d6ca9d4e1218a2b6ea5f
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: I28295ee6e64ab9674449a98e8d036b39d3d538bf
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: I461899d72e92e5c5db9f64261206ec0e6704e0c8
|
|
- Some older documentation files are not updated yet.
Change-Id: If1717e12308f0e2e76c10024e6eebe68ddeddc9f
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Change-Id: I5b3ee68c8155e62ca6c68e84132d78af0615ee4d
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
+ Move papi_exec resource out of for loops.
Change-Id: I39d75ad2552986f82f7e2c8e3aae2fbc07f042e0
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Change-Id: I3ecad09d2659124a6c07d642a453845ed1779258
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
+ Introduce two explicit handling modes to save memory in scale test.
+ Connect in async mode for both handling modes (to avoid reconnects).
+ Support both pre- and post-37758 VPP PAPI async behavior.
+ Use control-ping in dumps to emulate sync mode.
+ Do not use it for single reply to avoid VPP-2033.
+ Fix call sites to get their replies with correct handling mode.
+ Drain enqueued replies to avoid subsequent errors.
+ Retry if read returns None too early.
+ Update docstrings.
- Complexity issues reported by pylint postponed, needs larger refactor.
- Explicit replace of VAT is done in subsequent changes.
Ticket: CSIT-1547 CSIT-1671
Change-Id: I3c63fa5c578975cc4dd7fce0babe3ab04ec15ed3
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Used black (with line length overridden to 80).
Also manually removed u from inside f strigs.
Change-Id: I5be354e744a697ac68f781b08159de6d1fbdc07e
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: I23f2ab678e6666a1423620fa373261d822030bc8
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: Ia7eefc28645c78ad346d294099ef6258faa9814f
|
|
- Static content will be removed separately
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: I0992e941e1c24552837eaf7e8d3f6564b3cb21c8
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: I1c13afd8cb377c76d65f7099b9652b96613ad087
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: Ibfab25ee82eaa207987f4070cf2386ea9d0781cd
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: Ibb8a59fd4e08bc736494c7ab18e689a55f7a98dc
|
|
- Due to divergence from original design path the RAW was never
consumed. It adds too much code complexity and requires processing
on both storage and compute. Removing entirely to make modeling
efficient.
- log (apparently SSH) section will never be consumed in the way it is
coded in model. This section is also not part of model schema itself
due to the point above.
- Introducing telemetry section that is going to carry telemetry
items required for CDash.
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: I7e0256c6c9715de8ee559eed29dce96329aac97d
|
|
Signed-off-by: Viliam Luc <vluc@cisco.com>
Change-Id: Icd8a5e892bb2c31c01851f731482c692cea4d70a
|
|
- removed remnant debugging code
+ fix compatibility with Metric naming convention
+ fix failures when perf-stat command returns wrong data
+ better handling when perf-stat command fails
Signed-off-by: Viliam Luc <vluc@cisco.com>
Change-Id: I23a064ed09c8cf0394abb5306fb04f4e33c20aa4
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: I9aa88fb094b03888fc30d84edc1deaa406075db4
|
|
Flipping logic depends whether TG and DUT ports are reordered.
# If TG and DUT ports are reordered -> flip
# If TG reordered and DUT not reordered -> don't flip
# If DUT reordered and TG not reordered -> don't flip
# If DUT and TG not reordered -> flip
Change-Id: I7202a17e03e48d112ed5b98e3ef0e3816ce50f14
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Also add 10k tunnels IPSec policy mode test with spd fast path enabled.
spd fast path is to reduce policy adding time, refer to patchset [1] & [2]
for details.
[1] https://gerrit.fd.io/r/c/vpp/+/36095
[2] https://gerrit.fd.io/r/c/vpp/+/36097
Signed-off-by: xinfeng zhao <xinfengx.zhao@intel.com>
Signed-off-by: yulong pei <yulong.pei@intel.com>
Change-Id: I211b38a60d273c46e68a5b5f712090037977c6e6
|
|
After some experiments using simulator (based on real measurment data),
coeff value of 5 gives nice balance between exploration and precision.
Change-Id: I6c416ed5bf3796b2e952ae4d17ab6e6bce5e1b5f
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Without this, far-off samples are considered too important
leading to wrong integration.
This should fix the symptom of secondary stats being too different.
Change-Id: I66496ac25da95eae1f83154f32eb56f941cb20ce
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Add VPP wireguard async mode test suite to use QAT device for crypto.
Also change keyword ipsechw to cryptohw in suite_setup.robot since currently
crypto device is not only used by IPSec.
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
Signed-off-by: xinfeng zhao <xinfengx.zhao@intel.com>
Change-Id: Ibdadb3b09c04b7181415ffd4a248abddc6289075
|
|
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
Change-Id: I7abd546e67fdbe481b204bb6a1ec7e9c654dcdae
|
|
Signed-off-by: Viliam Luc <vluc@cisco.com>
Change-Id: I17ced17a309cc0ac21c5fc94e570c89a456339e2
|
|
Signed-off-by: pmikus <peter.mikus@protonmail.ch>
Change-Id: If898b07e7406bfe28a63e4a793cf2f4e08a53e9d
|
|
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: Ib3e9b9bb7937557f9880fb230856eb96534446d0
|
|
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: I12e7515f7dd04dd493c566ce85ed17f87e2563bb
|
|
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: I276e12e881a4db0a82d5e03107fe153d02a762c4
|
|
Updated for what is available in (last common ancestor of) 2206.
(So newer messages such as ipsec_spd_entry_add_del_v2 are not used yet.)
+ Removed messgages documented as unused.
- Did not check if more become unused.
+ Restored alphabetical order.
Change-Id: I4191c3f8629106f52ce387d03f30f9f973ffbefe
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
+ The main speedup comes from abandoning copy.deepcopy(),
doing shallow list copies (at most) and introcuding copy_fast().
+ Turn into dataclasses whenever possible, use type hints.
+ Simplify the partition search code,
+ It is now clearer the time complexity is O(N*(N+n)),
where N is number of samples, and n is the average size
of the last group of the current record glist.
+ Used black for formatting, so no u"" anymore.
+ Update metadata for 0.3.0 release.
Change-Id: I302203b4d42aeb22be1128e2fe72353a44eae5d0
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Signed-off-by: Viliam Luc <vluc@cisco.com>
Change-Id: I935ee5cf5bd7e760ef29fdc338f981e4e7cc333f
|
|
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: I0a84ce88b318c488ba2d519b20237c88b9f9f1e6
|
|
Change-Id: Id80b5611bb9a5d4398d4b2d00e92cda469949607
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
+ print testpmd and l3fwd pid after start
Testpmd and l3fwd utility can be started but link might not be
ready yet. This fix start the utility on all DUTs and do the
check later.
Signed-off-by: Viliam Luc <vluc@cisco.com>
Change-Id: If476e22f206d9a6a0dd399879a88eafedca92bb6
|
|
The test scenario is about two private networks which communication was
protected by ipsec.
The test suites are to show performance impovement about ipsec esp flow
lookup that offloaded by NIC flow director and rss.
Verified on 3n-clx and Intel E810 NIC environment, with 64B ipsec packet
flow, performance improved ~31% with 1C2T, ~110% with 2C4T, ~250% with 4C8T.
Signed-off-by: xinfeng zhao <xinfengx.zhao@intel.com>
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
Change-Id: I30aec8c5115e5a6fbef88c11d1bef2624029d1b9
|
|
Some tests (bonding) are failing as VPP refuses startup.conf.
That is because the config parameter support has been removed:
https://gerrit.fd.io/r/c/vpp/+/34822
That is VPP commit 8c8531c15c9195c2c82695eff6e70887de67fd88.
According to commit message, it was just a workaround
for Cisco ENIC cards, which are no longer tested in CSIT.
+ Delete 1 robot and 1 python keyword as they become unused.
Change-Id: If53e324a27cfbd4659e49ab8bcd03253642dde96
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: Iba133914882485191ad962e36fae70b6fda0af98
|
|
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: Ie358593f9977d04aca9e50a14e0d14158e1b0cf1
|
|
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: I53226ff6a345ac99c1e2e110712b2ff6197efe95
|
|
Signed-off-by: Cui,Cheng <cheng.cui@intel.com>
Change-Id: I71e257e5f16a887a40ff83d782a9ebdee9d05c34
Signed-off-by: Cui,Cheng <cheng.cui@intel.com>
|
|
The tarball is unnecessarily big, remove superflous directories.
Change-Id: I55f6fae64973cfac34a1dc18cbdc688cd416e8b9
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech>
|
|
- This is actually bug not a feature.
- AB to be added later
- Tested on TREX and iPerf3
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: Ib6f2d13e3b9401a9fb5759e42a8a310ee11b9d41
|
|
The implementation of GTPU offload rx is to use ip4_gtpu flow MARK action
of NIC and vpp flow REDIRECT_TO_NODE and BUFFER_ADVANCE fuction to direct
received gtpu flow to gtpu4-flow-input graph node, skipped ethernet-input,
ip4-input, ip4-lookup, ip4-local, ip4-udp-lookup normal graph node
processing.
Verified on 3n-clx and Intel E810 NIC environment, single core with 64B packet,
performance improve ~33% that compare with pure software way.
Signed-off-by: xinfeng zhao <xinfengx.zhao@intel.com>
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
Change-Id: I2af4589448bdb1729e4ce206a8cf3a1239c61af8
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
|
|
Signed-off-by: Peter Mikus <pmikus@cisco.com>
Change-Id: Id0e7d31dc1368140c2c829fb2fcab009fbbed26d
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: Id4d84aa7268080843b099fd7ab9851234612968b
|
|
- No support for IMIX.
+ Fix a bad bug in padding (most ASTF profiles had wrong frame sizes).
+ Fix a big typo in TCP PPS profiles (s->c was not data, just RST).
+ Control transaction size via ASTF_N_DATA_FRAMES env variable.
- Default value 5 leads to transactions smaller than before.
+ It ensures transaction is one burst (per direction) even for jumbo.
+ Edit autogen to set supported frame sizes based on suite id.
+ Both TCP and UDP use the same values:
+ 64B for CPS (exact for UDP, nominal for TCP).
+ 100B, 1518B and 9000B for TPUT and PPS.
- TCP TPUT achievable minimum is 70B.
+ Used 100B to leave room for possible IPv6 ASTF tests.
+ Separate function for code reused by vpp and trex tests.
- I do not really like the new "copy and edit" approach added here.
+ But it is a quick edit, better autogen refactor is low priority.
+ Consider both established and transitory sessions as valid.
- Mostly for compatibility with 2202 behavior and to avoid ramp-ups.
- Assuming both session states have similar enough VPP CPU overhead.
+ Added a TODO to investigate and maybe reconsider later.
+ Update the state timeout value to 240s.
+ That is the default for TCP (for transitory state).
- UDP could keep using 300s.
+ But I prefer UDP and TCP to behave as similarly as possible.
+ Use TRex tunables to get the exact frame size (for data packets).
- It is not clear why the recipe for MSS has to be this complicated.
+ Move code away from profile init, as frame size is not known there.
+ Change internal profile API, so values related to MSS are passed.
+ Lower ramp-up rate for TCP TPUT tests.
+ Because without lower rate, jumbo fails on packet loss in ramp-up.
+ UDP TPUT ramp-up rate also lowered (just to keep suites more similar).
+ Distinguish one-direction and aggregated average frame size.
+ Update keyword documentation where the distiction matters.
+ One-direction is needed for turning bandwidth limit to TPS limit.
+ Aggregated is needed for correct NDRPDR bandwidth result value.
- TCP TPUT will always be few percent below bidirectional maximum.
+ That is unavoidable, as one direction sends more control packets.
+ Add runtime consistency checks so future refactors are safer.
+ Fail if padding requested would be negative.
+ Fail if suite claims unexpected values for packets per transaction.
+ Edit the 4 types of ASTF profiles to keep them similar to each other.
+ Move UDP TPUT limit value from a field back to direct argument.
+ Stop pretending first UDP packet is not data.
+ Apply small improvements where convenient.
+ Replace "aggregate" with "aggregated" where possible.
+ To lower probability of any future typos in variable names.
+ Avoid calling Set Numeric Frame Sizes twice.
+ Code formatting, keyword documentation, code comments, ...
+ Add TODOs for less important code quality improvements.
- Postpone updating of methodology pages to a subsequent change.
Change-Id: I4b381e5210e69669f972326202fdcc5a2c9c923b
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
The previous path computation code was only correct
for tests with DUT (and TG).
There, we only want to filter by DUT NICs, but not TG NICs
(e.g. to allow testing E810 DUT with MLX TG).
But for tests without DUT (still with TG),
we do indeed want to filter by NICs on TG, so this patch does that.
Change-Id: Ia482c6327af7bca26353ad6e503ae224d90cb3a4
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
This reverts commit 22bfd741614145d86751df2d61ff77af9db63627.
Reason for revert: OMFG this patch is causing 90% perf degradation on AWS
Change-Id: Id081f61ae849678b6b877fbf03309c4b1b0c6df6
Signed-off-by: Peter Mikus <pmikus@cisco.com>
|