aboutsummaryrefslogtreecommitdiffstats
path: root/fr
AgeCommit message (Collapse)AuthorFilesLines
2021-02-04fr: Use dumpcap instead lf tcpdump to capture hdlc net-devicesHarald Welte1-3/+3
Like GBProxy_Tests, we need to use dumpcap instead of tcpdump. Change-Id: I36d5524b11e10ac687c65345aedc4a99e7fda8d8
2021-02-04fr: Don't write TIMEROP_START to ConsoleLogHarald Welte1-1/+1
There is quite a number of them, as we start 100 UE components plus associated structure underneath. Let's avoid spamming the console. Change-Id: I6621ac6094de310e974ce0438d01fca868719eb1
2021-02-04fr: Randomize packet size used in TC_ul2dl_ud()Harald Welte1-1/+15
This allows us to test with a variety of packet sizes up to [close to] the MTU of the underlying transport (configurable by modulepar) Change-Id: I8e38ecf6b270c81bd73ee43b1fa0b259a999c14b
2021-02-04fr: Add TC_ul2dl_udHarald Welte1-0/+42
This test case tests NS-UL-UNITDATA transmit and expectes the exact same LLC PDU to be sent back as NS-DL-UNITDATA. This way we can verify that both uplink and downlink paths are working, and that no messages are lost. It works stable here on my laptop, showing that if we test the TTCN3 NS + BSSGP code over a E1 line against another instance (fr against fr-net) works reliable. Change-Id: Ic115af02207c9b9f4c84fa75890048acb6856c79
2021-02-04fr: UE_CT: Store BVCI + CellId in each UEHarald Welte1-4/+8
Change-Id: I1a25a03e17b3d957763512dc4acda134338df3be
2021-02-04fr/fr-net: Use four parallel FR NS-VC instead of just oneHarald Welte1-2/+26
Related: OS#4974 Change-Id: I366ceea525c09236f5432ba043a757e593d4ed0e
2021-02-03fr: TC_bvc_bringupHarald Welte1-0/+12
Change-Id: I91d61e01158802496a1058748c51506f0472b24e
2021-02-03fr: Generalize and document functionsHarald Welte1-20/+36
This prepares a lot of the FR_Tests functionality to be re-used in more test cases. Change-Id: I95b759ad85d9d1cba1533122672e7ff5628bc014
2021-02-03fr: Add component id in logs; add 15s timeout waiting for BVC-UNBLOCKHarald Welte1-2/+11
Change-Id: If62d54517ee9af828d6fa56d83073260ceced1e8
2021-02-03fr: Make TC_foo() actually terminateHarald Welte1-3/+7
Rather than running in an endless loop (useful for manual tets), terminate each UE_CT after sending 50 packets. Also adjust the ramping and timeouts in a way that it manages to terminate ahead of the guard timeout. Change-Id: I7f40f5c59d399d528072986d067b5014fbd085c4
2021-02-03fr: Implement per-UE guard timeoutHarald Welte1-2/+13
Change-Id: Ie52b1923ca81f7717be446891df1bfdb67ed7ed7
2021-02-03fr: Don't wait indefinitely, but wait until all UE components terminateHarald Welte1-4/+14
Change-Id: Ifdf9c7fb625ccdb467b3b87f79431da18fea5f2e
2021-01-30fr + gbproxy: Start to use AF_PACKET ENOBUFS work-aroundHarald Welte1-0/+1
If send() on AF_PACKET returns ENOBUFS, usleep for 2ms and re-try until it succeeds. Change-Id: Id7bdd2c690eae3bec1df7634944cd73f0bf0b29a Closes: OS#5343
2021-01-17remove unused NS_CodecPort.ttcn/NS_CodecPort_CtrlFunctAlexander Couzens2-2/+2
Change-Id: If885353ec8947309fc7d3ad780448cfc7402e1a4
2020-11-16FR/FRNET: Introduce simplistic simulation of UEsHarald Welte1-0/+75
Let's start a number of per-UE/TLLI component on each BVC, and generate some uplink traffic with random-payload 512-byte LLC frames. The FRNET(SGSN) side simply ignores all of those by means of a CreateCallback. Change-Id: I1b25b4a650d831bb07e9623b76e6c3dcdd71ac88
2020-11-16BSSGP_Emulation: Allow a "CreateCb" to handle unknown inbound TLLIHarald Welte1-1/+2
The existing BSSGP Code assumed that the TLLIs were always known "a priori" by the test case. With the newly-introduced create_cb, the user can provide a function to handle any incoming messages for an unknown TLLL. The default handler behaves like before: fail + terminate. Change-Id: Ice0e145f5a6518ff79547dd851042b7965f38e00
2020-11-12FR/FRNET: Reduce console log verbosityHarald Welte1-1/+1
When doing load simulatin, the amount of console printing becomes a serious throttle factor. We don't need to see the decode of every message. Change-Id: Ic988d1e1d60c9d03d5b70e9b38f109b47569b89e
2020-11-12FR/FRNET: Dynamically create number of BVC at runtimeHarald Welte1-24/+35
We don't want the number of NSE and the number of BVC to be a compile-time choice, but rather something dynamic at rutime. This allows configuration files to define those details. Change-Id: I55b44481b5322deaf78619c1689462d716ddfcec
2020-11-12NS_Emulation: Support multiple NS-VC within one NSE (NS-VCG)Harald Welte1-8/+12
This is something we need to simulate more complex scenarios, particularly in the context of frame relay. Change-Id: If1220852785853f8a5d8de183d5053ddd6ccb958
2020-10-16FR/FRNET: add Frame Relay emulationHarald Welte5-0/+178
Add frame relay testcase for BSS and SGSN side. The test cases require hdlc interfaces (based on dadhi with super channels and no lmi). Change-Id: I95d64dc26a8d2ff02d6cf2bfcd22a97e5481f957
generated by cgit v1.2.3 (git 2.25.1) at 2024-07-04 19:35:04 +0000