aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2013-06-18Implemented estimated C/I algorithm for link adaptation.link_adaptationIvan Kluchnikov8-11/+274
Description of algorithm: 1. Start with initial coding scheme. 2. Transmit RLC blocks until the end of the reporting window (Packet Downlink Ack/Nack message is received) 3. Estimate the average C/I for the previous reporting window and change CS, if it is necessary. 4. If there is more data to transmit return to step 2. Added configuration options to enable/disable link adaptation and setting maximum C/I level for CS1, CS2, CS3, CS4.
2013-05-13Added timing advance support for up and downlink TBFsAndreas Eversberg7-11/+185
The timing advance of any TBF is stored when it ends. Whenever a new TBF with the same TLLI is created (downlink TBF), the stored TA is recalled. This algorithm assumes that the mobile does not move too fast during transfer. Also the mobile must start a connection in order to get correct initial timing advance. This algorithm does not implement the timing advance procedure as defined in TS 04.60. To implement the standard timing advance procedure, the BTS must decode RACH on certain bursts, the mobile is expected to send them. This requires much more complexity to a transceiver like USRP/UmTRX or Calypso BTS. The algorithm was tested at TA >= 8 and works quite well.
2013-05-09Allow to enable realtime priority for the BTS processAndreas Eversberg1-1/+22
I ported the Holger's scheduling patch from sysmobts to osmo-pcu. This is usefull, if PCU uses direct access to the DSP of sysmobts. The latency to respond to a PH-READY_TO_SEND.ind may not be higher than 18ms. Currently we are using nice to increase our priority but for a heavily loaded cell this is not enough. Add an option to enable realtime scheduling and use it in the screenrc. Linux offers two realtime scheduling classes these are SCHED_FIFO and SCHED_RR. For SCHED_FIFO the process is running as long as possible (potentially taking all the CPU and never yielding it), for SCHED_RR the process can still be pre-empted at the end of the timeslice. Using SCHED_RR appears to be the more safe option as a run-a-way sysmobts process will not be able to take all the CPU time.
2013-04-24We shouldn't include the VTY tests in make check by default.Ivan Kluchnikov1-2/+2
2013-04-06Added conditional python-based tests for VTY/config handlingKat4-0/+56
2013-03-19Make the code handle SIGTERM.Holger Hans Peter Freyther1-0/+1
This way the pcu can be easily stopped with a sysvinit script.
2013-03-17Fix: Correctly display SGSN remote IP at debuggingAndreas Eversberg1-1/+4
2013-03-17Introduce new file for various measurementsAndreas Eversberg8-49/+271
The measurements include: - DL bandwidth usage - DL packet loss rate - DL measurements by mobile - UL measurements by BTS In order to receive DL measurements from mobile, it must be enabled via system information message at BSC.
2013-03-17Add RSSI of data blocks to PCU interface0.2.00.2Andreas Eversberg7-11/+14
This bumps the PCU API version to 5 and requires osmo-bts >= 0.2.0.
2013-03-16Use default 'alpha' value of 0Andreas Eversberg1-1/+1
Since we don't know the RX and TX parameters of the BTS that might be used with PCU, the MS should not adapt the TX power from the RX level. So the MS should always transmits with same power. Finding an 'alpha' and 'gamma' value that will result in a constant RX level at the BTS is a task of deployment.
2013-03-10PCU: respect the PCU-side "local port" as configured via L1 IFHarald Welte3-7/+11
This makes sure that the UDP local port of the Gb link is actually set to what is configured via OML from OpenBSC.
2013-03-01Fixed decoding and encoding of CSN_RECURSIVE_TARRAY, CSN_RECURSIVE_TARRAY_1, ↵Ivan Kluchnikov1-4/+4
CSN_RECURSIVE_TARRAY_2. We should offset pointer to data after each iteration.
2013-02-04Fixed memory leaks caused by not freeing msgb in udp_read_cb() function.Ivan Kluchnikov1-0/+1
2013-01-25Fixed closing of PDCH when enabling sysmo-bts' direct DSP accessAndreas Eversberg2-2/+2
If not compiled with this support, the closing function does not exists, so it must not be called. Removed a "#warning", because closing is already supported.
2013-01-23misc: Use the $enableval in the macro and not "yes"Holger Hans Peter Freyther1-2/+2
--enable-sysmocom-bts=no does not work because the enableval variable is not used.
2013-01-17pcu_main: Add '-V' to obtain the (compile) version of osmo-pcu0.1.00.1Harald Welte1-1/+6
2013-01-17VTY: complete the copyright messageHarald Welte1-1/+2
2013-01-16Use PCU's talloc context to allocate libosmogb instancesAndreas Eversberg1-2/+2
This is usefull to identifiy memory leaks while using libosmogb.
2013-01-16Free each fl1h (sysmo-bts' layer 1 handler) when closing BTS<->PCU socketAndreas Eversberg2-1/+10
2013-01-16Fixed memory leaks caused by not freeing bitvectorAndreas Eversberg2-0/+5
Especially each data message from SGSN caused two memory leaks, which resulted in increasing memory usage while receiving date from SGSN.
2013-01-16Use global PCU's talloc context for all talloc callsAndreas Eversberg2-1/+5
This allows to track memory leaks of all memory that is allocated by talloc. Especially bitvector allocations is tracked now.
2013-01-15Various fixes for multi slot allocation algotihmAndreas Eversberg1-19/+44
2013-01-15Get rid of allocating first timeslot at tfi_allocAndreas Eversberg4-37/+37
This simpliefies the allocation process. tfi_alloc is responsible to allocate a TFI, not a time slot. The first time slot available depends on multislot class and on other ongoing TBF (concurrent TBFs), so it is part of the allocation algorithm to select it.
2013-01-14Adding missing initialization of logging to RLCMACTest.cppAndreas Eversberg1-0/+9
2013-01-11add missing examples/Makefile.amHarald Welte1-0/+4
2013-01-11install osmo-pcu.cfg to /etc/osmocom/Harald Welte2-1/+2
2013-01-11add example configuration file as per Andreas' suggestionHarald Welte1-0/+6
2013-01-11make sure to register NS protocol to the VTYHarald Welte1-0/+1
Without calling gprs_ns_vty_init(), the NS specific VTY commands are not activated.
2012-12-18Adding direct access to DSP of sysmoBTS for PDTCH trafficeAndreas Eversberg12-10/+1285
In order to use this feature, sysmoBTS requires option "-M", otherwise the traffic is forwarded through socket interface. This is essential, if PCU runs on processor of sysmoBTS. The reaction time and delay of PDTCH frames could heavily degrade proper packet flow.
2012-12-18Fix: Align TLLI to 32 bit margin, before converting itAndreas Eversberg1-1/+3
2012-12-18Fix of debug names and initial log levelsAndreas Eversberg1-7/+7
2012-12-18Doc: Update on README and tbf.txtAndreas Eversberg2-4/+2
2012-12-18Rework on multislot allocation algorithmAndreas Eversberg3-31/+69
The slots are choosen in a way that later selection of slots for concurrent TBFs will have the same downlink/uplink control channel (PACCH). This is required for polling acknowledge of new TBF assignment on PACCH. The MS will always have the same PACCH uplink slot while beeing in packet transfer mode.
2012-12-18Added paging PS support by Ivan KluchnikovAndreas Eversberg4-3/+82
Original code: c7e7f6868b6f24346424dee904f4e76d3f216ff4 (The code was committed earlier, but got lost somehow.) I added IMSI, so the paging request is sent in correct paging group. Also I excluded rest octets from pseudo length. It is tested and it work.
2012-12-18Use seperate function to parse IMSI from BSSGP messageAndreas Eversberg1-20/+31
2012-12-18Fix: Don't use TIMING_ADVANCE_INDEX in assignment messages, if not usedAndreas Eversberg3-18/+39
Some MS will leave packet transfer mode, if TAI is given, but no timing advance control messages are sent on PTCCH. This could results in permanent loss of larger packets.
2012-12-18Fix: Drop ongoing UL TBF, if another UL TBF is established for same TLLIAndreas Eversberg1-2/+12
This happens, if the mobile looses coverage and establishes a subsequent uplink TBF.
2012-12-18Fix: Handling of final acknowledge on downlink TBFsAndreas Eversberg1-10/+11
It is essential to initialize the first_fin_ack variable. Also it is essential not to free TBF, in case the final ack is not received, but all frames have been acked. In this special case, we resend the final block again and again (and poll), until we receive the final ack.
2012-12-18Fix: Be sure to check if the PACKET DOWNLINK ACK belongs to the polled MSAndreas Eversberg1-2/+8
2012-12-18Adding flow chart diagram of ongoing TBFs and their eventsAndreas Eversberg4-4/+180
It is quite essential. It shows how TBFs are related and helps to estimate states and timers (timeouts) of the MS. In order to use it, it must be defined by a switch at gprs_rlcmac.h.
2012-12-18Fix: Free CS paging messagesAndreas Eversberg1-3/+8
2012-12-18Fix: Removed potential double free bug when receiving NS messagesAndreas Eversberg1-2/+0
2012-10-10Ported fixes and improvements from Wireshark for RLC/MAC control block ↵Ivan Kluchnikov4-128/+342
decoding. Part 1. Generic improvements to the csn.1 dissector include: • Added a flag to the type descriptor structure so that any type can (potentially) flagged as xxx_OR_NULL. This was specifically needed for UINT_OR_NULL for the ms capabilities struct. • Changed the CSN_SERIALIZE so that the length of the length field can be specified. • For CSN_NEXT_EXIST removed the requirement that the next type be CSN_END, to allow truncation of multiple IEs. • For CSN_LEFT_ALIGNED_VAR_BITMAP corrected the handling of bit_offset. • Added a new type CSN_PADDING_BITS to make inspection of padding bits more convenient. • Improved the CSN_RECURSIVE_TARRAY_1 to show the array index. Improvements to the gsm_rlcmac dissector include: • Rework the definition of EGPRS Ack/Nack Description so that the length IE (and absence thereof) is handled correctly. • Added Padding Bits definitions to all PDUs • Change the Multislot_capability_t and Content_t definitions to allow truncation at any IE
2012-10-05Merge branch 'jolly'Ivan Kluchnikov19-570/+3283
Conflicts: src/gprs_bssgp_pcu.cpp src/gprs_rlcmac.cpp src/gprs_rlcmac_data.cpp src/gprs_rlcmac_sched.cpp
2012-10-03Allow setting of seperate coding schemes for uplink and downlinkAndreas Eversberg6-13/+23
2012-10-03Replace switch/case construct by a structure, to define coding schemesAndreas Eversberg3-16/+28
A new attribute at TBF instance indicates the current scheme used.
2012-10-03Fix: Poll correctly for downlink acknowlege, to prevent window stallingAndreas Eversberg1-8/+7
2012-09-28Fix: Cleanly open and close NS instanceAndreas Eversberg2-8/+19
2012-09-27Use PCH confirm from BTS to start downlink packet flowAndreas Eversberg6-10/+99
Since we don't know when the IMM.ASS message is sent on it's paging group on PCH, we will wait for confirm from BTS and start packet flow then.
2012-09-27Free existing UL/DL TBF, if RACH has been received from MSAndreas Eversberg1-0/+17
In this case the mobile has lost existing flows, so it make sense to free them. The TFI(s) can be re-used immidiately, because they are not associated by MS anymore.