aboutsummaryrefslogtreecommitdiffstats
path: root/src
AgeCommit message (Collapse)AuthorFilesLines
2013-12-26alloc: Move the uplink ts selection/pre-assignment out of the codeHolger Hans Peter Freyther1-87/+100
Create a select_ul_slots which is very (95%) similar to the select_dl_slots handling. This needs to be refactored and the todo for multiple uplink slots should be handled too. Add a warning about a potential failure on a busy PCU.
2013-12-26alloc: Move the selection of the first enabled pdch to a new methodHolger Hans Peter Freyther1-14/+23
2013-12-26alloc: Move the tx window calculation to a new methodHolger Hans Peter Freyther1-18/+24
2013-12-26alloc: Move upating of the rx window max for Type==1 to a new methodHolger Hans Peter Freyther1-10/+21
2013-12-26alloc: Move the collision handling to a new methodHolger Hans Peter Freyther1-64/+88
The naming of RX/TX still looks fishy and it should use DL/UL instead. Work on pointers right now. We could introduce a struct with the window min and max and pass this struct. The usage of 'D' appears wrong in the ul_usage..
2013-12-26alloc: Move the selection of downlink slots to a new methodHolger Hans Peter Freyther1-67/+86
This code could evolve into working for Uplink and Downlink. The only different for the final Uplink assignment is that free USFs are collected.
2013-12-25alloc: Simplify the assignment of Trb and TtbHolger Hans Peter Freyther1-6/+2
Put the cases for MS_A/MS_B and MS_A/MS_C together.
2013-12-25alloc: Add a note that the tx window handling differs from rxHolger Hans Peter Freyther1-0/+2
It looks like the code is different (without a reason) a good reason.
2013-12-25tbf: Kill the tsc member as it duplicates dataHolger Hans Peter Freyther5-8/+12
We can just use first_ts and the trx/pdch to extract this information. Avoid duplication of data.
2013-12-25tbf: Fix typo and call it sameHolger Hans Peter Freyther1-1/+1
2013-12-25bts: Rename ts and trx to ts_no and trx_no as we operate on numberHolger Hans Peter Freyther1-6/+6
2013-12-25alloc/tests: Create an allocation test for various scenariousHolger Hans Peter Freyther1-5/+0
The allocation in the TBF/BTS code is quite complex. In parts this is due the assignment and requests occuring under differen circumstances. Attempt to re-create the commono scenarios. Remove the bogus msclass check in gprs_rlcmac_tbf::update as the allocation code will check the ms class anyway.
2013-12-25sched: Attempt to improve the fairness and schedule UL/AL ACK/ASSHolger Hans Peter Freyther2-14/+23
It is possible that certain UL ACK messages are not sent when there are many many uplink and downlink assignments. Try to be more fair and schedule them round-robin. This way no starvation should occur.
2013-12-25sched: Remove unused bts parameter from the internal methodHolger Hans Peter Freyther1-2/+2
2013-12-25llc: Calculate the average queuing delay of the LLC queuesDaniel Willmann3-4/+28
Use a formula like it is used with TCP. This can help to make decisions about if frames should be dropped or not at the time we enqueue them. This code will store two timeval structs in fron the of the actual data and compute the average at the time of the dequeue.
2013-12-25llc: Initialize the LLC frame with garbage to detect wrong usageHolger Hans Peter Freyther1-0/+2
2013-12-18llc: Count the number of frames queued inside the LLC queueHolger Hans Peter Freyther3-2/+10
2013-12-18rlc: Use sizeof() for the memset instead of ARRAY_SIZEHolger Hans Peter Freyther1-1/+1
In this case both will give the same result but it is better to use sizeof. But it is better to use the raw number of bytes instead of the number of elements.
2013-12-12l1if: Include string.h to fix possible coverity issueHolger Hans Peter Freyther1-0/+1
Coverity warns about the below lines of code. strerror is supposed to return a char* and %s should work on char. Let's see if including string.h will make coverity more happy. The pre-processes file looked sane as well. LOGP(DL1IF, LOGL_ERROR, "error writing to L1 msg_queue: %s\n", strerror(errno)); Coverity: CID 1040951, CID 1040950, CID 1040952
2013-12-12decoding: Cosmetic - change rbb decodingDaniel Willmann1-2/+4
2013-12-12rlc: Make the update loop more understandableDaniel Willmann1-7/+11
Add bitnum_to_bsn() as a convenience function to get the BSN, use it in the update handling and ignore rbb for values outside of our tx window.
2013-12-12decoding: Use 'I' and 'R' in rbb array for DLDaniel Willmann3-3/+3
We want to match up rbb decoding and encoding so it helps to use the same chars.
2013-12-12encoding: Factor out encode_rbb to help testingDaniel Willmann2-13/+24
2013-12-12encoding: Use ul_window ssn()/update_rbb() methods when encoding ul ACKDaniel Willmann1-20/+18
Use the ssn() and update_rbb() methods of gprs_rlc_ul_window when encoding the ACK/NACK packet in write_packet_uplink_ack()
2013-12-12encoding: Change wording to match that of 3GPP TS 04.60Daniel Willmann1-1/+1
The spec (in ch. 9.1.8.1) speaks of RECEIVED or INVALID in the RLC receiver, so change the wording in the log message to match.
2013-12-12rlc: Manage the received block bitmap in the ul_windowDaniel Willmann2-0/+20
Added two methods to gprs_rlc_ul_window * ssn() returns the starting sequence number * update_rbb() returns an array of chars representing the state of the received block bitmap. Each element is either 'I'nvalid or 'R'eceived. The rbb is generated from v_n rbb[63] relates to BSN ssn-1 ... rbb[0] relates to BSN ssn-64
2013-12-04rlc: Make the RLC types only operate on the BSNHolger Hans Peter Freyther4-116/+112
The code has an internal optimization to only use window_size space. This means that the caller needed to know that only half of the size was used. Change the API to work on the BSN and do the mapping internally. The compiler should have plenty of opportunity to propagate the constant(s) but this has not been verified.
2013-12-04tbf: Separate the handling for rh->si and call in case of re-transmissionHolger Hans Peter Freyther2-8/+15
In case of a retransmission ack the window again..
2013-11-26tbf/sched: We pick the _last_ entry, rotate the listsHolger Hans Peter Freyther3-0/+13
We always pick the _last_ entry from the lists. Let's rotate so we make it a bit more fair.
2013-11-26tbf/rlc: Move raising of V(Q) into the ul window codeHolger Hans Peter Freyther3-18/+36
2013-11-26rlc/tbf: Move the code to raise V(R) into the ul_window handlingHolger Hans Peter Freyther3-14/+22
2013-11-26tbf/rlc: Move the putting of a block into the rlc codeHolger Hans Peter Freyther3-2/+8
2013-11-26tbf: Move the extraction of the TLLI into a separate methodHolger Hans Peter Freyther2-38/+49
2013-11-26tbf/rlc: Remove the side-effect from the loop.Holger Hans Peter Freyther1-7/+8
2013-11-26tbf: Reduce complexity and deal with only one way of assignmentHolger Hans Peter Freyther1-16/+4
The normal mode of the PCU doesn't really work right now. There is no place to have certain kind of features that are not built and not unit-tested.
2013-11-26tbf/rlc: Move the check if something is in the window out.Holger Hans Peter Freyther2-6/+16
2013-11-26rlc: The bsn has never been used. Remove it from the Xl_windowHolger Hans Peter Freyther1-2/+0
2013-11-26rlc: Test the basic of the gprs_rlc_v_n code for remembering the stateHolger Hans Peter Freyther3-3/+11
2013-11-26tbf/rlc: Move the v_n handling into a dedicated objectHolger Hans Peter Freyther4-7/+36
2013-11-26tbf/rlc: Big change and move the window handling out to the rlcHolger Hans Peter Freyther5-119/+242
The send and receive window is now managed by an external object. There are some issues that can only be solved with C++11 but it is progres and removes some of the spaghetti code. For GPRS the sns and ws is hardcoded. Move that into the window code.
2013-11-26tbf: Update the legend and mention invalid as wellHolger Hans Peter Freyther1-1/+1
We should not see it but the legend should be complete.
2013-11-26rlc: Use uint16_t for the BSN in all cases.Holger Hans Peter Freyther1-2/+2
The sns is only 128 for GPRS (it can be bigger for EDGE) so the uint8_t was enough but it is bad to have inconsistencies for both of them.
2013-11-26tbf/rlc: Move the counting of unacked frames into the RLC V_B classHolger Hans Peter Freyther3-8/+21
Whenwe receive a final ack we assume that all the unacked frames have been acked now. Move the counting to V_B and now the caller can remove the BSN and the "lost" variable which has always been zero.
2013-11-26tbf: Split the handling of DL ACK/NACK into two separate partsHolger Hans Peter Freyther2-59/+69
Split the handling of the final ack and the handling of frames that is moving frames.
2013-11-26tbf/rlc: Change the code and generate the entire state in the V_BHolger Hans Peter Freyther3-15/+21
2013-11-26tbf/rlc: Move the moving of V(A) into the V_B codeHolger Hans Peter Freyther3-8/+23
Move the code that moves the V_B to the first not acked frame. Return how many indexes the V_B was moved and update the V_A in the caller.
2013-11-26tbf/rlc: Move the parsing of RBB to Decoding, move window marking outHolger Hans Peter Freyther5-26/+61
Move the parsing of the bitbmap out of the TBF code into Decoding. Move the updating of the V_B into the V_B class. Add some comments about handling the mod_sns, mod_sns_half parameters by using template code.
2013-11-26tbf: Fix the indention of this routineHolger Hans Peter Freyther1-1/+1
2013-11-26debug_diagram: Remove the special debug_diagram compilation modeHolger Hans Peter Freyther6-175/+0
This approach is somehow flawed. We need/want to debug problems on systems with real traffic and re-compiling it with debug_diagram is not an option. All internal logging needs to be expressive enough so we can understand what is going on (e.g. create a script to post-process the output).
2013-11-26tbf: Simplify the create_dl_acked_block into two partsHolger Hans Peter Freyther2-13/+22
Separate the resend from creating a new block. It shows that the first_fin_ack is always set to FALSE for re-sending. This might not be what we want.
generated by cgit v1.2.3 (git 2.25.1) at 2024-09-22 14:09:56 +0000