5g-team
/
srsRAN_4G
mirror of https://github.com/pvnis/srsRAN_4G.git
2
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
8,493 Commits
2 Branches
0 Tags
243 MiB
fix-cqi
master
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'e6683b7387'
${ noResults }
Commit Graph

1115 Commits (e6683b7387455ff3fd90732808cb43accad42e66)

Author SHA1 Message Date
Andre Puschmann 483a216bd5 ue,stack: refactor handling of radio bearears in UE stack 
this is a rather large commit that is hard to split because
it touches quite a few components.

It's a preparation patch for adding NR split bearers in the next
step.

We realized that managing RLC and PDCP bearers for both NR and LTE
in the same entity doesn't work. This is because we use the LCID
as a key for all accesses. With NR dual connectivity however we
can have the same LCID active at the same time for both LTE and NR
carriers.

The patch solves that by creating a dedicated NR instance for RLC/PDCP
in the stack. But then the question arises for UL traffic on, e.g. LCID 4
what PDCP instance the GW should use for pushing SDUs. It doesnt' know
that. And in fact it doesn't need to. It just needs to know EPS
bearer IDs. So the next change was to remove the knowledge of what
LCIDs are from the GW. Make is agnostic and only work on EPS bearer IDs.

The handling and mapping between EPS bearer IDs and LCIDs for LTE
or NR (mainly PDCP for pushing data) is done in the Stack because
it has access to both.

The NAS also has a EPS bearer map but only knows about default and
dedicated bearers. It doesn't know on which logical channels they
are transmitted.
 3 years ago
Francisco e70b8c2472 sched,bugfix: fix SI scheduler logging 3 years ago
Andre Puschmann 4b069d9b86 enb,mac: fix concurrent access to cell structs 
this patch moves the rwlock that protects the UE database outside
to also protect the cell struct.

It also adds a missing write guard when setting the cell
 3 years ago
Andre Puschmann ff72c78745 enb,mac: fix indention 3 years ago
Pedro Alvarez aae266e848 Added configuration option for S1AP SCTP bind port 3 years ago
yagoda ce30b7e1ce fixing eMBMS after new ue_db architecture and scheduler changes 3 years ago
Pedro Alvarez 181b903117 Small log fix for RNTI in integrity failure notifications 3 years ago
Ismael Gomez dd9889e07c Add carrier index to uplink PCAP captures 3 years ago
Bedran Karakoc c280ac5ad6 Introducing extended NGAP functionality 
Remove unused logger for now in ngap::ue to pass tests
 3 years ago
Francisco 5a82a63bb5 bugfix,srsenb: fix intraenb handover softbuffer allocation during handover 3 years ago
Francisco 2a009eb340 sched,feature: Enable configuring maximum coderate for SIB allocations 3 years ago
Francisco 6ababf401b feature,sched: allow setting initial DL CQI in enb.conf 3 years ago
Francisco 4cf5a8b25a fix: avoid configuring the phy mbsfn from the rrc, when phy is not fully initialized 3 years ago
Francisco f55e1c5a73 fix,srsenb: start stack before the radio to avoid stack methods to be called when stack is not fully initialized 3 years ago
Francisco e8983b88f0 refactor: use enb_cc_idx instead of ue_cc_idx in mac::ue 3 years ago
Francisco 69b7a7eb19 bugfix,srsenb: the lcids beloging to SRB1 and SRB2 were not being reactivated correctly during intra-eNB handover 3 years ago
Francisco a0ba2d576d remove unused return variable 3 years ago
Francisco 048d487921 bugfix,gtpu: send end marker via the forwarding tunnel, when the source enb receives ue ctxt release command. Improve GTPU unit testing 3 years ago
Francisco 8cbbebdea2 bugfix,scheduler: make sure UL harq pending data is cleared up when the last retx falls in a measurement gap 3 years ago
Francisco 51c9fc514d refactor,sched: fix log lines related with meas gap collisions 3 years ago
Pedro Alvarez 163976f050 Fix releasing on on RRC and not on S1AP on integrity failure. 3 years ago
Pedro Alvarez a53b321ab8 Release UE upon reception of integrity error. 3 years ago
Pedro Alvarez 299d2ee35d Added interface to make the PDCP notify the RRC of integrity check 
failures.
 3 years ago
Francisco 5c7ea6f6df bugfix,srsenb,rrc: remove redundant calls to setup PDCP/RLC bearer state during reestablishment 3 years ago
Francisco fd3901e4a2 bugfix,srsenb,rrc: delete old context right away after sending Reestablishment msg. 
Following ReestablishmentRequests should use the most recent rnti as old context.
 3 years ago
Francisco 7e77da6a15 mac,optimization: use rnti memory pool to create mac::ue objects 3 years ago
Francisco fd1d73666c bugfix,srsenb,mac: avoid sanity check for UL PDU length if CRC=KO 3 years ago
Francisco 76004a3054 enb,mac,refactor: mac::ue now uses byte_buffer_pool for UL PDUs 3 years ago
Francisco 75eebd6b5d bugfix,gtpu: close forwarding tunnel during reestablishment due to handover failure 3 years ago
Francisco 597f993706 sched,improvement: filter out Msg3 UL SNR estimates from the TPC control loop 3 years ago
Andre Puschmann c0be8187c8 rlc interface: convert entire read_pdu() interface to return uint32_t 
we had it returning int but had a bug in using the return value properly,
i.e. handling when -1 was returned in RLC TM.

Thinking about it more, it doesn't make sense to have a negative return
value here anyway. Either the RLC can return a PDU or not. If it can't the
returned lenght is zero.
 3 years ago
Andre Puschmann aaffdeb326 rrc,rlc: add method to signal protocol failure to RRC 
general-purpose method for lower layers to signal protocol
failures to higher layers, i.e. RRC.

In the current case, implement a direct release of the UE (enb) or
a reestablishment (UE).
 3 years ago
Andre Puschmann 80a89ec4bf enb,mac: protect concurrent access to ue_db from different threads 
race condition between PHY worker(s) and Stack thread when
accessing ue_db member. RWlock needs to be hold.

partly fixes #2932
 3 years ago
faluco 3df72cb6ab Move out a free call in a RT thread when doing lots of rachs. 3 years ago
Francisco ceaef3523f sched,feature: Enable setting the target PHR below which the maximum UL grant size starts decreasing 3 years ago
Francisco 486823e5de sched,bugfix: The PHR handling scheme now considers the number of PRBs of the UL grant that carried the PHR. Before, it was assuming the most conservative scenario of Nprb=1 3 years ago
Francisco f7aef3ffc1 sched,bugfix: fix nof_retx update when UL HARQs are resumed 3 years ago
Ismael Gomez 5cef780789 Radio-link KO notification. Measgap fix 3 years ago
Francisco 92e272ce3e limit minimum UL grant size to accommodate both BSR and RLC headers 3 years ago
Francisco 89406b07d6 make eNB UL SNR initial value an exponential average alpha configurable 3 years ago
Francisco 2a2c56391c sched,refactor: Improve DL/UL grant log lines in the scheduler to print CFI as well 3 years ago
Francisco 47494a668d rrc,bugfix: fix reestablishment crash due to pointer invalidation 3 years ago
Francisco 3f75538660 rrc,feature: Use t301 to set time that the enb waits for the UE RRCConnectionSetupComplete or ReestablishmentComplete 3 years ago
Francisco 0349949cce bugfix, handover: disable inactivity timer during s1 handover in target enb 3 years ago
Francisco 19af55b63d bugfix, handover: Avoid sending UEContextReleaseRequest from Target eNB during S1 Handover due to UL KOs 3 years ago
Pedro Alvarez c23034e1a7 Make sure the S1AP state is reset after sending a successful/unsuccessful message 3 years ago
Pedro Alvarez 18e2149d68 Sending Initial Context Setup Response after RRC reconfiguration. 3 years ago
Francisco 2cdda3b6db sched,enhancement: allow PUSCH allocations when PHICH falls in measurement Gap by resuming UL HARQ 3 years ago
Andre Puschmann fcba67ca84 enb,rrc: fix logging text 3 years ago
Andre Puschmann 8322887135 enb,rrc: increase RRC PDU queue size to 128 3 years ago
Francisco a8790abd7e fix centos7 tuple compilation issue 3 years ago
Francisco 200006c4b7 bugfix,scheduler: avoid offset to increase decrease unboundedly when mcs is equal to 0 or max_mcs 3 years ago
Francisco 0dafe4dd58 fix coderate derivation in PDCCH based on aggregation level 3 years ago
Francisco a587d20252 allow specification of subset of valid measurement gap offsets in rr.conf 3 years ago
Francisco 0996e50dce bugfix, s1ap: allow source enb to release old UE ctxt in case a connectionRequest arrives with repeated m-TMSI 3 years ago
Francisco fe0dc77c03 bugfix, s1ap: fix uninitialized variable access in s1ap_test 4 years ago
Francisco 7076dbfef8 feature, scheduler: Add configuration parameter to set minimum TTI interval between TPCs 4 years ago
Francisco 4627ae9802 bugfix, scheduler: effectively disable adaptive mcs when target bler is unspecified 4 years ago
Francisco c75777c618 bugfix, sched: avoid compilation issue due to uninitialized variable access 4 years ago
Francisco 329cf02cd0 bugfix, sched: ensure max_coderate is respected in scheduler 4 years ago
Francisco a8cccd7a10 bugfix, sched: fix aggregation level derivation to avoid L > 3 4 years ago
Francisco dedd9c09a7 feature, scheduler: Add abitility to set minimum aggregation level in PDCCH, and disable adaptive aggregation level based on target BLER 4 years ago
Francisco fb4a363abd fix formatting errors in scheduler 4 years ago
Francisco 1f1665bab7 improve adaptive offset adjustment log lines 4 years ago
Francisco c5029fc266 feature - implement adaptive link adaptation in DL and UL 4 years ago
Francisco 8811b7c9a7 fix handling of CRNTI CE in the RRC. The scheduler should not allocate more grants for the new rnti after this point. 4 years ago
Francisco 59513c0eed fix bearer reactivation in case the RRC session is recovered after max KO 4 years ago
Andre Puschmann fbb151af2d enb,rrc: fix bearer re-activation after max KO 
this patch makes sure that if RBs have been deactivated after
max KO on either the DL or the UL, they are reactivated if either
of them is working again.

UL/DL are always activated together, even if just one of them
recovers. The KO counter (or timer), however, is not stopped implicitly.
 4 years ago
Francisco 59db527481 fix reestablishment reject logging level 4 years ago
Francisco a17e3b71e7 set minimum snr under which the UL SNR estimate is not updated 4 years ago
Francisco 8df6ed07c6 sched small fixes - use 0.930 instead of 0.932 for max coderate. Allow computing mcs/tbs based on max coderate directly. Other small fixes 4 years ago
Francisco 41bc7e4810 avoid letting repeated s1ap::user_release calls to deem the user as in zombie state. Use timer instead to auto-remove user from s1ap if no response from mme arrives 4 years ago
Francisco e07ef0aa18 avoid releasing ues doing s1 handover before TS1relocoverall expiry 4 years ago
Francisco b093576634 avoid sending UEContextReleaseRequest from Source eNB during S1 Handover. 4 years ago
Francisco f2a56c9139 simplification of multiqueue design. 
- the consumer does multi-staged waiting:
  1. spins first across all queues in a RR fashion
  2. each queue access is done with a try_lock.
  3. if the try_lock fails, it increases the number of spins needed
  2. if no queue had data, the consumer sleeps for 100 usec.
- no differentiation between queues, in terms of notification features
 4 years ago
Francisco 2d6a878826 change interface to create multiqueue queues 4 years ago
Francisco f705735093 optimization - only allow some multiqueue queues to notify the multiqueue consumer of the arrival of tasks 4 years ago
yagoda f90f6f0c7e adding MAC 0 padding support 4 years ago
Francisco a83c0ebe3b increase starting CFI when subframe contains control allocations (e.g. SIB or RAR) when cell bandwidth <= 5 MHz 4 years ago
Francisco 0a1ae0e4ad redesign and fix enb rrc message logs 4 years ago
Francisco 8d30c88330 bugfix - avoid calling rlf release timer handler concurrently 4 years ago
Francisco f0f99c1e4f update sched benchmark, created unit test to cover rbg search algorithms, improved algorithms to search for empty rbgs 4 years ago
Francisco 22f1f2d56c move ngap files into their own folders 4 years ago
Francisco 1488606156 refactor - create s1ap dedicated folder 4 years ago
Francisco 0e0835a805 move scheduler rbg/prb/cce mask and interval types and associated helper functions to a separate file 4 years ago
Francisco 7c0649bc24 delete erab if gtpu tunnel creation failed 4 years ago
Francisco c6d543efdf fix conversions of drb id to erab id in srsenb 4 years ago
Andre Puschmann 167200f5cd rrc_bearer_cfg: replace fixed E-RAB to DRB mapping 
also make sure we don't assign LCIDs beyond the possible
number.

possible fix for https://github.com/srsran/srsRAN/issues/658

Co-authored-by: herlesupreeth <herlesupreeth@gmail.com>
Co-authored-by: Francisco <francisco.paisana@softwareradiosystems.com>
 4 years ago
Francisco fa34aaa16e ensure the DL mask count is reduced enough to fit required bytes 4 years ago
Francisco 4d3ff0d139 implementation of basic search for optimal rbgmask for subband CQI 4 years ago
Francisco f8b6eae2bf use of min DL cqi across all PRBS to compute required DL PRBs 4 years ago
Francisco 8564996eaf fix sched_dl_cqi method to compute cqi-optimal rbgmask 4 years ago
Francisco 485c6ee129 subband CQI feature - add K parameter in sched_interface::ue_cfg_t and update it from RRC 4 years ago
Francisco 84ad6dd8ee Integrated new sched_dl_cqi class in sched_ue_cell 
- extend sched_dl_cqi interface to allow getting tti when cqi was last updated
- extend sched_dl_cqi to quickly get average cqi across the whole bandwidth
 4 years ago
Francisco 96ab16f1c7 s1ap fix - release old rnti (instead of new one) when a new rnti with the same tmsi is found. Furthermore, we now make sure to warn log if the enb is unable to release the old rnti 4 years ago
Francisco 6ed617f429 s1ap, bugfix - check if MME-UE-S1AP-ID has been yet assigned before sending UL NAS Transport. If it hasn't log error and abort. 4 years ago
Francisco 32228389a9 fix latest static analysis warnings in rrc paging 4 years ago
Francisco 46cfdaf9e5 Handle case when old rnti gets removed during RRC reestablishment 
This is achieved via:
- keep restarting the inactivity timer whenever a ReestablishmentRequest is received
- in case the RRC ReestablishmentComplete is very late and the old rnti was removed anyway,
  abort RRC reestablishment procedure and release new rnti
 4 years ago
Andre Puschmann b2825d1280 mac_controller: after maxRetx disable ALL UE bearers 
the reasoning here is that the only way to recover
from the maxRetx event is the UE attempting a reestablishment.

No further traffic, neither control nor data is scheduled
until then.
 4 years ago
Andre Puschmann b61be7878a enb,rrc: split RLF counter and timer handling for DL/UL/RLC 
this patch splits the counter and timer handling for PHY DL, PHY UL,
and RLC errors and makes sure that, for example, a successful DL
does not cancel the UL RLF timer, and vice versa.

They all use the same timeout value which is user-configurable.
 4 years ago