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.
- 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
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>
- 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
- renaming: sched_cqi -> sched_dl_cqi
- testing: add unit test for sched_dl_cqi wideband cqi mode.
- feature: remove time exponential averaging of dl cqi
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
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.
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.
With the new design, each queue created in the multiqueue object has its own mutex.
Pushing tasks to separate queues will, therefore, not cause contention.
There will be, however, still contention between the popping thread and the pushing threads.
Paging messages can only be transmitted in a subset of subframes - {0, 4, 5, 9}
I avoid now creating objects for the subframes that are not part of this subset
- moved paging record handling to separate class
- parallel access to pending pcch messages by phy workers based on TTI and without common lock
- asn1 pcch message packing now takes place in stack thread, to avoid real-time issues
the NR support in the eNB isn't finished yet but loading
all DL objects even for 4G only isn't required. So only
even attempt to load them if a NR cell is configured.
This feature is useful in scenarios where eNB is behind a NAT and EPC on public cloud
Since eNB's GTP-U can only bind to IP address visible in it's host (usually a private IP beind a NAT), it advertises the same to UE in S1AP message
resulting in traffic from EPC to eNB (DL traffic) not able to reach eNB. This commit solves this issue by providing a means to advertise a different IP address than the one GTP-U UDP socket is bound to.
This IP is usually advertised in InitialContextSetupResponse and E-RABSetupResponse.
In order to make use of this feature, use the following parameter in enb.conf
Example:
gtp_advertise_addr = 213.61.141.82
- use const char* instead of std::string in enumerated<>::to_string() to avoid mallocs.
- Remove the use of "typedef", and use "using" keyword instead.
- Fix rrc_nr::setup_release_c<>::to_string() broken linkage.