1. If VNF is slower than PNF, next scheduling will be done after resync.
2. If VNF is unstable in slot progress, next scheduling will be done in next stable status.
3. SFN/SLOT indications are queued in nrUE.
4. get_queue is used instead of unquque for FIFO operation.
This commit includes several changes.
First, it updates the way we fill the uplink
UCI indication. This still has some todo comments
since all the fields in this ind were hardcoded.
It also updates the pack and unpack functions for
the NR_UCI incations. Lastly, it includes updates
to the gNB UCI scheduler. The updates in the gNB
were cherry picked from develop. There are two updates
to the VNF which allow the VNF timing/slots to be
adjusted based on the received timing_info packets
from the VNF. This adjustment was disabled in IISc's
nFAPI code. For now, it is improving the timing
between the VNF and PNF.
The previous two commits did not correctly check
the gNB sfn and slot. The code now correctly sets
the slot in which the PUSCH is SCHEDULED, not the slot
the rx is expected. Also, the sfn/slot is only updated
if we do in fact schedule a PUSCH.
Although the timing isnt perfect, if the VNF adjusts
in the middle of scheduling a DCI the gNB would try to
send two DCIs for a certain sfn/slot; this commit will
stop that from happening. The timing issue still persits.
This commit properly is filing the dci_format but the
rx_ind and dl_ind has not been updated in the NR UE
prior to sending to gNB. The chnages in this commit
to properly fill the dci_format include:
- filling the dci canidates every slot indiction
- checking for a dl_tti/ul_tti containing a dci
- when dci payload is received we use the filled
dci canidates to update the dci_format
Also, a bug in LTE UE is found where we are indexing
passed the end of ue_sdu array. We added a check for this.
Furthermore, in the nr_ue_scheduler we were utilizing
tx_req but the memory wasnt properly initialized.
Also started handling ul_dci that is received in NR UE.
We now receive dci format 01, 10, and 11 at the NR UE.
The dl_info being filled in NRUE was not memset to 0, that
has been added. Additionally, we will need to get the dci_format
based on the length in the dl_config structs and map it.
For the moment the dci_format values are hard coded.
Also, there is not a field for dci_format in
nFAPI so there is a way it is detected by the NR UE.
For now, we have hardcoded the value in the stub when
we receive the DCI. Lastly, updated the nr_harq_handling
in the gNB.
This commit includes the changes made to gNB.
All of these changes were cherry picked from
the Eurecom/develop branch to ensure the
CFRA procedure is working. Without these changes
there were several seg. faults.
This commit contains the changes to the OAI code in
the NR UE and in nFAPI.
- Filling crc_ind for msg3
- Updating pack/unpack of CRC_ind
- Creating phy abstraction CRC_ind handling in VNF
- Removal of old comments
- Only sending on RACH in phy_stub
- Calling nr_msg3_transmitted in scheduler for NSA mode
The send_standlone_msg will now be able to send
any type of uplink message and we fill the RX
indication in the stub_schedule_response and
are trying to send to gNB. Need to update the NR nfapi.
- Hardcoding the freq_index when filling RACH
- Not handling DCI unless dci_rel_15 global is allocated
- Verifies ra_rnti matches in prach pdu and RACH
- Updates handling of DCI and RX from dl and tx reqs
Changed NFAPI_mode check in the gNB scheduler
to ensure that the frame is correctly updated for
msg3. Also added comments in NR UE for handling
the RAR and implementing the dci/rx_ind handover
for the RAR procedure.
The fix added in this commit may not be correct and
should be reviewed by Eurecom. For now we set the
future frame based on the PRACH recevied frame + 1.
Also fixed the copying of RACH ind into UL_INFO global.
RAR is being passed to NR UE.
Several other changes in this commit include:
- Setting the number of PDUS in the ul_config
to be zero when we will it (inside of the
get_ul_config() function.)
- We reset the nprach flag to 0 for each slot.
- We return out of the prach scheduler if
mac->ul_config == NULL
- Added error checking for array indexing
- Created a dummy scheduler response function.
- Allocated ul_config_req initially
Also calling the new L1_nsa_prach_procedures()
function after using the nr_ue_scheduler. Added
some error checking in the nr_ue_scheduler. Added
some error checking in the process_nsa_msg in both
the NR and LTE UE
Had to hack the code to schedule the prach.
Need to verify this is ok because ul_config = NULL.
Also had to change logic for calling fill_schedule_response.
However, we are not properly calling the
nr_ue_get_rach() function. The RA is being
properly initialized. In the nr_ue_get_rach()
the pRACH is not scheduled. Tried to add in
the scheduling of this, but it crashes here.
The other changes include revering the
nr_ue_get_rach() changes from previous commit.
Also, changed a few log levels and reverted
a previous change in the NR UE RRC layer that
was incorrect.
Also removed several warnings by properly
declaring nr-nFAPI functions and removed
declaration of unused functions. Also, reverted
eNB configuration file changes.
-fsanitize=address detected that we were trying to access bytes
past the end of a block of malloc'ed memory.
Specifically, in this code:
```
} else if (ENB_NAS_USE_TUN) {
if( LOG_DEBUGFLAG(DEBUG_PDCP) )
log_dump(PDCP, pdcpData, sizeToWrite, LOG_DUMP_CHAR,"PDCP
output to be sent to TUN interface: \n");
ret = write(nas_sock_fd[0], pdcpData, sizeToWrite);
```
-fsanitize=address said:
```
==80==ERROR: AddressSanitizer: heap-buffer-overflow on address
0x61100004ffdc at pc 0x7f57c5f576a5 bp 0x7f57bb53c240 sp
0x7f57bb53b9e8
READ of size 108 at 0x61100004ffdc thread T7
0x61100004ffdc is located 0 bytes to the right of 220-byte region
[0x61100004ff00,0x61100004ffdc)
```
So, the code was trying to access the first byte after a block of
heap memory.
sizeToWrite was calculated like this:
```
int sizeToWrite= sizeof (pdcp_data_ind_header_t) +
pdcpHead->data_size;
```
There were a few other similar invocations of write() in the same
function used the wrong size. That sizeToWrite calculation
should be used only when the header is being sent, too, which
happens in only one place in this function.
With this commit, our tests pass and -fsanitize=address is happy.
The global mbms_rab_id was defined as `uint16_t` in
eNB_scheduler_mch.c but `int` in other places. So, code that
wrotes this global would clobber the two bytes that follow the
variable in memory.
Found by `-fsanitize=address`.
For some reason the htons invocation got deleted which broke
the VNF P7 interface.
Also fixed some build issues, not sure how these got onto the
develop branch to begin with.
Removing references to memcpy_dl_config_req and
memcpy_tx_req because these functions are leaking
memory. For our standalone mode we have created
new/modified memcpy_dl/tx_req functions.
David Kim