maki/openairinterface5g
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Select best out of multiple TDAs

Browse Source 
A new function get_best_ul_tda() allows to select the "best" TDA out of
a number of TDA. Specifically, get_best_ul_tda() considers in how many
RBs a given TDA might be allocated, and multiplies this number of RBs
with the number of symbols in the TDA to form a "score". The TDA with
the best "score" is selected, considering it maximizes resource usage.

In phytest scheduler, we use this function to verify that the given TDA
can be really allocated in all selected slots.
This commit is contained in:
Robert Schmidt
2025-08-14 12:50:03 +02:00
parent 71e01bbaa9
commit 6981783ee1
3 changed files with 84 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c
+7
View File
@@ -262,6 +262,13 @@ void nr_ul_preprocessor_phytest(gNB_MAC_INST *nr_mac, post_process_pusch_t *pp_p
// TODO implement beam procedures for phy-test mode
int beam = 0;
const NR_tda_info_t *tda_p;
const int n_tda = get_num_ul_tda(nr_mac, sched_slot, tda_info.k2, &tda_p);
DevAssert(n_tda > 0);
/* check only the first TDA: we are only interested in finding out if this TDA fits completely */
int rb_s = rbStart, rb_l = rbSize;
get_best_ul_tda(nr_mac, beam, tda_p, 1, sched_frame, sched_slot, &rb_s, &rb_l);
DevAssert(rb_s == rbStart && rb_l == rbSize);
const int buffer_index = ul_buffer_index(sched_frame, sched_slot, slots_frame, nr_mac->vrb_map_UL_size);
uint16_t *vrb_map_UL = &nr_mac->common_channels[CC_id].vrb_map_UL[beam][buffer_index * MAX_BWP_SIZE];
openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_ulsch.c
+69 -3
View File
@@ -92,6 +92,65 @@ int get_num_ul_tda(gNB_MAC_INST *nrmac, int slot, int k2, const NR_tda_info_t **
return diff;
}
static void get_max_rb_range(const uint16_t *vrb_map_ul, const uint16_t *ulprbbl, uint16_t mask, int *rb_start, int *rb_len)
{
int best_start = *rb_start;
int best_len = 0;
int current_start = *rb_start;
int current_len = 0;
for (int rb = *rb_start; rb < *rb_len; ++rb) {
if (ulprbbl[rb] == 0 && (mask & vrb_map_ul[rb]) == 0) {
current_len++;
continue;
}
/* this RB is blocked for UL, or already in use */
if (current_len > best_len) {
best_start = current_start;
best_len = current_len;
}
current_start = rb + 1; /* will start in next RB, or updated later */
current_len = 0;
}
if (current_len > best_len) {
best_start = current_start;
best_len = current_len;
}
*rb_start = best_start;
*rb_len = best_len;
}
const NR_tda_info_t *get_best_ul_tda(const gNB_MAC_INST *nrmac, int beam, const NR_tda_info_t *tdas, int n_tda, int frame, int slot, int *rb_start, int *rb_len)
{
/* there is a mixed slot only when in TDD */
const frame_structure_t *fs = &nrmac->frame_structure;
const int index = ul_buffer_index(frame, slot, fs->numb_slots_frame, nrmac->vrb_map_UL_size);
uint16_t *vrb_map_UL = &nrmac->common_channels[0].vrb_map_UL[beam][index * MAX_BWP_SIZE];
DevAssert(n_tda <= 16);
const NR_tda_info_t *best_tda = tdas;
uint64_t score = 0;
int check_rb_start = *rb_start;
int check_rb_len = *rb_len;
for (int i = 0; i < n_tda; ++i, tdas++) {
int start = check_rb_start;
int len = check_rb_len;
uint16_t tda_mask = SL_to_bitmap(tdas->startSymbolIndex, tdas->nrOfSymbols);
get_max_rb_range(vrb_map_UL, nrmac->ulprbbl, tda_mask, &start, &len);
uint64_t s = (uint64_t)tdas->nrOfSymbols * len;
if (s > score) {
best_tda = tdas;
score = s;
*rb_start = start;
*rb_len = len;
}
LOG_D(NR_MAC, "%4d.%2d tda k2 %ld mask 0x%04x has PRB start/len %d/%d score %ld\n", frame, slot, tdas->k2, tda_mask, start, len, s);
}
return best_tda;
}
bwp_info_t get_pusch_bwp_start_size(NR_UE_info_t *UE)
{
NR_UE_UL_BWP_t *ul_bwp = &UE->current_UL_BWP;
@@ -2600,19 +2659,26 @@ static void nr_ulsch_preprocessor(gNB_MAC_INST *nr_mac, post_process_pusch_t *pp
* "cell-specific", i.e., the UE will add it on the computation. */
int k2 = fs_get_diff(fs, *next, current) - koffset;
DevAssert(k2 > 0);
// we assume that all beams have the same symbol utilization in all RBs for
// simplification (but this might not be true). Otherwise, we would need to
// check it separately for all beams and give a list of TDAs (one for each
// beam) in pf_ul(), which is complex. We should rewrite pf_ul() to
// schedule for one beam only instead (and for one BWP).
int beam = 0;
const NR_tda_info_t *tda_info = NULL;
int n_tda = get_num_ul_tda(nr_mac, next->s, k2, &tda_info);
if (n_tda == 0) /* no TDA fulfills this */
break;
/* TODO: get_num_ul_tda() might indicate multiple TDAs to use (e.g., one
* handling SRS), this will be fixed in a follow-up commit. */
int rb_start = 0;
int rb_len = bw;
tda_info = get_best_ul_tda(nr_mac, beam, tda_info, n_tda, next->f, next->s, &rb_start, &rb_len);
DevAssert(tda_info->valid_tda);
int tda = seq_arr_dist(&nr_mac->ul_tda, seq_arr_front(&nr_mac->ul_tda), tda_info);
AssertFatal(tda >= 0 && tda < 16, "illegal TDA index %d\n", tda);
int len[num_beams];
for (int i = 0; i < num_beams; i++)
len[i] = bw;
len[i] = rb_len;
/* proportional fair scheduling algorithm */
int sched = pf_ul(nr_mac, pp_pusch, tda, tda_info, nr_mac->UE_info.connected_ue_list, max_dci, num_beams, len);
LOG_D(NR_MAC, "run pf_ul() at %4d.%2d with tda %d k2 %d (ULSCH at %4d.%2d) scheduled %d last_dl %d\n", frame, slot, tda, k2, next->f, next->s, sched, last_dl);
openair2/LAYER2/NR_MAC_gNB/mac_proto.h
+8
View File
@@ -413,6 +413,14 @@ bool add_UE_to_list(int list_size, NR_UE_info_t *list[list_size], NR_UE_info_t *
NR_UE_info_t *remove_UE_from_list(int list_size, NR_UE_info_t *list[list_size], rnti_t rnti);
int get_dl_tda(const gNB_MAC_INST *nrmac, int slot);
int get_num_ul_tda(gNB_MAC_INST *nrmac, int slot, int k2, const NR_tda_info_t **first_idx);
const NR_tda_info_t *get_best_ul_tda(const gNB_MAC_INST *nrmac,
int beam,
const NR_tda_info_t *tdas,
int n_tda,
int frame,
int slot,
int *rb_start,
int *rb_len);
int get_cce_index(const gNB_MAC_INST *nrmac,
const int CC_id,