Commit 8e24272a authored by Laurent THOMAS's avatar Laurent THOMAS Committed by Robert Schmidt

Fix multiple errors in nr_dlsch_channel_level_median() and nr_dlsch_channel_level()

UBsan fails with errors without this commit. The commit changes:

- avoid two possible integer overflows in nr_dlsch_channel_level() and nr_dlsch_channel_level_median()
- nr_dlsch_channel_level_median() is simplified
- use MAX_ANT in avg to handle more than 4x4 config
parent 3f3a9869
...@@ -156,7 +156,12 @@ static void nr_dlsch_extract_rbs(uint32_t rxdataF_sz, ...@@ -156,7 +156,12 @@ static void nr_dlsch_extract_rbs(uint32_t rxdataF_sz,
uint16_t dlDmrsSymbPos, uint16_t dlDmrsSymbPos,
int chest_time_type); int chest_time_type);
static void nr_dlsch_channel_level_median(uint32_t rx_size_symbol, int32_t dl_ch_estimates_ext[][rx_size_symbol], int32_t *median, int n_tx, int n_rx, int length); static void nr_dlsch_channel_level_median(uint32_t rx_size_symbol,
int32_t dl_ch_estimates_ext[][rx_size_symbol],
int32_t median[MAX_ANT][MAX_ANT],
int n_tx,
int n_rx,
int length);
/** \brief This function performs channel compensation (matched filtering) on the received RBs for this allocation. In addition, it computes the squared-magnitude of the channel with weightings for /** \brief This function performs channel compensation (matched filtering) on the received RBs for this allocation. In addition, it computes the squared-magnitude of the channel with weightings for
16QAM/64QAM detection as well as dual-stream detection (cross-correlation) 16QAM/64QAM detection as well as dual-stream detection (cross-correlation)
...@@ -205,7 +210,7 @@ static void nr_dlsch_channel_level(uint32_t rx_size_symbol, ...@@ -205,7 +210,7 @@ static void nr_dlsch_channel_level(uint32_t rx_size_symbol,
int32_t dl_ch_estimates_ext[][rx_size_symbol], int32_t dl_ch_estimates_ext[][rx_size_symbol],
NR_DL_FRAME_PARMS *frame_parms, NR_DL_FRAME_PARMS *frame_parms,
uint8_t n_tx, uint8_t n_tx,
int32_t *avg, int32_t avg[MAX_ANT][MAX_ANT],
uint8_t symbol, uint8_t symbol,
uint32_t len, uint32_t len,
unsigned short nb_rb); unsigned short nb_rb);
...@@ -271,7 +276,6 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, ...@@ -271,7 +276,6 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue,
const int nr_slot_rx = proc->nr_slot_rx; const int nr_slot_rx = proc->nr_slot_rx;
const int gNB_id = proc->gNB_id; const int gNB_id = proc->gNB_id;
int avg[16];
uint8_t slot = 0; uint8_t slot = 0;
int32_t codeword_TB0 = -1; int32_t codeword_TB0 = -1;
...@@ -431,27 +435,28 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, ...@@ -431,27 +435,28 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue,
if (meas_enabled) if (meas_enabled)
start_meas(&meas); start_meas(&meas);
if (first_symbol_flag == 1) { if (first_symbol_flag == 1) {
int32_t avg[MAX_ANT][MAX_ANT];
nr_dlsch_channel_level(rx_size_symbol, dl_ch_estimates_ext, frame_parms, nl, avg, symbol, nb_re_pdsch, nb_rb_pdsch); nr_dlsch_channel_level(rx_size_symbol, dl_ch_estimates_ext, frame_parms, nl, avg, symbol, nb_re_pdsch, nb_rb_pdsch);
int avgs = 0; int avgs = 0;
int32_t median[16]; int32_t median[MAX_ANT][MAX_ANT];
for (int aatx = 0; aatx < nl; aatx++) for (int aatx = 0; aatx < nl; aatx++)
for (int aarx = 0; aarx < n_rx; aarx++) { for (int aarx = 0; aarx < n_rx; aarx++) {
// LOG_I(PHY, "nb_rb %d len %d avg_%d_%d Power per SC is %d\n",nb_rb, len,aarx, aatx,avg[aatx*n_rx+aarx]); // LOG_I(PHY, "nb_rb %d len %d avg_%d_%d Power per SC is %d\n",nb_rb, len,aarx, aatx,avg[aatx*n_rx+aarx]);
avgs = cmax(avgs, avg[(aatx * n_rx) + aarx]); avgs = cmax(avgs, avg[aatx][aarx]);
// LOG_I(PHY, "avgs Power per SC is %d\n", avgs); // LOG_I(PHY, "avgs Power per SC is %d\n", avgs);
median[(aatx * n_rx) + aarx] = avg[(aatx * n_rx) + aarx]; median[aatx][aarx] = avg[aatx][aarx];
} }
if (nl > 1) { if (nl > 1) {
nr_dlsch_channel_level_median(rx_size_symbol, dl_ch_estimates_ext, median, nl, n_rx, nb_re_pdsch); nr_dlsch_channel_level_median(rx_size_symbol, dl_ch_estimates_ext, median, nl, n_rx, nb_re_pdsch);
for (int aatx = 0; aatx < nl; aatx++) { for (int aatx = 0; aatx < nl; aatx++) {
for (int aarx = 0; aarx < n_rx; aarx++) { for (int aarx = 0; aarx < n_rx; aarx++) {
avgs = cmax(avgs, median[aatx*n_rx + aarx]); avgs = cmax(avgs, median[aatx][aarx]);
}
} }
} }
} *log2_maxh = (log2_approx(avgs) / 2) + 1;
*log2_maxh = (log2_approx(avgs)/2) + 1; // LOG_I(PHY, "avgs Power per SC is %d lg2_maxh %d\n", avgs, log2_maxh);
//LOG_I(PHY, "avgs Power per SC is %d lg2_maxh %d\n", avgs, log2_maxh); LOG_D(PHY, "[DLSCH] AbsSubframe %d.%d log2_maxh = %d (%d)\n", frame % 1024, nr_slot_rx, *log2_maxh, avgs);
LOG_D(PHY, "[DLSCH] AbsSubframe %d.%d log2_maxh = %d (%d,%d)\n", frame % 1024, nr_slot_rx, *log2_maxh, avg[0], avgs);
} }
if (meas_enabled) { if (meas_enabled) {
stop_meas(&meas); stop_meas(&meas);
...@@ -465,15 +470,7 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, ...@@ -465,15 +470,7 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue,
meas.p_time / (cpuf * 1000.0)); meas.p_time / (cpuf * 1000.0));
} }
#if T_TRACER #if T_TRACER
T(T_UE_PHY_PDSCH_ENERGY, T(T_UE_PHY_PDSCH_ENERGY, T_INT(gNB_id), T_INT(0), T_INT(frame % 1024), T_INT(nr_slot_rx));
T_INT(gNB_id),
T_INT(0),
T_INT(frame % 1024),
T_INT(nr_slot_rx),
T_INT(avg[0]),
T_INT(avg[1]),
T_INT(avg[2]),
T_INT(avg[3]));
#endif #endif
//---------------------------------------------------------- //----------------------------------------------------------
...@@ -1108,7 +1105,7 @@ void nr_dlsch_channel_level(uint32_t rx_size_symbol, ...@@ -1108,7 +1105,7 @@ void nr_dlsch_channel_level(uint32_t rx_size_symbol,
int32_t dl_ch_estimates_ext[][rx_size_symbol], int32_t dl_ch_estimates_ext[][rx_size_symbol],
NR_DL_FRAME_PARMS *frame_parms, NR_DL_FRAME_PARMS *frame_parms,
uint8_t n_tx, uint8_t n_tx,
int32_t *avg, int32_t avg[MAX_ANT][MAX_ANT],
uint8_t symbol, uint8_t symbol,
uint32_t len, uint32_t len,
unsigned short nb_rb) unsigned short nb_rb)
...@@ -1137,62 +1134,45 @@ void nr_dlsch_channel_level(uint32_t rx_size_symbol, ...@@ -1137,62 +1134,45 @@ void nr_dlsch_channel_level(uint32_t rx_size_symbol,
avg128D = simde_mm_add_epi32(avg128D,simde_mm_srai_epi32(simde_mm_madd_epi16(dl_ch128[2],dl_ch128[2]),x)); avg128D = simde_mm_add_epi32(avg128D,simde_mm_srai_epi32(simde_mm_madd_epi16(dl_ch128[2],dl_ch128[2]),x));
dl_ch128+=3; dl_ch128+=3;
} }
int32_t *tmp = (int32_t *)&avg128D;
avg[(aatx*frame_parms->nb_antennas_rx)+aarx] =(((int32_t*)&avg128D)[0] + avg[aatx][aarx] = ((int64_t)tmp[0] + tmp[1] + tmp[2] + tmp[3]) / y;
((int32_t*)&avg128D)[1] +
((int32_t*)&avg128D)[2] +
((int32_t*)&avg128D)[3])/y;
// printf("Channel level : %d\n",avg[(aatx<<1)+aarx]); // printf("Channel level : %d\n",avg[(aatx<<1)+aarx]);
} }
} }
simde_mm_empty();
simde_m_empty();
} }
static void nr_dlsch_channel_level_median(uint32_t rx_size_symbol, int32_t dl_ch_estimates_ext[][rx_size_symbol], int32_t *median, int n_tx, int n_rx, int length) static void nr_dlsch_channel_level_median(uint32_t rx_size_symbol,
int32_t dl_ch_estimates_ext[][rx_size_symbol],
int32_t median[MAX_ANT][MAX_ANT],
int n_tx,
int n_rx,
int length)
{ {
for (int aatx = 0; aatx < n_tx; aatx++) {
for (int aarx = 0; aarx < n_rx; aarx++) {
int64_t max = median[aatx][aarx]; // initialize the med point for max
int64_t min = median[aatx][aarx]; // initialize the med point for min
simde__m128i *dl_ch128 = (simde__m128i *)dl_ch_estimates_ext[aatx * n_rx + aarx];
const int length2 = length >> 2; // length = number of REs, hence length2=nb_REs*(32/128) in SIMD loop
short ii; for (int ii = 0; ii < length2; ii++) {
int aatx,aarx; simde__m128i norm128D =
int length2; simde_mm_srai_epi32(simde_mm_madd_epi16(*dl_ch128, *dl_ch128), 2); //[|H_0|²/4 |H_1|²/4 |H_2|²/4 |H_3|²/4]
int max = 0, min=0; int32_t *tmp = (int32_t *)&norm128D;
int norm_pack; int64_t norm_pack = (int64_t)tmp[0] + tmp[1] + tmp[2] + tmp[3];
simde__m128i *dl_ch128, norm128D;
for (aatx=0; aatx<n_tx; aatx++) {
for (aarx=0; aarx<n_rx; aarx++) {
max = median[aatx*n_rx + aarx];//initialize the med point for max
min = median[aatx*n_rx + aarx];//initialize the med point for min
norm128D = simde_mm_setzero_si128();
dl_ch128=(simde__m128i *)dl_ch_estimates_ext[aatx*n_rx + aarx];
length2 = length>>2;//length = number of REs, hence length2=nb_REs*(32/128) in SIMD loop
for (ii=0;ii<length2;ii++) {
norm128D = simde_mm_srai_epi32(simde_mm_madd_epi16(dl_ch128[0],dl_ch128[0]), 2);//[|H_0|²/4 |H_1|²/4 |H_2|²/4 |H_3|²/4]
//print_ints("norm128D",&norm128D[0]);
norm_pack = ((int32_t*)&norm128D)[0] +
((int32_t*)&norm128D)[1] +
((int32_t*)&norm128D)[2] +
((int32_t*)&norm128D)[3];// compute the sum
if (norm_pack > max) if (norm_pack > max)
max = norm_pack;//store values more than max max = norm_pack;
if (norm_pack < min) if (norm_pack < min)
min = norm_pack;//store values less than min min = norm_pack;
dl_ch128+=1; dl_ch128+=1;
} }
median[aatx*n_rx + aarx] = (max+min)>>1; median[aatx][aarx] = (max + min) >> 1;
//printf("Channel level median [%d]: %d max = %d min = %d\n",aatx*n_rx + aarx, median[aatx*n_rx + aarx],max,min); //printf("Channel level median [%d]: %d max = %d min = %d\n",aatx*n_rx + aarx, median[aatx*n_rx + aarx],max,min);
} }
} }
simde_mm_empty();
simde_m_empty();
} }
//============================================================================================== //==============================================================================================
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment