merged with branch fix-frame-rx-and-memory and conflicts solved

cmake_targets/CMakeLists.txt

@@ -2397,7 +2397,7 @@ add_executable(nr-uesoftmodem
   ${s1ap_h}
 #  ${OPENAIR_BIN_DIR}/messages_xml.h
   ${OPENAIR_DIR}/common/utils/threadPool/thread-pool.c
-  ${OPENAIR_DIR}/executables//nr-uesoftmodem.c
+  ${OPENAIR_DIR}/executables/nr-uesoftmodem.c
   ${OPENAIR_DIR}/executables/nr-ue.c
   ${OPENAIR1_DIR}/SIMULATION/TOOLS/taus.c
   ${OPENAIR_TARGETS}/ARCH/COMMON/common_lib.c
@@ -2431,8 +2431,8 @@ add_executable(nr-uesoftmodem-nos1
   ${s1ap_h}
 #  ${OPENAIR_BIN_DIR}/messages_xml.h
   ${OPENAIR_TARGETS}/RT/USER/rt_wrapper.c
-  ${OPENAIR_TARGETS}/RT/USER/nr-ue.c
-  ${OPENAIR_TARGETS}/RT/USER/nr-uesoftmodem.c
+  ${OPENAIR_DIR}/executables/nr-ue.c
+  ${OPENAIR_DIR}/executables/nr-uesoftmodem.c
   ${OPENAIR1_DIR}/SIMULATION/TOOLS/taus.c
   #${OPENAIR_TARGETS}/COMMON/create_tasks_ue.c
   ${OPENAIR_TARGETS}/ARCH/COMMON/common_lib.c

cmake_targets/autotests/test_case_list.xml

@@ -1117,6 +1117,26 @@
       <nruns>3</nruns>
     </testCase> 
     
+    <testCase id="015107">
+      <class>execution</class>
+      <desc>nr_ulschsim Test cases. (Test1: 106 PRB),
+                                    (Test2: 217 PRB),
+                                    (Test3: 273 PRB)</desc>
+      <pre_compile_prog></pre_compile_prog>
+      <compile_prog>$OPENAIR_DIR/cmake_targets/build_oai</compile_prog>
+      <compile_prog_args> --phy_simulators  -c </compile_prog_args>
+      <pre_exec>$OPENAIR_DIR/cmake_targets/autotests/tools/free_mem.bash</pre_exec>
+      <pre_exec_args></pre_exec_args>
+      <main_exec> $OPENAIR_DIR/targets/bin/nr_ulschsim.Rel15</main_exec>
+      <main_exec_args>-R 106 -m9 -s13 -n100
+                      -R 217 -m15 -s15 -n100
+                      -R 273 -m19 -s20 -n100</main_exec_args>
+      <tags>nr_ulschsim.test1 nr_ulschsim.test2 nr_ulschsim.test3</tags>
+      <search_expr_true>PUSCH test OK</search_expr_true>
+      <search_expr_false>segmentation fault|assertion|exiting|fatal</search_expr_false>
+      <nruns>3</nruns>
+    </testCase>
+
     <testCase id="015110">
       <class>execution</class>
       <desc>dlsim_tm4 test cases (Test 1: 10 MHz, R2.FDD (MCS 5), EVA5, -1dB), 

common/utils/LOG/log.h

@@ -45,14 +45,15 @@
 #include <time.h>
 #include <stdint.h>
 #ifndef __STDC_FORMAT_MACROS
-#define __STDC_FORMAT_MACROS
+  #define __STDC_FORMAT_MACROS
 #endif
 #include <inttypes.h>
 #ifndef _GNU_SOURCE
-#define _GNU_SOURCE
+  #define _GNU_SOURCE
 #endif
 #include <pthread.h>
 #include "T.h"
+#include <common/utils/utils.h>
 /*----------------------------------------------------------------------------*/
 
 #ifdef __cplusplus
@@ -165,7 +166,7 @@ extern "C" {
     {"NAS",         DEBUG_NAS},\
     {"UE_TIMING",   UE_TIMING},\
     {NULL,-1}\
-}
+  }
 
 
 
@@ -256,7 +257,7 @@ typedef struct {
   log_component_t         log_component[MAX_LOG_COMPONENTS];
   char                    level2string[NUM_LOG_LEVEL];
   int                     flag;
-    char*                   filelog_name;
+  char                   *filelog_name;
   uint64_t                debug_mask;
   uint64_t                dump_mask;
 } log_t;
@@ -277,9 +278,9 @@ void log_set_instance_type (log_instance_type_t instance);
 log_t *g_log;
 #else
 #ifdef __cplusplus
-   extern "C" {
+extern "C" {
 #endif
-extern log_t *g_log;
+  extern log_t *g_log;
 #ifdef __cplusplus
 }
 #endif
@@ -289,7 +290,7 @@ extern log_t *g_log;
 /*----------------------------------------------------------------------------*/
 int  logInit (void);
 void logRecord_mt(const char *file, const char *func, int line,int comp, int level, const char *format, ...) __attribute__ ((format (printf, 6, 7)));
-void log_dump(int component, void *buffer, int buffsize,int datatype, const char* format, ... );
+void log_dump(int component, void *buffer, int buffsize,int datatype, const char *format, ... );
 int  set_log(int component, int level);
 void set_glog(int level);
 
@@ -344,10 +345,10 @@ int32_t write_file_matlab(const char *fname, const char *vname, void *data, int
 /*   optname                            help                                                paramflags       XXXptr                    defXXXval              type   numelt */
 /*------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------*/
 #define LOG_GLOBALPARAMS_DESC { \
-{LOG_CONFIG_STRING_GLOBAL_LOG_LEVEL,    "Default log level for all componemts\n",              0,  	      strptr:(char **)&gloglevel,    defstrval:log_level_names[2].name,    TYPE_STRING,    0}, \
-{LOG_CONFIG_STRING_GLOBAL_LOG_ONLINE,   "Default console output option, for all components\n", 0,  	      iptr:&(consolelog),            defintval:1,                          TYPE_INT,       0}, \
-{LOG_CONFIG_STRING_GLOBAL_LOG_OPTIONS,  LOG_CONFIG_HELP_OPTIONS,                               0,  	      strlistptr:NULL,               defstrlistval:NULL,                   TYPE_STRINGLIST,0} \
-}
+    {LOG_CONFIG_STRING_GLOBAL_LOG_LEVEL,    "Default log level for all componemts\n",              0,         strptr:(char **)&gloglevel,    defstrval:log_level_names[2].name,    TYPE_STRING,    0}, \
+    {LOG_CONFIG_STRING_GLOBAL_LOG_ONLINE,   "Default console output option, for all components\n", 0,         iptr:&(consolelog),            defintval:1,                          TYPE_INT,       0}, \
+    {LOG_CONFIG_STRING_GLOBAL_LOG_OPTIONS,  LOG_CONFIG_HELP_OPTIONS,                               0,         strlistptr:NULL,               defstrlistval:NULL,                   TYPE_STRINGLIST,0} \
+  }
 
 #define LOG_OPTIONS_IDX   2
 /*----------------------------------------------------------------------------------*/
@@ -360,22 +361,22 @@ int32_t write_file_matlab(const char *fname, const char *vname, void *data, int
 // debugging macros
 #define LOG_F  LOG_I           /* because  LOG_F was originaly to dump a message or buffer but is also used as a regular level...., to dump use LOG_DUMPMSG */
 #  if T_TRACER
-     /* per component, level dependant macros */
+/* per component, level dependant macros */
 #    define LOG_E(c, x...) do { if (T_stdout) { if( g_log->log_component[c].level >= OAILOG_ERR    ) logRecord_mt(__FILE__, __FUNCTION__, __LINE__,c, OAILOG_ERR, x)     ;} else { T(T_LEGACY_ ## c ## _ERROR, T_PRINTF(x))   ;}} while (0)
 #    define LOG_W(c, x...) do { if (T_stdout) { if( g_log->log_component[c].level >= OAILOG_WARNING) logRecord_mt(__FILE__, __FUNCTION__, __LINE__,c, OAILOG_WARNING, x) ;} else { T(T_LEGACY_ ## c ## _WARNING, T_PRINTF(x)) ;}} while (0)
 #    define LOG_I(c, x...) do { if (T_stdout) { if( g_log->log_component[c].level >= OAILOG_INFO   ) logRecord_mt(__FILE__, __FUNCTION__, __LINE__,c, OAILOG_INFO, x)    ;} else { T(T_LEGACY_ ## c ## _INFO, T_PRINTF(x))    ;}} while (0)
 #    define LOG_D(c, x...) do { if (T_stdout) { if( g_log->log_component[c].level >= OAILOG_DEBUG  ) logRecord_mt(__FILE__, __FUNCTION__, __LINE__,c, OAILOG_DEBUG, x)   ;} else { T(T_LEGACY_ ## c ## _DEBUG, T_PRINTF(x))   ;}} while (0)
 #    define LOG_T(c, x...) do { if (T_stdout) { if( g_log->log_component[c].level >= OAILOG_TRACE  ) logRecord_mt(__FILE__, __FUNCTION__, __LINE__,c, OAILOG_TRACE, x)   ;} else { T(T_LEGACY_ ## c ## _TRACE, T_PRINTF(x))   ;}} while (0)
-    /* macro used to dump a buffer or a message as in openair2/RRC/LTE/RRC_eNB.c, replaces LOG_F macro */
+/* macro used to dump a buffer or a message as in openair2/RRC/LTE/RRC_eNB.c, replaces LOG_F macro */
 #    define LOG_DUMPMSG(c, f, b, s, x...) do {  if(g_log->dump_mask & f) log_dump(c, b, s, LOG_DUMP_CHAR, x)  ;}   while (0)  /* */
 #    define nfapi_log(FILE, FNC, LN, COMP, LVL, F...)  do { if (T_stdout) { logRecord_mt(__FILE__, __FUNCTION__, __LINE__,COMP, LVL, F)  ;}}   while (0)  /* */
-     /* bitmask dependant macros, to isolate debugging code */
+/* bitmask dependant macros, to isolate debugging code */
 #    define LOG_DEBUGFLAG(D) (g_log->debug_mask & D)
 
-     /* bitmask dependant macros, to generate debug file such as matlab file or message dump */
+/* bitmask dependant macros, to generate debug file such as matlab file or message dump */
 #    define LOG_DUMPFLAG(D) (g_log->dump_mask & D)
 #    define LOG_M(file, vector, data, len, dec, format) do { write_file_matlab(file, vector, data, len, dec, format);} while(0)/* */
-     /* define variable only used in LOG macro's */
+/* define variable only used in LOG macro's */
 #    define LOG_VAR(A,B) A B
 #  else /* T_TRACER: remove all debugging and tracing messages, except errors */
 #    define LOG_I(c, x...) do {logRecord_mt(__FILE__, __FUNCTION__, __LINE__,c, OAILOG_INFO, x) ; } while(0)/* */
@@ -425,7 +426,7 @@ static __inline__ uint64_t rdtsc(void) {
 
 extern double cpuf;
 
-static inline uint64_t checkTCPU(int timeout, char * file, int line) {
+static inline uint64_t checkTCPU(int timeout, char *file, int line) {
   static __thread uint64_t  lastCPUTime=0;
   static __thread uint64_t  last=0;
   uint64_t cur=rdtsc();
@@ -434,25 +435,29 @@ static inline uint64_t checkTCPU(int timeout, char * file, int line) {
   uint64_t CPUTime=CPUt.tv_sec*1000*1000+CPUt.tv_nsec/1000;
   double microCycles=(double)(cpuf*1000);
   int duration=(int)((cur-last)/microCycles);
+
   if ( last!=0 && duration > timeout ) {
     //struct timespec ts;
     //clock_gettime(CLOCK_MONOTONIC, &ts);
     printf("%s:%d lte-ue delay %d (exceed %d), CPU for this period: %lld\n", file, line,
            duration, timeout, (long long)CPUTime-lastCPUTime );
   }
+
   last=cur;
   lastCPUTime=CPUTime;
   return cur;
 }
 
-static inline unsigned long long checkT(int timeout, char * file, int line) {
+static inline unsigned long long checkT(int timeout, char *file, int line) {
   static __thread unsigned long long last=0;
   unsigned long long cur=rdtsc();
   int microCycles=(int)(cpuf*1000);
   int duration=(int)((cur-last)/microCycles);
+
   if ( last!=0 && duration > timeout )
     printf("%s:%d lte-ue delay %d (exceed %d)\n", file, line,
            duration, timeout);
+
   last=cur;
   return cur;
 }
@@ -463,7 +468,7 @@ typedef struct m {
   uint64_t maxArray[11];
 } Meas;
 
-static inline void printMeas(char * txt, Meas *M, int period) {
+static inline void printMeas(char *txt, Meas *M, int period) {
   if (M->iterations%period == 0 ) {
     char txt2[512];
     sprintf(txt2,"%s avg=%" PRIu64 " iterations=%" PRIu64 " max=%"
@@ -481,13 +486,13 @@ static inline void printMeas(char * txt, Meas *M, int period) {
   }
 }
 
-static inline int cmpint(const void* a, const void* b) {
-    uint64_t* aa=(uint64_t*)a;
-    uint64_t* bb=(uint64_t*)b;
+static inline int cmpint(const void *a, const void *b) {
+  uint64_t *aa=(uint64_t *)a;
+  uint64_t *bb=(uint64_t *)b;
   return (int)(*aa-*bb);
 }
 
-static inline void updateTimes(uint64_t start, Meas *M, int period, char * txt) {
+static inline void updateTimes(uint64_t start, Meas *M, int period, char *txt) {
   if (start!=0) {
     uint64_t end=rdtsc();
     long long diff=(end-start)/(cpuf*1000);

common/utils/threadPool/thread-pool.h

@@ -92,7 +92,7 @@ static inline void pushNotifiedFIFO(notifiedFIFO_t *nf, notifiedFIFO_elt_t *msg)
   if (nf->outF == NULL)
     nf->outF = msg;
 
-  if (nf->inF)
+  if (nf->inF != NULL)
     nf->inF->next = msg;
 
   nf->inF = msg;
@@ -103,7 +103,7 @@ static inline void pushNotifiedFIFO(notifiedFIFO_t *nf, notifiedFIFO_elt_t *msg)
 static inline  notifiedFIFO_elt_t *pullNotifiedFIFO(notifiedFIFO_t *nf) {
   mutexlock(nf->lockF);
 
-  while(!nf->outF)
+  while(nf->outF == NULL)
     condwait(nf->notifF, nf->lockF);
 
   notifiedFIFO_elt_t *ret=nf->outF;
@@ -124,8 +124,12 @@ static inline  notifiedFIFO_elt_t *pollNotifiedFIFO(notifiedFIFO_t *nf) {
 
   notifiedFIFO_elt_t *ret=nf->outF;
 
-  if (ret!=NULL)
+  if (ret!=NULL) {
+    if (nf->outF==nf->outF->next)
+      LOG_E(TMR,"Circular list in thread pool: push several times the same buffer is forbidden\n");
+
     nf->outF=nf->outF->next;
+  }
 
   if (nf->outF==NULL)
     nf->inF=NULL;

common/utils/utils.h

@@ -3,6 +3,9 @@
 
 #include <stdint.h>
 #include <sys/types.h>
+#ifdef __cplusplus
+extern "C" {
+#endif
 
 void *calloc_or_fail(size_t size);
 void *malloc_or_fail(size_t size);
@@ -14,4 +17,17 @@ int hex_string_to_hex_value (uint8_t *hex_value, const char *hex_string, int siz
 
 char *itoa(int i);
 
+#define findInList(keY, result, list, element_type) {\
+    int i;\
+    for (i=0; i<sizeof(list)/sizeof(element_type) ; i++)\
+      if (list[i].key==keY) {\
+        result=list[i].val;\
+        break;\
+      }\
+    AssertFatal(i < sizeof(list)/sizeof(element_type), "List %s doesn't contain %s\n",#list, #keY); \
+  }
+#ifdef __cplusplus
+}
+#endif
+
 #endif

executables/nr-ue.c

@@ -138,13 +138,9 @@ PHY_VARS_NR_UE *init_nr_ue_vars(NR_DL_FRAME_PARMS *frame_parms,
 
 {
   PHY_VARS_NR_UE *ue;
-
-  if (frame_parms!=(NR_DL_FRAME_PARMS *)NULL) { // if we want to give initial frame parms, allocate the PHY_VARS_UE structure and put them in
   ue = (PHY_VARS_NR_UE *)malloc(sizeof(PHY_VARS_NR_UE));
   memset(ue,0,sizeof(PHY_VARS_NR_UE));
   memcpy(&(ue->frame_parms), frame_parms, sizeof(NR_DL_FRAME_PARMS));
-  } else ue = PHY_vars_UE_g[UE_id][0];
-
   ue->Mod_id      = UE_id;
   ue->mac_enabled = 1;
   // initialize all signal buffers
@@ -160,6 +156,7 @@ PHY_VARS_NR_UE *init_nr_ue_vars(NR_DL_FRAME_PARMS *frame_parms,
  */
 
 typedef struct syncData_s {
+  UE_nr_rxtx_proc_t *proc;
   PHY_VARS_NR_UE *UE;
 } syncData_t;
 
@@ -167,18 +164,13 @@ static void UE_synch(void *arg) {
   syncData_t *syncD=(syncData_t *) arg;
   int i, hw_slot_offset;
   PHY_VARS_NR_UE *UE = syncD->UE;
-  int current_band = 0;
-  int current_offset = 0;
   sync_mode_t sync_mode = pbch;
   int CC_id = UE->CC_id;
   int freq_offset=0;
   UE->is_synchronized = 0;
 
   if (UE->UE_scan == 0) {
-    int ind;
-
     get_band(downlink_frequency[CC_id][0], &UE->frame_parms.eutra_band,   &uplink_frequency_offset[CC_id][0], &UE->frame_parms.frame_type);
-    
     LOG_I( PHY, "[SCHED][UE] Check absolute frequency DL %"PRIu32", UL %"PRIu32" (oai_exit %d, rx_num_channels %d)\n",
            downlink_frequency[0][0], downlink_frequency[0][0]+uplink_frequency_offset[0][0],
            oai_exit, openair0_cfg[0].rx_num_channels);
@@ -197,7 +189,6 @@ static void UE_synch(void *arg) {
 
     sync_mode = pbch;
   } else {
-    current_band=0;
     LOG_E(PHY,"Fixme!\n");
     /*
     for (i=0; i<openair0_cfg[UE->rf_map.card].rx_num_channels; i++) {
@@ -245,11 +236,10 @@ static void UE_synch(void *arg) {
 
       break;
     */
-    
     case pbch:
       LOG_I(PHY, "[UE thread Synch] Running Initial Synch (mode %d)\n",UE->mode);
 
-      if (nr_initial_sync( UE, UE->mode ) == 0) {
+      if (nr_initial_sync( syncD->proc, UE, UE->mode ) == 0) {
         freq_offset = UE->common_vars.freq_offset; // frequency offset computed with pss in initial sync
         hw_slot_offset = (UE->rx_offset<<1) / UE->frame_parms.samples_per_slot;
         LOG_I(PHY,"Got synch: hw_slot_offset %d, carrier off %d Hz, rxgain %d (DL %u, UL %u), UE_scan_carrier %d\n",
@@ -377,7 +367,6 @@ void processSlotRX( PHY_VARS_NR_UE *UE, UE_nr_rxtx_proc_t *proc) {
     UE_mac->scheduled_response.slot = proc->nr_tti_rx;
     nr_ue_scheduled_response(&UE_mac->scheduled_response);
     //write_output("uerxdata_frame.m", "uerxdata_frame", UE->common_vars.rxdata[0], UE->frame_parms.samples_per_frame, 1, 1);
-    printf("Processing slot %d\n",proc->nr_tti_rx);
 #ifdef UE_SLOT_PARALLELISATION
     phy_procedures_slot_parallelization_nrUE_RX( UE, proc, 0, 0, 1, UE->mode, no_relay, NULL );
 #else
@@ -548,10 +537,6 @@ void *UE_thread(void *arg) {
   AssertFatal(UE->rfdevice.trx_start_func(&UE->rfdevice) == 0, "Could not start the device\n");
   notifiedFIFO_t nf;
   initNotifiedFIFO(&nf);
-  bool syncRunning=false;
-  notifiedFIFO_elt_t *syncMsg=newNotifiedFIFO_elt(sizeof(syncData_t),0,&nf,UE_synch);
-  syncData_t *syncD=(syncData_t *)NotifiedFifoData(syncMsg);
-  syncD->UE=UE;
   int nbSlotProcessing=0;
   int thread_idx=0;
   notifiedFIFO_elt_t *processingMsg[RX_NB_TH];
@@ -562,8 +547,13 @@ void *UE_thread(void *arg) {
     tmp->UE=UE;
   }
 
+  bool syncRunning=false;
+  notifiedFIFO_elt_t *syncMsg=newNotifiedFIFO_elt(sizeof(syncData_t),0,&nf,UE_synch);
+  syncData_t *syncD=(syncData_t *)NotifiedFifoData(syncMsg);
+  syncD->UE=UE;
+  syncD->proc=&((processingData_t *)NotifiedFifoData(processingMsg[0]))->proc;
   const int nb_slot_frame = 10*UE->frame_parms.slots_per_subframe;
-  int absolute_slot=-1;
+  int absolute_slot, decoded_frame_rx=INT_MAX, trashed_frames=0;
 
   while (!oai_exit) {
     if (!syncD->UE->is_synchronized) {
@@ -572,11 +562,14 @@ void *UE_thread(void *arg) {
 
         if (res) {
           syncRunning=false;
-        } else
+        } else {
           trashFrame(UE, &timestamp);
+          trashed_frames++;
+        }
       } else {
         readFrames(UE, &timestamp);
         pushTpool(Tpool, syncMsg);
+        trashed_frames=0;
         syncRunning=true;
       }
 
@@ -588,7 +581,6 @@ void *UE_thread(void *arg) {
       syncInFrame(UE, &timestamp);
       UE->rx_offset=0;
       UE->time_sync_cell=0;
-      //printf("first stream frame rx %d\n",UE->proc.proc_rxtx[0].frame_rx);
       // read in first symbol
       AssertFatal (UE->frame_parms.ofdm_symbol_size+UE->frame_parms.nb_prefix_samples0 ==
                    UE->rfdevice.trx_read_func(&UE->rfdevice,
@@ -596,12 +588,21 @@ void *UE_thread(void *arg) {
                                               (void **)UE->common_vars.rxdata,
                                               UE->frame_parms.ofdm_symbol_size+UE->frame_parms.nb_prefix_samples0,
                                               UE->frame_parms.nb_antennas_rx),"");
+      // we have the decoded frame index in the return of the synch process
+      // and we shifted above to the first slot of next frame
+      // the synch thread proc context is hard linked to regular processing thread context, thread id  = 0
+      UE_nr_rxtx_proc_t *proc=&(((processingData_t *)NotifiedFifoData(processingMsg[0]))->proc);
+      // shift the frame index with all the frames we trashed meanwhile we perform the synch search
+      proc->decoded_frame_rx=(proc->decoded_frame_rx + trashed_frames) % MAX_FRAME_NUMBER;
+      decoded_frame_rx=proc->decoded_frame_rx;
+      // we do ++ first in the regular processing, so it will be 0;
+      absolute_slot=decoded_frame_rx*nb_slot_frame + nb_slot_frame -1;
       continue;
     }
 
     absolute_slot++;
     thread_idx = absolute_slot % RX_NB_TH;
-    int slot_nr=absolute_slot % nb_slot_frame;
+    int slot_nr = absolute_slot % nb_slot_frame;
     UE_nr_rxtx_proc_t *proc=&(((processingData_t *)NotifiedFifoData(processingMsg[thread_idx]))->proc);
     // update thread index for received subframe
     proc->nr_tti_rx= slot_nr;
@@ -611,6 +612,7 @@ void *UE_thread(void *arg) {
     proc->subframe_tx=proc->nr_tti_rx;
     proc->frame_rx = ( absolute_slot/nb_slot_frame ) % MAX_FRAME_NUMBER;
     proc->frame_tx = ( (absolute_slot + DURATION_RX_TO_TX) /nb_slot_frame ) % MAX_FRAME_NUMBER;
+    proc->decoded_frame_rx=-1;
     LOG_D(PHY,"Process slot %d thread Idx %d \n", slot_nr, thread_idx);
 
     for (int i=0; i<UE->frame_parms.nb_antennas_rx; i++)
@@ -671,21 +673,37 @@ void *UE_thread(void *arg) {
     proc->timestamp_tx = timestamp+
                          (DURATION_RX_TO_TX*UE->frame_parms.samples_per_slot)-
                          UE->frame_parms.ofdm_symbol_size-UE->frame_parms.nb_prefix_samples0;
+    notifiedFIFO_elt_t *res;
 
     if (getenv("RFSIMULATOR")) {
       // FixMe: Wait previous thread is done, because race conditions seems too bad
       // in case of actual RF board, the overlap between threads mitigate the issue
-      while (!tryPullTpool(&nf, Tpool)) {
-        nbSlotProcessing--;
+      // We must receive one message, that proves the slot processing is done
+      while ((res=tryPullTpool(&nf, Tpool)) == NULL)
         usleep(200);
-      }
+
+      nbSlotProcessing--;
+      processingData_t *tmp=(processingData_t *)res->msgData;
+
+      if (tmp->proc.decoded_frame_rx != -1)
+        decoded_frame_rx=tmp->proc.decoded_frame_rx;
     }
 
-    while (nbSlotProcessing >= RX_NB_TH && !tryPullTpool(&nf, Tpool)) {
+    while (nbSlotProcessing >= RX_NB_TH && (res=tryPullTpool(&nf, Tpool)) != NULL ) {
       nbSlotProcessing--;
+      processingData_t *tmp=(processingData_t *)res->msgData;
+
+      if (tmp->proc.decoded_frame_rx != -1)
+        decoded_frame_rx=tmp->proc.decoded_frame_rx;
+
       usleep(200);
     }
 
+    if (  decoded_frame_rx != proc->frame_rx &&
+          ((decoded_frame_rx+1) % MAX_FRAME_NUMBER) != proc->frame_rx )
+      LOG_E(PHY,"Decoded frame index (%d) is not compatible with current context (%d), UE should go back to synch mode\n",
+            decoded_frame_rx,  proc->frame_rx);
+
     pushTpool(Tpool, processingMsg[thread_idx]);
   } // while !oai_exit
 
@@ -709,7 +727,6 @@ void init_UE(int nb_inst) {
     //    UE->rfdevice.type      = NONE_DEV;
     //PHY_VARS_NR_UE *UE = PHY_vars_UE_g[inst][0];
     LOG_I(PHY,"Initializing memory for UE instance %d (%p)\n",inst,PHY_vars_UE_g[inst]);
-    PHY_vars_UE_g[inst][0] = init_nr_ue_vars(NULL,inst,0);
     PHY_VARS_NR_UE *UE = PHY_vars_UE_g[inst][0];
     AssertFatal((UE->if_inst = nr_ue_if_module_init(inst)) != NULL, "can not initial IF module\n");
     nr_l3_init_ue();

executables/nr-uesoftmodem.h

@@ -28,6 +28,7 @@
 #include "PHY/defs_nr_UE.h"
 #include "SIMULATION/ETH_TRANSPORT/proto.h"
 #include <threadPool/thread-pool.h>
+#include <openair2/LAYER2/NR_MAC_gNB/mac_proto.h>
 
 
 #if defined(ENABLE_ITTI)

openair1/PHY/NR_TRANSPORT/nr_dci_tools.c

@@ -46,7 +46,7 @@ void nr_fill_cce_list(NR_gNB_DCI_ALLOC_t* dci_alloc, uint16_t n_shift, uint8_t m
   uint8_t bsize = pdcch_params->reg_bundle_size;
   uint8_t R = pdcch_params->interleaver_size;
   uint16_t N_reg = pdcch_params->n_rb * pdcch_params->n_symb;
-  uint16_t Y, N_cce, M_s_max, n_CI=0, tmp, C;
+  uint16_t Y, N_cce, M_s_max, n_CI=0, tmp, C=0;
   uint16_t n_RNTI = (pdcch_params->search_space_type == NFAPI_NR_SEARCH_SPACE_TYPE_UE_SPECIFIC)? pdcch_params->rnti:0;
   uint32_t A[3]={39827,39829,39839};
 
@@ -64,7 +64,7 @@ void nr_fill_cce_list(NR_gNB_DCI_ALLOC_t* dci_alloc, uint16_t n_shift, uint8_t m
   }
 
   if (pdcch_params->cr_mapping_type == NFAPI_NR_CCE_REG_MAPPING_INTERLEAVED) {
-    AssertFatal((N_reg%(bsize*R))==0, "CCE to REG interleaving: Invalid configuration leading to non integer C (N_reg %d, bsize %d R %d)\n",
+    AssertFatal((N_reg%(bsize*R))==0, "CCE to REG interleaving: Invalid configuration leading to non integer C (N_reg %us, bsize %d R %d)\n",
     N_reg, bsize, R);
     C = N_reg/(bsize*R);
   }

openair1/PHY/NR_TRANSPORT/nr_dlsch.c

@@ -215,7 +215,7 @@ uint8_t nr_generate_pdsch(NR_gNB_DLSCH_t dlsch,
   nr_dlsch_encoding(harq->pdu, slot, &dlsch, &frame_parms);
 #ifdef DEBUG_DLSCH
 printf("PDSCH encoding:\nPayload:\n");
-for (int i=0; i<TBS>>7; i++) {
+for (int i=0; i<harq->B>>7; i++) {
   for (int j=0; j<16; j++)
     printf("0x%02x\t", harq->pdu[(i<<4)+j]);
   printf("\n");
@@ -308,16 +308,13 @@ for (int i=0; i<n_dmrs>>4; i++) {
 
 
     // Non interleaved VRB to PRB mapping
-  uint16_t start_sc = frame_parms.first_carrier_offset + frame_parms.ssb_start_subcarrier;
+ uint16_t start_sc = frame_parms.first_carrier_offset + rel15->start_prb*NR_NB_SC_PER_RB;
  if (start_sc >= frame_parms.ofdm_symbol_size)
    start_sc -= frame_parms.ofdm_symbol_size;
- /*rel15->start_prb*NR_NB_SC_PER_RB +
-  ((pdcch_params.search_space_type == NFAPI_NR_SEARCH_SPACE_TYPE_COMMON) && (pdcch_params.dci_format == NFAPI_NR_DL_DCI_FORMAT_1_0))?\
-       ((frame_parms.ssb_start_subcarrier/NR_NB_SC_PER_RB + pdcch_params.rb_offset)*NR_NB_SC_PER_RB) : 0;*/
 
 #ifdef DEBUG_DLSCH_MAPPING
-printf("PDSCH resource mapping started (start SC %d\tstart symbol %d\tN_PRB %d\tnb_symbols %d)\n",
-start_sc, rel15->start_symbol, rel15->n_prb, rel15->nb_symbols);
+ printf("PDSCH resource mapping started (start SC %d\tstart symbol %d\tN_PRB %d\tnb_symbols %d)\n",
+	start_sc, rel15->start_symbol, rel15->n_prb, rel15->nb_symbols);
 #endif
 
   for (int ap=0; ap<rel15->nb_layers; ap++) {

openair1/PHY/NR_TRANSPORT/nr_tbs_tools.c

@@ -126,7 +126,7 @@ void nr_get_tbs(nfapi_nr_dl_config_dlsch_pdu *dlsch_pdu,
   uint16_t N_RE, Ninfo, Ninfo_prime, C, TBS=0, R;
   uint8_t table_idx, Qm, n, scale;
 
-  table_idx = get_table_idx(mcs_table, dci_format, rnti_type, ss_type);
+  table_idx = 1; //get_table_idx(mcs_table, dci_format, rnti_type, ss_type);
   scale = ((table_idx==2)&&((Imcs==20)||(Imcs==26)))?11:10;
   
   N_RE = min(156, N_RE_prime)*dlsch_rel15->n_prb;

openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c

@@ -1990,7 +1990,7 @@ unsigned short nr_dlsch_extract_rbs_single(int **rxdataF,
 					   int **dl_ch_estimates_ext,
 					   unsigned short pmi,
 					   unsigned char *pmi_ext,
-                                        unsigned int *rb_alloc,
+					   unsigned int *rb_alloc, //unused in NR
 					   unsigned char symbol,
 					   unsigned short start_rb,
 					   unsigned short nb_rb_pdsch,
@@ -2008,20 +2008,8 @@ unsigned short nr_dlsch_extract_rbs_single(int **rxdataF,
 
   unsigned char pilots=0,j=0;
 
-  //symbol_mod = (symbol>=(7-frame_parms->Ncp)) ? symbol-(7-frame_parms->Ncp) : symbol;
-  pilots = (symbol==2) ? 1 : 0; //to updated from config
-  //l=symbol;
-  //nsymb = (frame_parms->Ncp==NORMAL) ? 14:12;
-  k = frame_parms->first_carrier_offset + 516; //0
-
- /* if (frame_parms->frame_type == TDD) {  // TDD
-    sss_symb = nsymb-1;
-    pss_symb = 2;
-  } else {
-    sss_symb = (nsymb>>1)-2;
-    pss_symb = (nsymb>>1)-1;
-  }*/
-
+  pilots = (symbol==2) ? 1 : 0; //to updated from config!!!
+  k = frame_parms->first_carrier_offset + 12*start_rb; 
 
   for (aarx=0; aarx<frame_parms->nb_antennas_rx; aarx++) {
 

openair1/PHY/NR_UE_TRANSPORT/nr_initial_sync.c

@@ -102,7 +102,7 @@ void free_list(NR_UE_SSB *node) {
 }
 
 
-int nr_pbch_detection(PHY_VARS_NR_UE *ue, int pbch_initial_symbol, runmode_t mode)
+int nr_pbch_detection(UE_nr_rxtx_proc_t * proc, PHY_VARS_NR_UE *ue, int pbch_initial_symbol, runmode_t mode)
 {
   NR_DL_FRAME_PARMS *frame_parms=&ue->frame_parms;
   int ret =-1;
@@ -160,7 +160,7 @@ int nr_pbch_detection(PHY_VARS_NR_UE *ue, int pbch_initial_symbol, runmode_t mod
 #endif
 
     ret = nr_rx_pbch(ue,
-	             0,
+	             proc,
 		     ue->pbch_vars[0],
 		     frame_parms,
 		     0,
@@ -201,7 +201,7 @@ int nr_pbch_detection(PHY_VARS_NR_UE *ue, int pbch_initial_symbol, runmode_t mod
 char duplex_string[2][4] = {"FDD","TDD"};
 char prefix_string[2][9] = {"NORMAL","EXTENDED"};
 
-int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode)
+int nr_initial_sync(UE_nr_rxtx_proc_t *proc, PHY_VARS_NR_UE *ue, runmode_t mode)
 {
 
   int32_t sync_pos, sync_pos_frame; // k_ssb, N_ssb_crb, sync_pos2,
@@ -276,7 +276,6 @@ int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode)
 	}
     }
 
-
     /* check that SSS/PBCH block is continuous inside the received buffer */
     if (sync_pos < (NR_NUMBER_OF_SUBFRAMES_PER_FRAME*fp->samples_per_subframe - (NB_SYMBOLS_PBCH * fp->ofdm_symbol_size))) {
 
@@ -307,7 +306,7 @@ int nr_initial_sync(PHY_VARS_NR_UE *ue, runmode_t mode)
       //nr_init_frame_parms_ue(fp,NR_MU_1,NORMAL,n_ssb_crb,0);
 
       nr_gold_pbch(ue);
-      ret = nr_pbch_detection(ue,1,mode);  // start pbch detection at first symbol after pss
+      ret = nr_pbch_detection(proc, ue,1,mode);  // start pbch detection at first symbol after pss
 
       if (ret == 0) {
         // sync at symbol ue->symbol_offset

openair1/PHY/NR_UE_TRANSPORT/nr_pbch.c

@@ -410,13 +410,15 @@ unsigned char sign(int8_t x) {
 uint8_t pbch_deinterleaving_pattern[32] = {28,0,31,30,7,29,25,27,5,8,24,9,10,11,12,13,1,4,3,14,15,16,17,2,26,18,19,20,21,22,6,23};
 
 int nr_rx_pbch( PHY_VARS_NR_UE *ue,
-                int Ns,
+                UE_nr_rxtx_proc_t *proc,
                 NR_UE_PBCH *nr_ue_pbch_vars,
                 NR_DL_FRAME_PARMS *frame_parms,
                 uint8_t eNB_id,
                 uint8_t i_ssb,
                 MIMO_mode_t mimo_mode,
                 uint32_t high_speed_flag) {
+
+  int Ns = proc->nr_tti_rx;
   NR_UE_COMMON *nr_ue_common_vars = &ue->common_vars;
   int max_h=0;
   int symbol;
@@ -592,6 +594,7 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue,
 #endif
 
   ue->dl_indication.rx_ind = &ue->rx_ind; //  hang on rx_ind instance
+  ue->dl_indication.proc=proc;
   //ue->rx_ind.sfn_slot = 0;  //should be set by higher-1-layer, i.e. clean_and_set_if_instance()
   ue->rx_ind.rx_indication_body[0].pdu_type = FAPI_NR_RX_PDU_TYPE_MIB;
   ue->rx_ind.rx_indication_body[0].mib_pdu.pdu = &decoded_output[0];

openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h

@@ -1137,7 +1137,7 @@ int rx_sss(PHY_VARS_NR_UE *phy_vars_ue,int32_t *tot_metric,uint8_t *flip_max,uin
   \returns number of tx antennas or -1 if error
 */
 int nr_rx_pbch( PHY_VARS_NR_UE *ue,
-		     int subframe_rx,
+		     UE_nr_rxtx_proc_t *proc,
 		     NR_UE_PBCH *nr_ue_pbch_vars,
 		     NR_DL_FRAME_PARMS *frame_parms,
 		     uint8_t eNB_id,
@@ -1145,7 +1145,8 @@ int nr_rx_pbch( PHY_VARS_NR_UE *ue,
 		     MIMO_mode_t mimo_mode,
 		     uint32_t high_speed_flag);
 
-int nr_pbch_detection(PHY_VARS_NR_UE *ue,
+int nr_pbch_detection(UE_nr_rxtx_proc_t *proc,
+		      PHY_VARS_NR_UE *ue,
 		      int pbch_initial_symbol,
 		      runmode_t mode);
 
@@ -1433,7 +1434,8 @@ void generate_RIV_tables(void);
   @param phy_vars_ue Pointer to UE variables
   @param mode current running mode
 */
-int nr_initial_sync(PHY_VARS_NR_UE *phy_vars_ue, runmode_t mode);
+int nr_initial_sync(UE_nr_rxtx_proc_t *proc,
+                    PHY_VARS_NR_UE *phy_vars_ue, runmode_t mode);
 
 
 /*!

openair1/PHY/NR_UE_TRANSPORT/nr_transport_ue.h

@@ -277,7 +277,7 @@ typedef struct {
   /// soft bits for each received segment ("w"-sequence)(for definition see 36-212 V8.6 2009-03, p.15)
   int16_t w[MAX_NUM_NR_DLSCH_SEGMENTS][3*8448];
   /// for abstraction soft bits for each received segment ("w"-sequence)(for definition see 36-212 V8.6 2009-03, p.15)
-  double w_abs[MAX_NUM_NR_DLSCH_SEGMENTS][3*8448];
+  //double w_abs[MAX_NUM_NR_DLSCH_SEGMENTS][3*8448];
   /// soft bits for each received segment ("d"-sequence)(for definition see 36-212 V8.6 2009-03, p.15)
   int16_t *d[MAX_NUM_NR_DLSCH_SEGMENTS];
   /// LDPC processing buffers

openair1/PHY/NR_UE_TRANSPORT/pss_nr.c

@@ -402,7 +402,7 @@ void init_context_pss_nr(NR_DL_FRAME_PARMS *frame_parms_ue)
      assert(0);
     }
 
-    size = NR_NUMBER_OF_SUBFRAMES_PER_FRAME*sizeof(int64_t)*frame_parms_ue->samples_per_subframe;
+    size = sizeof(int64_t)*(frame_parms_ue->samples_per_frame + (2*ofdm_symbol_size));
     q = (int64_t*)malloc16(size);
     if (q != NULL) {
       pss_corr_ue[i] = q;

openair1/PHY/TOOLS/cdot_prod.c

@@ -158,122 +158,6 @@ int32_t dot_product(int16_t *x,
 #endif
 }
 
-int64_t dot_product64(int16_t *x,
-		      int16_t *y,
-		      uint32_t N, //must be a multiple of 8
-		      uint8_t output_shift)
-{
-
-  uint32_t n;
-
-#if defined(__x86_64__) || defined(__i386__)
-  __m128i *x128,*y128,mmtmp1,mmtmp2,mmtmp3,mmcumul,mmcumul_re,mmcumul_im;
-  __m128i minus_i = _mm_set_epi16(-1,1,-1,1,-1,1,-1,1);
-  int64_t result;
-
-  x128 = (__m128i*) x;
-  y128 = (__m128i*) y;
-
-  mmcumul_re = _mm_setzero_si128();
-  mmcumul_im = _mm_setzero_si128();
-
-  for (n=0; n<(N>>2); n++) {
-
-//    printf("n=%d, x128=%p, y128=%p\n",n,x128,y128);
-       // print_shorts("x",&x128[0]);
-       // print_shorts("y",&y128[0]);
-
-    // this computes Re(z) = Re(x)*Re(y) + Im(x)*Im(y)
-    mmtmp1 = _mm_madd_epi16(x128[0],y128[0]);
-      //  print_ints("retmp",&mmtmp1);
-    // mmtmp1 contains real part of 4 consecutive outputs (32-bit)
-    // shift and accumulate results
-    mmtmp1 = _mm_srai_epi32(mmtmp1,output_shift);
-    mmcumul_re = _mm_add_epi32(mmcumul_re,mmtmp1);
-        //print_ints("re",&mmcumul_re);
-
-
-    // this computes Im(z) = Re(x)*Im(y) - Re(y)*Im(x)
-    mmtmp2 = _mm_shufflelo_epi16(y128[0],_MM_SHUFFLE(2,3,0,1));
-        //print_shorts("y",&mmtmp2);
-    mmtmp2 = _mm_shufflehi_epi16(mmtmp2,_MM_SHUFFLE(2,3,0,1));
-        //print_shorts("y",&mmtmp2);
-    mmtmp2 = _mm_sign_epi16(mmtmp2,minus_i);
-          //  print_shorts("y",&mmtmp2);
-
-    mmtmp3 = _mm_madd_epi16(x128[0],mmtmp2);
-            //print_ints("imtmp",&mmtmp3);
-    // mmtmp3 contains imag part of 4 consecutive outputs (32-bit)
-    // shift and accumulate results
-    mmtmp3 = _mm_srai_epi32(mmtmp3,output_shift);
-    mmcumul_im = _mm_add_epi32(mmcumul_im,mmtmp3);
-        //print_ints("im",&mmcumul_im);
-
-    x128++;
-    y128++;
-  }
-
-  // this gives Re Re Im Im
-  mmcumul = _mm_hadd_epi32(mmcumul_re,mmcumul_im);
-    //print_ints("cumul1",&mmcumul);
-  // this gives Re Im Re Im
-  mmcumul = _mm_hadd_epi32(mmcumul,mmcumul);
-    //print_ints("cumul2",&mmcumul);
-
-  //mmcumul = _mm_srai_epi32(mmcumul,output_shift);
-  // extract the lower half
-  result = _mm_extract_epi64(mmcumul,0);
-  //printf("result: (%d,%d)\n",((int32_t*)&result)[0],((int32_t*)&result)[1]); 
-  _mm_empty();
-  _m_empty();
- 
-  return(result);
-
-#elif defined(__arm__)
-  int16x4_t *x_128=(int16x4_t*)x;
-  int16x4_t *y_128=(int16x4_t*)y;
-  int32x4_t tmp_re,tmp_im;
-  int32x4_t tmp_re1,tmp_im1;
-  int32x4_t re_cumul,im_cumul;
-  int32x2_t re_cumul2,im_cumul2;
-  int32x4_t shift = vdupq_n_s32(-output_shift); 
-  int32x2x2_t result2;
-  int16_t conjug[4]__attribute__((aligned(16))) = {-1,1,-1,1} ;
-
-  re_cumul = vdupq_n_s32(0);
-  im_cumul = vdupq_n_s32(0); 
-
-  for (n=0; n<(N>>2); n++) {
-
-    tmp_re  = vmull_s16(*x_128++, *y_128++);
-    //tmp_re = [Re(x[0])Re(y[0]) Im(x[0])Im(y[0]) Re(x[1])Re(y[1]) Im(x[1])Im(y[1])] 
-    tmp_re1 = vmull_s16(*x_128++, *y_128++);
-    //tmp_re1 = [Re(x1[1])Re(x2[1]) Im(x1[1])Im(x2[1]) Re(x1[1])Re(x2[2]) Im(x1[1])Im(x2[2])] 
-    tmp_re  = vcombine_s32(vpadd_s32(vget_low_s32(tmp_re),vget_high_s32(tmp_re)),
-                           vpadd_s32(vget_low_s32(tmp_re1),vget_high_s32(tmp_re1)));
-    //tmp_re = [Re(ch[0])Re(rx[0])+Im(ch[0])Im(ch[0]) Re(ch[1])Re(rx[1])+Im(ch[1])Im(ch[1]) Re(ch[2])Re(rx[2])+Im(ch[2]) Im(ch[2]) Re(ch[3])Re(rx[3])+Im(ch[3])Im(ch[3])] 
-
-    tmp_im  = vmull_s16(vrev32_s16(vmul_s16(*x_128++,*(int16x4_t*)conjug)),*y_128++);
-    //tmp_im = [-Im(ch[0])Re(rx[0]) Re(ch[0])Im(rx[0]) -Im(ch[1])Re(rx[1]) Re(ch[1])Im(rx[1])]
-    tmp_im1 = vmull_s16(vrev32_s16(vmul_s16(*x_128++,*(int16x4_t*)conjug)),*y_128++);
-    //tmp_im1 = [-Im(ch[2])Re(rx[2]) Re(ch[2])Im(rx[2]) -Im(ch[3])Re(rx[3]) Re(ch[3])Im(rx[3])]
-    tmp_im  = vcombine_s32(vpadd_s32(vget_low_s32(tmp_im),vget_high_s32(tmp_im)),
-                           vpadd_s32(vget_low_s32(tmp_im1),vget_high_s32(tmp_im1)));
-    //tmp_im = [-Im(ch[0])Re(rx[0])+Re(ch[0])Im(rx[0]) -Im(ch[1])Re(rx[1])+Re(ch[1])Im(rx[1]) -Im(ch[2])Re(rx[2])+Re(ch[2])Im(rx[2]) -Im(ch[3])Re(rx[3])+Re(ch[3])Im(rx[3])]
-
-    re_cumul = vqaddq_s32(re_cumul,vqshlq_s32(tmp_re,shift));
-    im_cumul = vqaddq_s32(im_cumul,vqshlq_s32(tmp_im,shift));
-  }
-  
-  re_cumul2 = vpadd_s32(vget_low_s32(re_cumul),vget_high_s32(re_cumul));
-  im_cumul2 = vpadd_s32(vget_low_s32(im_cumul),vget_high_s32(im_cumul));
-  re_cumul2 = vpadd_s32(re_cumul2,re_cumul2);
-  im_cumul2 = vpadd_s32(im_cumul2,im_cumul2);
-  result2   = vzip_s32(re_cumul2,im_cumul2);
-  return(vget_lane_s32(result2.val[0],0));
-#endif
-}
-
 
 
 #ifdef MAIN

openair1/PHY/TOOLS/tools_defs.h

@@ -354,10 +354,116 @@ int32_t dot_product(int16_t *x,
                     uint32_t N, //must be a multiple of 8
                     uint8_t output_shift);
 
-int64_t dot_product64(int16_t *x,
+static inline int64_t dot_product64(int16_t *x,
 		      int16_t *y,
 		      uint32_t N, //must be a multiple of 8
-		      uint8_t output_shift);
+		      uint8_t output_shift)
+{
+
+#if defined(__x86_64__) || defined(__i386__)
+  __m128i *x128,*y128,mmtmp1,mmtmp2,mmtmp3,mmcumul,mmcumul_re,mmcumul_im;
+  __m128i minus_i = _mm_set_epi16(-1,1,-1,1,-1,1,-1,1);
+  int64_t result;
+
+  x128 = (__m128i*) x;
+  y128 = (__m128i*) y;
+
+  mmcumul_re = _mm_setzero_si128();
+  mmcumul_im = _mm_setzero_si128();
+
+  __m128i*end=x128+(N>>2);
+  for (__m128i* inPtr=x128; inPtr < end ; inPtr++) {
+
+//    printf("n=%d, x128=%p, y128=%p\n",n,x128,y128);
+       // print_shorts("x",&x128[0]);
+       // print_shorts("y",&y128[0]);
+
+    // this computes Re(z) = Re(x)*Re(y) + Im(x)*Im(y)
+    mmtmp1 = _mm_madd_epi16(*inPtr,*y128);
+      //  print_ints("retmp",&mmtmp1);
+    // mmtmp1 contains real part of 4 consecutive outputs (32-bit)
+    // shift and accumulate results
+    mmtmp1 = _mm_srai_epi32(mmtmp1,output_shift);
+    mmcumul_re = _mm_add_epi32(mmcumul_re,mmtmp1);
+        //print_ints("re",&mmcumul_re);
+
+
+    // this computes Im(z) = Re(x)*Im(y) - Re(y)*Im(x)
+    mmtmp2 = _mm_shufflelo_epi16(*y128,_MM_SHUFFLE(2,3,0,1));
+        //print_shorts("y",&mmtmp2);
+    mmtmp2 = _mm_shufflehi_epi16(mmtmp2,_MM_SHUFFLE(2,3,0,1));
+        //print_shorts("y",&mmtmp2);
+    mmtmp2 = _mm_sign_epi16(mmtmp2,minus_i);
+          //  print_shorts("y",&mmtmp2);
+
+    mmtmp3 = _mm_madd_epi16(*inPtr,mmtmp2);
+            //print_ints("imtmp",&mmtmp3);
+    // mmtmp3 contains imag part of 4 consecutive outputs (32-bit)
+    // shift and accumulate results
+    mmtmp3 = _mm_srai_epi32(mmtmp3,output_shift);
+    mmcumul_im = _mm_add_epi32(mmcumul_im,mmtmp3);
+        //print_ints("im",&mmcumul_im);
+    y128++;
+  }
+
+  // this gives Re Re Im Im
+  mmcumul = _mm_hadd_epi32(mmcumul_re,mmcumul_im);
+    //print_ints("cumul1",&mmcumul);
+  // this gives Re Im Re Im
+  mmcumul = _mm_hadd_epi32(mmcumul,mmcumul);
+    //print_ints("cumul2",&mmcumul);
+
+  //mmcumul = _mm_srai_epi32(mmcumul,output_shift);
+  // extract the lower half
+  result = _mm_extract_epi64(mmcumul,0);
+  //printf("result: (%d,%d)\n",((int32_t*)&result)[0],((int32_t*)&result)[1]); 
+ 
+  return(result);
+
+#elif defined(__arm__)
+  int16x4_t *x_128=(int16x4_t*)x;
+  int16x4_t *y_128=(int16x4_t*)y;
+  int32x4_t tmp_re,tmp_im;
+  int32x4_t tmp_re1,tmp_im1;
+  int32x4_t re_cumul,im_cumul;
+  int32x2_t re_cumul2,im_cumul2;
+  int32x4_t shift = vdupq_n_s32(-output_shift); 
+  int32x2x2_t result2;
+  int16_t conjug[4]__attribute__((aligned(16))) = {-1,1,-1,1} ;
+
+  re_cumul = vdupq_n_s32(0);
+  im_cumul = vdupq_n_s32(0); 
+
+  for (n=0; n<(N>>2); n++) {
+
+    tmp_re  = vmull_s16(*x_128++, *y_128++);
+    //tmp_re = [Re(x[0])Re(y[0]) Im(x[0])Im(y[0]) Re(x[1])Re(y[1]) Im(x[1])Im(y[1])] 
+    tmp_re1 = vmull_s16(*x_128++, *y_128++);
+    //tmp_re1 = [Re(x1[1])Re(x2[1]) Im(x1[1])Im(x2[1]) Re(x1[1])Re(x2[2]) Im(x1[1])Im(x2[2])] 
+    tmp_re  = vcombine_s32(vpadd_s32(vget_low_s32(tmp_re),vget_high_s32(tmp_re)),
+                           vpadd_s32(vget_low_s32(tmp_re1),vget_high_s32(tmp_re1)));
+    //tmp_re = [Re(ch[0])Re(rx[0])+Im(ch[0])Im(ch[0]) Re(ch[1])Re(rx[1])+Im(ch[1])Im(ch[1]) Re(ch[2])Re(rx[2])+Im(ch[2]) Im(ch[2]) Re(ch[3])Re(rx[3])+Im(ch[3])Im(ch[3])] 
+
+    tmp_im  = vmull_s16(vrev32_s16(vmul_s16(*x_128++,*(int16x4_t*)conjug)),*y_128++);
+    //tmp_im = [-Im(ch[0])Re(rx[0]) Re(ch[0])Im(rx[0]) -Im(ch[1])Re(rx[1]) Re(ch[1])Im(rx[1])]
+    tmp_im1 = vmull_s16(vrev32_s16(vmul_s16(*x_128++,*(int16x4_t*)conjug)),*y_128++);
+    //tmp_im1 = [-Im(ch[2])Re(rx[2]) Re(ch[2])Im(rx[2]) -Im(ch[3])Re(rx[3]) Re(ch[3])Im(rx[3])]
+    tmp_im  = vcombine_s32(vpadd_s32(vget_low_s32(tmp_im),vget_high_s32(tmp_im)),
+                           vpadd_s32(vget_low_s32(tmp_im1),vget_high_s32(tmp_im1)));
+    //tmp_im = [-Im(ch[0])Re(rx[0])+Re(ch[0])Im(rx[0]) -Im(ch[1])Re(rx[1])+Re(ch[1])Im(rx[1]) -Im(ch[2])Re(rx[2])+Re(ch[2])Im(rx[2]) -Im(ch[3])Re(rx[3])+Re(ch[3])Im(rx[3])]
+
+    re_cumul = vqaddq_s32(re_cumul,vqshlq_s32(tmp_re,shift));
+    im_cumul = vqaddq_s32(im_cumul,vqshlq_s32(tmp_im,shift));
+  }
+  
+  re_cumul2 = vpadd_s32(vget_low_s32(re_cumul),vget_high_s32(re_cumul));
+  im_cumul2 = vpadd_s32(vget_low_s32(im_cumul),vget_high_s32(im_cumul));
+  re_cumul2 = vpadd_s32(re_cumul2,re_cumul2);
+  im_cumul2 = vpadd_s32(im_cumul2,im_cumul2);
+  result2   = vzip_s32(re_cumul2,im_cumul2);
+  return(vget_lane_s32(result2.val[0],0));
+#endif
+}
 
 void dft12(int16_t *x,int16_t *y);
 void dft24(int16_t *x,int16_t *y,uint8_t scale_flag);

openair1/PHY/defs_nr_UE.h

@@ -157,109 +157,6 @@
 
 #include "targets/ARCH/COMMON/common_lib.h"
 
-#include "NR_IF_Module.h"
-
-/// Context data structure for RX/TX portion of subframe processing
-typedef struct {
-  /// index of the current UE RX/TX proc
-  int                  proc_id;
-  /// Component Carrier index
-  uint8_t              CC_id;
-  /// timestamp transmitted to HW
-  openair0_timestamp timestamp_tx;
-  //#ifdef UE_NR_PHY_DEMO
-  /// NR TTI index within subframe_tx [0 .. ttis_per_subframe - 1] to act upon for transmission
-  int nr_tti_tx;
-  /// NR TTI index within subframe_rx [0 .. ttis_per_subframe - 1] to act upon for reception
-  int nr_tti_rx;
-  //#endif
-  /// subframe to act upon for transmission
-  int subframe_tx;
-  /// subframe to act upon for reception
-  int subframe_rx;
-  /// frame to act upon for transmission
-  int frame_tx;
-  /// frame to act upon for reception
-  int frame_rx;
-  /// \brief Instance count for RXn-TXnp4 processing thread.
-  /// \internal This variable is protected by \ref mutex_rxtx.
-  int instance_cnt_rxtx;
-  /// pthread structure for RXn-TXnp4 processing thread
-  pthread_t pthread_rxtx;
-  /// pthread attributes for RXn-TXnp4 processing thread
-  pthread_attr_t attr_rxtx;
-  /// condition variable for tx processing thread
-  pthread_cond_t cond_rxtx;
-  /// mutex for RXn-TXnp4 processing thread
-  pthread_mutex_t mutex_rxtx;
-  /// scheduling parameters for RXn-TXnp4 thread
-  struct sched_param sched_param_rxtx;
-
-  /// internal This variable is protected by ref mutex_fep_slot1.
-  //int instance_cnt_slot0_dl_processing;
-  int instance_cnt_slot1_dl_processing;
-  /// pthread descriptor fep_slot1 thread
-  //pthread_t pthread_slot0_dl_processing;
-  pthread_t pthread_slot1_dl_processing;
-  /// pthread attributes for fep_slot1 processing thread
-  // pthread_attr_t attr_slot0_dl_processing;
-  pthread_attr_t attr_slot1_dl_processing;
-  /// condition variable for UE fep_slot1 thread;
-  //pthread_cond_t cond_slot0_dl_processing;
-  pthread_cond_t cond_slot1_dl_processing;
-  /// mutex for UE synch thread
-  //pthread_mutex_t mutex_slot0_dl_processing;
-  pthread_mutex_t mutex_slot1_dl_processing;
-  //int instance_cnt_slot0_dl_processing;
-  int instance_cnt_dlsch_td;
-  /// pthread descriptor fep_slot1 thread
-  //pthread_t pthread_slot0_dl_processing;
-  pthread_t pthread_dlsch_td;
-  /// pthread attributes for fep_slot1 processing thread
-  // pthread_attr_t attr_slot0_dl_processing;
-  pthread_attr_t attr_dlsch_td;
-  /// condition variable for UE fep_slot1 thread;
-  //pthread_cond_t cond_slot0_dl_processing;
-  pthread_cond_t cond_dlsch_td;
-  /// mutex for UE synch thread
-  //pthread_mutex_t mutex_slot0_dl_processing;
-  pthread_mutex_t mutex_dlsch_td;
-  //
-  uint8_t chan_est_pilot0_slot1_available;
-  uint8_t chan_est_slot1_available;
-  uint8_t llr_slot1_available;
-  uint8_t dci_slot0_available;
-  uint8_t first_symbol_available;
-  uint8_t decoder_thread_available;
-  uint8_t decoder_main_available;
-  uint8_t decoder_switch;
-  int counter_decoder;
-  uint8_t channel_level;
-  int eNB_id;
-  int harq_pid;
-  int llr8_flag;
-  /// scheduling parameters for fep_slot1 thread
-  struct sched_param sched_param_fep_slot1;
-
-  int sub_frame_start;
-  int sub_frame_step;
-  unsigned long long gotIQs;
-  uint8_t decoder_thread_available1;
-  int instance_cnt_dlsch_td1;
-  /// pthread descriptor fep_slot1 thread
-  //pthread_t pthread_slot0_dl_processing;
-  pthread_t pthread_dlsch_td1;
-  /// pthread attributes for fep_slot1 processing thread
-  // pthread_attr_t attr_slot0_dl_processing;
-  pthread_attr_t attr_dlsch_td1;
-  /// condition variable for UE fep_slot1 thread;
-  //pthread_cond_t cond_slot0_dl_processing;
-  pthread_cond_t cond_dlsch_td1;
-  /// mutex for UE synch thread
-  //pthread_mutex_t mutex_slot0_dl_processing;
-  pthread_mutex_t mutex_dlsch_td1;
-  int dci_err_cnt;
-} UE_nr_rxtx_proc_t;
 
 /// Context data structure for eNB subframe processing
 typedef struct {
@@ -961,6 +858,9 @@ typedef struct UE_NR_SCAN_INFO_s {
   int32_t freq_offset_Hz[3][10];
 } UE_NR_SCAN_INFO_t;
 
+
+#include "NR_IF_Module.h"
+
 /// Top-level PHY Data Structure for UE
 typedef struct {
   /// \brief Module ID indicator for this instance

openair1/PHY/thread_NR_UE.h

+#ifndef __thread_NR_UE__
+#define __thread_NR_UE__
+#include <pthread.h>
+#include <targets/ARCH/COMMON/common_lib.h>
+/// Context data structure for RX/TX portion of subframe processing
+typedef struct {
+  /// index of the current UE RX/TX proc
+  int                  proc_id;
+  /// Component Carrier index
+  uint8_t              CC_id;
+  /// timestamp transmitted to HW
+  openair0_timestamp timestamp_tx;
+  //#ifdef UE_NR_PHY_DEMO
+  /// NR TTI index within subframe_tx [0 .. ttis_per_subframe - 1] to act upon for transmission
+  int nr_tti_tx;
+  /// NR TTI index within subframe_rx [0 .. ttis_per_subframe - 1] to act upon for reception
+  int nr_tti_rx;
+  //#endif
+  /// subframe to act upon for transmission
+  int subframe_tx;
+  /// subframe to act upon for reception
+  int subframe_rx;
+  /// frame to act upon for transmission
+  int frame_tx;
+  /// frame to act upon for reception
+  int frame_rx;
+  int decoded_frame_rx;
+  /// internal This variable is protected by ref mutex_fep_slot1.
+  //int instance_cnt_slot0_dl_processing;
+  int instance_cnt_slot1_dl_processing;
+  /// pthread descriptor fep_slot1 thread
+  //pthread_t pthread_slot0_dl_processing;
+  pthread_t pthread_slot1_dl_processing;
+  /// pthread attributes for fep_slot1 processing thread
+  // pthread_attr_t attr_slot0_dl_processing;
+  pthread_attr_t attr_slot1_dl_processing;
+  /// condition variable for UE fep_slot1 thread;
+  //pthread_cond_t cond_slot0_dl_processing;
+  pthread_cond_t cond_slot1_dl_processing;
+  /// mutex for UE synch thread
+  //pthread_mutex_t mutex_slot0_dl_processing;
+  pthread_mutex_t mutex_slot1_dl_processing;
+  //int instance_cnt_slot0_dl_processing;
+  int instance_cnt_dlsch_td;
+  /// pthread descriptor fep_slot1 thread
+  //pthread_t pthread_slot0_dl_processing;
+  pthread_t pthread_dlsch_td;
+  /// pthread attributes for fep_slot1 processing thread
+  // pthread_attr_t attr_slot0_dl_processing;
+  pthread_attr_t attr_dlsch_td;
+  /// condition variable for UE fep_slot1 thread;
+  //pthread_cond_t cond_slot0_dl_processing;
+  pthread_cond_t cond_dlsch_td;
+  /// mutex for UE synch thread
+  uint8_t chan_est_pilot0_slot1_available;
+  uint8_t chan_est_slot1_available;
+  uint8_t llr_slot1_available;
+  uint8_t dci_slot0_available;
+  uint8_t first_symbol_available;
+  uint8_t decoder_thread_available;
+  uint8_t decoder_main_available;
+  uint8_t decoder_switch;
+  int counter_decoder;
+  uint8_t channel_level;
+  int eNB_id;
+  int harq_pid;
+  int llr8_flag;
+  /// scheduling parameters for fep_slot1 thread
+  struct sched_param sched_param_fep_slot1;
+
+  int sub_frame_start;
+  int sub_frame_step;
+  unsigned long long gotIQs;
+  uint8_t decoder_thread_available1;
+  int dci_err_cnt;
+} UE_nr_rxtx_proc_t;
+
+#endif

openair1/SIMULATION/NR_PHY/dlschsim.c

@@ -174,7 +174,7 @@ int main(int argc, char **argv) {
 	//logInit();
 	randominit(0);
 
-	while ((c = getopt(argc, argv, "df:hpg:i:j:n:l:m:r:s:S:y:z:M:N:F:R:P:")) != -1) {
+	while ((c = getopt(argc, argv, "df:hpg:i:j:n:l:m:r:s:S:y:z:M:N:F:R:P:L:")) != -1) {
 		switch (c) {
 		case 'f':
 			write_output_file = 1;
@@ -315,6 +315,10 @@ int main(int argc, char **argv) {
 
 			break;
 
+		case 'L':
+		  loglvl = atoi(optarg);
+		  break;
+
 		case 'm':
 			Imcs = atoi(optarg);
 			break;

openair1/SIMULATION/NR_PHY/dlsim.c

@@ -684,7 +684,8 @@ int main(int argc, char **argv)
 	  LOG_M("rxsig1.m","rxs1", UE->common_vars.rxdata[1],frame_length_complex_samples,1,1);
       }
       if (UE->is_synchronized == 0) {
-	ret = nr_initial_sync(UE, normal_txrx);
+	UE_nr_rxtx_proc_t proc={0};
+	ret = nr_initial_sync(&proc, UE, normal_txrx);
 	printf("nr_initial_sync1 returns %d\n",ret);
 	if (ret<0) n_errors++;
       }

openair1/SIMULATION/NR_PHY/pbchsim.c

@@ -600,7 +600,8 @@ int main(int argc, char **argv)
       }
 
       if (UE->is_synchronized == 0) {
-	ret = nr_initial_sync(UE, normal_txrx);
+	UE_nr_rxtx_proc_t proc={0};
+	ret = nr_initial_sync(&proc, UE, normal_txrx);
 	printf("nr_initial_sync1 returns %d\n",ret);
 	if (ret<0) n_errors++;
       }
@@ -621,9 +622,10 @@ int main(int argc, char **argv)
 
           nr_pbch_channel_estimation(UE,0,0,i,i-(UE->symbol_offset+1),ssb_index%8,n_hf);
         }
+	UE_nr_rxtx_proc_t proc={0};
 
         ret = nr_rx_pbch(UE,
-	                 0,
+	                 &proc,
 		         UE->pbch_vars[0],
 		         frame_parms,
 		         0,

openair2/LAYER2/NR_MAC_UE/mac_proto.h

@@ -35,6 +35,7 @@
 
 #include "mac_defs.h"
 #include "mac.h"
+#include <openair1/PHY/defs_nr_UE.h>
 
 /**\brief decode mib pdu in NR_UE, from if_module ul_ind with P7 tx_ind message
    \param module_id      module id
@@ -46,6 +47,7 @@
    \param pdu_length     length of pdu
    \param cell_id        cell id */
 int8_t nr_ue_decode_mib(
+    UE_nr_rxtx_proc_t *proc,
     module_id_t module_id, 
     int cc_id, 
     uint8_t gNB_index, 

openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c

@@ -193,6 +193,7 @@ int8_t nr_ue_process_dlsch(module_id_t module_id, int cc_id, uint8_t gNB_index,
 }
 
 int8_t nr_ue_decode_mib(
+	UE_nr_rxtx_proc_t *proc,
 	module_id_t module_id,
 	int 		cc_id,
 	uint8_t 	gNB_index,
@@ -594,6 +595,7 @@ int8_t nr_ue_decode_mib(
 	    if(mac->if_module != NULL && mac->if_module->phy_config_request != NULL){
 		mac->if_module->phy_config_request(&mac->phy_config);
 	    }
+	    proc->decoded_frame_rx=frame;
     //}
     return 0;
 }

openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c

@@ -42,9 +42,10 @@
 static nr_ue_if_module_t *nr_ue_if_module_inst[MAX_IF_MODULES];
 
 //  L2 Abstraction Layer
-int handle_bcch_bch(module_id_t module_id, int cc_id, unsigned int gNB_index, uint8_t *pduP, unsigned int additional_bits, uint32_t ssb_index, uint32_t ssb_length, uint16_t cell_id){
+int handle_bcch_bch(UE_nr_rxtx_proc_t *proc, module_id_t module_id, int cc_id, unsigned int gNB_index, uint8_t *pduP, unsigned int additional_bits, uint32_t ssb_index, uint32_t ssb_length, uint16_t cell_id){
 
-  return nr_ue_decode_mib( module_id,
+  return nr_ue_decode_mib( proc,
+		           module_id,
 			   cc_id,
 			   gNB_index,
 			   additional_bits,
@@ -197,7 +198,8 @@ int nr_ue_dl_indication(nr_downlink_indication_t *dl_info){
     for(i=0; i<dl_info->rx_ind->number_pdus; ++i){
       switch(dl_info->rx_ind->rx_indication_body[i].pdu_type){
       case FAPI_NR_RX_PDU_TYPE_MIB:
-	ret_mask |= (handle_bcch_bch(dl_info->module_id, dl_info->cc_id, dl_info->gNB_index,
+	ret_mask |= (handle_bcch_bch( dl_info->proc,
+				     dl_info->module_id, dl_info->cc_id, dl_info->gNB_index,
 				     (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.pdu,
 				     (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.additional_bits,
 				     (dl_info->rx_ind->rx_indication_body+i)->mib_pdu.ssb_index,

openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.h

@@ -34,6 +34,7 @@
 #define __NR_IF_MODULE_H__
 
 #include "platform_types.h"
+#include <openair1/PHY/thread_NR_UE.h>
 #include "fapi_nr_ue_interface.h"
 
 
@@ -64,6 +65,7 @@ typedef struct {
     frame_t frame;
     /// slot
     int slot;
+    UE_nr_rxtx_proc_t * proc;
 
     /// NR UE FAPI-like P7 message, direction: L1 to L2
     /// data reception indication structure
@@ -209,7 +211,7 @@ int nr_ue_dcireq(nr_dcireq_t *dcireq);
    \param ssb_index       SSB index within 0 - (L_ssb-1) corresponding to 38.331 ch.13 parameter i
    \param ssb_length      corresponding to L1 parameter L_ssb 
    \param cell_id         cell id */
-int handle_bcch_bch(module_id_t module_id, int cc_id, unsigned int gNB_index, uint8_t *pduP, unsigned int additional_bits, uint32_t ssb_index, uint32_t ssb_length, uint16_t cell_id);
+int handle_bcch_bch(UE_nr_rxtx_proc_t *proc, module_id_t module_id, int cc_id, unsigned int gNB_index, uint8_t *pduP, unsigned int additional_bits, uint32_t ssb_index, uint32_t ssb_length, uint16_t cell_id);
 
 //  TODO check
 /**\brief handle BCCH-DL-SCH message from dl_indication