Add 5G preprocessor with NVS slicing

CMakeLists.txt

@@ -1366,6 +1366,7 @@ set (MAC_NR_SRC
   ${NR_GNB_MAC_DIR}/mac_rrc_dl_handler.c
   ${NR_GNB_MAC_DIR}/mac_rrc_ul_direct.c
   ${NR_GNB_MAC_DIR}/mac_rrc_ul_f1ap.c
+  ${NR_GNB_MAC_DIR}/slicing/nr_slicing.c
  )
 
 

doc/SW_archi.md

@@ -274,7 +274,7 @@ nr_preprocessor_phytest()], multiple users in FR1
   2)  Checks the quantity of waiting data in RLC
   3)  Either set up resource allocation directly (e.g., for a single UE,
       phytest), or call into a function to perform actual resource allocation.
-      Currently, this is done using pf_dl() which implements a basic
+      Currently, this is done using nr_pf_dl() which implements a basic
       proportional fair scheduler:
       * for every UE, check for retransmission and allocate as necessary
       * Calculate the PF coefficient and put eligible UEs into a list

openair1/SIMULATION/NR_PHY/dlsim.c

@@ -723,7 +723,7 @@ int main(int argc, char **argv)
   nr_mac_add_test_ue(RC.nrmac[0], secondaryCellGroup->spCellConfig->reconfigurationWithSync->newUE_Identity, secondaryCellGroup);
   // reset preprocessor to the one of DLSIM after it has been set during
   // nr_mac_config_scc()
-  gNB_mac->pre_processor_dl = nr_dlsim_preprocessor;
+  gNB_mac->pre_processor_dl.dl = nr_dlsim_preprocessor;
   phy_init_nr_gNB(gNB);
   N_RB_DL = gNB->frame_parms.N_RB_DL;
   NR_UE_info_t *UE_info = RC.nrmac[0]->UE_info.list[0];

openair2/LAYER2/NR_MAC_gNB/config.c

@@ -818,7 +818,7 @@ void nr_mac_config_scc(gNB_MAC_INST *nrmac, NR_ServingCellConfigCommon_t *scc, c
   }
 
   if (get_softmodem_params()->phy_test) {
-    nrmac->pre_processor_dl = nr_preprocessor_phytest;
+    nrmac->pre_processor_dl.dl = nr_preprocessor_phytest;
     nrmac->pre_processor_ul = nr_ul_preprocessor_phytest;
   } else {
     nrmac->pre_processor_dl = nr_init_fr1_dlsch_preprocessor(0);

openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_dlsch.c

@@ -317,7 +317,7 @@ int nr_write_ce_dlsch_pdu(module_id_t module_idP,
   return offset;
 }
 
-static void nr_store_dlsch_buffer(module_id_t module_id, frame_t frame, sub_frame_t slot)
+void nr_store_dlsch_buffer(module_id_t module_id, frame_t frame, sub_frame_t slot)
 {
   UE_iterator(RC.nrmac[module_id]->UE_info.list, UE) {
     NR_UE_sched_ctrl_t *sched_ctrl = &UE->UE_sched_ctrl;
@@ -589,13 +589,30 @@ static int comparator(const void *p, const void *q) {
   return ((UEsched_t*)p)->coef < ((UEsched_t*)q)->coef;
 }
 
-static void pf_dl(module_id_t module_id,
-                  frame_t frame,
-                  sub_frame_t slot,
-                  NR_UE_info_t **UE_list,
-                  int max_num_ue,
-                  int n_rb_sched,
-                  uint16_t *rballoc_mask)
+static void *nr_pf_dl_setup(void)
+{
+  void *data = malloc(MAX_MOBILES_PER_GNB * sizeof(float));
+  AssertFatal(data, "%s(): could not allocate data\n", __func__);
+  for (int i = 0; i < MAX_MOBILES_PER_ENB; i++)
+    *(float *) data = 0.0f;
+  return data;
+}
+static void nr_pf_dl_unset(void **data)
+{
+  DevAssert(data);
+  if (*data)
+    free(*data);
+  *data = NULL;
+}
+
+static void nr_pf_dl(module_id_t module_id,
+                     frame_t frame,
+                     sub_frame_t slot,
+                     NR_UE_info_t **UE_list,
+                     int max_num_ue,
+                     int n_rb_sched,
+                     uint16_t *rballoc_mask,
+                     void *data)
 {
   gNB_MAC_INST *mac = RC.nrmac[module_id];
   NR_ServingCellConfigCommon_t *scc=mac->common_channels[0].ServingCellConfigCommon;
@@ -827,6 +844,13 @@ static void pf_dl(module_id_t module_id,
     iterator++;
   }
 }
+nr_dl_sched_algo_t nr_proportional_fair_wbcqi_dl = {
+  .name  = "nr_proportional_fair_wbcqi_dl",
+  .setup = nr_pf_dl_setup,
+  .unset = nr_pf_dl_unset,
+  .run   = nr_pf_dl,
+  .data  = NULL
+};
 
 static void nr_fr1_dlsch_preprocessor(module_id_t module_id, frame_t frame, sub_frame_t slot)
 {
@@ -878,16 +902,17 @@ static void nr_fr1_dlsch_preprocessor(module_id_t module_id, frame_t frame, sub_
   int max_sched_ues = bw / (average_agg_level * NR_NB_REG_PER_CCE);
 
   /* proportional fair scheduling algorithm */
-  pf_dl(module_id,
-        frame,
-        slot,
-        UE_info->list,
-        max_sched_ues,
-        n_rb_sched,
-        rballoc_mask);
+  RC.nrmac[module_id]->pre_processor_dl.dl_algo.run(module_id,
+                                                    frame,
+                                                    slot,
+                                                    UE_info->list,
+                                                    max_sched_ues,
+                                                    n_rb_sched,
+                                                    rballoc_mask,
+                                                    RC.nrmac[module_id]->pre_processor_dl.dl_algo.data);
 }
 
-nr_pp_impl_dl nr_init_fr1_dlsch_preprocessor(int CC_id) {
+nr_pp_impl_param_dl_t nr_init_fr1_dlsch_preprocessor(int CC_id) {
   /* during initialization: no mutex needed */
   /* in the PF algorithm, we have to use the TBsize to compute the coefficient.
    * This would include the number of DMRS symbols, which in turn depends on
@@ -912,7 +937,12 @@ nr_pp_impl_dl nr_init_fr1_dlsch_preprocessor(int CC_id) {
     }
   }
 
-  return nr_fr1_dlsch_preprocessor;
+  nr_pp_impl_param_dl_t impl;
+  memset(&impl, 0, sizeof(impl));
+  impl.dl = nr_fr1_dlsch_preprocessor;
+  impl.dl_algo = nr_proportional_fair_wbcqi_dl;
+  impl.dl_algo.data = impl.dl_algo.setup();
+  return impl;
 }
 
 void nr_schedule_ue_spec(module_id_t module_id,
@@ -930,7 +960,10 @@ void nr_schedule_ue_spec(module_id_t module_id,
     return;
 
   /* PREPROCESSOR */
-  gNB_mac->pre_processor_dl(module_id, frame, slot);
+  pthread_mutex_lock(&gNB_mac->UE_info.mutex);
+  gNB_mac->pre_processor_dl.dl(module_id, frame, slot);
+  pthread_mutex_unlock(&gNB_mac->UE_info.mutex);
+
   const int CC_id = 0;
   NR_ServingCellConfigCommon_t *scc = gNB_mac->common_channels[CC_id].ServingCellConfigCommon;
   NR_UEs_t *UE_info = &gNB_mac->UE_info;

openair2/LAYER2/NR_MAC_gNB/mac_proto.h

@@ -76,7 +76,7 @@ void nr_schedule_ue_spec(module_id_t module_id,
                          nfapi_nr_tx_data_request_t *TX_req);
 
 /* \brief default FR1 DL preprocessor init routine, returns preprocessor to call */
-nr_pp_impl_dl nr_init_fr1_dlsch_preprocessor(int CC_id);
+nr_pp_impl_param_dl_t nr_init_fr1_dlsch_preprocessor(int CC_id);
 
 void schedule_nr_sib1(module_id_t module_idP,
                       frame_t frameP,
@@ -179,7 +179,7 @@ int nr_acknack_scheduling(gNB_MAC_INST *mac,
 int get_pdsch_to_harq_feedback(NR_PUCCH_Config_t *pucch_Config,
                                nr_dci_format_t dci_format,
                                uint8_t *pdsch_to_harq_feedback);
-  
+
 int nr_get_pucch_resource(NR_ControlResourceSet_t *coreset,
                           NR_PUCCH_Config_t *pucch_Config,
                           int CCEIndex);

openair2/LAYER2/NR_MAC_gNB/main.c

@@ -259,7 +259,7 @@ void mac_top_init_gNB(ngran_node_t node_type,
       uid_linear_allocator_init(&RC.nrmac[i]->UE_info.uid_allocator);
 
       if (get_softmodem_params()->phy_test) {
-        RC.nrmac[i]->pre_processor_dl = nr_preprocessor_phytest;
+        RC.nrmac[i]->pre_processor_dl.dl = nr_preprocessor_phytest;
         RC.nrmac[i]->pre_processor_ul = nr_ul_preprocessor_phytest;
       } else {
         RC.nrmac[i]->pre_processor_dl = nr_init_fr1_dlsch_preprocessor(0);

openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h

@@ -716,6 +716,8 @@ typedef struct {
   float ul_thr_ue;
   float dl_thr_ue;
   long pdsch_HARQ_ACK_Codebook;
+  /// Assoc slice
+  slice_id_t dl_id;
 } NR_UE_info_t;
 
 typedef struct {
@@ -730,6 +732,19 @@ typedef struct {
 
 #define UE_iterator(BaSe, VaR) NR_UE_info_t ** VaR##pptr=BaSe, *VaR; while ((VaR=*(VaR##pptr++)))
 
+/**
+ * definition of a scheduling algorithm implementation used in the
+ * default DL scheduler
+ */
+typedef struct {
+  char *name;
+  void *(*setup)(void);
+  void (*unset)(void **);
+  void (*run)(
+      module_id_t, frame_t, sub_frame_t, NR_UE_info_t **, int, int, uint16_t *, void *);
+  void *data;
+} nr_dl_sched_algo_t;
+
 typedef void (*nr_pp_impl_dl)(module_id_t mod_id,
                               frame_t frame,
                               sub_frame_t slot);
@@ -737,6 +752,43 @@ typedef bool (*nr_pp_impl_ul)(module_id_t mod_id,
                               frame_t frame,
                               sub_frame_t slot);
 
+struct nr_slice_info_s;
+struct nr_slice_s;
+typedef struct {
+  int algorithm;
+
+  /// inform the slice algorithm about a new UE
+  void (*add_UE)(struct nr_slice_info_s *s, NR_UE_info_t **UE_list);
+  /// inform the slice algorithm about a UE that disconnected
+  void (*remove_UE)(struct nr_slice_info_s *s, NR_UE_info_t* rm_ue, int idx);
+  /// move a UE to a slice in DL/UL, -1 means don't move (no-op).
+  void (*move_UE)(struct nr_slice_info_s *s, NR_UE_info_t* assoc_ue, int old_idx, int new_idx);
+  /// get UE associated slice's index
+  int (*get_UE_slice_idx)(struct nr_slice_info_s *s, uint16_t rnti);
+  /// get UE's index from the slice
+  int (*get_UE_idx)(struct nr_slice_s *si, uint16_t rnti);
+
+  /// Adds a new slice through admission control. slice_params are
+  /// algorithm-specific parameters. sched is either a default_sched_ul_algo_t
+  /// or default_sched_dl_algo_t, depending on whether this implementation
+  /// handles UL/DL. If slice at index exists, updates existing
+  /// slice. Returns index of new slice or -1 on failure.
+  int (*addmod_slice)(struct nr_slice_info_s *s,
+                      int id,
+                      char *label,
+                      void *sched,
+                      void *slice_params);
+  /// Returns slice through slice_idx. 1 if successful, 0 if not.
+  int (*remove_slice)(struct nr_slice_info_s *s, uint8_t slice_idx);
+
+  nr_pp_impl_dl dl;
+  nr_dl_sched_algo_t dl_algo;
+
+  void (*destroy)(struct nr_slice_info_s **s);
+
+  struct nr_slice_info_s *slices;
+} nr_pp_impl_param_dl_t;
+
 typedef struct f1_config_t {
   f1ap_setup_req_t *setup_req;
   f1ap_setup_resp_t *setup_resp;
@@ -835,7 +887,7 @@ typedef struct gNB_MAC_INST_s {
   uint32_t ulsch_max_frame_inactivity;
 
   /// DL preprocessor for differentiated scheduling
-  nr_pp_impl_dl pre_processor_dl;
+  nr_pp_impl_param_dl_t pre_processor_dl;
   /// UL preprocessor for differentiated scheduling
   nr_pp_impl_ul pre_processor_ul;
 

openair2/LAYER2/NR_MAC_gNB/slicing/nr_slicing.c

+/*
+ * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The OpenAirInterface Software Alliance licenses this file to You under
+ * the OAI Public License, Version 1.1  (the "License"); you may not use this file
+ * except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *      http://www.openairinterface.org/?page_id=698
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *-------------------------------------------------------------------------------
+ * For more information about the OpenAirInterface (OAI) Software Alliance:
+ *      contact@openairinterface.org
+ */
+
+/*!
+ * \file   nr_slicing.c
+ * \brief  Generic NR Slicing helper functions and Static Slicing Implementation
+ * \author Robert Schmidt
+ * \date   2021
+ * \email  robert.schmidt@eurecom.fr
+ */
+
+#define _GNU_SOURCE
+#include <stdlib.h>
+#include <dlfcn.h>
+
+#include "assertions.h"
+#include "common/utils/LOG/log.h"
+
+#include "NR_MAC_COMMON/nr_mac_extern.h"
+#include "NR_MAC_COMMON/nr_mac.h"
+#include "openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h"
+#include "openair2/LAYER2/NR_MAC_gNB/mac_proto.h"
+#include "openair2/LAYER2/RLC/rlc.h"
+
+#include "nr_slicing.h"
+#include "nr_slicing_internal.h"
+
+#include "executables/softmodem-common.h"
+
+extern RAN_CONTEXT_t RC;
+
+#define RET_FAIL(ret, x...) do { LOG_E(MAC, x); return ret; } while (0)
+
+int nr_slicing_get_UE_slice_idx(nr_slice_info_t *si, uint16_t rnti) {
+  for (int s_len = 0; s_len < si->num; s_len++) {
+    // Using UE_iterator will stop at NULL => not work for 2 UEs
+    for (int i = 0; i < MAX_MOBILES_PER_GNB; i++) {
+      if (si->s[s_len]->UE_list[i] != NULL) {
+        if (si->s[s_len]->UE_list[i]->rnti == rnti) {
+          return si->s[s_len]->id;
+        }
+      }
+    }
+  }
+  LOG_E(NR_MAC, "cannot find slice idx for UE rnti 0x%04x\n", rnti);
+  return -99;
+}
+
+int nr_slicing_get_UE_idx(nr_slice_t *si, uint16_t rnti) {
+  for (int i = 0; i < MAX_MOBILES_PER_GNB; i++) {
+    // Check for NULL because w.r.t 2 UEs, the former UE may already be moved!
+    if (si->UE_list[i] != NULL) {
+      if (si->UE_list[i]->rnti == rnti)
+        return i;
+    }
+  }
+  LOG_E(NR_MAC, "cannot find ue idx for UE rnti 0x%04x\n", rnti);
+  return -99;
+}
+
+void nr_slicing_add_UE(nr_slice_info_t *si, NR_UE_info_t **UE_list) {
+  // Add all the connected UEs to the first slice 0
+  UE_iterator(UE_list, UE) {
+    if (UE) {
+      UE->dl_id = 0;
+      for (int i = 0; i < MAX_MOBILES_PER_GNB; i++) {
+        if (si->s[0]->UE_list[i] == NULL) {
+          si->s[0]->UE_list[i] = UE;
+          LOG_D(NR_MAC, "%s(), add UE_list[%d], rnti 0x%04x to slice idx 0\n", __func__, i, si->s[0]->UE_list[i]->rnti);
+          break;
+        }
+      }
+    }
+  }
+
+}
+
+void nr_slicing_remove_UE(nr_slice_info_t *si, NR_UE_info_t* rm_ue, int idx) {
+  for (int i = 0; i < MAX_MOBILES_PER_GNB; i++) {
+    // Check for NULL because w.r.t 2 UEs, the former UE may already be moved!
+    if(si->s[idx]->UE_list[i] != NULL) {
+      if (si->s[idx]->UE_list[i]->rnti == rm_ue->rnti) {
+        si->s[idx]->UE_list[i] = NULL;
+        rm_ue->dl_id = -1;
+        break;
+      }
+    }
+  }
+}
+
+void nr_slicing_move_UE(nr_slice_info_t *si, NR_UE_info_t* assoc_ue, int old_idx, int new_idx) {
+  DevAssert(new_idx >= -1 && new_idx < si->num);
+  DevAssert(old_idx >= -1 && old_idx < si->num);
+
+  // remove UE from old slice
+  nr_slicing_remove_UE(si, assoc_ue, old_idx);
+
+  // add UE to new slice
+  for (int i = 0; i < MAX_MOBILES_PER_GNB; i++) {
+    if (si->s[new_idx]->UE_list[i] == NULL) {
+      assoc_ue->dl_id = si->s[new_idx]->id;
+      si->s[new_idx]->UE_list[i] = assoc_ue;
+      break;
+    }
+  }
+}
+
+int _nr_exists_slice(uint8_t n, nr_slice_t **s, int id) {
+  for (int i = 0; i < n; ++i)
+    if (s[i]->id == id)
+      return i;
+  return -1;
+}
+
+nr_slice_t *_nr_add_slice(uint8_t *n, nr_slice_t **s) {
+  s[*n] = calloc(1, sizeof(nr_slice_t));
+  if (!s[*n])
+    return NULL;
+  create_nr_list(&s[*n]->UEs, MAX_MOBILES_PER_GNB);
+  *n += 1;
+  return s[*n - 1];
+}
+
+nr_slice_t *_nr_remove_slice(uint8_t *n, nr_slice_t **s, int idx) {
+  if (idx >= *n)
+    return NULL;
+
+  nr_slice_t *sr = s[idx];
+
+  for (int i = idx + 1; i < *n; ++i)
+    s[i - 1] = s[i];
+  *n -= 1;
+  s[*n] = NULL;
+
+  if (sr->label)
+    free(sr->label);
+
+  return sr;
+}
+
+
+/************************* NVS Slicing Implementation **************************/
+
+typedef struct {
+  float exp; // exponential weight. mov. avg for weight calc
+  int   rb;  // number of RBs this slice has been scheduled in last round
+  float eff; // effective rate for rate slices
+  float beta_eff; // averaging coeff so we average over roughly one second
+  int   active;   // activity state for rate slices
+} _nvs_int_t;
+
+int _nvs_nr_admission_control(const nr_slice_info_t *si,
+                              const nvs_nr_slice_param_t *p,
+                              int idx)
+{
+  if (p->type != NVS_RATE && p->type != NVS_RES)
+    RET_FAIL(-1, "%s(): invalid slice type %d\n", __func__, p->type);
+  if (p->type == NVS_RATE && p->Mbps_reserved > p->Mbps_reference)
+    RET_FAIL(-1,
+             "%s(): a rate slice cannot reserve more than the reference rate\n",
+             __func__);
+  if (p->type == NVS_RES && p->pct_reserved > 1.0f)
+    RET_FAIL(-1, "%s(): cannot reserve more than 1.0\n", __func__);
+  float sum_req = 0.0f;
+  for (int i = 0; i < si->num; ++i) {
+    const nvs_nr_slice_param_t *sp = i == idx ? p : si->s[i]->algo_data;
+    if (sp->type == NVS_RATE)
+      sum_req += sp->Mbps_reserved / sp->Mbps_reference;
+    else
+      sum_req += sp->pct_reserved;
+  }
+  if (idx < 0) { /* not an existing slice */
+    if (p->type == NVS_RATE)
+      sum_req += p->Mbps_reserved / p->Mbps_reference;
+    else
+      sum_req += p->pct_reserved;
+  }
+  if (sum_req > 1.0)
+    RET_FAIL(-3,
+             "%s(): admission control failed: sum of resources is %f > 1.0\n",
+             __func__, sum_req);
+  return 0;
+}
+
+int addmod_nvs_nr_slice_dl(nr_slice_info_t *si,
+                           int id,
+                           char *label,
+                           void *algo,
+                           void *slice_params_dl)
+{
+  nvs_nr_slice_param_t *dl = slice_params_dl;
+  int index = _nr_exists_slice(si->num, si->s, id);
+  if (index < 0 && si->num >= MAX_NVS_SLICES)
+    RET_FAIL(-2, "%s(): cannot handle more than %d slices\n", __func__, MAX_NVS_SLICES);
+
+  if (index < 0 && !dl)
+    RET_FAIL(-100, "%s(): no parameters for new slice %d, aborting\n", __func__, id);
+
+  if (dl) {
+    int rc = _nvs_nr_admission_control(si, dl, index);
+    if (rc < 0)
+      return rc;
+  }
+
+  nr_slice_t *s = NULL;
+  if (index >= 0) {
+    s = si->s[index];
+    if (label) {
+      if (s->label) free(s->label);
+      s->label = label;
+    }
+    if (algo && s->dl_algo.run != ((nr_dl_sched_algo_t*)algo)->run) {
+      s->dl_algo.unset(&s->dl_algo.data);
+      s->dl_algo = *(nr_dl_sched_algo_t*) algo;
+      if (!s->dl_algo.data)
+        s->dl_algo.data = s->dl_algo.setup();
+    }
+    if (dl) {
+      free(s->algo_data);
+      s->algo_data = dl;
+    } else { /* we have no parameters: we are done */
+      return index;
+    }
+  } else {
+    if (!algo)
+      RET_FAIL(-14, "%s(): no scheduler algorithm provided\n", __func__);
+
+    s = _nr_add_slice(&si->num, si->s);
+    if (!s)
+      RET_FAIL(-4, "%s(): cannot allocate memory for slice\n", __func__);
+    s->int_data = malloc(sizeof(_nvs_int_t));
+    if (!s->int_data)
+      RET_FAIL(-5, "%s(): cannot allocate memory for slice internal data\n", __func__);
+
+    s->id = id;
+    s->label = label;
+    s->dl_algo = *(nr_dl_sched_algo_t*) algo;
+    if (!s->dl_algo.data)
+      s->dl_algo.data = s->dl_algo.setup();
+    s->algo_data = dl;
+  }
+
+  _nvs_int_t *nvs_p = s->int_data;
+  /* reset all slice-internal parameters */
+  nvs_p->rb = 0;
+  nvs_p->active = 0;
+  if (dl->type == NVS_RATE) {
+    nvs_p->exp = dl->Mbps_reserved / dl->Mbps_reference;
+    nvs_p->eff = dl->Mbps_reference;
+  } else {
+    nvs_p->exp = dl->pct_reserved;
+    nvs_p->eff = 0; // not used
+  }
+  // scale beta so we (roughly) average the eff rate over 1s
+  nvs_p->beta_eff = BETA / nvs_p->exp;
+
+  return index < 0 ? si->num - 1 : index;
+}
+
+//int addmod_nvs_slice_ul(nr_slice_info_t *si,
+//                        int id,
+//                        char *label,
+//                        void *slice_params_ul) {
+//  nvs_nr_slice_param_t *sp = slice_params_ul;
+//  int index = _nr_exists_slice(si->num, si->s, id);
+//  if (index < 0 && si->num >= MAX_NVS_SLICES)
+//    RET_FAIL(-2, "%s(): cannot handle more than %d slices\n", __func__, MAX_NVS_SLICES);
+//
+//  int rc = _nvs_admission_control(si->num, si->s, sp, index);
+//  if (rc < 0)
+//    return rc;
+//
+//  nr_slice_t *ns = NULL;
+//  if (index < 0) {
+//    ns = _add_slice(&si->num, si->s);
+//    if (!ns)
+//      RET_FAIL(-4, "%s(): cannot allocate memory for slice\n", __func__);
+//    ns->id = id;
+//    ns->int_data = malloc(sizeof(_nvs_int_t));
+//    if (!ns->int_data)
+//      RET_FAIL(-5, "%s(): cannot allocate memory for slice internal data\n",
+//               __func__);
+//  } else {
+//    ns = si->s[index];
+//    free(ns->algo_data);
+//  }
+//  if (label) {
+//    if (ns->label)
+//      free(ns->label);
+//    ns->label = label;
+//  }
+//  ns->algo_data = sp;
+//  _nvs_int_t *nvs_p = ns->int_data;
+//  nvs_p->rb = 0;
+//  nvs_p->active = 0;
+//  if (sp->type == NVS_RATE) {
+//    nvs_p->exp = sp->Mbps_reserved;
+//    nvs_p->eff = sp->Mbps_reference;
+//  } else {
+//    nvs_p->exp = sp->pct_reserved;
+//    nvs_p->eff = 0; // not used
+//  }
+//
+//  return si->num - 1;
+//}
+
+int remove_nvs_nr_slice_dl(nr_slice_info_t *si, uint8_t slice_idx)
+{
+  if (slice_idx == 0 && si->num <= 1)
+    return 0;
+  UE_iterator(si->s[slice_idx]->UE_list, rm_ue) {
+    nr_slicing_remove_UE(si, rm_ue, slice_idx);
+    rm_ue->dl_id = si->s[0]->id;
+    LOG_D(NR_MAC, "%s(), move UE rnti 0x%04x in slice ID %d idx %d to slice ID %d idx %d\n",
+          __func__, rm_ue->rnti, si->s[slice_idx]->id, slice_idx, si->s[0]->id, 0);
+    for (int i = 0; i < MAX_MOBILES_PER_GNB; i++) {
+      if (si->s[0]->UE_list[i] == NULL) {
+        si->s[0]->UE_list[i] = rm_ue;
+        break;
+      }
+    }
+  }
+
+  nr_slice_t *sr = _nr_remove_slice(&si->num, si->s, slice_idx);
+  if (!sr)
+    return 0;
+  free(sr->algo_data);
+  free(sr->int_data);
+  sr->dl_algo.unset(&sr->dl_algo.data);
+  free(sr);
+  return 1;
+}
+
+//int remove_nvs_slice_ul(nr_slice_info_t *si, uint8_t slice_idx) {
+//  if (slice_idx == 0)
+//    return 0;
+//  nr_slice_t *sr = _remove_slice(&si->num, si->s, si->UE_assoc_slice, slice_idx);
+//  if (!sr)
+//    return 0;
+//  free(sr->algo_data);
+//  free(sr->int_data);
+//  free(sr);
+//  return 1;
+//}
+
+extern void nr_store_dlsch_buffer(module_id_t, frame_t, sub_frame_t);
+
+void nvs_nr_dl(module_id_t mod_id,
+               frame_t frame,
+               sub_frame_t slot)
+{
+  gNB_MAC_INST *nrmac = RC.nrmac[mod_id];
+  NR_UEs_t *UE_info = &nrmac->UE_info;
+  if (UE_info->list[0] == NULL) /* no UEs at all -> don't bother */
+    return;
+
+  /* check if we are supposed to schedule something */
+  NR_ServingCellConfigCommon_t *scc = RC.nrmac[mod_id]->common_channels[0].ServingCellConfigCommon;
+  const int CC_id = 0;
+  /* Get bwpSize and TDAfrom the first UE */
+  /* This is temporary and it assumes all UEs have the same BWP and TDA*/
+  NR_UE_info_t *first_UE=UE_info->list[0];
+  NR_UE_sched_ctrl_t *sched_ctrl = &first_UE->UE_sched_ctrl;
+  NR_UE_DL_BWP_t *current_BWP = &first_UE->current_DL_BWP;
+  const int tda = get_dl_tda(RC.nrmac[mod_id], scc, slot);
+  int startSymbolIndex, nrOfSymbols;
+  const int coresetid = sched_ctrl->coreset->controlResourceSetId;
+  const struct NR_PDSCH_TimeDomainResourceAllocationList *tdaList = get_dl_tdalist(current_BWP, coresetid, sched_ctrl->search_space->searchSpaceType->present, TYPE_C_RNTI_);
+  AssertFatal(tda < tdaList->list.count, "time_domain_allocation %d>=%d\n", tda, tdaList->list.count);
+  const int startSymbolAndLength = tdaList->list.array[tda]->startSymbolAndLength;
+  SLIV2SL(startSymbolAndLength, &startSymbolIndex, &nrOfSymbols);
+
+  const uint16_t bwpSize = coresetid == 0 ? RC.nrmac[mod_id]->cset0_bwp_size : current_BWP->BWPSize;
+  const uint16_t BWPStart = coresetid == 0 ? RC.nrmac[mod_id]->cset0_bwp_start : current_BWP->BWPStart;
+
+  const uint16_t slbitmap = SL_to_bitmap(startSymbolIndex, nrOfSymbols);
+  uint16_t *vrb_map = RC.nrmac[mod_id]->common_channels[CC_id].vrb_map;
+  uint16_t rballoc_mask[bwpSize];
+  int n_rb_sched = 0;
+
+  for (int i = 0; i < bwpSize; i++) {
+    // calculate mask: init with "NOT" vrb_map:
+    // if any RB in vrb_map is blocked (1), the current RBG will be 0
+    rballoc_mask[i] = (~vrb_map[i+BWPStart])&0x3fff; //bitwise not and 14 symbols
+
+    // if all the pdsch symbols are free
+    if ((rballoc_mask[i]&slbitmap) == slbitmap) {
+      n_rb_sched++;
+    }
+  }
+
+  /* Retrieve amount of data to send */
+  nr_store_dlsch_buffer(mod_id, frame, slot);
+
+  nr_slice_info_t *si = RC.nrmac[mod_id]->pre_processor_dl.slices;
+  int bytes_last_round = 0;
+  UE_iterator(UE_info->list, UE) {
+    const NR_UE_sched_ctrl_t *sched_ctrl = &UE->UE_sched_ctrl;
+
+    bytes_last_round += UE->mac_stats.dl.current_bytes;
+
+    int s_idx;
+    for (s_idx = 0; s_idx < si->num; s_idx++)
+      if (si->s[s_idx]->id == UE->dl_id)
+        break;
+    DevAssert(s_idx >= 0 && s_idx < si->num);
+
+    /* if UE has data or retransmission, mark respective slice as active */
+    const int retx_pid = sched_ctrl->retrans_dl_harq.head;
+    const bool active = sched_ctrl->num_total_bytes > 0 || retx_pid >= 0;
+    ((_nvs_int_t *)si->s[s_idx]->int_data)->active |= active;
+  }
+
+  float maxw = 0.0f;
+  int maxidx = -1;
+  for (int i = 0; i < si->num; ++i) {
+    nr_slice_t *s = si->s[i];
+    nvs_nr_slice_param_t *p = s->algo_data;
+    _nvs_int_t *ip = s->int_data;
+
+    float w = 0.0f;
+    if (p->type == NVS_RATE) {
+      /* if this slice has been marked as inactive, disable to prevent that
+       * it's exp rate is uselessly driven down */
+      if (!ip->active)
+        continue;
+      float inst = 0.0f;
+      if (ip->rb > 0) { /* it was scheduled last round */
+        /* inst rate: B in last round * 8(bit) / 1000000 (Mbps) * 1000 (1ms) */
+        inst = (float) bytes_last_round * 8 / 1000;
+        ip->eff = (1.0f - ip->beta_eff) * ip->eff + ip->beta_eff * inst;
+        //LOG_W(NR_MAC, "i %d slice %d ip->rb %d inst %f ip->eff %f\n", i, s->id, ip->rb, inst, ip->eff);
+        ip->rb = 0;
+      }
+      ip->exp = (1 - BETA) * ip->exp + BETA * inst;
+      const float rsv = p->Mbps_reserved * min(1.0f, ip->eff / p->Mbps_reference);
+      w = rsv / ip->exp;
+    } else {
+      float inst = (float)ip->rb / bwpSize;
+      ip->exp = (1.0f - BETA) * ip->exp + BETA * inst;
+      w = p->pct_reserved / ip->exp;
+    }
+    //LOG_I(NR_MAC, "i %d slice %d type %d ip->exp %f w %f\n", i, s->id, p->type, ip->exp, w);
+    ip->rb = 0;
+    if (w > maxw + 0.001f) {
+      maxw = w;
+      maxidx = i;
+    }
+  }
+
+  if (maxidx < 0)
+    return;
+
+  ((_nvs_int_t *)si->s[maxidx]->int_data)->rb = n_rb_sched;
+
+  //int rbg_rem = n_rb_sched;
+  if (si->s[maxidx]->UE_list[0] != NULL) {
+    /*rbg_rem = */
+    LOG_D(NR_MAC, "%4d.%2d scheduling slice idx %d ID %d (first UE rnti 0x%04x)\n", frame, slot, maxidx, si->s[maxidx]->id, si->s[maxidx]->UE_list[0]->rnti);
+    si->s[maxidx]->dl_algo.run(mod_id,
+                               frame,
+                               slot,
+                               si->s[maxidx]->UE_list,
+                               2, // max_num_ue
+                               n_rb_sched,
+                               rballoc_mask,
+                               si->s[maxidx]->dl_algo.data);
+  } else {
+    LOG_D(NR_MAC, "%4d.%2d not scheduling slice idx %d ID %d (no UEs)\n", frame, slot, maxidx, si->s[maxidx]->id);
+  }
+  // TODO
+  //if (rbg_rem == n_rbg_sched) // if no RBGs have been used mark as inactive
+  //  ((_nvs_int_t *)si->s[maxidx]->int_data)->active = 0;
+}
+
+/*
+void nvs_ul(module_id_t mod_id,
+               int CC_id,
+               frame_t frame,
+               sub_frame_t subframe,
+               frame_t sched_frame,
+               sub_frame_t sched_subframe) {
+  ulsch_scheduler_pre_processor(mod_id, CC_id, frame, subframe, sched_frame, sched_subframe);
+}
+*/
+
+void nvs_nr_destroy(nr_slice_info_t **si) {
+  const int n_dl = (*si)->num;
+  (*si)->num = 0;
+  for (int i = 0; i < n_dl; ++i) {
+    nr_slice_t *s = (*si)->s[i];
+    if (s->label)
+      free(s->label);
+    free(s->algo_data);
+    free(s->int_data);
+    free(s);
+  }
+  free((*si)->s);
+}
+
+nr_pp_impl_param_dl_t nvs_nr_dl_init(module_id_t mod_id, int CC_id)
+{
+  nr_slice_info_t *si = calloc(1, sizeof(nr_slice_info_t));
+  DevAssert(si);
+
+  si->num = 0;
+  si->s = calloc(MAX_NVS_SLICES, sizeof(nr_slice_t));
+  DevAssert(si->s);
+  for (int i = 0; i < MAX_MOBILES_PER_GNB; ++i)
+    si->UE_assoc_slice[i] = -1;
+
+  /* insert default slice, all resources */
+  nvs_nr_slice_param_t *dlp = malloc(sizeof(nvs_nr_slice_param_t));
+  DevAssert(dlp);
+  dlp->type = NVS_RES;
+  dlp->pct_reserved = 1.0f;
+  nr_dl_sched_algo_t *algo = &RC.nrmac[mod_id]->pre_processor_dl.dl_algo;
+  algo->data = NULL;
+  const int rc = addmod_nvs_nr_slice_dl(si, 0, strdup("default"), algo, dlp);
+  DevAssert(0 == rc);
+  NR_UEs_t *UE_info = &RC.nrmac[mod_id]->UE_info;
+  nr_slicing_add_UE(si, UE_info->list);
+
+  nr_pp_impl_param_dl_t nvs;
+  nvs.algorithm = NVS_SLICING;
+  nvs.add_UE = nr_slicing_add_UE;
+  nvs.remove_UE = nr_slicing_remove_UE;
+  nvs.move_UE = nr_slicing_move_UE;
+  nvs.get_UE_slice_idx = nr_slicing_get_UE_slice_idx;
+  nvs.get_UE_idx = nr_slicing_get_UE_idx;
+  nvs.addmod_slice = addmod_nvs_nr_slice_dl;
+  nvs.remove_slice = remove_nvs_nr_slice_dl;
+  nvs.dl = nvs_nr_dl;
+  // current DL algo becomes default scheduler
+  nvs.dl_algo = *algo;
+  nvs.destroy = nvs_nr_destroy;
+  nvs.slices = si;
+
+  return nvs;
+}

openair2/LAYER2/NR_MAC_gNB/slicing/nr_slicing.h

+/*
+ * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The OpenAirInterface Software Alliance licenses this file to You under
+ * the OAI Public License, Version 1.1  (the "License"); you may not use this file
+ * except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *      http://www.openairinterface.org/?page_id=698
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *-------------------------------------------------------------------------------
+ * For more information about the OpenAirInterface (OAI) Software Alliance:
+ *      contact@openairinterface.org
+ */
+
+/*!
+ * \file   nr_slicing.h
+ * \brief  General NR slice definition and helper parameters
+ * \author Robert Schmidt
+ * \date   2021
+ * \email  robert.schmidt@eurecom.fr
+ */
+
+#ifndef NR_SLICING_H__
+#define NR_SLICING_H__
+
+#include "openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h"
+
+typedef struct nr_slice_s {
+  /// Arbitrary ID
+  slice_id_t id;
+  /// Arbitrary label
+  char *label;
+
+  nr_dl_sched_algo_t dl_algo;
+
+  /// A specific algorithm's implementation parameters
+  void *algo_data;
+  /// Internal data that might be kept alongside a slice's params
+  void *int_data;
+
+  // list of users in this slice
+  NR_list_t UEs;
+  NR_UE_info_t *UE_list[MAX_MOBILES_PER_GNB+1];
+} nr_slice_t;
+
+typedef struct nr_slice_info_s {
+  uint8_t num;
+  nr_slice_t **s;
+  uint8_t UE_assoc_slice[MAX_MOBILES_PER_GNB+1];
+} nr_slice_info_t;
+
+#define NVS_SLICING 20
+/* arbitrary upper limit, increase if you want to instantiate more slices */
+#define MAX_NVS_SLICES 10
+/* window for slice weight averaging -> 1s for fine granularity */
+#define BETA 0.001f
+typedef struct {
+  enum nvs_type {NVS_RATE, NVS_RES} type;
+  union {
+    struct { float Mbps_reserved; float Mbps_reference; };
+    struct { float pct_reserved; };
+  };
+} nvs_nr_slice_param_t;
+nr_pp_impl_param_dl_t nvs_nr_dl_init(module_id_t mod_id, int CC_id);
+
+#endif /* NR_SLICING_H__ */

openair2/LAYER2/NR_MAC_gNB/slicing/nr_slicing_internal.h

+/*
+ * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The OpenAirInterface Software Alliance licenses this file to You under
+ * the OAI Public License, Version 1.1  (the "License"); you may not use this file
+ * except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *      http://www.openairinterface.org/?page_id=698
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *-------------------------------------------------------------------------------
+ * For more information about the OpenAirInterface (OAI) Software Alliance:
+ *      contact@openairinterface.org
+ */
+
+/*!
+ * \file   nr_slicing_internal.h
+ * \brief  Internal NR slice helper functions
+ * \author Robert Schmidt
+ * \date   2021
+ * \email  robert.schmidt@eurecom.fr
+ */
+
+#ifndef NR_SLICING_INTERNAL_H__
+#define NR_SLICING_INTERNAL_H__
+
+#include "nr_slicing.h"
+
+void nr_slicing_add_UE(nr_slice_info_t *si, NR_UE_info_t **UE_list);
+
+void nr_slicing_remove_UE(nr_slice_info_t *si, NR_UE_info_t* rm_ue, int idx);
+
+void nr_slicing_move_UE(nr_slice_info_t *si, NR_UE_info_t* assoc_ue, int old_idx, int new_idx);
+
+int nr_slicing_get_UE_slice_idx(nr_slice_info_t *si, uint16_t rnti);
+
+int nr_slicing_get_UE_idx(nr_slice_t *si, uint16_t rnti);
+
+nr_slice_t *_nr_add_slice(uint8_t *n, nr_slice_t **s);
+nr_slice_t *_nr_remove_slice(uint8_t *n, nr_slice_t **s, int idx);
+
+#endif /* NR_SLICING_INTERNAL_H__ */