Added automatic generation of rte core masks for eal_init

plugins/DPDKReaderModule.cpp

@@ -14,6 +14,8 @@
 
 #include "appmodel/DataReceiverModule.hpp"
 #include "appmodel/DPDKReaderConf.hpp"
+#include "appmodel/DPDKPortConfiguration.hpp"
+#include "confmodel/ProcessingResource.hpp"
 #include "confmodel/NetworkDevice.hpp"
 #include "confmodel/QueueWithSourceId.hpp"
 
@@ -141,6 +143,11 @@ DPDKReaderModule::do_configure(const data_t& /*args*/)
   // Construct the pcie devices allowed mask
   std::string first_pcie_addr;
   bool is_first_pcie_addr = true;
+  std::vector<uint16_t> rte_cores;
+
+  // FIXME: Hardcoding core 0 for GARP. Replace with better param.
+  rte_cores.push_back(0);
+
   std::vector<const confmodel::DetectorToDaqConnection*> d2d_conns;
   for (auto res : res_set) {
     auto connection = res->cast<confmodel::DetectorToDaqConnection>();
@@ -172,8 +179,15 @@ DPDKReaderModule::do_configure(const data_t& /*args*/)
     }
     eal_params.push_back("-a");
     eal_params.push_back(net_device->get_pcie_addr());
+
+    for ( const auto* proc_res : receiver->get_configuration()->get_used_lcores() ) {
+      rte_cores.insert(rte_cores.end(), proc_res->get_cpu_cores().begin(), proc_res->get_cpu_cores().end());
+    }
   }
 
+  eal_params.push_back("-l");
+  eal_params.push_back(fmt::format("{}", fmt::join(rte_cores,",")));
+
 
   // Use the first pcie device id as file prefix
   // FIXME: Review this strategy - should work in most of cases, but it could be 

scripts/udp_receiver.py → scripts/dpdklibs_udp_receiver.py

@@ -9,6 +9,7 @@
 N_STREAM = 128
 FRAME_SIZE = 7200
 FRAME_TS_GAP = 2048
+FRAME_TS_GAP = 2000
 
 def print_header(wib_frame,prefix="\t"):
     header = wib_frame.get_daqheader()
@@ -35,20 +36,34 @@ def dump_data(data):
 @click.command()
 @click.option('-d', '--dump', is_flag=True, default=False)
 @click.option('-c', '--count', type=int, default=None)
-def main(dump, count):
+@click.option('-p', '--port', type=int, default=None)
+@click.option('-g', '--gap', type=int, default=None)
+@click.option('-f', '--frame-type', type=click.Choice(['wib', 'tde']), default='wib')
+def main(dump, count, port, gap, frame_type):
     s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM, 0)
 
-    s.bind(('0.0.0.0', 0x4444))
+    s.bind(('0.0.0.0', port))
+
     prev_stream = {}
     i=0
     dtstart = time.time()
     dtlast = dtstart
     sampling = 100000
+    match frame_type:
+        case 'wib':
+            frame_class = fddetdataformats.WIBEthFrame
+            port = port if not port is None else 0x4444
+            gap = gap if not gap is None else 2048
+        case 'tde':
+            frame_class = fddetdataformats.TDEEthFrame
+            port = port if not port is None else 54323
+            gap = gap if not gap is None else 2000
+
     print('Starting receiver')
     while (count==None or i<count):
     # while i<10:
         data, address = s.recvfrom(20000)
-        wf = fddetdataformats.WIBEthFrame(data)
+        wf = frame_class(data)
         header = wf.get_daqheader()
 
 

src/IfaceWrapper.cpp

@@ -93,9 +93,8 @@ IfaceWrapper::IfaceWrapper(
 
 
   // Here is my list of cores
-  std::vector<uint16_t> rte_cores;
-  for( auto proc_res : iface_cfg->get_used_lcores()) {
-    rte_cores.insert(rte_cores.end(), proc_res->get_cpu_cores().begin(), proc_res->get_cpu_cores().end());
+  for( const auto* proc_res : iface_cfg->get_used_lcores()) {
+    m_rte_cores.insert(m_rte_cores.end(), proc_res->get_cpu_cores().begin(), proc_res->get_cpu_cores().end());
   }
 
   // iterate through active streams
@@ -125,11 +124,11 @@ IfaceWrapper::IfaceWrapper(
     m_num_frames_rxq[rx_q] = { 0 };
     m_num_bytes_rxq[rx_q] = { 0 };
 
-    m_rx_core_map[rte_cores[core_idx]][rx_q] = tx_ip;
+    m_rx_core_map[m_rte_cores[core_idx]][rx_q] = tx_ip;
     m_stream_id_to_source_id[rx_q] = strm_src;
 
     ++rx_q;
-    if ( ++core_idx == rte_cores.size()) {
+    if ( ++core_idx == m_rte_cores.size()) {
       core_idx = 0;
     }
   }

src/IfaceWrapper.hpp

@@ -63,10 +63,12 @@ class IfaceWrapper
   void setup_interface();
   void setup_flow_steering();
   void setup_xstats();
-
+  
   void enable_flow() { m_lcore_enable_flow.store(true);}
   void disable_flow() { m_lcore_enable_flow.store(false);}
 
+  const std::vector<uint16_t>& get_rte_cores() const { return m_rte_cores; }
+
 protected:
   //iface_conf_t m_cfg;
   int m_iface_id;
@@ -93,6 +95,7 @@ class IfaceWrapper
   std::set<std::string> m_ips;
   std::set<int> m_rx_qs;
   std::set<int> m_tx_qs;
+  std::vector<uint16_t> m_rte_cores;
 
   // CPU core ID -> [queue -> ip]
   std::map<int, std::map<int, std::string>> m_rx_core_map;