Protect the muon decoding from unexpected data

Currently the decoding throws exceptions when a raw bank doesn't look as expected. These are replaced with error counters, and processing of the rest of the raw banks continues.

In addition, some assumptions (that seem to be violated with the current data) are now checked, preventing SIGABRT (invalid subspan arguments).

Adresses #242 but there are most certainly more issues as infrequent SIGABRT happens still in the MuonMon task.

/cc @samarian @mstahl

Muon/MuonDAQ/src/components/MuonRawInUpgradeToHits.cpp

@@ -51,45 +51,10 @@ namespace LHCb::MuonUpgrade::DAQ {
     ByteType single_bit_position[8] = {0x80, 0x40, 0x20, 0x10, 0x08, 0x04, 0x02, 0x01};
   }
 
-  namespace EC::RawToHits {
-
-    enum class ErrorCode : StatusCode::code_t {
-      BAD_MAGIC = 10,
-      BANK_TOO_SHORT,
-      PADDING_TOO_LONG,
-      TOO_MANY_HITS,
-      INVALID_TELL1
-    };
-    struct ErrorCategory : StatusCode::Category {
-      const char* name() const override { return "MuonRawBankDecoding"; }
-      bool        isRecoverable( StatusCode::code_t ) const override { return false; }
-      std::string message( StatusCode::code_t code ) const override {
-        switch ( static_cast<ErrorCode>( code ) ) {
-        case ErrorCode::BAD_MAGIC:
-          return "Incorrect Magic pattern in raw bank";
-        case ErrorCode::BANK_TOO_SHORT:
-          return "Muon bank is too short";
-        case ErrorCode::TOO_MANY_HITS:
-          return "Muon bank has too many hits for its size";
-        default:
-          return StatusCode::default_category().message( code );
-        }
-      }
-    };
-  } // namespace EC::RawToHits
 } // namespace LHCb::MuonUpgrade::DAQ
 
-STATUSCODE_ENUM_DECL( LHCb::MuonUpgrade::DAQ::EC::RawToHits::ErrorCode )
-STATUSCODE_ENUM_IMPL( LHCb::MuonUpgrade::DAQ::EC::RawToHits::ErrorCode,
-                      LHCb::MuonUpgrade::DAQ::EC::RawToHits::ErrorCategory )
-
 namespace LHCb::MuonUpgrade::DAQ {
   namespace {
-    [[gnu::noreturn]] void throw_exception( EC::RawToHits::ErrorCode ec, const char* tag ) {
-      auto sc = StatusCode( ec );
-      throw GaudiException{sc.message(), tag, std::move( sc )};
-    }
-#define OOPS( x ) throw_exception( x, __PRETTY_FUNCTION__ )
     template <typename Iterator>
     Iterator addCoordsCrossingMap( Iterator first, Iterator last, CommonMuonHits& commonHits,
                                    const ComputeTilePosition& compute, size_t /* nStations */ ) {
@@ -206,6 +171,11 @@ namespace LHCb::MuonUpgrade::DAQ {
     mutable unsigned int                                local_wrong_hits_counters[44][16] = {};
     mutable unsigned int                                local_acquired_frame[44][16]      = {};
     mutable Gaudi::Accumulators::MsgCounter<MSG::ERROR> m_invalid_add{this, "invalid add"};
+    mutable Gaudi::Accumulators::MsgCounter<MSG::ERROR> m_bad_magic{this, "Incorrect Magic pattern in raw bank"};
+    mutable Gaudi::Accumulators::MsgCounter<MSG::ERROR> m_bank_too_short{this, "Muon bank is too short"};
+    mutable Gaudi::Accumulators::MsgCounter<MSG::ERROR> m_too_many_hits{this,
+                                                                        "Muon bank has too many hits for its size"};
+    mutable Gaudi::Accumulators::MsgCounter<MSG::ERROR> m_bad_size_of_link{this, "Size of link is bad (<= 6)"};
     mutable Gaudi::Accumulators::Counter<>              m_Tell40Size[44]{
         {this, "Tell_001_0 size"}, {this, "Tell_001_1 size"}, {this, "Tell_002_0 size"}, {this, "Tell_002_1 size"},
         {this, "Tell_003_0 size"}, {this, "Tell_003_1 size"}, {this, "Tell_004_0 size"}, {this, "Tell_004_1 size"},
@@ -335,7 +305,10 @@ namespace LHCb::MuonUpgrade::DAQ {
     // std::vector<LHCb::span<ByteType>> lista_link;
     // lista_link.reserve( 64 );
     for ( const auto& raw_bank : muon_banks ) {
-      if ( RawBank::MagicPattern != raw_bank->magic() ) { OOPS( EC::RawToHits::ErrorCode::BAD_MAGIC ); }
+      if ( RawBank::MagicPattern != raw_bank->magic() ) {
+        ++m_bad_magic;
+        continue;
+      }
 
       // the Tell40 number is hidden in source ID lower part
       unsigned int tell40PCI       = raw_bank->sourceID() & 0x00FF;
@@ -350,7 +323,10 @@ namespace LHCb::MuonUpgrade::DAQ {
         debug() << " reading tell40 num " << TNumber << " PCI n " << PCINumber << " "
                 << " bank size" << (unsigned int)raw_bank->size() << endmsg;
 
-      if ( (unsigned int)raw_bank->size() < 1 ) { OOPS( EC::RawToHits::ErrorCode::BANK_TOO_SHORT ); }
+      if ( (unsigned int)raw_bank->size() < 1 ) {
+        ++m_bank_too_short;
+        continue;
+      }
       auto         range8             = raw_bank->range<ByteType>();
       auto         range_data         = range8.subspan( 1 );
       unsigned int link_start_pointer = 0;
@@ -367,7 +343,10 @@ namespace LHCb::MuonUpgrade::DAQ {
         debug() << " Synch event " << synch_evt << " aligned fiber info stored " << align_info << endmsg;
       }
       if ( synch_evt ) continue;
-      if ( (unsigned int)raw_bank->size() < 1 + 3 * align_info ) { OOPS( EC::RawToHits::ErrorCode::BANK_TOO_SHORT ); }
+      if ( (unsigned int)raw_bank->size() < 1 + 3 * align_info ) {
+        ++m_bank_too_short;
+        continue;
+      }
       unsigned int number_of_readout_fibers = 0;
       unsigned int map_connected_fibers[24] = {};
 
@@ -376,7 +355,10 @@ namespace LHCb::MuonUpgrade::DAQ {
       if ( !align_info ) {
         for ( unsigned int i = 0; i < active_links; i++ ) { map_connected_fibers[i] = i; }
         number_of_readout_fibers = active_links;
-        if ( (unsigned int)raw_bank->size() < active_links + 1 ) { OOPS( EC::RawToHits::ErrorCode::BANK_TOO_SHORT ); }
+        if ( (unsigned int)raw_bank->size() < active_links + 1 ) {
+          ++m_bank_too_short;
+          continue;
+        }
 
       } else {
         link_start_pointer = link_start_pointer + 3;
@@ -418,8 +400,10 @@ namespace LHCb::MuonUpgrade::DAQ {
           for ( int i = 0; i < 24; i++ ) { verbose() << i << " " << align_vector[i] << endmsg; }
         }
         number_of_readout_fibers = readout_fibers;
-        if ( (unsigned int)raw_bank->size() < number_of_readout_fibers + 1 + 3 )
-          OOPS( EC::RawToHits::ErrorCode::BANK_TOO_SHORT );
+        if ( (unsigned int)raw_bank->size() < number_of_readout_fibers + 1 + 3 ) {
+          ++m_bank_too_short;
+          continue;
+        }
       }
       if ( (unsigned int)raw_bank->size() > number_of_readout_fibers + 1 + 3 * align_info ) {
         if ( msgLevel( MSG::VERBOSE ) ) {
@@ -460,7 +444,25 @@ namespace LHCb::MuonUpgrade::DAQ {
         if ( size_of_link > 1 ) {
           // get two subpart (there is an overlap due to the 4 bits shift
           auto range_link_HitsMap = range_data.subspan( link_start_pointer, 7 );
-          auto range_link_TDC     = range_data.subspan( link_start_pointer + 6, size_of_link - 6 );
+          if ( size_of_link <= 6 ) {
+            // Example from data: link = 0, size_of_link = 4, link_start_pointer = 0, raw_bank->m_length = 53
+            // Raw bank data:
+            // {m_magic = 52171, m_length = 53, m_type = 13 '\r', m_version = 3 '\003', m_sourceID = 26628, m_data =
+            // {2420323592}}
+            //
+            // all words: [3525579, 1745093389, 2420323592, 35, 4255762696, 224, 4278255661, 0,
+            // 1879048192, 0, 2154116872, 34, 4287088648, 0]
+            //
+            // in hex:
+            //   cb cb 35 00 0d 03 04 68  <- MDF header
+            // payload:
+            //   08 35 43 90 23 00 00 00 08 c9 a9 fd e0 00 00 00
+            //   2d 00 01 ff 00 00 00 00 00 00 00 70 00 00 00 00
+            //   08 37 65 80 22 00 00 00 08 c8 87 ff 00 00 00 00
+            ++m_bad_size_of_link;
+            break;
+          }
+          auto range_link_TDC = range_data.subspan( link_start_pointer + 6, size_of_link - 6 );
 
           bool         first_hitmap_byte  = false;
           bool         last_hitmap_byte   = true;