stop run if first RDH is wrong

Author: sy-c

doc/releaseNotes.md

@@ -531,7 +531,7 @@ This file describes the main feature changes for each readout.exe released versi
 - Disabled unused RDMA features (still available by switch in CMake)
 - Added feature for memory banks real time monitoring. Enabled by setting membanksMonitorRate in /etc/o2.d/readout-defaults.cfg. Output (high-rate text) can be seen locally in real time with `tail -f /tmp/readout-monitor-mempool-(id)`. The status of each page in the bank is displayed.
 
-## v2.17.2 - 08/03/2023
+s## v2.17.2 - 08/03/2023
 - Updated naming of files saved to disk when equipment-*.saveErrorPagesMax is set. Now includes timestamp, equipment id, and file counter: e.g. /tmp/readout-t1678288819-eq1-superpage.1.raw
 
 ## v2.18.0 - 21/03/2023
@@ -606,3 +606,6 @@ This file describes the main feature changes for each readout.exe released versi
 
 ## v2.21.8 - 29/11/2023
 - Fixed spurious readout warning("TF non-contiguous ordering") when doing Start/Stop/Start with DD is enabled.
+
+## next version
+- When equipment-*-rdhCheckFirstOrbit is set (default), the run is stopped if the first RDH received for an equipment does not passes validation, as in the case when first orbits mismatch.

src/ReadoutEquipment.cxx

@@ -438,6 +438,17 @@ Thread::CallbackResult ReadoutEquipment::threadCallback(void* arg)
 	}
         updatePageStateFromDataBlockContainerReference(nextBlock, MemoryPage::PageState::InEquipment);
 
+        // here some test code to artifically corrupt RDH of block X of equipment id Y
+	// used to test error conditions, eg first RDH received is wrong
+        /*
+	int X=1;
+	int Y=33;
+        if ((ptr->currentBlockId==X)&&(ptr->id==Y)) {
+	  o2::Header::RAWDataHeader* h = (o2::Header::RAWDataHeader*) nextBlock->getData()->data;
+	  h->version=99;
+	}
+        */
+
 	// handle RDH-formatted data
 	if (ptr->cfgRdhUseFirstInPageEnabled) {
           if ((ptr->processRdh(nextBlock)) && ptr->cfgDropPagesWithError) {
@@ -852,6 +863,12 @@ int ReadoutEquipment::processRdh(DataBlockContainerReference& block)
 
     if (tagDatablockFromRdh(h, blockHeader) == 0) {
       blockHeader.isRdhFormat = 1;
+    } else {
+      if ((cfgRdhCheckFirstOrbit) && (currentBlockId == 0)) {
+        // if 1st RDH received is wrong, it's the same as a wrong orbit
+        theLog.log(LogErrorSupport_(3241), "Equipment %s : first RDH is wrong", name.c_str());
+        isFatalError++;
+      }
     }
 
     if (cfgRdhDumpFirstInPageEnabled) {