Fix nasa#411, rework exception handling in CFE

Move exception handling to a PSP function.  In this approach
the CFE only logs data after the event as a background job.

Replaces the CFE_ES_ProcessCoreException with a simple notification
that causes the ES background job to run, which in turn polls the
PSP for logged exceptions and writes entries to the ES ER log.

Both the PSP execption scan and the ER log file dump are converted
to background jobs.

cmake/sample_defs/cpu1_platform_cfg.h

@@ -565,7 +565,7 @@
 **       in the error log. Any context information beyond this size will
 **       be truncated.
 */
-#define CFE_PLATFORM_ES_ER_LOG_MAX_CONTEXT_SIZE     128
+#define CFE_PLATFORM_ES_ER_LOG_MAX_CONTEXT_SIZE     256
 
 
 /**

cmake/target/inc/target_config.h

@@ -51,10 +51,15 @@ typedef void (*System_1HzISRFunc_t)(void);
 
 
 /**
- * Prototype for exception ISR function implemented in CFE ES
+ * Prototype for notification function implemented in CFE ES
  * The PSP should call this when exceptions occur.
+ *
+ * NOTE: the PSP must call this routine only from a context where
+ * it is possible to use OSAL primitives.  This means it must _not_
+ * be called from an ISR/signal context on systems where these are
+ * restricted.
  */
-typedef void (*System_ExceptionFunc_t)(uint32  HostTaskId, const char *ReasonString, const uint32 *ContextPointer, uint32 ContextSize);
+typedef void (*System_NotifyFunc_t)(void);
 
 /**
  * Abstract pointer to a module API
@@ -87,9 +92,9 @@ typedef const struct
    System_MainFunc_t SystemMain;
 
    /**
-    * Exception handler function. Called from PSP during exception handling.
+    * Notification function. Called from PSP after async event handling.
     */
-   System_ExceptionFunc_t SystemExceptionISR;
+   System_NotifyFunc_t SystemNotify;
 
    /*
     * Sizes of memory segments required by the CFE based on the current config

cmake/target/src/target_config.c

@@ -37,6 +37,7 @@
 #include "cfe_platform_cfg.h"
 #include "cfe_es.h"
 #include "cfe_time.h"
+#include "private/cfe_es_resetdata_typedef.h"
 #include "cfecfs_version_info.h"
 #include "cfecfs_build_info.h"
 
@@ -61,7 +62,7 @@ Target_CfeConfigData GLOBAL_CFE_CONFIGDATA =
        */
       .System1HzISR = CFE_TIME_Local1HzISR,
       .SystemMain = CFE_ES_Main,
-      .SystemExceptionISR = CFE_ES_ProcessCoreException,
+      .SystemNotify = CFE_ES_ProcessAsyncEvent,
 
       /*
        * Default values for Startup file.
@@ -73,7 +74,7 @@ Target_CfeConfigData GLOBAL_CFE_CONFIGDATA =
        * Sizes of other memory segments
        */
       .CdsSize = CFE_PLATFORM_ES_CDS_SIZE,
-      .ResetAreaSize = CFE_PLATFORM_ES_RESET_AREA_SIZE,
+      .ResetAreaSize = sizeof(CFE_ES_ResetData_t),
       .UserReservedSize = CFE_PLATFORM_ES_USER_RESERVED_SIZE,
 
       .RamDiskSectorSize = CFE_PLATFORM_ES_RAM_DISK_SECTOR_SIZE,

fsw/cfe-core/src/es/cfe_es_api.c

@@ -43,6 +43,7 @@
 #include "cfe_es_events.h"
 #include "cfe_es_cds.h"
 #include "cfe_es_cds_mempool.h"
+#include "cfe_es_task.h"
 #include "cfe_psp.h"
 #include "cfe_es_log.h"
 
@@ -96,7 +97,7 @@ int32 CFE_ES_ResetCFE(uint32 ResetType)
            */
            CFE_ES_WriteToERLog(CFE_ES_LogEntryType_CORE, CFE_PSP_RST_TYPE_POWERON,
                                          CFE_PSP_RST_SUBTYPE_RESET_COMMAND,
-                                         "POWER ON RESET due to max proc resets (Commanded).", NULL,0 );
+                                         "POWER ON RESET due to max proc resets (Commanded).");
            /*
            ** Call the BSP reset routine
            */
@@ -116,7 +117,7 @@ int32 CFE_ES_ResetCFE(uint32 ResetType)
            */
            CFE_ES_WriteToERLog(CFE_ES_LogEntryType_CORE, CFE_PSP_RST_TYPE_PROCESSOR,
                                        CFE_PSP_RST_SUBTYPE_RESET_COMMAND,
-                                       "PROCESSOR RESET called from CFE_ES_ResetCFE (Commanded).", NULL,0 );
+                                       "PROCESSOR RESET called from CFE_ES_ResetCFE (Commanded).");
            /*
            ** Call the BSP reset routine
            */
@@ -140,7 +141,7 @@ int32 CFE_ES_ResetCFE(uint32 ResetType)
        */
        CFE_ES_WriteToERLog(CFE_ES_LogEntryType_CORE, CFE_PSP_RST_TYPE_POWERON,
                                        CFE_PSP_RST_SUBTYPE_RESET_COMMAND,
-                                       "POWERON RESET called from CFE_ES_ResetCFE (Commanded).", NULL,0 );
+                                       "POWERON RESET called from CFE_ES_ResetCFE (Commanded).");
 
        /*
        ** Call the BSP reset routine
@@ -1716,126 +1717,13 @@ void CFE_ES_UnlockSharedData(const char *FunctionName, int32 LineNumber)
 }/* end CFE_ES_UnlockSharedData */
 
 /******************************************************************************
-**  Function:  CFE_ES_ProcessCoreException() - See API and header file for details
+**  Function:  CFE_ES_ProcessAsyncEvent()
+**
+**  Purpose:
+**    Called by the PSP to notify CFE ES that an asynchronous event occurred.
 */
-void CFE_ES_ProcessCoreException(uint32  HostTaskId,     const char *ReasonString,
-                                 const uint32 *ContextPointer, uint32 ContextSize)
+void CFE_ES_ProcessAsyncEvent(void)
 {
-    uint32                 i;
-    int32                  Status;
-    OS_task_prop_t         TaskProp;
-    CFE_ES_TaskInfo_t      EsTaskInfo;
-    uint32                 FoundExceptionTask = 0;
-    uint32                 ExceptionTaskID = 0;
-
-    /*
-    ** If a loadable cFE Application caused the reset and it's
-    ** exception action is set to Restart the App rather than cause a
-    ** processor reset, then just reset the App.
-    */
-
-    /*
-    ** We have the Host Task Id ( vxWorks, RTEMS, etc ). Search
-    ** the OSAPI to see if a match can be found.
-    */
-    for ( i = 0; i < OS_MAX_TASKS; i++ )
-    {
-       if (CFE_ES_Global.TaskTable[i].RecordUsed == true)
-       {
-          ExceptionTaskID = CFE_ES_Global.TaskTable[i].TaskId;
-          Status = OS_TaskGetInfo (ExceptionTaskID, &TaskProp);
-
-          if ( Status == OS_SUCCESS && TaskProp.OStask_id == HostTaskId )
-          {
-             FoundExceptionTask = 1;
-             break;
-          }
-       }
-    }
-
-    /*
-    ** If the Task is found in the OS, see if the cFE App ID associated with it can be found.
-    */
-    if ( FoundExceptionTask == 1 )
-    {
-       Status = CFE_ES_GetTaskInfo( &EsTaskInfo, ExceptionTaskID );
-       /*
-       ** The App ID was found, now see if the ExceptionAction is set for a reset
-       */
-       if ( Status == CFE_SUCCESS )
-       {
-          if ( CFE_ES_Global.AppTable[EsTaskInfo.AppId].StartParams.ExceptionAction == CFE_ES_ExceptionAction_RESTART_APP )
-          {
-
-             /*
-             ** Log the Application reset
-             */
-             CFE_ES_WriteToERLog(CFE_ES_LogEntryType_CORE, CFE_ES_APP_RESTART,
-                            CFE_PSP_RST_SUBTYPE_EXCEPTION, (char *)ReasonString,
-                            ContextPointer, ContextSize );
-
-             /*
-             ** Finally restart the App! This call is just a request
-             ** to ES.
-             */
-             CFE_ES_RestartApp(EsTaskInfo.AppId );
-
-             /*
-             ** Return to avoid the Processor Restart Logic
-             */
-             return;
-
-          } /* end if ExceptionAction */
-
-       } /* end if */
-
-    } /* End if FoundExceptionTask */
-
-    /*
-    ** If we made it here, which means that we need to do a processor reset
-    */
-
-    /*
-    ** Before doing a Processor reset, check to see
-    ** if the maximum number has been exceeded
-    */
-    if ( CFE_ES_ResetDataPtr->ResetVars.ProcessorResetCount >=
-         CFE_ES_ResetDataPtr->ResetVars.MaxProcessorResetCount )
-    {
-        /*
-        ** Log the reset in the ER Log. The log will be wiped out, but it's good to have
-        ** the entry just in case something fails.
-        */
-        CFE_ES_WriteToERLog(CFE_ES_LogEntryType_CORE,  CFE_PSP_RST_TYPE_POWERON,
-                            CFE_PSP_RST_SUBTYPE_EXCEPTION, (char *)ReasonString,
-                            ContextPointer, ContextSize );
-
-        /*
-        ** Call the BSP reset routine to do a Poweron Reset
-        */
-        CFE_PSP_Restart(CFE_PSP_RST_TYPE_POWERON);
-
-    }
-    else /* Do a processor reset */
-    {
-        /*
-        ** Update the reset variables
-        */
-        CFE_ES_ResetDataPtr->ResetVars.ProcessorResetCount++;
-        CFE_ES_ResetDataPtr->ResetVars.ES_CausedReset = true;
-
-        /*
-        ** Log the reset in the ER Log
-        */
-        CFE_ES_WriteToERLog(CFE_ES_LogEntryType_CORE, CFE_PSP_RST_TYPE_PROCESSOR,
-                            CFE_PSP_RST_SUBTYPE_EXCEPTION, (char *)ReasonString,
-                            ContextPointer, ContextSize );
-
-        /*
-        ** Need to do a processor reset
-        */
-        CFE_PSP_Restart(CFE_PSP_RST_TYPE_PROCESSOR);
-
-    } /* end if */
-
-} /* End of CFE_ES_ProcessCoreException */
+    /* This just wakes up the background task to log/handle the event. */
+    CFE_ES_BackgroundWakeup();
+}

fsw/cfe-core/src/es/cfe_es_apps.h

@@ -198,6 +198,16 @@ int32 CFE_ES_AppDumpAllInfo(void);
 */
 bool CFE_ES_RunAppTableScan(uint32 ElapsedTime, void *Arg);
 
+/*
+** Scan for new exceptions stored in the PSP
+*/
+bool CFE_ES_RunExceptionScan(uint32 ElapsedTime, void *Arg);
+
+/*
+** Check if ER log dump request is pending
+*/
+bool CFE_ES_RunERLogDump(uint32 ElapsedTime, void *Arg);
+
 /*
 ** Perform the requested control action for an application
 */

fsw/cfe-core/src/es/cfe_es_backgroundtask.c

@@ -82,6 +82,18 @@ const CFE_ES_BackgroundJobEntry_t CFE_ES_BACKGROUND_JOB_TABLE[] =
                 .JobArg = &CFE_ES_TaskData.BackgroundPerfDumpState,
                 .ActivePeriod = CFE_PLATFORM_ES_PERF_CHILD_MS_DELAY,
                 .IdlePeriod = CFE_PLATFORM_ES_PERF_CHILD_MS_DELAY * 1000
+        },
+        {   /* Check for exceptions stored in the PSP */
+                .RunFunc = CFE_ES_RunExceptionScan,
+                .JobArg = NULL,
+                .ActivePeriod = CFE_PLATFORM_ES_APP_SCAN_RATE,
+                .IdlePeriod = CFE_PLATFORM_ES_APP_SCAN_RATE
+        },
+        {   /* Check for ER log write requests */
+                .RunFunc = CFE_ES_RunERLogDump,
+                .JobArg = &CFE_ES_TaskData.BackgroundERLogDumpState,
+                .ActivePeriod = CFE_PLATFORM_ES_APP_SCAN_RATE,
+                .IdlePeriod = CFE_PLATFORM_ES_APP_SCAN_RATE
         }
 };