Do not fail on Resource Occupied error indirect scom

There is a race in the indirect scom status reading where
we can catch a 001 pib status that is not a real error.
Note - Fix already added in P8 but forwardport got missed

Change-Id: Ida6fe97fb9a0281640c6cc089482dada5dad1c33
CQ: SW384238
Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/38063
Tested-by: Jenkins Server <pfd-jenkins+hostboot@us.ibm.com>
Tested-by: Jenkins OP Build CI <op-jenkins+hostboot@us.ibm.com>
Tested-by: FSP CI Jenkins <fsp-CI-jenkins+hostboot@us.ibm.com>
Reviewed-by: Thi N. Tran <thi@us.ibm.com>
Reviewed-by: Christian R. Geddes <crgeddes@us.ibm.com>
Reviewed-by: Matthew A. Ploetz <maploetz@us.ibm.com>

1 files changed, 38 insertions, 11 deletions

diff --git a/src/usr/scom/scom.C b/src/usr/scom/scom.C
index b52c33ba0..cca8b45d6 100644
--- a/src/usr/scom/scom.C
+++ b/src/usr/scom/scom.C
@@ -359,11 +359,10 @@ errlHndl_t doForm0IndirectScom(DeviceFW::OperationType i_opType,
                 //  or we saw an error, then we're done
                 if (scomout.done || scomout.piberr)
                 {
-                    // we should never see this error code so we are most
-                    //  likely going to fail, but since the hardware team
-                    //  cannot explain why we get this we're going to
-                    //  poll for awhile just in case it could work with
-                    //  a retry
+                    // there is a small chance for a race if we check the
+                    //  status very quickly, the hardware sets the status
+                    //  to 001=Resource Occupied when the command first
+                    //  starts so keep polling
                     if( scomout.piberr != PIB::PIB_RESOURCE_OCCUPIED )
                     {
                         break;
@@ -426,6 +425,8 @@ errlHndl_t doForm0IndirectScom(DeviceFW::OperationType i_opType,
                 //Add this target to the FFDC
                 ERRORLOG::ErrlUserDetailsTarget(i_target,"IndSCOM Target")
                   .addToLog(l_err);
+
+                l_err->collectTrace( SCOM_COMP_NAME, 256);
             }
             // if we got a timeout, create an errorlog.
             else if( scomout.done == 0 )
@@ -459,6 +460,8 @@ errlHndl_t doForm0IndirectScom(DeviceFW::OperationType i_opType,
                 //Add this target to the FFDC
                 ERRORLOG::ErrlUserDetailsTarget(i_target,"IndSCOM Target")
                   .addToLog(l_err);
+
+                l_err->collectTrace( SCOM_COMP_NAME, 256);
             }
             else // It worked
             {
@@ -508,7 +511,14 @@ errlHndl_t doForm0IndirectScom(DeviceFW::OperationType i_opType,
                 //  or we saw an error, then we're done
                 if (scomout.done || scomout.piberr)
                 {
-                    break;
+                    // there is a small chance for a race if we check the
+                    //  status very quickly, the hardware sets the status
+                    //  to 001=Resource Occupied when the command first
+                    //  starts so keep polling
+                    if( scomout.piberr != PIB::PIB_RESOURCE_OCCUPIED )
+                    {
+                        break;
+                    }
                 }
 
                 nanosleep( 0, 10000 ); //sleep for 10,000 ns
@@ -541,15 +551,30 @@ errlHndl_t doForm0IndirectScom(DeviceFW::OperationType i_opType,
                                       i_addr,
                                       scomout.data64);
 
-                //Add the callouts for the specific PCB/PIB error
-                PIB::addFruCallouts( i_target,
-                                     scomout.piberr,
-                                     i_addr,
-                                     l_err );
+                // we should never hit this so if we do we are going
+                //  to blame hardware
+                if( scomout.piberr == PIB::PIB_RESOURCE_OCCUPIED )
+                {
+                    SCOM::UdPibInfo(scomout.piberr).addToLog(l_err);
+                    l_err->addHwCallout( i_target,
+                                         HWAS::SRCI_PRIORITY_HIGH,
+                                         HWAS::NO_DECONFIG,
+                                         HWAS::GARD_NULL );
+                }
+                else
+                {
+                    //Add the callouts for the specific PCB/PIB error
+                    PIB::addFruCallouts( i_target,
+                                         scomout.piberr,
+                                         i_addr,
+                                         l_err );
+                }
 
                 //Add this target to the FFDC
                 ERRORLOG::ErrlUserDetailsTarget(i_target,"IndSCOM Target")
                   .addToLog(l_err);
+
+                l_err->collectTrace( SCOM_COMP_NAME, 256);
             }
             // if we got a timeout, create an errorlog.
             else if( scomout.done == 0 )
@@ -583,6 +608,8 @@ errlHndl_t doForm0IndirectScom(DeviceFW::OperationType i_opType,
                 //Add this target to the FFDC
                 ERRORLOG::ErrlUserDetailsTarget(i_target,"IndSCOM Target")
                   .addToLog(l_err);
+
+                l_err->collectTrace( SCOM_COMP_NAME, 256);
             }
         } // end of write
     } while(0);