Need to consider Portable-AOT cases during code patching

Portable-AOT makes some code sequences different from the
expectation of the current CPU type. When that sequence is
patched,  different possible scenarios have to be considered.
Otherwise, doing only according to the current CPU type may
run into crashes.

Signed-off-by: Julian <zlwang@ca.ibm.com>

runtime/compiler/runtime/Trampoline.cpp

@@ -1,5 +1,5 @@
 /*******************************************************************************
- * Copyright (c) 2000, 2020 IBM Corp. and others
+ * Copyright (c) 2000, 2021 IBM Corp. and others
  *
  * This program and the accompanying materials are made available under
  * the terms of the Eclipse Public License 2.0 which accompanies this
@@ -480,10 +480,13 @@ bool ppcCodePatching(void *method, void *callSite, void *currentPC, void *curren
 
          oldBits = *(int32_t *)((uint8_t *)callSite - encodingStartOffset);     // The load of the first IPIC cache slot or rldimi
 #if defined(TR_TARGET_64BIT)
-         if (TR::Compiler->target.cpu.isAtLeast(OMR_PROCESSOR_PPC_P10))
+
+         distance = *(int32_t *)((uint8_t *)callSite - encodingStartOffset - 4);
+         // Testing if it is really a paddi case
+         if (TR::Compiler->target.cpu.isAtLeast(OMR_PROCESSOR_PPC_P10) &&
+             (((distance >> 20) & 0x00000FFF) == 0x061))
             {
             // oldBits is the latter half of paddi
-            distance = *(int32_t *)((uint8_t *)callSite - encodingStartOffset - 4);
             distance = (distance & 0x0003FFFF) << 16;  // Getting the top-18-bits and shifted into place
             distance |= oldBits & 0x0000FFFF;          // Concatenate with the bottom-16-bits
             distance = (distance << 30) >> 30;         // sign-extend