Switch ordering of function call optimizations

Moving up PruneUnusedArgumentPass right after function cloning, then doing InstructionCombining, can potentially remove alloca pointers passed as function arguments, but is used.
This frees up memory on the stack frame since those allocas no longer need to be stored on private memory.

Author: dlei6g

IGC/Compiler/CISACodeGen/ShaderCodeGen.cpp

@@ -685,6 +685,30 @@ static void AddLegalizationPasses(CodeGenContext& ctx, IGCPassManager& mpm, PSSi
         mpm.add(llvm::createGlobalDCEPass());
         mpm.add(new PurgeMetaDataUtils());
         mpm.add(createGenXCodeGenModulePass());
+
+        // Light cleanup for subroutines after cloning. Note that the constant
+        // propogation order is reversed, compared to the opt sequence in
+        // OptimizeIR. There is a substantial gain with CFG simplification after
+        // interprocedural constant propagation.
+        if (!isOptDisabled)
+        {
+            mpm.add(createPruneUnusedArgumentsPass());
+#if LLVM_VERSION_MAJOR >= 12
+            mpm.add(createIPSCCPPass());
+#else
+            if (!ctx.m_hasStackCalls)
+            {
+                // Don't run IPConstantProp when stackcalls are present.
+                // Let global constants be relocated inside stack funcs.
+                // We cannot process SLM constants inside stackcalls, so don't propagate them.
+                mpm.add(createIPConstantPropagationPass());
+            }
+            mpm.add(createConstantPropagationPass());
+#endif
+            mpm.add(createDeadCodeEliminationPass());
+            mpm.add(createCFGSimplificationPass());
+            mpm.add(createIGCInstructionCombiningPass());
+        }
     }
 
     // Remove all uses of implicit arg instrinsics after inlining by lowering them to kernel args
@@ -793,30 +817,6 @@ static void AddLegalizationPasses(CodeGenContext& ctx, IGCPassManager& mpm, PSSi
         mpm.add(new StatelessToStateful(hasBufOff));
     }
 
-    // Light cleanup for subroutines after cloning. Note that the constant
-    // propogation order is reversed, compared to the opt sequence in
-    // OptimizeIR. There is a substantial gain with CFG simplification after
-    // interprocedural constant propagation.
-    if (ctx.m_enableSubroutine && !isOptDisabled)
-    {
-        mpm.add(createPruneUnusedArgumentsPass());
-
-#if LLVM_VERSION_MAJOR >= 12
-        mpm.add(createIPSCCPPass());
-#else
-        if (!ctx.m_hasStackCalls)
-        {
-            // Don't run IPConstantProp when stackcalls are present.
-            // Let global constants be relocated inside stack funcs.
-            // We cannot process SLM constants inside stackcalls, so don't propagate them.
-            mpm.add(createIPConstantPropagationPass());
-        }
-        mpm.add(createConstantPropagationPass());
-#endif
-
-        mpm.add(createDeadCodeEliminationPass());
-        mpm.add(createCFGSimplificationPass());
-    }
     // Since we don't support switch statements, switch lowering is needed after the last CFG simplication
     mpm.add(llvm::createLowerSwitchPass());
 

IGC/Compiler/Optimizer/OpenCLPasses/ProgramScopeConstants/ProgramScopeConstantAnalysis.cpp

@@ -415,7 +415,6 @@ void ProgramScopeConstantAnalysis::addData(Constant* initializer,
             const unsigned pointedToAddrSpace = WalkCastsToFindNamedAddrSpace(initializer);
 
             IGC_ASSERT(addressSpace == ADDRESS_SPACE_GLOBAL || addressSpace == ADDRESS_SPACE_CONSTANT);
-            IGC_ASSERT_MESSAGE(pointerSize == 8, "Can global var pointer ever be 32bits?");
 
             // We can only patch global and constant pointers.
             if (pointedToAddrSpace == ADDRESS_SPACE_GLOBAL || pointedToAddrSpace == ADDRESS_SPACE_CONSTANT)