added function that tests if an instruction waits for VA_SDST=0

Author: mihajlovicana

llvm/lib/Target/AMDGPU/AMDGPUInsertDelayAlu.cpp

@@ -14,9 +14,13 @@
 #include "AMDGPU.h"
 #include "GCNSubtarget.h"
 #include "MCTargetDesc/AMDGPUMCTargetDesc.h"
+#include "SIDefines.h"
 #include "SIInstrInfo.h"
+#include "Utils/AMDGPUBaseInfo.h"
 #include "llvm/ADT/SetVector.h"
+#include "llvm/CodeGen/MachineInstr.h"
 #include "llvm/MC/MCRegister.h"
+#include "llvm/Support/ErrorHandling.h"
 
 using namespace llvm;
 
@@ -57,6 +61,19 @@ class AMDGPUInsertDelayAlu : public MachineFunctionPass {
     return false;
   }
 
+  static bool instructionWaitsForSALUWrites(const MachineInstr &MI) {
+    // These instruction types wait for VA_SDST==0 before issuing.
+    // S_CBRANCH_EXECZ and S_CBRANCH_VCCZ are covered by SALU flag
+    const uint64_t VA_SDST_0 = SIInstrFlags::SALU | SIInstrFlags::EXP |
+                               SIInstrFlags::DS | SIInstrFlags::SMRD |
+                               SIInstrFlags::MIMG | SIInstrFlags::VIMAGE |
+                               SIInstrFlags::VSAMPLE;
+
+    if (MI.getDesc().TSFlags & VA_SDST_0)
+      return true;
+    return false;
+  }
+
   // Types of delay that can be encoded in an s_delay_alu instruction.
   enum DelayType { VALU, TRANS, SALU, OTHER };
 
@@ -365,7 +382,7 @@ class AMDGPUInsertDelayAlu : public MachineFunctionPass {
 
       DelayType Type = getDelayType(MI.getDesc().TSFlags);
 
-      if (SII->isSALU(MI.getOpcode())) {
+      if (instructionWaitsForSALUWrites(MI)) {
         auto It = State.find(lastSGPRfromVALU);
         if (It != State.end()) {
           DelayInfo Info = It->getSecond();