[AMDGPU] Fix whole wavefront mode

[Flakebi]

Summary

We cannot move wwm over exec copies because the exec register needs an exact exec mask.

[arsenm]

llvm/lib/Target/AMDGPU/SIWholeQuadMode.cpp

526–527

Can't this whole function be replaced with SIInstrInfo::mayReadExec?

llvm/test/CodeGen/AMDGPU/wqm.mir

119–147

Should be able to remove all of this (except tracksRegLiveness: true)

[Flakebi]

Use mayReadEXEC as suggested by Matt.

[nhaehnle]

LGTM

llvm/lib/Target/AMDGPU/SIWholeQuadMode.cpp

void markInstruction(MachineInstr &MI, char Flag,
                        std::vector<WorkItem> &Worklist);
   void markInstructionUses(const MachineInstr &MI, char Flag,
                            std::vector<WorkItem> &Worklist);
   char scanInstructions(MachineFunction &MF, std::vector<WorkItem> &Worklist);
   void propagateInstruction(MachineInstr &MI, std::vector<WorkItem> &Worklist);
   void propagateBlock(MachineBasicBlock &MBB, std::vector<WorkItem> &Worklist);
   char analyzeFunction(MachineFunction &MF);
 
-  bool requiresCorrectState(const MachineInstr &MI) const;
-
   MachineBasicBlock::iterator saveSCC(MachineBasicBlock &MBB,
                                       MachineBasicBlock::iterator Before);
   MachineBasicBlock::iterator
   prepareInsertion(MachineBasicBlock &MBB, MachineBasicBlock::iterator First,
                    MachineBasicBlock::iterator Last, bool PreferLast,
                    bool SaveSCC);
   void toExact(MachineBasicBlock &MBB, MachineBasicBlock::iterator Before,
                unsigned SaveWQM, unsigned LiveMaskReg);
while (!Worklist.empty()) {
     if (WI.MI)
       propagateInstruction(*WI.MI, Worklist);
     else
       propagateBlock(*WI.MBB, Worklist);
   }
 
   return GlobalFlags;
 }
 
-/// Whether \p MI really requires the exec state computed during analysis.
-///
-/// Scalar instructions must occasionally be marked WQM for correct propagation
-/// (e.g. thread masks leading up to branches), but when it comes to actual
-/// execution, they don't care about EXEC.
-bool SIWholeQuadMode::requiresCorrectState(const MachineInstr &MI) const {
-  if (MI.isTerminator())
-    return true;
-
-  // Skip instructions that are not affected by EXEC
-  if (TII->isScalarUnit(MI))
-    return false;
-
-  // Generic instructions such as COPY will either disappear by register
-  // coalescing or be lowered to SALU or VALU instructions.
-  if (MI.isTransient()) {
-    if (MI.getNumExplicitOperands() >= 1) {
-      const MachineOperand &Op = MI.getOperand(0);
-      if (Op.isReg()) {
-        if (TRI->isSGPRReg(*MRI, Op.getReg())) {
-          // SGPR instructions are not affected by EXEC
-          return false;
-        }
-      }
-    }
-  }
-
-  return true;
-}
-
 MachineBasicBlock::iterator
 SIWholeQuadMode::saveSCC(MachineBasicBlock &MBB,
                          MachineBasicBlock::iterator Before) {
   Register SaveReg = MRI->createVirtualRegister(&AMDGPU::SReg_32_XM0RegClass);
 
   MachineInstr *Save =
       BuildMI(MBB, Before, DebugLoc(), TII->get(AMDGPU::COPY), SaveReg)
           .addReg(AMDGPU::SCC);
   MachineInstr *Restore =
for (;;) {
     if (FirstWWM == IE)
       FirstWWM = II;
 
     // First, figure out the allowed states (Needs) based on the propagated
     // flags.
     if (II != IE) {
       MachineInstr &MI = *II;
 
-      if (requiresCorrectState(MI)) {
+      if (MI.isTerminator() || TII->mayReadEXEC(*MRI, MI)) {
         auto III = Instructions.find(&MI);
         if (III != Instructions.end()) {
           if (III->second.Needs & StateWWM)
             Needs = StateWWM;
           else if (III->second.Needs & StateWQM)
             Needs = StateWQM;
           else
             Needs &= ~III->second.Disabled;

llvm/test/CodeGen/AMDGPU/atomic_optimizations_local_pointer.ll

 ; GFX7LESS-NEXT:    buffer_wbinvl1
 ; GFX7LESS-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7LESS-NEXT:    s_mov_b32 s2, -1
 ; GFX7LESS-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7LESS-NEXT:    s_endpgm
 ;
 ; GFX8-LABEL: add_i32_varying:
 ; GFX8:       ; %bb.0: ; %entry
+; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX8-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX8-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX8-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX8-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX8-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX8-NEXT:    s_not_b64 exec, exec
 ; GFX8-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX8-NEXT:    s_not_b64 exec, exec

 ; GFX8-NEXT:    s_mov_b32 s2, -1
 ; GFX8-NEXT:    s_nop 0
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX8-NEXT:    s_endpgm
 ;
 ; GFX9-LABEL: add_i32_varying:
 ; GFX9:       ; %bb.0: ; %entry
+; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX9-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX9-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX9-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX9-NEXT:    s_not_b64 exec, exec
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    s_not_b64 exec, exec

 ; GFX9-NEXT:    s_mov_b32 s2, -1
 ; GFX9-NEXT:    s_nop 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX9-NEXT:    s_endpgm
 ;
 ; GFX1064-LABEL: add_i32_varying:
 ; GFX1064:       ; %bb.0: ; %entry
+; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1064-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1064-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX1064-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX1064-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1064-NEXT:    v_mbcnt_hi_u32_b32_e64 v0, s3, v0
 ; GFX1064-NEXT:    s_not_b64 exec, exec
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1064-NEXT:    s_not_b64 exec, exec

 ; GFX1064-NEXT:    s_mov_b32 s3, 0x31016000
 ; GFX1064-NEXT:    s_nop 1
 ; GFX1064-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX1064-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX1064-NEXT:    s_endpgm
 ;
 ; GFX1032-LABEL: add_i32_varying:
 ; GFX1032:       ; %bb.0: ; %entry
+; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    ; implicit-def: $vcc_hi
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1032-NEXT:    s_or_saveexec_b32 s2, -1
-; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1032-NEXT:    s_mov_b32 exec_lo, s2
 ; GFX1032-NEXT:    v_cmp_ne_u32_e64 s2, 1, 0
 ; GFX1032-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    s_or_saveexec_b32 s4, -1

 ; GFX7LESS-NEXT:    buffer_wbinvl1
 ; GFX7LESS-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7LESS-NEXT:    s_mov_b32 s2, -1
 ; GFX7LESS-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7LESS-NEXT:    s_endpgm
 ;
 ; GFX8-LABEL: add_i32_varying_gfx1032:
 ; GFX8:       ; %bb.0: ; %entry
+; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX8-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX8-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX8-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX8-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX8-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX8-NEXT:    s_not_b64 exec, exec
 ; GFX8-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX8-NEXT:    s_not_b64 exec, exec

 ; GFX8-NEXT:    s_mov_b32 s2, -1
 ; GFX8-NEXT:    s_nop 0
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX8-NEXT:    s_endpgm
 ;
 ; GFX9-LABEL: add_i32_varying_gfx1032:
 ; GFX9:       ; %bb.0: ; %entry
+; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX9-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX9-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX9-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX9-NEXT:    s_not_b64 exec, exec
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    s_not_b64 exec, exec

 ; GFX9-NEXT:    s_mov_b32 s2, -1
 ; GFX9-NEXT:    s_nop 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX9-NEXT:    s_endpgm
 ;
 ; GFX1064-LABEL: add_i32_varying_gfx1032:
 ; GFX1064:       ; %bb.0: ; %entry
+; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1064-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1064-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX1064-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX1064-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1064-NEXT:    v_mbcnt_hi_u32_b32_e64 v0, s3, v0
 ; GFX1064-NEXT:    s_not_b64 exec, exec
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1064-NEXT:    s_not_b64 exec, exec

 ; GFX1064-NEXT:    s_mov_b32 s3, 0x31016000
 ; GFX1064-NEXT:    s_nop 1
 ; GFX1064-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX1064-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX1064-NEXT:    s_endpgm
 ;
 ; GFX1032-LABEL: add_i32_varying_gfx1032:
 ; GFX1032:       ; %bb.0: ; %entry
+; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    ; implicit-def: $vcc_hi
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1032-NEXT:    s_or_saveexec_b32 s2, -1
-; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1032-NEXT:    s_mov_b32 exec_lo, s2
 ; GFX1032-NEXT:    v_cmp_ne_u32_e64 s2, 1, 0
 ; GFX1032-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    s_or_saveexec_b32 s4, -1

 ; GFX7LESS-NEXT:    buffer_wbinvl1
 ; GFX7LESS-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7LESS-NEXT:    s_mov_b32 s2, -1
 ; GFX7LESS-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7LESS-NEXT:    s_endpgm
 ;
 ; GFX8-LABEL: add_i32_varying_gfx1064:
 ; GFX8:       ; %bb.0: ; %entry
+; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX8-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX8-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX8-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX8-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX8-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX8-NEXT:    s_not_b64 exec, exec
 ; GFX8-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX8-NEXT:    s_not_b64 exec, exec

 ; GFX8-NEXT:    s_mov_b32 s2, -1
 ; GFX8-NEXT:    s_nop 0
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX8-NEXT:    s_endpgm
 ;
 ; GFX9-LABEL: add_i32_varying_gfx1064:
 ; GFX9:       ; %bb.0: ; %entry
+; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX9-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX9-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX9-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX9-NEXT:    s_not_b64 exec, exec
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    s_not_b64 exec, exec

 ; GFX9-NEXT:    s_mov_b32 s2, -1
 ; GFX9-NEXT:    s_nop 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX9-NEXT:    s_endpgm
 ;
 ; GFX1064-LABEL: add_i32_varying_gfx1064:
 ; GFX1064:       ; %bb.0: ; %entry
+; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1064-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1064-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX1064-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX1064-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1064-NEXT:    v_mbcnt_hi_u32_b32_e64 v0, s3, v0
 ; GFX1064-NEXT:    s_not_b64 exec, exec
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1064-NEXT:    s_not_b64 exec, exec

 ; GFX1064-NEXT:    s_mov_b32 s3, 0x31016000
 ; GFX1064-NEXT:    s_nop 1
 ; GFX1064-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX1064-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX1064-NEXT:    s_endpgm
 ;
 ; GFX1032-LABEL: add_i32_varying_gfx1064:
 ; GFX1032:       ; %bb.0: ; %entry
+; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    ; implicit-def: $vcc_hi
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1032-NEXT:    s_or_saveexec_b32 s2, -1
-; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1032-NEXT:    s_mov_b32 exec_lo, s2
 ; GFX1032-NEXT:    v_cmp_ne_u32_e64 s2, 1, 0
 ; GFX1032-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    s_or_saveexec_b32 s4, -1

 ; GFX7LESS-NEXT:    buffer_wbinvl1
 ; GFX7LESS-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7LESS-NEXT:    s_mov_b32 s2, -1
 ; GFX7LESS-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7LESS-NEXT:    s_endpgm
 ;
 ; GFX8-LABEL: sub_i32_varying:
 ; GFX8:       ; %bb.0: ; %entry
+; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX8-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX8-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX8-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX8-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX8-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX8-NEXT:    s_not_b64 exec, exec
 ; GFX8-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX8-NEXT:    s_not_b64 exec, exec

 ; GFX8-NEXT:    s_mov_b32 s2, -1
 ; GFX8-NEXT:    s_nop 0
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX8-NEXT:    s_endpgm
 ;
 ; GFX9-LABEL: sub_i32_varying:
 ; GFX9:       ; %bb.0: ; %entry
+; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX9-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX9-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX9-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX9-NEXT:    s_not_b64 exec, exec
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    s_not_b64 exec, exec

 ; GFX9-NEXT:    s_mov_b32 s2, -1
 ; GFX9-NEXT:    s_nop 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX9-NEXT:    s_endpgm
 ;
 ; GFX1064-LABEL: sub_i32_varying:
 ; GFX1064:       ; %bb.0: ; %entry
+; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1064-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1064-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX1064-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX1064-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1064-NEXT:    v_mbcnt_hi_u32_b32_e64 v0, s3, v0
 ; GFX1064-NEXT:    s_not_b64 exec, exec
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1064-NEXT:    s_not_b64 exec, exec

 ; GFX1064-NEXT:    s_mov_b32 s3, 0x31016000
 ; GFX1064-NEXT:    s_nop 1
 ; GFX1064-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX1064-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX1064-NEXT:    s_endpgm
 ;
 ; GFX1032-LABEL: sub_i32_varying:
 ; GFX1032:       ; %bb.0: ; %entry
+; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    ; implicit-def: $vcc_hi
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1032-NEXT:    s_or_saveexec_b32 s2, -1
-; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1032-NEXT:    s_mov_b32 exec_lo, s2
 ; GFX1032-NEXT:    v_cmp_ne_u32_e64 s2, 1, 0
 ; GFX1032-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    s_or_saveexec_b32 s4, -1

 ; GFX7LESS-NEXT:    buffer_wbinvl1
 ; GFX7LESS-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7LESS-NEXT:    s_mov_b32 s2, -1
 ; GFX7LESS-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7LESS-NEXT:    s_endpgm
 ;
 ; GFX8-LABEL: or_i32_varying:
 ; GFX8:       ; %bb.0: ; %entry
+; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX8-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX8-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX8-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX8-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX8-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX8-NEXT:    s_not_b64 exec, exec
 ; GFX8-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX8-NEXT:    s_not_b64 exec, exec

 ; GFX8-NEXT:    s_mov_b32 s2, -1
 ; GFX8-NEXT:    s_nop 0
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX8-NEXT:    s_endpgm
 ;
 ; GFX9-LABEL: or_i32_varying:
 ; GFX9:       ; %bb.0: ; %entry
+; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX9-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX9-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX9-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX9-NEXT:    s_not_b64 exec, exec
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    s_not_b64 exec, exec

 ; GFX9-NEXT:    s_mov_b32 s2, -1
 ; GFX9-NEXT:    s_nop 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX9-NEXT:    s_endpgm
 ;
 ; GFX1064-LABEL: or_i32_varying:
 ; GFX1064:       ; %bb.0: ; %entry
+; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1064-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1064-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX1064-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX1064-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1064-NEXT:    v_mbcnt_hi_u32_b32_e64 v0, s3, v0
 ; GFX1064-NEXT:    s_not_b64 exec, exec
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1064-NEXT:    s_not_b64 exec, exec

 ; GFX1064-NEXT:    s_mov_b32 s3, 0x31016000
 ; GFX1064-NEXT:    s_nop 1
 ; GFX1064-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX1064-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX1064-NEXT:    s_endpgm
 ;
 ; GFX1032-LABEL: or_i32_varying:
 ; GFX1032:       ; %bb.0: ; %entry
+; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    ; implicit-def: $vcc_hi
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1032-NEXT:    s_or_saveexec_b32 s2, -1
-; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1032-NEXT:    s_mov_b32 exec_lo, s2
 ; GFX1032-NEXT:    v_cmp_ne_u32_e64 s2, 1, 0
 ; GFX1032-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    s_or_saveexec_b32 s4, -1

 ; GFX7LESS-NEXT:    buffer_wbinvl1
 ; GFX7LESS-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7LESS-NEXT:    s_mov_b32 s2, -1
 ; GFX7LESS-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7LESS-NEXT:    s_endpgm
 ;
 ; GFX8-LABEL: xor_i32_varying:
 ; GFX8:       ; %bb.0: ; %entry
+; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX8-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX8-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX8-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX8-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX8-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX8-NEXT:    s_not_b64 exec, exec
 ; GFX8-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX8-NEXT:    s_not_b64 exec, exec

 ; GFX8-NEXT:    s_mov_b32 s2, -1
 ; GFX8-NEXT:    s_nop 0
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX8-NEXT:    s_endpgm
 ;
 ; GFX9-LABEL: xor_i32_varying:
 ; GFX9:       ; %bb.0: ; %entry
+; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX9-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX9-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX9-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX9-NEXT:    s_not_b64 exec, exec
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    s_not_b64 exec, exec

 ; GFX9-NEXT:    s_mov_b32 s2, -1
 ; GFX9-NEXT:    s_nop 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX9-NEXT:    s_endpgm
 ;
 ; GFX1064-LABEL: xor_i32_varying:
 ; GFX1064:       ; %bb.0: ; %entry
+; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1064-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1064-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX1064-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX1064-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1064-NEXT:    v_mbcnt_hi_u32_b32_e64 v0, s3, v0
 ; GFX1064-NEXT:    s_not_b64 exec, exec
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1064-NEXT:    s_not_b64 exec, exec

 ; GFX1064-NEXT:    s_mov_b32 s3, 0x31016000
 ; GFX1064-NEXT:    s_nop 1
 ; GFX1064-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX1064-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX1064-NEXT:    s_endpgm
 ;
 ; GFX1032-LABEL: xor_i32_varying:
 ; GFX1032:       ; %bb.0: ; %entry
+; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    ; implicit-def: $vcc_hi
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1032-NEXT:    s_or_saveexec_b32 s2, -1
-; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1032-NEXT:    s_mov_b32 exec_lo, s2
 ; GFX1032-NEXT:    v_cmp_ne_u32_e64 s2, 1, 0
 ; GFX1032-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    s_or_saveexec_b32 s4, -1

 ; GFX7LESS-NEXT:    buffer_wbinvl1
 ; GFX7LESS-NEXT:    s_mov_b32 s3, 0xf000
 ; GFX7LESS-NEXT:    s_mov_b32 s2, -1
 ; GFX7LESS-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX7LESS-NEXT:    s_endpgm
 ;
 ; GFX8-LABEL: umax_i32_varying:
 ; GFX8:       ; %bb.0: ; %entry
+; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX8-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX8-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX8-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX8-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX8-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX8-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX8-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX8-NEXT:    s_not_b64 exec, exec
 ; GFX8-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX8-NEXT:    s_not_b64 exec, exec

 ; GFX8-NEXT:    s_mov_b32 s2, -1
 ; GFX8-NEXT:    s_nop 0
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX8-NEXT:    s_endpgm
 ;
 ; GFX9-LABEL: umax_i32_varying:
 ; GFX9:       ; %bb.0: ; %entry
+; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX9-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX9-NEXT:    v_mbcnt_lo_u32_b32 v0, s2, 0
 ; GFX9-NEXT:    v_mbcnt_hi_u32_b32 v0, s3, v0
 ; GFX9-NEXT:    s_not_b64 exec, exec
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    s_not_b64 exec, exec

 ; GFX9-NEXT:    s_mov_b32 s2, -1
 ; GFX9-NEXT:    s_nop 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX9-NEXT:    s_endpgm
 ;
 ; GFX1064-LABEL: umax_i32_varying:
 ; GFX1064:       ; %bb.0: ; %entry
+; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1064-NEXT:    s_or_saveexec_b64 s[2:3], -1
-; GFX1064-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1064-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1064-NEXT:    s_mov_b64 exec, s[2:3]
 ; GFX1064-NEXT:    v_cmp_ne_u32_e64 s[2:3], 1, 0
 ; GFX1064-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1064-NEXT:    v_mbcnt_hi_u32_b32_e64 v0, s3, v0
 ; GFX1064-NEXT:    s_not_b64 exec, exec
 ; GFX1064-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1064-NEXT:    s_not_b64 exec, exec

 ; GFX1064-NEXT:    s_mov_b32 s3, 0x31016000
 ; GFX1064-NEXT:    s_nop 1
 ; GFX1064-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX1064-NEXT:    buffer_store_dword v0, off, s[0:3], 0
 ; GFX1064-NEXT:    s_endpgm
 ;
 ; GFX1032-LABEL: umax_i32_varying:
 ; GFX1032:       ; %bb.0: ; %entry
+; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    ; implicit-def: $vcc_hi
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX1032-NEXT:    s_or_saveexec_b32 s2, -1
-; GFX1032-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x24
 ; GFX1032-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX1032-NEXT:    s_mov_b32 exec_lo, s2
 ; GFX1032-NEXT:    v_cmp_ne_u32_e64 s2, 1, 0
 ; GFX1032-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, s2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX1032-NEXT:    s_not_b32 exec_lo, exec_lo
 ; GFX1032-NEXT:    s_or_saveexec_b32 s4, -1

llvm/test/CodeGen/AMDGPU/wqm.mir

bb.0:
     %10:vgpr_32 = V_SET_INACTIVE_B32 %11, undef %12:sreg_32, implicit $exec
     %14:vgpr_32 = COPY %7
     %13:vgpr_32 = V_MOV_B32_dpp %14, killed %10, 323, 12, 15, 0, implicit $exec
     early-clobber %15:vgpr_32 = WWM killed %13, implicit $exec
     BUFFER_STORE_DWORD_OFFSET_exact killed %15, %6, %7, 4, 0, 0, 0, 0, 0, implicit $exec
     S_ENDPGM 0
 
 ...
+
+---
+# Ensure that wwm is not put around an EXEC copy
+#CHECK-LABEL: name: copy_exec
+#CHECK: %7:sreg_64 = COPY $exec
+#CHECK-NEXT: %14:sreg_64 = ENTER_WWM -1, implicit-def $exec, implicit-def $scc, implicit $exec
+#CHECK-NEXT: %8:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
+#CHECK-NEXT: $exec = EXIT_WWM %14
+#CHECK-NEXT: %9:vgpr_32 = V_MBCNT_LO_U32_B32_e64 %7.sub0, 0, implicit $exec
+name:            copy_exec
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3
+
+    %3:sgpr_32 = COPY $sgpr3
+    %2:sgpr_32 = COPY $sgpr2
+    %1:sgpr_32 = COPY $sgpr1
+    %0:sgpr_32 = COPY $sgpr0
+    %4:sgpr_128 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1, %2, %subreg.sub2, %3, %subreg.sub3
+    %5:sreg_32 = S_MOV_B32 0
+    %6:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %4, %5, 0, 0, 0, 0, 0, 0, implicit $exec
+
+    %8:sreg_64 = COPY $exec
+    %9:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
+    %10:vgpr_32 = V_MBCNT_LO_U32_B32_e64 %8.sub0:sreg_64, 0, implicit $exec
+    %11:vgpr_32 = V_MOV_B32_dpp %9:vgpr_32, %10:vgpr_32, 312, 15, 15, 0, implicit $exec
+    %12:sreg_32 = V_READLANE_B32 %11:vgpr_32, 63
+    early-clobber %13:sreg_32 = WWM %9:vgpr_32, implicit $exec
+
+    %14:vgpr_32 = COPY %13
+    BUFFER_STORE_DWORD_OFFSET_exact killed %14, %4, %5, 4, 0, 0, 0, 0, 0, implicit $exec
+    S_ENDPGM 0
+
+...