AMDGPU: Force skip over SMRD, VMEM and s_waitcnt instructions

[pendingchaos]

Summary

This fixes a large Dawn of War 3 performance regression with RADV from Mesa 19.0 to master which was caused by creating less code in some branches.

[pendingchaos]

Seems I forgot to update a few test cases. I'll add a updated revision soon.

[pendingchaos]

This second version updates various tests.

[nhaehnle]

LGTM

lib/Target/AMDGPU/SIInsertSkips.cpp

for (MachineBasicBlock::const_iterator I = MBB.begin(), E = MBB.end();
       // when EXEC = 0. We should skip the loop lest it becomes infinite.
       if (I->getOpcode() == AMDGPU::S_CBRANCH_VCCNZ ||
           I->getOpcode() == AMDGPU::S_CBRANCH_VCCZ)
         return true;
 
       if (TII->hasUnwantedEffectsWhenEXECEmpty(*I))
         return true;
 
+      // These instructions are potentially expensive even if EXEC = 0.
+      if (TII->isSMRD(*I) || TII->isVMEM(*I) || I->getOpcode() == AMDGPU::S_WAITCNT)
+        return true;
+
       ++NumInstr;
       if (NumInstr >= SkipThreshold)
         return true;
     }
   }
 
   return false;
 }