[AMDGPU] Add wave barrier builtin

The wave barrier represents the discardable barrier. Its main purpose is to
carry convergent attribute, thus preventing illegal CFG optimizations. All lanes
in a wave come to convergence point simultaneously with SIMT, thus no special
instruction is needed in the ISA. The barrier is discarded during code generation.

Differential Revision: https://reviews.llvm.org/D26585

llvm-svn: 287007

3 files changed, 20 insertions, 0 deletions

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUMCInstLower.cpp b/llvm/lib/Target/AMDGPU/AMDGPUMCInstLower.cpp
index cb259a38389..7d56355074b 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUMCInstLower.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUMCInstLower.cpp
@@ -196,6 +196,12 @@ void AMDGPUAsmPrinter::EmitInstruction(const MachineInstr *MI) {
       return;
     }
 
+    if (MI->getOpcode() == AMDGPU::WAVE_BARRIER) {
+      if (isVerbose())
+        OutStreamer->emitRawComment(" wave barrier");
+      return;
+    }
+
     MCInst TmpInst;
     MCInstLowering.lower(MI, TmpInst);
     EmitToStreamer(*OutStreamer, TmpInst);
diff --git a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
index 17b3265bed0..22143605402 100644
--- a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
@@ -3492,6 +3492,9 @@ unsigned SIInstrInfo::getInstSizeInBytes(const MachineInstr &MI) const {
   if (DescSize != 0 && DescSize != 4)
     return DescSize;
 
+  if (Opc == AMDGPU::WAVE_BARRIER)
+    return 0;
+
   // 4-byte instructions may have a 32-bit literal encoded after them. Check
   // operands that coud ever be literals.
   if (isVALU(MI) || isSALU(MI)) {
diff --git a/llvm/lib/Target/AMDGPU/SIInstructions.td b/llvm/lib/Target/AMDGPU/SIInstructions.td
index 15f3ac55faf..423599d8ccf 100644
--- a/llvm/lib/Target/AMDGPU/SIInstructions.td
+++ b/llvm/lib/Target/AMDGPU/SIInstructions.td
@@ -137,6 +137,17 @@ def S_ANDN2_B64_term : PseudoInstSI<(outs SReg_64:$dst),
   let isTerminator = 1;
 }
 
+def WAVE_BARRIER : SPseudoInstSI<(outs), (ins),
+  [(int_amdgcn_wave_barrier)]> {
+  let SchedRW = [];
+  let hasNoSchedulingInfo = 1;
+  let hasSideEffects = 1;
+  let mayLoad = 1;
+  let mayStore = 1;
+  let isBarrier = 1;
+  let isConvergent = 1;
+}
+
 // SI pseudo instructions. These are used by the CFG structurizer pass
 // and should be lowered to ISA instructions prior to codegen.