[clang] [llvm] [AMDGPU] Introduce asyncmark/wait intrinsics (PR #173259)

Jay Foad via cfe-commits cfe-commits at lists.llvm.org
Mon Dec 22 07:18:02 PST 2025


================
@@ -566,6 +569,35 @@ class SIInsertWaitcnts {
     return VmemReadMapping[getVmemType(Inst)];
   }
 
+  bool hasCPolAsyncBit(const MachineInstr &MI) const {
+    const MachineOperand *CPol = TII->getNamedOperand(MI, AMDGPU::OpName::cpol);
+    if (!CPol || !CPol->isImm())
+      return false;
+    return CPol->getImm() & AMDGPU::CPol::ASYNC_pregfx12;
+  }
+
+  // FIXME: For GFX1250, this should also check for usesASYNC_CNT
+  bool isAsync(const MachineInstr &MI) const {
+    if (!SIInstrInfo::isLDSDMA(MI))
+      return false;
+    if (SIInstrInfo::usesASYNC_CNT(MI)) {
----------------
jayfoad wrote:

Nit: don't need braces

https://github.com/llvm/llvm-project/pull/173259


More information about the cfe-commits mailing list