[llvm] r357063 - [AMDGPU][MC] Corrected handling of tied src for atomic return MUBUF opcodes

[Dmitry Preobrazhensky via llvm-commits]

Author: dpreobra
Date: Wed Mar 27 06:07:41 2019
New Revision: 357063

URL: http://llvm.org/viewvc/llvm-project?rev=357063&view=rev
Log:
[AMDGPU][MC] Corrected handling of tied src for atomic return MUBUF opcodes

See bug 40917: https://bugs.llvm.org/show_bug.cgi?id=40917

Reviewers: artem.tamazov, arsenm

Differential Revision: https://reviews.llvm.org/D59305

Modified:
    llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
    llvm/trunk/test/MC/AMDGPU/mubuf.s

Modified: llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp?rev=357063&r1=357062&r2=357063&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp (original)
+++ llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp Wed Mar 27 06:07:41 2019
@@ -4911,13 +4911,19 @@ void AMDGPUAsmParser::cvtMubufImpl(MCIns
   bool HasLdsModifier = false;
   OptionalImmIndexMap OptionalIdx;
   assert(IsAtomicReturn ? IsAtomic : true);
+  unsigned FirstOperandIdx = 1;
 
-  for (unsigned i = 1, e = Operands.size(); i != e; ++i) {
+  for (unsigned i = FirstOperandIdx, e = Operands.size(); i != e; ++i) {
     AMDGPUOperand &Op = ((AMDGPUOperand &)*Operands[i]);
 
     // Add the register arguments
     if (Op.isReg()) {
       Op.addRegOperands(Inst, 1);
+      // Insert a tied src for atomic return dst.
+      // This cannot be postponed as subsequent calls to
+      // addImmOperands rely on correct number of MC operands.
+      if (IsAtomicReturn && i == FirstOperandIdx)
+        Op.addRegOperands(Inst, 1);
       continue;
     }
 
@@ -4955,12 +4961,6 @@ void AMDGPUAsmParser::cvtMubufImpl(MCIns
     }
   }
 
-  // Copy $vdata_in operand and insert as $vdata for MUBUF_Atomic RTN insns.
-  if (IsAtomicReturn) {
-    MCInst::iterator I = Inst.begin(); // $vdata_in is always at the beginning.
-    Inst.insert(I, *I);
-  }
-
   addOptionalImmOperand(Inst, Operands, OptionalIdx, AMDGPUOperand::ImmTyOffset);
   if (!IsAtomic) { // glc is hard-coded.
     addOptionalImmOperand(Inst, Operands, OptionalIdx, AMDGPUOperand::ImmTyGLC);

Modified: llvm/trunk/test/MC/AMDGPU/mubuf.s
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/MC/AMDGPU/mubuf.s?rev=357063&r1=357062&r2=357063&view=diff
==============================================================================
--- llvm/trunk/test/MC/AMDGPU/mubuf.s (original)
+++ llvm/trunk/test/MC/AMDGPU/mubuf.s Wed Mar 27 06:07:41 2019
@@ -711,6 +711,14 @@ buffer_atomic_inc v1, v[2:3], s[8:11], 5
 // SICI: buffer_atomic_inc v1, v[2:3], s[8:11], 56 idxen offen offset:4 glc slc ; encoding: [0x04,0x70,0xf0,0xe0,0x02,0x01,0x42,0xb8]
 // VI:   buffer_atomic_inc v1, v[2:3], s[8:11], 56 idxen offen offset:4 glc slc ; encoding: [0x04,0x70,0x2e,0xe1,0x02,0x01,0x02,0xb8]
 
+buffer_atomic_add v5, off, s[8:11], 0.5 offset:4095 glc
+// SICI: buffer_atomic_add v5, off, s[8:11], 0.5 offset:4095 glc ; encoding: [0xff,0x4f,0xc8,0xe0,0x00,0x05,0x02,0xf0]
+// VI:   buffer_atomic_add v5, off, s[8:11], 0.5 offset:4095 glc ; encoding: [0xff,0x4f,0x08,0xe1,0x00,0x05,0x02,0xf0]
+
+buffer_atomic_add v5, off, s[8:11], 0.15915494 offset:4095 glc
+// NOSICI: error: invalid operand for instruction
+// VI:   buffer_atomic_add v5, off, s[8:11], 0.15915494 offset:4095 glc ; encoding: [0xff,0x4f,0x08,0xe1,0x00,0x05,0x02,0xf8]
+
 //===----------------------------------------------------------------------===//
 // Lds support
 //===----------------------------------------------------------------------===//