AMDGPU: Extract full 64-bit subregister and use subregs

Instead of extracting both 32-bit components from the 128-bit
register. This produces fewer copies and is easier for
the copy peephole optimizer to understand and see the actual uses
as extracts from a reg_sequence.

This avoids needing to handle subregister composing in the
PeepholeOptimizer's ValueTracker for this case.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247162 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/AMDGPU/SIInstrInfo.cpp b/lib/Target/AMDGPU/SIInstrInfo.cpp
index 80b541061e310c7291410b1555ab5a4fc07c8b17..bf1c35023fb416eee84569d56e56497618128612 100644 (file)
--- a/lib/Target/AMDGPU/SIInstrInfo.cpp
+++ b/lib/Target/AMDGPU/SIInstrInfo.cpp
@@ -1828,15 +1828,10 @@ void SIInstrInfo::legalizeOperands(MachineInstr *MI) const {
     }
 
     MachineBasicBlock &MBB = *MI->getParent();
-    // Extract the ptr from the resource descriptor.
-
-    // SRsrcPtrLo = srsrc:sub0
-    unsigned SRsrcPtrLo = buildExtractSubReg(MI, MRI, *SRsrc,
-        &AMDGPU::VReg_128RegClass, AMDGPU::sub0, &AMDGPU::VGPR_32RegClass);
 
-    // SRsrcPtrHi = srsrc:sub1
-    unsigned SRsrcPtrHi = buildExtractSubReg(MI, MRI, *SRsrc,
-        &AMDGPU::VReg_128RegClass, AMDGPU::sub1, &AMDGPU::VGPR_32RegClass);
+    // Extract the ptr from the resource descriptor.
+    unsigned SRsrcPtr = buildExtractSubReg(MI, MRI, *SRsrc,
+      &AMDGPU::VReg_128RegClass, AMDGPU::sub0_sub1, &AMDGPU::VReg_64RegClass);
 
     // Create an empty resource descriptor
     unsigned Zero64 = MRI.createVirtualRegister(&AMDGPU::SReg_64RegClass);
@@ -1861,36 +1856,39 @@ void SIInstrInfo::legalizeOperands(MachineInstr *MI) const {
             .addImm(RsrcDataFormat >> 32);
 
     // NewSRsrc = {Zero64, SRsrcFormat}
-    BuildMI(MBB, MI, MI->getDebugLoc(), get(AMDGPU::REG_SEQUENCE),
-            NewSRsrc)
-            .addReg(Zero64)
-            .addImm(AMDGPU::sub0_sub1)
-            .addReg(SRsrcFormatLo)
-            .addImm(AMDGPU::sub2)
-            .addReg(SRsrcFormatHi)
-            .addImm(AMDGPU::sub3);
+    BuildMI(MBB, MI, MI->getDebugLoc(), get(AMDGPU::REG_SEQUENCE), NewSRsrc)
+      .addReg(Zero64)
+      .addImm(AMDGPU::sub0_sub1)
+      .addReg(SRsrcFormatLo)
+      .addImm(AMDGPU::sub2)
+      .addReg(SRsrcFormatHi)
+      .addImm(AMDGPU::sub3);
 
     MachineOperand *VAddr = getNamedOperand(*MI, AMDGPU::OpName::vaddr);
     unsigned NewVAddr = MRI.createVirtualRegister(&AMDGPU::VReg_64RegClass);
-    unsigned NewVAddrLo;
-    unsigned NewVAddrHi;
     if (VAddr) {
       // This is already an ADDR64 instruction so we need to add the pointer
       // extracted from the resource descriptor to the current value of VAddr.
-      NewVAddrLo = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
-      NewVAddrHi = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
+      unsigned NewVAddrLo = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
+      unsigned NewVAddrHi = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
 
-      // NewVaddrLo = SRsrcPtrLo + VAddr:sub0
+      // NewVaddrLo = SRsrcPtr:sub0 + VAddr:sub0
       DebugLoc DL = MI->getDebugLoc();
       BuildMI(MBB, MI, DL, get(AMDGPU::V_ADD_I32_e32), NewVAddrLo)
-        .addReg(SRsrcPtrLo)
+        .addReg(SRsrcPtr, 0, AMDGPU::sub0)
         .addReg(VAddr->getReg(), 0, AMDGPU::sub0);
 
-      // NewVaddrHi = SRsrcPtrHi + VAddr:sub1
+      // NewVaddrHi = SRsrcPtr:sub1 + VAddr:sub1
       BuildMI(MBB, MI, DL, get(AMDGPU::V_ADDC_U32_e32), NewVAddrHi)
-        .addReg(SRsrcPtrHi)
+        .addReg(SRsrcPtr, 0, AMDGPU::sub1)
         .addReg(VAddr->getReg(), 0, AMDGPU::sub1);
 
+      // NewVaddr = {NewVaddrHi, NewVaddrLo}
+      BuildMI(MBB, MI, MI->getDebugLoc(), get(AMDGPU::REG_SEQUENCE), NewVAddr)
+        .addReg(NewVAddrLo)
+        .addImm(AMDGPU::sub0)
+        .addReg(NewVAddrHi)
+        .addImm(AMDGPU::sub1);
     } else {
       // This instructions is the _OFFSET variant, so we need to convert it to
       // ADDR64.
@@ -1917,21 +1915,17 @@ void SIInstrInfo::legalizeOperands(MachineInstr *MI) const {
       MI->removeFromParent();
       MI = Addr64;
 
-      NewVAddrLo = SRsrcPtrLo;
-      NewVAddrHi = SRsrcPtrHi;
+      // NewVaddr = {NewVaddrHi, NewVaddrLo}
+      BuildMI(MBB, MI, MI->getDebugLoc(), get(AMDGPU::REG_SEQUENCE), NewVAddr)
+        .addReg(SRsrcPtr, 0, AMDGPU::sub0)
+        .addImm(AMDGPU::sub0)
+        .addReg(SRsrcPtr, 0, AMDGPU::sub1)
+        .addImm(AMDGPU::sub1);
+
       VAddr = getNamedOperand(*MI, AMDGPU::OpName::vaddr);
       SRsrc = getNamedOperand(*MI, AMDGPU::OpName::srsrc);
     }
 
-    // NewVaddr = {NewVaddrHi, NewVaddrLo}
-    BuildMI(MBB, MI, MI->getDebugLoc(), get(AMDGPU::REG_SEQUENCE),
-            NewVAddr)
-            .addReg(NewVAddrLo)
-            .addImm(AMDGPU::sub0)
-            .addReg(NewVAddrHi)
-            .addImm(AMDGPU::sub1);
-
-
     // Update the instruction to use NewVaddr
     VAddr->setReg(NewVAddr);
     // Update the instruction to use NewSRsrc