[Thumb-1] Synthesize TBB/TBH instructions to make use of compressed jump tables

[Reapplying r284580 and r285917 with fix and testing to ensure emitted jump tables for Thumb-1 have 4-byte alignment]

The TBB and TBH instructions in Thumb-2 allow jump tables to be compressed into sequences of bytes or shorts respectively. These instructions do not exist in Thumb-1, however it is possible to synthesize them out of a sequence of other instructions.

It turns out this sequence is so short that it's almost never a lose for performance and is ALWAYS a significant win for code size.

TBB example:
Before: lsls r0, r0, #2    After: add  r0, pc
        adr  r1, .LJTI0_0         ldrb r0, [r0, #6]
        ldr  r0, [r0, r1]         lsls r0, r0, #1
        mov  pc, r0               add  pc, r0
  => No change in prologue code size or dynamic instruction count. Jump table shrunk by a factor of 4.

The only case that can increase dynamic instruction count is the TBH case:

Before: lsls r0, r4, #2    After: lsls r4, r4, #1
        adr  r1, .LJTI0_0         add  r4, pc
        ldr  r0, [r0, r1]         ldrh r4, [r4, #6]
        mov  pc, r0               lsls r4, r4, #1
                                  add  pc, r4
  => 1 more instruction in prologue. Jump table shrunk by a factor of 2.

So there is an argument that this should be disabled when optimizing for performance (and a TBH needs to be generated). I'm not so sure about that in practice, because on small cores with Thumb-1 performance is often tied to code size. But I'm willing to turn it off when optimizing for performance if people want (also note that TBHs are fairly rare in practice!)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@285690 91177308-0d34-0410-b5e6-96231b3b80d8

Author: James Molloy

lib/Target/ARM/ARMAsmPrinter.cpp

@@ -1166,6 +1166,9 @@ void ARMAsmPrinter::EmitJumpTableTBInst(const MachineInstr *MI,
   const MachineOperand &MO1 = MI->getOperand(1);
   unsigned JTI = MO1.getIndex();
 
+  if (Subtarget->isThumb1Only())
+    EmitAlignment(2);
+  
   MCSymbol *JTISymbol = GetARMJTIPICJumpTableLabel(JTI);
   OutStreamer->EmitLabel(JTISymbol);
 
@@ -1712,6 +1715,83 @@ void ARMAsmPrinter::EmitInstruction(const MachineInstr *MI) {
                                      .addReg(0));
     return;
   }
+  case ARM::tTBB_JT:
+  case ARM::tTBH_JT: {
+
+    bool Is8Bit = MI->getOpcode() == ARM::tTBB_JT;
+    unsigned Base = MI->getOperand(0).getReg();
+    unsigned Idx = MI->getOperand(1).getReg();
+    assert(MI->getOperand(1).isKill() && "We need the index register as scratch!");
+
+    // Multiply up idx if necessary.
+    if (!Is8Bit)
+      EmitToStreamer(*OutStreamer, MCInstBuilder(ARM::tLSLri)
+                                       .addReg(Idx)
+                                       .addReg(ARM::CPSR)
+                                       .addReg(Idx)
+                                       .addImm(1)
+                                       // Add predicate operands.
+                                       .addImm(ARMCC::AL)
+                                       .addReg(0));
+
+    if (Base == ARM::PC) {
+      // TBB [base, idx] =
+      //    ADDS idx, idx, base
+      //    LDRB idx, [idx, #4] ; or LDRH if TBH
+      //    LSLS idx, #1
+      //    ADDS pc, pc, idx
+
+      EmitToStreamer(*OutStreamer, MCInstBuilder(ARM::tADDhirr)
+                                       .addReg(Idx)
+                                       .addReg(Idx)
+                                       .addReg(Base)
+                                       // Add predicate operands.
+                                       .addImm(ARMCC::AL)
+                                       .addReg(0));
+
+      unsigned Opc = Is8Bit ? ARM::tLDRBi : ARM::tLDRHi;
+      EmitToStreamer(*OutStreamer, MCInstBuilder(Opc)
+                                       .addReg(Idx)
+                                       .addReg(Idx)
+                                       .addImm(Is8Bit ? 4 : 2)
+                                       // Add predicate operands.
+                                       .addImm(ARMCC::AL)
+                                       .addReg(0));
+    } else {
+      // TBB [base, idx] =
+      //    LDRB idx, [base, idx] ; or LDRH if TBH
+      //    LSLS idx, #1
+      //    ADDS pc, pc, idx
+
+      unsigned Opc = Is8Bit ? ARM::tLDRBr : ARM::tLDRHr;
+      EmitToStreamer(*OutStreamer, MCInstBuilder(Opc)
+                                       .addReg(Idx)
+                                       .addReg(Base)
+                                       .addReg(Idx)
+                                       // Add predicate operands.
+                                       .addImm(ARMCC::AL)
+                                       .addReg(0));
+    }
+
+    EmitToStreamer(*OutStreamer, MCInstBuilder(ARM::tLSLri)
+                                     .addReg(Idx)
+                                     .addReg(ARM::CPSR)
+                                     .addReg(Idx)
+                                     .addImm(1)
+                                     // Add predicate operands.
+                                     .addImm(ARMCC::AL)
+                                     .addReg(0));
+
+    OutStreamer->EmitLabel(GetCPISymbol(MI->getOperand(3).getImm()));
+    EmitToStreamer(*OutStreamer, MCInstBuilder(ARM::tADDhirr)
+                                     .addReg(ARM::PC)
+                                     .addReg(ARM::PC)
+                                     .addReg(Idx)
+                                     // Add predicate operands.
+                                     .addImm(ARMCC::AL)
+                                     .addReg(0));
+    return;
+  }
   case ARM::tBR_JTr:
   case ARM::BR_JTr: {
     // Lower and emit the instruction itself, then the jump table following it.

lib/Target/ARM/ARMConstantIslandPass.cpp

@@ -58,6 +58,11 @@ static cl::opt<unsigned>
 CPMaxIteration("arm-constant-island-max-iteration", cl::Hidden, cl::init(30),
           cl::desc("The max number of iteration for converge"));
 
+static cl::opt<bool> SynthesizeThumb1TBB(
+    "arm-synthesize-thumb-1-tbb", cl::Hidden, cl::init(true),
+    cl::desc("Use compressed jump tables in Thumb-1 by synthesizing an "
+             "equivalent to the TBB/TBH instructions"));
+
 namespace {
   /// ARMConstantIslands - Due to limited PC-relative displacements, ARM
   /// requires constant pool entries to be scattered among the instructions
@@ -189,6 +194,7 @@ namespace {
     bool isThumb;
     bool isThumb1;
     bool isThumb2;
+    bool isPositionIndependentOrROPI;
   public:
     static char ID;
     ARMConstantIslands() : MachineFunctionPass(ID) {}
@@ -319,13 +325,16 @@ bool ARMConstantIslands::runOnMachineFunction(MachineFunction &mf) {
 
   STI = &static_cast<const ARMSubtarget &>(MF->getSubtarget());
   TII = STI->getInstrInfo();
+  isPositionIndependentOrROPI =
+      STI->getTargetLowering()->isPositionIndependent() || STI->isROPI();
   AFI = MF->getInfo<ARMFunctionInfo>();
 
   isThumb = AFI->isThumbFunction();
   isThumb1 = AFI->isThumb1OnlyFunction();
   isThumb2 = AFI->isThumb2Function();
 
   HasFarJump = false;
+  bool GenerateTBB = isThumb2 || (isThumb1 && SynthesizeThumb1TBB);
 
   // This pass invalidates liveness information when it splits basic blocks.
   MF->getRegInfo().invalidateLiveness();
@@ -337,7 +346,7 @@ bool ARMConstantIslands::runOnMachineFunction(MachineFunction &mf) {
   // Try to reorder and otherwise adjust the block layout to make good use
   // of the TB[BH] instructions.
   bool MadeChange = false;
-  if (isThumb2 && AdjustJumpTableBlocks) {
+  if (GenerateTBB && AdjustJumpTableBlocks) {
     scanFunctionJumpTables();
     MadeChange |= reorderThumb2JumpTables();
     // Data is out of date, so clear it. It'll be re-computed later.
@@ -414,7 +423,7 @@ bool ARMConstantIslands::runOnMachineFunction(MachineFunction &mf) {
     MadeChange |= optimizeThumb2Branches();
 
   // Optimize jump tables using TBB / TBH.
-  if (isThumb2)
+  if (GenerateTBB)
     MadeChange |= optimizeThumb2JumpTables();
 
   // After a while, this might be made debug-only, but it is not expensive.
@@ -540,9 +549,11 @@ void ARMConstantIslands::doInitialJumpTablePlacement(
     case ARM::t2BR_JT:
       JTOpcode = ARM::JUMPTABLE_INSTS;
       break;
+    case ARM::tTBB_JT:
     case ARM::t2TBB_JT:
       JTOpcode = ARM::JUMPTABLE_TBB;
       break;
+    case ARM::tTBH_JT:
     case ARM::t2TBH_JT:
       JTOpcode = ARM::JUMPTABLE_TBH;
       break;
@@ -615,8 +626,9 @@ unsigned ARMConstantIslands::getCPELogAlign(const MachineInstr *CPEMI) {
   case ARM::CONSTPOOL_ENTRY:
     break;
   case ARM::JUMPTABLE_TBB:
-    return 0;
+    return isThumb1 ? 2 : 0;
   case ARM::JUMPTABLE_TBH:
+    return isThumb1 ? 2 : 1;
   case ARM::JUMPTABLE_INSTS:
     return 1;
   case ARM::JUMPTABLE_ADDRS:
@@ -638,7 +650,8 @@ unsigned ARMConstantIslands::getCPELogAlign(const MachineInstr *CPEMI) {
 void ARMConstantIslands::scanFunctionJumpTables() {
   for (MachineBasicBlock &MBB : *MF) {
     for (MachineInstr &I : MBB)
-      if (I.isBranch() && I.getOpcode() == ARM::t2BR_JT)
+      if (I.isBranch() &&
+          (I.getOpcode() == ARM::t2BR_JT || I.getOpcode() == ARM::tBR_JTr))
         T2JumpTables.push_back(&I);
   }
 }
@@ -679,6 +692,7 @@ initializeFunctionInfo(const std::vector<MachineInstr*> &CPEMIs) {
         default:
           continue;  // Ignore other JT branches
         case ARM::t2BR_JT:
+        case ARM::tBR_JTr:
           T2JumpTables.push_back(&I);
           continue;   // Does not get an entry in ImmBranches
         case ARM::Bcc:
@@ -1943,7 +1957,7 @@ bool ARMConstantIslands::preserveBaseRegister(MachineInstr *JumpMI,
 
   if (RemovableAdd) {
     RemovableAdd->eraseFromParent();
-    DeadSize += 4;
+    DeadSize += isThumb2 ? 4 : 2;
   } else if (BaseReg == EntryReg) {
     // The add wasn't removable, but clobbered the base for the TBB. So we can't
     // preserve it.
@@ -2010,25 +2024,82 @@ bool ARMConstantIslands::optimizeThumb2JumpTables() {
     if (!ByteOk && !HalfWordOk)
       continue;
 
+    CPUser &User = CPUsers[JumpTableUserIndices[JTI]];
     MachineBasicBlock *MBB = MI->getParent();
     if (!MI->getOperand(0).isKill()) // FIXME: needed now?
       continue;
-    unsigned IdxReg = MI->getOperand(1).getReg();
-    bool IdxRegKill = MI->getOperand(1).isKill();
 
-    CPUser &User = CPUsers[JumpTableUserIndices[JTI]];
     unsigned DeadSize = 0;
     bool CanDeleteLEA = false;
     bool BaseRegKill = false;
-    bool PreservedBaseReg =
+    
+    unsigned IdxReg = ~0U;
+    bool IdxRegKill = true;
+    if (isThumb2) {
+      IdxReg = MI->getOperand(1).getReg();
+      IdxRegKill = MI->getOperand(1).isKill();
+
+      bool PreservedBaseReg =
         preserveBaseRegister(MI, User.MI, DeadSize, CanDeleteLEA, BaseRegKill);
+      if (!jumpTableFollowsTB(MI, User.CPEMI) && !PreservedBaseReg)
+        continue;
+    } else {
+      // We're in thumb-1 mode, so we must have something like:
+      //   %idx = tLSLri %idx, 2
+      //   %base = tLEApcrelJT
+      //   %t = tLDRr %idx, %base
+      unsigned BaseReg = User.MI->getOperand(0).getReg();
+
+      if (User.MI->getIterator() == User.MI->getParent()->begin())
+        continue;
+      MachineInstr *Shift = User.MI->getPrevNode();
+      if (Shift->getOpcode() != ARM::tLSLri ||
+          Shift->getOperand(3).getImm() != 2 ||
+          !Shift->getOperand(2).isKill())
+        continue;
+      IdxReg = Shift->getOperand(2).getReg();
+      unsigned ShiftedIdxReg = Shift->getOperand(0).getReg();
 
-    if (!jumpTableFollowsTB(MI, User.CPEMI) && !PreservedBaseReg)
-      continue;
+      MachineInstr *Load = User.MI->getNextNode();
+      if (Load->getOpcode() != ARM::tLDRr)
+        continue;
+      if (Load->getOperand(1).getReg() != ShiftedIdxReg ||
+          Load->getOperand(2).getReg() != BaseReg ||
+          !Load->getOperand(1).isKill())
+        continue;
 
+      // If we're in PIC mode, there should be another ADD following.
+      if (isPositionIndependentOrROPI) {
+        MachineInstr *Add = Load->getNextNode();
+        if (Add->getOpcode() != ARM::tADDrr ||
+            Add->getOperand(2).getReg() != Load->getOperand(0).getReg() ||
+            Add->getOperand(3).getReg() != BaseReg ||
+            !Add->getOperand(2).isKill())
+          continue;
+        if (Add->getOperand(0).getReg() != MI->getOperand(0).getReg())
+          continue;
+
+        Add->eraseFromParent();
+        DeadSize += 2;
+      } else {
+        if (Load->getOperand(0).getReg() != MI->getOperand(0).getReg())
+          continue;
+      }
+      
+      
+      // Now safe to delete the load and lsl. The LEA will be removed later.
+      CanDeleteLEA = true;
+      Shift->eraseFromParent();
+      Load->eraseFromParent();
+      DeadSize += 4;
+    }
+    
     DEBUG(dbgs() << "Shrink JT: " << *MI);
     MachineInstr *CPEMI = User.CPEMI;
     unsigned Opc = ByteOk ? ARM::t2TBB_JT : ARM::t2TBH_JT;
+    if (!isThumb2)
+      Opc = ByteOk ? ARM::tTBB_JT : ARM::tTBH_JT;
+
     MachineBasicBlock::iterator MI_JT = MI;
     MachineInstr *NewJTMI =
         BuildMI(*MBB, MI_JT, MI->getDebugLoc(), TII->get(Opc))
@@ -2048,7 +2119,7 @@ bool ARMConstantIslands::optimizeThumb2JumpTables() {
 
       if (CanDeleteLEA)  {
         User.MI->eraseFromParent();
-        DeadSize += 4;
+        DeadSize += isThumb2 ? 4 : 2;
 
         // The LEA was eliminated, the TBB instruction becomes the only new user
         // of the jump table.
@@ -2164,9 +2235,16 @@ adjustJTTargetBlockForward(MachineBasicBlock *BB, MachineBasicBlock *JTBB) {
   // Add an unconditional branch from NewBB to BB.
   // There doesn't seem to be meaningful DebugInfo available; this doesn't
   // correspond directly to anything in the source.
-  assert (isThumb2 && "Adjusting for TB[BH] but not in Thumb2?");
-  BuildMI(NewBB, DebugLoc(), TII->get(ARM::t2B)).addMBB(BB)
-          .addImm(ARMCC::AL).addReg(0);
+  if (isThumb2)
+    BuildMI(NewBB, DebugLoc(), TII->get(ARM::t2B))
+        .addMBB(BB)
+        .addImm(ARMCC::AL)
+        .addReg(0);
+  else
+    BuildMI(NewBB, DebugLoc(), TII->get(ARM::tB))
+        .addMBB(BB)
+        .addImm(ARMCC::AL)
+        .addReg(0);
 
   // Update internal data structures to account for the newly inserted MBB.
   MF->RenumberBlocks(NewBB);

lib/Target/ARM/ARMInstrThumb.td

@@ -1315,6 +1315,18 @@ def tLEApcrelJT : tPseudoInst<(outs tGPR:$Rd),
                               (ins i32imm:$label, pred:$p),
                               2, IIC_iALUi, []>, Sched<[WriteALU]>;
 
+// Thumb-1 doesn't have the TBB or TBH instructions, but we can synthesize them
+// and make use of the same compressed jump table format as Thumb-2.
+let Size = 2 in {
+def tTBB_JT : tPseudoInst<(outs),
+        (ins tGPR:$base, tGPR:$index, i32imm:$jt, i32imm:$pclbl), 0, IIC_Br, []>,
+        Sched<[WriteBr]>;
+
+def tTBH_JT : tPseudoInst<(outs),
+        (ins tGPR:$base, tGPR:$index, i32imm:$jt, i32imm:$pclbl), 0, IIC_Br, []>,
+        Sched<[WriteBr]>;
+}
+
 //===----------------------------------------------------------------------===//
 // TLS Instructions
 //

test/CodeGen/ARM/arm-position-independence-jump-table.ll

@@ -8,9 +8,9 @@
 ; RUN: llc -relocation-model=ropi      -mtriple=thumbv7m--none-eabi -disable-block-placement < %s | FileCheck %s --check-prefix=CHECK --check-prefix=THUMB2
 ; RUN: llc -relocation-model=ropi-rwpi -mtriple=thumbv7m--none-eabi -disable-block-placement < %s | FileCheck %s --check-prefix=CHECK --check-prefix=THUMB2
 
-; RUN: llc -relocation-model=static    -mtriple=thumbv6m--none-eabi -disable-block-placement < %s | FileCheck %s --check-prefix=CHECK --check-prefix=THUMB1 --check-prefix=THUMB1_ABS
-; RUN: llc -relocation-model=ropi      -mtriple=thumbv6m--none-eabi -disable-block-placement < %s | FileCheck %s --check-prefix=CHECK --check-prefix=THUMB1 --check-prefix=THUMB1_PC
-; RUN: llc -relocation-model=ropi-rwpi -mtriple=thumbv6m--none-eabi -disable-block-placement < %s | FileCheck %s --check-prefix=CHECK --check-prefix=THUMB1 --check-prefix=THUMB1_PC
+; RUN: llc -relocation-model=static    -mtriple=thumbv6m--none-eabi -disable-block-placement < %s | FileCheck %s --check-prefix=CHECK --check-prefix=THUMB1
+; RUN: llc -relocation-model=ropi      -mtriple=thumbv6m--none-eabi -disable-block-placement < %s | FileCheck %s --check-prefix=CHECK --check-prefix=THUMB1
+; RUN: llc -relocation-model=ropi-rwpi -mtriple=thumbv6m--none-eabi -disable-block-placement < %s | FileCheck %s --check-prefix=CHECK --check-prefix=THUMB1
 
 
 declare void @exit0()
@@ -85,30 +85,23 @@ lab4:
 ; THUMB2: [[LBB4]]
 ; THUMB2-NEXT: b exit4
 
-; THUMB1: lsls    r[[R_TAB_INDEX:[0-9]+]], r{{[0-9]+}}, #2 
-; THUMB1: adr     r[[R_TAB_BASE:[0-9]+]], [[LJTI:\.LJTI[0-9]+_[0-9]+]]
-; THUMB1: ldr     r[[R_BB_ADDR:[0-9]+]], [r[[R_TAB_INDEX]], r[[R_TAB_BASE]]]
-; THUMB1_PC: adds    r[[R_BB_ADDR]], r[[R_BB_ADDR]], r[[R_TAB_BASE]]
-; THUMB1: mov     pc, r[[R_BB_ADDR]]
-; THUMB1: [[LJTI]]
-; THUMB1_ABS: .long [[LBB1:\.LBB[0-9]+_[0-9]+]]+1
-; THUMB1_ABS: .long [[LBB2:\.LBB[0-9]+_[0-9]+]]+1
-; THUMB1_ABS: .long [[LBB3:\.LBB[0-9]+_[0-9]+]]+1
-; THUMB1_ABS: .long [[LBB4:\.LBB[0-9]+_[0-9]+]]+1
-; THUMB1_PC:  .long [[LBB1:\.LBB[0-9]+_[0-9]+]]-[[LJTI]]
-; THUMB1_PC:  .long [[LBB2:\.LBB[0-9]+_[0-9]+]]-[[LJTI]]
-; THUMB1_PC:  .long [[LBB3:\.LBB[0-9]+_[0-9]+]]-[[LJTI]]
-; THUMB1_PC:  .long [[LBB4:\.LBB[0-9]+_[0-9]+]]-[[LJTI]]
+
+; THUMB1: add     r[[x:[0-9]+]], pc
+; THUMB1: ldrb    r[[x]], [r[[x]], #4]
+; THUMB1: lsls    r[[x]], r[[x]], #1
+; THUMB1: [[LCPI:\.LCPI[0-9]+_[0-9]+]]:
+; THUMB1: add     pc, r[[x]]
+; THUMB1: .p2align 2
+; THUMB1: .byte   ([[LBB1:\.LBB[0-9]+_[0-9]+]]-([[LCPI]]+4))/2
+; THUMB1: .byte   ([[LBB2:\.LBB[0-9]+_[0-9]+]]-([[LCPI]]+4))/2
+; THUMB1: .byte   ([[LBB3:\.LBB[0-9]+_[0-9]+]]-([[LCPI]]+4))/2
+; THUMB1: .byte   ([[LBB4:\.LBB[0-9]+_[0-9]+]]-([[LCPI]]+4))/2
 ; THUMB1: [[LBB1]]
 ; THUMB1-NEXT: bl exit1
-; THUMB1-NEXT: pop
 ; THUMB1: [[LBB2]]
 ; THUMB1-NEXT: bl exit2
-; THUMB1-NEXT: pop
 ; THUMB1: [[LBB3]]
 ; THUMB1-NEXT: bl exit3
-; THUMB1-NEXT: pop
 ; THUMB1: [[LBB4]]
 ; THUMB1-NEXT: bl exit4
-; THUMB1-NEXT: pop
 }