[LoongArch] Ensure pcaddu18i and jirl adjacency in tail calls for correct relocation (#113932)

Prior to this patch, both `pcaddu18i` and `jirl` were marked as
scheduling boundaries to prevent instruction reordering that would
disrupt their adjacency. However, in certain cases, epilogues were still
being inserted between these two instructions, breaking the required
proximity. This patch ensures that `pcaddu18i` and `jirl` remain
adjacent even in the presence of epilogues, maintaining correct
relocation behavior for tail calls on LoongArch.

Author: heiher

llvm/lib/Target/LoongArch/LoongArchExpandPseudoInsts.cpp

@@ -165,11 +165,9 @@ bool LoongArchPreRAExpandPseudo::expandMI(
   case LoongArch::PseudoLA_TLS_DESC_LARGE:
     return expandLoadAddressTLSDesc(MBB, MBBI, NextMBBI, /*Large=*/true);
   case LoongArch::PseudoCALL:
-  case LoongArch::PseudoCALL_MEDIUM:
   case LoongArch::PseudoCALL_LARGE:
     return expandFunctionCALL(MBB, MBBI, NextMBBI, /*IsTailCall=*/false);
   case LoongArch::PseudoTAIL:
-  case LoongArch::PseudoTAIL_MEDIUM:
   case LoongArch::PseudoTAIL_LARGE:
     return expandFunctionCALL(MBB, MBBI, NextMBBI, /*IsTailCall=*/true);
   case LoongArch::PseudoBRIND:
@@ -545,7 +543,7 @@ bool LoongArchPreRAExpandPseudo::expandFunctionCALL(
 
   switch (MF->getTarget().getCodeModel()) {
   default:
-    report_fatal_error("Unsupported code model");
+    report_fatal_error("Unexpected code model");
     break;
   case CodeModel::Small: {
     // CALL:
@@ -556,31 +554,6 @@ bool LoongArchPreRAExpandPseudo::expandFunctionCALL(
     CALL = BuildMI(MBB, MBBI, DL, TII->get(Opcode)).add(Func);
     break;
   }
-  case CodeModel::Medium: {
-    // CALL:
-    // pcaddu18i $ra, %call36(func)
-    // jirl      $ra, $ra, 0
-    // TAIL:
-    // pcaddu18i $scratch, %call36(func)
-    // jirl      $r0, $scratch, 0
-    Opcode =
-        IsTailCall ? LoongArch::PseudoJIRL_TAIL : LoongArch::PseudoJIRL_CALL;
-    Register ScratchReg =
-        IsTailCall
-            ? MF->getRegInfo().createVirtualRegister(&LoongArch::GPRRegClass)
-            : LoongArch::R1;
-    MachineInstrBuilder MIB =
-        BuildMI(MBB, MBBI, DL, TII->get(LoongArch::PCADDU18I), ScratchReg);
-
-    CALL =
-        BuildMI(MBB, MBBI, DL, TII->get(Opcode)).addReg(ScratchReg).addImm(0);
-
-    if (Func.isSymbol())
-      MIB.addExternalSymbol(Func.getSymbolName(), LoongArchII::MO_CALL36);
-    else
-      MIB.addDisp(Func, 0, LoongArchII::MO_CALL36);
-    break;
-  }
   case CodeModel::Large: {
     // Emit the 5-insn large address load sequence, either directly or
     // indirectly in case of going through the GOT, then JIRL_TAIL or
@@ -671,6 +644,10 @@ class LoongArchExpandPseudo : public MachineFunctionPass {
                 MachineBasicBlock::iterator &NextMBBI);
   bool expandCopyCFR(MachineBasicBlock &MBB, MachineBasicBlock::iterator MBBI,
                      MachineBasicBlock::iterator &NextMBBI);
+  bool expandFunctionCALL(MachineBasicBlock &MBB,
+                          MachineBasicBlock::iterator MBBI,
+                          MachineBasicBlock::iterator &NextMBBI,
+                          bool IsTailCall);
 };
 
 char LoongArchExpandPseudo::ID = 0;
@@ -705,6 +682,10 @@ bool LoongArchExpandPseudo::expandMI(MachineBasicBlock &MBB,
   switch (MBBI->getOpcode()) {
   case LoongArch::PseudoCopyCFR:
     return expandCopyCFR(MBB, MBBI, NextMBBI);
+  case LoongArch::PseudoCALL_MEDIUM:
+    return expandFunctionCALL(MBB, MBBI, NextMBBI, /*IsTailCall=*/false);
+  case LoongArch::PseudoTAIL_MEDIUM:
+    return expandFunctionCALL(MBB, MBBI, NextMBBI, /*IsTailCall=*/true);
   }
 
   return false;
@@ -763,6 +744,54 @@ bool LoongArchExpandPseudo::expandCopyCFR(
   return true;
 }
 
+bool LoongArchExpandPseudo::expandFunctionCALL(
+    MachineBasicBlock &MBB, MachineBasicBlock::iterator MBBI,
+    MachineBasicBlock::iterator &NextMBBI, bool IsTailCall) {
+  MachineFunction *MF = MBB.getParent();
+  MachineInstr &MI = *MBBI;
+  DebugLoc DL = MI.getDebugLoc();
+  const MachineOperand &Func = MI.getOperand(0);
+  MachineInstrBuilder CALL;
+  unsigned Opcode;
+
+  switch (MF->getTarget().getCodeModel()) {
+  default:
+    report_fatal_error("Unexpected code model");
+    break;
+  case CodeModel::Medium: {
+    // CALL:
+    // pcaddu18i $ra, %call36(func)
+    // jirl      $ra, $ra, 0
+    // TAIL:
+    // pcaddu18i $t8, %call36(func)
+    // jirl      $r0, $t8, 0
+    Opcode =
+        IsTailCall ? LoongArch::PseudoJIRL_TAIL : LoongArch::PseudoJIRL_CALL;
+    Register ScratchReg = IsTailCall ? LoongArch::R20 : LoongArch::R1;
+    MachineInstrBuilder MIB =
+        BuildMI(MBB, MBBI, DL, TII->get(LoongArch::PCADDU18I), ScratchReg);
+
+    CALL =
+        BuildMI(MBB, MBBI, DL, TII->get(Opcode)).addReg(ScratchReg).addImm(0);
+
+    if (Func.isSymbol())
+      MIB.addExternalSymbol(Func.getSymbolName(), LoongArchII::MO_CALL36);
+    else
+      MIB.addDisp(Func, 0, LoongArchII::MO_CALL36);
+    break;
+  }
+  }
+
+  // Transfer implicit operands.
+  CALL.copyImplicitOps(MI);
+
+  // Transfer MI flags.
+  CALL.setMIFlags(MI.getFlags());
+
+  MI.eraseFromParent();
+  return true;
+}
+
 } // end namespace
 
 INITIALIZE_PASS(LoongArchPreRAExpandPseudo, "loongarch-prera-expand-pseudo",

llvm/lib/Target/LoongArch/LoongArchInstrInfo.cpp

@@ -391,9 +391,6 @@ bool LoongArchInstrInfo::isSchedulingBoundary(const MachineInstr &MI,
   //
   // The following instruction patterns are prohibited from being reordered:
   //
-  // * pcaddu18 $ra, %call36(s)
-  //   jirl     $ra, $ra, 0
-  //
   // * pcalau12i $a0, %pc_hi20(s)
   //   addi.d $a1, $zero, %pc_lo12(s)
   //   lu32i.d $a1, %pc64_lo20(s)
@@ -413,10 +410,6 @@ bool LoongArchInstrInfo::isSchedulingBoundary(const MachineInstr &MI,
   // boundaries, and the instructions between them are guaranteed to be
   // ordered according to data dependencies.
   switch (MI.getOpcode()) {
-  case LoongArch::PCADDU18I:
-    if (MI.getOperand(1).getTargetFlags() == LoongArchII::MO_CALL36)
-      return true;
-    break;
   case LoongArch::PCALAU12I: {
     auto AddI = std::next(MII);
     if (AddI == MIE || AddI->getOpcode() != LoongArch::ADDI_D)

llvm/lib/Target/LoongArch/LoongArchInstrInfo.td

@@ -1484,7 +1484,7 @@ def : Pat<(loongarch_call tglobaladdr:$func), (PseudoCALL tglobaladdr:$func)>;
 def : Pat<(loongarch_call texternalsym:$func), (PseudoCALL texternalsym:$func)>;
 
 // Function call with 'Medium' code model.
-let isCall = 1, Defs = [R1] in
+let isCall = 1, Defs = [R1, R20], Size = 8 in
 def PseudoCALL_MEDIUM : Pseudo<(outs), (ins bare_symbol:$func)>;
 
 let Predicates = [IsLA64] in {
@@ -1533,7 +1533,8 @@ def : Pat<(loongarch_tail (iPTR texternalsym:$dst)),
           (PseudoTAIL texternalsym:$dst)>;
 
 // Tail call with 'Medium' code model.
-let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, Uses = [R3] in
+let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1,
+    Uses = [R3], Defs = [R20], Size = 8 in
 def PseudoTAIL_MEDIUM : Pseudo<(outs), (ins bare_symbol:$dst)>;
 
 let Predicates = [IsLA64] in {

llvm/test/CodeGen/LoongArch/code-models.ll

@@ -105,8 +105,8 @@ define i32 @caller_tail(i32 %i) nounwind {
 ;
 ; MEDIUM-LABEL: caller_tail:
 ; MEDIUM:       # %bb.0: # %entry
-; MEDIUM-NEXT:    pcaddu18i $a1, %call36(callee_tail)
-; MEDIUM-NEXT:    jr $a1
+; MEDIUM-NEXT:    pcaddu18i $t8, %call36(callee_tail)
+; MEDIUM-NEXT:    jr $t8
 ;
 ; LARGE-LABEL: caller_tail:
 ; LARGE:       # %bb.0: # %entry

llvm/test/CodeGen/LoongArch/expand-adjacency.ll

@@ -27,11 +27,8 @@ declare i32 @callee_tail(i32 %i)
 
 define i32 @caller_call_tail(i32 %i) nounwind {
 ; MEDIUM-LABEL: caller_call_tail:
-; MEDIUM:         pcaddu18i $a1, %call36(callee_tail)
-; MEDIUM-NEXT:    ld.d {{.*}}
-; MEDIUM-NEXT:    ld.d {{.*}}
-; MEDIUM-NEXT:    addi.d {{.*}}
-; MEDIUM-NEXT:    jr $a1
+; MEDIUM:         pcaddu18i $t8, %call36(callee_tail)
+; MEDIUM-NEXT:    jr $t8
 ;
 ; LARGE-LABEL: caller_call_tail:
 ; LARGE:         pcalau12i [[REG1:\$[a-z0-9]+]], %got_pc_hi20(callee_tail)