[RISCV] Add CFI information for vector callee-saved registers

llvm/lib/Target/RISCV/RISCVFrameLowering.cpp

@@ -435,6 +435,33 @@ void RISCVFrameLowering::adjustStackForRVV(MachineFunction &MF,
                Flag, getStackAlign());
 }
 
+static void appendScalableVectorExpression(const TargetRegisterInfo &TRI,
+                                           SmallVectorImpl<char> &Expr,
+                                           int FixedOffset, int ScalableOffset,
+                                           llvm::raw_string_ostream &Comment) {
+  unsigned DwarfVLenB = TRI.getDwarfRegNum(RISCV::VLENB, true);
+  uint8_t Buffer[16];
+  if (FixedOffset) {
+    Expr.push_back(dwarf::DW_OP_consts);
+    Expr.append(Buffer, Buffer + encodeSLEB128(FixedOffset, Buffer));
+    Expr.push_back((uint8_t)dwarf::DW_OP_plus);
+    Comment << (FixedOffset < 0 ? " - " : " + ") << std::abs(FixedOffset);
+  }
+
+  Expr.push_back((uint8_t)dwarf::DW_OP_consts);
+  Expr.append(Buffer, Buffer + encodeSLEB128(ScalableOffset, Buffer));
+
+  Expr.push_back((uint8_t)dwarf::DW_OP_bregx);
+  Expr.append(Buffer, Buffer + encodeULEB128(DwarfVLenB, Buffer));
+  Expr.push_back(0);
+
+  Expr.push_back((uint8_t)dwarf::DW_OP_mul);
+  Expr.push_back((uint8_t)dwarf::DW_OP_plus);
+
+  Comment << (ScalableOffset < 0 ? " - " : " + ") << std::abs(ScalableOffset)
+          << " * vlenb";
+}
+
 static MCCFIInstruction createDefCFAExpression(const TargetRegisterInfo &TRI,
                                                Register Reg,
                                                uint64_t FixedOffset,
@@ -452,30 +479,38 @@ static MCCFIInstruction createDefCFAExpression(const TargetRegisterInfo &TRI,
   else
     Comment << printReg(Reg, &TRI);
 
-  uint8_t buffer[16];
-  if (FixedOffset) {
-    Expr.push_back(dwarf::DW_OP_consts);
-    Expr.append(buffer, buffer + encodeSLEB128(FixedOffset, buffer));
-    Expr.push_back((uint8_t)dwarf::DW_OP_plus);
-    Comment << " + " << FixedOffset;
-  }
+  appendScalableVectorExpression(TRI, Expr, FixedOffset, ScalableOffset,
+                                 Comment);
 
-  Expr.push_back((uint8_t)dwarf::DW_OP_consts);
-  Expr.append(buffer, buffer + encodeSLEB128(ScalableOffset, buffer));
+  SmallString<64> DefCfaExpr;
+  uint8_t Buffer[16];
+  DefCfaExpr.push_back(dwarf::DW_CFA_def_cfa_expression);
+  DefCfaExpr.append(Buffer, Buffer + encodeULEB128(Expr.size(), Buffer));
+  DefCfaExpr.append(Expr.str());
 
-  unsigned DwarfVlenb = TRI.getDwarfRegNum(RISCV::VLENB, true);
-  Expr.push_back((uint8_t)dwarf::DW_OP_bregx);
-  Expr.append(buffer, buffer + encodeULEB128(DwarfVlenb, buffer));
-  Expr.push_back(0);
+  return MCCFIInstruction::createEscape(nullptr, DefCfaExpr.str(), SMLoc(),
+                                        Comment.str());
+}
 
-  Expr.push_back((uint8_t)dwarf::DW_OP_mul);
-  Expr.push_back((uint8_t)dwarf::DW_OP_plus);
+static MCCFIInstruction createDefCFAOffset(const TargetRegisterInfo &TRI,
+                                           Register Reg, uint64_t FixedOffset,
+                                           uint64_t ScalableOffset) {
+  assert(ScalableOffset != 0 && "Did not need to adjust CFA for RVV");
+  SmallString<64> Expr;
+  std::string CommentBuffer;
+  llvm::raw_string_ostream Comment(CommentBuffer);
+  Comment << printReg(Reg, &TRI) << "  @ cfa";
 
-  Comment << " + " << ScalableOffset << " * vlenb";
+  // Build up the expression (FixedOffset + ScalableOffset * VLENB).
+  appendScalableVectorExpression(TRI, Expr, FixedOffset, ScalableOffset,
+                                 Comment);
 
   SmallString<64> DefCfaExpr;
-  DefCfaExpr.push_back(dwarf::DW_CFA_def_cfa_expression);
-  DefCfaExpr.append(buffer, buffer + encodeULEB128(Expr.size(), buffer));
+  uint8_t Buffer[16];
+  unsigned DwarfReg = TRI.getDwarfRegNum(Reg, true);
+  DefCfaExpr.push_back(dwarf::DW_CFA_expression);
+  DefCfaExpr.append(Buffer, Buffer + encodeULEB128(DwarfReg, Buffer));
+  DefCfaExpr.append(Buffer, Buffer + encodeULEB128(Expr.size(), Buffer));
   DefCfaExpr.append(Expr.str());
 
   return MCCFIInstruction::createEscape(nullptr, DefCfaExpr.str(), SMLoc(),
@@ -671,6 +706,9 @@ void RISCVFrameLowering::emitPrologue(MachineFunction &MF,
           .addCFIIndex(CFIIndex)
           .setMIFlag(MachineInstr::FrameSetup);
     }
+
+    std::advance(MBBI, getRVVCalleeSavedInfo(MF, CSI).size());
+    emitCalleeSavedRVVPrologCFI(MBB, MBBI, hasFP(MF));
   }
 
   if (hasFP(MF)) {
@@ -1492,6 +1530,41 @@ bool RISCVFrameLowering::spillCalleeSavedRegisters(
   return true;
 }
 
+void RISCVFrameLowering::emitCalleeSavedRVVPrologCFI(
+    MachineBasicBlock &MBB, MachineBasicBlock::iterator MI, bool HasFP) const {
+  MachineFunction *MF = MBB.getParent();
+  const MachineFrameInfo &MFI = MF->getFrameInfo();
+  RISCVMachineFunctionInfo *RVFI = MF->getInfo<RISCVMachineFunctionInfo>();
+  const TargetInstrInfo &TII = *STI.getInstrInfo();
+  DebugLoc DL = MBB.findDebugLoc(MI);
+
+  const auto &RVVCSI = getRVVCalleeSavedInfo(*MF, MFI.getCalleeSavedInfo());
+  if (RVVCSI.empty())
+    return;
+
+  uint64_t FixedSize = getStackSizeWithRVVPadding(*MF);
+  if (!HasFP) {
+    uint64_t ScalarLocalVarSize =
+        MFI.getStackSize() - RVFI->getCalleeSavedStackSize() -
+        RVFI->getRVPushStackSize() - RVFI->getVarArgsSaveSize() +
+        RVFI->getRVVPadding();
+    FixedSize -= ScalarLocalVarSize;
+  }
+
+  for (auto &CS : RVVCSI) {
+    // Insert the spill to the stack frame.
+    int FI = CS.getFrameIdx();
+    if (FI >= 0 && MFI.getStackID(FI) == TargetStackID::ScalableVector) {
+      unsigned CFIIndex = MF->addFrameInst(
+          createDefCFAOffset(*STI.getRegisterInfo(), CS.getReg(), -FixedSize,
+                             MFI.getObjectOffset(FI) / 8));
+      BuildMI(MBB, MI, DL, TII.get(TargetOpcode::CFI_INSTRUCTION))
+          .addCFIIndex(CFIIndex)
+          .setMIFlag(MachineInstr::FrameSetup);
+    }
+  }
+}
+
 bool RISCVFrameLowering::restoreCalleeSavedRegisters(
     MachineBasicBlock &MBB, MachineBasicBlock::iterator MI,
     MutableArrayRef<CalleeSavedInfo> CSI, const TargetRegisterInfo *TRI) const {

llvm/lib/Target/RISCV/RISCVFrameLowering.h

@@ -88,6 +88,9 @@ class RISCVFrameLowering : public TargetFrameLowering {
   void adjustStackForRVV(MachineFunction &MF, MachineBasicBlock &MBB,
                          MachineBasicBlock::iterator MBBI, const DebugLoc &DL,
                          int64_t Amount, MachineInstr::MIFlag Flag) const;
+  void emitCalleeSavedRVVPrologCFI(MachineBasicBlock &MBB,
+                                   MachineBasicBlock::iterator MI,
+                                   bool HasFP) const;
   std::pair<int64_t, Align>
   assignRVVStackObjectOffsets(MachineFunction &MF) const;
 };

llvm/test/CodeGen/RISCV/rvv-cfi-info.ll

@@ -0,0 +1,111 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv64 -mattr=+v,+m -verify-machineinstrs < %s \
+; RUN:   | FileCheck -check-prefix=OMIT-FP %s
+; RUN: llc -mtriple=riscv64 -mattr=+v,+m -verify-machineinstrs -frame-pointer=all < %s \
+; RUN:   | FileCheck -check-prefix=NO-OMIT-FP %s
+
+define riscv_vector_cc <vscale x 1 x i32> @test_vector_callee_cfi(<vscale x 1 x i32> %va) {
+; OMIT-FP-LABEL: test_vector_callee_cfi:
+; OMIT-FP:       # %bb.0: # %entry
+; OMIT-FP-NEXT:    addi sp, sp, -16
+; OMIT-FP-NEXT:    .cfi_def_cfa_offset 16
+; OMIT-FP-NEXT:    csrr a0, vlenb
+; OMIT-FP-NEXT:    slli a0, a0, 3
+; OMIT-FP-NEXT:    sub sp, sp, a0
+; OMIT-FP-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; OMIT-FP-NEXT:    csrr a0, vlenb
+; OMIT-FP-NEXT:    li a1, 6
+; OMIT-FP-NEXT:    mul a0, a0, a1
+; OMIT-FP-NEXT:    add a0, sp, a0
+; OMIT-FP-NEXT:    addi a0, a0, 16
+; OMIT-FP-NEXT:    vs1r.v v1, (a0) # Unknown-size Folded Spill
+; OMIT-FP-NEXT:    csrr a0, vlenb
+; OMIT-FP-NEXT:    slli a0, a0, 2
+; OMIT-FP-NEXT:    add a0, sp, a0
+; OMIT-FP-NEXT:    addi a0, a0, 16
+; OMIT-FP-NEXT:    vs2r.v v2, (a0) # Unknown-size Folded Spill
+; OMIT-FP-NEXT:    addi a0, sp, 16
+; OMIT-FP-NEXT:    vs4r.v v4, (a0) # Unknown-size Folded Spill
+; OMIT-FP-NEXT:    .cfi_escape 0x10, 0x61, 0x08, 0x11, 0x7e, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # $v1 @ cfa - 2 * vlenb
+; OMIT-FP-NEXT:    .cfi_escape 0x10, 0x62, 0x08, 0x11, 0x7c, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # $v2m2 @ cfa - 4 * vlenb
+; OMIT-FP-NEXT:    .cfi_escape 0x10, 0x64, 0x08, 0x11, 0x78, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # $v4m4 @ cfa - 8 * vlenb
+; OMIT-FP-NEXT:    #APP
+; OMIT-FP-NEXT:    #NO_APP
+; OMIT-FP-NEXT:    csrr a0, vlenb
+; OMIT-FP-NEXT:    li a1, 6
+; OMIT-FP-NEXT:    mul a0, a0, a1
+; OMIT-FP-NEXT:    add a0, sp, a0
+; OMIT-FP-NEXT:    addi a0, a0, 16
+; OMIT-FP-NEXT:    vl1r.v v1, (a0) # Unknown-size Folded Reload
+; OMIT-FP-NEXT:    csrr a0, vlenb
+; OMIT-FP-NEXT:    slli a0, a0, 2
+; OMIT-FP-NEXT:    add a0, sp, a0
+; OMIT-FP-NEXT:    addi a0, a0, 16
+; OMIT-FP-NEXT:    vl2r.v v2, (a0) # Unknown-size Folded Reload
+; OMIT-FP-NEXT:    addi a0, sp, 16
+; OMIT-FP-NEXT:    vl4r.v v4, (a0) # Unknown-size Folded Reload
+; OMIT-FP-NEXT:    csrr a0, vlenb
+; OMIT-FP-NEXT:    slli a0, a0, 3
+; OMIT-FP-NEXT:    add sp, sp, a0
+; OMIT-FP-NEXT:    addi sp, sp, 16
+; OMIT-FP-NEXT:    ret
+;
+; NO-OMIT-FP-LABEL: test_vector_callee_cfi:
+; NO-OMIT-FP:       # %bb.0: # %entry
+; NO-OMIT-FP-NEXT:    addi sp, sp, -32
+; NO-OMIT-FP-NEXT:    .cfi_def_cfa_offset 32
+; NO-OMIT-FP-NEXT:    sd ra, 24(sp) # 8-byte Folded Spill
+; NO-OMIT-FP-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
+; NO-OMIT-FP-NEXT:    .cfi_offset ra, -8
+; NO-OMIT-FP-NEXT:    .cfi_offset s0, -16
+; NO-OMIT-FP-NEXT:    addi s0, sp, 32
+; NO-OMIT-FP-NEXT:    .cfi_def_cfa s0, 0
+; NO-OMIT-FP-NEXT:    csrr a0, vlenb
+; NO-OMIT-FP-NEXT:    slli a0, a0, 3
+; NO-OMIT-FP-NEXT:    sub sp, sp, a0
+; NO-OMIT-FP-NEXT:    csrr a0, vlenb
+; NO-OMIT-FP-NEXT:    slli a0, a0, 1
+; NO-OMIT-FP-NEXT:    sub a0, s0, a0
+; NO-OMIT-FP-NEXT:    addi a0, a0, -32
+; NO-OMIT-FP-NEXT:    vs1r.v v1, (a0) # Unknown-size Folded Spill
+; NO-OMIT-FP-NEXT:    csrr a0, vlenb
+; NO-OMIT-FP-NEXT:    slli a0, a0, 2
+; NO-OMIT-FP-NEXT:    sub a0, s0, a0
+; NO-OMIT-FP-NEXT:    addi a0, a0, -32
+; NO-OMIT-FP-NEXT:    vs2r.v v2, (a0) # Unknown-size Folded Spill
+; NO-OMIT-FP-NEXT:    csrr a0, vlenb
+; NO-OMIT-FP-NEXT:    slli a0, a0, 3
+; NO-OMIT-FP-NEXT:    sub a0, s0, a0
+; NO-OMIT-FP-NEXT:    addi a0, a0, -32
+; NO-OMIT-FP-NEXT:    vs4r.v v4, (a0) # Unknown-size Folded Spill
+; NO-OMIT-FP-NEXT:    .cfi_escape 0x10, 0x61, 0x0b, 0x11, 0x60, 0x22, 0x11, 0x7e, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # $v1 @ cfa - 32 - 2 * vlenb
+; NO-OMIT-FP-NEXT:    .cfi_escape 0x10, 0x62, 0x0b, 0x11, 0x60, 0x22, 0x11, 0x7c, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # $v2m2 @ cfa - 32 - 4 * vlenb
+; NO-OMIT-FP-NEXT:    .cfi_escape 0x10, 0x64, 0x0b, 0x11, 0x60, 0x22, 0x11, 0x78, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # $v4m4 @ cfa - 32 - 8 * vlenb
+; NO-OMIT-FP-NEXT:    #APP
+; NO-OMIT-FP-NEXT:    #NO_APP
+; NO-OMIT-FP-NEXT:    csrr a0, vlenb
+; NO-OMIT-FP-NEXT:    slli a0, a0, 1
+; NO-OMIT-FP-NEXT:    sub a0, s0, a0
+; NO-OMIT-FP-NEXT:    addi a0, a0, -32
+; NO-OMIT-FP-NEXT:    vl1r.v v1, (a0) # Unknown-size Folded Reload
+; NO-OMIT-FP-NEXT:    csrr a0, vlenb
+; NO-OMIT-FP-NEXT:    slli a0, a0, 2
+; NO-OMIT-FP-NEXT:    sub a0, s0, a0
+; NO-OMIT-FP-NEXT:    addi a0, a0, -32
+; NO-OMIT-FP-NEXT:    vl2r.v v2, (a0) # Unknown-size Folded Reload
+; NO-OMIT-FP-NEXT:    csrr a0, vlenb
+; NO-OMIT-FP-NEXT:    slli a0, a0, 3
+; NO-OMIT-FP-NEXT:    sub a0, s0, a0
+; NO-OMIT-FP-NEXT:    addi a0, a0, -32
+; NO-OMIT-FP-NEXT:    vl4r.v v4, (a0) # Unknown-size Folded Reload
+; NO-OMIT-FP-NEXT:    addi sp, s0, -32
+; NO-OMIT-FP-NEXT:    ld ra, 24(sp) # 8-byte Folded Reload
+; NO-OMIT-FP-NEXT:    ld s0, 16(sp) # 8-byte Folded Reload
+; NO-OMIT-FP-NEXT:    addi sp, sp, 32
+; NO-OMIT-FP-NEXT:    ret
+entry:
+  call void asm sideeffect "",
+  "~{v1},~{v2},~{v3},~{v4},~{v5},~{v6},~{v7}"()
+
+  ret <vscale x 1 x i32> %va
+}

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-insert-subvector.ll

@@ -628,6 +628,7 @@ define void @insert_v2i64_nxv16i64_hi(ptr %psv, ptr %out) {
 ; RV32-NEXT:    vs8r.v v8, (a0)
 ; RV32-NEXT:    vs8r.v v16, (a1)
 ; RV32-NEXT:    addi sp, s0, -80
+; RV32-NEXT:    .cfi_def_cfa sp, 80
 ; RV32-NEXT:    lw ra, 76(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    lw s0, 72(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 80
@@ -661,6 +662,7 @@ define void @insert_v2i64_nxv16i64_hi(ptr %psv, ptr %out) {
 ; RV64-NEXT:    vs8r.v v8, (a0)
 ; RV64-NEXT:    vs8r.v v16, (a1)
 ; RV64-NEXT:    addi sp, s0, -80
+; RV64-NEXT:    .cfi_def_cfa sp, 80
 ; RV64-NEXT:    ld ra, 72(sp) # 8-byte Folded Reload
 ; RV64-NEXT:    ld s0, 64(sp) # 8-byte Folded Reload
 ; RV64-NEXT:    addi sp, sp, 80