[LoongArch][RISCV] Support R_LARCH_{ADD,SUB}_ULEB128/R_RISCV_{SET,SUB}_ULEB128 for .uleb128 directives

This patch is originally from three upstream commits:
1, R_LARCH_{ADD,SUB}_ULEB128 are originally landed from b57159c(llvm#76433).
2, R_RISCV_{SET,SUB}_ULEB128 are originally supported from 1df5ea2. Among it, we change
the default behaviour of `-riscv-uleb128-reloc` to not produce uleb128 reloc, in order
to avoid any other side-effects due to the updated implementation of `MCAssembler::relaxLEB()`
function. And at the same time, we ensure that this patch can't introduce new default traits
(such as the generation for uleb128 reloc) on RISCV in this version.
3, Fix invalid-sleb.s in original commit d7398a3.

Change-Id: Ie687b7d8483c76cf647141162641db1a9d819a04

Author: Ami-zhang

llvm/include/llvm/BinaryFormat/ELFRelocs/RISCV.def

@@ -55,3 +55,5 @@ ELF_RELOC(R_RISCV_SET32,             56)
 ELF_RELOC(R_RISCV_32_PCREL,          57)
 ELF_RELOC(R_RISCV_IRELATIVE,         58)
 ELF_RELOC(R_RISCV_PLT32,             59)
+ELF_RELOC(R_RISCV_SET_ULEB128,       60)
+ELF_RELOC(R_RISCV_SUB_ULEB128,       61)

llvm/include/llvm/MC/MCAsmBackend.h

@@ -21,6 +21,7 @@ class MCAlignFragment;
 class MCDwarfCallFrameFragment;
 class MCDwarfLineAddrFragment;
 class MCFragment;
+class MCLEBFragment;
 class MCRelaxableFragment;
 class MCSymbol;
 class MCAsmLayout;
@@ -194,6 +195,13 @@ class MCAsmBackend {
     return false;
   }
 
+  // Defined by linker relaxation targets to possibly emit LEB128 relocations
+  // and set Value at the relocated location.
+  virtual std::pair<bool, bool>
+  relaxLEB128(MCLEBFragment &LF, MCAsmLayout &Layout, int64_t &Value) const {
+    return std::make_pair(false, false);
+  }
+
   /// @}
 
   /// Returns the minimum size of a nop in bytes on this target. The assembler

llvm/include/llvm/MC/MCFixup.h

@@ -25,6 +25,7 @@ enum MCFixupKind {
   FK_Data_4,      ///< A four-byte fixup.
   FK_Data_8,      ///< A eight-byte fixup.
   FK_Data_6b,     ///< A six-bits fixup.
+  FK_Data_leb128, ///< A leb128 fixup.
   FK_PCRel_1,     ///< A one-byte pc relative fixup.
   FK_PCRel_2,     ///< A two-byte pc relative fixup.
   FK_PCRel_4,     ///< A four-byte pc relative fixup.

llvm/include/llvm/MC/MCFragment.h

@@ -428,7 +428,7 @@ class MCOrgFragment : public MCFragment {
   }
 };
 
-class MCLEBFragment : public MCFragment {
+class MCLEBFragment final : public MCEncodedFragmentWithFixups<10, 1> {
   /// True if this is a sleb128, false if uleb128.
   bool IsSigned;
 
@@ -439,17 +439,16 @@ class MCLEBFragment : public MCFragment {
 
 public:
   MCLEBFragment(const MCExpr &Value_, bool IsSigned_, MCSection *Sec = nullptr)
-      : MCFragment(FT_LEB, false, Sec), IsSigned(IsSigned_), Value(&Value_) {
+      : MCEncodedFragmentWithFixups<10, 1>(FT_LEB, false, Sec),
+        IsSigned(IsSigned_), Value(&Value_) {
     Contents.push_back(0);
   }
 
   const MCExpr &getValue() const { return *Value; }
+  void setValue(const MCExpr *Expr) { Value = Expr; }
 
   bool isSigned() const { return IsSigned; }
 
-  SmallString<8> &getContents() { return Contents; }
-  const SmallString<8> &getContents() const { return Contents; }
-
   /// @}
 
   static bool classof(const MCFragment *F) {

llvm/lib/MC/MCAsmBackend.cpp

@@ -89,6 +89,7 @@ const MCFixupKindInfo &MCAsmBackend::getFixupKindInfo(MCFixupKind Kind) const {
       {"FK_Data_4", 0, 32, 0},
       {"FK_Data_8", 0, 64, 0},
       {"FK_Data_6b", 0, 6, 0},
+      {"FK_Data_leb128", 0, 0, 0},
       {"FK_PCRel_1", 0, 8, MCFixupKindInfo::FKF_IsPCRel},
       {"FK_PCRel_2", 0, 16, MCFixupKindInfo::FKF_IsPCRel},
       {"FK_PCRel_4", 0, 32, MCFixupKindInfo::FKF_IsPCRel},

llvm/lib/MC/MCAssembler.cpp

@@ -918,6 +918,12 @@ void MCAssembler::layout(MCAsmLayout &Layout) {
         Contents = DF.getContents();
         break;
       }
+      case MCFragment::FT_LEB: {
+        auto &LF = cast<MCLEBFragment>(Frag);
+        Fixups = LF.getFixups();
+        Contents = LF.getContents();
+        break;
+      }
       case MCFragment::FT_PseudoProbe: {
         MCPseudoProbeAddrFragment &PF = cast<MCPseudoProbeAddrFragment>(Frag);
         Fixups = PF.getFixups();
@@ -1006,22 +1012,41 @@ bool MCAssembler::relaxInstruction(MCAsmLayout &Layout,
 }
 
 bool MCAssembler::relaxLEB(MCAsmLayout &Layout, MCLEBFragment &LF) {
-  uint64_t OldSize = LF.getContents().size();
+  const unsigned OldSize = static_cast<unsigned>(LF.getContents().size());
+  unsigned PadTo = OldSize;
   int64_t Value;
-  bool Abs = LF.getValue().evaluateKnownAbsolute(Value, Layout);
-  if (!Abs)
-    report_fatal_error("sleb128 and uleb128 expressions must be absolute");
-  SmallString<8> &Data = LF.getContents();
+  SmallVectorImpl<char> &Data = LF.getContents();
+  LF.getFixups().clear();
+  // Use evaluateKnownAbsolute for Mach-O as a hack: .subsections_via_symbols
+  // requires that .uleb128 A-B is foldable where A and B reside in different
+  // fragments. This is used by __gcc_except_table.
+  bool Abs = getSubsectionsViaSymbols()
+                 ? LF.getValue().evaluateKnownAbsolute(Value, Layout)
+                 : LF.getValue().evaluateAsAbsolute(Value, Layout);
+  if (!Abs) {
+    bool Relaxed, UseZeroPad;
+    std::tie(Relaxed, UseZeroPad) = getBackend().relaxLEB128(LF, Layout, Value);
+    if (!Relaxed) {
+      getContext().reportError(LF.getValue().getLoc(),
+                               Twine(LF.isSigned() ? ".s" : ".u") +
+                                   "leb128 expression is not absolute");
+      LF.setValue(MCConstantExpr::create(0, Context));
+    }
+    uint8_t Tmp[10]; // maximum size: ceil(64/7)
+    PadTo = std::max(PadTo, encodeULEB128(uint64_t(Value), Tmp));
+    if (UseZeroPad)
+      Value = 0;
+  }
   Data.clear();
   raw_svector_ostream OSE(Data);
   // The compiler can generate EH table assembly that is impossible to assemble
   // without either adding padding to an LEB fragment or adding extra padding
   // to a later alignment fragment. To accommodate such tables, relaxation can
   // only increase an LEB fragment size here, not decrease it. See PR35809.
   if (LF.isSigned())
-    encodeSLEB128(Value, OSE, OldSize);
+    encodeSLEB128(Value, OSE, PadTo);
   else
-    encodeULEB128(Value, OSE, OldSize);
+    encodeULEB128(Value, OSE, PadTo);
   return OldSize != LF.getContents().size();
 }
 

llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchAsmBackend.cpp

@@ -92,6 +92,7 @@ static uint64_t adjustFixupValue(const MCFixup &Fixup, uint64_t Value,
   case FK_Data_2:
   case FK_Data_4:
   case FK_Data_8:
+  case FK_Data_leb128:
     return Value;
   case LoongArch::fixup_loongarch_b16: {
     if (!isInt<18>(Value))
@@ -129,6 +130,15 @@ static uint64_t adjustFixupValue(const MCFixup &Fixup, uint64_t Value,
   }
 }
 
+static void fixupLeb128(MCContext &Ctx, const MCFixup &Fixup,
+                        MutableArrayRef<char> Data, uint64_t Value) {
+  unsigned I;
+  for (I = 0; I != Data.size() && Value; ++I, Value >>= 7)
+    Data[I] |= uint8_t(Value & 0x7f);
+  if (Value)
+    Ctx.reportError(Fixup.getLoc(), "Invalid uleb128 value!");
+}
+
 void LoongArchAsmBackend::applyFixup(const MCAssembler &Asm,
                                      const MCFixup &Fixup,
                                      const MCValue &Target,
@@ -144,6 +154,10 @@ void LoongArchAsmBackend::applyFixup(const MCAssembler &Asm,
   MCFixupKindInfo Info = getFixupKindInfo(Kind);
   MCContext &Ctx = Asm.getContext();
 
+  // Fixup leb128 separately.
+  if (Fixup.getTargetKind() == FK_Data_leb128)
+    return fixupLeb128(Ctx, Fixup, Data, Value);
+
   // Apply any target-specific value adjustments.
   Value = adjustFixupValue(Fixup, Value, Ctx);
 
@@ -173,6 +187,7 @@ bool LoongArchAsmBackend::shouldForceRelocation(const MCAssembler &Asm,
   case FK_Data_2:
   case FK_Data_4:
   case FK_Data_8:
+  case FK_Data_leb128:
     return !Target.isAbsolute();
   }
 }
@@ -202,9 +217,24 @@ getRelocPairForSize(unsigned Size) {
     return std::make_pair(
         MCFixupKind(FirstLiteralRelocationKind + ELF::R_LARCH_ADD64),
         MCFixupKind(FirstLiteralRelocationKind + ELF::R_LARCH_SUB64));
+  case 128:
+    return std::make_pair(
+        MCFixupKind(FirstLiteralRelocationKind + ELF::R_LARCH_ADD_ULEB128),
+        MCFixupKind(FirstLiteralRelocationKind + ELF::R_LARCH_SUB_ULEB128));
   }
 }
 
+std::pair<bool, bool> LoongArchAsmBackend::relaxLEB128(MCLEBFragment &LF,
+                                                       MCAsmLayout &Layout,
+                                                       int64_t &Value) const {
+  const MCExpr &Expr = LF.getValue();
+  if (LF.isSigned() || !Expr.evaluateKnownAbsolute(Value, Layout))
+    return std::make_pair(false, false);
+  LF.getFixups().push_back(
+      MCFixup::create(0, &Expr, FK_Data_leb128, Expr.getLoc()));
+  return std::make_pair(true, true);
+}
+
 bool LoongArchAsmBackend::writeNopData(raw_ostream &OS, uint64_t Count,
                                        const MCSubtargetInfo *STI) const {
   // We mostly follow binutils' convention here: align to 4-byte boundary with a
@@ -226,21 +256,27 @@ bool LoongArchAsmBackend::handleAddSubRelocations(const MCAsmLayout &Layout,
                                                   uint64_t &FixedValue) const {
   std::pair<MCFixupKind, MCFixupKind> FK;
   uint64_t FixedValueA, FixedValueB;
-  const MCSection &SecA = Target.getSymA()->getSymbol().getSection();
-  const MCSection &SecB = Target.getSymB()->getSymbol().getSection();
-
-  // We need record relocation if SecA != SecB. Usually SecB is same as the
-  // section of Fixup, which will be record the relocation as PCRel. If SecB
-  // is not same as the section of Fixup, it will report error. Just return
-  // false and then this work can be finished by handleFixup.
-  if (&SecA != &SecB)
-    return false;
-
-  // In SecA == SecB case. If the linker relaxation is enabled, we need record
-  // the ADD, SUB relocations. Otherwise the FixedValue has already been
-  // calculated out in evaluateFixup, return true and avoid record relocations.
-  if (!STI.hasFeature(LoongArch::FeatureRelax))
-    return true;
+  const MCSymbol &SA = Target.getSymA()->getSymbol();
+  const MCSymbol &SB = Target.getSymB()->getSymbol();
+
+  bool force = !SA.isInSection() || !SB.isInSection();
+  if (!force) {
+    const MCSection &SecA = SA.getSection();
+    const MCSection &SecB = SB.getSection();
+
+    // We need record relocation if SecA != SecB. Usually SecB is same as the
+    // section of Fixup, which will be record the relocation as PCRel. If SecB
+    // is not same as the section of Fixup, it will report error. Just return
+    // false and then this work can be finished by handleFixup.
+    if (&SecA != &SecB)
+      return false;
+
+    // In SecA == SecB case. If the linker relaxation is enabled, we need record
+    // the ADD, SUB relocations. Otherwise the FixedValue has already been calc-
+    // ulated out in evaluateFixup, return true and avoid record relocations.
+    if (!STI.hasFeature(LoongArch::FeatureRelax))
+      return true;
+  }
 
   switch (Fixup.getKind()) {
   case llvm::FK_Data_1:
@@ -255,6 +291,9 @@ bool LoongArchAsmBackend::handleAddSubRelocations(const MCAsmLayout &Layout,
   case llvm::FK_Data_8:
     FK = getRelocPairForSize(64);
     break;
+  case llvm::FK_Data_leb128:
+    FK = getRelocPairForSize(128);
+    break;
   default:
     llvm_unreachable("unsupported fixup size");
   }

llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchAsmBackend.h

@@ -65,6 +65,9 @@ class LoongArchAsmBackend : public MCAsmBackend {
   void relaxInstruction(MCInst &Inst,
                         const MCSubtargetInfo &STI) const override {}
 
+  std::pair<bool, bool> relaxLEB128(MCLEBFragment &LF, MCAsmLayout &Layout,
+                                    int64_t &Value) const override;
+
   bool writeNopData(raw_ostream &OS, uint64_t Count,
                     const MCSubtargetInfo *STI) const override;
 

llvm/lib/Target/RISCV/MCTargetDesc/RISCVAsmBackend.cpp

@@ -19,6 +19,7 @@
 #include "llvm/MC/MCObjectWriter.h"
 #include "llvm/MC/MCSymbol.h"
 #include "llvm/MC/MCValue.h"
+#include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Endian.h"
 #include "llvm/Support/EndianStream.h"
 #include "llvm/Support/ErrorHandling.h"
@@ -27,6 +28,13 @@
 
 using namespace llvm;
 
+// Temporary workaround for old linkers that do not support ULEB128 relocations,
+// which are abused by DWARF v5 DW_LLE_offset_pair/DW_RLE_offset_pair
+// implemented in Clang/LLVM.
+static cl::opt<bool> ULEB128Reloc(
+    "riscv-uleb128-reloc", cl::init(false), cl::Hidden,
+    cl::desc("Emit R_RISCV_SET_ULEB128/E_RISCV_SUB_ULEB128 if appropriate"));
+
 std::optional<MCFixupKind> RISCVAsmBackend::getFixupKind(StringRef Name) const {
   if (STI.getTargetTriple().isOSBinFormatELF()) {
     unsigned Type;
@@ -126,6 +134,7 @@ bool RISCVAsmBackend::shouldForceRelocation(const MCAssembler &Asm,
   case FK_Data_2:
   case FK_Data_4:
   case FK_Data_8:
+  case FK_Data_leb128:
     if (Target.isAbsolute())
       return false;
     break;
@@ -330,6 +339,19 @@ bool RISCVAsmBackend::relaxDwarfCFA(MCDwarfCallFrameFragment &DF,
   return true;
 }
 
+std::pair<bool, bool> RISCVAsmBackend::relaxLEB128(MCLEBFragment &LF,
+                                                   MCAsmLayout &Layout,
+                                                   int64_t &Value) const {
+  if (LF.isSigned())
+    return std::make_pair(false, false);
+  const MCExpr &Expr = LF.getValue();
+  if (ULEB128Reloc) {
+    LF.getFixups().push_back(
+        MCFixup::create(0, &Expr, FK_Data_leb128, Expr.getLoc()));
+  }
+  return std::make_pair(Expr.evaluateKnownAbsolute(Value, Layout), false);
+}
+
 // Given a compressed control flow instruction this function returns
 // the expanded instruction.
 unsigned RISCVAsmBackend::getRelaxedOpcode(unsigned Op) const {
@@ -416,6 +438,7 @@ static uint64_t adjustFixupValue(const MCFixup &Fixup, uint64_t Value,
   case FK_Data_4:
   case FK_Data_8:
   case FK_Data_6b:
+  case FK_Data_leb128:
     return Value;
   case RISCV::fixup_riscv_set_6b:
     return Value & 0x03;
@@ -596,6 +619,10 @@ bool RISCVAsmBackend::handleAddSubRelocations(const MCAsmLayout &Layout,
     TA = ELF::R_RISCV_ADD64;
     TB = ELF::R_RISCV_SUB64;
     break;
+  case llvm::FK_Data_leb128:
+    TA = ELF::R_RISCV_SET_ULEB128;
+    TB = ELF::R_RISCV_SUB_ULEB128;
+    break;
   default:
     llvm_unreachable("unsupported fixup size");
   }

llvm/lib/Target/RISCV/MCTargetDesc/RISCVAsmBackend.h

@@ -99,6 +99,8 @@ class RISCVAsmBackend : public MCAsmBackend {
                           bool &WasRelaxed) const override;
   bool relaxDwarfCFA(MCDwarfCallFrameFragment &DF, MCAsmLayout &Layout,
                      bool &WasRelaxed) const override;
+  std::pair<bool, bool> relaxLEB128(MCLEBFragment &LF, MCAsmLayout &Layout,
+                                    int64_t &Value) const override;
 
   bool writeNopData(raw_ostream &OS, uint64_t Count,
                     const MCSubtargetInfo *STI) const override;

llvm/test/MC/ELF/RISCV/gen-dwarf.s

@@ -9,7 +9,7 @@
 ## emit special opcodes to make .debug_line smaller, but we don't do this for
 ## consistency.
 
-# RUN: llvm-mc -filetype=obj -triple=riscv64 -g -dwarf-version=5 -mattr=+relax < %s -o %t
+# RUN: llvm-mc -filetype=obj -triple=riscv64 -g -dwarf-version=5 -mattr=+relax -riscv-uleb128-reloc=1 < %s -o %t
 # RUN: llvm-dwarfdump -eh-frame -debug-line -debug-rnglists -v %t | FileCheck %s
 # RUN: llvm-readobj -r -x .eh_frame %t | FileCheck %s --check-prefix=RELOC
 
@@ -48,9 +48,10 @@
 # RELOC-NEXT:   0x34 R_RISCV_32_PCREL <null> 0x0
 # RELOC-NEXT: }
 
-## TODO A section needs two relocations.
 # RELOC:      Section ([[#]]) .rela.debug_rnglists {
 # RELOC-NEXT:   0xD R_RISCV_64 .text.foo 0x0
+# RELOC-NEXT:   0x15 R_RISCV_SET_ULEB128 <null> 0x0
+# RELOC-NEXT:   0x15 R_RISCV_SUB_ULEB128 .text.foo 0x0
 # RELOC-NEXT:   0x17 R_RISCV_64 .text.bar 0x0
 # RELOC-NEXT: }