[CodeGen] Use assembler expressions to lay out the EH LSDA

[rprichard]

Summary

Rely on the assembler to finalize the layout of the DWARF/Itanium
exception-handling LSDA. Rather than calculate the exact size of each
thing in the LSDA, use assembler directives:

• To emit the offset to the TTBase label:

.uleb128 .Lttbase0-.Lttbaseref0
.Lttbaseref0:

• To emit the size of the call site table:

.uleb128 .Lcst_end0-.Lcst_begin0
.Lcst_begin0:
... call site table entries ...
.Lcst_end0:

• To align the type info table:

... action table ...
.balign 4
.long _ZTIi
.long _ZTIl
.Lttbase0:

Using assembler directives simplifies the compiler and allows switching
the encoding of offsets in the call site table from udata4 to uleb128 for
a large code size savings. (This commit does not change the encoding.)

The combination of the uleb128 followed by a balign creates an unfortunate
dependency cycle that the assembler must sometimes resolve either by
padding an LEB or by inserting zero padding before the type table. See
PR35809 or GNU as bug 4029.

[rprichard]

I described this change on the llvm-dev mailing list a month ago: https://lists.llvm.org/pipermail/llvm-dev/2018-January/120178.html.

[rprichard]

Remove the -layout-eh-table-in-assembler option; emit uleb128 code
offsets unconditionally.

Also: update LLVM tests to account for the encoding change.

Comments:

• The SJLJ call site encoding value is still udata4, even though the two fields in each entry are encoded with uleb128. IIRC, I checked GCC, and it specified the encoding as uleb128. I suspect the SJLJ decoder ignores the encoding, but I haven't checked.
• I removed the comments about the "16-byte bundle". I *think* it originated from Itanium documentation, where it refers to a VLIW instruction bundle.
• check-llvm passes. I also built a stage2 toolchain and verified that its check-libcxx and check-libcxxabi pass on a 64-bit Linux machine.

[rprichard]

Split this patch into two parts:

• a patch that uses assembler directives to lay out the LSDA
• a patch that switches the CST encoding to uleb128 and omits TTBase for empty type tables

[rprichard]

Remove the AsmPrinter::EmitPaddedULEB128 and
MCStreamer::EmitPaddedULEB128IntValue functions, which are now unused.

Fixup comments above MCStreamer::Emit[US]LEB128IntValue.

[• espindola]

LGTM. Thanks!

[• espindola]

r324749

Thank you so much for cleaning this up.

include/llvm/CodeGen/AsmPrinter.h

public:
   void EmitInt32(int Value) const;
 
   /// Emit something like ".long Hi-Lo" where the size in bytes of the directive
   /// is specified by Size and Hi/Lo specify the labels.  This implicitly uses
   /// .set if it is available.
   void EmitLabelDifference(const MCSymbol *Hi, const MCSymbol *Lo,
                            unsigned Size) const;
 
+  /// Emit something like ".uleb128 Hi-Lo".
+  void EmitLabelDifferenceAsULEB128(const MCSymbol *Hi,
+                                    const MCSymbol *Lo) const;
+
   /// Emit something like ".long Label+Offset" where the size in bytes of the
   /// directive is specified by Size and Label specifies the label.  This
   /// implicitly uses .set if it is available.
   void EmitLabelPlusOffset(const MCSymbol *Label, uint64_t Offset,
                            unsigned Size, bool IsSectionRelative = false) const;
 
   /// Emit something like ".long Label" where the size in bytes of the directive
   /// is specified by Size and Label specifies the label.

include/llvm/MC/MCObjectStreamer.h

public:
   /// Emit the absolute difference between \c Hi and \c Lo, as long as we can
   /// compute it.  Currently, that requires that both symbols are in the same
   /// data fragment.  Otherwise, do nothing and return \c false.
   ///
   /// \pre Offset of \c Hi is greater than the offset \c Lo.
   void emitAbsoluteSymbolDiff(const MCSymbol *Hi, const MCSymbol *Lo,
                               unsigned Size) override;
 
+  void emitAbsoluteSymbolDiffAsULEB128(const MCSymbol *Hi,
+                                       const MCSymbol *Lo) override;
+
   bool mayHaveInstructions(MCSection &Sec) const override;
 };
 
 } // end namespace llvm
 
 #endif

include/llvm/MC/MCStreamer.h

public:
   virtual void EmitCVFPOData(const MCSymbol *ProcSym, SMLoc Loc = {}) {}
 
   /// Emit the absolute difference between two symbols.
   ///
   /// \pre Offset of \c Hi is greater than the offset \c Lo.
   virtual void emitAbsoluteSymbolDiff(const MCSymbol *Hi, const MCSymbol *Lo,
                                       unsigned Size);
 
+  /// Emit the absolute difference between two symbols encoded with ULEB128.
+  virtual void emitAbsoluteSymbolDiffAsULEB128(const MCSymbol *Hi,
+                                               const MCSymbol *Lo);
+
   virtual MCSymbol *getDwarfLineTableSymbol(unsigned CUID);
   virtual void EmitCFISections(bool EH, bool Debug);
   void EmitCFIStartProc(bool IsSimple);
   void EmitCFIEndProc();
   virtual void EmitCFIDefCfa(int64_t Register, int64_t Offset);
   virtual void EmitCFIDefCfaOffset(int64_t Offset);
   virtual void EmitCFIDefCfaRegister(int64_t Register);
   virtual void EmitCFIOffset(int64_t Register, int64_t Offset);

lib/CodeGen/AsmPrinter/ARMException.cpp

else if (shouldEmitPersonality) {
     // Emit actual exception table
     emitExceptionTable();
   }
 
   if (Asm->MAI->getExceptionHandlingType() == ExceptionHandling::ARM)
     ATS.emitFnEnd();
 }
 
-void ARMException::emitTypeInfos(unsigned TTypeEncoding) {
+void ARMException::emitTypeInfos(unsigned TTypeEncoding, MCSymbol *TTBaseLabel) {
   const MachineFunction *MF = Asm->MF;
   const std::vector<const GlobalValue *> &TypeInfos = MF->getTypeInfos();
   const std::vector<unsigned> &FilterIds = MF->getFilterIds();
 
   bool VerboseAsm = Asm->OutStreamer->isVerboseAsm();
 
   int Entry = 0;
   // Emit the Catch TypeInfos.
   if (VerboseAsm && !TypeInfos.empty()) {
     Asm->OutStreamer->AddComment(">> Catch TypeInfos <<");
     Asm->OutStreamer->AddBlankLine();
     Entry = TypeInfos.size();
   }
 
   for (const GlobalValue *GV : reverse(TypeInfos)) {
     if (VerboseAsm)
       Asm->OutStreamer->AddComment("TypeInfo " + Twine(Entry--));
     Asm->EmitTTypeReference(GV, TTypeEncoding);
   }
 
+  Asm->OutStreamer->EmitLabel(TTBaseLabel);
+
   // Emit the Exception Specifications.
   if (VerboseAsm && !FilterIds.empty()) {
     Asm->OutStreamer->AddComment(">> Filter TypeInfos <<");
     Asm->OutStreamer->AddBlankLine();
     Entry = 0;
   }
   for (std::vector<unsigned>::const_iterator
          I = FilterIds.begin(), E = FilterIds.end(); I < E; ++I) {

lib/CodeGen/AsmPrinter/AsmPrinterDwarf.cpp

 
 void AsmPrinter::EmitULEB128(uint64_t Value, const char *Desc) const {
   if (isVerbose() && Desc)
     OutStreamer->AddComment(Desc);
 
   OutStreamer->EmitULEB128IntValue(Value);
 }
 
+/// Emit something like ".uleb128 Hi-Lo".
+void AsmPrinter::EmitLabelDifferenceAsULEB128(const MCSymbol *Hi,
+                                              const MCSymbol *Lo) const {
+  OutStreamer->emitAbsoluteSymbolDiffAsULEB128(Hi, Lo);
+}
+
 static const char *DecodeDWARFEncoding(unsigned Encoding) {
   switch (Encoding) {
   case dwarf::DW_EH_PE_absptr:
     return "absptr";
   case dwarf::DW_EH_PE_omit:
     return "omit";
   case dwarf::DW_EH_PE_pcrel:
     return "pcrel";
+  case dwarf::DW_EH_PE_uleb128:
+    return "uleb128";
+  case dwarf::DW_EH_PE_sleb128:
+    return "sleb128";
   case dwarf::DW_EH_PE_udata4:
     return "udata4";
   case dwarf::DW_EH_PE_udata8:
     return "udata8";
   case dwarf::DW_EH_PE_sdata4:
     return "sdata4";
   case dwarf::DW_EH_PE_sdata8:
     return "sdata8";

lib/CodeGen/AsmPrinter/DwarfException.h

public:
   /// Gather and emit post-function exception information.
   void endFunction(const MachineFunction *) override;
 
   void beginFragment(const MachineBasicBlock *MBB,
                      ExceptionSymbolProvider ESP) override;
 };
 
 class LLVM_LIBRARY_VISIBILITY ARMException : public DwarfCFIExceptionBase {
-  void emitTypeInfos(unsigned TTypeEncoding) override;
+  void emitTypeInfos(unsigned TTypeEncoding, MCSymbol *TTBaseLabel) override;
   ARMTargetStreamer &getTargetStreamer();
 
 public:
   //===--------------------------------------------------------------------===//
   // Main entry points.
   //
   ARMException(AsmPrinter *A);
   ~ARMException() override;

lib/CodeGen/AsmPrinter/EHStreamer.h

struct CallSiteEntry {
     // LPad contains the landing pad start labels.
     const LandingPadInfo *LPad; // Null indicates that there is no landing pad.
 
     unsigned Action;
   };
 
   /// Compute the actions table and gather the first action index for each
   /// landing pad site.
-  unsigned computeActionsTable(const SmallVectorImpl<const LandingPadInfo*>&LPs,
-                               SmallVectorImpl<ActionEntry> &Actions,
-                               SmallVectorImpl<unsigned> &FirstActions);
+  void computeActionsTable(const SmallVectorImpl<const LandingPadInfo*> &LPs,
 
   void computePadMap(const SmallVectorImpl<const LandingPadInfo *> &LandingPads,
                      RangeMapType &PadMap);
 
   /// Compute the call-site table.  The entry for an invoke has a try-range
   /// containing the call, a non-zero landing pad and an appropriate action.
   /// The entry for an ordinary call has a try-range containing the call and
   /// zero for the landing pad and the action.  Calls marked 'nounwind' have
protected:
   ///     exception.  If it matches then the exception and type id are passed
   ///     on to the landing pad.  Otherwise the next action is looked up.  This
   ///     chain is terminated with a next action of zero.  If no type id is
   ///     found the frame is unwound and handling continues.
   ///  3. Type id table contains references to all the C++ typeinfo for all
   ///     catches in the function.  This tables is reversed indexed base 1.
   void emitExceptionTable();
 
-  virtual void emitTypeInfos(unsigned TTypeEncoding);
+  virtual void emitTypeInfos(unsigned TTypeEncoding, MCSymbol *TTBaseLabel);
 
   // Helpers for identifying what kind of clause an EH typeid or selector
   // corresponds to. Negative selectors are for filter clauses, the zero
   // selector is for cleanups, and positive selectors are for catch clauses.
   static bool isFilterEHSelector(int Selector) { return Selector < 0; }
   static bool isCleanupEHSelector(int Selector) { return Selector == 0; }
   static bool isCatchEHSelector(int Selector) { return Selector > 0; }
 

lib/CodeGen/AsmPrinter/EHStreamer.cpp

for (; Count != MinSize; ++Count)
     if (LIds[Count] != RIds[Count])
       return Count;
 
   return Count;
 }
 
 /// Compute the actions table and gather the first action index for each landing
 /// pad site.
-unsigned EHStreamer::
+void EHStreamer::
 computeActionsTable(const SmallVectorImpl<const LandingPadInfo*> &LandingPads,
                     SmallVectorImpl<ActionEntry> &Actions,
                     SmallVectorImpl<unsigned> &FirstActions) {
   // The action table follows the call-site table in the LSDA. The individual
   // records are of two types:
   //
   //   * Catch clause
   //   * Exception specification
for (SmallVectorImpl<const LandingPadInfo *>::const_iterator
     // indicates that there are no actions.
     FirstActions.push_back(FirstAction);
 
     // Compute this sites contribution to size.
     SizeActions += SizeSiteActions;
 
     PrevLPI = LPI;
   }
-
-  return SizeActions;
 }
 
 /// Return `true' if this is a call to a function marked `nounwind'. Return
 /// `false' otherwise.
 bool EHStreamer::callToNoUnwindFunction(const MachineInstr *MI) {
   assert(MI->isCall() && "This should be a call instruction!");
 
   bool MarkedNoUnwind = false;
void EHStreamer::emitExceptionTable() {
   std::sort(LandingPads.begin(), LandingPads.end(),
             [](const LandingPadInfo *L,
                const LandingPadInfo *R) { return L->TypeIds < R->TypeIds; });
 
   // Compute the actions table and gather the first action index for each
   // landing pad site.
   SmallVector<ActionEntry, 32> Actions;
   SmallVector<unsigned, 64> FirstActions;
-  unsigned SizeActions =
-    computeActionsTable(LandingPads, Actions, FirstActions);
 
   // Compute the call-site table.
   SmallVector<CallSiteEntry, 64> CallSites;
   computeCallSiteTable(CallSites, LandingPads, FirstActions);
 
-  // Final tallies.
-
-  // Call sites.
   bool IsSJLJ = Asm->MAI->getExceptionHandlingType() == ExceptionHandling::SjLj;
   bool HaveTTData = IsSJLJ ? (!TypeInfos.empty() || !FilterIds.empty()) : true;
 
-  unsigned CallSiteTableLength;
-  if (IsSJLJ)
-    CallSiteTableLength = 0;
-  else {
-    unsigned SiteStartSize  = 4; // dwarf::DW_EH_PE_udata4
-    unsigned SiteLengthSize = 4; // dwarf::DW_EH_PE_udata4
-    unsigned LandingPadSize = 4; // dwarf::DW_EH_PE_udata4
-    CallSiteTableLength =
-      CallSites.size() * (SiteStartSize + SiteLengthSize + LandingPadSize);
-  }
-
-  for (unsigned i = 0, e = CallSites.size(); i < e; ++i) {
-    CallSiteTableLength += getULEB128Size(CallSites[i].Action);
-    if (IsSJLJ)
-      CallSiteTableLength += getULEB128Size(i);
-  }
-
   // Type infos.
   MCSection *LSDASection = Asm->getObjFileLowering().getLSDASection();
   unsigned TTypeEncoding;
-  unsigned TypeFormatSize;
 
   if (!HaveTTData) {
     // For SjLj exceptions, if there is no TypeInfo, then we just explicitly say
     // that we're omitting that bit.
     TTypeEncoding = dwarf::DW_EH_PE_omit;
-    // dwarf::DW_EH_PE_absptr
-    TypeFormatSize = Asm->getDataLayout().getPointerSize();
   } else {
     // Okay, we have actual filters or typeinfos to emit.  As such, we need to
     // pick a type encoding for them.  We're about to emit a list of pointers to
     // typeinfo objects at the end of the LSDA.  However, unless we're in static
     // mode, this reference will require a relocation by the dynamic linker.
     //
     // Because of this, we have a couple of options:
     //
if (!HaveTTData) {
     //      need to tell the personality function of the unwinder to indirect
     //      through the dyld stub.
     //
     // FIXME: When (3) is actually implemented, we'll have to emit the stubs
     // somewhere.  This predicate should be moved to a shared location that is
     // in target-independent code.
     //
     TTypeEncoding = Asm->getObjFileLowering().getTTypeEncoding();
-    TypeFormatSize = Asm->GetSizeOfEncodedValue(TTypeEncoding);
   }
 
   // Begin the exception table.
   // Sometimes we want not to emit the data into separate section (e.g. ARM
   // EHABI). In this case LSDASection will be NULL.
   if (LSDASection)
     Asm->OutStreamer->SwitchSection(LSDASection);
   Asm->EmitAlignment(2);
 
   // Emit the LSDA.
   MCSymbol *GCCETSym =
     Asm->OutContext.getOrCreateSymbol(Twine("GCC_except_table")+
                                       Twine(Asm->getFunctionNumber()));
   Asm->OutStreamer->EmitLabel(GCCETSym);
   Asm->OutStreamer->EmitLabel(Asm->getCurExceptionSym());
 
   // Emit the LSDA header.
   Asm->EmitEncodingByte(dwarf::DW_EH_PE_omit, "@LPStart");
   Asm->EmitEncodingByte(TTypeEncoding, "@TType");
 
-  // The type infos need to be aligned. GCC does this by inserting padding just
-  // before the type infos. However, this changes the size of the exception
-  // table, so you need to take this into account when you output the exception
-  // table size. However, the size is output using a variable length encoding.
-  // So by increasing the size by inserting padding, you may increase the number
-  // of bytes used for writing the size. If it increases, say by one byte, then
-  // you now need to output one less byte of padding to get the type infos
-  // aligned. However this decreases the size of the exception table. This
-  // changes the value you have to output for the exception table size. Due to
-  // the variable length encoding, the number of bytes used for writing the
-  // length may decrease. If so, you then have to increase the amount of
-  // padding. And so on. If you look carefully at the GCC code you will see that
-  // it indeed does this in a loop, going on and on until the values stabilize.
-  // We chose another solution: don't output padding inside the table like GCC
-  // does, instead output it before the table.
-  unsigned SizeTypes = TypeInfos.size() * TypeFormatSize;
-  unsigned CallSiteTableLengthSize = getULEB128Size(CallSiteTableLength);
-  unsigned TTypeBaseOffset =
-    sizeof(int8_t) +                            // Call site format
-    CallSiteTableLengthSize +                   // Call site table length size
-    CallSiteTableLength +                       // Call site table length
-    SizeActions +                               // Actions size
-    SizeTypes;
-  unsigned TTypeBaseOffsetSize = getULEB128Size(TTypeBaseOffset);
-  unsigned TotalSize =
-    sizeof(int8_t) +                            // LPStart format
-    sizeof(int8_t) +                            // TType format
-    (HaveTTData ? TTypeBaseOffsetSize : 0) +    // TType base offset size
-    TTypeBaseOffset;                            // TType base offset
-  unsigned PadBytes = (4 - TotalSize) & 3;
-
+  MCSymbol *TTBaseLabel = nullptr;
   if (HaveTTData) {
-    // Account for any extra padding that will be added to the call site table
-    // length.
-    Asm->EmitPaddedULEB128(TTypeBaseOffset, TTypeBaseOffsetSize + PadBytes,
-                           "@TType base offset");
-    PadBytes = 0;
+    // N.B.: There is a dependency loop between the size of the TTBase uleb128
+    // here and the amount of padding before the aligned type table. The
+    // assembler must sometimes pad this uleb128 or insert extra padding before
+    // the type table. See PR35809 or GNU as bug 4029.
+    MCSymbol *TTBaseRefLabel = Asm->createTempSymbol("ttbaseref");
+    TTBaseLabel = Asm->createTempSymbol("ttbase");
+    Asm->EmitLabelDifferenceAsULEB128(TTBaseLabel, TTBaseRefLabel);
+    Asm->OutStreamer->EmitLabel(TTBaseRefLabel);
   }
 
   bool VerboseAsm = Asm->OutStreamer->isVerboseAsm();
 
-  // SjLj Exception handling
-  if (IsSJLJ) {
-    Asm->EmitEncodingByte(dwarf::DW_EH_PE_udata4, "Call site");
+  // Emit the landing pad call site table.
+  MCSymbol *CstBeginLabel = Asm->createTempSymbol("cst_begin");
+  MCSymbol *CstEndLabel = Asm->createTempSymbol("cst_end");
+  Asm->EmitLabelDifferenceAsULEB128(CstEndLabel, CstBeginLabel);
+  Asm->OutStreamer->EmitLabel(CstBeginLabel);
 
-    // Add extra padding if it wasn't added to the TType base offset.
-    Asm->EmitPaddedULEB128(CallSiteTableLength,
-                           CallSiteTableLengthSize + PadBytes,
-                           "Call site table length");
-
-    // Emit the landing pad site information.
+  // SjLj Exception handling
+  if (IsSJLJ) {
     unsigned idx = 0;
     for (SmallVectorImpl<CallSiteEntry>::const_iterator
          I = CallSites.begin(), E = CallSites.end(); I != E; ++I, ++idx) {
       const CallSiteEntry &S = *I;
 
-      // Offset of the landing pad, counted in 16-byte bundles relative to the
-      // @LPStart address.
+      // Index of the call site entry.
       if (VerboseAsm) {
         Asm->OutStreamer->AddComment(">> Call Site " + Twine(idx) + " <<");
         Asm->OutStreamer->AddComment("  On exception at call site "+Twine(idx));
       }
       Asm->EmitULEB128(idx);
 
       // Offset of the first associated action record, relative to the start of
       // the action table. This value is biased by 1 (1 indicates the start of
if (IsSJLJ) {
     //
     //   * The position of the call-site.
     //   * The position of the landing pad.
     //   * The first action record for that call site.
     //
     // A missing entry in the call-site table indicates that a call is not
     // supposed to throw.
 
-    // Emit the landing pad call site table.
-    Asm->EmitEncodingByte(dwarf::DW_EH_PE_udata4, "Call site");
-
-    // Add extra padding if it wasn't added to the TType base offset.
-    Asm->EmitPaddedULEB128(CallSiteTableLength,
-                           CallSiteTableLengthSize + PadBytes,
-                           "Call site table length");
-
     unsigned Entry = 0;
     for (SmallVectorImpl<CallSiteEntry>::const_iterator
          I = CallSites.begin(), E = CallSites.end(); I != E; ++I) {
       const CallSiteEntry &S = *I;
 
       MCSymbol *EHFuncBeginSym = Asm->getFunctionBegin();
 
       MCSymbol *BeginLabel = S.BeginLabel;
       if (!BeginLabel)
         BeginLabel = EHFuncBeginSym;
       MCSymbol *EndLabel = S.EndLabel;
       if (!EndLabel)
         EndLabel = Asm->getFunctionEnd();
 
-      // Offset of the call site relative to the previous call site, counted in
-      // number of 16-byte bundles. The first call site is counted relative to
-      // the start of the procedure fragment.
+      // Offset of the call site relative to the start of the procedure.
       if (VerboseAsm)
         Asm->OutStreamer->AddComment(">> Call Site " + Twine(++Entry) + " <<");
       Asm->EmitLabelDifference(BeginLabel, EHFuncBeginSym, 4);
       if (VerboseAsm)
         Asm->OutStreamer->AddComment(Twine("  Call between ") +
                                      BeginLabel->getName() + " and " +
                                      EndLabel->getName());
       Asm->EmitLabelDifference(EndLabel, BeginLabel, 4);
 
-      // Offset of the landing pad, counted in 16-byte bundles relative to the
-      // @LPStart address.
+      // Offset of the landing pad relative to the start of the procedure.
       if (!S.LPad) {
         if (VerboseAsm)
           Asm->OutStreamer->AddComment("    has no landing pad");
         Asm->OutStreamer->EmitIntValue(0, 4/*size*/);
       } else {
         if (VerboseAsm)
           Asm->OutStreamer->AddComment(Twine("    jumps to ") +
                                        S.LPad->LandingPadLabel->getName());
         Asm->EmitLabelDifference(S.LPad->LandingPadLabel, EHFuncBeginSym, 4);
       }
 
       // Offset of the first associated action record, relative to the start of
       // the action table. This value is biased by 1 (1 indicates the start of
       // the action table), and 0 indicates that there are no actions.
       if (VerboseAsm) {
         if (S.Action == 0)
           Asm->OutStreamer->AddComment("  On action: cleanup");
         else
           Asm->OutStreamer->AddComment("  On action: " +
                                        Twine((S.Action - 1) / 2 + 1));
       }
       Asm->EmitULEB128(S.Action);
     }
   }
+  Asm->OutStreamer->EmitLabel(CstEndLabel);
 
   // Emit the Action Table.
   int Entry = 0;
   for (SmallVectorImpl<ActionEntry>::const_iterator
          I = Actions.begin(), E = Actions.end(); I != E; ++I) {
     const ActionEntry &Action = *I;
 
     if (VerboseAsm) {
if (VerboseAsm) {
       } else {
         unsigned NextAction = Entry + (Action.NextAction + 1) / 2;
         Asm->OutStreamer->AddComment("  Continue to action "+Twine(NextAction));
       }
     }
     Asm->EmitSLEB128(Action.NextAction);
   }
 
-  emitTypeInfos(TTypeEncoding);
+  if (HaveTTData) {
+    Asm->EmitAlignment(2);
+    emitTypeInfos(TTypeEncoding, TTBaseLabel);
+  }
 
   Asm->EmitAlignment(2);
 }
 
-void EHStreamer::emitTypeInfos(unsigned TTypeEncoding) {
+void EHStreamer::emitTypeInfos(unsigned TTypeEncoding, MCSymbol *TTBaseLabel) {
   const MachineFunction *MF = Asm->MF;
   const std::vector<const GlobalValue *> &TypeInfos = MF->getTypeInfos();
   const std::vector<unsigned> &FilterIds = MF->getFilterIds();
 
   bool VerboseAsm = Asm->OutStreamer->isVerboseAsm();
 
   int Entry = 0;
   // Emit the Catch TypeInfos.
   if (VerboseAsm && !TypeInfos.empty()) {
     Asm->OutStreamer->AddComment(">> Catch TypeInfos <<");
     Asm->OutStreamer->AddBlankLine();
     Entry = TypeInfos.size();
   }
 
   for (const GlobalValue *GV : make_range(TypeInfos.rbegin(),
                                           TypeInfos.rend())) {
     if (VerboseAsm)
       Asm->OutStreamer->AddComment("TypeInfo " + Twine(Entry--));
     Asm->EmitTTypeReference(GV, TTypeEncoding);
   }
 
+  Asm->OutStreamer->EmitLabel(TTBaseLabel);
+
   // Emit the Exception Specifications.
   if (VerboseAsm && !FilterIds.empty()) {
     Asm->OutStreamer->AddComment(">> Filter TypeInfos <<");
     Asm->OutStreamer->AddBlankLine();
     Entry = 0;
   }
   for (std::vector<unsigned>::const_iterator
          I = FilterIds.begin(), E = FilterIds.end(); I < E; ++I) {

lib/MC/MCAsmStreamer.cpp

 }
 
 void MCAsmStreamer::EmitULEB128Value(const MCExpr *Value) {
   int64_t IntValue;
   if (Value->evaluateAsAbsolute(IntValue)) {
     EmitULEB128IntValue(IntValue);
     return;
   }
-  OS << ".uleb128 ";
+  OS << "\t.uleb128 ";
   Value->print(OS, MAI);
   EmitEOL();
 }
 
 void MCAsmStreamer::EmitSLEB128Value(const MCExpr *Value) {
   int64_t IntValue;
   if (Value->evaluateAsAbsolute(IntValue)) {
     EmitSLEB128IntValue(IntValue);
     return;
   }
-  OS << ".sleb128 ";
+  OS << "\t.sleb128 ";
   Value->print(OS, MAI);
   EmitEOL();
 }
 
 void MCAsmStreamer::EmitDTPRel64Value(const MCExpr *Value) {
   assert(MAI->getDTPRel64Directive() != nullptr);
   OS << MAI->getDTPRel64Directive();
   Value->print(OS, MAI);

lib/MC/MCObjectStreamer.cpp

void MCObjectStreamer::flushPendingLabels(MCFragment *F, uint64_t FOffset) {
   }
   for (MCSymbol *Sym : PendingLabels) {
     Sym->setFragment(F);
     Sym->setOffset(FOffset);
   }
   PendingLabels.clear();
 }
 
+// As a compile-time optimization, avoid allocating and evaluating an MCExpr
+// tree for (Hi - Lo) when Hi and Lo are offsets into the same fragment.
+static Optional<uint64_t> absoluteSymbolDiff(const MCSymbol *Hi,
+                                             const MCSymbol *Lo) {
+  if (!Hi->getFragment() || Hi->getFragment() != Lo->getFragment() ||
+      Hi->isVariable() || Lo->isVariable())
+    return None;
+
+  return Hi->getOffset() - Lo->getOffset();
+}
+
 void MCObjectStreamer::emitAbsoluteSymbolDiff(const MCSymbol *Hi,
                                               const MCSymbol *Lo,
                                               unsigned Size) {
-  // If not assigned to the same (valid) fragment, fallback.
-  if (!Hi->getFragment() || Hi->getFragment() != Lo->getFragment() ||
-      Hi->isVariable() || Lo->isVariable()) {
-    MCStreamer::emitAbsoluteSymbolDiff(Hi, Lo, Size);
+  if (Optional<uint64_t> Diff = absoluteSymbolDiff(Hi, Lo)) {
+    EmitIntValue(*Diff, Size);
     return;
   }
+  MCStreamer::emitAbsoluteSymbolDiff(Hi, Lo, Size);
+}
 
-  EmitIntValue(Hi->getOffset() - Lo->getOffset(), Size);
+void MCObjectStreamer::emitAbsoluteSymbolDiffAsULEB128(const MCSymbol *Hi,
+                                                       const MCSymbol *Lo) {
+  if (Optional<uint64_t> Diff = absoluteSymbolDiff(Hi, Lo)) {
+    EmitULEB128IntValue(*Diff);
+    return;
+  }
+  MCStreamer::emitAbsoluteSymbolDiffAsULEB128(Hi, Lo);
 }
 
 void MCObjectStreamer::reset() {
   if (Assembler)
     Assembler->reset();
   CurInsertionPoint = MCSection::iterator();
   EmitEHFrame = true;
   EmitDebugFrame = false;

lib/MC/MCStreamer.cpp

void MCStreamer::emitAbsoluteSymbolDiff(const MCSymbol *Hi, const MCSymbol *Lo,
   }
 
   // Otherwise, emit with .set (aka assignment).
   MCSymbol *SetLabel = Context.createTempSymbol("set", true);
   EmitAssignment(SetLabel, Diff);
   EmitSymbolValue(SetLabel, Size);
 }
 
+void MCStreamer::emitAbsoluteSymbolDiffAsULEB128(const MCSymbol *Hi,
+                                                 const MCSymbol *Lo) {
+  // Get the Hi-Lo expression.
+  const MCExpr *Diff =
+      MCBinaryExpr::createSub(MCSymbolRefExpr::create(Hi, Context),
+                              MCSymbolRefExpr::create(Lo, Context), Context);
+
+  EmitULEB128Value(Diff);
+}
+
 void MCStreamer::EmitAssemblerFlag(MCAssemblerFlag Flag) {}
 void MCStreamer::EmitThumbFunc(MCSymbol *Func) {}
 void MCStreamer::EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue) {}
 void MCStreamer::BeginCOFFSymbolDef(const MCSymbol *Symbol) {
   llvm_unreachable("this directive only supported on COFF targets");
 }
 void MCStreamer::EndCOFFSymbolDef() {
   llvm_unreachable("this directive only supported on COFF targets");

test/CodeGen/ARM/ehabi-handlerdata-nounwind.ll

 
 ; CHECK:   .personality __gxx_personality_v0
 ; CHECK:   .handlerdata
 ; CHECK:   .p2align 2
 ; CHECK-LABEL: GCC_except_table0:
 ; CHECK-LABEL: .Lexception0:
 ; CHECK:   .byte 255                     @ @LPStart Encoding = omit
 ; CHECK:   .byte 0                       @ @TType Encoding = absptr
-; CHECK:   .asciz
+; CHECK:   .uleb128 .Lttbase
 ; CHECK:   .byte 3                       @ Call site Encoding = udata4
 ; CHECK:   .fnend

test/CodeGen/ARM/ehabi-handlerdata.ll

 ; CHECK:   .fnstart
 ; CHECK:   .personality __gxx_personality_v0
 ; CHECK:   .handlerdata
 ; CHECK:   .p2align 2
 ; CHECK-LABEL: GCC_except_table0:
 ; CHECK-LABEL: .Lexception0:
 ; CHECK:   .byte 255                     @ @LPStart Encoding = omit
 ; CHECK:   .byte 0                       @ @TType Encoding = absptr
-; CHECK:   .asciz
+; CHECK:   .uleb128 .Lttbase
 ; CHECK:   .byte 3                       @ Call site Encoding = udata4
+; CHECK:   .uleb128 .Lcst_end
 ; CHECK:   .long
 ; CHECK:   .long
 ; CHECK:   .long
 ; CHECK:   .fnend

test/CodeGen/X86/eh-unknown.ll

 ; CHECK: .Lfunc_begin0:
 ; CHECK: .seh_handler __unknown_ehpersonality, @unwind, @except
 ; CHECK: callq throwit
 ; CHECK: .Lfunc_end0:
 ; CHECK: .seh_handlerdata
 ; CHECK: .Lexception0:
 ; CHECK:  .byte   255                     # @LPStart Encoding = omit
 ; CHECK:  .byte   0                       # @TType Encoding = absptr
-; CHECK:  .asciz  "\217\200"              # @TType base offset
+; CHECK:  .uleb128 .Lttbase0-.Lttbaseref0
+; CHECK:  .Lttbaseref0:
 ; CHECK:  .byte   3                       # Call site Encoding = udata4
-; CHECK:  .byte   13                      # Call site table length
-; CHECK:  .long   .Lfunc_begin0-.Lfunc_begin0 # >> Call Site 1 <<
-; CHECK:  .long   .Lfunc_end0-.Lfunc_begin0 #   Call between .Lfunc_begin0 and .Lfunc_end0
+; CHECK:  .uleb128 .Lcst_end0-.Lcst_begin0
+; CHECK:  .Lcst_begin0:
+; CHECK:  .long .Lfunc_begin0-.Lfunc_begin0 # >> Call Site 1 <<
+; CHECK:  .long .Lfunc_end0-.Lfunc_begin0 #   Call between .Lfunc_begin0 and .Lfunc_end0
 ; CHECK:  .long   0                       #     has no landing pad
 ; CHECK:  .byte   0                       #   On action: cleanup
+; CHECK:  .Lcst_end0:

test/CodeGen/X86/patchpoint-invoke.ll

%0 = landingpad { i8*, i32 }
           catch i8* null
   ret i64 0
 }
 
 ; Verify that the exception table was emitted:
 ; CHECK:      [[EXCEPTION_LABEL]]:
 ; CHECK-NEXT: .byte 255
 ; CHECK-NEXT: .byte 3
-; CHECK-NEXT: .byte 21
+; CHECK-NEXT: .uleb128 .Lttbase{{[0-9]+}}-[[TTBASEREF:.Lttbaseref[0-9]+]]
+; CHECK-NEXT: [[TTBASEREF]]:
 ; CHECK-NEXT: .byte 3
-; CHECK-NEXT: .byte 13
+; CHECK-NEXT: .uleb128 .Lcst_end{{[0-9]+}}-[[CST_BEGIN:.Lcst_begin[0-9]+]]
+; CHECK-NEXT: [[CST_BEGIN]]:
 ; Verify that the unwind data covers the entire patchpoint region:
 ; CHECK-NEXT: .long .Ltmp0-[[FUNC_BEGIN]]
 ; CHECK-NEXT: .long [[PP_END]]-.Ltmp0
 
 
 ; Verify that the stackmap section got emitted:
 ; CHECK-LABEL: __LLVM_StackMaps:
 ; Header

test/CodeGen/XCore/exception.ll

Resume:
   resume { i8*, i32 } %0
 Exit:
   ret void
 }
 
 ; CHECK: [[LSDA]]:
 ; CHECK: .byte  255
 ; CHECK: .byte  0
-; CHECK: .asciiz
+; CHECK: .uleb128 [[TTBASE:.Lttbase[0-9]+]]-[[TTBASEREF:.Lttbaseref[0-9]+]]
+; CHECK: [[TTBASEREF]]:
 ; CHECK: .byte  3
-; CHECK: .byte  26
+; CHECK: .uleb128 [[CST_END:.Lcst_end[0-9]+]]-[[CST_BEGIN:.Lcst_begin[0-9]+]]
+; CHECK: [[CST_BEGIN]]:
 ; CHECK: .long [[PRE_G]]-[[START]]
 ; CHECK: .long [[POST_G]]-[[PRE_G]]
 ; CHECK: .long [[LANDING]]-[[START]]
 ; CHECK: .byte 5
 ; CHECK: .long [[POST_G]]-[[START]]
 ; CHECK: .long [[END]]-[[POST_G]]
 ; CHECK: .long 0
 ; CHECK: .byte 0
+; CHECK: [[CST_END]]:
+; CHECK: .byte 0
 ; CHECK: .byte 0
 ; CHECK: .byte 1
 ; CHECK: .byte 125
 ; CHECK: .byte 2
 ; CHECK: .byte 125
+; CHECK: .p2align 2
 ; CHECK: .long _ZTIi
 ; CHECK: .long _ZTId
+; CHECK: [[TTBASE]]: