DWARF: high_pc computation (WebAssembly#2595)

kripken · cadcode · commit 058ab5016f6c · 2020-01-23T17:14:25.000+08:00
Update high_pc values. These are interesting as they
may be a relative offset compared to the low_pc.

For functions we already had both a start and an end. Add
such tracking for instructions as well.
diff --git a/src/passes/Print.cpp b/src/passes/Print.cpp
@@ -1443,8 +1443,8 @@ struct PrintSExpression : public OverriddenVisitor<PrintSExpression> {
         auto iter = currFunction->expressionLocations.find(curr);
         if (iter != currFunction->expressionLocations.end()) {
           Colors::grey(o);
-          o << ";; code offset: 0x" << std::hex << iter->second << std::dec
-            << '\n';
+          o << ";; code offset: 0x" << std::hex << iter->second.start
+            << std::dec << '\n';
           restoreNormalColor(o);
           doIndent(o, indent);
         }
diff --git a/src/wasm-binary.h b/src/wasm-binary.h
@@ -1014,6 +1014,7 @@ class WasmBinaryWriter {
   void writeSourceMapEpilog();
   void writeDebugLocation(const Function::DebugLocation& loc);
   void writeDebugLocation(Expression* curr, Function* func);
+  void writeDebugLocationEnd(Expression* curr, Function* func);
 
   // helpers
   void writeInlineString(const char* name);
diff --git a/src/wasm-stack.h b/src/wasm-stack.h
@@ -92,6 +92,9 @@ class BinaryInstWriter : public OverriddenVisitor<BinaryInstWriter> {
       parent.writeDebugLocation(curr, func);
     }
     OverriddenVisitor<BinaryInstWriter>::visit(curr);
+    if (func && !sourceMap) {
+      parent.writeDebugLocationEnd(curr, func);
+    }
   }
 
   void visitBlock(Block* curr);
diff --git a/src/wasm.h b/src/wasm.h
@@ -1167,7 +1167,7 @@ struct BinaryLocations {
   struct Span {
     uint32_t start, end;
   };
-  std::unordered_map<Expression*, uint32_t> expressions;
+  std::unordered_map<Expression*, Span> expressions;
   std::unordered_map<Function*, Span> functions;
 };
 
@@ -1225,7 +1225,7 @@ class Function : public Importable {
   std::set<DebugLocation> epilogLocation;
 
   // General debugging info support: track instructions and the function itself.
-  std::unordered_map<Expression*, uint32_t> expressionLocations;
+  std::unordered_map<Expression*, BinaryLocations::Span> expressionLocations;
   BinaryLocations::Span funcLocation;
 
   size_t getNumParams();
diff --git a/src/wasm/wasm-binary.cpp b/src/wasm/wasm-binary.cpp
@@ -160,7 +160,8 @@ void WasmBinaryWriter::finishSection(int32_t start) {
     // We are relative to the section start.
     auto totalAdjustment = adjustmentForLEBShrinking + body;
     for (auto& pair : binaryLocations.expressions) {
-      pair.second -= totalAdjustment;
+      pair.second.start -= totalAdjustment;
+      pair.second.end -= totalAdjustment;
     }
     for (auto& pair : binaryLocations.functions) {
       pair.second.start -= totalAdjustment;
@@ -339,7 +340,9 @@ void WasmBinaryWriter::writeFunctions() {
       for (auto* curr : binaryLocationTrackedExpressionsForFunc) {
         // We added the binary locations, adjust them: they must be relative
         // to the code section.
-        binaryLocations.expressions[curr] -= adjustmentForLEBShrinking;
+        auto& span = binaryLocations.expressions[curr];
+        span.start -= adjustmentForLEBShrinking;
+        span.end -= adjustmentForLEBShrinking;
       }
     }
     if (!binaryLocationTrackedExpressionsForFunc.empty()) {
@@ -708,11 +711,20 @@ void WasmBinaryWriter::writeDebugLocation(Expression* curr, Function* func) {
   // If this is an instruction in a function, and if the original wasm had
   // binary locations tracked, then track it in the output as well.
   if (func && !func->expressionLocations.empty()) {
-    binaryLocations.expressions[curr] = o.size();
+    binaryLocations.expressions[curr] =
+      BinaryLocations::Span{uint32_t(o.size()), 0};
     binaryLocationTrackedExpressionsForFunc.push_back(curr);
   }
 }
 
+void WasmBinaryWriter::writeDebugLocationEnd(Expression* curr, Function* func) {
+  if (func && !func->expressionLocations.empty()) {
+    auto& span = binaryLocations.expressions.at(curr);
+    assert(span.end == 0);
+    span.end = o.size();
+  }
+}
+
 void WasmBinaryWriter::writeInlineString(const char* name) {
   int32_t size = strlen(name);
   o << U32LEB(size);
@@ -2293,7 +2305,9 @@ BinaryConsts::ASTNodes WasmBinaryBuilder::readExpression(Expression*& curr) {
       currFunction->debugLocations[curr] = *currDebugLocation.begin();
     }
     if (DWARF && currFunction) {
-      currFunction->expressionLocations[curr] = startPos - codeSectionLocation;
+      currFunction->expressionLocations[curr] =
+        BinaryLocations::Span{uint32_t(startPos - codeSectionLocation),
+                              uint32_t(pos - codeSectionLocation)};
     }
   }
   BYN_TRACE("zz recurse from " << depth-- << " at " << pos << std::endl);
diff --git a/src/wasm/wasm-debug.cpp b/src/wasm/wasm-debug.cpp
@@ -328,45 +328,59 @@ struct LineState {
   }
 };
 
-// Represents a mapping of addresses to expressions.
+// Represents a mapping of addresses to expressions. We track beginnings and
+// endings of expressions separately, since the end of one (which is one past
+// the end in DWARF notation) overlaps with the beginning of the next, and also
+// to let us use contextual information (we may know we are looking up the end
+// of an instruction).
 struct AddrExprMap {
-  std::unordered_map<uint32_t, Expression*> map;
+  std::unordered_map<uint32_t, Expression*> startMap;
+  std::unordered_map<uint32_t, Expression*> endMap;
 
   // Construct the map from the binaryLocations loaded from the wasm.
   AddrExprMap(const Module& wasm) {
     for (auto& func : wasm.functions) {
       for (auto pair : func->expressionLocations) {
-        assert(map.count(pair.second) == 0);
-        map[pair.second] = pair.first;
+        add(pair.first, pair.second);
       }
     }
   }
 
   // Construct the map from new binaryLocations just written
   AddrExprMap(const BinaryLocations& newLocations) {
     for (auto pair : newLocations.expressions) {
-      assert(map.count(pair.second) == 0);
-      map[pair.second] = pair.first;
+      add(pair.first, pair.second);
     }
   }
 
-  Expression* get(uint32_t addr) const {
-    auto iter = map.find(addr);
-    if (iter != map.end()) {
+  Expression* getStart(uint32_t addr) const {
+    auto iter = startMap.find(addr);
+    if (iter != startMap.end()) {
       return iter->second;
     }
     return nullptr;
   }
 
-  void dump() const {
-    std::cout << "  (size: " << map.size() << ")\n";
-    for (auto pair : map) {
-      std::cout << "  " << pair.first << " => " << pair.second << '\n';
+  Expression* getEnd(uint32_t addr) const {
+    auto iter = endMap.find(addr);
+    if (iter != endMap.end()) {
+      return iter->second;
     }
+    return nullptr;
+  }
+
+private:
+  void add(Expression* expr, BinaryLocations::Span span) {
+    assert(startMap.count(span.start) == 0);
+    startMap[span.start] = expr;
+    assert(endMap.count(span.end) == 0);
+    endMap[span.end] = expr;
   }
 };
 
-// Represents a mapping of addresses to expressions.
+// Represents a mapping of addresses to expressions. Note that we use a single
+// map for the start and end addresses, since there is no chance of a function's
+// start overlapping with another's end (there is the size LEB in the middle).
 struct FuncAddrMap {
   std::unordered_map<uint32_t, Function*> map;
 
@@ -415,10 +429,6 @@ struct LocationUpdater {
   // TODO: for memory efficiency, we may want to do this in a streaming manner,
   //       binary to binary, without YAML IR.
 
-  // TODO: apparently DWARF offsets may be into the middle of instructions...
-  //       we may need to track their spans too
-  // https://github.com/WebAssembly/debugging/issues/9#issuecomment-567720872
-
   LocationUpdater(Module& wasm, const BinaryLocations& newLocations)
     : wasm(wasm), newLocations(newLocations), oldExprAddrMap(wasm),
       newExprAddrMap(newLocations), oldFuncAddrMap(wasm) {}
@@ -427,10 +437,21 @@ struct LocationUpdater {
   // address, or if there was but if that instruction no longer exists, return
   // 0. Otherwise, return the new updated location.
   uint32_t getNewExprAddr(uint32_t oldAddr) const {
-    if (auto* expr = oldExprAddrMap.get(oldAddr)) {
+    if (auto* expr = oldExprAddrMap.getStart(oldAddr)) {
       auto iter = newLocations.expressions.find(expr);
       if (iter != newLocations.expressions.end()) {
-        uint32_t newAddr = iter->second;
+        uint32_t newAddr = iter->second.start;
+        return newAddr;
+      }
+    }
+    return 0;
+  }
+
+  uint32_t getNewExprEndAddr(uint32_t oldAddr) const {
+    if (auto* expr = oldExprAddrMap.getEnd(oldAddr)) {
+      auto iter = newLocations.expressions.find(expr);
+      if (iter != newLocations.expressions.end()) {
+        uint32_t newAddr = iter->second.end;
         return newAddr;
       }
     }
@@ -529,6 +550,76 @@ static void iterContextAndYAML(const T& contextList, U& yamlList, W func) {
   assert(yamlValue == yamlList.end());
 }
 
+static void updateDIE(const llvm::DWARFDebugInfoEntry& DIE,
+                      llvm::DWARFYAML::Entry& yamlEntry,
+                      const llvm::DWARFAbbreviationDeclaration* abbrevDecl,
+                      const LocationUpdater& locationUpdater) {
+  auto tag = DIE.getTag();
+  // Pairs of low/high_pc require some special handling, as the high
+  // may be an offset relative to the low. First, process the low_pcs.
+  uint32_t oldLowPC = 0, newLowPC = 0;
+  iterContextAndYAML(
+    abbrevDecl->attributes(),
+    yamlEntry.Values,
+    [&](const llvm::DWARFAbbreviationDeclaration::AttributeSpec& attrSpec,
+        llvm::DWARFYAML::FormValue& yamlValue) {
+      auto attr = attrSpec.Attr;
+      if (attr != llvm::dwarf::DW_AT_low_pc) {
+        return;
+      }
+      uint32_t oldValue = yamlValue.Value, newValue = 0;
+      if (tag == llvm::dwarf::DW_TAG_GNU_call_site ||
+          tag == llvm::dwarf::DW_TAG_inlined_subroutine ||
+          tag == llvm::dwarf::DW_TAG_lexical_block ||
+          tag == llvm::dwarf::DW_TAG_label) {
+        newValue = locationUpdater.getNewExprAddr(oldValue);
+      } else if (tag == llvm::dwarf::DW_TAG_compile_unit ||
+                 tag == llvm::dwarf::DW_TAG_subprogram) {
+        newValue = locationUpdater.getNewFuncAddr(oldValue);
+      } else {
+        Fatal() << "unknown tag with low_pc "
+                << llvm::dwarf::TagString(tag).str();
+      }
+      oldLowPC = oldValue;
+      newLowPC = newValue;
+      yamlValue.Value = newValue;
+    });
+  // Next, process the high_pcs.
+  // TODO: do this more efficiently, without a second traversal (but that's a
+  //       little tricky given the special double-traversal we have).
+  iterContextAndYAML(
+    abbrevDecl->attributes(),
+    yamlEntry.Values,
+    [&](const llvm::DWARFAbbreviationDeclaration::AttributeSpec& attrSpec,
+        llvm::DWARFYAML::FormValue& yamlValue) {
+      auto attr = attrSpec.Attr;
+      if (attr != llvm::dwarf::DW_AT_high_pc) {
+        return;
+      }
+      uint32_t oldValue = yamlValue.Value, newValue = 0;
+      bool isRelative = attrSpec.Form == llvm::dwarf::DW_FORM_data4;
+      if (isRelative) {
+        oldValue += oldLowPC;
+      }
+      if (tag == llvm::dwarf::DW_TAG_GNU_call_site ||
+          tag == llvm::dwarf::DW_TAG_inlined_subroutine ||
+          tag == llvm::dwarf::DW_TAG_lexical_block ||
+          tag == llvm::dwarf::DW_TAG_label) {
+        newValue = locationUpdater.getNewExprEndAddr(oldValue);
+      } else if (tag == llvm::dwarf::DW_TAG_compile_unit ||
+                 tag == llvm::dwarf::DW_TAG_subprogram) {
+        newValue = locationUpdater.getNewFuncAddr(oldValue);
+      } else {
+        Fatal() << "unknown tag with low_pc "
+                << llvm::dwarf::TagString(tag).str();
+      }
+      if (isRelative) {
+        newValue -= newLowPC;
+      }
+      yamlValue.Value = newValue;
+    });
+}
+
 static void updateCompileUnits(const BinaryenDWARFInfo& info,
                                llvm::DWARFYAML::Data& yaml,
                                const LocationUpdater& locationUpdater) {
@@ -545,36 +636,12 @@ static void updateCompileUnits(const BinaryenDWARFInfo& info,
         yamlUnit.Entries,
         [&](const llvm::DWARFDebugInfoEntry& DIE,
             llvm::DWARFYAML::Entry& yamlEntry) {
-          auto tag = DIE.getTag();
           // Process the entries in each relevant DIE, looking for attributes to
           // change.
           auto abbrevDecl = DIE.getAbbreviationDeclarationPtr();
           if (abbrevDecl) {
-            iterContextAndYAML(
-              abbrevDecl->attributes(),
-              yamlEntry.Values,
-              [&](const llvm::DWARFAbbreviationDeclaration::AttributeSpec&
-                    attrSpec,
-                  llvm::DWARFYAML::FormValue& yamlValue) {
-                if (attrSpec.Attr == llvm::dwarf::DW_AT_low_pc) {
-                  if (tag == llvm::dwarf::DW_TAG_GNU_call_site ||
-                      tag == llvm::dwarf::DW_TAG_inlined_subroutine ||
-                      tag == llvm::dwarf::DW_TAG_lexical_block ||
-                      tag == llvm::dwarf::DW_TAG_label) {
-                    // low_pc in certain tags represent expressions.
-                    yamlValue.Value =
-                      locationUpdater.getNewExprAddr(yamlValue.Value);
-                  } else if (tag == llvm::dwarf::DW_TAG_compile_unit ||
-                             tag == llvm::dwarf::DW_TAG_subprogram) {
-                    // low_pc in certain tags represent function.
-                    yamlValue.Value =
-                      locationUpdater.getNewFuncAddr(yamlValue.Value);
-                  } else {
-                    Fatal() << "unknown tag with low_pc "
-                            << llvm::dwarf::TagString(tag).str();
-                  }
-                }
-              });
+            // This is relevant; look for things to update.
+            updateDIE(DIE, yamlEntry, abbrevDecl, locationUpdater);
           }
         });
     });
diff --git a/test/passes/dwarfdump_roundtrip_dwarfdump.bin.txt b/test/passes/dwarfdump_roundtrip_dwarfdump.bin.txt
@@ -138,11 +138,11 @@ Abbrev table for offset: 0x00000000
               DW_AT_stmt_list [DW_FORM_sec_offset]	(0x00000000)
               DW_AT_comp_dir [DW_FORM_strp]	( .debug_str[0x0000009b] = "/usr/local/google/home/azakai/Dev/emscripten")
               DW_AT_low_pc [DW_FORM_addr]	(0x0000000000000000)
-              DW_AT_high_pc [DW_FORM_data4]	(0x00000002)
+              DW_AT_high_pc [DW_FORM_data4]	(0x00000000)
 
 0x00000026:   DW_TAG_subprogram [2]  
                 DW_AT_low_pc [DW_FORM_addr]	(0x0000000000000000)
-                DW_AT_high_pc [DW_FORM_data4]	(0x00000002)
+                DW_AT_high_pc [DW_FORM_data4]	(0x00000000)
                 DW_AT_linkage_name [DW_FORM_strp]	( .debug_str[0x000000c8] = "_Z3foov")
                 DW_AT_name [DW_FORM_strp]	( .debug_str[0x000000d0] = "foo")
                 DW_AT_decl_file [DW_FORM_data1]	("/usr/local/google/home/azakai/Dev/emscripten/a.cpp")
diff --git a/test/passes/fannkuch3.bin.txt b/test/passes/fannkuch3.bin.txt
@@ -2534,7 +2534,7 @@ Abbrev table for offset: 0x00000000
 
 0x00000082:   DW_TAG_subprogram [10] *
                 DW_AT_low_pc [DW_FORM_addr]	(0x0000000000000006)
-                DW_AT_high_pc [DW_FORM_data4]	(0x00000397)
+                DW_AT_high_pc [DW_FORM_data4]	(0x00000383)
                 DW_AT_frame_base [DW_FORM_exprloc]	(DW_OP_WASM_location 0x1 +0, DW_OP_stack_value)
                 DW_AT_GNU_all_call_sites [DW_FORM_flag_present]	(true)
                 DW_AT_linkage_name [DW_FORM_strp]	( .debug_str[0x00000166] = "_Z15fannkuch_workerPv")
@@ -2818,7 +2818,7 @@ Abbrev table for offset: 0x00000000
 
 0x0000023b:   DW_TAG_subprogram [23] *
                 DW_AT_low_pc [DW_FORM_addr]	(0x000000000000038b)
-                DW_AT_high_pc [DW_FORM_data4]	(0x00000342)
+                DW_AT_high_pc [DW_FORM_data4]	(0x000002fb)
                 DW_AT_frame_base [DW_FORM_exprloc]	(DW_OP_WASM_location 0x0 +2, DW_OP_stack_value)
                 DW_AT_GNU_all_call_sites [DW_FORM_flag_present]	(true)
                 DW_AT_name [DW_FORM_strp]	( .debug_str[0x0000018c] = "main")
@@ -2851,7 +2851,7 @@ Abbrev table for offset: 0x00000000
 0x00000278:     DW_TAG_inlined_subroutine [24] *
                   DW_AT_abstract_origin [DW_FORM_ref4]	(cu + 0x01a8 => {0x000001a8} "_ZL8fannkuchi")
                   DW_AT_low_pc [DW_FORM_addr]	(0x00000000000003c6)
-                  DW_AT_high_pc [DW_FORM_data4]	(0x000002c8)
+                  DW_AT_high_pc [DW_FORM_data4]	(0x0000026d)
                   DW_AT_call_file [DW_FORM_data1]	("/usr/local/google/home/azakai/Dev/emscripten/tests/fannkuch.cpp")
                   DW_AT_call_line [DW_FORM_data1]	(159)
                   DW_AT_call_column [DW_FORM_data1]	(0x29)
diff --git a/test/passes/fannkuch3_manyopts.bin.txt b/test/passes/fannkuch3_manyopts.bin.txt
@@ -2534,7 +2534,7 @@ Abbrev table for offset: 0x00000000
 
 0x00000082:   DW_TAG_subprogram [10] *
                 DW_AT_low_pc [DW_FORM_addr]	(0x0000000000000007)
-                DW_AT_high_pc [DW_FORM_data4]	(0x00000397)
+                DW_AT_high_pc [DW_FORM_data4]	(0x00000353)
                 DW_AT_frame_base [DW_FORM_exprloc]	(DW_OP_WASM_location 0x1 +0, DW_OP_stack_value)
                 DW_AT_GNU_all_call_sites [DW_FORM_flag_present]	(true)
                 DW_AT_linkage_name [DW_FORM_strp]	( .debug_str[0x00000166] = "_Z15fannkuch_workerPv")
@@ -2818,7 +2818,7 @@ Abbrev table for offset: 0x00000000
 
 0x0000023b:   DW_TAG_subprogram [23] *
                 DW_AT_low_pc [DW_FORM_addr]	(0x000000000000035c)
-                DW_AT_high_pc [DW_FORM_data4]	(0x00000342)
+                DW_AT_high_pc [DW_FORM_data4]	(0x000002cf)
                 DW_AT_frame_base [DW_FORM_exprloc]	(DW_OP_WASM_location 0x0 +2, DW_OP_stack_value)
                 DW_AT_GNU_all_call_sites [DW_FORM_flag_present]	(true)
                 DW_AT_name [DW_FORM_strp]	( .debug_str[0x0000018c] = "main")
@@ -2851,7 +2851,7 @@ Abbrev table for offset: 0x00000000
 0x00000278:     DW_TAG_inlined_subroutine [24] *
                   DW_AT_abstract_origin [DW_FORM_ref4]	(cu + 0x01a8 => {0x000001a8} "_ZL8fannkuchi")
                   DW_AT_low_pc [DW_FORM_addr]	(0x0000000000000000)
-                  DW_AT_high_pc [DW_FORM_data4]	(0x000002c8)
+                  DW_AT_high_pc [DW_FORM_data4]	(0x00000000)
                   DW_AT_call_file [DW_FORM_data1]	("/usr/local/google/home/azakai/Dev/emscripten/tests/fannkuch.cpp")
                   DW_AT_call_line [DW_FORM_data1]	(159)
                   DW_AT_call_column [DW_FORM_data1]	(0x29)
diff --git a/test/passes/fib2.bin.txt b/test/passes/fib2.bin.txt
diff --git a/test/passes/ignore_missing_func.bin.txt b/test/passes/ignore_missing_func.bin.txt

Original file line number	Diff line number	Diff line change
`@@ -92,6 +92,9 @@ class BinaryInstWriter : public OverriddenVisitor<BinaryInstWriter> {`
`92`	`92`	`parent.writeDebugLocation(curr, func);`
`93`	`93`	`}`
`94`	`94`	`OverriddenVisitor<BinaryInstWriter>::visit(curr);`
	`95`	`+ if (func && !sourceMap) {`
	`96`	`+ parent.writeDebugLocationEnd(curr, func);`
	`97`	`+ }`
`95`	`98`	`}`
`96`	`99`
`97`	`100`	`void visitBlock(Block* curr);`