[𝘀𝗽𝗿] initial version

Created using spr 1.3.4

compiler-rt/lib/sanitizer_common/CMakeLists.txt

@@ -88,6 +88,7 @@ set(SANITIZER_SYMBOLIZER_SOURCES
   sanitizer_symbolizer_libcdep.cpp
   sanitizer_symbolizer_mac.cpp
   sanitizer_symbolizer_markup.cpp
+  sanitizer_symbolizer_markup_fuchsia.cpp
   sanitizer_symbolizer_posix_libcdep.cpp
   sanitizer_symbolizer_report.cpp
   sanitizer_symbolizer_report_fuchsia.cpp
@@ -195,6 +196,7 @@ set(SANITIZER_IMPL_HEADERS
   sanitizer_symbolizer_internal.h
   sanitizer_symbolizer_libbacktrace.h
   sanitizer_symbolizer_mac.h
+  sanitizer_symbolizer_markup.h
   sanitizer_syscall_generic.inc
   sanitizer_syscall_linux_aarch64.inc
   sanitizer_syscall_linux_arm.inc

compiler-rt/lib/sanitizer_common/sanitizer_flags.inc

@@ -275,3 +275,7 @@ COMMON_FLAG(bool, test_only_emulate_no_memorymap, false,
 // program.
 COMMON_FLAG(bool, test_only_replace_dlopen_main_program, false,
             "TEST ONLY replace dlopen(<main program>,...) with dlopen(NULL)")
+
+COMMON_FLAG(bool, enable_symbolizer_markup, SANITIZER_FUCHSIA,
+            "Use sanitizer symbolizer markup, available on Linux "
+            "and always set true for fuchsia.")

compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cpp

@@ -16,6 +16,7 @@
 #include "sanitizer_file.h"
 #include "sanitizer_flags.h"
 #include "sanitizer_fuchsia.h"
+#include "sanitizer_symbolizer_markup.h"
 
 namespace __sanitizer {
 
@@ -62,6 +63,9 @@ const char *StackTracePrinter::StripFunctionName(const char *function) {
 #if !SANITIZER_SYMBOLIZER_MARKUP
 
 StackTracePrinter *StackTracePrinter::NewStackTracePrinter() {
+  if (common_flags()->enable_symbolizer_markup)
+    return new (GetGlobalLowLevelAllocator()) MarkupStackTracePrinter();
+
   return new (GetGlobalLowLevelAllocator()) FormattedStackTracePrinter();
 }
 

compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h

@@ -154,6 +154,8 @@ class Symbolizer final {
 
   void InvalidateModuleList();
 
+  const ListOfModules &GetRefreshedListOfModules();
+
  private:
   // GetModuleNameAndOffsetForPC has to return a string to the caller.
   // Since the corresponding module might get unloaded later, we should create

compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cpp

@@ -191,6 +191,13 @@ void Symbolizer::RefreshModules() {
   modules_fresh_ = true;
 }
 
+const ListOfModules &Symbolizer::GetRefreshedListOfModules() {
+  if (!modules_fresh_)
+    RefreshModules();
+
+  return modules_;
+}
+
 static const LoadedModule *SearchForModule(const ListOfModules &modules,
                                            uptr address) {
   for (uptr i = 0; i < modules.size(); i++) {

compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_markup.cpp

@@ -8,105 +8,169 @@
 //
 // This file is shared between various sanitizers' runtime libraries.
 //
-// Implementation of offline markup symbolizer.
+// This generic support for offline symbolizing is based on the
+// Fuchsia port.  We don't do any actual symbolization per se.
+// Instead, we emit text containing raw addresses and raw linkage
+// symbol names, embedded in Fuchsia's symbolization markup format.
+// See the spec at:
+// https://llvm.org/docs/SymbolizerMarkupFormat.html
 //===----------------------------------------------------------------------===//
 
-#include "sanitizer_platform.h"
+#include "sanitizer_symbolizer_markup.h"
 
-#if SANITIZER_SYMBOLIZER_MARKUP
-
-#  include "sanitizer_common.h"
-#  include "sanitizer_stacktrace_printer.h"
-#  include "sanitizer_symbolizer.h"
-#  include "sanitizer_symbolizer_markup_constants.h"
+#include "sanitizer_common.h"
+#include "sanitizer_stacktrace_printer.h"
+#include "sanitizer_symbolizer.h"
+#include "sanitizer_symbolizer_markup_constants.h"
 
 namespace __sanitizer {
 
-// This generic support for offline symbolizing is based on the
-// Fuchsia port.  We don't do any actual symbolization per se.
-// Instead, we emit text containing raw addresses and raw linkage
-// symbol names, embedded in Fuchsia's symbolization markup format.
-// Fuchsia's logging infrastructure emits enough information about
-// process memory layout that a post-processing filter can do the
-// symbolization and pretty-print the markup.  See the spec at:
-// https://fuchsia.googlesource.com/zircon/+/master/docs/symbolizer_markup.md
-
-// This is used by UBSan for type names, and by ASan for global variable names.
-// It's expected to return a static buffer that will be reused on each call.
-const char *Symbolizer::Demangle(const char *name) {
-  static char buffer[kFormatDemangleMax];
-  internal_snprintf(buffer, sizeof(buffer), kFormatDemangle, name);
-  return buffer;
+void MarkupStackTracePrinter::RenderData(InternalScopedString *buffer,
+                                         const char *format, const DataInfo *DI,
+                                         const char *strip_path_prefix) {
+  RenderContext(buffer);
+  buffer->AppendF(kFormatData, DI->start);
 }
 
-// This is used mostly for suppression matching.  Making it work
-// would enable "interceptor_via_lib" suppressions.  It's also used
-// once in UBSan to say "in module ..." in a message that also
-// includes an address in the module, so post-processing can already
-// pretty-print that so as to indicate the module.
-bool Symbolizer::GetModuleNameAndOffsetForPC(uptr pc, const char **module_name,
-                                             uptr *module_address) {
+bool MarkupStackTracePrinter::RenderNeedsSymbolization(const char *format) {
   return false;
 }
 
-// This is mainly used by hwasan for online symbolization. This isn't needed
-// since hwasan can always just dump stack frames for offline symbolization.
-bool Symbolizer::SymbolizeFrame(uptr addr, FrameInfo *info) { return false; }
-
-// This is used in some places for suppression checking, which we
-// don't really support for Fuchsia.  It's also used in UBSan to
-// identify a PC location to a function name, so we always fill in
-// the function member with a string containing markup around the PC
-// value.
-// TODO(mcgrathr): Under SANITIZER_GO, it's currently used by TSan
-// to render stack frames, but that should be changed to use
-// RenderStackFrame.
-SymbolizedStack *Symbolizer::SymbolizePC(uptr addr) {
-  SymbolizedStack *s = SymbolizedStack::New(addr);
+// We don't support the stack_trace_format flag at all.
+void MarkupStackTracePrinter::RenderFrame(InternalScopedString *buffer,
+                                          const char *format, int frame_no,
+                                          uptr address, const AddressInfo *info,
+                                          bool vs_style,
+                                          const char *strip_path_prefix) {
+  CHECK(!RenderNeedsSymbolization(format));
+  RenderContext(buffer);
+  buffer->AppendF(kFormatFrame, frame_no, address);
+}
+
+bool MarkupSymbolizerTool::SymbolizePC(uptr addr, SymbolizedStack *stack) {
   char buffer[kFormatFunctionMax];
   internal_snprintf(buffer, sizeof(buffer), kFormatFunction, addr);
-  s->info.function = internal_strdup(buffer);
-  return s;
+  stack->info.function = internal_strdup(buffer);
+  return true;
 }
 
-// Always claim we succeeded, so that RenderDataInfo will be called.
-bool Symbolizer::SymbolizeData(uptr addr, DataInfo *info) {
+bool MarkupSymbolizerTool::SymbolizeData(uptr addr, DataInfo *info) {
   info->Clear();
   info->start = addr;
   return true;
 }
 
-class MarkupStackTracePrinter : public StackTracePrinter {
-  // We ignore the format argument to __sanitizer_symbolize_global.
-  void RenderData(InternalScopedString *buffer, const char *format,
-                  const DataInfo *DI, const char *strip_path_prefix) override {
-    buffer->AppendF(kFormatData, DI->start);
-  }
+const char *MarkupSymbolizerTool::Demangle(const char *name) {
+  static char buffer[kFormatDemangleMax];
+  internal_snprintf(buffer, sizeof(buffer), kFormatDemangle, name);
+  return buffer;
+}
 
-  bool RenderNeedsSymbolization(const char *format) override { return false; }
+// Fuchsia's implementation of symbolizer markup doesn't need to emit contextual
+// elements at this point.
+// Fuchsia's logging infrastructure emits enough information about
+// process memory layout that a post-processing filter can do the
+// symbolization and pretty-print the markup.
+#if !SANITIZER_FUCHSIA
+
+// Simplier view of a LoadedModule. It only holds information necessary to
+// identify unique modules.
+struct RenderedModule {
+  char *full_name;
+  uptr base_address;
+  u8 uuid[kModuleUUIDSize];  // BuildId
+};
 
-  // We don't support the stack_trace_format flag at all.
-  void RenderFrame(InternalScopedString *buffer, const char *format,
-                   int frame_no, uptr address, const AddressInfo *info,
-                   bool vs_style, const char *strip_path_prefix) override {
-    CHECK(!RenderNeedsSymbolization(format));
-    buffer->AppendF(kFormatFrame, frame_no, address);
-  }
+static bool ModulesEq(const LoadedModule &module,
+                      const RenderedModule &renderedModule) {
+  return module.base_address() == renderedModule.base_address &&
+         internal_memcmp(module.uuid(), renderedModule.uuid,
+                         module.uuid_size()) == 0 &&
+         internal_strcmp(module.full_name(), renderedModule.full_name) == 0;
+}
 
- protected:
-  ~MarkupStackTracePrinter();
-};
+static bool ModuleHasBeenRendered(
+    const LoadedModule &module,
+    const InternalMmapVectorNoCtor<RenderedModule> &renderedModules) {
+  for (const auto &renderedModule : renderedModules)
+    if (ModulesEq(module, renderedModule))
+      return true;
 
-StackTracePrinter *StackTracePrinter::NewStackTracePrinter() {
-  return new (GetGlobalLowLevelAllocator()) MarkupStackTracePrinter();
+  return false;
 }
 
-Symbolizer *Symbolizer::PlatformInit() {
-  return new (symbolizer_allocator_) Symbolizer({});
+static void RenderModule(InternalScopedString *buffer,
+                         const LoadedModule &module, uptr moduleId) {
+  InternalScopedString buildIdBuffer;
+  for (uptr i = 0; i < module.uuid_size(); i++)
+    buildIdBuffer.AppendF("%02x", module.uuid()[i]);
+
+  buffer->AppendF(kFormatModule, moduleId, module.full_name(),
+                  buildIdBuffer.data());
+  buffer->Append("\n");
 }
 
-void Symbolizer::LateInitialize() { Symbolizer::GetOrInit(); }
+static void RenderMmaps(InternalScopedString *buffer,
+                        const LoadedModule &module, uptr moduleId) {
+  InternalScopedString accessBuffer;
 
-}  // namespace __sanitizer
+  // All module mmaps are readable at least
+  for (const auto &range : module.ranges()) {
+    accessBuffer.Append("r");
+    if (range.writable)
+      accessBuffer.Append("w");
+    if (range.executable)
+      accessBuffer.Append("x");
+
+    //{{{mmap:%starting_addr:%size_in_hex:load:%moduleId:r%(w|x):%relative_addr}}}
+
+    // module.base_address == dlpi_addr
+    // range.beg == dlpi_addr + p_vaddr
+    // relative address == p_vaddr == range.beg - module.base_address
+    buffer->AppendF(kFormatMmap, range.beg, range.end - range.beg, moduleId,
+                    accessBuffer.data(), range.beg - module.base_address());
+
+    buffer->Append("\n");
+    accessBuffer.clear();
+  }
+}
 
-#endif  // SANITIZER_SYMBOLIZER_MARKUP
+void MarkupStackTracePrinter::RenderContext(InternalScopedString *buffer) {
+  // Keeps track of the modules that have been rendered.
+  static bool initialized = false;
+  static InternalMmapVectorNoCtor<RenderedModule> renderedModules;
+  if (!initialized) {
+    // arbitrary initial size, counting the main module plus some important libs
+    // like libc.
+    renderedModules.Initialize(3);
+    initialized = true;
+  }
+
+  if (renderedModules.size() == 0)
+    buffer->Append("{{{reset}}}\n");
+
+  const auto &modules = Symbolizer::GetOrInit()->GetRefreshedListOfModules();
+
+  for (const auto &module : modules) {
+    if (ModuleHasBeenRendered(module, renderedModules))
+      continue;
+
+    // symbolizer markup id, used to refer to this modules from other contextual
+    // elements
+    uptr moduleId = renderedModules.size();
+
+    RenderModule(buffer, module, moduleId);
+    RenderMmaps(buffer, module, moduleId);
+
+    RenderedModule renderedModule{
+        internal_strdup(module.full_name()), module.base_address(), {}};
+
+    // kModuleUUIDSize is the size of curModule.uuid
+    CHECK_GE(kModuleUUIDSize, module.uuid_size());
+    internal_memcpy(renderedModule.uuid, module.uuid(), module.uuid_size());
+    renderedModules.push_back(renderedModule);
+  }
+}
+#endif  // !SANITIZER_FUCHSIA
+
+}  // namespace __sanitizer

compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_markup.h

@@ -0,0 +1,68 @@
+//===-- sanitizer_symbolizer_markup.h -----------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+//  This file is shared between various sanitizers' runtime libraries.
+//
+//  Header for the offline markup symbolizer.
+//===----------------------------------------------------------------------===//
+#ifndef SANITIZER_SYMBOLIZER_MARKUP_H
+#define SANITIZER_SYMBOLIZER_MARKUP_H
+
+#include "sanitizer_common.h"
+#include "sanitizer_stacktrace_printer.h"
+#include "sanitizer_symbolizer.h"
+#include "sanitizer_symbolizer_internal.h"
+
+namespace __sanitizer {
+
+class MarkupStackTracePrinter : public StackTracePrinter {
+ public:
+  // We don't support the stack_trace_format flag at all.
+  void RenderFrame(InternalScopedString *buffer, const char *format,
+                   int frame_no, uptr address, const AddressInfo *info,
+                   bool vs_style, const char *strip_path_prefix = "") override;
+
+  bool RenderNeedsSymbolization(const char *format) override;
+
+  // We ignore the format argument to __sanitizer_symbolize_global.
+  void RenderData(InternalScopedString *buffer, const char *format,
+                  const DataInfo *DI,
+                  const char *strip_path_prefix = "") override;
+
+ private:
+  void RenderContext(InternalScopedString *buffer);
+
+ protected:
+  ~MarkupStackTracePrinter() {}
+};
+
+class MarkupSymbolizerTool final : public SymbolizerTool {
+ public:
+  // This is used in some places for suppression checking, which we
+  // don't really support for Fuchsia.  It's also used in UBSan to
+  // identify a PC location to a function name, so we always fill in
+  // the function member with a string containing markup around the PC
+  // value.
+  // TODO(mcgrathr): Under SANITIZER_GO, it's currently used by TSan
+  // to render stack frames, but that should be changed to use
+  // RenderStackFrame.
+  bool SymbolizePC(uptr addr, SymbolizedStack *stack) override;
+
+  // Always claim we succeeded, so that RenderDataInfo will be called.
+  bool SymbolizeData(uptr addr, DataInfo *info) override;
+
+  // May return NULL if demangling failed.
+  // This is used by UBSan for type names, and by ASan for global variable
+  // names. It's expected to return a static buffer that will be reused on each
+  // call.
+  const char *Demangle(const char *name) override;
+};
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_SYMBOLIZER_MARKUP_H

compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_markup_constants.h

@@ -35,6 +35,12 @@ constexpr const char *kFormatData = "{{{data:%p}}}";
 // One frame in a backtrace (printed on a line by itself).
 constexpr const char *kFormatFrame = "{{{bt:%u:%p}}}";
 
+// Module contextual element.
+constexpr const char *kFormatModule = "{{{module:%d:%s:elf:%s}}}";
+
+// mmap for a module segment.
+constexpr const char *kFormatMmap = "{{{mmap:%p:0x%x:load:%d:%s:0x%x}}}";
+
 // Dump trigger element.
 #define FORMAT_DUMPFILE "{{{dumpfile:%s:%s}}}"