[lldb] Refactor InstrumentationRuntimeAsan and add a new plugin (#69388)

[lldb] Refactor InstrumentationRuntimeAsan and add a new plugin
InstrumentationRuntimeLibsanitizers.

This commit refactors InstrumentationRuntimeASan by pulling out reusable
code into a separate ReportRetriever class. The purpose of the
refactoring is to allow reuse of the ReportRetriever class in another
plugin.

The commit also adds InstrumentationRuntimeASanLibsanitizers, a new runtime
plugin for ASan. The plugin provides the same
functionality as InstrumentationRuntimeASan, but provides a different
set of symbols/library names to search for while activating the plugin.

rdar://112491689

Author: usama54321

lldb/include/lldb/lldb-enumerations.h

@@ -527,6 +527,7 @@ enum InstrumentationRuntimeType {
   eInstrumentationRuntimeTypeUndefinedBehaviorSanitizer = 0x0002,
   eInstrumentationRuntimeTypeMainThreadChecker = 0x0003,
   eInstrumentationRuntimeTypeSwiftRuntimeReporting = 0x0004,
+  eInstrumentationRuntimeTypeLibsanitizersAsan = 0x0005,
   eNumInstrumentationRuntimeTypes
 };
 

lldb/source/Plugins/InstrumentationRuntime/ASan/CMakeLists.txt

@@ -4,10 +4,7 @@ add_lldb_library(lldbPluginInstrumentationRuntimeASan PLUGIN
   LINK_LIBS
     lldbBreakpoint
     lldbCore
-    lldbExpression
-    lldbInterpreter
     lldbSymbol
     lldbTarget
-  LINK_COMPONENTS
-    Support
+    lldbPluginInstrumentationRuntimeUtility
   )

lldb/source/Plugins/InstrumentationRuntime/ASan/InstrumentationRuntimeASan.cpp

@@ -9,23 +9,14 @@
 #include "InstrumentationRuntimeASan.h"
 
 #include "lldb/Breakpoint/StoppointCallbackContext.h"
-#include "lldb/Core/Debugger.h"
 #include "lldb/Core/Module.h"
 #include "lldb/Core/PluginInterface.h"
 #include "lldb/Core/PluginManager.h"
-#include "lldb/Core/ValueObject.h"
-#include "lldb/Expression/UserExpression.h"
-#include "lldb/Host/StreamFile.h"
-#include "lldb/Interpreter/CommandReturnObject.h"
 #include "lldb/Symbol/Symbol.h"
-#include "lldb/Target/InstrumentationRuntimeStopInfo.h"
-#include "lldb/Target/StopInfo.h"
-#include "lldb/Target/Target.h"
-#include "lldb/Target/Thread.h"
+#include "lldb/Target/Process.h"
 #include "lldb/Utility/RegularExpression.h"
-#include "lldb/Utility/Stream.h"
 
-#include "llvm/ADT/StringSwitch.h"
+#include "Plugins/InstrumentationRuntime/Utility/ReportRetriever.h"
 
 using namespace lldb;
 using namespace lldb_private;
@@ -69,169 +60,6 @@ bool InstrumentationRuntimeASan::CheckIfRuntimeIsValid(
   return symbol != nullptr;
 }
 
-const char *address_sanitizer_retrieve_report_data_prefix = R"(
-extern "C"
-{
-int __asan_report_present();
-void *__asan_get_report_pc();
-void *__asan_get_report_bp();
-void *__asan_get_report_sp();
-void *__asan_get_report_address();
-const char *__asan_get_report_description();
-int __asan_get_report_access_type();
-size_t __asan_get_report_access_size();
-}
-)";
-
-const char *address_sanitizer_retrieve_report_data_command = R"(
-struct {
-    int present;
-    int access_type;
-    void *pc;
-    void *bp;
-    void *sp;
-    void *address;
-    size_t access_size;
-    const char *description;
-} t;
-
-t.present = __asan_report_present();
-t.access_type = __asan_get_report_access_type();
-t.pc = __asan_get_report_pc();
-t.bp = __asan_get_report_bp();
-t.sp = __asan_get_report_sp();
-t.address = __asan_get_report_address();
-t.access_size = __asan_get_report_access_size();
-t.description = __asan_get_report_description();
-t
-)";
-
-StructuredData::ObjectSP InstrumentationRuntimeASan::RetrieveReportData() {
-  ProcessSP process_sp = GetProcessSP();
-  if (!process_sp)
-    return StructuredData::ObjectSP();
-
-  ThreadSP thread_sp =
-      process_sp->GetThreadList().GetExpressionExecutionThread();
-  StackFrameSP frame_sp =
-      thread_sp->GetSelectedFrame(DoNoSelectMostRelevantFrame);
-
-  if (!frame_sp)
-    return StructuredData::ObjectSP();
-
-  EvaluateExpressionOptions options;
-  options.SetUnwindOnError(true);
-  options.SetTryAllThreads(true);
-  options.SetStopOthers(true);
-  options.SetIgnoreBreakpoints(true);
-  options.SetTimeout(process_sp->GetUtilityExpressionTimeout());
-  options.SetPrefix(address_sanitizer_retrieve_report_data_prefix);
-  options.SetAutoApplyFixIts(false);
-  options.SetLanguage(eLanguageTypeObjC_plus_plus);
-
-  ValueObjectSP return_value_sp;
-  ExecutionContext exe_ctx;
-  Status eval_error;
-  frame_sp->CalculateExecutionContext(exe_ctx);
-  ExpressionResults result = UserExpression::Evaluate(
-      exe_ctx, options, address_sanitizer_retrieve_report_data_command, "",
-      return_value_sp, eval_error);
-  if (result != eExpressionCompleted) {
-    StreamString ss;
-    ss << "cannot evaluate AddressSanitizer expression:\n";
-    ss << eval_error.AsCString();
-    Debugger::ReportWarning(ss.GetString().str(),
-                            process_sp->GetTarget().GetDebugger().GetID());
-    return StructuredData::ObjectSP();
-  }
-
-  int present = return_value_sp->GetValueForExpressionPath(".present")
-                    ->GetValueAsUnsigned(0);
-  if (present != 1)
-    return StructuredData::ObjectSP();
-
-  addr_t pc =
-      return_value_sp->GetValueForExpressionPath(".pc")->GetValueAsUnsigned(0);
-  addr_t bp =
-  return_value_sp->GetValueForExpressionPath(".bp")->GetValueAsUnsigned(0);
-  addr_t sp =
-      return_value_sp->GetValueForExpressionPath(".sp")->GetValueAsUnsigned(0);
-  addr_t address = return_value_sp->GetValueForExpressionPath(".address")
-                       ->GetValueAsUnsigned(0);
-  addr_t access_type =
-      return_value_sp->GetValueForExpressionPath(".access_type")
-          ->GetValueAsUnsigned(0);
-  addr_t access_size =
-      return_value_sp->GetValueForExpressionPath(".access_size")
-          ->GetValueAsUnsigned(0);
-  addr_t description_ptr =
-      return_value_sp->GetValueForExpressionPath(".description")
-          ->GetValueAsUnsigned(0);
-  std::string description;
-  Status error;
-  process_sp->ReadCStringFromMemory(description_ptr, description, error);
-
-  StructuredData::Dictionary *dict = new StructuredData::Dictionary();
-  dict->AddStringItem("instrumentation_class", "AddressSanitizer");
-  dict->AddStringItem("stop_type", "fatal_error");
-  dict->AddIntegerItem("pc", pc);
-  dict->AddIntegerItem("bp", bp);
-  dict->AddIntegerItem("sp", sp);
-  dict->AddIntegerItem("address", address);
-  dict->AddIntegerItem("access_type", access_type);
-  dict->AddIntegerItem("access_size", access_size);
-  dict->AddStringItem("description", description);
-
-  return StructuredData::ObjectSP(dict);
-}
-
-std::string
-InstrumentationRuntimeASan::FormatDescription(StructuredData::ObjectSP report) {
-  std::string description = std::string(report->GetAsDictionary()
-                                            ->GetValueForKey("description")
-                                            ->GetAsString()
-                                            ->GetValue());
-  return llvm::StringSwitch<std::string>(description)
-      .Case("heap-use-after-free", "Use of deallocated memory")
-      .Case("heap-buffer-overflow", "Heap buffer overflow")
-      .Case("stack-buffer-underflow", "Stack buffer underflow")
-      .Case("initialization-order-fiasco", "Initialization order problem")
-      .Case("stack-buffer-overflow", "Stack buffer overflow")
-      .Case("stack-use-after-return", "Use of stack memory after return")
-      .Case("use-after-poison", "Use of poisoned memory")
-      .Case("container-overflow", "Container overflow")
-      .Case("stack-use-after-scope", "Use of out-of-scope stack memory")
-      .Case("global-buffer-overflow", "Global buffer overflow")
-      .Case("unknown-crash", "Invalid memory access")
-      .Case("stack-overflow", "Stack space exhausted")
-      .Case("null-deref", "Dereference of null pointer")
-      .Case("wild-jump", "Jump to non-executable address")
-      .Case("wild-addr-write", "Write through wild pointer")
-      .Case("wild-addr-read", "Read from wild pointer")
-      .Case("wild-addr", "Access through wild pointer")
-      .Case("signal", "Deadly signal")
-      .Case("double-free", "Deallocation of freed memory")
-      .Case("new-delete-type-mismatch",
-            "Deallocation size different from allocation size")
-      .Case("bad-free", "Deallocation of non-allocated memory")
-      .Case("alloc-dealloc-mismatch",
-            "Mismatch between allocation and deallocation APIs")
-      .Case("bad-malloc_usable_size", "Invalid argument to malloc_usable_size")
-      .Case("bad-__sanitizer_get_allocated_size",
-            "Invalid argument to __sanitizer_get_allocated_size")
-      .Case("param-overlap",
-            "Call to function disallowing overlapping memory ranges")
-      .Case("negative-size-param", "Negative size used when accessing memory")
-      .Case("bad-__sanitizer_annotate_contiguous_container",
-            "Invalid argument to __sanitizer_annotate_contiguous_container")
-      .Case("odr-violation", "Symbol defined in multiple translation units")
-      .Case(
-          "invalid-pointer-pair",
-          "Comparison or arithmetic on pointers from different memory regions")
-      // for unknown report codes just show the code
-      .Default("AddressSanitizer detected: " + description);
-}
-
 bool InstrumentationRuntimeASan::NotifyBreakpointHit(
     void *baton, StoppointCallbackContext *context, user_id_t break_id,
     user_id_t break_loc_id) {
@@ -244,32 +72,8 @@ bool InstrumentationRuntimeASan::NotifyBreakpointHit(
 
   ProcessSP process_sp = instance->GetProcessSP();
 
-  if (process_sp->GetModIDRef().IsLastResumeForUserExpression())
-    return false;
-
-  StructuredData::ObjectSP report = instance->RetrieveReportData();
-  std::string description;
-  if (report) {
-    description = instance->FormatDescription(report);
-  }
-  // Make sure this is the right process
-  if (process_sp && process_sp == context->exe_ctx_ref.GetProcessSP()) {
-    ThreadSP thread_sp = context->exe_ctx_ref.GetThreadSP();
-    if (thread_sp)
-      thread_sp->SetStopInfo(InstrumentationRuntimeStopInfo::
-                                 CreateStopReasonWithInstrumentationData(
-                                     *thread_sp, description, report));
-
-    StreamFileSP stream_sp(
-        process_sp->GetTarget().GetDebugger().GetOutputStreamSP());
-    if (stream_sp) {
-      stream_sp->Printf("AddressSanitizer report breakpoint hit. Use 'thread "
-                        "info -s' to get extended information about the "
-                        "report.\n");
-    }
-    return true; // Return true to stop the target
-  } else
-    return false; // Let target run
+  return ReportRetriever::NotifyBreakpointHit(process_sp, context, break_id,
+                                              break_loc_id);
 }
 
 void InstrumentationRuntimeASan::Activate() {
@@ -280,29 +84,14 @@ void InstrumentationRuntimeASan::Activate() {
   if (!process_sp)
     return;
 
-  ConstString symbol_name("_ZN6__asanL7AsanDieEv");
-  const Symbol *symbol = GetRuntimeModuleSP()->FindFirstSymbolWithNameAndType(
-      symbol_name, eSymbolTypeCode);
-
-  if (symbol == nullptr)
-    return;
+  Breakpoint *breakpoint = ReportRetriever::SetupBreakpoint(
+      GetRuntimeModuleSP(), process_sp, ConstString("_ZN6__asanL7AsanDieEv"));
 
-  if (!symbol->ValueIsAddress() || !symbol->GetAddressRef().IsValid())
+  if (!breakpoint)
     return;
 
-  Target &target = process_sp->GetTarget();
-  addr_t symbol_address = symbol->GetAddressRef().GetOpcodeLoadAddress(&target);
-
-  if (symbol_address == LLDB_INVALID_ADDRESS)
-    return;
-
-  const bool internal = true;
-  const bool hardware = false;
   const bool sync = false;
-  Breakpoint *breakpoint =
-      process_sp->GetTarget()
-          .CreateBreakpoint(symbol_address, internal, hardware)
-          .get();
+
   breakpoint->SetCallback(InstrumentationRuntimeASan::NotifyBreakpointHit, this,
                           sync);
   breakpoint->SetBreakpointKind("address-sanitizer-report");
@@ -312,12 +101,13 @@ void InstrumentationRuntimeASan::Activate() {
 }
 
 void InstrumentationRuntimeASan::Deactivate() {
-  if (GetBreakpointID() != LLDB_INVALID_BREAK_ID) {
-    ProcessSP process_sp = GetProcessSP();
-    if (process_sp) {
-      process_sp->GetTarget().RemoveBreakpointByID(GetBreakpointID());
-      SetBreakpointID(LLDB_INVALID_BREAK_ID);
-    }
-  }
   SetActive(false);
+
+  if (GetBreakpointID() == LLDB_INVALID_BREAK_ID)
+    return;
+
+  if (ProcessSP process_sp = GetProcessSP()) {
+    process_sp->GetTarget().RemoveBreakpointByID(GetBreakpointID());
+    SetBreakpointID(LLDB_INVALID_BREAK_ID);
+  }
 }

lldb/source/Plugins/InstrumentationRuntime/ASan/InstrumentationRuntimeASan.h

@@ -10,9 +10,6 @@
 #define LLDB_SOURCE_PLUGINS_INSTRUMENTATIONRUNTIME_ASAN_INSTRUMENTATIONRUNTIMEASAN_H
 
 #include "lldb/Target/InstrumentationRuntime.h"
-#include "lldb/Target/Process.h"
-#include "lldb/Utility/StructuredData.h"
-#include "lldb/lldb-private.h"
 
 namespace lldb_private {
 
@@ -51,10 +48,6 @@ class InstrumentationRuntimeASan : public lldb_private::InstrumentationRuntime {
                                   StoppointCallbackContext *context,
                                   lldb::user_id_t break_id,
                                   lldb::user_id_t break_loc_id);
-
-  StructuredData::ObjectSP RetrieveReportData();
-
-  std::string FormatDescription(StructuredData::ObjectSP report);
 };
 
 } // namespace lldb_private

lldb/source/Plugins/InstrumentationRuntime/ASanLibsanitizers/CMakeLists.txt

@@ -0,0 +1,10 @@
+add_lldb_library(lldbPluginInstrumentationRuntimeASanLibsanitizers PLUGIN
+  InstrumentationRuntimeASanLibsanitizers.cpp
+
+  LINK_LIBS
+    lldbBreakpoint
+    lldbCore
+    lldbSymbol
+    lldbTarget
+    lldbPluginInstrumentationRuntimeUtility
+  )