[lldb] Introduce StackFrameRecognizer

This patch introduces a concept of "frame recognizer" and "recognized frame". This should be an extensible mechanism that retrieves information about special frames based on ABI, arguments or other special properties of that frame, even without source code. A few examples where that could be useful could be 1) objc_exception_throw, where we'd like to get the current exception, 2) terminate_with_reason and extracting the current terminate string, 3) recognizing Objective-C frames and automatically extracting the receiver+selector, or perhaps all arguments (based on selector).

Differential Revision: https://reviews.llvm.org/D44603

llvm-svn: 345678

1 files changed, 386 insertions, 0 deletions

diff --git a/lldb/source/Commands/CommandObjectFrame.cpp b/lldb/source/Commands/CommandObjectFrame.cpp
index 5424bcd..0dface6 100644
--- a/lldb/source/Commands/CommandObjectFrame.cpp
+++ b/lldb/source/Commands/CommandObjectFrame.cpp
@@ -24,6 +24,7 @@
 #include "lldb/DataFormatters/ValueObjectPrinter.h"
 #include "lldb/Host/Host.h"
 #include "lldb/Host/OptionParser.h"
+#include "lldb/Host/StringConvert.h"
 #include "lldb/Interpreter/CommandInterpreter.h"
 #include "lldb/Interpreter/CommandReturnObject.h"
 #include "lldb/Interpreter/OptionGroupFormat.h"
@@ -40,6 +41,7 @@
 #include "lldb/Symbol/VariableList.h"
 #include "lldb/Target/Process.h"
 #include "lldb/Target/StackFrame.h"
+#include "lldb/Target/StackFrameRecognizer.h"
 #include "lldb/Target/StopInfo.h"
 #include "lldb/Target/Target.h"
 #include "lldb/Target/Thread.h"
@@ -715,6 +717,23 @@ protected:
       result.SetStatus(eReturnStatusSuccessFinishResult);
     }
 
+    if (m_option_variable.show_recognized_args) {
+      auto recognized_frame = frame->GetRecognizedFrame();
+      if (recognized_frame) {
+        ValueObjectListSP recognized_arg_list =
+            recognized_frame->GetRecognizedArguments();
+        if (recognized_arg_list) {
+          for (auto &rec_value_sp : recognized_arg_list->GetObjects()) {
+            options.SetFormat(m_option_format.GetFormat());
+            options.SetVariableFormatDisplayLanguage(
+                rec_value_sp->GetPreferredDisplayLanguage());
+            options.SetRootValueObjectName(rec_value_sp->GetName().AsCString());
+            rec_value_sp->Dump(result.GetOutputStream(), options);
+          }
+        }
+      }
+    }
+
     if (m_interpreter.TruncationWarningNecessary()) {
       result.GetOutputStream().Printf(m_interpreter.TruncationWarningText(),
                                       m_cmd_name.c_str());
@@ -738,6 +757,368 @@ protected:
   OptionGroupValueObjectDisplay m_varobj_options;
 };
 
+#pragma mark CommandObjectFrameRecognizer
+
+static OptionDefinition g_frame_recognizer_add_options[] = {
+    // clang-format off
+  { LLDB_OPT_SET_ALL, false, "shlib",         's', OptionParser::eRequiredArgument, nullptr, {}, CommandCompletions::eModuleCompletion, eArgTypeShlibName,   "Name of the module or shared library that this recognizer applies to." },
+  { LLDB_OPT_SET_ALL, false, "function",      'n', OptionParser::eRequiredArgument, nullptr, {}, CommandCompletions::eSymbolCompletion, eArgTypeName,        "Name of the function that this recognizer applies to." },
+  { LLDB_OPT_SET_2,   false, "python-class",  'l', OptionParser::eRequiredArgument, nullptr, {}, 0,                                     eArgTypePythonClass, "Give the name of a Python class to use for this frame recognizer." },
+  { LLDB_OPT_SET_ALL, false, "regex",         'x', OptionParser::eNoArgument,       nullptr, {}, 0,                                     eArgTypeNone,        "Function name and module name are actually regular expressions." }
+    // clang-format on
+};
+
+class CommandObjectFrameRecognizerAdd : public CommandObjectParsed {
+private:
+  class CommandOptions : public Options {
+  public:
+    CommandOptions() : Options() {}
+    ~CommandOptions() override = default;
+
+    Status SetOptionValue(uint32_t option_idx, llvm::StringRef option_arg,
+                          ExecutionContext *execution_context) override {
+      Status error;
+      const int short_option = m_getopt_table[option_idx].val;
+
+      switch (short_option) {
+      case 'l':
+        m_class_name = std::string(option_arg);
+        break;
+      case 's':
+        m_module = std::string(option_arg);
+        break;
+      case 'n':
+        m_function = std::string(option_arg);
+        break;
+      case 'x':
+        m_regex = true;
+        break;
+      default:
+        error.SetErrorStringWithFormat("unrecognized option '%c'",
+                                       short_option);
+        break;
+      }
+
+      return error;
+    }
+
+    void OptionParsingStarting(ExecutionContext *execution_context) override {
+      m_module = "";
+      m_function = "";
+      m_class_name = "";
+      m_regex = false;
+    }
+
+    llvm::ArrayRef<OptionDefinition> GetDefinitions() override {
+      return llvm::makeArrayRef(g_frame_recognizer_add_options);
+    }
+
+    // Instance variables to hold the values for command options.
+    std::string m_class_name;
+    std::string m_module;
+    std::string m_function;
+    bool m_regex;
+  };
+
+  CommandOptions m_options;
+
+  Options *GetOptions() override { return &m_options; }
+
+protected:
+  bool DoExecute(Args &command, CommandReturnObject &result) override;
+
+public:
+  CommandObjectFrameRecognizerAdd(CommandInterpreter &interpreter)
+      : CommandObjectParsed(interpreter, "frame recognizer add",
+                            "Add a new frame recognizer.", nullptr),
+        m_options() {
+    SetHelpLong(R"(
+Frame recognizers allow for retrieving information about special frames based on
+ABI, arguments or other special properties of that frame, even without source
+code or debug info. Currently, one use case is to extract function arguments
+that would otherwise be unaccesible, or augment existing arguments.
+
+Adding a custom frame recognizer is possible by implementing a Python class
+and using the 'frame recognizer add' command. The Python class should have a
+'get_recognized_arguments' method and it will receive an argument of type
+lldb.SBFrame representing the current frame that we are trying to recognize.
+The method should return a (possibly empty) list of lldb.SBValue objects that
+represent the recognized arguments.
+
+An example of a recognizer that retrieves the file descriptor values from libc
+functions 'read', 'write' and 'close' follows:
+
+  class LibcFdRecognizer(object):
+    def get_recognized_arguments(self, frame):
+      if frame.name in ["read", "write", "close"]:
+        fd = frame.EvaluateExpression("$arg1").unsigned
+        value = lldb.target.CreateValueFromExpression("fd", "(int)%d" % fd)
+        return [value]
+      return []
+
+The file containing this implementation can be imported via 'command script
+import' and then we can register this recognizer with 'frame recognizer add'.
+It's important to restrict the recognizer to the libc library (which is
+libsystem_kernel.dylib on macOS) to avoid matching functions with the same name
+in other modules:
+
+(lldb) command script import .../fd_recognizer.py
+(lldb) frame recognizer add -l fd_recognizer.LibcFdRecognizer -n read -s libsystem_kernel.dylib
+
+When the program is stopped at the beginning of the 'read' function in libc, we
+can view the recognizer arguments in 'frame variable':
+
+(lldb) b read
+(lldb) r
+Process 1234 stopped
+* thread #1, queue = 'com.apple.main-thread', stop reason = breakpoint 1.3
+    frame #0: 0x00007fff06013ca0 libsystem_kernel.dylib`read
+(lldb) frame variable
+(int) fd = 3
+
+    )");
+  }
+  ~CommandObjectFrameRecognizerAdd() override = default;
+};
+
+bool CommandObjectFrameRecognizerAdd::DoExecute(Args &command,
+                                                CommandReturnObject &result) {
+  if (m_options.m_class_name.empty()) {
+    result.AppendErrorWithFormat(
+        "%s needs a Python class name (-l argument).\n", m_cmd_name.c_str());
+    result.SetStatus(eReturnStatusFailed);
+    return false;
+  }
+
+  if (m_options.m_module.empty()) {
+    result.AppendErrorWithFormat("%s needs a module name (-s argument).\n",
+                                 m_cmd_name.c_str());
+    result.SetStatus(eReturnStatusFailed);
+    return false;
+  }
+
+  if (m_options.m_function.empty()) {
+    result.AppendErrorWithFormat("%s needs a function name (-n argument).\n",
+                                 m_cmd_name.c_str());
+    result.SetStatus(eReturnStatusFailed);
+    return false;
+  }
+
+  ScriptInterpreter *interpreter = m_interpreter.GetScriptInterpreter();
+
+  if (interpreter &&
+      !interpreter->CheckObjectExists(m_options.m_class_name.c_str())) {
+    result.AppendWarning(
+        "The provided class does not exist - please define it "
+        "before attempting to use this frame recognizer");
+  }
+
+  StackFrameRecognizerSP recognizer_sp =
+      StackFrameRecognizerSP(new ScriptedStackFrameRecognizer(
+          interpreter, m_options.m_class_name.c_str()));
+  if (m_options.m_regex) {
+    auto module =
+        RegularExpressionSP(new RegularExpression(m_options.m_module));
+    auto func =
+        RegularExpressionSP(new RegularExpression(m_options.m_function));
+    StackFrameRecognizerManager::AddRecognizer(recognizer_sp, module, func);
+  } else {
+    auto module = ConstString(m_options.m_module);
+    auto func = ConstString(m_options.m_function);
+    StackFrameRecognizerManager::AddRecognizer(recognizer_sp, module, func);
+  }
+
+  result.SetStatus(eReturnStatusSuccessFinishNoResult);
+  return result.Succeeded();
+}
+
+class CommandObjectFrameRecognizerClear : public CommandObjectParsed {
+public:
+  CommandObjectFrameRecognizerClear(CommandInterpreter &interpreter)
+      : CommandObjectParsed(interpreter, "frame recognizer clear",
+                           "Delete all frame recognizers.", nullptr) {}
+
+  ~CommandObjectFrameRecognizerClear() override = default;
+
+protected:
+  bool DoExecute(Args &command, CommandReturnObject &result) override {
+    StackFrameRecognizerManager::RemoveAllRecognizers();
+    result.SetStatus(eReturnStatusSuccessFinishResult);
+    return result.Succeeded();
+  }
+};
+
+class CommandObjectFrameRecognizerDelete : public CommandObjectParsed {
+ public:
+  CommandObjectFrameRecognizerDelete(CommandInterpreter &interpreter)
+      : CommandObjectParsed(interpreter, "frame recognizer delete",
+                            "Delete an existing frame recognizer.", nullptr) {}
+
+  ~CommandObjectFrameRecognizerDelete() override = default;
+
+ protected:
+  bool DoExecute(Args &command, CommandReturnObject &result) override {
+    if (command.GetArgumentCount() == 0) {
+      if (!m_interpreter.Confirm(
+              "About to delete all frame recognizers, do you want to do that?",
+              true)) {
+        result.AppendMessage("Operation cancelled...");
+        result.SetStatus(eReturnStatusFailed);
+        return false;
+      }
+
+      StackFrameRecognizerManager::RemoveAllRecognizers();
+      result.SetStatus(eReturnStatusSuccessFinishResult);
+      return result.Succeeded();
+    }
+
+    if (command.GetArgumentCount() != 1) {
+      result.AppendErrorWithFormat("'%s' takes zero or one arguments.\n",
+                                   m_cmd_name.c_str());
+      result.SetStatus(eReturnStatusFailed);
+      return false;
+    }
+
+    uint32_t recognizer_id =
+        StringConvert::ToUInt32(command.GetArgumentAtIndex(0), 0, 0);
+
+    StackFrameRecognizerManager::RemoveRecognizerWithID(recognizer_id);
+    result.SetStatus(eReturnStatusSuccessFinishResult);
+    return result.Succeeded();
+  }
+};
+
+class CommandObjectFrameRecognizerList : public CommandObjectParsed {
+ public:
+  CommandObjectFrameRecognizerList(CommandInterpreter &interpreter)
+      : CommandObjectParsed(interpreter, "frame recognizer list",
+                            "Show a list of active frame recognizers.",
+                            nullptr) {}
+
+  ~CommandObjectFrameRecognizerList() override = default;
+
+ protected:
+  bool DoExecute(Args &command, CommandReturnObject &result) override {
+    bool any_printed = false;
+    StackFrameRecognizerManager::ForEach(
+        [&result, &any_printed](uint32_t recognizer_id, std::string name,
+                                std::string function, std::string symbol,
+                                bool regexp) {
+          if (name == "") name = "(internal)";
+          result.GetOutputStream().Printf(
+              "%d: %s, module %s, function %s%s\n", recognizer_id, name.c_str(),
+              function.c_str(), symbol.c_str(), regexp ? " (regexp)" : "");
+          any_printed = true;
+        });
+
+    if (any_printed)
+      result.SetStatus(eReturnStatusSuccessFinishResult);
+    else {
+      result.GetOutputStream().PutCString("no matching results found.\n");
+      result.SetStatus(eReturnStatusSuccessFinishNoResult);
+    }
+    return result.Succeeded();
+  }
+};
+
+class CommandObjectFrameRecognizerInfo : public CommandObjectParsed {
+ public:
+  CommandObjectFrameRecognizerInfo(CommandInterpreter &interpreter)
+      : CommandObjectParsed(
+            interpreter, "frame recognizer info",
+            "Show which frame recognizer is applied a stack frame (if any).",
+            nullptr) {
+    CommandArgumentEntry arg;
+    CommandArgumentData index_arg;
+
+    // Define the first (and only) variant of this arg.
+    index_arg.arg_type = eArgTypeFrameIndex;
+    index_arg.arg_repetition = eArgRepeatPlain;
+
+    // There is only one variant this argument could be; put it into the
+    // argument entry.
+    arg.push_back(index_arg);
+
+    // Push the data for the first argument into the m_arguments vector.
+    m_arguments.push_back(arg);
+  }
+
+  ~CommandObjectFrameRecognizerInfo() override = default;
+
+ protected:
+  bool DoExecute(Args &command, CommandReturnObject &result) override {
+    Process *process = m_exe_ctx.GetProcessPtr();
+    if (process == nullptr) {
+      result.AppendError("no process");
+      result.SetStatus(eReturnStatusFailed);
+      return false;
+    }
+    Thread *thread = m_exe_ctx.GetThreadPtr();
+    if (thread == nullptr) {
+      result.AppendError("no thread");
+      result.SetStatus(eReturnStatusFailed);
+      return false;
+    }
+    if (command.GetArgumentCount() != 1) {
+      result.AppendErrorWithFormat(
+          "'%s' takes exactly one frame index argument.\n", m_cmd_name.c_str());
+      result.SetStatus(eReturnStatusFailed);
+      return false;
+    }
+
+    uint32_t frame_index =
+        StringConvert::ToUInt32(command.GetArgumentAtIndex(0), 0, 0);
+    StackFrameSP frame_sp = thread->GetStackFrameAtIndex(frame_index);
+    if (!frame_sp) {
+      result.AppendErrorWithFormat("no frame with index %u", frame_index);
+      result.SetStatus(eReturnStatusFailed);
+      return false;
+    }
+
+    auto recognizer =
+        StackFrameRecognizerManager::GetRecognizerForFrame(frame_sp);
+
+    Stream &output_stream = result.GetOutputStream();
+    output_stream.Printf("frame %d ", frame_index);
+    if (recognizer) {
+      output_stream << "is recognized by ";
+      output_stream << recognizer->GetName();
+    } else {
+      output_stream << "not recognized by any recognizer";
+    }
+    output_stream.EOL();
+    result.SetStatus(eReturnStatusSuccessFinishResult);
+    return result.Succeeded();
+  }
+};
+
+class CommandObjectFrameRecognizer : public CommandObjectMultiword {
+ public:
+  CommandObjectFrameRecognizer(CommandInterpreter &interpreter)
+      : CommandObjectMultiword(
+            interpreter, "frame recognizer",
+            "Commands for editing and viewing frame recognizers.",
+            "frame recognizer [<sub-command-options>] ") {
+    LoadSubCommand(
+        "add",
+        CommandObjectSP(new CommandObjectFrameRecognizerAdd(interpreter)));
+    LoadSubCommand(
+        "clear",
+        CommandObjectSP(new CommandObjectFrameRecognizerClear(interpreter)));
+    LoadSubCommand(
+        "delete",
+        CommandObjectSP(new CommandObjectFrameRecognizerDelete(interpreter)));
+    LoadSubCommand(
+        "list",
+        CommandObjectSP(new CommandObjectFrameRecognizerList(interpreter)));
+    LoadSubCommand(
+        "info",
+        CommandObjectSP(new CommandObjectFrameRecognizerInfo(interpreter)));
+  }
+
+  ~CommandObjectFrameRecognizer() override = default;
+};
+
 #pragma mark CommandObjectMultiwordFrame
 
 //-------------------------------------------------------------------------
@@ -758,6 +1139,11 @@ CommandObjectMultiwordFrame::CommandObjectMultiwordFrame(
                  CommandObjectSP(new CommandObjectFrameSelect(interpreter)));
   LoadSubCommand("variable",
                  CommandObjectSP(new CommandObjectFrameVariable(interpreter)));
+#ifndef LLDB_DISABLE_PYTHON
+  LoadSubCommand(
+      "recognizer",
+      CommandObjectSP(new CommandObjectFrameRecognizer(interpreter)));
+#endif
 }
 
 CommandObjectMultiwordFrame::~CommandObjectMultiwordFrame() = default;