[lldb-dap] Always stop on enrty for attaching

Recently upon debugging a program with thousands of threads, lldb-dap would hang at a `threads` request sent right after receiving the `configurationDone` response. Soon after it will end the debug session with "Process <pid> ex
ited with status = -1 (0xffffffff) lost connection".

This is because LLDB is still in the middle of resuming all the threads. And requesting threads will require stopp
ing the process. From the gdb-remote log it ended up getting `lldb::StateType::eStateInvalid` and just exit with s
tatus -1.

I don't think it's reasonable to allow getting threads from a running process. The alternative will be reject the
`threads` request if the process is not stopped.

Author: kusmour

lldb/test/API/tools/lldb-dap/attach/TestDAP_attach.py

@@ -1,5 +1,5 @@
 """
-Test lldb-dap setBreakpoints request
+Test lldb-dap attach request
 """
 
 
@@ -35,6 +35,9 @@ def set_and_hit_breakpoint(self, continueToExit=True):
         self.assertEqual(
             len(breakpoint_ids), len(lines), "expect correct number of breakpoints"
         )
+        # Send a configurationDone request when process is ready to continue
+        self.dap_server.request_configurationDone()
+        self.assertTrue(self.verify_stop_exception_info("signal SIGSTOP"))
         self.continue_to_breakpoints(breakpoint_ids)
         if continueToExit:
             self.continue_to_exit()
@@ -175,6 +178,9 @@ def test_commands(self):
         functions = ["main"]
         breakpoint_ids = self.set_function_breakpoints(functions)
         self.assertEqual(len(breakpoint_ids), len(functions), "expect one breakpoint")
+        # Execute the configurationDone even if this is not real attaching
+        self.dap_server.request_configurationDone()
+        self.assertTrue(self.verify_stop_exception_info("signal SIGSTOP"))
         self.continue_to_breakpoints(breakpoint_ids)
         output = self.collect_console(timeout_secs=10, pattern=stopCommands[-1])
         self.verify_commands("stopCommands", output, stopCommands)

lldb/test/API/tools/lldb-dap/attach/TestDAP_attachByPortNum.py

@@ -32,6 +32,9 @@ def set_and_hit_breakpoint(self, continueToExit=True):
         self.assertEqual(
             len(breakpoint_ids), len(lines), "expect correct number of breakpoints"
         )
+        # Send a configurationDone request when process is ready to continue
+        self.dap_server.request_configurationDone()
+        self.assertTrue(self.verify_stop_exception_info("signal SIGSTOP"))
         self.continue_to_breakpoints(breakpoint_ids)
         if continueToExit:
             self.continue_to_exit()

lldb/tools/lldb-dap/Handler/AttachRequestHandler.cpp

@@ -73,9 +73,11 @@ void AttachRequestHandler::operator()(const llvm::json::Object &request) const {
   llvm::StringRef core_file = GetString(arguments, "coreFile").value_or("");
   const uint64_t timeout_seconds =
       GetInteger<uint64_t>(arguments, "timeout").value_or(30);
-  dap.stop_at_entry = core_file.empty()
-                          ? GetBoolean(arguments, "stopOnEntry").value_or(false)
-                          : true;
+  // Clients like VS Code sends threads request right after receiving
+  // configurationDone reponse where the process might be resuming.
+  // Getting threads list on a running process is not supported by LLDB.
+  // Always stop the process after attaching.
+  dap.stop_at_entry = true;
   dap.configuration.postRunCommands = GetStrings(arguments, "postRunCommands");
   const llvm::StringRef debuggerRoot =
       GetString(arguments, "debuggerRoot").value_or("");