Add the ability to debug through an exec into ld

[clayborg]

Summary

A previous commit enabled LLDB to be able to debug a program launched via ld: https://reviews.llvm.org/D108061.

This commit adds the ability to debug a program launched via ld when it happens during an exec into the dynamic loader. There was an issue where after the exec we would locate the rendezvous structure right away but it didn't contain any valid values and we would try to set the dyanamic loader breakpoint at address zero. This patch fixes that and adds a test.

[yinghuitan]

lldb/source/Plugins/DynamicLoader/POSIX-DYLD/DynamicLoaderPOSIXDYLD.cpp

312

Since zero break address is considered invalid, maybe move m_rendezvous.GetBreakAddress() != 0 check into DYLDRendezvous::IsValid() method?

[clayborg]

lldb/source/Plugins/DynamicLoader/POSIX-DYLD/DynamicLoaderPOSIXDYLD.cpp

312

I wanted to be safe and not mess up anyone else that might be using m_rendezvous.IsValid() as it isn't super clear what IsValid() should return. So to be super safe and not affect any other code, I would prefer to leave it this way unless one of the original authors of this file has a different opinion.

[wallace]

nice!

lldb/source/Plugins/DynamicLoader/POSIX-DYLD/DynamicLoaderPOSIXDYLD.cpp

LLDB_LOG(log,
              m_dyld_bid,
              m_process ? m_process->GetID() : LLDB_INVALID_PROCESS_ID);
     return true;
   }
 
   addr_t break_addr;
   Target &target = m_process->GetTarget();
   BreakpointSP dyld_break;
-  if (m_rendezvous.IsValid()) {
+  if (m_rendezvous.IsValid() && m_rendezvous.GetBreakAddress() != 0) {
     break_addr = m_rendezvous.GetBreakAddress();
     LLDB_LOG(log, "Setting rendezvous break address for pid {0} at {1:x}",
              m_process ? m_process->GetID() : LLDB_INVALID_PROCESS_ID,
              break_addr);
     dyld_break = target.CreateBreakpoint(break_addr, true, false);
   } else {
     LLDB_LOG(log, "Rendezvous structure is not set up yet. "
                   "Trying to locate rendezvous breakpoint in the interpreter "
if (status.Fail() || info.GetMapped() != MemoryRegionInfo::eYes ||
     Log *log = GetLog(LLDBLog::DynamicLoader);
     LLDB_LOG(log, "Failed to get interpreter region info: {0}", status);
     return nullptr;
   }
 
   FileSpec file(info.GetName().GetCString());
   ModuleSpec module_spec(file, target.GetArchitecture());
 
-  if (ModuleSP module_sp = target.GetOrCreateModule(module_spec, 
+  if (ModuleSP module_sp = target.GetOrCreateModule(module_spec,
                                                     true /* notify */)) {
     UpdateLoadedSections(module_sp, LLDB_INVALID_ADDRESS, m_interpreter_base,
                          false);
     m_interpreter_module = module_sp;
     return module_sp;
   }
   return nullptr;
 }

lldb/test/API/functionalities/dyld-exec-linux/Makefile

+CXX_SOURCES := main.cpp
+include Makefile.rules

lldb/test/API/functionalities/dyld-exec-linux/TestDyldExecLinux.py

+"""
+Test that LLDB can launch a linux executable and then execs into the dynamic
+loader into this program again.
+"""
+
+import lldb
+import os
+
+from lldbsuite.test.decorators import *
+from lldbsuite.test.lldbtest import *
+from lldbsuite.test import lldbutil
+
+class TestLinux64ExecViaDynamicLoader(TestBase):
+    mydir = TestBase.compute_mydir(__file__)
+
+    @skipIf(oslist=no_match(['linux']))
+    @no_debug_info_test
+    @skipIf(oslist=["linux"], archs=["arm"])
+    def test(self):
+        self.build()
+
+        # Extracts path of the interpreter.
+        exe = self.getBuildArtifact("a.out")
+
+        spec = lldb.SBModuleSpec()
+        spec.SetFileSpec(lldb.SBFileSpec(exe))
+        interp_section = lldb.SBModule(spec).FindSection(".interp")
+        if not interp_section:
+          return
+        section_data = interp_section.GetSectionData()
+        error = lldb.SBError()
+        dyld_path = section_data.GetString(error,0)
+        if error.Fail():
+          return
+
+        target = self.dbg.CreateTarget(exe)
+        self.assertTrue(target, VALID_TARGET)
+
+        # Set a breakpoint in the main function that will get hit after the
+        # program exec's via the dynamic loader. The breakpoint will only get
+        # hit if we can successfully read the shared library lists in the
+        # DynamicLoaderPOSIXDYLD.cpp when we exec into the dynamic loader.
+        breakpoint_main = target.BreakpointCreateBySourceRegex("// Break here", lldb.SBFileSpec("main.cpp"))
+        # Setup our launch info to supply the dynamic loader path to the
+        # program so it gets two args:
+        # - path to a.out
+        # - path to dynamic loader
+        launch_info = lldb.SBLaunchInfo([dyld_path])
+        error = lldb.SBError()
+        process = target.Launch(launch_info, error)
+        self.assertSuccess(error)
+
+        threads = lldbutil.get_stopped_threads(process, lldb.eStopReasonExec)
+        self.assertEqual(len(threads), 1, "We got a thread stopped for exec.")
+
+        process.Continue();
+
+        # Stopped on main here.
+        self.assertEqual(process.GetState(), lldb.eStateStopped)
+        thread = process.GetSelectedThread()
+        self.assertIn("main", thread.GetFrameAtIndex(0).GetDisplayFunctionName())

lldb/test/API/functionalities/dyld-exec-linux/main.cpp

+#include <unistd.h>
+
+int main(int argc, const char *argv[], const char *envp[]) {
+  if (argc == 2) {
+    // If we have two arguments the first is the path to this executable,
+    // the second is the path to the linux dynamic loader that we should
+    // exec with. We want to re-run this problem under the dynamic loader
+    // and make sure we can hit the breakpoint in the "else".
+    const char *interpreter = argv[1];
+    const char *this_program = argv[0];
+    const char *exec_argv[3] = {interpreter, this_program, nullptr};
+    execve(interpreter, (char *const *)exec_argv, (char *const *)envp);
+  }
+  // Break here
+  return 0;
+}