Avoid crash with Ravenscar tasks

When using Ravenscar, gdb can crash if the user sets a breakpoint very
early in task startup.  This happens because gdb thinks the runtime is
initialized, but in practice the particular task isn't sufficiently
initialized.  This patch avoids the issue by turning an assertion into
an early return.

I tested this using the AdaCore internal test suite.  I don't know how
to test Ravenscar using the FSF test suite.

diff --git a/gdb/ravenscar-thread.c b/gdb/ravenscar-thread.c
index 1718c367ff63c568db8c65c7f2f2d696c7a34b08..fd4bf453cac5364b4a1bb80bb490a9bfe78b7193 100644 (file)
--- a/gdb/ravenscar-thread.c
+++ b/gdb/ravenscar-thread.c
@@ -302,10 +302,12 @@ ravenscar_thread_target::add_active_thread ()
   if (!runtime_initialized ())
     return nullptr;
 
-  /* Make sure we set m_base_ptid before calling active_task
-     as the latter relies on it.  */
+  /* It's possible for runtime_initialized to return true but for it
+     not to be fully initialized.  For example, this can happen for a
+     breakpoint placed at the task's beginning.  */
   ptid_t active_ptid = active_task (base_cpu);
-  gdb_assert (active_ptid != null_ptid);
+  if (active_ptid == null_ptid)
+    return nullptr;
 
   /* The running thread may not have been added to
      system.tasking.debug's list yet; so ravenscar_update_thread_list
@@ -437,7 +439,9 @@ ravenscar_thread_target::wait (ptid_t ptid,
     {
       m_base_ptid = event_ptid;
       this->update_thread_list ();
-      return this->add_active_thread ()->ptid;
+      thread_info *thr = this->add_active_thread ();
+      if (thr != nullptr)
+       return thr->ptid;
     }
   return event_ptid;
 }