[review,v2] Handle pending stops from the Windows kernel

Message ID 20191126171131.93A2F28174@gnutoolchain-gerrit.osci.io
State New, archived
Headers

Commit Message

Simon Marchi (Code Review) Nov. 26, 2019, 5:11 p.m. UTC
  Change URL: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/414
......................................................................

Handle pending stops from the Windows kernel

PR gdb/22992 concerns an assertion failure in gdb when debugging a
certain inferior:

    int finish_step_over(execution_control_state*): Assertion `ecs->event_thread->control.trap_expected' failed.

Initially the investigation centered on the discovery that gdb was not
suspending other threads when attempting to single-step.  This
oversight is corrected in this patch: now, when stepping a thread, gdb
will call SuspendThread on all other threads.

However, the bug persisted even after this change.  In particular,
WaitForDebugEvent could see a stop for a thread that was ostensibly
suspended.  Our theory of what is happening here is that there are
actually simultaneous breakpoint hits, and the Windows kernel queues
the events, causing the second stop to be reported on a suspended
thread.

In Windows 10 or later gdb could use the DBG_REPLY_LATER flag to
ContinueDebugEvent to request that such events be re-reported later.
However, relying on that did not seem advisable, so this patch instead
arranges to queue such "pending" stops, and then to report them later,
once the step has completed.

In the PR, Pedro pointed out that it's best in this scenario to
implement the stopped_by_sw_breakpoint method, so this patch does this
as well.

gdb/ChangeLog
2019-11-26  Tom Tromey  <tromey@adacore.com>

	PR gdb/22992
	* windows-nat.c (current_event): Update comment.
	(last_wait_event, desired_stop_thread_id): New globals.
	(struct pending_stop): New.
	(pending_stops): New global.
	(windows_nat_target) <stopped_by_sw_breakpoint>
	<supports_stopped_by_sw_breakpoint>: New methods.
	(windows_fetch_one_register): Add assertions.  Adjust PC.
	(windows_continue): Handle pending stops.  Suspend other threads
	when stepping.  Use last_wait_event
	(wait_for_debug_event): New function.
	(get_windows_debug_event): Use wait_for_debug_event.  Handle
	pending stops.  Queue spurious stops.
	(windows_nat_target::wait): Set stopped_at_breakpoint.
	(windows_nat_target::kill): Use wait_for_debug_event.
	* nat/windows-nat.h (struct windows_thread_info)
	<stopped_at_breakpoint>: New field.
	* nat/windows-nat.c (windows_thread_info::resume): Clear
	stopped_at_breakpoint.

Change-Id: I5d0dd64c3d2d8220b534d3e02aeaa6f6815264ab
---
M gdb/ChangeLog
M gdb/nat/windows-nat.c
M gdb/nat/windows-nat.h
M gdb/windows-nat.c
4 files changed, 214 insertions(+), 13 deletions(-)
  

Patch

diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index 14cb238..bb441a1 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,5 +1,27 @@ 
 2019-11-26  Tom Tromey  <tromey@adacore.com>
 
+	PR gdb/22992
+	* windows-nat.c (current_event): Update comment.
+	(last_wait_event, desired_stop_thread_id): New globals.
+	(struct pending_stop): New.
+	(pending_stops): New global.
+	(windows_nat_target) <stopped_by_sw_breakpoint>
+	<supports_stopped_by_sw_breakpoint>: New methods.
+	(windows_fetch_one_register): Add assertions.  Adjust PC.
+	(windows_continue): Handle pending stops.  Suspend other threads
+	when stepping.  Use last_wait_event
+	(wait_for_debug_event): New function.
+	(get_windows_debug_event): Use wait_for_debug_event.  Handle
+	pending stops.  Queue spurious stops.
+	(windows_nat_target::wait): Set stopped_at_breakpoint.
+	(windows_nat_target::kill): Use wait_for_debug_event.
+	* nat/windows-nat.h (struct windows_thread_info)
+	<stopped_at_breakpoint>: New field.
+	* nat/windows-nat.c (windows_thread_info::resume): Clear
+	stopped_at_breakpoint.
+
+2019-11-26  Tom Tromey  <tromey@adacore.com>
+
 	* windows-nat.c (enum thread_disposition_type): New.
 	(thread_rec): Replace "get_context" parameter with "disposition";
 	change type.
diff --git a/gdb/nat/windows-nat.c b/gdb/nat/windows-nat.c
index a98ff42..59b0fac 100644
--- a/gdb/nat/windows-nat.c
+++ b/gdb/nat/windows-nat.c
@@ -49,6 +49,8 @@ 
 {
   if (suspended > 0)
     {
+      stopped_at_breakpoint = false;
+
       if (ResumeThread (h) == (DWORD) -1)
 	{
 	  DWORD err = GetLastError ();
diff --git a/gdb/nat/windows-nat.h b/gdb/nat/windows-nat.h
index 488c9fb..d64ef8e 100644
--- a/gdb/nat/windows-nat.h
+++ b/gdb/nat/windows-nat.h
@@ -71,6 +71,10 @@ 
      inferior thread.  */
   bool reload_context = false;
 
+  /* True if this thread is currently stopped at a breakpoint.  This
+     is used to offset the PC when needed.  */
+  bool stopped_at_breakpoint = false;
+
   /* The name of the thread, allocated by xmalloc.  */
   gdb::unique_xmalloc_ptr<char> name;
 };
diff --git a/gdb/windows-nat.c b/gdb/windows-nat.c
index abd9d55..8448d1c 100644
--- a/gdb/windows-nat.c
+++ b/gdb/windows-nat.c
@@ -218,8 +218,17 @@ 
 
 /* The process and thread handles for the above context.  */
 
-static DEBUG_EVENT current_event;	/* The current debug event from
-					   WaitForDebugEvent */
+/* The current debug event from WaitForDebugEvent or from a pending
+   stop.  */
+static DEBUG_EVENT current_event;
+
+/* The most recent event from WaitForDebugEvent.  Unlike
+   current_event, this is guaranteed never to come from a pending
+   stop.  This is important because only data from the most recent
+   event from WaitForDebugEvent can be used when calling
+   ContinueDebugEvent.  */
+static DEBUG_EVENT last_wait_event;
+
 static HANDLE current_process_handle;	/* Currently executing process */
 static windows_thread_info *current_thread;	/* Info on currently selected thread */
 
@@ -289,6 +298,37 @@ 
 
 #endif /* 0 */
 
+/* The ID of the thread for which we anticipate a stop event.
+   Normally this is -1, meaning we'll accept an event in any
+   thread.  */
+static DWORD desired_stop_thread_id = -1;
+
+/* A single pending stop.  See "pending_stops" for more
+   information.  */
+struct pending_stop
+{
+  /* The thread id.  */
+  DWORD thread_id;
+
+  /* The target waitstatus we computed.  */
+  target_waitstatus status;
+
+  /* The event.  A few fields of this can be referenced after a stop,
+     and it seemed simplest to store the entire event.  */
+  DEBUG_EVENT event;
+};
+
+/* A vector of pending stops.  Sometimes, Windows will report a stop
+   on a thread that has been ostensibly suspended.  We believe what
+   happens here is that two threads hit a breakpoint simultaneously,
+   and the Windows kernel queues the stop events.  However, this can
+   result in the strange effect of trying to single step thread A --
+   leaving all other threads suspended -- and then seeing a stop in
+   thread B.  To handle this scenario, we queue all such "pending"
+   stops here, and then process them once the step has completed.  See
+   PR gdb/22992.  */
+static std::vector<pending_stop> pending_stops;
+
 struct windows_nat_target final : public x86_nat_target<inf_child_target>
 {
   void close () override;
@@ -307,6 +347,16 @@ 
   void fetch_registers (struct regcache *, int) override;
   void store_registers (struct regcache *, int) override;
 
+  bool stopped_by_sw_breakpoint () override
+  {
+    return current_thread->stopped_at_breakpoint;
+  }
+
+  bool supports_stopped_by_sw_breakpoint () override
+  {
+    return true;
+  }
+
   enum target_xfer_status xfer_partial (enum target_object object,
 					const char *annex,
 					gdb_byte *readbuf,
@@ -542,6 +592,10 @@ 
   struct gdbarch *gdbarch = regcache->arch ();
   struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
 
+  gdb_assert (!gdbarch_read_pc_p (gdbarch));
+  gdb_assert (gdbarch_pc_regnum (gdbarch) >= 0);
+  gdb_assert (!gdbarch_write_pc_p (gdbarch));
+
   if (r == I387_FISEG_REGNUM (tdep))
     {
       long l = *((long *) context_offset) & 0xffff;
@@ -561,7 +615,28 @@ 
       regcache->raw_supply (r, (char *) &l);
     }
   else
-    regcache->raw_supply (r, context_offset);
+    {
+      if (th->stopped_at_breakpoint && r == gdbarch_pc_regnum (gdbarch))
+	{
+	  int size = register_size (gdbarch, r);
+	  if (size == 4)
+	    {
+	      uint32_t value;
+	      memcpy (&value, context_offset, size);
+	      value -= gdbarch_decr_pc_after_break (gdbarch);
+	      memcpy (context_offset, &value, size);
+	    }
+	  else
+	    {
+	      gdb_assert (size == 8);
+	      uint64_t value;
+	      memcpy (&value, context_offset, size);
+	      value -= gdbarch_decr_pc_after_break (gdbarch);
+	      memcpy (context_offset, &value, size);
+	    }
+	}
+      regcache->raw_supply (r, context_offset);
+    }
 }
 
 void
@@ -1297,16 +1372,36 @@ 
 {
   BOOL res;
 
+  desired_stop_thread_id = id;
+
+  /* If there are pending stops, and we might plausibly hit one of
+     them, we don't want to actually continue the inferior -- we just
+     want to report the stop.  In this case, we just pretend to
+     continue.  See the comment by the definition of "pending_stops"
+     for details on why this is needed.  */
+  for (const auto &item : pending_stops)
+    {
+      if (desired_stop_thread_id == -1
+	  || desired_stop_thread_id == item.thread_id)
+	{
+	  DEBUG_EVENTS (("windows_continue - pending stop anticipated, "
+			 "desired=0x%x, item=0x%x\n",
+			 desired_stop_thread_id, item.thread_id));
+	  return TRUE;
+	}
+    }
+
   DEBUG_EVENTS (("ContinueDebugEvent (cpid=%d, ctid=0x%x, %s);\n",
-		  (unsigned) current_event.dwProcessId,
-		  (unsigned) current_event.dwThreadId,
+		  (unsigned) last_wait_event.dwProcessId,
+		  (unsigned) last_wait_event.dwThreadId,
 		  continue_status == DBG_CONTINUE ?
 		  "DBG_CONTINUE" : "DBG_EXCEPTION_NOT_HANDLED"));
 
   for (windows_thread_info *th : thread_list)
-    if ((id == -1 || id == (int) th->tid)
-	&& th->suspended)
+    if (id == -1 || id == (int) th->tid)
       {
+	if (!th->suspended)
+	  continue;
 	if (debug_registers_changed)
 	  {
 	    th->context.ContextFlags |= CONTEXT_DEBUG_REGISTERS;
@@ -1333,9 +1428,15 @@ 
 	  }
 	th->resume ();
       }
+    else
+      {
+	/* When single-stepping a specific thread, other threads must
+	   be suspended.  */
+	th->suspend ();
+      }
 
-  res = ContinueDebugEvent (current_event.dwProcessId,
-			    current_event.dwThreadId,
+  res = ContinueDebugEvent (last_wait_event.dwProcessId,
+			    last_wait_event.dwThreadId,
 			    continue_status);
 
   if (!res)
@@ -1487,6 +1588,17 @@ 
   return TRUE;
 }
 
+/* A wrapper for WaitForDebugEvent that sets "last_wait_event"
+   appropriately.  */
+static BOOL
+wait_for_debug_event (DEBUG_EVENT *event, DWORD timeout)
+{
+  BOOL result = WaitForDebugEvent (event, timeout);
+  if (result)
+    last_wait_event = *event;
+  return result;
+}
+
 /* Get the next event from the child.  Returns a non-zero thread id if the event
    requires handling by WFI (or whatever).  */
 static int
@@ -1499,9 +1611,36 @@ 
   static windows_thread_info dummy_thread_info (0, 0, 0);
   DWORD thread_id = 0;
 
+  /* If there is a relevant pending stop, report it now.  See the
+     comment by the definition of "pending_stops" for details on why
+     this is needed.  */
+  for (auto iter = pending_stops.begin ();
+       iter != pending_stops.end ();
+       ++iter)
+    {
+      if (desired_stop_thread_id == -1
+	  || desired_stop_thread_id == iter->thread_id)
+	{
+	  thread_id = iter->thread_id;
+	  *ourstatus = iter->status;
+	  current_event = iter->event;
+
+	  inferior_ptid = ptid_t (current_event.dwProcessId, thread_id, 0);
+	  current_thread = thread_rec (thread_id, INVALIDATE_CONTEXT);
+	  current_thread->reload_context = 1;
+
+	  DEBUG_EVENTS (("get_windows_debug_event - "
+			 "pending stop found in 0x%x (desired=0x%x)\n",
+			 thread_id, desired_stop_thread_id));
+
+	  pending_stops.erase (iter);
+	  return thread_id;
+	}
+    }
+
   last_sig = GDB_SIGNAL_0;
 
-  if (!(debug_event = WaitForDebugEvent (&current_event, 1000)))
+  if (!(debug_event = wait_for_debug_event (&current_event, 1000)))
     goto out;
 
   event_count++;
@@ -1671,7 +1810,27 @@ 
 
   if (!thread_id || saw_create != 1)
     {
-      CHECK (windows_continue (continue_status, -1, 0));
+      CHECK (windows_continue (continue_status, desired_stop_thread_id, 0));
+    }
+  else if (desired_stop_thread_id != -1 && desired_stop_thread_id != thread_id)
+    {
+      /* Pending stop.  See the comment by the definition of
+	 "pending_stops" for details on why this is needed.  */
+      DEBUG_EVENTS (("get_windows_debug_event - "
+		     "unexpected stop in 0x%x (expecting 0x%x)\n",
+		     thread_id, desired_stop_thread_id));
+
+      if (current_event.dwDebugEventCode == EXCEPTION_DEBUG_EVENT
+	  && (current_event.u.Exception.ExceptionRecord.ExceptionCode
+	      == EXCEPTION_BREAKPOINT)
+	  && windows_initialization_done)
+	{
+	  th = thread_rec (thread_id, INVALIDATE_CONTEXT);
+	  th->stopped_at_breakpoint = true;
+	}
+      pending_stops.push_back ({thread_id, *ourstatus, current_event});
+      thread_id = 0;
+      CHECK (windows_continue (continue_status, desired_stop_thread_id, 0));
     }
   else
     {
@@ -1733,7 +1892,21 @@ 
       SetConsoleCtrlHandler (&ctrl_c_handler, FALSE);
 
       if (retval)
-	return ptid_t (current_event.dwProcessId, retval, 0);
+	{
+	  ptid_t result = ptid_t (current_event.dwProcessId, retval, 0);
+
+	  if (current_thread != nullptr)
+	    {
+	      current_thread->stopped_at_breakpoint = false;
+	      if (current_event.dwDebugEventCode == EXCEPTION_DEBUG_EVENT
+		  && (current_event.u.Exception.ExceptionRecord.ExceptionCode
+		      == EXCEPTION_BREAKPOINT)
+		  && windows_initialization_done)
+		current_thread->stopped_at_breakpoint = true;
+	    }
+
+	  return result;
+	}
       else
 	{
 	  int detach = 0;
@@ -2873,7 +3046,7 @@ 
     {
       if (!windows_continue (DBG_CONTINUE, -1, 1))
 	break;
-      if (!WaitForDebugEvent (&current_event, INFINITE))
+      if (!wait_for_debug_event (&current_event, INFINITE))
 	break;
       if (current_event.dwDebugEventCode == EXIT_PROCESS_DEBUG_EVENT)
 	break;