public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Pedro Alves <pedro@palves.net>
To: gdb-patches@sourceware.org
Cc: Andrew Burgess <aburgess@redhat.com>
Subject: [FYI/pushed v4 25/25] Cancel execution command on thread exit, when stepping, nexting, etc.
Date: Mon, 13 Nov 2023 15:04:27 +0000	[thread overview]
Message-ID: <20231113150427.477431-26-pedro@palves.net> (raw)
In-Reply-To: <20231113150427.477431-1-pedro@palves.net>

If your target has no support for TARGET_WAITKIND_NO_RESUMED events
(and no way to support them, such as the yet-unsubmitted AMDGPU
target), and you step over thread exit with scheduler-locking on, this
is what you get:

 (gdb) n
 [Thread ... exited]
 *hang*

Getting back the prompt by typing Ctrl-C may not even work, since no
inferior thread is running to receive the SIGINT.  Even if it works,
it seems unnecessarily harsh.  If you started an execution command for
which there's a clear thread of interest (step, next, until, etc.),
and that thread disappears, then I think it's more user friendly if
GDB just detects the situation and aborts the command, giving back the
prompt.

That is what this commit implements.  It does this by explicitly
requesting the target to report thread exit events whenever the main
resumed thread has a thread_fsm.  Note that unlike stepping over a
breakpoint, we don't need to enable clone events in this case.

With this patch, we get:

 (gdb) n
 [Thread 0x7ffff7d89700 (LWP 3961883) exited]
 Command aborted, thread exited.
 (gdb)

Reviewed-By: Andrew Burgess <aburgess@redhat.com>
Change-Id: I901ab64c91d10830590b2dac217b5264635a2b95
---
 gdb/infrun.c                                  | 73 ++++++++++++++---
 .../gdb.threads/step-over-thread-exit.exp     | 81 +++++++++++++------
 2 files changed, 118 insertions(+), 36 deletions(-)

diff --git a/gdb/infrun.c b/gdb/infrun.c
index 943ea88538c..62b306ff347 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -1956,6 +1956,22 @@ displaced_step_prepare (thread_info *thread)
   return status;
 }
 
+/* True if any thread of TARGET that matches RESUME_PTID requires
+   target_thread_events enabled.  This assumes TARGET does not support
+   target thread options.  */
+
+static bool
+any_thread_needs_target_thread_events (process_stratum_target *target,
+				       ptid_t resume_ptid)
+{
+  for (thread_info *tp : all_non_exited_threads (target, resume_ptid))
+    if (displaced_step_in_progress_thread (tp)
+	|| schedlock_applies (tp)
+	|| tp->thread_fsm () != nullptr)
+      return true;
+  return false;
+}
+
 /* Maybe disable thread-{cloned,created,exited} event reporting after
    a step-over (either in-line or displaced) finishes.  */
 
@@ -1979,9 +1995,10 @@ update_thread_events_after_step_over (thread_info *event_thread,
   else
     {
       /* We can only control the target-wide target_thread_events
-	 setting.  Disable it, but only if other threads don't need it
-	 enabled.  */
-      if (!displaced_step_in_progress_any_thread ())
+	 setting.  Disable it, but only if other threads in the target
+	 don't need it enabled.  */
+      process_stratum_target *target = event_thread->inf->process_target ();
+      if (!any_thread_needs_target_thread_events (target, minus_one_ptid))
 	target_thread_events (false);
     }
 }
@@ -2559,12 +2576,25 @@ do_target_resume (ptid_t resume_ptid, bool step, enum gdb_signal sig)
       else
 	target_thread_events (true);
     }
+  else if (tp->thread_fsm () != nullptr)
+    {
+      gdb_thread_options options = GDB_THREAD_OPTION_EXIT;
+      if (target_supports_set_thread_options (options))
+	tp->set_thread_options (options);
+      else
+	target_thread_events (true);
+    }
   else
     {
       if (target_supports_set_thread_options (0))
 	tp->set_thread_options (0);
-      else if (!displaced_step_in_progress_any_thread ())
-	target_thread_events (false);
+      else
+	{
+	  process_stratum_target *resume_target = tp->inf->process_target ();
+	  if (!any_thread_needs_target_thread_events (resume_target,
+						      resume_ptid))
+	    target_thread_events (false);
+	}
     }
 
   /* If we're resuming more than one thread simultaneously, then any
@@ -5842,6 +5872,13 @@ handle_thread_exited (execution_control_state *ecs)
   ecs->event_thread->stepping_over_breakpoint = 0;
   ecs->event_thread->stepping_over_watchpoint = 0;
 
+  /* If the thread had an FSM, then abort the command.  But only after
+     finishing the step over, as in non-stop mode, aborting this
+     thread's command should not interfere with other threads.  We
+     must check this before finish_step over, however, which may
+     update the thread list and delete the event thread.  */
+  bool abort_cmd = (ecs->event_thread->thread_fsm () != nullptr);
+
   /* Maybe the thread was doing a step-over, if so release
      resources and start any further pending step-overs.
 
@@ -5855,6 +5892,13 @@ handle_thread_exited (execution_control_state *ecs)
      the event thread has exited.  */
   gdb_assert (ret == 0);
 
+  if (abort_cmd)
+    {
+      delete_thread (ecs->event_thread);
+      ecs->event_thread = nullptr;
+      return false;
+    }
+
   /* If finish_step_over started a new in-line step-over, don't
      try to restart anything else.  */
   if (step_over_info_valid_p ())
@@ -9287,7 +9331,8 @@ normal_stop ()
       if (inferior_ptid != null_ptid)
 	finish_ptid = ptid_t (inferior_ptid.pid ());
     }
-  else if (last.kind () != TARGET_WAITKIND_NO_RESUMED)
+  else if (last.kind () != TARGET_WAITKIND_NO_RESUMED
+	   && last.kind () != TARGET_WAITKIND_THREAD_EXITED)
     finish_ptid = inferior_ptid;
 
   gdb::optional<scoped_finish_thread_state> maybe_finish_thread_state;
@@ -9330,7 +9375,8 @@ normal_stop ()
     {
       if ((last.kind () != TARGET_WAITKIND_SIGNALLED
 	   && last.kind () != TARGET_WAITKIND_EXITED
-	   && last.kind () != TARGET_WAITKIND_NO_RESUMED)
+	   && last.kind () != TARGET_WAITKIND_NO_RESUMED
+	   && last.kind () != TARGET_WAITKIND_THREAD_EXITED)
 	  && target_has_execution ()
 	  && previous_thread != inferior_thread ())
 	{
@@ -9346,7 +9392,8 @@ normal_stop ()
       update_previous_thread ();
     }
 
-  if (last.kind () == TARGET_WAITKIND_NO_RESUMED)
+  if (last.kind () == TARGET_WAITKIND_NO_RESUMED
+      || last.kind () == TARGET_WAITKIND_THREAD_EXITED)
     {
       stop_print_frame = false;
 
@@ -9354,7 +9401,12 @@ normal_stop ()
 	if (current_ui->prompt_state == PROMPT_BLOCKED)
 	  {
 	    target_terminal::ours_for_output ();
-	    gdb_printf (_("No unwaited-for children left.\n"));
+	    if (last.kind () == TARGET_WAITKIND_NO_RESUMED)
+	      gdb_printf (_("No unwaited-for children left.\n"));
+	    else if (last.kind () == TARGET_WAITKIND_THREAD_EXITED)
+	      gdb_printf (_("Command aborted, thread exited.\n"));
+	    else
+	      gdb_assert_not_reached ("unhandled");
 	  }
     }
 
@@ -9437,7 +9489,8 @@ normal_stop ()
     {
       if (last.kind () != TARGET_WAITKIND_SIGNALLED
 	  && last.kind () != TARGET_WAITKIND_EXITED
-	  && last.kind () != TARGET_WAITKIND_NO_RESUMED)
+	  && last.kind () != TARGET_WAITKIND_NO_RESUMED
+	  && last.kind () != TARGET_WAITKIND_THREAD_EXITED)
 	/* Delete the breakpoint we stopped at, if it wants to be deleted.
 	   Delete any breakpoint that is to be deleted at the next stop.  */
 	breakpoint_auto_delete (inferior_thread ()->control.stop_bpstat);
diff --git a/gdb/testsuite/gdb.threads/step-over-thread-exit.exp b/gdb/testsuite/gdb.threads/step-over-thread-exit.exp
index ed8534cf518..615bd838763 100644
--- a/gdb/testsuite/gdb.threads/step-over-thread-exit.exp
+++ b/gdb/testsuite/gdb.threads/step-over-thread-exit.exp
@@ -29,7 +29,7 @@ if { [build_executable "failed to prepare" $testfile \
 # NS_STOP_ALL is only used if testing "set non-stop on", and indicates
 # whether to have GDB explicitly stop all threads before continuing to
 # thread exit.
-proc test {displaced-stepping non-stop target-non-stop schedlock ns_stop_all} {
+proc test {displaced-stepping non-stop target-non-stop schedlock cmd ns_stop_all} {
     if {${non-stop} == "off" && $ns_stop_all} {
 	error "invalid arguments"
     }
@@ -72,31 +72,58 @@ proc test {displaced-stepping non-stop target-non-stop schedlock ns_stop_all} {
 
 	gdb_test_no_output "set scheduler-locking ${schedlock}"
 
-	gdb_test "continue" \
-	    "No unwaited-for children left." \
-	    "continue stops when thread exits"
+	if {$cmd == "continue"} {
+	    gdb_test "continue" \
+		"No unwaited-for children left." \
+		"continue stops when thread exits"
+	} else {
+	    gdb_test_multiple $cmd "command aborts when thread exits" {
+		-re "Command aborted, thread exited\\.\r\n$::gdb_prompt " {
+		    pass $gdb_test_name
+		}
+	    }
+	}
     } else {
 	gdb_test_no_output "set scheduler-locking ${schedlock}"
 
-	for { set i 0 } { $i < 100 } { incr i } {
-	    with_test_prefix "iter $i" {
-		set ok 0
-		set thread "<unknown>"
-		gdb_test_multiple "continue" "" {
-		    -re -wrap "Thread ($::decimal) .*hit Breakpoint $::decimal.* my_exit_syscall .*" {
-			set thread $expect_out(1,string)
-			set ok 1
-		    }
+	if {$cmd != "continue"} {
+	    set thread "<unknown>"
+	    gdb_test_multiple "continue" "" {
+		-re -wrap "Thread ($::decimal) .*hit Breakpoint $::decimal.* my_exit_syscall .*" {
+		    set thread $expect_out(1,string)
 		}
-		if {!${ok}} {
-		    # Exit if there's a failure to avoid lengthy
-		    # timeouts.
-		    break
+	    }
+	    if {${non-stop}} {
+		gdb_test -nopass "thread $thread" "Switching to thread .*" \
+		    "switch to event thread"
+	    }
+
+	    gdb_test_multiple $cmd "command aborts when thread exits" {
+		-re "Command aborted, thread exited\\.\r\n$::gdb_prompt " {
+		    pass $gdb_test_name
 		}
+	    }
+	} else {
+	    for { set i 0 } { $i < 100 } { incr i } {
+		with_test_prefix "iter $i" {
+		    set ok 0
+		    set thread "<unknown>"
+		    gdb_test_multiple "continue" "" {
+			-re -wrap "Thread ($::decimal) .*hit Breakpoint $::decimal.* my_exit_syscall .*" {
+			    set thread $expect_out(1,string)
+			    set ok 1
+			}
+		    }
+		    if {!${ok}} {
+			# Exit if there's a failure to avoid lengthy
+			# timeouts.
+			break
+		    }
 
-		if {${non-stop}} {
-		    gdb_test "thread $thread" "Switching to thread .*" \
-			"switch to event thread"
+		    if {${non-stop}} {
+			gdb_test -nopass "thread $thread" "Switching to thread .*" \
+			    "switch to event thread"
+		    }
 		}
 	    }
 	}
@@ -112,13 +139,15 @@ foreach_with_prefix displaced-stepping {off auto} {
 	    }
 
 	    foreach_with_prefix schedlock {off on} {
-		if {${non-stop} == "on"} {
-		    foreach_with_prefix ns_stop_all {0 1} {
-			test ${displaced-stepping} ${non-stop} ${target-non-stop} \
-			    ${schedlock} ${ns_stop_all}
+		foreach_with_prefix cmd {"next" "continue"} {
+		    if {${non-stop} == "on"} {
+			foreach_with_prefix ns_stop_all {0 1} {
+			    test ${displaced-stepping} ${non-stop} ${target-non-stop} \
+				${schedlock} ${cmd} ${ns_stop_all}
+			}
+		    } else {
+			test ${displaced-stepping} ${non-stop} ${target-non-stop} ${schedlock} ${cmd} 0
 		    }
-		} else {
-		    test ${displaced-stepping} ${non-stop} ${target-non-stop} ${schedlock} 0
 		}
 	    }
 	}
-- 
2.34.1


  parent reply	other threads:[~2023-11-13 15:06 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-11-13 15:04 [FYI/pushed v4 00/25] Step over thread clone and thread exit Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 01/25] Add "maint info linux-lwps" command Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 02/25] gdb/linux: Delete all other LWPs immediately on ptrace exec event Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 03/25] Step over clone syscall w/ breakpoint, TARGET_WAITKIND_THREAD_CLONED Pedro Alves
2023-11-14 12:55   ` Guinevere Larsen
2023-11-14 13:26     ` Pedro Alves
2023-11-14 16:29       ` Guinevere Larsen
2023-11-14 16:44         ` Luis Machado
2023-11-14 13:28     ` Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 04/25] Support clone events in the remote protocol Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 05/25] Avoid duplicate QThreadEvents packets Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 06/25] Thread options & clone events (core + remote) Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 07/25] Thread options & clone events (native Linux) Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 08/25] Thread options & clone events (Linux GDBserver) Pedro Alves
2024-02-06 11:04   ` Luis Machado
2024-02-06 14:57     ` Tom Tromey
2024-02-06 15:12       ` Luis Machado
2024-02-07  8:59       ` Luis Machado
2024-02-07 15:43         ` Tom Tromey
2024-02-07 17:10           ` Simon Marchi
2024-02-07 18:05             ` Luis Machado
2024-02-07 18:18               ` Tom Tromey
2024-02-07 18:56                 ` Pedro Alves
2024-02-07 20:11                   ` Pedro Alves
2024-02-08  8:57                     ` Luis Machado
2024-02-08 10:53                       ` Pedro Alves
2024-02-08 11:47                         ` Luis Machado
2024-02-08 14:58                     ` Tom Tromey
2024-02-07 18:06             ` Tom Tromey
2023-11-13 15:04 ` [FYI/pushed v4 09/25] gdbserver: Hide and don't detach pending clone children Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 10/25] Remove gdb/19675 kfails (displaced stepping + clone) Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 11/25] all-stop/synchronous RSP support thread-exit events Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 12/25] gdbserver/linux-low.cc: Ignore event_ptid if TARGET_WAITKIND_IGNORE Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 13/25] Move deleting thread on TARGET_WAITKIND_THREAD_EXITED to core Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 14/25] Introduce GDB_THREAD_OPTION_EXIT thread option, fix step-over-thread-exit Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 15/25] Implement GDB_THREAD_OPTION_EXIT support for Linux GDBserver Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 16/25] Implement GDB_THREAD_OPTION_EXIT support for native Linux Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 17/25] gdb: clear step over information on thread exit (PR gdb/27338) Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 18/25] stop_all_threads: (re-)enable async before waiting for stops Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 19/25] gdbserver: Queue no-resumed event after thread exit Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 20/25] Don't resume new threads if scheduler-locking is in effect Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 21/25] Report thread exit event for leader if reporting thread exit events Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 22/25] gdb/testsuite/lib/my-syscalls.S: Refactor new SYSCALL macro Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 23/25] Testcases for stepping over thread exit syscall (PR gdb/27338) Pedro Alves
2023-11-13 15:04 ` [FYI/pushed v4 24/25] Document remote clone events, and QThreadOptions packet Pedro Alves
2023-11-13 15:04 ` Pedro Alves [this message]
2023-11-13 19:28 ` [FYI/pushed v4 00/25] Step over thread clone and thread exit Tom de Vries
2023-11-14 10:51   ` Pedro Alves
2023-11-14 13:39     ` Tom de Vries

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20231113150427.477431-26-pedro@palves.net \
    --to=pedro@palves.net \
    --cc=aburgess@redhat.com \
    --cc=gdb-patches@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).