From 658b3049fafea5a2b4fa34bc0e23fea0e6bfb28d Mon Sep 17 00:00:00 2001 From: Pedro Alves Date: Mon, 28 Jun 2021 14:05:54 +0100 Subject: stop_all_threads: (re-)enable async before waiting for stops Running the gdb.threads/step-over-thread-exit-while-stop-all-threads.exp testcase added later in the series against gdbserver, after the TARGET_WAITKIND_NO_RESUMED fix from the following patch, would run into an infinite loop in stop_all_threads, leading to a timeout: FAIL: gdb.threads/step-over-thread-exit-while-stop-all-threads.exp: displaced-stepping=off: target-non-stop=on: iter 0: continue (timeout) The is really a latent bug, and it is about the fact that stop_all_threads stops listening to events from a target as soon as it sees a TARGET_WAITKIND_NO_RESUMED, ignoring that TARGET_WAITKIND_NO_RESUMED may be delayed. handle_no_resumed knows how to handle delayed no-resumed events, but stop_all_threads was never taught to. In more detail, here's what happens with that testcase: #1 - Multiple threads report breakpoint hits to gdb. #2 - gdb picks one events, and it's for thread 1. All other stops are left pending. thread 1 needs to move past a breakpoint, so gdb stops all threads to start an inline step over for thread 1. While stopping threads, some of the threads that were still running report events that are also left pending. #2 - gdb steps thread 1 #3 - Thread 1 exits while stepping (it steps over an exit syscall), gdbserver reports thread exit for thread 1 #4 - Thread 1 was the last resumed thread, so gdbserver also reports no-resumed: [remote] Notification received: Stop:w0;p3445d0.3445d3 [remote] Sending packet: $vStopped#55 [remote] Packet received: N [remote] Sending packet: $vStopped#55 [remote] Packet received: OK #5 - gdb processes the thread exit for thread 1, finishes the step over and restarts threads. #6 - gdb picks the next event to process out of one of the resumed threads with pending events: [infrun] random_resumed_with_pending_wait_status: Found 32 events, selecting #11 #7 - This is again a breakpoint hit and the breakpoint needs to be stepped over too, so gdb starts a step-over dance again. #8 - We reach stop_all_threads, which finds that some threads need to be stopped. #9 - wait_one finally consumes the no-resumed event queue by #4. Seeing this, wait_one disable target async, to stop listening for events out of the remote target. #10 - We still haven't seen all the stops expected, so stop_all_threads tries another iteration. #11 - Because the remote target is no longer async, and there are no other targets, wait_one return no-resumed immediately without polling the remote target. #12 - We still haven't seen all the stops expected, so stop_all_threads tries another iteration. goto #11, looping forever. Fix this by explicitly enabling/re-enabling target async on targets that can async, before waiting for stops. Change-Id: Ie3ffb0df89635585a6631aa842689cecc989e33f --- gdb/infrun.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 81 insertions(+) diff --git a/gdb/infrun.c b/gdb/infrun.c index 2866962..31321d7 100644 --- a/gdb/infrun.c +++ b/gdb/infrun.c @@ -5011,6 +5011,8 @@ wait_one () if (nfds == 0) { /* No waitable targets left. All must be stopped. */ + infrun_debug_printf ("no waitable targets left"); + target_waitstatus ws; ws.set_no_resumed (); return {nullptr, minus_one_ptid, std::move (ws)}; @@ -5269,6 +5271,83 @@ handle_one (const wait_one_event &event) return false; } +/* Helper for stop_all_threads. wait_one waits for events until it + sees a TARGET_WAITKIND_NO_RESUMED event. When it sees one, it + disables target_async for the target to stop waiting for events + from it. TARGET_WAITKIND_NO_RESUMED can be delayed though, + consider, debugging against gdbserver: + + #1 - Threads 1-5 are running, and thread 1 hits a breakpoint. + + #2 - gdb processes the breakpoint hit for thread 1, stops all + threads, and steps thread 1 over the breakpoint. while + stopping threads, some other threads reported interesting + events, which were left pending in the thread's objects + (infrun's queue). + + #2 - Thread 1 exits (it stepped an exit syscall), and gdbserver + reports the thread exit for thread 1. The event ends up in + remote's stop reply queue. + + #3 - That was the last resumed thread, so gdbserver reports + no-resumed, and that event also ends up in remote's stop + reply queue, queued after the thread exit from #2. + + #4 - gdb processes the thread exit event, which finishes the + step-over, and so gdb restarts all threads (threads with + pending events are left marked resumed, but aren't set + executing). The no-resumed event is still left pending in + the remote stop reply queue. + + #5 - Since there are now resumed threads with pending breakpoint + hits, gdb picks one at random to process next. + + #5 - gdb picks the breakpoint hit for thread 2 this time, and that + breakpoint also needs to be stepped over, so gdb stops all + threads again. + + #6 - stop_all_threads counts number of expected stops and calls + wait_one once for each. + + #7 - The first wait_one call collects the no-resumed event from #3 + above. + + #9 - Seeing the no-resumed event, wait_one disables target async + for the remote target, to stop waiting for events from it. + wait_one from here on always return no-resumed directly + without reaching the target. + + #10 - stop_all_threads still hasn't seen all the stops it expects, + so it does another pass. + + #11 - Since the remote target is not async (disabled in #9), + wait_one doesn't wait on it, so it won't see the expected + stops, and instead returns no-resumed directly. + + #12 - stop_all_threads still haven't seen all the stops, so it + does another pass. goto #b, looping forever. + + To handle this, we explicitly (re-)enable target async on all + targets that can async every time stop_all_threads goes wait for + the expected stops. */ + +static void +reenable_target_async () +{ + for (inferior *inf : all_inferiors ()) + { + process_stratum_target *target = inf->process_target (); + if (target != nullptr + && target->threads_executing + && target->can_async_p () + && !target->is_async_p ()) + { + switch_to_inferior_no_thread (inf); + target_async (1); + } + } +} + /* See infrun.h. */ void @@ -5395,6 +5474,8 @@ stop_all_threads (const char *reason, inferior *inf) if (pass > 0) pass = -1; + reenable_target_async (); + for (int i = 0; i < waits_needed; i++) { wait_one_event event = wait_one (); -- cgit v1.1