From: Pedro Alves <pedro@palves.net>
To: gdb-patches@sourceware.org
Subject: [PATCH v2 23/47] Windows gdbserver: Fix scheduler-locking
Date: Mon, 19 May 2025 14:22:44 +0100 [thread overview]
Message-ID: <20250519132308.3553663-24-pedro@palves.net> (raw)
In-Reply-To: <20250519132308.3553663-1-pedro@palves.net>
This rewrites win32_process_target::resume such that scheduler-locking
is implemented properly.
It also uses the new get_last_debug_event_ptid function to avoid
considering passing a signal to the wrong thread, like done for the
native side in a previous patch.
Note this code/comment being removed:
- /* Yes, we're ignoring resume_info[0].thread. It'd be tricky to make
- the Windows resume code do the right thing for thread switching. */
- tid = windows_process.current_event.dwThreadId;
This meant that scheduler-locking currently is broken badly unless you
stay in the thread that last reported an event. If you switch to a
different thread from the one that last reported an event and step,
you get a spurious SIGTRAP in the thread that last reported a stop,
not the one that you tried to step:
(gdb) t 1
[Switching to thread 1 (Thread 3908)]
#0 0x00007fffc768d6e4 in ntdll!ZwDelayExecution () from target:C:/Windows/SYSTEM32/ntdll.dll
(gdb) set scheduler-locking on
(gdb) set disassemble-next-line on
(gdb) frame
#0 0x00007fffc768d6e4 in ntdll!ZwDelayExecution () from target:C:/Windows/SYSTEM32/ntdll.dll
=> 0x00007fffc768d6e4 <ntdll!ZwDelayExecution+20>: c3 ret
(gdb) si
Thread 3 received signal SIGTRAP, Trace/breakpoint trap.
[Switching to Thread 2660]
0x00007fffc4e4e92e in KERNELBASE!EncodeRemotePointer () from target:C:/Windows/System32/KernelBase.dll
=> 0x00007fffc4e4e92e <KERNELBASE!EncodeRemotePointer+8254>: eb 78 jmp 0x7fffc4e4e9a8 <KERNELBASE!EncodeRemotePointer+8376>
(gdb)
Note how we switched to thread 1, stepped, and GDBserver still stepped
thread 3... This is fixed by this patch. We now get:
(gdb) info threads
Id Target Id Frame
1 Thread 920 0x00007ffe0372d6e4 in ntdll!ZwDelayExecution () from target:C:/Windows/SYSTEM32/ntdll.dll
2 Thread 8528 0x00007ffe03730ad4 in ntdll!ZwWaitForWorkViaWorkerFactory () from target:C:/Windows/SYSTEM32/ntdll.dll
3 Thread 3128 0x00007ffe03730ad4 in ntdll!ZwWaitForWorkViaWorkerFactory () from target:C:/Windows/SYSTEM32/ntdll.dll
* 4 Thread 7164 0x00007ffe0102e929 in KERNELBASE!EncodeRemotePointer () from target:C:/Windows/System32/KernelBase.dll
5 Thread 8348 0x00007ffe0372d6e4 in ntdll!ZwDelayExecution () from target:C:/Windows/SYSTEM32/ntdll.dll
6 Thread 2064 0x00007ffe0372d6e4 in ntdll!ZwDelayExecution () from target:C:/Windows/SYSTEM32/ntdll.dll
(gdb) t 1
[Switching to thread 1 (Thread 920)]
#0 0x00007ffe0372d6e4 in ntdll!ZwDelayExecution () from target:C:/Windows/SYSTEM32/ntdll.dll
(gdb) set scheduler-locking on
(gdb) si
0x00007ffe0372d6e4 in ntdll!ZwDelayExecution () from target:C:/Windows/SYSTEM32/ntdll.dll
(gdb) si
0x00007ffe00f9b44e in SleepEx () from target:C:/Windows/System32/KernelBase.dll
(gdb) si
0x00007ffe00f9b453 in SleepEx () from target:C:/Windows/System32/KernelBase.dll
I.e., we kept stepping the right thread, thread 1.
Note we stopped again at 0x00007ffe0372d6e4 the first time (same PC
the thread already was at before the first stepi) because the thread
had been stopped at a syscall, so that's normal:
(gdb) disassemble
Dump of assembler code for function ntdll!ZwDelayExecution:
0x00007ffe0372d6d0 <+0>: mov %rcx,%r10
0x00007ffe0372d6d3 <+3>: mov $0x34,%eax
0x00007ffe0372d6d8 <+8>: testb $0x1,0x7ffe0308
0x00007ffe0372d6e0 <+16>: jne 0x7ffe0372d6e5 <ntdll!ZwDelayExecution+21>
0x00007ffe0372d6e2 <+18>: syscall
=> 0x00007ffe0372d6e4 <+20>: ret
Change-Id: I44f4fe4cb98592517569c6716b9d189f42db25a0
---
gdbserver/win32-low.cc | 158 ++++++++++++++++++++++-------------------
1 file changed, 85 insertions(+), 73 deletions(-)
diff --git a/gdbserver/win32-low.cc b/gdbserver/win32-low.cc
index 32464024722..2e3e4442012 100644
--- a/gdbserver/win32-low.cc
+++ b/gdbserver/win32-low.cc
@@ -405,19 +405,13 @@ continue_one_thread (thread_info *thread, int thread_id)
}
}
+/* Helper for win32_process_target::kill. Resume all suspended
+ threads that match THREAD_ID, and continue the last debug event
+ with CONTINUE_STATUS. */
+
static BOOL
-child_continue (DWORD continue_status, int thread_id)
+child_continue_for_kill (DWORD continue_status, int thread_id)
{
- process_info *proc = find_process_pid (windows_process.process_id);
- for (thread_info &thread : proc->thread_list ())
- {
- auto *th = static_cast<windows_thread_info *> (thread.target_data ());
- if ((thread_id == -1 || thread_id == (int) th->tid)
- && !th->suspended
- && th->pending_stop.status.kind () != TARGET_WAITKIND_IGNORE)
- return TRUE;
- }
-
/* The inferior will only continue after the ContinueDebugEvent
call. */
for_each_thread ([&] (thread_info *thread)
@@ -676,7 +670,7 @@ win32_process_target::kill (process_info *process)
TerminateProcess (windows_process.handle, 0);
for (;;)
{
- if (!child_continue (DBG_CONTINUE, -1))
+ if (!child_continue_for_kill (DBG_CONTINUE, -1))
break;
if (!wait_for_debug_event (&windows_process.current_event, INFINITE))
break;
@@ -743,90 +737,108 @@ win32_process_target::thread_alive (ptid_t ptid)
return find_thread_ptid (ptid) != NULL;
}
-/* Resume the inferior process. RESUME_INFO describes how we want
- to resume. */
-void
-win32_process_target::resume (thread_resume *resume_info, size_t n)
-{
- DWORD tid;
- enum gdb_signal sig;
- int step;
- windows_thread_info *th;
- DWORD continue_status = DBG_CONTINUE;
- ptid_t ptid;
-
- /* This handles the very limited set of resume packets that GDB can
- currently produce. */
-
- if (n == 1 && resume_info[0].thread == minus_one_ptid)
- tid = -1;
- else if (n > 1)
- tid = -1;
- else
- /* Yes, we're ignoring resume_info[0].thread. It'd be tricky to make
- the Windows resume code do the right thing for thread switching. */
- tid = windows_process.current_event.dwThreadId;
+/* Helper for win32_process_target::resume. Resume THREAD, with
+ signal SIG, and stepping if STEP. If resuming the thread that last
+ reported an event, sets CONTINUE_STATUS appropriately to the status
+ that should be passed to ContinueDebugEvent. */
- if (resume_info[0].thread != minus_one_ptid)
- {
- sig = gdb_signal_from_host (resume_info[0].sig);
- step = resume_info[0].kind == resume_step;
- }
- else
- {
- sig = GDB_SIGNAL_0;
- step = 0;
- }
+static void
+resume_one_thread (thread_info *thread, bool step, gdb_signal sig,
+ DWORD *continue_status)
+{
+ auto *th = static_cast<windows_thread_info *> (thread->target_data ());
if (sig != GDB_SIGNAL_0)
{
- if (windows_process.current_event.dwDebugEventCode
- != EXCEPTION_DEBUG_EVENT)
+ /* Allow continuing with the same signal that interrupted us.
+ Otherwise complain. */
+ if (thread->id != get_last_debug_event_ptid ())
+ {
+ /* ContinueDebugEvent will be for a different thread. */
+ OUTMSG (("Cannot continue with signal %d here. "
+ "Not last-event thread", sig));
+ }
+ else if (windows_process.current_event.dwDebugEventCode
+ != EXCEPTION_DEBUG_EVENT)
{
- OUTMSG (("Cannot continue with signal %s here.\n",
+ OUTMSG (("Cannot continue with signal %s here. "
+ "Not stopped for EXCEPTION_DEBUG_EVENT.\n",
gdb_signal_to_string (sig)));
}
else if (sig == windows_process.last_sig)
- continue_status = DBG_EXCEPTION_NOT_HANDLED;
+ *continue_status = DBG_EXCEPTION_NOT_HANDLED;
else
OUTMSG (("Can only continue with received signal %s.\n",
gdb_signal_to_string (windows_process.last_sig)));
}
- windows_process.last_sig = GDB_SIGNAL_0;
+ win32_prepare_to_resume (th);
- /* Get context for the currently selected thread. */
- ptid = debug_event_ptid (&windows_process.current_event);
- th = windows_process.find_thread (ptid);
- if (th)
+ DWORD *context_flags = windows_process.context_flags_ptr (th);
+ if (*context_flags)
{
- win32_prepare_to_resume (th);
+ /* Move register values from the inferior into the thread
+ context structure. */
+ regcache_invalidate ();
- DWORD *context_flags = windows_process.context_flags_ptr (th);
- if (*context_flags)
+ if (step)
{
- /* Move register values from the inferior into the thread
- context structure. */
- regcache_invalidate ();
+ if (the_low_target.single_step != NULL)
+ (*the_low_target.single_step) (th);
+ else
+ error ("Single stepping is not supported "
+ "in this configuration.\n");
+ }
+
+ win32_set_thread_context (th);
+ *context_flags = 0;
+ }
+
+ continue_one_thread (thread, th->tid);
+}
+
+/* Resume the inferior process. RESUME_INFO describes how we want
+ to resume. */
+void
+win32_process_target::resume (thread_resume *resume_info, size_t n)
+{
+ DWORD continue_status = DBG_CONTINUE;
+ bool any_pending = false;
+
+ for_each_thread ([&] (thread_info *thread)
+ {
+ auto *th = static_cast<windows_thread_info *> (thread->target_data ());
- if (step)
+ for (int ndx = 0; ndx < n; ndx++)
+ {
+ thread_resume &r = resume_info[ndx];
+ ptid_t ptid = r.thread;
+ gdb_signal sig = gdb_signal_from_host (r.sig);
+ bool step = r.kind == resume_step;
+
+ if (ptid == minus_one_ptid || ptid == thread->id
+ /* Handle both 'pPID' and 'pPID.-1' as meaning 'all threads
+ of PID'. */
+ || (ptid.pid () == thread->id.pid ()
+ && (ptid.is_pid () || ptid.lwp () == -1)))
{
- if (the_low_target.single_step != NULL)
- (*the_low_target.single_step) (th);
- else
- error ("Single stepping is not supported "
- "in this configuration.\n");
- }
+ /* Ignore (wildcard) resume requests for already-resumed
+ threads. */
+ if (!th->suspended)
+ continue;
- win32_set_thread_context (th);
- *context_flags = 0;
+ resume_one_thread (thread, step, sig, &continue_status);
+ break;
+ }
}
- }
- /* Allow continuing with the same signal that interrupted us.
- Otherwise complain. */
+ if (!th->suspended
+ && th->pending_stop.status.kind () != TARGET_WAITKIND_IGNORE)
+ any_pending = true;
+ });
- child_continue (continue_status, tid);
+ if (!any_pending)
+ continue_last_debug_event (continue_status, debug_threads);
}
/* See nat/windows-nat.h. */
--
2.49.0
next prev parent reply other threads:[~2025-05-19 13:31 UTC|newest]
Thread overview: 97+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-19 13:22 [PATCH v2 00/47] Windows non-stop mode Pedro Alves
2025-05-19 13:22 ` [PATCH v2 01/47] Make default_gdb_exit resilient to failed closes Pedro Alves
2025-05-19 13:56 ` Andrew Burgess
2025-06-06 13:56 ` Pedro Alves
2025-05-19 13:22 ` [PATCH v2 02/47] Add test for continuing with some threads running Pedro Alves
2025-05-21 19:36 ` Kevin Buettner
2026-04-02 13:07 ` Pedro Alves
2025-05-19 13:22 ` [PATCH v2 03/47] infrun: Remove unnecessary currently_stepping call Pedro Alves
2025-05-21 19:44 ` Kevin Buettner
2026-04-02 13:17 ` Pedro Alves
2025-05-19 13:22 ` [PATCH v2 04/47] infrun: Split currently_stepping, fix sw watchpoints issue Pedro Alves
2026-04-02 13:33 ` Pedro Alves
2025-05-19 13:22 ` [PATCH v2 05/47] thread_info::executing+resumed -> thread_info::internal_state Pedro Alves
2026-04-06 18:01 ` Pedro Alves
2025-05-19 13:22 ` [PATCH v2 06/47] Windows gdb: Dead code in windows_nat_target::do_initial_windows_stuff Pedro Alves
2025-05-19 13:22 ` [PATCH v2 07/47] Windows gdb: Eliminate global current_process.dr[8] global Pedro Alves
2025-05-28 19:09 ` Tom Tromey
2026-04-06 19:44 ` Pedro Alves
2025-05-19 13:22 ` [PATCH v2 08/47] Windows gdb+gdbserver: New find_thread, replaces thread_rec(DONT_INVALIDATE_CONTEXT) Pedro Alves
2025-05-19 13:22 ` [PATCH v2 09/47] Windows gdb: handle_output_debug_string return type Pedro Alves
2025-05-19 13:22 ` [PATCH v2 10/47] Windows gdb: Eliminate reload_context Pedro Alves
2025-05-19 13:22 ` [PATCH v2 11/47] Windows gdb+gdbserver: Eliminate thread_rec(INVALIDATE_CONTEXT) calls Pedro Alves
2025-05-19 13:22 ` [PATCH v2 12/47] Windows gdb+gdbserver: Eliminate DONT_SUSPEND Pedro Alves
2025-05-19 13:22 ` [PATCH v2 13/47] Windows gdb+gdbserver: Eliminate windows_process_info::thread_rec Pedro Alves
2025-05-19 13:22 ` [PATCH v2 14/47] Windows gdb: Simplify windows_nat_target::wait Pedro Alves
2025-05-28 19:16 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 15/47] Windows gdb+gdbserver: Move suspending thread to when returning event Pedro Alves
2025-05-28 19:17 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 16/47] Windows gdb: Introduce continue_last_debug_event_main_thread Pedro Alves
2025-05-28 19:18 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 17/47] Windows gdb: Introduce windows_continue_flags Pedro Alves
2025-05-19 13:22 ` [PATCH v2 18/47] Windows gdb: Factor code out of windows_nat_target::windows_continue Pedro Alves
2025-05-19 13:22 ` [PATCH v2 19/47] Windows gdb: Pending stop and current_event Pedro Alves
2025-05-19 13:22 ` [PATCH v2 20/47] Windows gdb+gdbserver: Elim desired_stop_thread_id / rework pending_stops Pedro Alves
2025-05-30 20:41 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 21/47] Windows gdb+gdbserver: Introduce get_last_debug_event_ptid Pedro Alves
2025-05-28 19:21 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 22/47] Windows gdb: Can't pass signal to thread other than last stopped thread Pedro Alves
2025-05-28 19:22 ` Tom Tromey
2025-05-19 13:22 ` Pedro Alves [this message]
2025-05-30 20:37 ` [PATCH v2 23/47] Windows gdbserver: Fix scheduler-locking Tom Tromey
2025-05-19 13:22 ` [PATCH v2 24/47] Windows gdb: Enable "set scheduler-locking on" Pedro Alves
2025-05-19 13:22 ` [PATCH v2 25/47] Windows gdbserver: Eliminate soft-interrupt mechanism Pedro Alves
2025-05-19 13:22 ` [PATCH v2 26/47] Windows gdb+gdbserver: Make current_event per-thread state Pedro Alves
2025-05-28 19:30 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 27/47] Windows gdb+gdbserver: Make last_sig " Pedro Alves
2025-05-28 19:31 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 28/47] Windows gdb+gdbserver: Make siginfo_er " Pedro Alves
2025-05-28 19:33 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 29/47] Add backpointer from windows_thread_info to windows_process_info Pedro Alves
2025-05-19 13:22 ` [PATCH v2 30/47] Windows gdb+gdbserver: Share $_siginfo reading code Pedro Alves
2025-05-19 13:22 ` [PATCH v2 31/47] Windows gdb+gdbserver: Eliminate struct pending_stop Pedro Alves
2025-05-28 19:36 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 32/47] Windows gdb: Change serial_event management Pedro Alves
2025-05-28 19:37 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 33/47] Windows gdb: cygwin_set_dr => windows_set_dr, etc Pedro Alves
2025-05-19 13:22 ` [PATCH v2 34/47] Windows gdb: Avoid writing debug registers if watchpoint hit pending Pedro Alves
2025-05-30 20:43 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 35/47] Windows gdb+gdbserver: Check whether DBG_REPLY_LATER is available Pedro Alves
2025-05-19 13:22 ` [PATCH v2 36/47] linux-nat: Factor out get_detach_signal code to common code Pedro Alves
2025-05-28 19:44 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 37/47] Windows GDB: make windows_thread_info be private thread_info data Pedro Alves
2025-05-28 19:52 ` Tom Tromey
2025-05-19 13:22 ` [PATCH v2 38/47] Introduce windows_nat::event_code_to_string Pedro Alves
2025-05-28 19:53 ` Tom Tromey
2025-05-19 13:23 ` [PATCH v2 39/47] Windows gdb: Add non-stop support Pedro Alves
2025-06-05 16:21 ` Tom Tromey
2025-05-19 13:23 ` [PATCH v2 40/47] Windows gdb: Eliminate invalidate_context Pedro Alves
2025-05-28 19:54 ` Tom Tromey
2025-05-19 13:23 ` [PATCH v2 41/47] Windows gdb: Watchpoints while running (internal vs external stops) Pedro Alves
2025-05-30 20:50 ` Tom Tromey
2025-05-19 13:23 ` [PATCH v2 42/47] gdb_test_multiple: Anchor prompt match if -lbl Pedro Alves
2025-05-21 15:19 ` Tom de Vries
2025-05-27 22:41 ` Pedro Alves
2025-05-27 23:20 ` Pedro Alves
2025-05-28 11:59 ` [PATCH v2] of " Pedro Alves
2025-06-05 16:37 ` Pedro Alves
2025-06-05 17:20 ` [PATCH v3] " Pedro Alves
2025-06-06 9:58 ` Tom de Vries
2025-06-06 13:53 ` Pedro Alves
2025-05-19 13:23 ` [PATCH v2 43/47] Windows gdb: extra thread info => show exiting Pedro Alves
2025-05-28 19:58 ` Tom Tromey
2025-05-19 13:23 ` [PATCH v2 44/47] Add gdb.threads/leader-exit-schedlock.exp Pedro Alves
2025-05-29 16:09 ` Tom Tromey
2025-05-19 13:23 ` [PATCH v2 45/47] infrun: with AS+NS, prefer process exit over thread exit Pedro Alves
2025-05-19 13:23 ` [PATCH v2 46/47] Windows gdb: Always non-stop (default to "maint set target-non-stop on") Pedro Alves
2025-05-29 16:02 ` Tom Tromey
2025-05-19 13:23 ` [PATCH v2 47/47] Mention Windows scheduler-locking and non-stop support in NEWS Pedro Alves
2025-05-19 14:07 ` Eli Zaretskii
2025-06-05 17:57 ` [PATCH v2 00/47] Windows non-stop mode Tom Tromey
2025-06-11 22:06 ` [PATCH] Improve attach on Windows (was: Re: [PATCH v2 00/47] Windows non-stop mode) Pedro Alves
2026-04-02 12:21 ` [PATCH] Improve attach on Windows Pedro Alves
2026-04-02 18:52 ` Tom Tromey
2025-06-11 23:51 ` [PATCH v2 00/47] Windows non-stop mode Pedro Alves
2025-06-12 19:23 ` Tom Tromey
2025-06-13 10:34 ` Pedro Alves
2025-06-13 14:23 ` Tom Tromey
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=20250519132308.3553663-24-pedro@palves.net \
--to=pedro@palves.net \
--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