linux-kernel - [PATCH 1/3] PM / Sleep: Replace mutex_[un]lock(&system_transition_mutex) with [un]lock_system

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-Id: <20250520032545.29558-2-zhangzihuan@kylinos.cn>
Date: Tue, 20 May 2025 11:25:43 +0800
From: Zihuan Zhang <zhangzihuan@...inos.cn>
To: rafael@...nel.org,
	len.brown@...el.com,
	pavel@...nel.org
Cc: linux-pm@...r.kernel.org,
	linux-kernel@...r.kernel.org,
	akpm@...ux-foundation.org,
	tzungbi@...nel.org,
	a.fatoum@...gutronix.de,
	jani.nikula@...el.com,
	joel.granados@...nel.org,
	paulmck@...nel.org,
	zhangguopeng@...inos.cn,
	linux@...ssschuh.net,
	Zihuan Zhang <zhangzihuan@...inos.cn>
Subject: [PATCH 1/3] PM / Sleep: Replace mutex_[un]lock(&system_transition_mutex) with [un]lock_system_sleep()

These function currently calls mutex_lock(&system_transition_mutex) and
mutex_unlock() directly to protect the resume-from-disk operation from
concurrent suspend/resume transitions.

However, this is inconsistent with the rest of the power management
code, where lock_system_sleep() and unlock_system_sleep() are used to
wrap suspend/hibernate transitions. These wrapper functions not only
acquire system_transition_mutex, but also set PF_NOFREEZE for the
calling thread, ensuring that it is not subject to freezing during
suspend.

This change replaces the raw mutex_lock()/unlock() with the standard
lock_system_sleep()/unlock_system_sleep() wrapper pair, bringing it in
line with the locking pattern used by hibernate(), pm_suspend(), and
other similar entry points.

Benefits of this change:

 - Ensures the thread performing software resume is marked PF_NOFREEZE,
   which is important during early resume paths where freezing is
active.
 - Improves code clarity by making the locking intent more explicit.
 - Unifies suspend/hibernate locking style across the kernel power
   subsystem.
 - Reduces the risk of future maintenance issues due to inconsistent
   locking.

No functional change is expected at runtime, since the lock order and
coverage remain the same. This is a straightforward cleanup and
consistency fix.

Signed-off-by: Zihuan Zhang <zhangzihuan@...inos.cn>
---
 kernel/power/hibernate.c | 5 +++--
 kernel/reboot.c          | 5 +++--
 2 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c
index 23c0f4e6cb2f..cfaa92f24857 100644
--- a/kernel/power/hibernate.c
+++ b/kernel/power/hibernate.c
@@ -988,6 +988,7 @@ static int __init find_resume_device(void)
 
 static int software_resume(void)
 {
+	unsigned int sleep_flags;
 	int error;
 
 	pm_pr_dbg("Hibernation image partition %d:%d present\n",
@@ -995,7 +996,7 @@ static int software_resume(void)
 
 	pm_pr_dbg("Looking for hibernation image.\n");
 
-	mutex_lock(&system_transition_mutex);
+	sleep_flags = lock_system_sleep();
 	error = swsusp_check(true);
 	if (error)
 		goto Unlock;
@@ -1050,7 +1051,7 @@ static int software_resume(void)
 	hibernate_release();
 	/* For success case, the suspend path will release the lock */
  Unlock:
-	mutex_unlock(&system_transition_mutex);
+	unlock_system_sleep(sleep_flags);
 	pm_pr_dbg("Hibernation image not present or could not be loaded.\n");
 	return error;
  Close_Finish:
diff --git a/kernel/reboot.c b/kernel/reboot.c
index ec087827c85c..68ac7e377efb 100644
--- a/kernel/reboot.c
+++ b/kernel/reboot.c
@@ -729,6 +729,7 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd,
 		void __user *, arg)
 {
 	struct pid_namespace *pid_ns = task_active_pid_ns(current);
+	unsigned int sleep_flags;
 	char buffer[256];
 	int ret = 0;
 
@@ -761,7 +762,7 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd,
 		cmd = LINUX_REBOOT_CMD_HALT;
 	}
 
-	mutex_lock(&system_transition_mutex);
+	sleep_flags = lock_system_sleep();
 	switch (cmd) {
 	case LINUX_REBOOT_CMD_RESTART:
 		kernel_restart(NULL);
@@ -811,7 +812,7 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd,
 		ret = -EINVAL;
 		break;
 	}
-	mutex_unlock(&system_transition_mutex);
+	unlock_system_sleep(sleep_flags);
 	return ret;
 }
 
-- 
2.25.1