linux-kernel - [PATCH v2 4/4] PM: hibernate: don't set PF_FREEZER_SKIP flag when manipulating /dev/snapshot

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

Date:   Sun, 10 Jul 2022 11:25:47 +0900
From:   Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
To:     Greg KH <gregkh@...uxfoundation.org>,
        Oliver Neukum <oneukum@...e.com>,
        Wedson Almeida Filho <wedsonaf@...gle.com>,
        "Rafael J. Wysocki" <rjw@...k.pl>,
        Arjan van de Ven <arjan@...ux.intel.com>,
        Len Brown <len.brown@...el.com>,
        Dmitry Vyukov <dvyukov@...gle.com>
Cc:     linux-pm@...r.kernel.org, LKML <linux-kernel@...r.kernel.org>
Subject: [PATCH v2 4/4] PM: hibernate: don't set PF_FREEZER_SKIP flag when
 manipulating /dev/snapshot

Since khungtaskd skips threads with PF_FREEZER_SKIP flag set, currently
we can't report unbounded uninterruptible sleep when something went wrong
while manipulating /dev/snapshot interface.

Let's change snapshot_{open,read,write}() to use mutex_lock_killable()
and change snapshot_release() to use mutex_lock(), so that khungtaskd can
report unbounded uninterruptible sleep, by not setting PF_FREEZER_SKIP
flag.

Since /dev/snapshot is exclusive due to hibernate_acquire(), we could
choose mutex_trylock() for snapshot_{open,read,write}() as with
snapshot_ioctl(). But until we confirm that this patch does not
break something, let's stay mutex_lock_killable().

Signed-off-by: Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
Cc: "Rafael J. Wysocki" <rafael@...nel.org>
Cc: Len Brown <len.brown@...el.com>
Cc: Pavel Machek <pavel@....cz>
---
This patch is only compile tested. Need to review if somewhere depends
on PF_FREEZER_SKIP flag being set.

 kernel/power/user.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)

diff --git a/kernel/power/user.c b/kernel/power/user.c
index 32dd5a855e8c..9936efa07022 100644
--- a/kernel/power/user.c
+++ b/kernel/power/user.c
@@ -68,7 +68,8 @@ static int snapshot_open(struct inode *inode, struct file *filp)
 		break;
 	}
 
-	lock_system_sleep();
+	if (mutex_lock_killable(&system_transition_mutex))
+		return -EINTR;
 
 	if (!hibernate_acquire()) {
 		error = -EBUSY;
@@ -102,7 +103,7 @@ static int snapshot_open(struct inode *inode, struct file *filp)
 	data->dev = 0;
 
  Unlock:
-	unlock_system_sleep();
+	mutex_unlock(&system_transition_mutex);
 
 	return error;
 }
@@ -111,7 +112,7 @@ static int snapshot_release(struct inode *inode, struct file *filp)
 {
 	struct snapshot_data *data;
 
-	lock_system_sleep();
+	mutex_lock(&system_transition_mutex);
 
 	swsusp_free();
 	data = filp->private_data;
@@ -128,7 +129,7 @@ static int snapshot_release(struct inode *inode, struct file *filp)
 			PM_POST_HIBERNATION : PM_POST_RESTORE);
 	hibernate_release();
 
-	unlock_system_sleep();
+	mutex_unlock(&system_transition_mutex);
 
 	return 0;
 }
@@ -140,7 +141,8 @@ static ssize_t snapshot_read(struct file *filp, char __user *buf,
 	ssize_t res;
 	loff_t pg_offp = *offp & ~PAGE_MASK;
 
-	lock_system_sleep();
+	if (mutex_lock_killable(&system_transition_mutex))
+		return -EINTR;
 
 	data = filp->private_data;
 	if (!data->ready) {
@@ -161,7 +163,7 @@ static ssize_t snapshot_read(struct file *filp, char __user *buf,
 		*offp += res;
 
  Unlock:
-	unlock_system_sleep();
+	mutex_unlock(&system_transition_mutex);
 
 	return res;
 }
@@ -173,7 +175,8 @@ static ssize_t snapshot_write(struct file *filp, const char __user *buf,
 	ssize_t res;
 	loff_t pg_offp = *offp & ~PAGE_MASK;
 
-	lock_system_sleep();
+	if (mutex_lock_killable(&system_transition_mutex))
+		return -EINTR;
 
 	data = filp->private_data;
 
@@ -195,7 +198,7 @@ static ssize_t snapshot_write(struct file *filp, const char __user *buf,
 	if (res > 0)
 		*offp += res;
 unlock:
-	unlock_system_sleep();
+	mutex_unlock(&system_transition_mutex);
 
 	return res;
 }
-- 
2.18.4