Re: [PATCH v2 3/4] PM: hibernate: allow wait_for_device_probe() to timeout when resuming from hibernation

[Tetsuo Handa]

On 2022/07/12 3:14, Rafael J. Wysocki wrote:
> On Mon, Jul 11, 2022 at 1:21 PM Tetsuo Handa
> <penguin-kernel@xxxxxxxxxxxxxxxxxxx> wrote:
>>
>> On 2022/07/11 17:12, Greg KH wrote:
>>>                                                        creating a
>>> locking loop like this should be resolved first,
>>
>> Rafael and Arjan, can we agree with removing wait_for_device_probe() from snapshot_open() ?
> 
> No, we can't.

Then, can we defer wait_for_device_probe() till first write()/ioctl()
which is called without locks?

 kernel/power/user.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/kernel/power/user.c b/kernel/power/user.c
index a00a728ddfc1..92aecb989c76 100644
--- a/kernel/power/user.c
+++ b/kernel/power/user.c
@@ -26,6 +26,7 @@
 
 #include "power.h"
 
+static bool need_wait;
 
 static struct snapshot_data {
 	struct snapshot_handle handle;
@@ -78,7 +79,7 @@ static int snapshot_open(struct inode *inode, struct file *filp)
 		 * Resuming.  We may need to wait for the image device to
 		 * appear.
 		 */
-		wait_for_device_probe();
+		need_wait = true;
 
 		data->swap = -1;
 		data->mode = O_WRONLY;
@@ -168,6 +169,11 @@ static ssize_t snapshot_write(struct file *filp, const char __user *buf,
 	ssize_t res;
 	loff_t pg_offp = *offp & ~PAGE_MASK;
 
+	if (need_wait) {
+		wait_for_device_probe();
+		need_wait = false;
+	}
+
 	lock_system_sleep();
 
 	data = filp->private_data;
@@ -244,6 +250,11 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd,
 	loff_t size;
 	sector_t offset;
 
+	if (need_wait) {
+		wait_for_device_probe();
+		need_wait = false;
+	}
+
 	if (_IOC_TYPE(cmd) != SNAPSHOT_IOC_MAGIC)
 		return -ENOTTY;
 	if (_IOC_NR(cmd) > SNAPSHOT_IOC_MAXNR)
-- 
2.18.4