Close block devices in a synchronous manner

fput() doesn't wait for all references on the disk to be unclaimed but instead
it only schedules a worker that is supposed to cleanup resources once the device
is released.

During cache initialization we open device at least twice - to check its
properties and then to actually use it as cache. But since we use the async
fput() after the probe, the device might still be in use once we try to open it
for the second time (the second open returns -EBUSY).

Using synchronous __fput_sync() to close the device fixes the issue

Signed-off-by: Michal Mielewczyk <michal.mielewczyk@huawei.com>

configure.d/1_bdev_release.conf

@@ -20,9 +20,12 @@ check() {
 	elif compile_module $cur_name "bdev_release(NULL);" "linux/blkdev.h"
 	then
 		echo $cur_name 3 >> $config_file_path
-	elif compile_module $cur_name "bdev_file_open_by_path(NULL, 0, NULL, NULL);" "linux/blkdev.h"
+	elif compile_module $cur_name "bdev_file_open_by_path(NULL, 0, NULL, NULL);__fput_sync(NULL);" "linux/blkdev.h"
 	then
 		echo $cur_name 4 >> $config_file_path
+	elif compile_module $cur_name "bdev_file_open_by_path(NULL, 0, NULL, NULL);" "linux/blkdev.h"
+	then
+		echo $cur_name 5 >> $config_file_path
 	else
 		echo $cur_name X >> $config_file_path
 	fi
@@ -40,6 +43,9 @@ apply() {
 		add_define "cas_bdev_release(handle, mode, holder) \\
 			bdev_release(handle)" ;;
 	"4")
+		add_define "cas_bdev_release(handle, mode, holder) \\
+			__fput_sync(handle)" ;;
+	"5")
 		add_define "cas_bdev_release(handle, mode, holder) \\
 			fput(handle)" ;;
 	*)