loop: fix zero sized loop for block special file

By default, /dev/sda is block special file from devtmpfs, getattr will
return file size as zero, causing loop failed for raw block device.

We can add bdev_statx() to return device size, however this may
introduce changes that are not acknowledged by user. Fix this problem by
reverting changes for block special file, file mapping host is set to
bdev inode while opening, and use i_size_read() directly to get device
size.

Fixes: 47b71ab ("loop: use vfs_getattr_nosec for accurate file size")
Reported-by: kernel test robot <[email protected]>
Closes: https://lore.kernel.org/oe-lkp/[email protected]
Signed-off-by: Yu Kuai <[email protected]>
Reviewed-by: Ming Lei <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
[axboe: fix spelling error]
Signed-off-by: Jens Axboe <[email protected]>

Author: YuKuai-huawei

drivers/block/loop.c

@@ -139,20 +139,26 @@ static int part_shift;
 
 static loff_t lo_calculate_size(struct loop_device *lo, struct file *file)
 {
-	struct kstat stat;
 	loff_t loopsize;
 	int ret;
 
-	/*
-	 * Get the accurate file size. This provides better results than
-	 * cached inode data, particularly for network filesystems where
-	 * metadata may be stale.
-	 */
-	ret = vfs_getattr_nosec(&file->f_path, &stat, STATX_SIZE, 0);
-	if (ret)
-		return 0;
+	if (S_ISBLK(file_inode(file)->i_mode)) {
+		loopsize = i_size_read(file->f_mapping->host);
+	} else {
+		struct kstat stat;
+
+		/*
+		 * Get the accurate file size. This provides better results than
+		 * cached inode data, particularly for network filesystems where
+		 * metadata may be stale.
+		 */
+		ret = vfs_getattr_nosec(&file->f_path, &stat, STATX_SIZE, 0);
+		if (ret)
+			return 0;
+
+		loopsize = stat.size;
+	}
 
-	loopsize = stat.size;
 	if (lo->lo_offset > 0)
 		loopsize -= lo->lo_offset;
 	/* offset is beyond i_size, weird but possible */