btrfs: ioctl: return device fsid from DEV_INFO ioctl

Currently user space utilizes dev info ioctl to grab the info of a
certain devid, this includes its device uuid.  But the returned info is
not enough to determine if a device is a seed.

Commit a26d60d ("btrfs: sysfs: add devinfo/fsid to retrieve actual
fsid from the device") exports the same value in sysfs so this is for
parity with ioctl.  Add a new member, fsid, into
btrfs_ioctl_dev_info_args, and populate the member with fsid value.

This should not cause any compatibility problem, following the
combinations:

- Old user space, old kernel
- Old user space, new kernel
  User space tool won't even check the new member.

- New user space, old kernel
  The kernel won't touch the new member, and user space tool should
  zero out its argument, thus the new member is all zero.

  User space tool can then know the kernel doesn't support this fsid
  reporting, and falls back to whatever they can.

- New user space, new kernel
  Go as planned.

  Would find the fsid member is no longer zero, and trust its value.

Reviewed-by: Anand Jain <[email protected]>
Signed-off-by: Qu Wenruo <[email protected]>
Signed-off-by: David Sterba <[email protected]>

Author: adam900710

fs/btrfs/ioctl.c

@@ -2859,6 +2859,7 @@ static long btrfs_ioctl_dev_info(struct btrfs_fs_info *fs_info,
 	di_args->bytes_used = btrfs_device_get_bytes_used(dev);
 	di_args->total_bytes = btrfs_device_get_total_bytes(dev);
 	memcpy(di_args->uuid, dev->uuid, sizeof(di_args->uuid));
+	memcpy(di_args->fsid, dev->fs_devices->fsid, BTRFS_UUID_SIZE);
 	if (dev->name)
 		strscpy(di_args->path, btrfs_dev_name(dev), sizeof(di_args->path));
 	else

include/uapi/linux/btrfs.h

@@ -245,7 +245,17 @@ struct btrfs_ioctl_dev_info_args {
 	__u8 uuid[BTRFS_UUID_SIZE];		/* in/out */
 	__u64 bytes_used;			/* out */
 	__u64 total_bytes;			/* out */
-	__u64 unused[379];			/* pad to 4k */
+	/*
+	 * Optional, out.
+	 *
+	 * Showing the fsid of the device, allowing user space to check if this
+	 * device is a seeding one.
+	 *
+	 * Introduced in v6.3, thus user space still needs to check if kernel
+	 * changed this value.  Older kernel will not touch the values here.
+	 */
+	__u8 fsid[BTRFS_UUID_SIZE];
+	__u64 unused[377];			/* pad to 4k */
 	__u8 path[BTRFS_DEVICE_PATH_NAME_MAX];	/* out */
 };