Comment 7 for bug 2037569

after exploring udisks2/libblockdev/libfdisk, I have three possible proposals for resolution:

1. hack udisks2 udisks_linux_partition_table_update
  * calling bd_part_get_disk_spec is always going to be a loop, unless
    * watches have been inhibited
    * or calling bd_part_get_disk_spec causes
      ID_PART_TABLE_TYPE to be set
  * disable everything under `if (!part_type && num_parts > 0)`, log
    and bail
2. add read-only API in libblockdev so we open the dev RO in libfdisk
3. add vtoc support to libfdisk + enable the s390 bits in libblockdev

only option 1 sounds plausible to me, though, for a Mantic timeframe.