Attached is a updated tested patch.
The ext2fs_default_journal_size logic is stolen from lib/ext2fs/mkjournal.c, where the input and output is in blocks. Generally the default block size is 4096 (/etc/mke2fs.conf) so I just multiplied the numbers with 4096. I call ext2fs_default_journal_size twice, to ensure that JOURNAL_SIZE is correct, as the "final size" could require a bigger journal size. df report the "overhead" as: /dev/loop2p1 571204 547844 11424 98% /mnt - Kristian Klausen |