Hi. I'm a bit confused about something, and was hoping for some clarification. I created a raw image file (a zeroed-out file) and used fdisk to partition it. However, even though it asks me exactly how large I want the partition to be, the actual size of the partition ends up being larger (as reported by fdisk itself). E.g., if I request 100M partition, it ends up being 105M. I'm confused by this because I didn't think that a partition had any kind of "add-on" structure; they are just space on the disk, with start and end locations recorded in the partition table, correct? So why does fdisk need to make my partitions larger than I requested? Or is there something funky here about the way fdisk calculates and reports the size?

I suspect you are falling victim to binary measurements ( http://en.wikipedia.org/wiki/Binary_prefix ) where
2^20 = 1048576 or about 1.05 * 1000000