Primary vs. Extended Partitions
A discussion of primary vs. extended partitions
Primary partitions have a disadvantage: You only have four of them. (Comes from my earlier days... 64k will always be sufficient... why more that 4 partitions on a 10MB disk?) Therefore, with bigger disks a specific type of primary partition was defined - the "extended partition", which is nothing more than a construct to create more partitions (itselves sometimes called extended partitions, sometimes "logical drives"). But still the space defined for the "extended partition" (the one in the primary partition table) is valid for the sum of all sub-partitions.
Primary partitions are old-school... I guess nowadays you get away with defining one huge "extended partition" and creating lots of sub-partitions in there. Technically speaking extended partitions take time to scan, since an auto-created sub-partition table is at the beginning of each sub-partition, with a single entry plus a pointer to the next sub-partition... but that's no real show stopper: How often do you need to scan those tables?
If you know that you will have four or less partitions on the disk, stick with simple primary partitions. If unsure, let at least p4 be an extended partition with the remainder of the disk.
Just my few cents ;-)