Partition Logic User Manual


Partitions are a way of dividing your physical hard disk up into smaller, logical sections (which become 'logical' disks to your operating system).  For example in Windows, the first partition is your C: drive, the second is your D: drive, etc.

Windows systems generally arrive configured such that there is only a single, large partition (C:) covering the entire physical disk.  It is often better to have more than one partition, for example a partition for Windows and software, and another partition for data and documents.  This makes it possible, for example, to reinstall Windows on the first partition without disturbing your data.  Another reason for needing more partitions is when a user wants to install other operating systems (such as Linux) on the same disk and choose which one to run at boot time.  Partition Logic and other tools like it enable you to shrink your C: partition and create more partitions in the empty space.

Partition information resides in a table, the first part of which is typically located in the first sector(s) of your disk.  The table is also known as a disk label.  There are different schemes for the format of this table, but on PC-compatible computers the primary table resides in the first sector, known as the MBR (Master Boot Record) sector.  Thus this scheme is alternately known as MS-DOS partitioning or MBR partitioning.  Other disk label types include Sun/BSD and EFI.  As Partition Logic only supports MS-DOS partitioning at present, the rest of this description relates only to MS-DOS.


Partition tables in the MS-DOS partitioning scheme contain space for 4 partition entries.  The first sector of an MS-DOS-labeled disk contains some executable code and data, and the 'primary' partition table.  The table entries are each 16 bytes in length, and describe the starting sector of the partition, the partition size in sectors, the partition's 'tag' (which is used as an indication of the partition's filesystem type), a flag indicating whether the partition is 'bootable', and some legacy information about the partition's geometry (its starting and ending head, cylinder, and sector).  Nowadays the geometry information is not especially relevant, and the fields are not big enough to describe large partitions on modern disks, but partitioning programs like Partition Logic attempt to ensure that they are correct nonetheless.  Modern operating systems like Windows and Linux mostly ignore the geometry information.

The MS-DOS partitioning scheme is fairly old and obsolete, though it is the most common type and used on all current PC-compatible machines.  Newer computers such as Intel Itanium and Apple's Intel-based Macintosh systems use the more modern EFI GPT disk label.  This scheme may migrate to PC-compatible computers in the next few years.


A number of years ago it became apparent that many people would want to use more than the 4 partitions allowed in the MS-DOS primary partition table.  Thus software engineers dreamed up an extension to this scheme which, while terribly messy and sometimes inconsistently implemented, allows for a large number of partition entries.  This extension makes use of 'extended' partition tables containing 'logical' partitions (the ones in the primary table are known as primary partitions).

A user may create either:

  • 0-3 primary partitions, and some number of logical partitions; or

  • 4 primary partitions

If you have 4 primary partitions, there is no more room in this scheme for creating additional partitions.  This is so because one primary partition entry is required to describe the first extended partition, if applicable, which will contain all logical partitions.

When a user creates the first logical partition, the primary partition entry for the extended partition is created.  This primary extended partition 'wraps around' all logical partitions, and space for a new, extended partition table is reserved at the beginning of the extended partition.  Then the entry for the logical partition is placed in the extended table.  By convention, extended partition tables use only 2 entries; thus when a second logical partition is created, a second extended partition is created to wrap around it (inside the first one), and the remaining free entry in the first extended table describes the second extended partition.  The table entry for the second logical partition resides in a new table at the start of the second extended partition, and so on.  In other words, for each logical partition, there is an extended partition which contains it, and extended partitions 'nest' within one another:

Nested extended and logical partitions.  3 primary partitions, followed by an extended partition, which contains a) a logical partition; and b) a further extended partition containing a second logical partition.

P=primary, E=extended, L=logical

This scheme is convoluted, and is sometimes implemented differently depending on the partitioning software used.

Copyright © 1999-2013 J. Andrew McLaughlin | Visopsys and are trademarks of J. Andrew McLaughlin.    Contact