High Performance File System
This article needs additional citations for verification. (August 2013) (Learn how and when to remove this template message)
|Full name||High Performance File System|
|Introduced||November 1989OS/2 1.2with|
|Partition identifier||0x07 (MBR)|
|Directory contents||B+ tree|
|File allocation||B+ tree|
|Bad blocks||B+ tree|
|Max. volume size||64 GiB (as implemented)|
2 TiB (theoretical)
|Max. file size||7.68 GiB|
|Max. number of files||Unlimited|
|Max. filename length||255 characters|
|Allowed characters in filenames||Double-byte from 0x0020 to 0xFFFF|
|Dates recorded||Access, Creation, Modified|
|Attributes||Read-only, hidden, system, archive|
|File system permissions||Yes (only in HPFS386)|
|Supported operating systems||OS/2, Windows NT, Linux, DragonFly BSD, eComStation, ArcaOS|
HPFS (High Performance File System) is a file system created specifically for the OS/2 operating system to improve upon the limitations of the FAT file system. It was written by Gordon Letwin and others at Microsoft and added to OS/2 version 1.2, at that time still a joint undertaking of Microsoft and IBM, and released in 1988.
Compared with FAT, HPFS provided a number of additional capabilities:
- Support for mixed case file names, in different code pages
- Support for long file names (255 characters as opposed to FAT's 8.3 naming scheme)
- More efficient use of disk space (files are not stored using multiple-sector clusters but on a per-sector basis)
- An internal architecture that keeps related items close to each other on the disk volume
- Less fragmentation of data
- Extent-based space allocation
- Separate datestamps for last modification, last access, and creation (as opposed to last-modification-only datestamp in then-times implementations of FAT)
- B+ tree structure for directories
- Root directory located at the midpoint, rather than at the beginning of the disk, for faster average access
IBM offers two kinds of IFS drivers for this file system:
- The standard one with a cache limited to 2 MiB
- HPFS386 provided with certain server versions of OS/2, or as added component for the server versions that did not come with it
HPFS386's cache is limited by the amount of available memory in OS/2's system memory arena and was implemented in 32-bit assembly language. HPFS386 is a ring 0 driver (allowing direct hardware access and direct interaction with the kernel) with built-in SMB networking properties that are usable by various server daemons, whereas HPFS is a ring 3 driver. Thus, HPFS386 is faster than HPFS and highly optimized for server applications. It is also highly tunable by experienced administrators.
Though IBM still had rights to HPFS, its agreement with Microsoft to continue licensing the HPFS386 version is contingent upon the company paying Microsoft a licensing fee for each copy sold. This was a result from the Microsoft and IBM collaboration that both IBM and Microsoft had right to use Windows and OS/2 technology. Microsoft used HPFS in Windows NT.
Native support under Windows
Windows 95 and its successors Windows 98 and Windows Me can read and write HPFS only when mapped via a network share; they cannot read it from a local disk. They listed the NTFS partitions of networked computers as "HPFS", because NTFS and HPFS share the same filesystem identification number in the partition table.
Windows NT 3.51 can also read and write from local HPFS formatted drives. Starting with Windows NT 4 the filesystem driver PINBALL.SYS enabling the read/write access is not shipped anymore. Later Windows versions do not ship with this driver. Note that this driver is limited to 4GB HPFS volumes.
Microsoft retained rights to OS/2 technologies, including the HPFS filesystem, after they ceased collaboration with IBM. Since Windows NT 3.1 was designed for more rigorous (enterprise-class) use than previous versions of Windows, it included support for HPFS (and NTFS) giving it a larger storage capacity than the FAT12 and FAT16 filesystems. However, since HPFS lacks a journal, any recovery after an unexpected shutdown or other error state takes progressively longer as the filesystem grows. A utility such as CHKDSK would need to scan each entry in the filesystem to ensure no errors are present, a problem which is vastly reduced on NTFS, which simply replays the journal.
- Ray Duncan (September 1989). "Design goals and implementation of the new High Performance File System". Microsoft Systems Journal. 4:5: 1–13.
- Bridges, Dan (November 1996). "Inside the High Performance File System - parts 1 to 6". Electronic Developer Magazine for OS/2. Vol. 4 no. 10. Retrieved 2016-07-18.
- Dieter Brors (1997). "HPFS unter Windows NT 4.0". c't Magazin für Computertechnik (German). 1: 306.
- Chris Graham. "Appendix G — HPFS internals". The Graham Utilities for OS/2 - Version 2. Archived from the original on 2006-02-10.
- How to get Windows NT 4, Windows 2000, and Windows XP to read HPFS partitions