The High Performance File System (HPFS) is an installable file system. It is a hierarchical system and supports multiple directories. In many cases, accessing files under HPFS is faster than accessing similar files under the FAT file system. During installation of the OS/2 operating system, users can install the HPFS on the hard disk they use to start their computer.

Features of HPFS include:

• Caching of directories, data, and file system data structures
  
• Multi-threaded I/O operations
  
• Write-behind logic
  
• Optional write-through
  
• Strategic allocation of directory structures
  
• Highly contiguous file allocation
  
• Enhanced recoverability
  
• Extended attribute support
  
• Long file name support
  
• Starting the OS/2 operating system from an HPFS disk

File names under HPFS can contain 255 characters (one must be the terminating NULL, "\0") and can contain characters that are not valid for the FAT file system-for example, spaces. Each element of a path name residing on an HPFS disk can also have up to 255 characters. The total path including drive, directories, and file name cannot exceed 260 characters (259 with the terminating NULL). For more information on long file name support by installable file systems see Long File Names.

HPFS provides extremely fast access to very large disk volumes. HPFS uses a memory cache divided into blocks of 2KB. Data that is read from and written to the disk is transferred through this cache. Frequently-used data will often be found in the cache, thereby saving the time that a disk-read operation would require. When a user request specifies data that is not present in the cache, HPFS selects the least-recently used discardable block and then fills the block with the requested data.

When a write-data request is received, it usually is not necessary that the data be immediately written to the disk. HPFS will copy such data into the block cache without actually performing the disk-write operation. When the data is in the cache, it is written to disk as a background activity (referred to as lazy writing) which enables the typical user-write operation to occur much faster than in file systems where all write operations are synchronous.

The High Performance File System consists of:

• The High Performance File System driver, HPFS.IFS 
  The High Performance File System lazy-write utility, CACHE.EXE
  
  The High Performance File System lazy-write startup program, STARTLW.DLL
  
  The High Performance File System utilities, UHPFS.DLL
  The user determines the amount of lazy-write support by setting the following parameters on the command line that calls CACHE.EXE:

  • MaxAge:

    When the data in a cache block exceeds the specified time the block is queued for writing to the disk. This reduces the amount of data lost due to inopportune system shutdowns.

  • DiskIdle and BufferIdle

    When no user I/O request (non-lazy-write) has been made for DiskIdle number of milliseconds, all cache blocks (in oldest-first order) that have not been touched for BufferIdle number of milliseconds are queued for writing to disk. This enables HPFS to write out user data during times of relative disk inactivity and to reduce the need for rewriting heavily used cached blocks.

  STARTLW.DLL contains the code that starts the lazy-write thread.

  ---

  [Back] [Next]