Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Asynchronous Buffer Purge

IP.com Disclosure Number: IPCOM000121186D
Original Publication Date: 1991-Jul-01
Included in the Prior Art Database: 2005-Apr-03
Document File: 1 page(s) / 43K

Publishing Venue

IBM

Related People

Laverdure, AM: AUTHOR [+3]

Abstract

Disclosed is a process which allows writing altered I/O buffers to one or more datasets residing on one or more physical output devices, while minimizing the amount of wait time normally required to complete this process.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 86% of the total text.

Asynchronous Buffer Purge

      Disclosed is a process which allows writing altered I/O buffers
to one or more datasets residing on one or more physical output
devices, while minimizing the amount of wait time normally required
to complete this process.

      As implemented, asynchronous buffer purge initially scans a set
of output buffers and creates a list of entries for the altered
buffers to be written.  Each of these entries will contain,
minimally, the address of the buffer to be written, the UCB/CCHHR for
the buffer, and a forward pointer to the next entry.  Upon completion
of the scan, the entries are then ordered into UCB/CCHHR (device
address/track address) sequence.

      Once sequenced, the entries are again scanned, and for each
output device in the list, a unit of work is created and the address
of the first entry on the write chain for this device is stored in
the UOW and the forward pointer in the last entry in the chain for
this device is set to zero, and a WRITE request for this UOW is
issued.  This process is repeated until we have issued WRITE requests
for all of the devices in the list of entries.  The address of the
first UOW created is saved and all subsequently created UOWs are
chained.

      After all WRITE requests have been issued, the address of the
first UOW is retrieved and individual WRITE checks are issued for it
and all other UOWs in the chain.  At this time, we will also issue a
WAIT if I/O has not completed for a UO...