Browse Prior Art Database

Asynchronous Low-Level Management Functions

IP.com Disclosure Number: IPCOM000113091D
Original Publication Date: 1994-Jul-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 4 page(s) / 148K

Publishing Venue

IBM

Related People

Mendelson, RN: AUTHOR [+2]

Abstract

Disclosed is a method for performing asynchronous low-level management functions across the Micro Channel* bus of a system unit in which the Management Focal Point does not reside.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 37% of the total text.

Asynchronous Low-Level Management Functions

      Disclosed is a method for performing asynchronous low-level
management functions across the Micro Channel* bus of a system unit
in which the Management Focal Point does not reside.

      In a computing system, there are many system initialization
functions which must be performed before the system can perform
useful work.  These functions include verification of the hardware
configuration, error logging, and the initialization of certain
hardware assemblies and software routines.  In some systems, these
functions are managed and coordinated from a central point, such as
the Management Focal Point (MFP) of an IBM PS/2* computer system.

      In a normal mode of operation, the MFP communicates with the
Managed Units by means of delivery pipes, having one pair of delivery
pipes for each Managed Unit.  Once the pipes are initialized, the MFP
communicates with the Managed Units using control elements passed
through the delivery pipes.  Control elements are used for all
management functions, including the initialization of delivery pipes
between Managed Units.

      However, one of the functions that cannot be performed through
the delivery pipes is called "unit bringup," which includes the
initialization of the pipes themselves.  In order to initialize the
delivery pipes, a set of low-level primitives is defined, allowing
the MFP to communicate with the Managed Units.  These low-level
management functions essentially pass management control elements to
a Managed Unit, to initialize the delivery pipes or other parts of
the Managed Unit requiring initialization.  MFP services need to be
able to send these low-level requests asynchronously to multiple
Managed Units and to receive and process replies asynchronously,
allowing overlapped processing of initialization requests by the
Managed Units to greatly improve the performance of initialization.

      Previously, the management of unit bringup had been handled by
services located within the System Unit, due to an inability of
bringup services to notify Managed Units of the location of such
services before initialization.  The Managed Unit had to know the
location of the MFP to direct replies back to the MFP in response to
low-level management requests.  Thus, in previous applications, the

Managed Unit, assuming that the MFP was located in the System Unit,
directed all replies to the System Unit.

      However, under some circumstances, it is desirable for the MFP
to be located in a unit other than the System Unit.  The MFP may
reside in a coprocessor.  Also, in the event of failure of the System
Unit, an alternate MFP could be defined in some other coprocessor.

      The following new Attention Codes are defined to allow the MFP
to reside outside the System Unit:

ATTN CODE   DESCRIPTION
 x'90'       Assign Signal Mechanism for low-level management
requests.
             The Command Int...