Method and apparatus for automatic generation of I/O configuration descriptions

• US 5,671,441 A
• Filed: 11/29/1994
• Issued: 09/23/1997
• Est. Priority Date: 11/29/1994
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Johnson, Francis Edward, Halma, Marten Jan, Glassen, Steven Gardner, Hefferon, Eugene Paul
• Primary Examiner(s): Harvey, Jack B.
• Assistant Examiner(s): PHAN, RAYMOND NGAN
• Application Number:
  US08/346,642
• Time in Patent Office: 1,029 Days
• Field of Search: 395/856, 395/857, 395/858, 395/862, 395/885, 395/828, 395/825, 395/182.02, 395/200.15, 395/830, 395/831, 395/838, 395/839, 395/200.13, 395/200.12, 395/670, 395/674, 395/700
• US Class Current: 710/8
• CPC Class Codes: G06F 9/4411 Configuring for operating with peripheral devices; Loading of device drivers

Abstract

Automatic machine methods and apparatus for determining which components of an I/O configuration are shared by other components of the configuration. The information can be obtained through the use of existing self-description facilities and unique identifiers. By noting which channel paths are used to obtain configuration-data records and examining the unique identifiers provided for each I/O items it can be determined which I/O devices are accessible through the same control unit, and which control units provide access to the same I/O device. Furthermore, by examining the unique identifiers provided, it can be determined which I/O subsystems and which control units or channel subsystems are accessible through the same dynamic switch and which dynamic switches provide access to the same I/O subsystem of channel subsystem. A programmable subchannel is provided to enable communication between a CEC and an electrically-connected I/O item not represented in the current I/O configuration description being used by the CEC. Without the programmed subchannel, the CEC cannot access any I/O item not represented in the CEC'"'"'s current I/O configuration description.

Claims

1. A method in a computer system with a central electronics complex (CEC), having at least one I/O channel subsystem, a plurality of channel paths and a plurality of I/O items, said I/O items being attachable to said subsystem via said channel paths in different ways as to provide a plurality of different system configurations, the method performed by the CEC for accessing information associated with said I/O items, said method comprising the following CEC-performed steps:

structuring in the CEC an amorphous subchannel not defined in a current I/O configuration definition for the I/O channel subsystem,

using the amorphous subchannel to test for the existence of each I/O item connected to one or more channel paths of said CEC by simulating a plurality of I/O item addresses by the amorphous subchannel without the CEC using the current I/O configuration definition for accessing the I/O item,

attempting to access the I/O item addressed by the using step,

determining characteristics of each of said I/O items by accessing self-description information provided for each of said I/O items;

determining the channel paths usable to access each of said I/O items;

storing the self-description information about said attached I/O items and storing information for each channel path connection to each I/O item in an I/O item configuration record;

examining said I/O item configuration record to determine if any of said I/O items are shared by said channel paths or said I/O items, and

using said I/O item configuration record to build an I/O configuration definition, wherein the amorphous subchannel previously contained no fixed I/O configuration definition information, and allows information about said I/O items to be stored, modified, or retrieved when I/O commands are issued.

2. The method of claim 1, wherein said self-description information provided by each of said I/O items has a unique identifier including type and distinguishing characteristics of said I/O item.

3. The method of claim 1 wherein said amorphous subchannel is a programmable subchannel used to contain parameters to enable use of the programmable subchannel for obtaining self-description information from I/O items;

and

said programmable subchannel having no fixed I/O configuration definition of its own.

4. The method of claim 1, wherein each of said channel paths has a channel identifier (CHPID) usable by said CEC;

and information exits about which CHPIDs are usable by said CEC.

5. The method of claim 1, wherein said I/O items and their channel paths are located by executing a program.

6. The method of claim 1, wherein I/O items are tested using the entire range of device addresses for a single link destination address and a single channel address, prior to testing a next link destination address, wherein the entire range of link destination addresses are tested prior to testing a next channel path address until all channel path addresses have been tested.

7. The method of claim 3 wherein an information gathering step is performed during which said parameters in said programmable subchannel are used to select each I/O item from which information is to be retrieved;

and

said information gathering step also using said parameters to enable the use of said programmable subchannel to determine if said I/O item is a dynamic switch or a control unit.

8. The method of claim 3, wherein said I/O items honor a Sense ID command independent of any type of I/O item.

9. The method of claim 3, wherein said programmable subchannel is programmed, and I/O commands are issued to each I/O item to determine presence and characteristics of each I/O item.

10. The method of claim 2, wherein said unique identifier provided by each of said I/O items

provides information about a node at which point said I/O item is attached to each of said channel paths;

said unique identifier identifying said node by the use of a plurality of node descriptors (NDs), a plurality of associated node qualifiers (NQs), a plurality of node-element descriptors (NEDs), and a plurality of associated node qualifiers (NEQs); and

said self-description information being retrieved by a plurality of commands to read said configuration data, set a plurality of interface identifiers, and read and store a node identifier.

11. In a computer system with a central electronics complex (CEC), a plurality of subchannels and a plurality of connected I/O items, a programmable subchannel initially having no fixed I/O configuration definition, said programmable subchannel being capable of containing parameters and being independent of an I/O configuration program (IOCP), said programmable subchannel further comprising:

a storage location for storing configuration information about a plurality of I/O items;

said configuration information being usable to establish logical paths which enable the CEC to use said I/O items; and

means for allowing information about said I/O items to be stored, modified, or retrieved when I/O commands are issued.

12. The programmable subchannel of claim 11, wherein a said programmable subchannel is identifiable by said system by a special type field provided in a subchannel information block (SCHIB) within said programmable subchannel.

13. The programmable subchannel of claim 11, containing addressing information for routing data between a CEC-resource logical partition and an I/O item.

14. The programmable subchannel of claim 11, wherein after said programmable subchannel contains configuration information, I/O commands are issued to said I/O item to determine presence and characteristics of said I/O item.

15. A mechanism for locating I/O items and their associated channel paths in a Central Electronic Complex (CEC) of a computer system, said mechanism comprising:

at least one channel subsystem attaching said I/O items via said channel paths in different ways to provide a plurality of different system configurations;

self-description information being provided by each of said I/O items;

means for collecting data associating said I/O items with channel paths;

means for collecting characteristics information for said I/O items;

said storage location,

a storage location for storing a configuration description indicating a plurality of I/O items and their characteristics and their attachment to said channel subsystem by channel paths, wherein said mechanism does not depend on any prior fixed I/O configuration definition in the system, and means for using said storage location to build an I/O configuration definition.

16. The mechanism of claim 15, further comprising means for examining said configuration description for determining if any of said I/O items share any of said channel paths, and if any of said channel paths share said I/O items.

17. The mechanism of claim 15, wherein each of said I/O items provide a unique identifier including type and characteristics of said I/O item;

each of said I/O items provides self-description information about a node by the use of a node descriptor (ND), a plurality of associated node qualifiers (NQs), a plurality of node-element descriptors (NEDs), and a plurality of associated node-element qualifiers (NEQs); and

said CEC including a plurality of commands to read said configuration description, set a plurality of interface identifiers, and read and store a node identifier associated with each of said I/O items.