Browse Prior Art Database

Arbitration Logic

IP.com Disclosure Number: IPCOM000059955D
Original Publication Date: 1986-Feb-01
Included in the Prior Art Database: 2005-Mar-08
Document File: 3 page(s) / 52K

Publishing Venue

IBM

Related People

Granger, Y: AUTHOR

Abstract

In a machine environment where several independent devices can request simultaneously a service from a shared resource (memory, interrupt logic, etc.), the hereafter described logic (Fig. 1) allows the arbitration between the different requests to be performed with the following characteristics: - Arbitration resolution done in one LSSD (Level Sensitive Scan Design) clock cycle. - Automatic adaptation to the activity of the requesting devices, thanks to a simple mechanism. The arbitration logic only looks at the "Service Request" leads from the attached devices and needs no additional information on the device configuration (device installed, or not). The arbitration mechanism does not spend time to poll inactive or not installed devices.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 3

Arbitration Logic

In a machine environment where several independent devices can request simultaneously a service from a shared resource (memory, interrupt logic, etc.), the hereafter described logic (Fig. 1) allows the arbitration between the different requests to be performed with the following characteristics: - Arbitration resolution done in one LSSD (Level Sensitive Scan Design) clock cycle. - Automatic adaptation to the activity of the requesting devices, thanks to a simple mechanism. The arbitration logic only looks at the "Service Request" leads from the attached devices and needs no additional information on the device configuration (device installed, or not).

The arbitration mechanism does not spend time to poll inactive or not installed devices. - Capability of handling priorities, if priority levels are assigned to the requesting devices. The arbitration logic is mainly comprised of three parts (Fig. 2): a)

The Service Request Register 1, which samples and stacks the requests from the attached devices. Sampling is done with LSSD non-overlapped clocks, in order to present a stable set of requests to the other logic parts, which use the same clocks. b) The arbitration Ring Shift Register 2, which comprises as many stages as there are devices to possibly arbitrate. Any device is associated with a given stage of the Ring Shift Register. A pointer shifts in the register and stops when it reaches the stage dedicated to the device to be selected. c) The Combinational Logic 3 which controls the Ring Shift Register.

Fig. 3 shows the Arbitration Logic Diagram, and Fig. 4 provides the timing diagram. At initialization (reset), the Shift Register pointer is set in stage 1 of the register, and the pointer stays in this position as long as requests are not presented to the arbitration logic. When one or several requests are loaded in the Service Request Register 1, and assuming that the requested resource is available, the arbitration logic operates and causes the pointer to be shifted to the stage corresponding to the device to be served. The selected device is the first requesting device found by the pointer in the predefined rotation way. The pointer stays in this position during the time needed by the selected device to per...