1.2.1. How the Multi-ICE server uses RPC

Multi-ICE uses the Open Network Computing (ONC) version of the RPC software. Because RPC uses TCP/IP as its transport mechanism, it is just as easy to connect to a Multi-ICE server over a network or modem as it is to connect locally. Sources for the client-side of the RPC-based TAPOp interface can be found on the installation disk.

When connecting to a Multi-ICE server running on the same workstation, a Multi-ICE client uses a shared memory procedure call mechanism. Because the difference between the mechanisms is hidden, this manual refers only to RPC connections.

The operation of the TAPOp interface follows this pattern:

  1. The server starts and is configured to use a particular target device.

  2. A client opens a connection to the required device.

  3. The client makes RPC calls to the server.

  4. The server scans data through the scan chains of the device and returns any results.

  5. The client disconnects.

The client can cycle through steps 2, 3, and 4 as many times as required to perform the task.

Each TAPOp call performs one TAP operation. For example, there are TAP operations to write a value to the Instruction Register (IR) or read a device scan chain. Because of the low level of the interface and the high overhead of RPC calls, you can batch up multiple RPC calls into macros that are run by the server. This is similar to JAVA applets downloaded from a web server to a client (browser), because the browser is faster than the link. In Multi-ICE, the client downloads macros to the server because the server is faster than the link. This gives a significant performance improvement.

Copyright © 1998-2002 ARM Limited. All rights reserved.ARM DUI 0154B
Non-Confidential