RapidIO: The Next Generation Communication Fabric For Embedded Application

7.6. ERROR DETECTION AND RECOVERY

One of the main applications targets of the RapidIO technology is highly reliable electronic equipment such as carrier class telecommunications equipment and enterprise storage systems. In these systems, it is extremely important to do everything reasonably possible to avoid errors and, when errors occur, to detect and attempt to recover from them. The RapidIO interconnect technology provides extensive error detection and recovery techniques by combining retry protocols with cyclic redundancy codes. RapidIO also employs carefully selected delimiter control characters and response timers to reduce the likelihood of undetected errors in systems.

7.6.1. Lost Packet Detection

Some types of errors, such as a lost request or response packet or a lost acknowledgment, result in a system with hung resources. To detect this type of error time-out counters are used. Because the expiration of one of these timers should indicate to the system that there is a problem, this time interval should be set long enough so that a false time-out is not signaled. The response to this error condition is implementation dependent.

The RapidIO specifications require time-out counters for the physical layer, which are called the port link time-out counters, and counters for the logical layer, which are called the port response time-out counters. The interpretation of the counter values is implementation dependent, based on a number of factors, including link clock rate, ...