Avalon Memory-Mapped Interface Definition
The Avalon® Memory-Mapped (Avalon® -MM) interface is an interface protocol for use in connecting master and slave components in an Qys system. The protocol connects address-based read/write interfaces typical of an Avalon® memory-mapped master that usually controls a number of Avalon® memory-mapped slave peripherals. The typical Avalon® -MM master is a microprocessor; typical slaves include: memories, UARTs and timers. In contrast to the Avalon® -MM interface, the Avalon® Streaming (Avalon® -ST) interface protocol is used for very high-bandwidth, unidirectional traffic connecting source-sink pairs.
The Avalon® -MM interface defines:
- A set of signal types
- The behavior of these signals
- The types of transfers supported by these signals
For example, you can use the Avalon® -MM interface to describe a traditional peripheral interface, such as SRAM, that supports only simple, fixed-cycle read/write transfers. On the other hand, the Avalon® -MM interface can also be used to describe a more complex pipelined interface capable of burst transfers.
Some of the prominent features of the Avalon® -MM interface are:
- Separate Address, Data and Control Lines—Provides the simplest interface to on-chip logic. By using dedicated address and data paths, Avalon® -MM peripherals do not need to decode data and address cycles.
- Up to 1024-bit Data Width—Supports even-power-of-two data paths, up to 1024 bits.
- Synchronous Operation—Provides an interface optimized for synchronous, on-chip peripherals. Synchronous operation simplifies the timing behavior of the Avalon® -MM interface, and facilitates integration with high-speed peripherals.
- Dynamic Bus Sizing—Handles the details of transferring data between peripherals with different data widths. Avalon® -MM peripherals with differing data widths can interface easily with no special design considerations.
- Simplicity—Provides an easy-to-understand interface protocol with a short learning curve.
- Low resource Utilization—Provides an interface architecture that conserves on-chip logic resources.
- High performance—Provides performance up to one-transfer-per-clock.