Leon2 Processor IP Core Technology (Part 3)

    Like general embedded system software development, Leon2 application system software development needs to be carried out in a cross-compilation environment. Leon2 application system software development mainly uses LECCS (Leon/ERC32 GUN cross-compilation system). LECCS is a multi-platform development system developed by Cynus, OAR and Gaisler Research. It is a freely available development system based on the GNU series with a series of additional "dot" tools. LECCS can run on multiple platforms such as Linux (-2.2.x or higher), Solaris (-2.7 or higher) and Windows (Cygwin-1.1.7 and higher). To run LECCS on the Windows platform, Cygwin must be installed. It is a Unix-level emulator.

    LECCS includes the following packages: GCC--3.2.3 C/C++ compiler, RTEMS--4.6.0-β C/C++ real-time kernel, independent C library, GDB-5.3 SPARC cross debugger, remote debugging monitor, GDB's DDD graphical front end (Unix only), GDB's GDB-TK image front end (Windows only), boot PROM builder and Leon debug support unit monitor, etc. It also includes some tools, such as cross assembler, assembly preprocessor, GNU linker, binary converter, etc.

The process of using LECCS to develop Leon2 application system software is as follows:

① Use GCC to compile and link the program;

②Use the simulator to debug the program;

③Debug the program on the remote target system;

④ Generate boot PROM program for independent application.

    LECCS supports two types of applications: one is the usual sequential CC++ program, and the other is a multi-tasking real-time CC++ program based on the REDEMS core.

   The software simulation tool can be TSIM. It is a general-purpose SPARC architecture simulator that can simulate ERC-32 and Leon-based computing systems. For performance and usage, please refer to the TSIM User Guide. Leon2 provides great hardware support for application system debugging, namely its debug support unit (DSU) and trace buffer. For specific usage, please refer to the Leon2 User Manual and Leon DSU Monitor User Manual.

4 Applications of Leon2

     Leon2 is used both at home and abroad.

    There are many applications abroad, such as digital dictation machines, fingerprint security systems, GPS processors, BIST (built-in self-test) and other products based on Leon. Here we briefly introduce the ThumbPod fingerprint security system of the University of California, Los Angeles, and the NJ1030 GPS baseband product of Nemerix. The ThumbPod security system is a security system that uses human fingerprints as identity authentication (digital identity). The hardware system of its prototype is composed of Xilinx's FPGA (XC2V1000), Micron's 256 MHz-DDR-SDRAM, Authentec's AF-S2 finger sensor, etc. The CPU uses the Leon2 processor core, and the AES (Advanced Encryption System) coprocessor is connected through the Leon2 coprocessor interface (CPI) to execute the AES algorithm to speed up the encryption speed. The maximum operating clock frequency of the system can reach 100 MHz, and the operating voltage can be 3.3/2.5/1.8/1.5 V and other voltages. The NJ1030 GPS baseband product targets GPS L1 carrier C/A code processing and provides support for both WAAS and EGNOS augmentation systems. It is compatible with the world's leading third-party GPS-RF front-end equipment.

    In China, the 771 Institute of the Ministry of Aerospace and the National Astronomical Observatory of the Chinese Academy of Sciences are currently developing or planning to develop satellite-based data processing systems based on Leon2. This data processing system no longer uses large, heavy and power-hungry processor chips (such as DSP), but directly puts Leon2 into FPGA. The system hardwareizes most of the processing functions and uses the rich resources of FPAG to put the entire system into a single FPGA chip. Leon2 only performs control management and some necessary calculations. The National Astronomical Observatory is planning to develop a satellite-based system based on Leon2, which will process 1,780 GB of data collected daily. These data are compressed to about 10 GB through pre-processing (image integration, radiation correction, geometric correction, magnetic map processing), wavelet compression, image format, etc., and then transmitted to the ground station.

Conclusion

    This paper introduces the technical features, software and hardware development process, and domestic and foreign application examples of the Leon2 open source processing core. The technical features are mainly SPARCV8 specification, AMBH2.0 bus structure, fault-tolerant design method and VHDL programming style; the software and hardware development tools and methods of the Leon2 application system are introduced; in the application examples, two foreign applications and two domestic applications are introduced, including the satellite image processing system that the author is currently developing.