MMM       MMM
         MMM       MMM
          M M     M M
          M  M   M  M
          M   M M   M
         MMM   M   MMM
  {e/si} MMM       MMM ulator

This project is a highly extensible {ARM} {e,si}mulator. It is capable of both simple simulation of various ARM cores (currently M0, M3) or in-circuit emulation (currently the Michigan Micro Mote platform)

In addition, this project is used as a teaching tool for embedded systems courses (currently at University of Michigan and University of Utah), both to understand internal core design and higher-level MCU usage.

• The simulator/ directory contains the code to simulate a processor. The simulator is built using tup, a make replacement. For details, please see simulator/README.build and the tup homepage: http://gittup.org/tup/ Core selection is achieved using tup variants. By default, a pipelined version of every supported core is created, see the simulator/configs/ directory for other options.

  • The simulator is licensed under the GPLv3

• The platforms directory contains additional support for the various cores. In particular, it contains software desgined to run on each of the unique cores supported by the simulator. In addition, the M3 platform contains some tools for programming M3 chips and in-circuit emulation.

  • The platforms are licensed under the MIT/XWin license

• The docs/ directories hold the course materials and some amount of dated information on the simulator design. It is not particularly useful currently

  • Documentation, course materials licensed under the MIT/Xwin license

• There is little documentation currently available. Hopefully things are mostly intuitive / have good error messages. Github issues are appreciated if there are any particular pain points that need addressing.
• Interrupts may be unreliable -- the code path is not heavily tested, though no known bugs currently exist.
• Course documentation is out of date, does not reflect current projects
  • Or even all of the original projects for that matter

• General clean-up of configuration options
• Larger / more interesting peripheral library
• Better gdb / recovery semantics
  • This likely correlates with BusFault-style work
  • This is why there is a temporary imbalance in memory semantics, supporting the try_read_byte function for gdb