wiki:Developer/OpenProjects

Version 341 (modified by C Rempel, on Jun 20, 2013 at 9:21:19 PM) (diff)

/* Debugging */

Open Projects

Welcome! Whether you're here because of Summer of Code (SOC) or just want to scratch an itch to hack, we invite you to peruse our projects and ask about them on the RTEMS mailing list or IRC. If you plan to submit a proposal to do something for the RTEMS Project as part of a SOC, see Getting Started for SoC Students. RTEMS projects span kernel hacking, adding support for a new board (BSP), improving the development environment, developing tests, and more.

If you want to get your feet wet with RTEMS then check out our small projects page? where you can find projects that require little coding skill and are appropriate for those new to RTEMS or open source software projects. If you are interested in one of these projects but are not able to code and test it yourself, consider sponsoring one of the core RTEMS developers to do it for you. Volunteering or sponsoring is how things get done -- users keep RTEMS development alive!

Most of these projects will take between a few weeks and a few months of effort by a person who is familiar with the general use of GNU/Linux and GNU tools. Many RTEMS projects are done by student or volunteer coders, so we try to define small projects or subtasks that can be completed and committed individually. Most of the projects are feasible as a Summer of Code project. Since some projects have multiple steps, students should work with prospective mentors to define the scope of work in their proposal. Similarly, some projects might be a starting point for a class project or graduate thesis.

Overview

The order of projects in the list does not reflect their importance, difficulty, or feasibility. Our project list is not exclusive: if you have an idea, solicit feedback from the project's mailing list or IRC channel; many developers sit in IRC and check it (and their email) infrequently throughout the day, so be patient! There may or may not be enough work on a project to constitute an SOC project, and some of these are past SOC projects. If you are interested in one of these, please ask on the mailing list or IRC.

Projects in Bold text are higher priority simply meaning that users or developers have expressed a lot of interest in such projects. Projects in Italic text have had some work done and may need an updated description, might be complete already, and may or may not have sufficient work remaining.

If you have a new project add it to the appropriate list below, link to a wiki page that follows the recommended Open Project Template?, and briefly (1-2 sentences) summarize the project and how it will improve RTEMS.

Testing

Testing a large body of software like RTEMS is in a continual state of improvement. There is always a need for more test cases and easier ways to run them all and decode the results. In addition, we want to be able to run all tests on as many hardware and simulator configurations as possible. Testing doesn't sound exciting to most people but when you combine the breadth of what we need to test with our desire for 100% instruction and branch path coverage, you get some very interesting and challenging work.

Some of the identified activities which would augment our testing capabilities are listed here:

  • RTEMS Testing?. Testing for RTEMS, including Unit, Regression operational and building a custom test harness. No prior knowledge of software testing is required. (SOCIS)
  • Test Coverage Analysis? - Improve coverage by adding more test cases. Eliminating dead code and reaching 100% coverage helps reduce the likelihood of new and recurrent bugs. (SOCIS)
  • Test the POSIX FIFO Implementation?.
  • Add support for gprof output to covoar?
  • Improve Testing of the GNU Tools? on RTEMS targets
  • POSIX Timing Tests?
  • Add support for gcov output? to covoar so tools like gcov and lcov can be used with RTEMS (another page here)
  • POSIX Compliance Test Suite?
  • Simulator Updates - This is a new idea for a project and has not been described beyond what is here. Ask on the rtems-devel mailing list for more advice. The RTEMS Project relies heavily on multiple FOSS simulators including qemu and skyeye. We have multiple BSPs which are tested on simulators. Each BSP will need to be tested against the current release and development version of each simulator. If there is a problem, then the issue will have to be identified (possibly using something like git bisect) and the student will have to fix it in the simulator. As an example of this, we believe some ARM BSPs which worked on older versions of Skyeye no longer work on the current version. In some cases, the RTEMS Project is carrying unmerged patches for a simulator (at least qemu m68k and arm) which need to be updated to their development head and submitted. You will try to get the simulator developers to include RTEMS hello world and ticker executables for BSPs their simulator supports in their simulator tests to avoid future breakage. And you will ensure that the scripts in rtems-testing/sim-scripts are working for the latest version of the simulator. We would also like instructions on running the leon BSPs on qemu. Please ask questions. This is a broad project. (SOCIS)
  • RTEMS Test Template? - Improve by adding functionality
  • RTEMS Test Specification? - Need a formal test specification. Ideally, inspired by http://softwaretestingstandard.org/ some notes about existing test specifications can be found http://www.rtems.org/wiki/index.php/GoogleCodeInProjects#Test_Documentation_Files_update .
  • RTEMS Test Screen Validation? - This requires expertise in a unit testing framework, probably DejaGnu?, and will probably need to be tied into rtems-testing.

Tools and Development Environment

RTEMS applications are cross-compiled on a development host to produce executables that are transferred to and executed on target systems. The projects in this section focus on the host side of that equation. This means they will run on a developer's computer and possibly communicate with embedded hardware.

The following areas have been identified for projects related to improving RTEMS development:

<!-- * Integrate CEXP? into main RTEMS distribution. Possible licensing issues, please apply to the RTL project instead -->

RTEMS Run-Time Oriented

The projects in this category are more focused on the development of software that runs on RTEMS on target hardware.

Executive (SuperCore?, SuperCoreCPU, libcpu): a.k.a. kernel

  • Improvements to SMP support? - Multiprocessing is of increasing importance in modern systems and we want RTEMS to remain competitive and useful. This is a large project and subtasks should be identified before writing any proposal.
    • Implement or integrate Atomic Operations. SMP code requires additional synchronization primitives that are not available currently in RTEMS.
  • Unified Interrupt and PCI APIs -- UnifiedAPIs?
  • kqueue(2) or taskqueue(9) is a project to port the kqueue(2) or taskqueue(9) API from FreeBSD.
  • Add cache manager support for architectures not having it.
  • TinyRTEMS is an umbrella term that corresponds to any activities or ideas that could shrink the code and data space requirements for RTEMS. The goal is to progressively lower the minimum CPU requirements.
  • Improve the RTEMS SuperCore Scheduler?
  • Paravirtualization? of RTEMS to make it suitable to be run as a guest OS in a hypervisor.
  • MMU Support? for RTEMS.
  • RTEMS Sequenced Initialization is a project to allow RTEMS initialization to be dynamically constructed based upon user requirements. It would be like C++ global constructors conceptually.
  • ISO9660 file system
  • Condition Variables? for the Classic API.

Debugging

Board Support Package (BSP)

  • RTEMS can always use more BSPs for Simulators?. Being able to test, debug, and perform coverage analysis on simulators is critical to the ongoing success of the project.

API Layers (POSIX, Classic, SAPI)

rtems-libbsd

  • Update the RTEMS TCP/IP stack - The networking stack is old and showing it. This project actively underway. At a high level, this effort requires porting the TCP/IP stack and providing support functional equivalents of multiple BSD kernel constructs. This project has many subprojects many of which are appropriate for SOC. It would be of great usefulness to the community to get as many of these does as an SOC project as possible.
  • port BSD USB stack

Languages

V8 is high performance JavaScript? engine. It was writen in C++ and it can be embedded into a C++ application. With V8 it would be easier for user to build web server with javascript and even run js framework node.js.

Libraries and Applications

  • Identify and implement the functionality currently missing in dup()
  • Port Transparent IPC (http://tipc.sourceforge.net/index.html) to RTEMS
  • Implement a Simple Line Editor. Existing code can be refactored for starting point.
  • Dynamic Object File Loading lets a base application with RTEMS dynamically load the rest of the application. The dynamic parts can be optional features and never loaded, or upgraded replacements for parts of the application.
  • RTEMS Toolkits - We are defining collections of libraries and support programs which make it easier to get started for certain types of applications. We haven't identified all potential toolkits or components. Each potential component must be evaluated for license and appropriateness for use in an embedded environment like RTEMS. We also should define some guidelines about creating and maintaining toolkits in general. Here are the toolkits areas identified so far:
  • Turn the current port of LWIP into a first class citizen. Submit port, make target independent, create maintenance plan.
  • IDL/COM? Support for RTEMS.
  • Make Addon Package? - write a tool to help other developers with Libraries, Languages, and Applications

Retired Projects

The following projects are complete or pending.

Obsolete Projects

Some projects have been proposed that are viewed as being of minor use. This list is meant to provide a way to avoid wasted effort on projects that are not widely desired. However, projects on this list might still be useful to someone, given a motivated individual to work on them.