Abstract for Freenix 2004
Bart proposed that the team submit a 3-5 page abstract to Freenix 2004, of which he and Keith are chairs. Possible topics:
- work rewriting the CanBusLinuxDriver (including CANtelope)
- overall progress on the FlightComputerSoftware since the Usenix paper.
Points to consider and areas of interest (from the above page)
- open source
- available online: Twiki, ?WebCVS
- project reports
- Usage and development experiences, good and bad
- Free and Open Source Software Engineering is a special focus this year
- spec and design methods [eh, whats that? :-) ]
- testing
- reliability
- performance
- Team governance, admin, management [about PSAS, Twiki]
- Planning and forecasting [eh, whats that? :-) ]
- Measuring progress and assessing quality
- Embedded systems
Submission deadline is 16-Dec-2003
CanBusLinuxDriver work
Brainstorm. Feel free to comment and edit.
Motivation
- brief description of FC/LTC architecture
- brief description of CAN bus
- Goal: all aspects of rocket open source (CAN King -> CANtalope)
Starting point
- MOPS board, Intel chip, Debian Linux 2.4.18
- existing open source CAN driver
- mods required
- device model, capabilities and limitations
- no support for RTR messages
- use of
read
for RTR breaks normal driver model - 1 device per message object doesn't make full use of the chip
- no support for
select
- may not support telemetry bandwidth
Driver rewrite
- learning to write a driver
- existing CAN driver
- prior Linux experience
- pointers to material on web
- Intel chip docs vs. experimentation with
sucan
- interrupts
- methodology
- borrow from existing driver
- build environment, system
ssh/scp
to FC/LTC- serial port,
minicom
for driverkprintfs
- general vs. specific use and requirements
- simplify simplify simplify
Tools
- CAN King, motivation for CANtalope
sucan
interactive chip driver- thanks to
iopl
- thanks to
- RocketView, unit testers, LTC, FC
Philosopy/Conclusions
- specific mods still useful to open source community? should we care?
- credit to previous driver
- make best use of what you got; no compromises for generality
- small steps
- learning experience
-- IanOsgood - 03 Dec 2003
Attachments: