Phil Wilshire, of www.RealTimeLinux.org reports . . .

This was probably the most productive day of the whole conference. Significant proposals were discussed and the path forward planned for the next year

The day started with Stuart Hughes of Zentropix, who used some example code to explain the differences between the RTAI and RTL API's. He also explained the need to have a strategy to support the original simple API found in RTL V1 and RTAI as well as supporting the EL/IX initiative by Cygnus for those needing a portable POSIX approach.

Manfred Hollstein, of Cygnus, then covered the role and objectives of EL/IX. This effort seeks to define a POSIX/ISO API with different levels of compliance, ranging from deeply embedded appliances (running eCos), to full scale Linux desktop machines. This should aid portability, and make development and deployment simpler.

Phil Wilshire, a prominent RealTime Linux advocate, then discussed the need for starting the port activity and defining a common API for user applications. Two API formats were proposed: one simple, to ease the introduction of RT Linux to new users; the other conforming to the EL/IX proposal, for Industrial use.

Nicholas McGuire, one of the organizers of the workshop, added a call for a resurrection of the documentation project. A group was set up to lobby for the introduction, at some stage, of a minimum set of kernel hooks for Real Time systems into the current and future Linux kernel structures.

Mailing lists with nominated leaders were to be set up to cover topics:

• Documentation project — leader: Nicholas McGuire — task: revive and upgrade the documentation.
• Kernel Interface — leader: Tomaz Motylewski — task: produce a “Linus” acceptable patch.
• Testing and Validation — leader: John Storrs — task: define Software and Hardware techniques for performance validation.
• API definition — leader: Arnold Niessen — task: define “Simple” and “Industry” API'S and get involved with the EL/IX project.
• RT Networking — leader: Douglas Niehaus — task: Discuss RT Ethernet and other networking requirements and options.
• Porting — leader: Bernhard Kuhn — task: Initiate and support porting projects.
• PPC M68K Arm7 StrongArm Alpha Drivers — leader: David Schleef — task: Collect and produce device drivers.

The afternoon continued with constructive discussions on the various mailing list topics.

The need to adopt low latency and soft real time patches into the kernel was discussed. If enough poeple make enough noise, we should have all the hooks we need in the kernel.

David Schleef remembered that he had looked at a combination of RTAI and RTL HAL layers and had, in fact, produced a “sort of” patch. This was to be dug out and, after review, offered to Linus as a possible kernel patch with the hope of eventual inclusion into the Kernel. This work would, ideally, be done with full cooperation of Victor and Paolo.

The mood at the end of this conference was very positive. Convincing arguments were presented showing how Linux can present considerable advantages to companies producing products in the the Embedded / Real Time markets. The adoption and input into the EL/IX initiative offers some interesting possibilities. Using this approach projects can be rapidly prototyped on well known desktop environments and then deployed on a wide variety of target systems. Reduced time-to-market and low feature overhead is combined with low cost target system options.

Having consumed 120 litres of coffee over the three days, the Real Time Linux conference ended later in the evening after some Real Beer.

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.