Re: Orocos and RTAI/LXRT
Peter Soetens <[email protected]>
| Newsgroups | gmane.science.robotics.orocos.user |
|---|---|
| Organization | KU Leuven |
| Message-ID | <[email protected]> |
just some comments to clarify.... On Monday 23 June 2003 21:18, Herman Bruyninckx wrote: > On Mon, 23 Jun 2003, Peter Soetens wrote: > > [... RTAI/LXRT...] > > > Does this mean we drop portability ? Certainly not ! Orocos still has an > > os-abstraction layer, and kernel-RTAI and kernel-RTLinux just fell out > > because of the 'incompleteness' they offer in their API. > > Some current problems: > 1. the RTLinux support for user space hard real-time is poor, in the > GPL version at least. So, no portability there. > 2. RTAI/LXRT currently only runs on ia32 systems (Intel, AMD, ...) > So, not PowerPC or ARM or Xscale, or... But since this is a demand driven 'market', ports to these platforms will become available when they're at least needed by someone who finds it worth investing time in it (and contributing it back) > 3. the LXRT calls are fully RTAI-only. This could be tackled by > neutral wrappers. (As is already done in the existing > OS-abstraction for all "normal" OS interactions.) One can not see in Orocos that one is using LXRT or not. OS primitives are hidden behind the OS abstraction layer, device drivers behind the FDI (Framework Device Interfaces). I know we got that one right. Peter -- ------------------------------------------------------------------------ Peter Soetens http://www.orocos.org Katholieke Universiteit Leuven Division Production Engineering, tel. +32 16 322773 Machine Design and Automation fax. +32 16 322987 Celestijnenlaan 300B [email protected] B-3001 Leuven Belgium http://www.mech.kuleuven.ac.be/pma ------------------------------------------------------------------------