Weekly Edition
Return to the Press page
| |||||||||||||
|
| |||||||||||||
16 papers on real-time and embedded Linux (Linux Devices)
Linux Devices is hosting 16
papers on real-time and embedded Linux. "LinuxDevices.com is
pleased to publish the proceedings from the Seventh Real-Time Linux
Workshop held in Lille, France, November 3-4, 2005, at the University for
Science and Technology of Lille (USTL). The papers span a broad range of
topics, ranging from fundamental real-time technologies to applications,
hardware, and tools. "
(Log in to post comments)
16 papers on real-time and embedded Linux (Linux Devices) Posted Nov 20, 2005 22:10 UTC (Sun) by mmarq (guest, #2332) [Link] I belive that the most important issue about RT and embedded is related with device drivers and their current "dissonance" as could be seen here: http://lwn.net/Articles/144269/
HALs and more HALs, Wrappers and more Wrappers other abstraction layers http://www.linuxdevices.com/files/rtlws-2005/JanKiszka.pdf (skins?)!!... isn't there some design alteration that could obviate all this mess ?
Concurs with the *mess* the fact that the design of some kind of "Dual-Kernel"(appears it to me), be it hypervisor or else, is almost inavoidable for a deterministic Real-Time approach!
In that perspective, as a laic, i've been pointing for a exokernel approach http://pdos.csail.mit.edu/exo/exo-slides/sld001.htm as a possible alternative, in consequence of other topics, and this one IMO offer nothing but a good reason to think about it again.
Extraordinary flexibity exibts those exokernel "things",as intented for OSS http://citeseer.ist.psu.edu/cache/papers/cs/4203/http:zSz...
And i, along with the majority in Linux community i belive, would be nothing but feeling safer and happier with a Linux structure covering all the "width" of a system hardware.
[obviating width: means nothing touches hardware without some Linux structure intervention, (there isn't a dual kernel)]
16 papers on real-time and embedded Linux (Linux Devices) Posted Nov 21, 2005 2:52 UTC (Mon) by mmarq (guest, #2332) [Link] opps... though related that was not the link i wanted to point
""Extraordinary flexibity exibts those exokernel "things",as intented for OSS http://citeseer.ist.psu.edu/cache/papers/cs/4203/http:zSz... ""
Adding to it, i think that a good defenition of this exo thing is a monolithic kernel completely skinned to the barebone, without compatibility problems due; http://cliki.tunes.org/Exokernel ...
And adding to the link that i intended to present above;
Dosent Xen effectivly operate in a similar way has described in the above paragraph, considering that it is a exokernel ?
So if "we" already gonna have a exo thing in the form a VMM, and probabily going to have another possibly different nanokernel for deterministic real time, why the waste ?
Aren't that going to be too many kernels wrapped up as a big one ? is it all linux or gonna be called Linux or what ?
|
|
||||||||||||