LinuxSA Mailing list archives

  From: Alan Kennington <akenning@dog.topology.org>
  To  : drewp@bigpond.com
  Date: Wed, 31 Mar 1999 14:14:55 +0930

Re: Linux in real time (was More Meeting Information)

Andrew,

I probably know much less about real-time OSs than most people, but...

When I talked about real-time software, I meant in the sense of
kernel modules being able to completely take control of the CPU.

I.e. linux does not provide real-time functionality at the process
level, but like most OSs, real-time can be achieved in device drivers
(i.e. kernel modules).
Of course, if you want _two_ kernel modules to achieve real-time 
functionality, then one is in trouble.

Therefore I was only referring to the fact that linux is ideal for
kernel module writing, and that the kernel can do what it likes
with the CPU. This is a much more limited meaning of "real-time"
than in the context of real-time OSs.

Also I should mention that I was not talking about microsecond-level
real-time. I only meant down to a few milliseconds.
I assume that the _real_ real-time will be done on
dedicated microcontrollers which are connected by various buses to
the main CPU, being a sort of x86 or something.

Therefore in the sorts of projects that I was talking about,
I envisage there being a box containing a board or two, which have
a CPU + RAM + 1 or more microcontrollers.
Then I was asserting that linux would be good for the CPU -- but
probably not for the _real_ real-time processors, which are the
uC chips (e.g. microchip PIC, philips XA, coldfire etc. etc.).

Cheers,
Alan Kennington.

-- 
Check out the LinuxSA web pages at http://www.linuxsa.org.au/
To unsubscribe from the LinuxSA list:
  mail linuxsa-request@linuxsa.org.au with "unsubscribe" as the subject