lkml.org 
[lkml]   [1998]   [Jan]   [29]   [last100]   RSS Feed
Views: [more markup]  [less markup]  [headers]  [forward] 
 
Messages in this thread
/
DateThu, 29 Jan 1998 13:55:47 +0100 (MET)
FromGabriel Paubert <>
SubjectRe: Suitable chipset for a pocket linux box

On Thu, 29 Jan 1998, Matti Aarnio wrote:

> > The worst thing is they removed the MMU :-(
> 
> 	No, the basic 68000 DOES NOT have MMU, thus they removed nothing.
> 	The first CPU in 68K series with internal MMU is 68030.
> 

True, but the 68020 (and 68010 AFAIR) is also designed to work with an
external MMU. Although the MMU chip for the 68010 was strange. 

> 	Now I wonder why you would want to have a MMU in an embedded
> 	application system..  I do think the MMU is something for
> 	multi-user time-shared systems, and not space-constrained,
> 	and possibly power-constrained gadgets.
> 	(Sure it would be "cool" to have Linux running at my wrist-
> 	 watch and show the time thru X, but why I should burn all
> 	 those wats of power ?  Just to keep my wrist warm ?)

Look at the docs of recent embedded RISC chips. Some MMU can
operate in two modes of operations:

- one is to define which areas of memory are
cacheable/write-through/write-back attributes. In the era of on chip
caches even for embedded processors, it is necessary.

- the other mode is a normal MMU mode (paging) for PDA...

And do not forget that keeping instructions and data in the cache
actually saves power, since in these chips very often the I/O buffers
dissipate a significant proportion of the power when they are enabled. 

  Gabriel

P.S: I now think we should stop this thread or make it private, it is not
about kernel development. 



\
 
 \ /
  Last update: 2005-03-22 12:41    [from the cache]
©2003-2008