Memory Protection (Attributes)

Chris Johns chrisj at rtems.org
Tue Dec 6 01:30:11 UTC 2011


On 6/12/11 11:58 AM, Peter Dufault wrote:
>
> On Dec 5, 2011, at 7:28 , Chris Johns wrote:
>
>> One thing Gedare and I have discussed when we meet earlier this year is
>> a process type solution is not what we see as a suitable model. We see
>> issues in offering that type of solution.
>
> What do you mean by "a process type solution"?
>
> If you mean a thread/task corresponds to a unix flavor process then I agree.
>

Yeap that is what I mean. Attempting to fit a process type environment 
into an embedded realtime box.

> I don't see the "use case" (boy, I personally hate that term)

Hehe. It is a simple way with 2 words to state what it is.

> of "memory protection available, memory protection completely irrelevant" as a valid "use case" under RTEMS.

I took "memory protection completely irrelevant" to mean not part of the 
solution I want in my system. If you do not want it there should be a 
zero cost involved in removing it.

> (I'd say programming model or something and not use case).

Some applications do not want or need the overhead of setting up and 
managing the memory. There is also an overhead of managing the different 
contexts or arenas. Other want it or are required to have it.

Chris



More information about the devel mailing list