[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [cdt-dev] IMemory and memory spaces
|
At 05:47 PM 5/10/2010, Pawel Piech wrote:
On 05/10/2010 03:42 PM, John Cortell wrote:
At 05:35 PM 5/10/2010, Pawel Piech wrote:
I've been prototyping the idea and it seems doable. Basically, I
have an IMemorySpaceDMContext that has as a parent an
IMemoryDMContext. This new interface simply has a method that
returns the memory space identifier. Down at the MI layer, the
command can use the new context interface to get the memory space
to use in the MI command.
Couldn't you have IMemorySpaceDMContext extend
IMemoryDMContext? I'm not sure whether parent-child relationship
is needed between memory contexts.
IMemorySpaceDMContext does extend IMemoryDMContext. But, in
practice, IMemoryDMContext is really just an MI control context
under the covers. By using a runtime parent-child relationship, I
can simply add the memory space characteristic to the context
object without touching the control context implementation. It may
make more sense when you see the patch.
It does make sense. What I don't know is if you have memory spaces,
do you still need to have MI control context to implement IMemoryDMContext.
Yep. The memory space stuff is just some additional qualification in
the underlying mi command. The rest of the memory commands have the
same requirements (need the same context).
John
