[patch 1/5] Staging: VME Framework for the Linux Kernel

Martyn Welch martyn.welch at gefanuc.com
Tue Aug 11 16:38:24 UTC 2009


Emilio G. Cota wrote:
> Martyn Welch wrote:
>   
>>> It wouldn't break it, the model simply couldn't give you more
>>> than 8 windows-->8 devices.
>>>   
>>>       
>> Unless the devices we the same and the driver reused one window.
>>     
>
> No, we agreed on that devices should know *nothing* about windows.
>
>   
Don't think I did. :-)

I agreed that it would be impossible to assign sole use of 8 resources 
to 17 different drivers. I feel that there could be room and indeed a 
need for both approaches. It may be more efficient for a driver to use 
and move 1 window to position it's self over the registers of each of a 
number of compatible devices as required. Such a driver could exist 
side-by-side with drivers using your approach, that utilize a layer 
which removes such control. Whilst I understand you can assign a fairly 
sizable chunk of memory mapped IO for VME, this is not always the case. 
In situations where the available memory mapped IO is limited, placing a 
large window to cover 2 smaller windows with dead space in between them 
will not be an adequate solution either.

Martyn
> E.
>   


-- 
Martyn Welch MEng MPhil MIET (Principal Software Engineer)   T:+44(0)1327322748
GE Fanuc Intelligent Platforms Ltd,        |Registered in England and Wales
Tove Valley Business Park, Towcester,      |(3828642) at 100 Barbirolli Square,
Northants, NN12 6PF, UK T:+44(0)1327359444 |Manchester,M2 3AB  VAT:GB 927559189



More information about the devel mailing list