Aha! thanks for the clarification.

How much penalty is it to use write-through instead of write-back? It 
feels like it could hurt performace pretty much?
- Karl

Steve Woodford wrote:
> On Thursday 31 March 2005 11:53, Karl Janmar wrote:
> 
> 
>>I have heard that the PXA255 had some problems with it's cache... has
>>somebody more info on this?
> 
> 
> Early steppings of the processor had problems when write-back caching was 
> enabled. NetBSD detects those steppings and switches to write-through 
> automatically.
> 
> Cheers, Steve