dma_cache_maint_contiguous should be patched as dma_cache_maint do

[Russell King - ARM Linux]

On Thu, Dec 10, 2009 at 04:21:29PM -0500, Nicolas Pitre wrote:
> On Thu, 10 Dec 2009, Russell King - ARM Linux wrote:
> 
> > If it is just ARMv6 which is affected by this (note that we can not rely
> > upon ID_MMFR3 telling us since this register has a different purpose
> > there) then we can solve this by just modifying the ARMv6 DMA cache
> > methods - which becomes easier if I get the rearrangement of the DMA
> > stuff finally sorted.  The *only* thing that remains in the way of that
> > is the stupid flush_ioremap_region() crap for just one MTD driver.
> 
> My suggestion: kill flush_ioremap_region() and substitute it by an ARMv5 
> cache range invalidate loop implementation local to that driver.
> 
> While at it, here's the patch:

Great, can it end up in the patch system?