[PATCHv6 3/5] common: dma-mapping: Introduce common remapping functions

[Catalin Marinas]

On Fri, Aug 08, 2014 at 11:45:56PM +0100, Andrew Morton wrote:
> On Fri,  8 Aug 2014 13:23:15 -0700 Laura Abbott <lauraa at codeaurora.org> wrote:
> > For architectures without coherent DMA, memory for DMA may
> > need to be remapped with coherent attributes. Factor out
> > the the remapping code from arm and put it in a
> > common location to reduce code duplication.
> > 
> > As part of this, the arm APIs are now migrated away from
> > ioremap_page_range to the common APIs which use map_vm_area for remapping.
> > This should be an equivalent change and using map_vm_area is more
> > correct as ioremap_page_range is intended to bring in io addresses
> > into the cpu space and not regular kernel managed memory.
> > 
> > ...
> >
> > @@ -267,3 +269,68 @@ int dma_common_mmap(struct device *dev, struct vm_area_struct *vma,
> >  	return ret;
> >  }
> >  EXPORT_SYMBOL(dma_common_mmap);
> > +
> > +/*
> > + * remaps an allocated contiguous region into another vm_area.
> > + * Cannot be used in non-sleeping contexts
> > + */
> > +
> > +void *dma_common_contiguous_remap(struct page *page, size_t size,
> > +			unsigned long vm_flags,
> > +			pgprot_t prot, const void *caller)
> > +{
> > +	int i;
> > +	struct page **pages;
> > +	void *ptr;
> > +
> > +	pages = kmalloc(sizeof(struct page *) << get_order(size), GFP_KERNEL);
> > +	if (!pages)
> > +		return NULL;
> > +
> > +	for (i = 0; i < (size >> PAGE_SHIFT); i++)
> > +		pages[i] = page + i;
> 
> Assumes a single mem_map[] array.  That's not the case for sparsemem
> (at least).

Good point. The "page" pointer (and memory range) passed to this
function has been allocated with alloc_pages(), so the range is
guaranteed to be physically contiguous but it does not imply a single
mem_map[] array. For arm64 with SPARSEMEM_VMEMMAP_ENABLE it's safe but
not all architectures use this (especially on 32-bit).

What about using pfn_to_page(pfn + i)?

Thanks.

-- 
Catalin