On 11/01/17 10:11, Geert Uytterhoeven wrote: > From: Takeshi Kihara <takeshi.kihara.df@xxxxxxxxxxx> > > This patch adds support for DMA_ATTR_SKIP_CPU_SYNC attribute for > dma_{un}map_{page,sg} functions family to swiotlb. > > DMA_ATTR_SKIP_CPU_SYNC allows platform code to skip synchronization of > the CPU cache for the given buffer assuming that it has been already > transferred to 'device' domain. > > Ported from IOMMU .{un}map_{sg,page} ops. > > Signed-off-by: Takeshi Kihara <takeshi.kihara.df@xxxxxxxxxxx> > Acked-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx> > Signed-off-by: Geert Uytterhoeven <geert+renesas@xxxxxxxxx> > --- > v2: > - Add Acked-by. > > Support for DMA_ATTR_SKIP_CPU_SYNC was included when porting the IOMMU > ops from arm to arm64 in commit 13b8629f651164d7 ("arm64: Add IOMMU > dma_ops"). > > Presumably it was an oversight that the existing swiotlb based > implementation didn't have support for DMA_ATTR_SKIP_CPU_SYNC yet? Less an oversight, more that nobody's wanted to use it until now ;) Personally I'd prefer flag tests to be "!(x)" rather than "(x) == 0", but the latter is already in place, so I'll leave the final word on style/consistency nitpicks to Catalin and Will. Reviewed-by: Robin Murphy <robin.murphy@xxxxxxx> > --- > arch/arm64/mm/dma-mapping.c | 12 ++++++++---- > 1 file changed, 8 insertions(+), 4 deletions(-) > > diff --git a/arch/arm64/mm/dma-mapping.c b/arch/arm64/mm/dma-mapping.c > index e04082700bb16c35..1d7d5d2881db7c19 100644 > --- a/arch/arm64/mm/dma-mapping.c > +++ b/arch/arm64/mm/dma-mapping.c > @@ -211,7 +211,8 @@ static dma_addr_t __swiotlb_map_page(struct device *dev, struct page *page, > dma_addr_t dev_addr; > > dev_addr = swiotlb_map_page(dev, page, offset, size, dir, attrs); > - if (!is_device_dma_coherent(dev)) > + if (!is_device_dma_coherent(dev) && > + (attrs & DMA_ATTR_SKIP_CPU_SYNC) == 0) > __dma_map_area(phys_to_virt(dma_to_phys(dev, dev_addr)), size, dir); > > return dev_addr; > @@ -222,7 +223,8 @@ static void __swiotlb_unmap_page(struct device *dev, dma_addr_t dev_addr, > size_t size, enum dma_data_direction dir, > unsigned long attrs) > { > - if (!is_device_dma_coherent(dev)) > + if (!is_device_dma_coherent(dev) && > + (attrs & DMA_ATTR_SKIP_CPU_SYNC) == 0) > __dma_unmap_area(phys_to_virt(dma_to_phys(dev, dev_addr)), size, dir); > swiotlb_unmap_page(dev, dev_addr, size, dir, attrs); > } > @@ -235,7 +237,8 @@ static int __swiotlb_map_sg_attrs(struct device *dev, struct scatterlist *sgl, > int i, ret; > > ret = swiotlb_map_sg_attrs(dev, sgl, nelems, dir, attrs); > - if (!is_device_dma_coherent(dev)) > + if (!is_device_dma_coherent(dev) && > + (attrs & DMA_ATTR_SKIP_CPU_SYNC) == 0) > for_each_sg(sgl, sg, ret, i) > __dma_map_area(phys_to_virt(dma_to_phys(dev, sg->dma_address)), > sg->length, dir); > @@ -251,7 +254,8 @@ static void __swiotlb_unmap_sg_attrs(struct device *dev, > struct scatterlist *sg; > int i; > > - if (!is_device_dma_coherent(dev)) > + if (!is_device_dma_coherent(dev) && > + (attrs & DMA_ATTR_SKIP_CPU_SYNC) == 0) > for_each_sg(sgl, sg, nelems, i) > __dma_unmap_area(phys_to_virt(dma_to_phys(dev, sg->dma_address)), > sg->length, dir); >