diff --git a/arch/arc/mm/dma.c b/arch/arc/mm/dma.c
index 29a46bb..01eaf88 100644
--- a/arch/arc/mm/dma.c
+++ b/arch/arc/mm/dma.c
@@ -17,18 +17,14 @@
  */
 
 #include <linux/dma-mapping.h>
-#include <linux/dma-debug.h>
-#include <linux/export.h>
 #include <asm/cache.h>
 #include <asm/cacheflush.h>
 
-/*
- * Helpers for Coherent DMA API.
- */
-void *dma_alloc_noncoherent(struct device *dev, size_t size,
-			    dma_addr_t *dma_handle, gfp_t gfp)
+
+static void *arc_dma_alloc(struct device *dev, size_t size,
+		dma_addr_t *dma_handle, gfp_t gfp, struct dma_attrs *attrs)
 {
-	void *paddr;
+	void *paddr, *kvaddr;
 
 	/* This is linear addr (0x8000_0000 based) */
 	paddr = alloc_pages_exact(size, gfp);
@@ -38,22 +34,6 @@
 	/* This is bus address, platform dependent */
 	*dma_handle = (dma_addr_t)paddr;
 
-	return paddr;
-}
-EXPORT_SYMBOL(dma_alloc_noncoherent);
-
-void dma_free_noncoherent(struct device *dev, size_t size, void *vaddr,
-			  dma_addr_t dma_handle)
-{
-	free_pages_exact((void *)dma_handle, size);
-}
-EXPORT_SYMBOL(dma_free_noncoherent);
-
-void *dma_alloc_coherent(struct device *dev, size_t size,
-			 dma_addr_t *dma_handle, gfp_t gfp)
-{
-	void *paddr, *kvaddr;
-
 	/*
 	 * IOC relies on all data (even coherent DMA data) being in cache
 	 * Thus allocate normal cached memory
@@ -65,22 +45,15 @@
 	 *   -For coherent data, Read/Write to buffers terminate early in cache
 	 *   (vs. always going to memory - thus are faster)
 	 */
-	if (is_isa_arcv2() && ioc_exists)
-		return dma_alloc_noncoherent(dev, size, dma_handle, gfp);
-
-	/* This is linear addr (0x8000_0000 based) */
-	paddr = alloc_pages_exact(size, gfp);
-	if (!paddr)
-		return NULL;
+	if ((is_isa_arcv2() && ioc_exists) ||
+	    dma_get_attr(DMA_ATTR_NON_CONSISTENT, attrs))
+		return paddr;
 
 	/* This is kernel Virtual address (0x7000_0000 based) */
 	kvaddr = ioremap_nocache((unsigned long)paddr, size);
 	if (kvaddr == NULL)
 		return NULL;
 
-	/* This is bus address, platform dependent */
-	*dma_handle = (dma_addr_t)paddr;
-
 	/*
 	 * Evict any existing L1 and/or L2 lines for the backing page
 	 * in case it was used earlier as a normal "cached" page.
@@ -95,26 +68,111 @@
 
 	return kvaddr;
 }
-EXPORT_SYMBOL(dma_alloc_coherent);
 
-void dma_free_coherent(struct device *dev, size_t size, void *kvaddr,
-		       dma_addr_t dma_handle)
+static void arc_dma_free(struct device *dev, size_t size, void *vaddr,
+		dma_addr_t dma_handle, struct dma_attrs *attrs)
 {
-	if (is_isa_arcv2() && ioc_exists)
-		return dma_free_noncoherent(dev, size, kvaddr, dma_handle);
-
-	iounmap((void __force __iomem *)kvaddr);
+	if (!dma_get_attr(DMA_ATTR_NON_CONSISTENT, attrs) &&
+	    !(is_isa_arcv2() && ioc_exists))
+		iounmap((void __force __iomem *)vaddr);
 
 	free_pages_exact((void *)dma_handle, size);
 }
-EXPORT_SYMBOL(dma_free_coherent);
 
 /*
- * Helper for streaming DMA...
+ * streaming DMA Mapping API...
+ * CPU accesses page via normal paddr, thus needs to explicitly made
+ * consistent before each use
  */
-void __arc_dma_cache_sync(unsigned long paddr, size_t size,
-			  enum dma_data_direction dir)
+static void _dma_cache_sync(unsigned long paddr, size_t size,
+		enum dma_data_direction dir)
 {
-	__inline_dma_cache_sync(paddr, size, dir);
+	switch (dir) {
+	case DMA_FROM_DEVICE:
+		dma_cache_inv(paddr, size);
+		break;
+	case DMA_TO_DEVICE:
+		dma_cache_wback(paddr, size);
+		break;
+	case DMA_BIDIRECTIONAL:
+		dma_cache_wback_inv(paddr, size);
+		break;
+	default:
+		pr_err("Invalid DMA dir [%d] for OP @ %lx\n", dir, paddr);
+	}
 }
-EXPORT_SYMBOL(__arc_dma_cache_sync);
+
+static dma_addr_t arc_dma_map_page(struct device *dev, struct page *page,
+		unsigned long offset, size_t size, enum dma_data_direction dir,
+		struct dma_attrs *attrs)
+{
+	unsigned long paddr = page_to_phys(page) + offset;
+	_dma_cache_sync(paddr, size, dir);
+	return (dma_addr_t)paddr;
+}
+
+static int arc_dma_map_sg(struct device *dev, struct scatterlist *sg,
+	   int nents, enum dma_data_direction dir, struct dma_attrs *attrs)
+{
+	struct scatterlist *s;
+	int i;
+
+	for_each_sg(sg, s, nents, i)
+		s->dma_address = dma_map_page(dev, sg_page(s), s->offset,
+					       s->length, dir);
+
+	return nents;
+}
+
+static void arc_dma_sync_single_for_cpu(struct device *dev,
+		dma_addr_t dma_handle, size_t size, enum dma_data_direction dir)
+{
+	_dma_cache_sync(dma_handle, size, DMA_FROM_DEVICE);
+}
+
+static void arc_dma_sync_single_for_device(struct device *dev,
+		dma_addr_t dma_handle, size_t size, enum dma_data_direction dir)
+{
+	_dma_cache_sync(dma_handle, size, DMA_TO_DEVICE);
+}
+
+static void arc_dma_sync_sg_for_cpu(struct device *dev,
+		struct scatterlist *sglist, int nelems,
+		enum dma_data_direction dir)
+{
+	int i;
+	struct scatterlist *sg;
+
+	for_each_sg(sglist, sg, nelems, i)
+		_dma_cache_sync((unsigned int)sg_virt(sg), sg->length, dir);
+}
+
+static void arc_dma_sync_sg_for_device(struct device *dev,
+		struct scatterlist *sglist, int nelems,
+		enum dma_data_direction dir)
+{
+	int i;
+	struct scatterlist *sg;
+
+	for_each_sg(sglist, sg, nelems, i)
+		_dma_cache_sync((unsigned int)sg_virt(sg), sg->length, dir);
+}
+
+static int arc_dma_supported(struct device *dev, u64 dma_mask)
+{
+	/* Support 32 bit DMA mask exclusively */
+	return dma_mask == DMA_BIT_MASK(32);
+}
+
+struct dma_map_ops arc_dma_ops = {
+	.alloc			= arc_dma_alloc,
+	.free			= arc_dma_free,
+	.map_page		= arc_dma_map_page,
+	.map_sg			= arc_dma_map_sg,
+	.sync_single_for_device	= arc_dma_sync_single_for_device,
+	.sync_single_for_cpu	= arc_dma_sync_single_for_cpu,
+	.sync_sg_for_cpu	= arc_dma_sync_sg_for_cpu,
+	.sync_sg_for_device	= arc_dma_sync_sg_for_device,
+	.dma_supported		= arc_dma_supported,
+};
+EXPORT_SYMBOL(arc_dma_ops);