arm: dma: Allow CMA pages to not have a kernel mapping.

Currently, there are use cases where not having any kernel
mapping is required; if the CMA memory needs to be used as
a pool which can have both cached and uncached mappings we
need to remove the mapping to avoid the multiple mapping
problem. Extend the dma APIs to use the DMA_ATTR_NO_KERNEL_MAPPING
with CMA. This doesn't end up saving any virtual address space
but the mapping will still not be present.

Change-Id: I64d21250abbe615c43e2b5b1272ee2b6d106705a
Signed-off-by: Laura Abbott <lauraa@codeaurora.org>
Signed-off-by: Mitchel Humpherys <mitchelh@codeaurora.org>
diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c
index ab1bd68..afaa39d 100644
--- a/arch/arm/mm/dma-mapping.c
+++ b/arch/arm/mm/dma-mapping.c
@@ -314,7 +314,8 @@
 core_initcall(consistent_init);
 
 static void *__alloc_from_contiguous(struct device *dev, size_t size,
-				     pgprot_t prot, struct page **ret_page);
+				     pgprot_t prot, struct page **ret_page,
+				     bool no_kernel_mapping);
 
 static struct arm_vmregion_head coherent_head = {
 	.vm_lock	= __SPIN_LOCK_UNLOCKED(&coherent_head.vm_lock),
@@ -343,7 +344,7 @@
 	if (!IS_ENABLED(CONFIG_CMA))
 		return 0;
 
-	ptr = __alloc_from_contiguous(NULL, size, prot, &page);
+	ptr = __alloc_from_contiguous(NULL, size, prot, &page, false);
 	if (ptr) {
 		coherent_head.vm_start = (unsigned long) ptr;
 		coherent_head.vm_end = (unsigned long) ptr + size;
@@ -522,12 +523,27 @@
 	return 0;
 }
 
-static void __dma_remap(struct page *page, size_t size, pgprot_t prot)
+static int __dma_clear_pte(pte_t *pte, pgtable_t token, unsigned long addr,
+			    void *data)
+{
+	pte_clear(&init_mm, addr, pte);
+	return 0;
+}
+
+static void __dma_remap(struct page *page, size_t size, pgprot_t prot,
+			bool no_kernel_map)
 {
 	unsigned long start = (unsigned long) page_address(page);
 	unsigned end = start + size;
+	int (*func)(pte_t *pte, pgtable_t token, unsigned long addr,
+			    void *data);
 
-	apply_to_page_range(&init_mm, start, size, __dma_update_pte, &prot);
+	if (no_kernel_map)
+		func = __dma_clear_pte;
+	else
+		func = __dma_update_pte;
+
+	apply_to_page_range(&init_mm, start, size, func, &prot);
 	dsb();
 	flush_tlb_kernel_range(start, end);
 }
@@ -604,7 +620,8 @@
 }
 
 static void *__alloc_from_contiguous(struct device *dev, size_t size,
-				     pgprot_t prot, struct page **ret_page)
+				     pgprot_t prot, struct page **ret_page,
+				     bool no_kernel_mapping)
 {
 	unsigned long order = get_order(size);
 	size_t count = size >> PAGE_SHIFT;
@@ -615,7 +632,7 @@
 		return NULL;
 
 	__dma_clear_buffer(page, size);
-	__dma_remap(page, size, prot);
+	__dma_remap(page, size, prot, no_kernel_mapping);
 
 	*ret_page = page;
 	return page_address(page);
@@ -624,7 +641,7 @@
 static void __free_from_contiguous(struct device *dev, struct page *page,
 				   size_t size)
 {
-	__dma_remap(page, size, pgprot_kernel);
+	__dma_remap(page, size, pgprot_kernel, false);
 	dma_release_from_contiguous(dev, page, size >> PAGE_SHIFT);
 }
 
@@ -649,7 +666,7 @@
 
 #define __alloc_remap_buffer(dev, size, gfp, prot, ret, c)	NULL
 #define __alloc_from_pool(dev, size, ret_page, c)		NULL
-#define __alloc_from_contiguous(dev, size, prot, ret)		NULL
+#define __alloc_from_contiguous(dev, size, prot, ret, w)	NULL
 #define __free_from_pool(cpu_addr, size)			0
 #define __free_from_contiguous(dev, page, size)			do { } while (0)
 #define __dma_free_remap(cpu_addr, size)			do { } while (0)
@@ -672,7 +689,8 @@
 
 
 static void *__dma_alloc(struct device *dev, size_t size, dma_addr_t *handle,
-			 gfp_t gfp, pgprot_t prot, const void *caller)
+			 gfp_t gfp, pgprot_t prot, const void *caller,
+			 bool no_kernel_mapping)
 {
 	u64 mask = get_coherent_dma_mask(dev);
 	struct page *page;
@@ -712,7 +730,8 @@
 	else if (gfp & GFP_ATOMIC)
 		addr = __alloc_from_pool(dev, size, &page, caller);
 	else
-		addr = __alloc_from_contiguous(dev, size, prot, &page);
+		addr = __alloc_from_contiguous(dev, size, prot, &page,
+						no_kernel_mapping);
 
 	if (addr)
 		*handle = pfn_to_dma(dev, page_to_pfn(page));
@@ -729,12 +748,14 @@
 {
 	pgprot_t prot = __get_dma_pgprot(attrs, pgprot_kernel);
 	void *memory;
+	bool no_kernel_mapping = dma_get_attr(DMA_ATTR_NO_KERNEL_MAPPING,
+					attrs);
 
 	if (dma_alloc_from_coherent(dev, size, handle, &memory))
 		return memory;
 
 	return __dma_alloc(dev, size, handle, gfp, prot,
-			   __builtin_return_address(0));
+			   __builtin_return_address(0), no_kernel_mapping);
 }
 
 /*