dma mapping fixes for 4.13-rc2:

- split the global dma coherent pool from the per-device pool.
    This fixes a regression in the earlier 4.13 pull requests where the
    global pool would override a per-device CMA pool. (Vladimir Murzin).
 -----BEGIN PGP SIGNATURE-----
 
 iQI/BAABCAApFiEEgdbnc3r/njty3Iq9D55TZVIEUYMFAll3l1sLHGhjaEBsc3Qu
 ZGUACgkQD55TZVIEUYN8BhAAqFxy2CrpEBk7gD2byOi9M4kTeXDYCESEoEAwuvTG
 Fesbw5zumliBR2cjt/qk/uIDZ93fP4BuHn89NtIfcGOD1LqYOyIPwUTpmb9AgicD
 y4eO1Gy/3DrG2haZcWYmDvq8yfSuR01H3ecY1KNsX1Y2kXxeBQfVKaUDR6fuix4+
 uCf98LzIWs3TYmj7h48LVB/oNnigvs0oljrB2dWrWVJHbgGYEpmdPjBEe6r95e5U
 5cHtPno5JA1lbBFt/nvsZl/NmzSd745SL3QwJsaVmSTf7oYnAuwyPI+5gqaoeQT6
 24947e8hJjuLhBpO7RiqnJY9QdPxT0XKclkCcjnRb5j3dB9KL09f9Dz60exyJzSe
 18V8+8+1m1BgvPsAOS/pLKYxKr9Kgzl9LFrFQaBkA5+7SPlywfV7HqaCkN/mKB4F
 XJoQyRDLlZiDStDKbrhGEAHG6oYaZXnkpQ5xDitSXcSkh9/2a/elsG3caUBRI5qP
 vKC0qvfBPjnHa/3lYNNoLgADB4tZCE3rRrVP6tqdHQbjuNUNK1wLNT7PiMfeoUVj
 Oqql4le0AKlsxO4vRjavOrtaW1bVT+eAYLEtdQfXWQDvhffriEW6r6I8PGqIOiCO
 OzxemCG2M6fcD9ho/VDpjo3Ei6tZylrxdTbrsm7ogQmo/U3ID9cfs452vIOYtCcB
 9so=
 =fJWP
 -----END PGP SIGNATURE-----

Merge tag 'dma-mapping-4.13-2' of git://git.infradead.org/users/hch/dma-mapping

Pull dma mapping fixes from Christoph Hellwig:
 "split the global dma coherent pool from the per-device pool.

  This fixes a regression in the earlier 4.13 pull requests where the
  global pool would override a per-device CMA pool (Vladimir Murzin)"

* tag 'dma-mapping-4.13-2' of git://git.infradead.org/users/hch/dma-mapping:
  ARM: NOMMU: Wire-up default DMA interface
  dma-coherent: introduce interface for default DMA pool

arch/arc/mm/dma.c

@@ -117,7 +117,7 @@ static int arc_dma_mmap(struct device *dev, struct vm_area_struct *vma,
 
 	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
 
-	if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret))
+	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
 		return ret;
 
 	if (off < count && user_count <= (count - off)) {

arch/arm/mm/dma-mapping-nommu.c

@@ -40,9 +40,21 @@ static void *arm_nommu_dma_alloc(struct device *dev, size_t size,
 
 {
 	const struct dma_map_ops *ops = &dma_noop_ops;
+	void *ret;
 
 	/*
-	 * We are here because:
+	 * Try generic allocator first if we are advertised that
+	 * consistency is not required.
+	 */
+
+	if (attrs & DMA_ATTR_NON_CONSISTENT)
+		return ops->alloc(dev, size, dma_handle, gfp, attrs);
+
+	ret = dma_alloc_from_global_coherent(size, dma_handle);
+
+	/*
+	 * dma_alloc_from_global_coherent() may fail because:
+	 *
 	 * - no consistent DMA region has been defined, so we can't
 	 *   continue.
 	 * - there is no space left in consistent DMA region, so we
@@ -50,11 +62,8 @@ static void *arm_nommu_dma_alloc(struct device *dev, size_t size,
 	 *   advertised that consistency is not required.
 	 */
 
-	if (attrs & DMA_ATTR_NON_CONSISTENT)
-		return ops->alloc(dev, size, dma_handle, gfp, attrs);
-
-	WARN_ON_ONCE(1);
-	return NULL;
+	WARN_ON_ONCE(ret == NULL);
+	return ret;
 }
 
 static void arm_nommu_dma_free(struct device *dev, size_t size,
@@ -63,14 +72,31 @@ static void arm_nommu_dma_free(struct device *dev, size_t size,
 {
 	const struct dma_map_ops *ops = &dma_noop_ops;
 
-	if (attrs & DMA_ATTR_NON_CONSISTENT)
+	if (attrs & DMA_ATTR_NON_CONSISTENT) {
 		ops->free(dev, size, cpu_addr, dma_addr, attrs);
-	else
-		WARN_ON_ONCE(1);
+	} else {
+		int ret = dma_release_from_global_coherent(get_order(size),
+							   cpu_addr);
+
+		WARN_ON_ONCE(ret == 0);
+	}
 
 	return;
 }
 
+static int arm_nommu_dma_mmap(struct device *dev, struct vm_area_struct *vma,
+			      void *cpu_addr, dma_addr_t dma_addr, size_t size,
+			      unsigned long attrs)
+{
+	int ret;
+
+	if (dma_mmap_from_global_coherent(vma, cpu_addr, size, &ret))
+		return ret;
+
+	return dma_common_mmap(dev, vma, cpu_addr, dma_addr, size);
+}
+
+
 static void __dma_page_cpu_to_dev(phys_addr_t paddr, size_t size,
 				  enum dma_data_direction dir)
 {
@@ -173,6 +199,7 @@ static void arm_nommu_dma_sync_sg_for_cpu(struct device *dev, struct scatterlist
 const struct dma_map_ops arm_nommu_dma_ops = {
 	.alloc			= arm_nommu_dma_alloc,
 	.free			= arm_nommu_dma_free,
+	.mmap			= arm_nommu_dma_mmap,
 	.map_page		= arm_nommu_dma_map_page,
 	.unmap_page		= arm_nommu_dma_unmap_page,
 	.map_sg			= arm_nommu_dma_map_sg,

arch/arm/mm/dma-mapping.c

@@ -851,7 +851,7 @@ static int __arm_dma_mmap(struct device *dev, struct vm_area_struct *vma,
 	unsigned long pfn = dma_to_pfn(dev, dma_addr);
 	unsigned long off = vma->vm_pgoff;
 
-	if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret))
+	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
 		return ret;
 
 	if (off < nr_pages && nr_vma_pages <= (nr_pages - off)) {

arch/arm64/mm/dma-mapping.c

@@ -329,7 +329,7 @@ static int __swiotlb_mmap(struct device *dev,
 	vma->vm_page_prot = __get_dma_pgprot(attrs, vma->vm_page_prot,
 					     is_device_dma_coherent(dev));
 
-	if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret))
+	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
 		return ret;
 
 	return __swiotlb_mmap_pfn(vma, pfn, size);
@@ -706,7 +706,7 @@ static int __iommu_mmap_attrs(struct device *dev, struct vm_area_struct *vma,
 	vma->vm_page_prot = __get_dma_pgprot(attrs, vma->vm_page_prot,
 					     is_device_dma_coherent(dev));
 
-	if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret))
+	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
 		return ret;
 
 	if (attrs & DMA_ATTR_FORCE_CONTIGUOUS) {

arch/mips/mm/dma-default.c

@@ -232,7 +232,7 @@ static int mips_dma_mmap(struct device *dev, struct vm_area_struct *vma,
 	else
 		vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
 
-	if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret))
+	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
 		return ret;
 
 	if (off < count && user_count <= (count - off)) {

drivers/base/dma-coherent.c

@@ -25,7 +25,7 @@ static inline struct dma_coherent_mem *dev_get_coherent_memory(struct device *de
 {
 	if (dev && dev->dma_mem)
 		return dev->dma_mem;
-	return dma_coherent_default_memory;
+	return NULL;
 }
 
 static inline dma_addr_t dma_get_device_base(struct device *dev,
@@ -165,9 +165,45 @@ void *dma_mark_declared_memory_occupied(struct device *dev,
 }
 EXPORT_SYMBOL(dma_mark_declared_memory_occupied);
 
+static void *__dma_alloc_from_coherent(struct dma_coherent_mem *mem,
+		ssize_t size, dma_addr_t *dma_handle)
+{
+	int order = get_order(size);
+	unsigned long flags;
+	int pageno;
+	int dma_memory_map;
+	void *ret;
+
+	spin_lock_irqsave(&mem->spinlock, flags);
+
+	if (unlikely(size > (mem->size << PAGE_SHIFT)))
+		goto err;
+
+	pageno = bitmap_find_free_region(mem->bitmap, mem->size, order);
+	if (unlikely(pageno < 0))
+		goto err;
+
+	/*
+	 * Memory was found in the coherent area.
+	 */
+	*dma_handle = mem->device_base + (pageno << PAGE_SHIFT);
+	ret = mem->virt_base + (pageno << PAGE_SHIFT);
+	dma_memory_map = (mem->flags & DMA_MEMORY_MAP);
+	spin_unlock_irqrestore(&mem->spinlock, flags);
+	if (dma_memory_map)
+		memset(ret, 0, size);
+	else
+		memset_io(ret, 0, size);
+
+	return ret;
+
+err:
+	spin_unlock_irqrestore(&mem->spinlock, flags);
+	return NULL;
+}
+
 /**
- * dma_alloc_from_coherent() - try to allocate memory from the per-device coherent area
- *
+ * dma_alloc_from_dev_coherent() - allocate memory from device coherent pool
  * @dev:	device from which we allocate memory
  * @size:	size of requested memory area
  * @dma_handle:	This will be filled with the correct dma handle
@@ -180,44 +216,18 @@ EXPORT_SYMBOL(dma_mark_declared_memory_occupied);
  * Returns 0 if dma_alloc_coherent should continue with allocating from
  * generic memory areas, or !0 if dma_alloc_coherent should return @ret.
  */
-int dma_alloc_from_coherent(struct device *dev, ssize_t size,
-				       dma_addr_t *dma_handle, void **ret)
+int dma_alloc_from_dev_coherent(struct device *dev, ssize_t size,
+		dma_addr_t *dma_handle, void **ret)
 {
 	struct dma_coherent_mem *mem = dev_get_coherent_memory(dev);
-	int order = get_order(size);
-	unsigned long flags;
-	int pageno;
-	int dma_memory_map;
 
 	if (!mem)
 		return 0;
 
-	*ret = NULL;
-	spin_lock_irqsave(&mem->spinlock, flags);
+	*ret = __dma_alloc_from_coherent(mem, size, dma_handle);
+	if (*ret)
+		return 1;
 
-	if (unlikely(size > (mem->size << PAGE_SHIFT)))
-		goto err;
-
-	pageno = bitmap_find_free_region(mem->bitmap, mem->size, order);
-	if (unlikely(pageno < 0))
-		goto err;
-
-	/*
-	 * Memory was found in the per-device area.
-	 */
-	*dma_handle = dma_get_device_base(dev, mem) + (pageno << PAGE_SHIFT);
-	*ret = mem->virt_base + (pageno << PAGE_SHIFT);
-	dma_memory_map = (mem->flags & DMA_MEMORY_MAP);
-	spin_unlock_irqrestore(&mem->spinlock, flags);
-	if (dma_memory_map)
-		memset(*ret, 0, size);
-	else
-		memset_io(*ret, 0, size);
-
-	return 1;
-
-err:
-	spin_unlock_irqrestore(&mem->spinlock, flags);
 	/*
 	 * In the case where the allocation can not be satisfied from the
 	 * per-device area, try to fall back to generic memory if the
@@ -225,25 +235,20 @@ err:
 	 */
 	return mem->flags & DMA_MEMORY_EXCLUSIVE;
 }
-EXPORT_SYMBOL(dma_alloc_from_coherent);
+EXPORT_SYMBOL(dma_alloc_from_dev_coherent);
 
-/**
- * dma_release_from_coherent() - try to free the memory allocated from per-device coherent memory pool
- * @dev:	device from which the memory was allocated
- * @order:	the order of pages allocated
- * @vaddr:	virtual address of allocated pages
- *
- * This checks whether the memory was allocated from the per-device
- * coherent memory pool and if so, releases that memory.
- *
- * Returns 1 if we correctly released the memory, or 0 if
- * dma_release_coherent() should proceed with releasing memory from
- * generic pools.
- */
-int dma_release_from_coherent(struct device *dev, int order, void *vaddr)
+void *dma_alloc_from_global_coherent(ssize_t size, dma_addr_t *dma_handle)
 {
-	struct dma_coherent_mem *mem = dev_get_coherent_memory(dev);
+	if (!dma_coherent_default_memory)
+		return NULL;
 
+	return __dma_alloc_from_coherent(dma_coherent_default_memory, size,
+			dma_handle);
+}
+
+static int __dma_release_from_coherent(struct dma_coherent_mem *mem,
+				       int order, void *vaddr)
+{
 	if (mem && vaddr >= mem->virt_base && vaddr <
 		   (mem->virt_base + (mem->size << PAGE_SHIFT))) {
 		int page = (vaddr - mem->virt_base) >> PAGE_SHIFT;
@@ -256,28 +261,39 @@ int dma_release_from_coherent(struct device *dev, int order, void *vaddr)
 	}
 	return 0;
 }
-EXPORT_SYMBOL(dma_release_from_coherent);
 
 /**
- * dma_mmap_from_coherent() - try to mmap the memory allocated from
- * per-device coherent memory pool to userspace
+ * dma_release_from_dev_coherent() - free memory to device coherent memory pool
  * @dev:	device from which the memory was allocated
- * @vma:	vm_area for the userspace memory
- * @vaddr:	cpu address returned by dma_alloc_from_coherent
- * @size:	size of the memory buffer allocated by dma_alloc_from_coherent
- * @ret:	result from remap_pfn_range()
+ * @order:	the order of pages allocated
+ * @vaddr:	virtual address of allocated pages
  *
  * This checks whether the memory was allocated from the per-device
- * coherent memory pool and if so, maps that memory to the provided vma.
+ * coherent memory pool and if so, releases that memory.
  *
- * Returns 1 if we correctly mapped the memory, or 0 if the caller should
- * proceed with mapping memory from generic pools.
+ * Returns 1 if we correctly released the memory, or 0 if the caller should
+ * proceed with releasing memory from generic pools.
  */
-int dma_mmap_from_coherent(struct device *dev, struct vm_area_struct *vma,
-			   void *vaddr, size_t size, int *ret)
+int dma_release_from_dev_coherent(struct device *dev, int order, void *vaddr)
 {
 	struct dma_coherent_mem *mem = dev_get_coherent_memory(dev);
 
+	return __dma_release_from_coherent(mem, order, vaddr);
+}
+EXPORT_SYMBOL(dma_release_from_dev_coherent);
+
+int dma_release_from_global_coherent(int order, void *vaddr)
+{
+	if (!dma_coherent_default_memory)
+		return 0;
+
+	return __dma_release_from_coherent(dma_coherent_default_memory, order,
+			vaddr);
+}
+
+static int __dma_mmap_from_coherent(struct dma_coherent_mem *mem,
+		struct vm_area_struct *vma, void *vaddr, size_t size, int *ret)
+{
 	if (mem && vaddr >= mem->virt_base && vaddr + size <=
 		   (mem->virt_base + (mem->size << PAGE_SHIFT))) {
 		unsigned long off = vma->vm_pgoff;
@@ -296,7 +312,39 @@ int dma_mmap_from_coherent(struct device *dev, struct vm_area_struct *vma,
 	}
 	return 0;
 }
-EXPORT_SYMBOL(dma_mmap_from_coherent);
+
+/**
+ * dma_mmap_from_dev_coherent() - mmap memory from the device coherent pool
+ * @dev:	device from which the memory was allocated
+ * @vma:	vm_area for the userspace memory
+ * @vaddr:	cpu address returned by dma_alloc_from_dev_coherent
+ * @size:	size of the memory buffer allocated
+ * @ret:	result from remap_pfn_range()
+ *
+ * This checks whether the memory was allocated from the per-device
+ * coherent memory pool and if so, maps that memory to the provided vma.
+ *
+ * Returns 1 if we correctly mapped the memory, or 0 if the caller should
+ * proceed with mapping memory from generic pools.
+ */
+int dma_mmap_from_dev_coherent(struct device *dev, struct vm_area_struct *vma,
+			   void *vaddr, size_t size, int *ret)
+{
+	struct dma_coherent_mem *mem = dev_get_coherent_memory(dev);
+
+	return __dma_mmap_from_coherent(mem, vma, vaddr, size, ret);
+}
+EXPORT_SYMBOL(dma_mmap_from_dev_coherent);
+
+int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *vaddr,
+				   size_t size, int *ret)
+{
+	if (!dma_coherent_default_memory)
+		return 0;
+
+	return __dma_mmap_from_coherent(dma_coherent_default_memory, vma,
+					vaddr, size, ret);
+}
 
 /*
  * Support for reserved memory regions defined in device tree

drivers/base/dma-mapping.c

@@ -235,7 +235,7 @@ int dma_common_mmap(struct device *dev, struct vm_area_struct *vma,
 
 	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
 
-	if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret))
+	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
 		return ret;
 
 	if (off < count && user_count <= (count - off)) {

include/linux/dma-mapping.h

@@ -157,16 +157,40 @@ static inline int is_device_dma_capable(struct device *dev)
  * These three functions are only for dma allocator.
  * Don't use them in device drivers.
  */
-int dma_alloc_from_coherent(struct device *dev, ssize_t size,
+int dma_alloc_from_dev_coherent(struct device *dev, ssize_t size,
 				       dma_addr_t *dma_handle, void **ret);
-int dma_release_from_coherent(struct device *dev, int order, void *vaddr);
+int dma_release_from_dev_coherent(struct device *dev, int order, void *vaddr);
 
-int dma_mmap_from_coherent(struct device *dev, struct vm_area_struct *vma,
+int dma_mmap_from_dev_coherent(struct device *dev, struct vm_area_struct *vma,
 			    void *cpu_addr, size_t size, int *ret);
+
+void *dma_alloc_from_global_coherent(ssize_t size, dma_addr_t *dma_handle);
+int dma_release_from_global_coherent(int order, void *vaddr);
+int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *cpu_addr,
+				  size_t size, int *ret);
+
 #else
-#define dma_alloc_from_coherent(dev, size, handle, ret) (0)
-#define dma_release_from_coherent(dev, order, vaddr) (0)
-#define dma_mmap_from_coherent(dev, vma, vaddr, order, ret) (0)
+#define dma_alloc_from_dev_coherent(dev, size, handle, ret) (0)
+#define dma_release_from_dev_coherent(dev, order, vaddr) (0)
+#define dma_mmap_from_dev_coherent(dev, vma, vaddr, order, ret) (0)
+
+static inline void *dma_alloc_from_global_coherent(ssize_t size,
+						   dma_addr_t *dma_handle)
+{
+	return NULL;
+}
+
+static inline int dma_release_from_global_coherent(int order, void *vaddr)
+{
+	return 0;
+}
+
+static inline int dma_mmap_from_global_coherent(struct vm_area_struct *vma,
+						void *cpu_addr, size_t size,
+						int *ret)
+{
+	return 0;
+}
 #endif /* CONFIG_HAVE_GENERIC_DMA_COHERENT */
 
 #ifdef CONFIG_HAS_DMA
@@ -481,7 +505,7 @@ static inline void *dma_alloc_attrs(struct device *dev, size_t size,
 
 	BUG_ON(!ops);
 
-	if (dma_alloc_from_coherent(dev, size, dma_handle, &cpu_addr))
+	if (dma_alloc_from_dev_coherent(dev, size, dma_handle, &cpu_addr))
 		return cpu_addr;
 
 	if (!arch_dma_alloc_attrs(&dev, &flag))
@@ -503,7 +527,7 @@ static inline void dma_free_attrs(struct device *dev, size_t size,
 	BUG_ON(!ops);
 	WARN_ON(irqs_disabled());
 
-	if (dma_release_from_coherent(dev, get_order(size), cpu_addr))
+	if (dma_release_from_dev_coherent(dev, get_order(size), cpu_addr))
 		return;
 
 	if (!ops->free || !cpu_addr)