Merge tag 'dma-mapping-5.15' of git://git.infradead.org/users/hch/dma-mapping

Pull dma-mapping updates from Christoph Hellwig:

 - fix debugfs initialization order (Anthony Iliopoulos)

 - use memory_intersects() directly (Kefeng Wang)

 - allow to return specific errors from ->map_sg (Logan Gunthorpe,
   Martin Oliveira)

 - turn the dma_map_sg return value into an unsigned int (me)

 - provide a common global coherent pool іmplementation (me)

* tag 'dma-mapping-5.15' of git://git.infradead.org/users/hch/dma-mapping: (31 commits)
  hexagon: use the generic global coherent pool
  dma-mapping: make the global coherent pool conditional
  dma-mapping: add a dma_init_global_coherent helper
  dma-mapping: simplify dma_init_coherent_memory
  dma-mapping: allow using the global coherent pool for !ARM
  ARM/nommu: use the generic dma-direct code for non-coherent devices
  dma-direct: add support for dma_coherent_default_memory
  dma-mapping: return an unsigned int from dma_map_sg{,_attrs}
  dma-mapping: disallow .map_sg operations from returning zero on error
  dma-mapping: return error code from dma_dummy_map_sg()
  x86/amd_gart: don't set failed sg dma_address to DMA_MAPPING_ERROR
  x86/amd_gart: return error code from gart_map_sg()
  xen: swiotlb: return error code from xen_swiotlb_map_sg()
  parisc: return error code from .map_sg() ops
  sparc/iommu: don't set failed sg dma_address to DMA_MAPPING_ERROR
  sparc/iommu: return error codes from .map_sg() ops
  s390/pci: don't set failed sg dma_address to DMA_MAPPING_ERROR
  s390/pci: return error code from s390_dma_map_sg()
  powerpc/iommu: don't set failed sg dma_address to DMA_MAPPING_ERROR
  powerpc/iommu: return error code from .map_sg() ops
  ...

6 files changed, 175 insertions, 110 deletions

diff --git a/kernel/dma/Kconfig b/kernel/dma/Kconfig
index 77b405508743..725cfd51762b 100644
--- a/kernel/dma/Kconfig
+++ b/kernel/dma/Kconfig
@@ -93,6 +93,10 @@ config DMA_COHERENT_POOL
 	select GENERIC_ALLOCATOR
 	bool
 
+config DMA_GLOBAL_POOL
+	select DMA_DECLARE_COHERENT
+	bool
+
 config DMA_REMAP
 	bool
 	depends on MMU
diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
index 794e76b03b34..25fc85a7aebe 100644
--- a/kernel/dma/coherent.c
+++ b/kernel/dma/coherent.c
@@ -20,8 +20,6 @@ struct dma_coherent_mem {
 	bool		use_dev_dma_pfn_offset;
 };
 
-static struct dma_coherent_mem *dma_coherent_default_memory __ro_after_init;
-
 static inline struct dma_coherent_mem *dev_get_coherent_memory(struct device *dev)
 {
 	if (dev && dev->dma_mem)
@@ -37,51 +35,44 @@ static inline dma_addr_t dma_get_device_base(struct device *dev,
 	return mem->device_base;
 }
 
-static int dma_init_coherent_memory(phys_addr_t phys_addr,
-		dma_addr_t device_addr, size_t size,
-		struct dma_coherent_mem **mem)
+static struct dma_coherent_mem *dma_init_coherent_memory(phys_addr_t phys_addr,
+		dma_addr_t device_addr, size_t size, bool use_dma_pfn_offset)
 {
-	struct dma_coherent_mem *dma_mem = NULL;
-	void *mem_base = NULL;
+	struct dma_coherent_mem *dma_mem;
 	int pages = size >> PAGE_SHIFT;
 	int bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
-	int ret;
+	void *mem_base;
 
-	if (!size) {
-		ret = -EINVAL;
-		goto out;
-	}
+	if (!size)
+		return ERR_PTR(-EINVAL);
 
 	mem_base = memremap(phys_addr, size, MEMREMAP_WC);
-	if (!mem_base) {
-		ret = -EINVAL;
-		goto out;
-	}
+	if (!mem_base)
+		return ERR_PTR(-EINVAL);
+
 	dma_mem = kzalloc(sizeof(struct dma_coherent_mem), GFP_KERNEL);
-	if (!dma_mem) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!dma_mem)
+		goto out_unmap_membase;
 	dma_mem->bitmap = kzalloc(bitmap_size, GFP_KERNEL);
-	if (!dma_mem->bitmap) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!dma_mem->bitmap)
+		goto out_free_dma_mem;
 
 	dma_mem->virt_base = mem_base;
 	dma_mem->device_base = device_addr;
 	dma_mem->pfn_base = PFN_DOWN(phys_addr);
 	dma_mem->size = pages;
+	dma_mem->use_dev_dma_pfn_offset = use_dma_pfn_offset;
 	spin_lock_init(&dma_mem->spinlock);
 
-	*mem = dma_mem;
-	return 0;
+	return dma_mem;
 
-out:
+out_free_dma_mem:
 	kfree(dma_mem);
-	if (mem_base)
-		memunmap(mem_base);
-	return ret;
+out_unmap_membase:
+	memunmap(mem_base);
+	pr_err("Reserved memory: failed to init DMA memory pool at %pa, size %zd MiB\n",
+		&phys_addr, size / SZ_1M);
+	return ERR_PTR(-ENOMEM);
 }
 
 static void dma_release_coherent_memory(struct dma_coherent_mem *mem)
@@ -130,9 +121,9 @@ int dma_declare_coherent_memory(struct device *dev, phys_addr_t phys_addr,
 	struct dma_coherent_mem *mem;
 	int ret;
 
-	ret = dma_init_coherent_memory(phys_addr, device_addr, size, &mem);
-	if (ret)
-		return ret;
+	mem = dma_init_coherent_memory(phys_addr, device_addr, size, false);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
 
 	ret = dma_assign_coherent_memory(dev, mem);
 	if (ret)
@@ -198,16 +189,6 @@ int dma_alloc_from_dev_coherent(struct device *dev, ssize_t size,
 	return 1;
 }
 
-void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
-				     dma_addr_t *dma_handle)
-{
-	if (!dma_coherent_default_memory)
-		return NULL;
-
-	return __dma_alloc_from_coherent(dev, dma_coherent_default_memory, size,
-					 dma_handle);
-}
-
 static int __dma_release_from_coherent(struct dma_coherent_mem *mem,
 				       int order, void *vaddr)
 {
@@ -243,15 +224,6 @@ int dma_release_from_dev_coherent(struct device *dev, int order, void *vaddr)
 	return __dma_release_from_coherent(mem, order, vaddr);
 }
 
-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)
 {
@@ -297,6 +269,28 @@ int dma_mmap_from_dev_coherent(struct device *dev, struct vm_area_struct *vma,
 	return __dma_mmap_from_coherent(mem, vma, vaddr, size, ret);
 }
 
+#ifdef CONFIG_DMA_GLOBAL_POOL
+static struct dma_coherent_mem *dma_coherent_default_memory __ro_after_init;
+
+void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
+				     dma_addr_t *dma_handle)
+{
+	if (!dma_coherent_default_memory)
+		return NULL;
+
+	return __dma_alloc_from_coherent(dev, dma_coherent_default_memory, size,
+					 dma_handle);
+}
+
+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);
+}
+
 int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *vaddr,
 				   size_t size, int *ret)
 {
@@ -307,6 +301,19 @@ int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *vaddr,
 					vaddr, size, ret);
 }
 
+int dma_init_global_coherent(phys_addr_t phys_addr, size_t size)
+{
+	struct dma_coherent_mem *mem;
+
+	mem = dma_init_coherent_memory(phys_addr, phys_addr, size, true);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
+	dma_coherent_default_memory = mem;
+	pr_info("DMA: default coherent area is set\n");
+	return 0;
+}
+#endif /* CONFIG_DMA_GLOBAL_POOL */
+
 /*
  * Support for reserved memory regions defined in device tree
  */
@@ -315,25 +322,22 @@ int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *vaddr,
 #include <linux/of_fdt.h>
 #include <linux/of_reserved_mem.h>
 
+#ifdef CONFIG_DMA_GLOBAL_POOL
 static struct reserved_mem *dma_reserved_default_memory __initdata;
+#endif
 
 static int rmem_dma_device_init(struct reserved_mem *rmem, struct device *dev)
 {
-	struct dma_coherent_mem *mem = rmem->priv;
-	int ret;
-
-	if (!mem) {
-		ret = dma_init_coherent_memory(rmem->base, rmem->base,
-					       rmem->size, &mem);
-		if (ret) {
-			pr_err("Reserved memory: failed to init DMA memory pool at %pa, size %ld MiB\n",
-				&rmem->base, (unsigned long)rmem->size / SZ_1M);
-			return ret;
-		}
+	if (!rmem->priv) {
+		struct dma_coherent_mem *mem;
+
+		mem = dma_init_coherent_memory(rmem->base, rmem->base,
+					       rmem->size, true);
+		if (IS_ERR(mem))
+			return PTR_ERR(mem);
+		rmem->priv = mem;
 	}
-	mem->use_dev_dma_pfn_offset = true;
-	rmem->priv = mem;
-	dma_assign_coherent_memory(dev, mem);
+	dma_assign_coherent_memory(dev, rmem->priv);
 	return 0;
 }
 
@@ -361,7 +365,9 @@ static int __init rmem_dma_setup(struct reserved_mem *rmem)
 		pr_err("Reserved memory: regions without no-map are not yet supported\n");
 		return -EINVAL;
 	}
+#endif
 
+#ifdef CONFIG_DMA_GLOBAL_POOL
 	if (of_get_flat_dt_prop(node, "linux,dma-default", NULL)) {
 		WARN(dma_reserved_default_memory,
 		     "Reserved memory: region for default DMA coherent area is redefined\n");
@@ -375,31 +381,16 @@ static int __init rmem_dma_setup(struct reserved_mem *rmem)
 	return 0;
 }
 
+#ifdef CONFIG_DMA_GLOBAL_POOL
 static int __init dma_init_reserved_memory(void)
 {
-	const struct reserved_mem_ops *ops;
-	int ret;
-
 	if (!dma_reserved_default_memory)
 		return -ENOMEM;
-
-	ops = dma_reserved_default_memory->ops;
-
-	/*
-	 * We rely on rmem_dma_device_init() does not propagate error of
-	 * dma_assign_coherent_memory() for "NULL" device.
-	 */
-	ret = ops->device_init(dma_reserved_default_memory, NULL);
-
-	if (!ret) {
-		dma_coherent_default_memory = dma_reserved_default_memory->priv;
-		pr_info("DMA: default coherent area is set\n");
-	}
-
-	return ret;
+	return dma_init_global_coherent(dma_reserved_default_memory->base,
+					dma_reserved_default_memory->size);
 }
-
 core_initcall(dma_init_reserved_memory);
+#endif /* CONFIG_DMA_GLOBAL_POOL */
 
 RESERVEDMEM_OF_DECLARE(dma, "shared-dma-pool", rmem_dma_setup);
 #endif
diff --git a/kernel/dma/debug.c b/kernel/dma/debug.c
index dadae6255d05..6c90c69e5311 100644
--- a/kernel/dma/debug.c
+++ b/kernel/dma/debug.c
@@ -792,7 +792,7 @@ static int dump_show(struct seq_file *seq, void *v)
 }
 DEFINE_SHOW_ATTRIBUTE(dump);
 
-static void dma_debug_fs_init(void)
+static int __init dma_debug_fs_init(void)
 {
 	struct dentry *dentry = debugfs_create_dir("dma-api", NULL);
 
@@ -805,7 +805,10 @@ static void dma_debug_fs_init(void)
 	debugfs_create_u32("nr_total_entries", 0444, dentry, &nr_total_entries);
 	debugfs_create_file("driver_filter", 0644, dentry, NULL, &filter_fops);
 	debugfs_create_file("dump", 0444, dentry, NULL, &dump_fops);
+
+	return 0;
 }
+core_initcall_sync(dma_debug_fs_init);
 
 static int device_dma_allocations(struct device *dev, struct dma_debug_entry **out_entry)
 {
@@ -890,8 +893,6 @@ static int dma_debug_init(void)
 		spin_lock_init(&dma_entry_hash[i].lock);
 	}
 
-	dma_debug_fs_init();
-
 	nr_pages = DIV_ROUND_UP(nr_prealloc_entries, DMA_DEBUG_DYNAMIC_ENTRIES);
 	for (i = 0; i < nr_pages; ++i)
 		dma_debug_create_entries(GFP_KERNEL);
@@ -1064,20 +1065,10 @@ static void check_for_stack(struct device *dev,
 	}
 }
 
-static inline bool overlap(void *addr, unsigned long len, void *start, void *end)
-{
-	unsigned long a1 = (unsigned long)addr;
-	unsigned long b1 = a1 + len;
-	unsigned long a2 = (unsigned long)start;
-	unsigned long b2 = (unsigned long)end;
-
-	return !(b1 <= a2 || a1 >= b2);
-}
-
 static void check_for_illegal_area(struct device *dev, void *addr, unsigned long len)
 {
-	if (overlap(addr, len, _stext, _etext) ||
-	    overlap(addr, len, __start_rodata, __end_rodata))
+	if (memory_intersects(_stext, _etext, addr, len) ||
+	    memory_intersects(__start_rodata, __end_rodata, addr, len))
 		err_printk(dev, NULL, "device driver maps memory from kernel text or rodata [addr=%p] [len=%lu]\n", addr, len);
 }
 
diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c
index f737e3347059..8dca4f97d12d 100644
--- a/kernel/dma/direct.c
+++ b/kernel/dma/direct.c
@@ -156,9 +156,14 @@ void *dma_direct_alloc(struct device *dev, size_t size,
 
 	if (!IS_ENABLED(CONFIG_ARCH_HAS_DMA_SET_UNCACHED) &&
 	    !IS_ENABLED(CONFIG_DMA_DIRECT_REMAP) &&
+	    !IS_ENABLED(CONFIG_DMA_GLOBAL_POOL) &&
 	    !dev_is_dma_coherent(dev))
 		return arch_dma_alloc(dev, size, dma_handle, gfp, attrs);
 
+	if (IS_ENABLED(CONFIG_DMA_GLOBAL_POOL) &&
+	    !dev_is_dma_coherent(dev))
+		return dma_alloc_from_global_coherent(dev, size, dma_handle);
+
 	/*
 	 * Remapping or decrypting memory may block. If either is required and
 	 * we can't block, allocate the memory from the atomic pools.
@@ -255,11 +260,19 @@ void dma_direct_free(struct device *dev, size_t size,
 
 	if (!IS_ENABLED(CONFIG_ARCH_HAS_DMA_SET_UNCACHED) &&
 	    !IS_ENABLED(CONFIG_DMA_DIRECT_REMAP) &&
+	    !IS_ENABLED(CONFIG_DMA_GLOBAL_POOL) &&
 	    !dev_is_dma_coherent(dev)) {
 		arch_dma_free(dev, size, cpu_addr, dma_addr, attrs);
 		return;
 	}
 
+	if (IS_ENABLED(CONFIG_DMA_GLOBAL_POOL) &&
+	    !dev_is_dma_coherent(dev)) {
+		if (!dma_release_from_global_coherent(page_order, cpu_addr))
+			WARN_ON_ONCE(1);
+		return;
+	}
+
 	/* If cpu_addr is not from an atomic pool, dma_free_from_pool() fails */
 	if (IS_ENABLED(CONFIG_DMA_COHERENT_POOL) &&
 	    dma_free_from_pool(dev, cpu_addr, PAGE_ALIGN(size)))
@@ -411,7 +424,7 @@ int dma_direct_map_sg(struct device *dev, struct scatterlist *sgl, int nents,
 
 out_unmap:
 	dma_direct_unmap_sg(dev, sgl, i, dir, attrs | DMA_ATTR_SKIP_CPU_SYNC);
-	return 0;
+	return -EIO;
 }
 
 dma_addr_t dma_direct_map_resource(struct device *dev, phys_addr_t paddr,
@@ -462,6 +475,8 @@ int dma_direct_mmap(struct device *dev, struct vm_area_struct *vma,
 
 	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
 		return ret;
+	if (dma_mmap_from_global_coherent(vma, cpu_addr, size, &ret))
+		return ret;
 
 	if (vma->vm_pgoff >= count || user_count > count - vma->vm_pgoff)
 		return -ENXIO;
diff --git a/kernel/dma/dummy.c b/kernel/dma/dummy.c
index eacd4c5b10bf..b492d59ac77e 100644
--- a/kernel/dma/dummy.c
+++ b/kernel/dma/dummy.c
@@ -22,7 +22,7 @@ static int dma_dummy_map_sg(struct device *dev, struct scatterlist *sgl,
 		int nelems, enum dma_data_direction dir,
 		unsigned long attrs)
 {
-	return 0;
+	return -EINVAL;
 }
 
 static int dma_dummy_supported(struct device *hwdev, u64 mask)
diff --git a/kernel/dma/mapping.c b/kernel/dma/mapping.c
index 2b06a809d0b9..7ee5284bff58 100644
--- a/kernel/dma/mapping.c
+++ b/kernel/dma/mapping.c
@@ -177,12 +177,8 @@ void dma_unmap_page_attrs(struct device *dev, dma_addr_t addr, size_t size,
 }
 EXPORT_SYMBOL(dma_unmap_page_attrs);
 
-/*
- * dma_maps_sg_attrs returns 0 on error and > 0 on success.
- * It should never return a value < 0.
- */
-int dma_map_sg_attrs(struct device *dev, struct scatterlist *sg, int nents,
-		enum dma_data_direction dir, unsigned long attrs)
+static int __dma_map_sg_attrs(struct device *dev, struct scatterlist *sg,
+	 int nents, enum dma_data_direction dir, unsigned long attrs)
 {
 	const struct dma_map_ops *ops = get_dma_ops(dev);
 	int ents;
@@ -197,13 +193,81 @@ int dma_map_sg_attrs(struct device *dev, struct scatterlist *sg, int nents,
 		ents = dma_direct_map_sg(dev, sg, nents, dir, attrs);
 	else
 		ents = ops->map_sg(dev, sg, nents, dir, attrs);
-	BUG_ON(ents < 0);
-	debug_dma_map_sg(dev, sg, nents, ents, dir);
+
+	if (ents > 0)
+		debug_dma_map_sg(dev, sg, nents, ents, dir);
+	else if (WARN_ON_ONCE(ents != -EINVAL && ents != -ENOMEM &&
+			      ents != -EIO))
+		return -EIO;
 
 	return ents;
 }
+
+/**
+ * dma_map_sg_attrs - Map the given buffer for DMA
+ * @dev:	The device for which to perform the DMA operation
+ * @sg:	The sg_table object describing the buffer
+ * @dir:	DMA direction
+ * @attrs:	Optional DMA attributes for the map operation
+ *
+ * Maps a buffer described by a scatterlist passed in the sg argument with
+ * nents segments for the @dir DMA operation by the @dev device.
+ *
+ * Returns the number of mapped entries (which can be less than nents)
+ * on success. Zero is returned for any error.
+ *
+ * dma_unmap_sg_attrs() should be used to unmap the buffer with the
+ * original sg and original nents (not the value returned by this funciton).
+ */
+unsigned int dma_map_sg_attrs(struct device *dev, struct scatterlist *sg,
+		    int nents, enum dma_data_direction dir, unsigned long attrs)
+{
+	int ret;
+
+	ret = __dma_map_sg_attrs(dev, sg, nents, dir, attrs);
+	if (ret < 0)
+		return 0;
+	return ret;
+}
 EXPORT_SYMBOL(dma_map_sg_attrs);
 
+/**
+ * dma_map_sgtable - Map the given buffer for DMA
+ * @dev:	The device for which to perform the DMA operation
+ * @sgt:	The sg_table object describing the buffer
+ * @dir:	DMA direction
+ * @attrs:	Optional DMA attributes for the map operation
+ *
+ * Maps a buffer described by a scatterlist stored in the given sg_table
+ * object for the @dir DMA operation by the @dev device. After success, the
+ * ownership for the buffer is transferred to the DMA domain.  One has to
+ * call dma_sync_sgtable_for_cpu() or dma_unmap_sgtable() to move the
+ * ownership of the buffer back to the CPU domain before touching the
+ * buffer by the CPU.
+ *
+ * Returns 0 on success or a negative error code on error. The following
+ * error codes are supported with the given meaning:
+ *
+ *   -EINVAL - An invalid argument, unaligned access or other error
+ *	       in usage. Will not succeed if retried.
+ *   -ENOMEM - Insufficient resources (like memory or IOVA space) to
+ *	       complete the mapping. Should succeed if retried later.
+ *   -EIO    - Legacy error code with an unknown meaning. eg. this is
+ *	       returned if a lower level call returned DMA_MAPPING_ERROR.
+ */
+int dma_map_sgtable(struct device *dev, struct sg_table *sgt,
+		    enum dma_data_direction dir, unsigned long attrs)
+{
+	int nents;
+
+	nents = __dma_map_sg_attrs(dev, sgt->sgl, sgt->orig_nents, dir, attrs);
+	if (nents < 0)
+		return nents;
+	sgt->nents = nents;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dma_map_sgtable);
+
 void dma_unmap_sg_attrs(struct device *dev, struct scatterlist *sg,
 				      int nents, enum dma_data_direction dir,
 				      unsigned long attrs)