Merge tag 'dma-mapping-4.20' of git://git.infradead.org/users/hch/dma-mapping

Pull dma mapping updates from Christoph Hellwig:
 "First batch of dma-mapping changes for 4.20.

  There will be a second PR as some big changes were only applied just
  before the end of the merge window, and I want to give them a few more
  days in linux-next.

  Summary:

   - mostly more consolidation of the direct mapping code, including
     converting over hexagon, and merging the coherent and non-coherent
     code into a single dma_map_ops instance (me)

   - cleanups for the dma_configure/dma_unconfigure callchains (me)

   - better handling of dma_masks in odd setups (me, Alexander Duyck)

   - better debugging of passing vmalloc address to the DMA API (Stephen
     Boyd)

   - CMA command line parsing fix (He Zhe)"

* tag 'dma-mapping-4.20' of git://git.infradead.org/users/hch/dma-mapping: (27 commits)
  dma-direct: respect DMA_ATTR_NO_WARN
  dma-mapping: translate __GFP_NOFAIL to DMA_ATTR_NO_WARN
  dma-direct: document the zone selection logic
  dma-debug: Check for drivers mapping invalid addresses in dma_map_single()
  dma-direct: fix return value of dma_direct_supported
  dma-mapping: move dma_default_get_required_mask under ifdef
  dma-direct: always allow dma mask <= physiscal memory size
  dma-direct: implement complete bus_dma_mask handling
  dma-direct: refine dma_direct_alloc zone selection
  dma-direct: add an explicit dma_direct_get_required_mask
  dma-mapping: make the get_required_mask method available unconditionally
  unicore32: remove swiotlb support
  Revert "dma-mapping: clear dev->dma_ops in arch_teardown_dma_ops"
  dma-mapping: support non-coherent devices in dma_common_get_sgtable
  dma-mapping: consolidate the dma mmap implementations
  dma-mapping: merge direct and noncoherent ops
  dma-mapping: move the dma_coherent flag to struct device
  MIPS: don't select DMA_MAYBE_COHERENT from DMA_PERDEV_COHERENT
  dma-mapping: add the missing ARCH_HAS_SYNC_DMA_FOR_CPU_ALL declaration
  dma-mapping: fix panic caused by passing empty cma command line argument
  ...

Author: torvalds

arch/arc/Kconfig

@@ -9,6 +9,7 @@
 config ARC
 	def_bool y
 	select ARC_TIMERS
+	select ARCH_HAS_DMA_COHERENT_TO_PFN
 	select ARCH_HAS_PTE_SPECIAL
 	select ARCH_HAS_SYNC_DMA_FOR_CPU
 	select ARCH_HAS_SYNC_DMA_FOR_DEVICE
@@ -17,8 +18,7 @@ config ARC
 	select BUILDTIME_EXTABLE_SORT
 	select CLONE_BACKWARDS
 	select COMMON_CLK
-	select DMA_NONCOHERENT_OPS
-	select DMA_NONCOHERENT_MMAP
+	select DMA_DIRECT_OPS
 	select GENERIC_ATOMIC64 if !ISA_ARCV2 || !(ARC_HAS_LL64 && ARC_HAS_LLSC)
 	select GENERIC_CLOCKEVENTS
 	select GENERIC_FIND_FIRST_BIT

arch/arc/mm/dma.c

@@ -84,29 +84,10 @@ void arch_dma_free(struct device *dev, size_t size, void *vaddr,
 	__free_pages(page, get_order(size));
 }
 
-int arch_dma_mmap(struct device *dev, struct vm_area_struct *vma,
-		void *cpu_addr, dma_addr_t dma_addr, size_t size,
-		unsigned long attrs)
+long arch_dma_coherent_to_pfn(struct device *dev, void *cpu_addr,
+		dma_addr_t dma_addr)
 {
-	unsigned long user_count = vma_pages(vma);
-	unsigned long count = PAGE_ALIGN(size) >> PAGE_SHIFT;
-	unsigned long pfn = __phys_to_pfn(dma_addr);
-	unsigned long off = vma->vm_pgoff;
-	int ret = -ENXIO;
-
-	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
-
-	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
-		return ret;
-
-	if (off < count && user_count <= (count - off)) {
-		ret = remap_pfn_range(vma, vma->vm_start,
-				      pfn + off,
-				      user_count << PAGE_SHIFT,
-				      vma->vm_page_prot);
-	}
-
-	return ret;
+	return __phys_to_pfn(dma_addr);
 }
 
 /*
@@ -167,21 +148,19 @@ void arch_sync_dma_for_cpu(struct device *dev, phys_addr_t paddr,
 }
 
 /*
- * Plug in coherent or noncoherent dma ops
+ * Plug in direct dma map ops.
  */
 void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
 			const struct iommu_ops *iommu, bool coherent)
 {
 	/*
 	 * IOC hardware snoops all DMA traffic keeping the caches consistent
 	 * with memory - eliding need for any explicit cache maintenance of
-	 * DMA buffers - so we can use dma_direct cache ops.
+	 * DMA buffers.
 	 */
-	if (is_isa_arcv2() && ioc_enable && coherent) {
-		set_dma_ops(dev, &dma_direct_ops);
-		dev_info(dev, "use dma_direct_ops cache ops\n");
-	} else {
-		set_dma_ops(dev, &dma_noncoherent_ops);
-		dev_info(dev, "use dma_noncoherent_ops cache ops\n");
-	}
+	if (is_isa_arcv2() && ioc_enable && coherent)
+		dev->dma_coherent = true;
+
+	dev_info(dev, "use %sncoherent DMA ops\n",
+		 dev->dma_coherent ? "" : "non");
 }

arch/arm/include/asm/dma-mapping.h

@@ -100,8 +100,10 @@ static inline unsigned long dma_max_pfn(struct device *dev)
 extern void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
 			       const struct iommu_ops *iommu, bool coherent);
 
+#ifdef CONFIG_MMU
 #define arch_teardown_dma_ops arch_teardown_dma_ops
 extern void arch_teardown_dma_ops(struct device *dev);
+#endif
 
 /* do not use this function in a driver */
 static inline bool is_device_dma_coherent(struct device *dev)

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

@@ -47,7 +47,8 @@ static void *arm_nommu_dma_alloc(struct device *dev, size_t size,
 	 */
 
 	if (attrs & DMA_ATTR_NON_CONSISTENT)
-		return dma_direct_alloc(dev, size, dma_handle, gfp, attrs);
+		return dma_direct_alloc_pages(dev, size, dma_handle, gfp,
+				attrs);
 
 	ret = dma_alloc_from_global_coherent(size, dma_handle);
 
@@ -70,7 +71,7 @@ static void arm_nommu_dma_free(struct device *dev, size_t size,
 			       unsigned long attrs)
 {
 	if (attrs & DMA_ATTR_NON_CONSISTENT) {
-		dma_direct_free(dev, size, cpu_addr, dma_addr, attrs);
+		dma_direct_free_pages(dev, size, cpu_addr, dma_addr, attrs);
 	} else {
 		int ret = dma_release_from_global_coherent(get_order(size),
 							   cpu_addr);
@@ -90,7 +91,7 @@ static int arm_nommu_dma_mmap(struct device *dev, struct vm_area_struct *vma,
 	if (dma_mmap_from_global_coherent(vma, cpu_addr, size, &ret))
 		return ret;
 
-	return dma_common_mmap(dev, vma, cpu_addr, dma_addr, size);
+	return dma_common_mmap(dev, vma, cpu_addr, dma_addr, size, attrs);
 }
 
 
@@ -237,7 +238,3 @@ void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
 
 	set_dma_ops(dev, dma_ops);
 }
-
-void arch_teardown_dma_ops(struct device *dev)
-{
-}

arch/c6x/Kconfig

@@ -9,7 +9,7 @@ config C6X
 	select ARCH_HAS_SYNC_DMA_FOR_CPU
 	select ARCH_HAS_SYNC_DMA_FOR_DEVICE
 	select CLKDEV_LOOKUP
-	select DMA_NONCOHERENT_OPS
+	select DMA_DIRECT_OPS
 	select GENERIC_ATOMIC64
 	select GENERIC_IRQ_SHOW
 	select HAVE_ARCH_TRACEHOOK

arch/hexagon/Kconfig

@@ -4,6 +4,7 @@ comment "Linux Kernel Configuration for Hexagon"
 
 config HEXAGON
 	def_bool y
+	select ARCH_HAS_SYNC_DMA_FOR_DEVICE
 	select ARCH_NO_PREEMPT
 	select HAVE_OPROFILE
 	# Other pending projects/to-do items.
@@ -29,6 +30,7 @@ config HEXAGON
 	select GENERIC_CLOCKEVENTS_BROADCAST
 	select MODULES_USE_ELF_RELA
 	select GENERIC_CPU_DEVICES
+	select DMA_DIRECT_OPS
 	---help---
 	  Qualcomm Hexagon is a processor architecture designed for high
 	  performance and low power across a wide variety of applications.

arch/hexagon/include/asm/Kbuild

@@ -6,6 +6,7 @@ generic-y += compat.h
 generic-y += current.h
 generic-y += device.h
 generic-y += div64.h
+generic-y += dma-mapping.h
 generic-y += emergency-restart.h
 generic-y += extable.h
 generic-y += fb.h

arch/hexagon/include/asm/dma-mapping.h

@@ -18,32 +18,19 @@
  * 02110-1301, USA.
  */
 
-#include <linux/dma-mapping.h>
-#include <linux/dma-direct.h>
+#include <linux/dma-noncoherent.h>
 #include <linux/bootmem.h>
 #include <linux/genalloc.h>
-#include <asm/dma-mapping.h>
 #include <linux/module.h>
 #include <asm/page.h>
 
-#define HEXAGON_MAPPING_ERROR	0
-
-const struct dma_map_ops *dma_ops;
-EXPORT_SYMBOL(dma_ops);
-
-static inline void *dma_addr_to_virt(dma_addr_t dma_addr)
-{
-	return phys_to_virt((unsigned long) dma_addr);
-}
-
 static struct gen_pool *coherent_pool;
 
 
 /* Allocates from a pool of uncached memory that was reserved at boot time */
 
-static void *hexagon_dma_alloc_coherent(struct device *dev, size_t size,
-				 dma_addr_t *dma_addr, gfp_t flag,
-				 unsigned long attrs)
+void *arch_dma_alloc(struct device *dev, size_t size, dma_addr_t *dma_addr,
+		gfp_t flag, unsigned long attrs)
 {
 	void *ret;
 
@@ -75,58 +62,17 @@ static void *hexagon_dma_alloc_coherent(struct device *dev, size_t size,
 	return ret;
 }
 
-static void hexagon_free_coherent(struct device *dev, size_t size, void *vaddr,
-				  dma_addr_t dma_addr, unsigned long attrs)
+void arch_dma_free(struct device *dev, size_t size, void *vaddr,
+		dma_addr_t dma_addr, unsigned long attrs)
 {
 	gen_pool_free(coherent_pool, (unsigned long) vaddr, size);
 }
 
-static int check_addr(const char *name, struct device *hwdev,
-		      dma_addr_t bus, size_t size)
-{
-	if (hwdev && hwdev->dma_mask && !dma_capable(hwdev, bus, size)) {
-		if (*hwdev->dma_mask >= DMA_BIT_MASK(32))
-			printk(KERN_ERR
-				"%s: overflow %Lx+%zu of device mask %Lx\n",
-				name, (long long)bus, size,
-				(long long)*hwdev->dma_mask);
-		return 0;
-	}
-	return 1;
-}
-
-static int hexagon_map_sg(struct device *hwdev, struct scatterlist *sg,
-			  int nents, enum dma_data_direction dir,
-			  unsigned long attrs)
+void arch_sync_dma_for_device(struct device *dev, phys_addr_t paddr,
+		size_t size, enum dma_data_direction dir)
 {
-	struct scatterlist *s;
-	int i;
-
-	WARN_ON(nents == 0 || sg[0].length == 0);
-
-	for_each_sg(sg, s, nents, i) {
-		s->dma_address = sg_phys(s);
-		if (!check_addr("map_sg", hwdev, s->dma_address, s->length))
-			return 0;
-
-		s->dma_length = s->length;
+	void *addr = phys_to_virt(paddr);
 
-		if (attrs & DMA_ATTR_SKIP_CPU_SYNC)
-			continue;
-
-		flush_dcache_range(dma_addr_to_virt(s->dma_address),
-				   dma_addr_to_virt(s->dma_address + s->length));
-	}
-
-	return nents;
-}
-
-/*
- * address is virtual
- */
-static inline void dma_sync(void *addr, size_t size,
-			    enum dma_data_direction dir)
-{
 	switch (dir) {
 	case DMA_TO_DEVICE:
 		hexagon_clean_dcache_range((unsigned long) addr,
@@ -144,76 +90,3 @@ static inline void dma_sync(void *addr, size_t size,
 		BUG();
 	}
 }
-
-/**
- * hexagon_map_page() - maps an address for device DMA
- * @dev:	pointer to DMA device
- * @page:	pointer to page struct of DMA memory
- * @offset:	offset within page
- * @size:	size of memory to map
- * @dir:	transfer direction
- * @attrs:	pointer to DMA attrs (not used)
- *
- * Called to map a memory address to a DMA address prior
- * to accesses to/from device.
- *
- * We don't particularly have many hoops to jump through
- * so far.  Straight translation between phys and virtual.
- *
- * DMA is not cache coherent so sync is necessary; this
- * seems to be a convenient place to do it.
- *
- */
-static dma_addr_t hexagon_map_page(struct device *dev, struct page *page,
-				   unsigned long offset, size_t size,
-				   enum dma_data_direction dir,
-				   unsigned long attrs)
-{
-	dma_addr_t bus = page_to_phys(page) + offset;
-	WARN_ON(size == 0);
-
-	if (!check_addr("map_single", dev, bus, size))
-		return HEXAGON_MAPPING_ERROR;
-
-	if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC))
-		dma_sync(dma_addr_to_virt(bus), size, dir);
-
-	return bus;
-}
-
-static void hexagon_sync_single_for_cpu(struct device *dev,
-					dma_addr_t dma_handle, size_t size,
-					enum dma_data_direction dir)
-{
-	dma_sync(dma_addr_to_virt(dma_handle), size, dir);
-}
-
-static void hexagon_sync_single_for_device(struct device *dev,
-					dma_addr_t dma_handle, size_t size,
-					enum dma_data_direction dir)
-{
-	dma_sync(dma_addr_to_virt(dma_handle), size, dir);
-}
-
-static int hexagon_mapping_error(struct device *dev, dma_addr_t dma_addr)
-{
-	return dma_addr == HEXAGON_MAPPING_ERROR;
-}
-
-const struct dma_map_ops hexagon_dma_ops = {
-	.alloc		= hexagon_dma_alloc_coherent,
-	.free		= hexagon_free_coherent,
-	.map_sg		= hexagon_map_sg,
-	.map_page	= hexagon_map_page,
-	.sync_single_for_cpu = hexagon_sync_single_for_cpu,
-	.sync_single_for_device = hexagon_sync_single_for_device,
-	.mapping_error	= hexagon_mapping_error,
-};
-
-void __init hexagon_dma_init(void)
-{
-	if (dma_ops)
-		return;
-
-	dma_ops = &hexagon_dma_ops;
-}

arch/hexagon/kernel/dma.c

@@ -10,8 +10,6 @@
 #include <linux/scatterlist.h>
 #include <linux/dma-debug.h>
 
-#define ARCH_HAS_DMA_GET_REQUIRED_MASK
-
 extern const struct dma_map_ops *dma_ops;
 extern struct ia64_machine_vector ia64_mv;
 extern void set_iommu_machvec(void);