dma-mapping: move dma_map_(un)single() to <linux/dma-mapping.h>

The implementation of dma_map_single() and dma_unmap_single() is
exactly the same for all the architectures that support them.

Factor them out to <linux/dma-mapping.h>, and make all drivers to
include <linux/dma-mapping.h> instead of <asm/dma-mapping.h>.

If we need to differentiate them for some architectures, we can
move the generic definitions to <asm-generic/dma-mapping.h>.

Add some comments to the helpers. The concept is quite similar to
the DMA-API of Linux kernel. Drivers are agnostic about what is
going on behind the scene. Just call dma_map_single() before the
DMA, and dma_unmap_single() after it.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
diff --git a/arch/arm/include/asm/dma-mapping.h b/arch/arm/include/asm/dma-mapping.h
index efdbed7..8bb3fa2 100644
--- a/arch/arm/include/asm/dma-mapping.h
+++ b/arch/arm/include/asm/dma-mapping.h
@@ -14,8 +14,6 @@
 #include <linux/types.h>
 #include <malloc.h>
 
-#define	dma_mapping_error(x, y)	0
-
 static inline void *dma_alloc_coherent(size_t len, unsigned long *handle)
 {
 	*handle = (unsigned long)memalign(ARCH_DMA_MINALIGN, ROUND(len, ARCH_DMA_MINALIGN));
@@ -27,28 +25,4 @@
 	free(addr);
 }
 
-static inline dma_addr_t dma_map_single(void *vaddr, size_t len,
-					enum dma_data_direction dir)
-{
-	unsigned long addr = (unsigned long)vaddr;
-
-	len = ALIGN(len, ARCH_DMA_MINALIGN);
-
-	if (dir == DMA_FROM_DEVICE)
-		invalidate_dcache_range(addr, addr + len);
-	else
-		flush_dcache_range(addr, addr + len);
-
-	return addr;
-}
-
-static inline void dma_unmap_single(dma_addr_t addr, size_t len,
-				    enum dma_data_direction dir)
-{
-	len = ALIGN(len, ARCH_DMA_MINALIGN);
-
-	if (dir != DMA_TO_DEVICE)
-		invalidate_dcache_range(addr, addr + len);
-}
-
 #endif /* __ASM_ARM_DMA_MAPPING_H */
diff --git a/arch/nds32/include/asm/dma-mapping.h b/arch/nds32/include/asm/dma-mapping.h
index 784f489..f8668f1 100644
--- a/arch/nds32/include/asm/dma-mapping.h
+++ b/arch/nds32/include/asm/dma-mapping.h
@@ -19,28 +19,4 @@
 	return (void *)*handle;
 }
 
-static inline dma_addr_t dma_map_single(void *vaddr, size_t len,
-					enum dma_data_direction dir)
-{
-	unsigned long addr = (unsigned long)vaddr;
-
-	len = ALIGN(len, ARCH_DMA_MINALIGN);
-
-	if (dir == DMA_FROM_DEVICE)
-		invalidate_dcache_range(addr, addr + len);
-	else
-		flush_dcache_range(addr, addr + len);
-
-	return addr;
-}
-
-static inline void dma_unmap_single(dma_addr_t addr, size_t len,
-				    enum dma_data_direction dir)
-{
-	len = ALIGN(len, ARCH_DMA_MINALIGN);
-
-	if (dir != DMA_TO_DEVICE)
-		invalidate_dcache_range(addr, addr + len);
-}
-
 #endif /* __ASM_NDS_DMA_MAPPING_H */
diff --git a/arch/riscv/include/asm/dma-mapping.h b/arch/riscv/include/asm/dma-mapping.h
index 1ac4a4f..6ecadab 100644
--- a/arch/riscv/include/asm/dma-mapping.h
+++ b/arch/riscv/include/asm/dma-mapping.h
@@ -16,8 +16,6 @@
 #include <linux/dma-direction.h>
 #include <malloc.h>
 
-#define dma_mapping_error(x, y)	0
-
 static inline void *dma_alloc_coherent(size_t len, unsigned long *handle)
 {
 	*handle = (unsigned long)memalign(ARCH_DMA_MINALIGN, len);
@@ -29,28 +27,4 @@
 	free(addr);
 }
 
-static inline dma_addr_t dma_map_single(void *vaddr, size_t len,
-					enum dma_data_direction dir)
-{
-	unsigned long addr = (unsigned long)vaddr;
-
-	len = ALIGN(len, ARCH_DMA_MINALIGN);
-
-	if (dir == DMA_FROM_DEVICE)
-		invalidate_dcache_range(addr, addr + len);
-	else
-		flush_dcache_range(addr, addr + len);
-
-	return addr;
-}
-
-static inline void dma_unmap_single(dma_addr_t addr, size_t len,
-				    enum dma_data_direction dir)
-{
-	len = ALIGN(len, ARCH_DMA_MINALIGN);
-
-	if (dir != DMA_TO_DEVICE)
-		invalidate_dcache_range(addr, addr + len);
-}
-
 #endif /* __ASM_RISCV_DMA_MAPPING_H */
diff --git a/arch/x86/include/asm/dma-mapping.h b/arch/x86/include/asm/dma-mapping.h
index 37704da..8be1003 100644
--- a/arch/x86/include/asm/dma-mapping.h
+++ b/arch/x86/include/asm/dma-mapping.h
@@ -14,8 +14,6 @@
 #include <linux/types.h>
 #include <malloc.h>
 
-#define	dma_mapping_error(x, y)	0
-
 static inline void *dma_alloc_coherent(size_t len, unsigned long *handle)
 {
 	*handle = (unsigned long)memalign(ARCH_DMA_MINALIGN, len);
@@ -27,28 +25,4 @@
 	free(addr);
 }
 
-static inline dma_addr_t dma_map_single(void *vaddr, size_t len,
-					enum dma_data_direction dir)
-{
-	unsigned long addr = (unsigned long)vaddr;
-
-	len = ALIGN(len, ARCH_DMA_MINALIGN);
-
-	if (dir == DMA_FROM_DEVICE)
-		invalidate_dcache_range(addr, addr + len);
-	else
-		flush_dcache_range(addr, addr + len);
-
-	return addr;
-}
-
-static inline void dma_unmap_single(dma_addr_t addr, size_t len,
-				    enum dma_data_direction dir)
-{
-	len = ALIGN(len, ARCH_DMA_MINALIGN);
-
-	if (dir != DMA_TO_DEVICE)
-		invalidate_dcache_range(addr, addr + len);
-}
-
 #endif /* __ASM_X86_DMA_MAPPING_H */
diff --git a/drivers/dma/ti/k3-udma.c b/drivers/dma/ti/k3-udma.c
index f274100..e587f1f 100644
--- a/drivers/dma/ti/k3-udma.c
+++ b/drivers/dma/ti/k3-udma.c
@@ -10,7 +10,7 @@
 #include <asm/io.h>
 #include <asm/bitops.h>
 #include <malloc.h>
-#include <asm/dma-mapping.h>
+#include <linux/dma-mapping.h>
 #include <dm.h>
 #include <dm/device_compat.h>
 #include <dm/devres.h>
diff --git a/drivers/mmc/tmio-common.c b/drivers/mmc/tmio-common.c
index 5321388..faf1819 100644
--- a/drivers/mmc/tmio-common.c
+++ b/drivers/mmc/tmio-common.c
@@ -4,7 +4,6 @@
  *   Author: Masahiro Yamada <yamada.masahiro@socionext.com>
  */
 
-#include <asm/dma-mapping.h>
 #include <common.h>
 #include <clk.h>
 #include <cpu_func.h>
@@ -14,7 +13,7 @@
 #include <dm/device_compat.h>
 #include <dm/pinctrl.h>
 #include <linux/compat.h>
-#include <linux/dma-direction.h>
+#include <linux/dma-mapping.h>
 #include <linux/io.h>
 #include <linux/sizes.h>
 #include <power/regulator.h>
diff --git a/drivers/mtd/nand/raw/denali.c b/drivers/mtd/nand/raw/denali.c
index 3e0ac39..b525b1b 100644
--- a/drivers/mtd/nand/raw/denali.c
+++ b/drivers/mtd/nand/raw/denali.c
@@ -5,14 +5,13 @@
  * Copyright (C) 2009-2010, Intel Corporation and its suppliers.
  */
 
-#include <asm/dma-mapping.h>
 #include <dm.h>
 #include <malloc.h>
 #include <nand.h>
 #include <dm/device_compat.h>
 #include <dm/devres.h>
 #include <linux/bitfield.h>
-#include <linux/dma-direction.h>
+#include <linux/dma-mapping.h>
 #include <linux/err.h>
 #include <linux/errno.h>
 #include <linux/io.h>
diff --git a/drivers/net/altera_tse.c b/drivers/net/altera_tse.c
index aabddd6..a511068 100644
--- a/drivers/net/altera_tse.c
+++ b/drivers/net/altera_tse.c
@@ -17,7 +17,7 @@
 #include <miiphy.h>
 #include <net.h>
 #include <asm/cache.h>
-#include <asm/dma-mapping.h>
+#include <linux/dma-mapping.h>
 #include <asm/io.h>
 #include "altera_tse.h"
 
diff --git a/drivers/net/ftmac110.c b/drivers/net/ftmac110.c
index 1fa93d4..8408240 100644
--- a/drivers/net/ftmac110.c
+++ b/drivers/net/ftmac110.c
@@ -12,7 +12,7 @@
 #include <net.h>
 #include <linux/errno.h>
 #include <asm/io.h>
-#include <asm/dma-mapping.h>
+#include <linux/dma-mapping.h>
 
 #if defined(CONFIG_MII) || defined(CONFIG_CMD_MII)
 #include <miiphy.h>
diff --git a/drivers/net/macb.c b/drivers/net/macb.c
index 7a2b1ab..631b53b 100644
--- a/drivers/net/macb.c
+++ b/drivers/net/macb.c
@@ -38,7 +38,7 @@
 
 #include <linux/mii.h>
 #include <asm/io.h>
-#include <asm/dma-mapping.h>
+#include <linux/dma-mapping.h>
 #include <asm/arch/clk.h>
 #include <linux/errno.h>
 
diff --git a/drivers/soc/ti/k3-navss-ringacc.c b/drivers/soc/ti/k3-navss-ringacc.c
index 8cbfe2b..c5661c5 100644
--- a/drivers/soc/ti/k3-navss-ringacc.c
+++ b/drivers/soc/ti/k3-navss-ringacc.c
@@ -9,7 +9,6 @@
 #include <cpu_func.h>
 #include <asm/io.h>
 #include <malloc.h>
-#include <asm/dma-mapping.h>
 #include <asm/bitops.h>
 #include <dm.h>
 #include <dm/device_compat.h>
@@ -17,6 +16,7 @@
 #include <dm/read.h>
 #include <dm/uclass.h>
 #include <linux/compat.h>
+#include <linux/dma-mapping.h>
 #include <linux/err.h>
 #include <linux/soc/ti/k3-navss-ringacc.h>
 #include <linux/soc/ti/ti_sci_protocol.h>
diff --git a/drivers/ufs/ufs.c b/drivers/ufs/ufs.c
index c9346c2..24e1bc2 100644
--- a/drivers/ufs/ufs.c
+++ b/drivers/ufs/ufs.c
@@ -19,7 +19,7 @@
 #include <hexdump.h>
 #include <scsi.h>
 
-#include <asm/dma-mapping.h>
+#include <linux/dma-mapping.h>
 
 #include "ufs.h"
 
diff --git a/drivers/usb/cdns3/gadget.c b/drivers/usb/cdns3/gadget.c
index 22e90a5..8377eb4 100644
--- a/drivers/usb/cdns3/gadget.c
+++ b/drivers/usb/cdns3/gadget.c
@@ -63,7 +63,7 @@
 #include <linux/usb/gadget.h>
 #include <linux/compat.h>
 #include <linux/iopoll.h>
-#include <asm/dma-mapping.h>
+#include <linux/dma-mapping.h>
 #include <linux/bitmap.h>
 #include <linux/bug.h>
 
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index 6ad1000..4ec3f6d 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -17,9 +17,9 @@
 #include <cpu_func.h>
 #include <malloc.h>
 #include <dwc3-uboot.h>
-#include <asm/dma-mapping.h>
 #include <dm/device_compat.h>
 #include <dm/devres.h>
+#include <linux/dma-mapping.h>
 #include <linux/err.h>
 #include <linux/ioport.h>
 #include <dm.h>
diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index 1502d67..e445c70 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -16,10 +16,10 @@
 #include <common.h>
 #include <cpu_func.h>
 #include <malloc.h>
-#include <asm/dma-mapping.h>
 #include <dm/device_compat.h>
 #include <dm/devres.h>
 #include <linux/bug.h>
+#include <linux/dma-mapping.h>
 #include <linux/list.h>
 
 #include <linux/usb/ch9.h>
diff --git a/drivers/usb/gadget/udc/udc-core.c b/drivers/usb/gadget/udc/udc-core.c
index 5288745..7f73926 100644
--- a/drivers/usb/gadget/udc/udc-core.c
+++ b/drivers/usb/gadget/udc/udc-core.c
@@ -18,7 +18,7 @@
 #include <linux/compat.h>
 #include <malloc.h>
 #include <asm/cache.h>
-#include <asm/dma-mapping.h>
+#include <linux/dma-mapping.h>
 #include <common.h>
 #include <dm.h>
 #include <dm/device-internal.h>
diff --git a/include/linux/dma-mapping.h b/include/linux/dma-mapping.h
new file mode 100644
index 0000000..20b6d60
--- /dev/null
+++ b/include/linux/dma-mapping.h
@@ -0,0 +1,63 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _LINUX_DMA_MAPPING_H
+#define _LINUX_DMA_MAPPING_H
+
+#include <linux/dma-direction.h>
+#include <linux/types.h>
+#include <asm/dma-mapping.h>
+#include <cpu_func.h>
+
+#define dma_mapping_error(x, y)	0
+
+/**
+ * Map a buffer to make it available to the DMA device
+ *
+ * Linux-like DMA API that is intended to be used from drivers. This hides the
+ * underlying cache operation from drivers. Call this before starting the DMA
+ * transfer. In most of architectures in U-Boot, the virtual address matches to
+ * the physical address (but we have exceptions like sandbox). U-Boot does not
+ * support iommu at the driver level, so it also matches to the DMA address.
+ * Hence, this helper currently just performs the cache operation, then returns
+ * straight-mapped dma_address, which is intended to be set to the register of
+ * the DMA device.
+ *
+ * @vaddr: address of the buffer
+ * @len: length of the buffer
+ * @dir: the direction of DMA
+ */
+static inline dma_addr_t dma_map_single(void *vaddr, size_t len,
+					enum dma_data_direction dir)
+{
+	unsigned long addr = (unsigned long)vaddr;
+
+	len = ALIGN(len, ARCH_DMA_MINALIGN);
+
+	if (dir == DMA_FROM_DEVICE)
+		invalidate_dcache_range(addr, addr + len);
+	else
+		flush_dcache_range(addr, addr + len);
+
+	return addr;
+}
+
+/**
+ * Unmap a buffer to make it available to CPU
+ *
+ * Linux-like DMA API that is intended to be used from drivers. This hides the
+ * underlying cache operation from drivers. Call this after finishin the DMA
+ * transfer.
+ *
+ * @addr: DMA address
+ * @len: length of the buffer
+ * @dir: the direction of DMA
+ */
+static inline void dma_unmap_single(dma_addr_t addr, size_t len,
+				    enum dma_data_direction dir)
+{
+	len = ALIGN(len, ARCH_DMA_MINALIGN);
+
+	if (dir != DMA_TO_DEVICE)
+		invalidate_dcache_range(addr, addr + len);
+}
+
+#endif