Merge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx:
  DMAENGINE: DMA40 U8500 platform configuration
  DMA: PL330: Add dma api driver

arch/arm/mach-ux500/clock.c

@@ -411,7 +411,7 @@ static struct clk_lookup u8500_common_clks[] = {
 	CLK(apetraceclk,	"apetrace",	NULL),
 	CLK(mcdeclk,	"mcde",		NULL),
 	CLK(ipi2clk,	"ipi2",		NULL),
-	CLK(dmaclk,	"dma40",	NULL),
+	CLK(dmaclk,	"dma40.0",	NULL),
 	CLK(b2r2clk,	"b2r2",		NULL),
 	CLK(tvclk,	"tv",		NULL),
 };

arch/arm/mach-ux500/cpu-db8500.c

@@ -32,6 +32,7 @@ static struct platform_device *platform_devs[] __initdata = {
 	&u8500_gpio_devs[6],
 	&u8500_gpio_devs[7],
 	&u8500_gpio_devs[8],
+	&u8500_dma40_device,
 };
 
 /* minimum static i/o mapping required to boot U8500 platforms */
@@ -71,6 +72,9 @@ void __init u8500_init_devices(void)
 {
 	ux500_init_devices();
 
+	if (cpu_is_u8500ed())
+		dma40_u8500ed_fixup();
+
 	/* Register the platform devices */
 	platform_add_devices(platform_devs, ARRAY_SIZE(platform_devs));
 

arch/arm/mach-ux500/devices-db8500.c

@@ -12,9 +12,13 @@
 #include <linux/gpio.h>
 #include <linux/amba/bus.h>
 
+#include <plat/ste_dma40.h>
+
 #include <mach/hardware.h>
 #include <mach/setup.h>
 
+#include "ste-dma40-db8500.h"
+
 static struct nmk_gpio_platform_data u8500_gpio_data[] = {
 	GPIO_DATA("GPIO-0-31", 0),
 	GPIO_DATA("GPIO-32-63", 32), /* 37..63 not routed to pin */
@@ -105,3 +109,108 @@ struct platform_device u8500_i2c4_device = {
 	.resource	= u8500_i2c4_resources,
 	.num_resources	= ARRAY_SIZE(u8500_i2c4_resources),
 };
+
+static struct resource dma40_resources[] = {
+	[0] = {
+		.start = U8500_DMA_BASE,
+		.end = U8500_DMA_BASE + SZ_4K - 1,
+		.flags = IORESOURCE_MEM,
+		.name = "base",
+	},
+	[1] = {
+		.start = U8500_DMA_LCPA_BASE,
+		.end = U8500_DMA_LCPA_BASE + SZ_4K - 1,
+		.flags = IORESOURCE_MEM,
+		.name = "lcpa",
+	},
+	[2] = {
+		.start = U8500_DMA_LCLA_BASE,
+		.end = U8500_DMA_LCLA_BASE + 16 * 1024 - 1,
+		.flags = IORESOURCE_MEM,
+		.name = "lcla",
+	},
+	[3] = {
+		.start = IRQ_DMA,
+		.end = IRQ_DMA,
+		.flags = IORESOURCE_IRQ}
+};
+
+/* Default configuration for physcial memcpy */
+struct stedma40_chan_cfg dma40_memcpy_conf_phy = {
+	.channel_type = (STEDMA40_CHANNEL_IN_PHY_MODE |
+			 STEDMA40_LOW_PRIORITY_CHANNEL |
+			 STEDMA40_PCHAN_BASIC_MODE),
+	.dir = STEDMA40_MEM_TO_MEM,
+
+	.src_info.endianess = STEDMA40_LITTLE_ENDIAN,
+	.src_info.data_width = STEDMA40_BYTE_WIDTH,
+	.src_info.psize = STEDMA40_PSIZE_PHY_1,
+
+	.dst_info.endianess = STEDMA40_LITTLE_ENDIAN,
+	.dst_info.data_width = STEDMA40_BYTE_WIDTH,
+	.dst_info.psize = STEDMA40_PSIZE_PHY_1,
+
+};
+/* Default configuration for logical memcpy */
+struct stedma40_chan_cfg dma40_memcpy_conf_log = {
+	.channel_type = (STEDMA40_CHANNEL_IN_LOG_MODE |
+			 STEDMA40_LOW_PRIORITY_CHANNEL |
+			 STEDMA40_LCHAN_SRC_LOG_DST_LOG |
+			 STEDMA40_NO_TIM_FOR_LINK),
+	.dir = STEDMA40_MEM_TO_MEM,
+
+	.src_info.endianess = STEDMA40_LITTLE_ENDIAN,
+	.src_info.data_width = STEDMA40_BYTE_WIDTH,
+	.src_info.psize = STEDMA40_PSIZE_LOG_1,
+
+	.dst_info.endianess = STEDMA40_LITTLE_ENDIAN,
+	.dst_info.data_width = STEDMA40_BYTE_WIDTH,
+	.dst_info.psize = STEDMA40_PSIZE_LOG_1,
+
+};
+
+/*
+ * Mapping between destination event lines and physical device address.
+ * The event line is tied to a device and therefor the address is constant.
+ */
+static const dma_addr_t dma40_tx_map[STEDMA40_NR_DEV];
+
+/* Mapping between source event lines and physical device address */
+static const dma_addr_t dma40_rx_map[STEDMA40_NR_DEV];
+
+/* Reserved event lines for memcpy only */
+static int dma40_memcpy_event[] = {
+	STEDMA40_MEMCPY_TX_1,
+	STEDMA40_MEMCPY_TX_2,
+	STEDMA40_MEMCPY_TX_3,
+	STEDMA40_MEMCPY_TX_4,
+};
+
+static struct stedma40_platform_data dma40_plat_data = {
+	.dev_len = STEDMA40_NR_DEV,
+	.dev_rx = dma40_rx_map,
+	.dev_tx = dma40_tx_map,
+	.memcpy = dma40_memcpy_event,
+	.memcpy_len = ARRAY_SIZE(dma40_memcpy_event),
+	.memcpy_conf_phy = &dma40_memcpy_conf_phy,
+	.memcpy_conf_log = &dma40_memcpy_conf_log,
+	.llis_per_log = 8,
+};
+
+struct platform_device u8500_dma40_device = {
+	.dev = {
+		.platform_data = &dma40_plat_data,
+	},
+	.name = "dma40",
+	.id = 0,
+	.num_resources = ARRAY_SIZE(dma40_resources),
+	.resource = dma40_resources
+};
+
+void dma40_u8500ed_fixup(void)
+{
+	dma40_plat_data.memcpy = NULL;
+	dma40_plat_data.memcpy_len = 0;
+	dma40_resources[0].start = U8500_DMA_BASE_ED;
+	dma40_resources[0].end = U8500_DMA_BASE_ED + SZ_4K - 1;
+}

arch/arm/mach-ux500/include/mach/db8500-regs.h

@@ -7,6 +7,18 @@
 #ifndef __MACH_DB8500_REGS_H
 #define __MACH_DB8500_REGS_H
 
+/* Base address and bank offsets for ESRAM */
+#define U8500_ESRAM_BASE	0x40000000
+#define U8500_ESRAM_BANK_SIZE	0x00020000
+#define U8500_ESRAM_BANK0	U8500_ESRAM_BASE
+#define U8500_ESRAM_BANK1	(U8500_ESRAM_BASE + U8500_ESRAM_BANK_SIZE)
+#define U8500_ESRAM_BANK2	(U8500_ESRAM_BANK1 + U8500_ESRAM_BANK_SIZE)
+#define U8500_ESRAM_BANK3	(U8500_ESRAM_BANK2 + U8500_ESRAM_BANK_SIZE)
+#define U8500_ESRAM_BANK4	(U8500_ESRAM_BANK3 + U8500_ESRAM_BANK_SIZE)
+/* Use bank 4 for DMA LCLA and LCPA */
+#define U8500_DMA_LCLA_BASE	U8500_ESRAM_BANK4
+#define U8500_DMA_LCPA_BASE	(U8500_ESRAM_BANK4 + 0x4000)
+
 #define U8500_PER3_BASE		0x80000000
 #define U8500_STM_BASE		0x80100000
 #define U8500_STM_REG_BASE	(U8500_STM_BASE + 0xF000)

arch/arm/mach-ux500/include/mach/devices.h

@@ -25,5 +25,8 @@ extern struct platform_device ux500_i2c3_device;
 
 extern struct platform_device u8500_i2c0_device;
 extern struct platform_device u8500_i2c4_device;
+extern struct platform_device u8500_dma40_device;
+
+void dma40_u8500ed_fixup(void);
 
 #endif

arch/arm/mach-ux500/ste-dma40-db8500.h

+/*
+ * arch/arm/mach-ux500/ste_dma40_db8500.h
+ * DB8500-SoC-specific configuration for DMA40
+ *
+ * Copyright (C) ST-Ericsson 2007-2010
+ * License terms: GNU General Public License (GPL) version 2
+ * Author: Per Friden <per.friden@stericsson.com>
+ * Author: Jonas Aaberg <jonas.aberg@stericsson.com>
+ */
+#ifndef STE_DMA40_DB8500_H
+#define STE_DMA40_DB8500_H
+
+#define STEDMA40_NR_DEV 64
+
+enum dma_src_dev_type {
+	STEDMA40_DEV_SPI0_RX = 0,
+	STEDMA40_DEV_SD_MMC0_RX = 1,
+	STEDMA40_DEV_SD_MMC1_RX = 2,
+	STEDMA40_DEV_SD_MMC2_RX = 3,
+	STEDMA40_DEV_I2C1_RX = 4,
+	STEDMA40_DEV_I2C3_RX = 5,
+	STEDMA40_DEV_I2C2_RX = 6,
+	STEDMA40_DEV_I2C4_RX = 7, /* Only on V1 */
+	STEDMA40_DEV_SSP0_RX = 8,
+	STEDMA40_DEV_SSP1_RX = 9,
+	STEDMA40_DEV_MCDE_RX = 10,
+	STEDMA40_DEV_UART2_RX = 11,
+	STEDMA40_DEV_UART1_RX = 12,
+	STEDMA40_DEV_UART0_RX = 13,
+	STEDMA40_DEV_MSP2_RX = 14,
+	STEDMA40_DEV_I2C0_RX = 15,
+	STEDMA40_DEV_USB_OTG_IEP_8 = 16,
+	STEDMA40_DEV_USB_OTG_IEP_1_9 = 17,
+	STEDMA40_DEV_USB_OTG_IEP_2_10 = 18,
+	STEDMA40_DEV_USB_OTG_IEP_3_11 = 19,
+	STEDMA40_DEV_SLIM0_CH0_RX_HSI_RX_CH0 = 20,
+	STEDMA40_DEV_SLIM0_CH1_RX_HSI_RX_CH1 = 21,
+	STEDMA40_DEV_SLIM0_CH2_RX_HSI_RX_CH2 = 22,
+	STEDMA40_DEV_SLIM0_CH3_RX_HSI_RX_CH3 = 23,
+	STEDMA40_DEV_SRC_SXA0_RX_TX = 24,
+	STEDMA40_DEV_SRC_SXA1_RX_TX = 25,
+	STEDMA40_DEV_SRC_SXA2_RX_TX = 26,
+	STEDMA40_DEV_SRC_SXA3_RX_TX = 27,
+	STEDMA40_DEV_SD_MM2_RX = 28,
+	STEDMA40_DEV_SD_MM0_RX = 29,
+	STEDMA40_DEV_MSP1_RX = 30,
+	/*
+	 * This channel is either SlimBus or MSP,
+	 * never both at the same time.
+	 */
+	STEDMA40_SLIM0_CH0_RX = 31,
+	STEDMA40_DEV_MSP0_RX = 31,
+	STEDMA40_DEV_SD_MM1_RX = 32,
+	STEDMA40_DEV_SPI2_RX = 33,
+	STEDMA40_DEV_I2C3_RX2 = 34,
+	STEDMA40_DEV_SPI1_RX = 35,
+	STEDMA40_DEV_USB_OTG_IEP_4_12 = 36,
+	STEDMA40_DEV_USB_OTG_IEP_5_13 = 37,
+	STEDMA40_DEV_USB_OTG_IEP_6_14 = 38,
+	STEDMA40_DEV_USB_OTG_IEP_7_15 = 39,
+	STEDMA40_DEV_SPI3_RX = 40,
+	STEDMA40_DEV_SD_MM3_RX = 41,
+	STEDMA40_DEV_SD_MM4_RX = 42,
+	STEDMA40_DEV_SD_MM5_RX = 43,
+	STEDMA40_DEV_SRC_SXA4_RX_TX = 44,
+	STEDMA40_DEV_SRC_SXA5_RX_TX = 45,
+	STEDMA40_DEV_SRC_SXA6_RX_TX = 46,
+	STEDMA40_DEV_SRC_SXA7_RX_TX = 47,
+	STEDMA40_DEV_CAC1_RX = 48,
+	/* RX channels 49 and 50 are unused */
+	STEDMA40_DEV_MSHC_RX = 51,
+	STEDMA40_DEV_SLIM1_CH0_RX_HSI_RX_CH4 = 52,
+	STEDMA40_DEV_SLIM1_CH1_RX_HSI_RX_CH5 = 53,
+	STEDMA40_DEV_SLIM1_CH2_RX_HSI_RX_CH6 = 54,
+	STEDMA40_DEV_SLIM1_CH3_RX_HSI_RX_CH7 = 55,
+	/* RX channels 56 thru 60 are unused */
+	STEDMA40_DEV_CAC0_RX = 61,
+	/* RX channels 62 and 63 are unused */
+};
+
+enum dma_dest_dev_type {
+	STEDMA40_DEV_SPI0_TX = 0,
+	STEDMA40_DEV_SD_MMC0_TX = 1,
+	STEDMA40_DEV_SD_MMC1_TX = 2,
+	STEDMA40_DEV_SD_MMC2_TX = 3,
+	STEDMA40_DEV_I2C1_TX = 4,
+	STEDMA40_DEV_I2C3_TX = 5,
+	STEDMA40_DEV_I2C2_TX = 6,
+	STEDMA50_DEV_I2C4_TX = 7, /* Only on V1 */
+	STEDMA40_DEV_SSP0_TX = 8,
+	STEDMA40_DEV_SSP1_TX = 9,
+	/* TX channel 10 is unused */
+	STEDMA40_DEV_UART2_TX = 11,
+	STEDMA40_DEV_UART1_TX = 12,
+	STEDMA40_DEV_UART0_TX= 13,
+	STEDMA40_DEV_MSP2_TX = 14,
+	STEDMA40_DEV_I2C0_TX = 15,
+	STEDMA40_DEV_USB_OTG_OEP_8 = 16,
+	STEDMA40_DEV_USB_OTG_OEP_1_9 = 17,
+	STEDMA40_DEV_USB_OTG_OEP_2_10= 18,
+	STEDMA40_DEV_USB_OTG_OEP_3_11 = 19,
+	STEDMA40_DEV_SLIM0_CH0_TX_HSI_TX_CH0 = 20,
+	STEDMA40_DEV_SLIM0_CH1_TX_HSI_TX_CH1 = 21,
+	STEDMA40_DEV_SLIM0_CH2_TX_HSI_TX_CH2 = 22,
+	STEDMA40_DEV_SLIM0_CH3_TX_HSI_TX_CH3 = 23,
+	STEDMA40_DEV_DST_SXA0_RX_TX = 24,
+	STEDMA40_DEV_DST_SXA1_RX_TX = 25,
+	STEDMA40_DEV_DST_SXA2_RX_TX = 26,
+	STEDMA40_DEV_DST_SXA3_RX_TX = 27,
+	STEDMA40_DEV_SD_MM2_TX = 28,
+	STEDMA40_DEV_SD_MM0_TX = 29,
+	STEDMA40_DEV_MSP1_TX = 30,
+	/*
+	 * This channel is either SlimBus or MSP,
+	 * never both at the same time.
+	 */
+	STEDMA40_SLIM0_CH0_TX = 31,
+	STEDMA40_DEV_MSP0_TX = 31,
+	STEDMA40_DEV_SD_MM1_TX = 32,
+	STEDMA40_DEV_SPI2_TX = 33,
+	/* Secondary I2C3 channel */
+	STEDMA40_DEV_I2C3_TX2 = 34,
+	STEDMA40_DEV_SPI1_TX = 35,
+	STEDMA40_DEV_USB_OTG_OEP_4_12 = 36,
+	STEDMA40_DEV_USB_OTG_OEP_5_13 = 37,
+	STEDMA40_DEV_USB_OTG_OEP_6_14 = 38,
+	STEDMA40_DEV_USB_OTG_OEP_7_15 = 39,
+	STEDMA40_DEV_SPI3_TX = 40,
+	STEDMA40_DEV_SD_MM3_TX = 41,
+	STEDMA40_DEV_SD_MM4_TX = 42,
+	STEDMA40_DEV_SD_MM5_TX = 43,
+	STEDMA40_DEV_DST_SXA4_RX_TX = 44,
+	STEDMA40_DEV_DST_SXA5_RX_TX = 45,
+	STEDMA40_DEV_DST_SXA6_RX_TX = 46,
+	STEDMA40_DEV_DST_SXA7_RX_TX = 47,
+	STEDMA40_DEV_CAC1_TX = 48,
+	STEDMA40_DEV_CAC1_TX_HAC1_TX = 49,
+	STEDMA40_DEV_HAC1_TX = 50,
+	STEDMA40_MEMXCPY_TX_0 = 51,
+	STEDMA40_DEV_SLIM1_CH0_TX_HSI_TX_CH4 = 52,
+	STEDMA40_DEV_SLIM1_CH1_TX_HSI_TX_CH5 = 53,
+	STEDMA40_DEV_SLIM1_CH2_TX_HSI_TX_CH6 = 54,
+	STEDMA40_DEV_SLIM1_CH3_TX_HSI_TX_CH7 = 55,
+	STEDMA40_MEMCPY_TX_1 = 56,
+	STEDMA40_MEMCPY_TX_2 = 57,
+	STEDMA40_MEMCPY_TX_3 = 58,
+	STEDMA40_MEMCPY_TX_4 = 59,
+	STEDMA40_MEMCPY_TX_5 = 60,
+	STEDMA40_DEV_CAC0_TX = 61,
+	STEDMA40_DEV_CAC0_TX_HAC0_TX = 62,
+	STEDMA40_DEV_HAC0_TX = 63,
+};
+
+#endif

drivers/dma/Kconfig

@@ -166,6 +166,15 @@ config TIMB_DMA
 config ARCH_HAS_ASYNC_TX_FIND_CHANNEL
 	bool
 
+config PL330_DMA
+	tristate "DMA API Driver for PL330"
+	select DMA_ENGINE
+	depends on PL330
+	help
+	  Select if your platform has one or more PL330 DMACs.
+	  You need to provide platform specific settings via
+	  platform_data for a dma-pl330 device.
+
 config DMA_ENGINE
 	bool
 

drivers/dma/Makefile

@@ -22,3 +22,4 @@ obj-$(CONFIG_COH901318) += coh901318.o coh901318_lli.o
 obj-$(CONFIG_AMCC_PPC440SPE_ADMA) += ppc4xx/
 obj-$(CONFIG_TIMB_DMA) += timb_dma.o
 obj-$(CONFIG_STE_DMA40) += ste_dma40.o ste_dma40_ll.o
+obj-$(CONFIG_PL330_DMA) += pl330.o

drivers/dma/pl330.c

+/* linux/drivers/dma/pl330.c
+ *
+ * Copyright (C) 2010 Samsung Electronics Co. Ltd.
+ *	Jaswinder Singh <jassi.brar@samsung.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/io.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/dmaengine.h>
+#include <linux/interrupt.h>
+#include <linux/amba/bus.h>
+#include <linux/amba/pl330.h>
+
+#define NR_DEFAULT_DESC	16
+
+enum desc_status {
+	/* In the DMAC pool */
+	FREE,
+	/*
+	 * Allocted to some channel during prep_xxx
+	 * Also may be sitting on the work_list.
+	 */
+	PREP,
+	/*
+	 * Sitting on the work_list and already submitted
+	 * to the PL330 core. Not more than two descriptors
+	 * of a channel can be BUSY at any time.
+	 */
+	BUSY,
+	/*
+	 * Sitting on the channel work_list but xfer done
+	 * by PL330 core
+	 */
+	DONE,
+};
+
+struct dma_pl330_chan {
+	/* Schedule desc completion */
+	struct tasklet_struct task;
+
+	/* DMA-Engine Channel */
+	struct dma_chan chan;
+
+	/* Last completed cookie */
+	dma_cookie_t completed;
+
+	/* List of to be xfered descriptors */
+	struct list_head work_list;
+
+	/* Pointer to the DMAC that manages this channel,
+	 * NULL if the channel is available to be acquired.
+	 * As the parent, this DMAC also provides descriptors
+	 * to the channel.
+	 */
+	struct dma_pl330_dmac *dmac;
+
+	/* To protect channel manipulation */
+	spinlock_t lock;
+
+	/* Token of a hardware channel thread of PL330 DMAC
+	 * NULL if the channel is available to be acquired.
+	 */
+	void *pl330_chid;
+};
+
+struct dma_pl330_dmac {
+	struct pl330_info pif;
+
+	/* DMA-Engine Device */
+	struct dma_device ddma;
+
+	/* Pool of descriptors available for the DMAC's channels */
+	struct list_head desc_pool;
+	/* To protect desc_pool manipulation */
+	spinlock_t pool_lock;
+
+	/* Peripheral channels connected to this DMAC */
+	struct dma_pl330_chan peripherals[0]; /* keep at end */
+};
+
+struct dma_pl330_desc {
+	/* To attach to a queue as child */
+	struct list_head node;
+
+	/* Descriptor for the DMA Engine API */
+	struct dma_async_tx_descriptor txd;
+
+	/* Xfer for PL330 core */
+	struct pl330_xfer px;
+
+	struct pl330_reqcfg rqcfg;
+	struct pl330_req req;
+
+	enum desc_status status;
+
+	/* The channel which currently holds this desc */
+	struct dma_pl330_chan *pchan;
+};
+
+static inline struct dma_pl330_chan *
+to_pchan(struct dma_chan *ch)
+{
+	if (!ch)
+		return NULL;
+
+	return container_of(ch, struct dma_pl330_chan, chan);
+}
+
+static inline struct dma_pl330_desc *
+to_desc(struct dma_async_tx_descriptor *tx)
+{
+	return container_of(tx, struct dma_pl330_desc, txd);
+}
+
+static inline void free_desc_list(struct list_head *list)
+{
+	struct dma_pl330_dmac *pdmac;
+	struct dma_pl330_desc *desc;
+	struct dma_pl330_chan *pch;
+	unsigned long flags;
+
+	if (list_empty(list))
+		return;
+
+	/* Finish off the work list */
+	list_for_each_entry(desc, list, node) {
+		dma_async_tx_callback callback;
+		void *param;
+
+		/* All desc in a list belong to same channel */
+		pch = desc->pchan;
+		callback = desc->txd.callback;
+		param = desc->txd.callback_param;
+
+		if (callback)
+			callback(param);
+
+		desc->pchan = NULL;
+	}
+
+	pdmac = pch->dmac;
+
+	spin_lock_irqsave(&pdmac->pool_lock, flags);
+	list_splice_tail_init(list, &pdmac->desc_pool);
+	spin_unlock_irqrestore(&pdmac->pool_lock, flags);
+}
+
+static inline void fill_queue(struct dma_pl330_chan *pch)
+{
+	struct dma_pl330_desc *desc;
+	int ret;
+
+	list_for_each_entry(desc, &pch->work_list, node) {
+
+		/* If already submitted */
+		if (desc->status == BUSY)
+			break;
+
+		ret = pl330_submit_req(pch->pl330_chid,
+						&desc->req);
+		if (!ret) {
+			desc->status = BUSY;
+			break;
+		} else if (ret == -EAGAIN) {
+			/* QFull or DMAC Dying */
+			break;
+		} else {
+			/* Unacceptable request */
+			desc->status = DONE;
+			dev_err(pch->dmac->pif.dev, "%s:%d Bad Desc(%d)\n",
+					__func__, __LINE__, desc->txd.cookie);
+			tasklet_schedule(&pch->task);
+		}
+	}
+}
+
+static void pl330_tasklet(unsigned long data)
+{
+	struct dma_pl330_chan *pch = (struct dma_pl330_chan *)data;
+	struct dma_pl330_desc *desc, *_dt;
+	unsigned long flags;
+	LIST_HEAD(list);
+
+	spin_lock_irqsave(&pch->lock, flags);
+
+	/* Pick up ripe tomatoes */
+	list_for_each_entry_safe(desc, _dt, &pch->work_list, node)
+		if (desc->status == DONE) {
+			pch->completed = desc->txd.cookie;
+			list_move_tail(&desc->node, &list);
+		}
+
+	/* Try to submit a req imm. next to the last completed cookie */
+	fill_queue(pch);
+
+	/* Make sure the PL330 Channel thread is active */
+	pl330_chan_ctrl(pch->pl330_chid, PL330_OP_START);
+
+	spin_unlock_irqrestore(&pch->lock, flags);
+
+	free_desc_list(&list);
+}
+
+static void dma_pl330_rqcb(void *token, enum pl330_op_err err)
+{
+	struct dma_pl330_desc *desc = token;
+	struct dma_pl330_chan *pch = desc->pchan;
+	unsigned long flags;
+
+	/* If desc aborted */
+	if (!pch)
+		return;
+
+	spin_lock_irqsave(&pch->lock, flags);
+
+	desc->status = DONE;
+
+	spin_unlock_irqrestore(&pch->lock, flags);
+
+	tasklet_schedule(&pch->task);
+}
+
+static int pl330_alloc_chan_resources(struct dma_chan *chan)
+{
+	struct dma_pl330_chan *pch = to_pchan(chan);
+	struct dma_pl330_dmac *pdmac = pch->dmac;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pch->lock, flags);
+
+	pch->completed = chan->cookie = 1;
+
+	pch->pl330_chid = pl330_request_channel(&pdmac->pif);
+	if (!pch->pl330_chid) {
+		spin_unlock_irqrestore(&pch->lock, flags);
+		return 0;
+	}
+
+	tasklet_init(&pch->task, pl330_tasklet, (unsigned long) pch);
+
+	spin_unlock_irqrestore(&pch->lock, flags);
+
+	return 1;
+}
+
+static int pl330_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd, unsigned long arg)
+{
+	struct dma_pl330_chan *pch = to_pchan(chan);
+	struct dma_pl330_desc *desc;
+	unsigned long flags;
+
+	/* Only supports DMA_TERMINATE_ALL */
+	if (cmd != DMA_TERMINATE_ALL)
+		return -ENXIO;
+
+	spin_lock_irqsave(&pch->lock, flags);
+
+	/* FLUSH the PL330 Channel thread */
+	pl330_chan_ctrl(pch->pl330_chid, PL330_OP_FLUSH);
+
+	/* Mark all desc done */
+	list_for_each_entry(desc, &pch->work_list, node)
+		desc->status = DONE;
+
+	spin_unlock_irqrestore(&pch->lock, flags);
+
+	pl330_tasklet((unsigned long) pch);
+
+	return 0;
+}
+
+static void pl330_free_chan_resources(struct dma_chan *chan)
+{
+	struct dma_pl330_chan *pch = to_pchan(chan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&pch->lock, flags);
+
+	tasklet_kill(&pch->task);
+
+	pl330_release_channel(pch->pl330_chid);
+	pch->pl330_chid = NULL;
+
+	spin_unlock_irqrestore(&pch->lock, flags);
+}
+
+static enum dma_status
+pl330_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
+		 struct dma_tx_state *txstate)
+{
+	struct dma_pl330_chan *pch = to_pchan(chan);
+	dma_cookie_t last_done, last_used;
+	int ret;
+
+	last_done = pch->completed;
+	last_used = chan->cookie;
+
+	ret = dma_async_is_complete(cookie, last_done, last_used);
+
+	dma_set_tx_state(txstate, last_done, last_used, 0);
+
+	return ret;
+}
+
+static void pl330_issue_pending(struct dma_chan *chan)
+{
+	pl330_tasklet((unsigned long) to_pchan(chan));
+}
+
+/*
+ * We returned the last one of the circular list of descriptor(s)
+ * from prep_xxx, so the argument to submit corresponds to the last
+ * descriptor of the list.
+ */
+static dma_cookie_t pl330_tx_submit(struct dma_async_tx_descriptor *tx)
+{
+	struct dma_pl330_desc *desc, *last = to_desc(tx);
+	struct dma_pl330_chan *pch = to_pchan(tx->chan);
+	dma_cookie_t cookie;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pch->lock, flags);
+
+	/* Assign cookies to all nodes */
+	cookie = tx->chan->cookie;
+
+	while (!list_empty(&last->node)) {
+		desc = list_entry(last->node.next, struct dma_pl330_desc, node);
+
+		if (++cookie < 0)
+			cookie = 1;
+		desc->txd.cookie = cookie;
+
+		list_move_tail(&desc->node, &pch->work_list);
+	}
+
+	if (++cookie < 0)
+		cookie = 1;
+	last->txd.cookie = cookie;
+
+	list_add_tail(&last->node, &pch->work_list);
+
+	tx->chan->cookie = cookie;
+
+	spin_unlock_irqrestore(&pch->lock, flags);
+
+	return cookie;
+}
+
+static inline void _init_desc(struct dma_pl330_desc *desc)
+{
+	desc->pchan = NULL;
+	desc->req.x = &desc->px;
+	desc->req.token = desc;
+	desc->rqcfg.swap = SWAP_NO;
+	desc->rqcfg.privileged = 0;
+	desc->rqcfg.insnaccess = 0;
+	desc->rqcfg.scctl = SCCTRL0;
+	desc->rqcfg.dcctl = DCCTRL0;
+	desc->req.cfg = &desc->rqcfg;
+	desc->req.xfer_cb = dma_pl330_rqcb;
+	desc->txd.tx_submit = pl330_tx_submit;
+
+	INIT_LIST_HEAD(&desc->node);
+}
+
+/* Returns the number of descriptors added to the DMAC pool */
+int add_desc(struct dma_pl330_dmac *pdmac, gfp_t flg, int count)
+{
+	struct dma_pl330_desc *desc;
+	unsigned long flags;
+	int i;
+
+	if (!pdmac)
+		return 0;
+
+	desc = kmalloc(count * sizeof(*desc), flg);
+	if (!desc)
+		return 0;
+
+	spin_lock_irqsave(&pdmac->pool_lock, flags);
+
+	for (i = 0; i < count; i++) {
+		_init_desc(&desc[i]);
+		list_add_tail(&desc[i].node, &pdmac->desc_pool);
+	}
+
+	spin_unlock_irqrestore(&pdmac->pool_lock, flags);
+
+	return count;
+}
+
+static struct dma_pl330_desc *
+pluck_desc(struct dma_pl330_dmac *pdmac)
+{
+	struct dma_pl330_desc *desc = NULL;
+	unsigned long flags;
+
+	if (!pdmac)
+		return NULL;
+
+	spin_lock_irqsave(&pdmac->pool_lock, flags);
+
+	if (!list_empty(&pdmac->desc_pool)) {
+		desc = list_entry(pdmac->desc_pool.next,
+				struct dma_pl330_desc, node);
+
+		list_del_init(&desc->node);
+
+		desc->status = PREP;
+		desc->txd.callback = NULL;
+	}
+
+	spin_unlock_irqrestore(&pdmac->pool_lock, flags);
+
+	return desc;
+}
+
+static struct dma_pl330_desc *pl330_get_desc(struct dma_pl330_chan *pch)
+{
+	struct dma_pl330_dmac *pdmac = pch->dmac;
+	struct dma_pl330_peri *peri = pch->chan.private;
+	struct dma_pl330_desc *desc;
+
+	/* Pluck one desc from the pool of DMAC */
+	desc = pluck_desc(pdmac);
+
+	/* If the DMAC pool is empty, alloc new */
+	if (!desc) {
+		if (!add_desc(pdmac, GFP_ATOMIC, 1))
+			return NULL;
+
+		/* Try again */
+		desc = pluck_desc(pdmac);
+		if (!desc) {
+			dev_err(pch->dmac->pif.dev,
+				"%s:%d ALERT!\n", __func__, __LINE__);
+			return NULL;
+		}
+	}
+
+	/* Initialize the descriptor */
+	desc->pchan = pch;
+	desc->txd.cookie = 0;
+	async_tx_ack(&desc->txd);
+
+	desc->req.rqtype = peri->rqtype;
+	desc->req.peri = peri->peri_id;
+
+	dma_async_tx_descriptor_init(&desc->txd, &pch->chan);
+
+	return desc;
+}
+
+static inline void fill_px(struct pl330_xfer *px,
+		dma_addr_t dst, dma_addr_t src, size_t len)
+{
+	px->next = NULL;
+	px->bytes = len;
+	px->dst_addr = dst;
+	px->src_addr = src;
+}
+
+static struct dma_pl330_desc *
+__pl330_prep_dma_memcpy(struct dma_pl330_chan *pch, dma_addr_t dst,
+		dma_addr_t src, size_t len)
+{
+	struct dma_pl330_desc *desc = pl330_get_desc(pch);
+
+	if (!desc) {
+		dev_err(pch->dmac->pif.dev, "%s:%d Unable to fetch desc\n",
+			__func__, __LINE__);
+		return NULL;
+	}
+
+	/*
+	 * Ideally we should lookout for reqs bigger than
+	 * those that can be programmed with 256 bytes of
+	 * MC buffer, but considering a req size is seldom
+	 * going to be word-unaligned and more than 200MB,
+	 * we take it easy.
+	 * Also, should the limit is reached we'd rather
+	 * have the platform increase MC buffer size than
+	 * complicating this API driver.
+	 */
+	fill_px(&desc->px, dst, src, len);
+
+	return desc;
+}
+
+/* Call after fixing burst size */
+static inline int get_burst_len(struct dma_pl330_desc *desc, size_t len)
+{
+	struct dma_pl330_chan *pch = desc->pchan;
+	struct pl330_info *pi = &pch->dmac->pif;
+	int burst_len;
+
+	burst_len = pi->pcfg.data_bus_width / 8;
+	burst_len *= pi->pcfg.data_buf_dep;
+	burst_len >>= desc->rqcfg.brst_size;
+
+	/* src/dst_burst_len can't be more than 16 */
+	if (burst_len > 16)
+		burst_len = 16;
+
+	while (burst_len > 1) {
+		if (!(len % (burst_len << desc->rqcfg.brst_size)))
+			break;
+		burst_len--;
+	}
+
+	return burst_len;
+}
+
+static struct dma_async_tx_descriptor *
+pl330_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dst,
+		dma_addr_t src, size_t len, unsigned long flags)
+{
+	struct dma_pl330_desc *desc;
+	struct dma_pl330_chan *pch = to_pchan(chan);
+	struct dma_pl330_peri *peri = chan->private;
+	struct pl330_info *pi;
+	int burst;
+
+	if (unlikely(!pch || !len || !peri))
+		return NULL;
+
+	if (peri->rqtype != MEMTOMEM)
+		return NULL;
+
+	pi = &pch->dmac->pif;
+
+	desc = __pl330_prep_dma_memcpy(pch, dst, src, len);
+	if (!desc)
+		return NULL;
+
+	desc->rqcfg.src_inc = 1;
+	desc->rqcfg.dst_inc = 1;
+
+	/* Select max possible burst size */
+	burst = pi->pcfg.data_bus_width / 8;
+
+	while (burst > 1) {
+		if (!(len % burst))
+			break;
+		burst /= 2;
+	}
+
+	desc->rqcfg.brst_size = 0;
+	while (burst != (1 << desc->rqcfg.brst_size))
+		desc->rqcfg.brst_size++;
+
+	desc->rqcfg.brst_len = get_burst_len(desc, len);
+
+	desc->txd.flags = flags;
+
+	return &desc->txd;
+}
+
+static struct dma_async_tx_descriptor *
+pl330_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
+		unsigned int sg_len, enum dma_data_direction direction,
+		unsigned long flg)
+{
+	struct dma_pl330_desc *first, *desc = NULL;
+	struct dma_pl330_chan *pch = to_pchan(chan);
+	struct dma_pl330_peri *peri = chan->private;
+	struct scatterlist *sg;
+	unsigned long flags;
+	int i, burst_size;
+	dma_addr_t addr;
+
+	if (unlikely(!pch || !sgl || !sg_len))
+		return NULL;
+
+	/* Make sure the direction is consistent */
+	if ((direction == DMA_TO_DEVICE &&
+				peri->rqtype != MEMTODEV) ||
+			(direction == DMA_FROM_DEVICE &&
+				peri->rqtype != DEVTOMEM)) {
+		dev_err(pch->dmac->pif.dev, "%s:%d Invalid Direction\n",
+				__func__, __LINE__);
+		return NULL;
+	}
+
+	addr = peri->fifo_addr;
+	burst_size = peri->burst_sz;
+
+	first = NULL;
+
+	for_each_sg(sgl, sg, sg_len, i) {
+
+		desc = pl330_get_desc(pch);
+		if (!desc) {
+			struct dma_pl330_dmac *pdmac = pch->dmac;
+
+			dev_err(pch->dmac->pif.dev,
+				"%s:%d Unable to fetch desc\n",
+				__func__, __LINE__);
+			if (!first)
+				return NULL;
+
+			spin_lock_irqsave(&pdmac->pool_lock, flags);
+
+			while (!list_empty(&first->node)) {
+				desc = list_entry(first->node.next,
+						struct dma_pl330_desc, node);
+				list_move_tail(&desc->node, &pdmac->desc_pool);
+			}
+
+			list_move_tail(&first->node, &pdmac->desc_pool);
+
+			spin_unlock_irqrestore(&pdmac->pool_lock, flags);
+
+			return NULL;
+		}
+
+		if (!first)
+			first = desc;
+		else
+			list_add_tail(&desc->node, &first->node);
+
+		if (direction == DMA_TO_DEVICE) {
+			desc->rqcfg.src_inc = 1;
+			desc->rqcfg.dst_inc = 0;
+			fill_px(&desc->px,
+				addr, sg_dma_address(sg), sg_dma_len(sg));
+		} else {
+			desc->rqcfg.src_inc = 0;
+			desc->rqcfg.dst_inc = 1;
+			fill_px(&desc->px,
+				sg_dma_address(sg), addr, sg_dma_len(sg));
+		}
+
+		desc->rqcfg.brst_size = burst_size;
+		desc->rqcfg.brst_len = 1;
+	}
+
+	/* Return the last desc in the chain */
+	desc->txd.flags = flg;
+	return &desc->txd;
+}
+
+static irqreturn_t pl330_irq_handler(int irq, void *data)
+{
+	if (pl330_update(data))
+		return IRQ_HANDLED;
+	else
+		return IRQ_NONE;
+}
+
+static int __devinit
+pl330_probe(struct amba_device *adev, struct amba_id *id)
+{
+	struct dma_pl330_platdata *pdat;
+	struct dma_pl330_dmac *pdmac;
+	struct dma_pl330_chan *pch;
+	struct pl330_info *pi;
+	struct dma_device *pd;
+	struct resource *res;
+	int i, ret, irq;
+
+	pdat = adev->dev.platform_data;
+
+	if (!pdat || !pdat->nr_valid_peri) {
+		dev_err(&adev->dev, "platform data missing\n");
+		return -ENODEV;
+	}
+
+	/* Allocate a new DMAC and its Channels */
+	pdmac = kzalloc(pdat->nr_valid_peri * sizeof(*pch)
+				+ sizeof(*pdmac), GFP_KERNEL);
+	if (!pdmac) {
+		dev_err(&adev->dev, "unable to allocate mem\n");
+		return -ENOMEM;
+	}
+
+	pi = &pdmac->pif;
+	pi->dev = &adev->dev;
+	pi->pl330_data = NULL;
+	pi->mcbufsz = pdat->mcbuf_sz;
+
+	res = &adev->res;
+	request_mem_region(res->start, resource_size(res), "dma-pl330");
+
+	pi->base = ioremap(res->start, resource_size(res));
+	if (!pi->base) {
+		ret = -ENXIO;
+		goto probe_err1;
+	}
+
+	irq = adev->irq[0];
+	ret = request_irq(irq, pl330_irq_handler, 0,
+			dev_name(&adev->dev), pi);
+	if (ret)
+		goto probe_err2;
+
+	ret = pl330_add(pi);
+	if (ret)
+		goto probe_err3;
+
+	INIT_LIST_HEAD(&pdmac->desc_pool);
+	spin_lock_init(&pdmac->pool_lock);
+
+	/* Create a descriptor pool of default size */
+	if (!add_desc(pdmac, GFP_KERNEL, NR_DEFAULT_DESC))
+		dev_warn(&adev->dev, "unable to allocate desc\n");
+
+	pd = &pdmac->ddma;
+	INIT_LIST_HEAD(&pd->channels);
+
+	/* Initialize channel parameters */
+	for (i = 0; i < pdat->nr_valid_peri; i++) {
+		struct dma_pl330_peri *peri = &pdat->peri[i];
+		pch = &pdmac->peripherals[i];
+
+		switch (peri->rqtype) {
+		case MEMTOMEM:
+			dma_cap_set(DMA_MEMCPY, pd->cap_mask);
+			break;
+		case MEMTODEV:
+		case DEVTOMEM:
+			dma_cap_set(DMA_SLAVE, pd->cap_mask);
+			break;
+		default:
+			dev_err(&adev->dev, "DEVTODEV Not Supported\n");
+			continue;
+		}
+
+		INIT_LIST_HEAD(&pch->work_list);
+		spin_lock_init(&pch->lock);
+		pch->pl330_chid = NULL;
+		pch->chan.private = peri;
+		pch->chan.device = pd;
+		pch->chan.chan_id = i;
+		pch->dmac = pdmac;
+
+		/* Add the channel to the DMAC list */
+		pd->chancnt++;
+		list_add_tail(&pch->chan.device_node, &pd->channels);
+	}
+
+	pd->dev = &adev->dev;
+
+	pd->device_alloc_chan_resources = pl330_alloc_chan_resources;
+	pd->device_free_chan_resources = pl330_free_chan_resources;
+	pd->device_prep_dma_memcpy = pl330_prep_dma_memcpy;
+	pd->device_tx_status = pl330_tx_status;
+	pd->device_prep_slave_sg = pl330_prep_slave_sg;
+	pd->device_control = pl330_control;
+	pd->device_issue_pending = pl330_issue_pending;
+
+	ret = dma_async_device_register(pd);
+	if (ret) {
+		dev_err(&adev->dev, "unable to register DMAC\n");
+		goto probe_err4;
+	}
+
+	amba_set_drvdata(adev, pdmac);
+
+	dev_info(&adev->dev,
+		"Loaded driver for PL330 DMAC-%d\n", adev->periphid);
+	dev_info(&adev->dev,
+		"\tDBUFF-%ux%ubytes Num_Chans-%u Num_Peri-%u Num_Events-%u\n",
+		pi->pcfg.data_buf_dep,
+		pi->pcfg.data_bus_width / 8, pi->pcfg.num_chan,
+		pi->pcfg.num_peri, pi->pcfg.num_events);
+
+	return 0;
+
+probe_err4:
+	pl330_del(pi);
+probe_err3:
+	free_irq(irq, pi);
+probe_err2:
+	iounmap(pi->base);
+probe_err1:
+	release_mem_region(res->start, resource_size(res));
+	kfree(pdmac);
+
+	return ret;
+}
+
+static int __devexit pl330_remove(struct amba_device *adev)
+{
+	struct dma_pl330_dmac *pdmac = amba_get_drvdata(adev);
+	struct dma_pl330_chan *pch, *_p;
+	struct pl330_info *pi;
+	struct resource *res;
+	int irq;
+
+	if (!pdmac)
+		return 0;
+
+	amba_set_drvdata(adev, NULL);
+
+	/* Idle the DMAC */
+	list_for_each_entry_safe(pch, _p, &pdmac->ddma.channels,
+			chan.device_node) {
+
+		/* Remove the channel */
+		list_del(&pch->chan.device_node);
+
+		/* Flush the channel */
+		pl330_control(&pch->chan, DMA_TERMINATE_ALL, 0);
+		pl330_free_chan_resources(&pch->chan);
+	}
+
+	pi = &pdmac->pif;
+
+	pl330_del(pi);
+
+	irq = adev->irq[0];
+	free_irq(irq, pi);
+
+	iounmap(pi->base);
+
+	res = &adev->res;
+	release_mem_region(res->start, resource_size(res));
+
+	kfree(pdmac);
+
+	return 0;
+}
+
+static struct amba_id pl330_ids[] = {
+	{
+		.id	= 0x00041330,
+		.mask	= 0x000fffff,
+	},
+	{ 0, 0 },
+};
+
+static struct amba_driver pl330_driver = {
+	.drv = {
+		.owner = THIS_MODULE,
+		.name = "dma-pl330",
+	},
+	.id_table = pl330_ids,
+	.probe = pl330_probe,
+	.remove = pl330_remove,
+};
+
+static int __init pl330_init(void)
+{
+	return amba_driver_register(&pl330_driver);
+}
+module_init(pl330_init);
+
+static void __exit pl330_exit(void)
+{
+	amba_driver_unregister(&pl330_driver);
+	return;
+}
+module_exit(pl330_exit);
+
+MODULE_AUTHOR("Jaswinder Singh <jassi.brar@samsung.com>");
+MODULE_DESCRIPTION("API Driver for PL330 DMAC");
+MODULE_LICENSE("GPL");

include/linux/amba/pl330.h

+/* linux/include/linux/amba/pl330.h
+ *
+ * Copyright (C) 2010 Samsung Electronics Co. Ltd.
+ *	Jaswinder Singh <jassi.brar@samsung.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef	__AMBA_PL330_H_
+#define	__AMBA_PL330_H_
+
+#include <asm/hardware/pl330.h>
+
+struct dma_pl330_peri {
+	/*
+	 * Peri_Req i/f of the DMAC that is
+	 * peripheral could be reached from.
+	 */
+	u8 peri_id; /* {0, 31} */
+	enum pl330_reqtype rqtype;
+
+	/* For M->D and D->M Channels */
+	int burst_sz; /* in power of 2 */
+	dma_addr_t fifo_addr;
+};
+
+struct dma_pl330_platdata {
+	/*
+	 * Number of valid peripherals connected to DMAC.
+	 * This may be different from the value read from
+	 * CR0, as the PL330 implementation might have 'holes'
+	 * in the peri list or the peri could also be reached
+	 * from another DMAC which the platform prefers.
+	 */
+	u8 nr_valid_peri;
+	/* Array of valid peripherals */
+	struct dma_pl330_peri *peri;
+	/* Bytes to allocate for MC buffer */
+	unsigned mcbuf_sz;
+};
+
+#endif	/* __AMBA_PL330_H_ */