[PATCH v5 0/4] ARM: S3C24XX: add dmaengine based dma-driver

[PATCH v5 0/4] ARM: S3C24XX: add dmaengine based dma-driver

[Heiko Stübner]

This series tries to provide a basic dmaengine driver for the s3c24xx
SoCs to subsequently retire the old one with custom API.

Since v4 an unnecessary second call to retrieve the dma cookie status was
removed.

Since v3 more smaller fixes were added, and memcpy operations now have a very
simple mechanism to try to use higher transfer widths.

Since v2 only some small fixes to the dma driver itself were added.

Since v1 three big changes happened (appart from fixing received comments):

For one the limitation of sg-lists to 1 element is gone.Secondly the
specifics of the virtual channels (target bus, handshake, etc) are not
encoded with channel ids but in the platformdata - making a later dt
binding easier, as these informations can there be gathered by the xlate
function. And lastly I also added a preliminary mechanism to handle the
specific needs for the earlier variants of the driver.

While s3c2443 and later can use any channel for any peripheral with just
marking the target-peripheral-id in a register, earlier SoCs can only use
specific channels and the target-ids also vary depending on the channel.

Therefore on newer SoCs the chansel element only needs to contain the
requested target-id, while on the older variants use it to encode both
the valid channels and the channel-specific peripheral-id.

On affected SoCs the target-id is always 3 bit wide, so we use these and
an additional 4th bit to hold the valid state. As the older SoCs all have
4 dma channels this can live happily in an u16 element.

The s3c24xx series will most likely not see any new offspring, but even
if it happens any new model will probably use the newer more flexible
variant of the dma controller, so this should not restrict anything
in the future.

Example:
SDI is valid on channels 0, 2 and 3 - with varying hw request sources.
For it the chansel field would look like

 ((BIT(3) | 1) << 3 * 4) | // channel 3, with request source 1
 ((BIT(3) | 2) << 2 * 4) | // channel 2, with request source 2
 ((BIT(3) | 2) << 0 * 4)   // channel 0, with request source 2

Heiko Stuebner (4):
  ARM: S3C24XX: number the dma clocks
  dmaengine: add driver for Samsung s3c24xx SoCs
  ARM: S3C24XX: add platform-devices for new dma driver for s3c2412 and s3c2443
  ARM: SAMSUNG: set s3c24xx_dma_filter for s3c64xx-spi0 device

 arch/arm/mach-s3c24xx/clock-s3c2412.c     |    8 +-
 arch/arm/mach-s3c24xx/common-s3c2443.c    |   12 +-
 arch/arm/mach-s3c24xx/common.c            |  106 +++
 arch/arm/mach-s3c24xx/common.h            |    3 +
 arch/arm/mach-s3c24xx/mach-jive.c         |    1 +
 arch/arm/mach-s3c24xx/mach-smdk2413.c     |    1 +
 arch/arm/mach-s3c24xx/mach-smdk2416.c     |    1 +
 arch/arm/mach-s3c24xx/mach-smdk2443.c     |    1 +
 arch/arm/mach-s3c24xx/mach-vstms.c        |    1 +
 arch/arm/plat-samsung/devs.c              |    5 +-
 drivers/dma/Kconfig                       |   12 +
 drivers/dma/Makefile                      |    1 +
 drivers/dma/s3c24xx-dma.c                 | 1340 +++++++++++++++++++++++++++++
 include/linux/platform_data/dma-s3c24xx.h |   43 +
 14 files changed, 1524 insertions(+), 11 deletions(-)
 create mode 100644 drivers/dma/s3c24xx-dma.c
 create mode 100644 include/linux/platform_data/dma-s3c24xx.h

-- 
1.7.10.4

[PATCH v5 1/4] ARM: S3C24XX: number the dma clocks

[Heiko Stübner]

Each dma channel has its own clock. The upcoming dma driver wants to
handle these itself and therefore needs to be able to get the correct
clock for a channel.

Therefore rename the dma clocks to "dma.X" for s3c2412, s3c2416 and
s3c2443. This does not change the behaviour for the old dma driver at
all, as all dma clocks are still always on and not handled by the old
dma driver at all.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
---
changes since v2:
- remove noise from commit message

 arch/arm/mach-s3c24xx/clock-s3c2412.c  |    8 ++++----
 arch/arm/mach-s3c24xx/common-s3c2443.c |   12 ++++++------
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/arch/arm/mach-s3c24xx/clock-s3c2412.c b/arch/arm/mach-s3c24xx/clock-s3c2412.c
index d8f253f..11b3b28 100644
--- a/arch/arm/mach-s3c24xx/clock-s3c2412.c
+++ b/arch/arm/mach-s3c24xx/clock-s3c2412.c
@@ -484,22 +484,22 @@ static struct clk init_clocks_disable[] = {
 
 static struct clk init_clocks[] = {
 	{
-		.name		= "dma",
+		.name		= "dma.0",
 		.parent		= &clk_h,
 		.enable		= s3c2412_clkcon_enable,
 		.ctrlbit	= S3C2412_CLKCON_DMA0,
 	}, {
-		.name		= "dma",
+		.name		= "dma.1",
 		.parent		= &clk_h,
 		.enable		= s3c2412_clkcon_enable,
 		.ctrlbit	= S3C2412_CLKCON_DMA1,
 	}, {
-		.name		= "dma",
+		.name		= "dma.2",
 		.parent		= &clk_h,
 		.enable		= s3c2412_clkcon_enable,
 		.ctrlbit	= S3C2412_CLKCON_DMA2,
 	}, {
-		.name		= "dma",
+		.name		= "dma.3",
 		.parent		= &clk_h,
 		.enable		= s3c2412_clkcon_enable,
 		.ctrlbit	= S3C2412_CLKCON_DMA3,
diff --git a/arch/arm/mach-s3c24xx/common-s3c2443.c b/arch/arm/mach-s3c24xx/common-s3c2443.c
index f6b9f2e..65d3eef 100644
--- a/arch/arm/mach-s3c24xx/common-s3c2443.c
+++ b/arch/arm/mach-s3c24xx/common-s3c2443.c
@@ -438,32 +438,32 @@ static struct clk init_clocks_off[] = {
 
 static struct clk init_clocks[] = {
 	{
-		.name		= "dma",
+		.name		= "dma.0",
 		.parent		= &clk_h,
 		.enable		= s3c2443_clkcon_enable_h,
 		.ctrlbit	= S3C2443_HCLKCON_DMA0,
 	}, {
-		.name		= "dma",
+		.name		= "dma.1",
 		.parent		= &clk_h,
 		.enable		= s3c2443_clkcon_enable_h,
 		.ctrlbit	= S3C2443_HCLKCON_DMA1,
 	}, {
-		.name		= "dma",
+		.name		= "dma.2",
 		.parent		= &clk_h,
 		.enable		= s3c2443_clkcon_enable_h,
 		.ctrlbit	= S3C2443_HCLKCON_DMA2,
 	}, {
-		.name		= "dma",
+		.name		= "dma.3",
 		.parent		= &clk_h,
 		.enable		= s3c2443_clkcon_enable_h,
 		.ctrlbit	= S3C2443_HCLKCON_DMA3,
 	}, {
-		.name		= "dma",
+		.name		= "dma.4",
 		.parent		= &clk_h,
 		.enable		= s3c2443_clkcon_enable_h,
 		.ctrlbit	= S3C2443_HCLKCON_DMA4,
 	}, {
-		.name		= "dma",
+		.name		= "dma.5",
 		.parent		= &clk_h,
 		.enable		= s3c2443_clkcon_enable_h,
 		.ctrlbit	= S3C2443_HCLKCON_DMA5,
-- 
1.7.10.4

[PATCH v5 2/4] dmaengine: add driver for Samsung s3c24xx SoCs

[Heiko Stübner]

This adds a new driver to support the s3c24xx dma using the dmaengine
and makes the old one in mach-s3c24xx obsolete in the long run.

Conceptually the s3c24xx-dma feels like a distant relative of the pl08x
with numerous virtual channels being mapped to a lot less physical ones.
The driver therefore borrows a lot from the amba-pl08x driver in this
regard. Functionality-wise the driver gains a memcpy ability in addition
to the slave_sg one.

The driver supports both the method for requesting the peripheral used
by SoCs before the S3C2443 and the different method for S3C2443 and later.

On earlier SoCs the hardware channels usable for specific peripherals is
constrainted while on later SoCs all channels can be used for any peripheral.

Tested on a s3c2416-based board, memcpy using the dmatest module and
slave_sg partially using the spi-s3c64xx driver.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
---
changes since v4:
- fix the unnecessary double retrieval of the cookie status
  pointed out by Vinod Koul
changes since v3:
- address comments from Tomasz Figa:
  - fixed a lot of nitpics
  - and added a rudimentary mechanism to select higher bus-widths for memcpy
- address more comments from Vinod Koul - reorder dma_control contents

changes since v2:
- add missing channel validation in s3c24xx_dma_phy_free when repurposing
  a physical channel for another virtual channel
- address comments from Vinod Koul:
  - fix naming of constants to prevent conflicts
  - better handle the unlikely case of an interrupt happening on an
    unused channel

changes since v1:
- address comments from Linus Walleij
- support sg-lists with more than 1 element
- add support for earlier s3c24xx SoCs
- redo the channeldata definitions:
  - no need for static ids for the virtual channels
  - provide a way to encode constraints for earlier variants
  - better base for later dt bindings

 drivers/dma/Kconfig                       |   12 +
 drivers/dma/Makefile                      |    1 +
 drivers/dma/s3c24xx-dma.c                 | 1340 +++++++++++++++++++++++++++++
 include/linux/platform_data/dma-s3c24xx.h |   43 +
 4 files changed, 1396 insertions(+)
 create mode 100644 drivers/dma/s3c24xx-dma.c
 create mode 100644 include/linux/platform_data/dma-s3c24xx.h

diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
index 526ec77..d639115 100644
--- a/drivers/dma/Kconfig
+++ b/drivers/dma/Kconfig
@@ -154,6 +154,18 @@ config TEGRA20_APB_DMA
 	  This DMA controller transfers data from memory to peripheral fifo
 	  or vice versa. It does not support memory to memory data transfer.
 
+config S3C24XX_DMAC
+	tristate "Samsung S3C24XX DMA support"
+	depends on ARCH_S3C24XX && !S3C24XX_DMA
+	select DMA_ENGINE
+	select DMA_VIRTUAL_CHANNELS
+	help
+	  Support for the Samsung S3C24XX DMA controller driver. The
+	  DMA controller is having multiple DMA channels which can be
+	  configured for different peripherals like audio, UART, SPI.
+	  The DMA controller can transfer data from memory to peripheral,
+	  periphal to memory, periphal to periphal and memory to memory.
+
 source "drivers/dma/sh/Kconfig"
 
 config COH901318
diff --git a/drivers/dma/Makefile b/drivers/dma/Makefile
index db89035..0ce2da9 100644
--- a/drivers/dma/Makefile
+++ b/drivers/dma/Makefile
@@ -30,6 +30,7 @@ obj-$(CONFIG_SIRF_DMA) += sirf-dma.o
 obj-$(CONFIG_TI_EDMA) += edma.o
 obj-$(CONFIG_STE_DMA40) += ste_dma40.o ste_dma40_ll.o
 obj-$(CONFIG_TEGRA20_APB_DMA) += tegra20-apb-dma.o
+obj-$(CONFIG_S3C24XX_DMAC) += s3c24xx-dma.o
 obj-$(CONFIG_PL330_DMA) += pl330.o
 obj-$(CONFIG_PCH_DMA) += pch_dma.o
 obj-$(CONFIG_AMBA_PL08X) += amba-pl08x.o
diff --git a/drivers/dma/s3c24xx-dma.c b/drivers/dma/s3c24xx-dma.c
new file mode 100644
index 0000000..56c9253
--- /dev/null
+++ b/drivers/dma/s3c24xx-dma.c
@@ -0,0 +1,1340 @@
+/*
+ * S3C24XX DMA handling
+ *
+ * Copyright (c) 2013 Heiko Stuebner <heiko@sntech.de>
+ *
+ * based on amba-pl08x.c
+ *
+ * Copyright (c) 2006 ARM Ltd.
+ * Copyright (c) 2010 ST-Ericsson SA
+ *
+ * Author: Peter Pearse <peter.pearse@arm.com>
+ * Author: Linus Walleij <linus.walleij@stericsson.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.
+ *
+ * The DMA controllers in S3C24XX SoCs have a varying number of DMA signals
+ * that can be routed to any of the 4 to 8 hardware-channels.
+ *
+ * Therefore on these DMA controllers the number of channels
+ * and the number of incoming DMA signals are two totally different things.
+ * It is usually not possible to theoretically handle all physical signals,
+ * so a multiplexing scheme with possible denial of use is necessary.
+ *
+ * Open items:
+ * - bursts
+ */
+
+#include <linux/platform_device.h>
+#include <linux/types.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/clk.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/platform_data/dma-s3c24xx.h>
+
+#include "dmaengine.h"
+#include "virt-dma.h"
+
+#define MAX_DMA_CHANNELS	8
+
+#define S3C24XX_DISRC			0x00
+#define S3C24XX_DISRCC			0x04
+#define S3C24XX_DISRCC_INC_INCREMENT	0
+#define S3C24XX_DISRCC_INC_FIXED	BIT(0)
+#define S3C24XX_DISRCC_LOC_AHB		0
+#define S3C24XX_DISRCC_LOC_APB		BIT(1)
+
+#define S3C24XX_DIDST			0x08
+#define S3C24XX_DIDSTC			0x0c
+#define S3C24XX_DIDSTC_INC_INCREMENT	0
+#define S3C24XX_DIDSTC_INC_FIXED	BIT(0)
+#define S3C24XX_DIDSTC_LOC_AHB		0
+#define S3C24XX_DIDSTC_LOC_APB		BIT(1)
+#define S3C24XX_DIDSTC_INT_TC0		0
+#define S3C24XX_DIDSTC_INT_RELOAD	BIT(2)
+
+#define S3C24XX_DCON			0x10
+
+#define S3C24XX_DCON_TC_MASK		0xfffff
+#define S3C24XX_DCON_DSZ_BYTE		(0 << 20)
+#define S3C24XX_DCON_DSZ_HALFWORD	(1 << 20)
+#define S3C24XX_DCON_DSZ_WORD		(2 << 20)
+#define S3C24XX_DCON_DSZ_MASK		(3 << 20)
+#define S3C24XX_DCON_DSZ_SHIFT		20
+#define S3C24XX_DCON_AUTORELOAD		0
+#define S3C24XX_DCON_NORELOAD		BIT(22)
+#define S3C24XX_DCON_HWTRIG		BIT(23)
+#define S3C24XX_DCON_HWSRC_SHIFT	24
+#define S3C24XX_DCON_SERV_SINGLE	0
+#define S3C24XX_DCON_SERV_WHOLE		BIT(27)
+#define S3C24XX_DCON_TSZ_UNIT		0
+#define S3C24XX_DCON_TSZ_BURST4		BIT(28)
+#define S3C24XX_DCON_INT		BIT(29)
+#define S3C24XX_DCON_SYNC_PCLK		0
+#define S3C24XX_DCON_SYNC_HCLK		BIT(30)
+#define S3C24XX_DCON_DEMAND		0
+#define S3C24XX_DCON_HANDSHAKE		BIT(31)
+
+#define S3C24XX_DSTAT			0x14
+#define S3C24XX_DSTAT_STAT_BUSY		BIT(20)
+#define S3C24XX_DSTAT_CURRTC_MASK	0xfffff
+
+#define S3C24XX_DMASKTRIG		0x20
+#define S3C24XX_DMASKTRIG_SWTRIG	BIT(0)
+#define S3C24XX_DMASKTRIG_ON		BIT(1)
+#define S3C24XX_DMASKTRIG_STOP		BIT(2)
+
+#define S3C24XX_DMAREQSEL		0x24
+#define S3C24XX_DMAREQSEL_HW		BIT(0)
+
+/*
+ * S3C2410, S3C2440 and S3C2442 SoCs cannot select any physical channel
+ * for a DMA source. Instead only specific channels are valid.
+ * All of these SoCs have 4 physical channels and the number of request
+ * source bits is 3. Additionally we also need 1 bit to mark the channel
+ * as valid.
+ * Therefore we separate the chansel element of the channel data into 4
+ * parts of 4 bits each, to hold the information if the channel is valid
+ * and the hw request source to use.
+ *
+ * Example:
+ * SDI is valid on channels 0, 2 and 3 - with varying hw request sources.
+ * For it the chansel field would look like
+ *
+ * ((BIT(3) | 1) << 3 * 4) | // channel 3, with request source 1
+ * ((BIT(3) | 2) << 2 * 4) | // channel 2, with request source 2
+ * ((BIT(3) | 2) << 0 * 4)   // channel 0, with request source 2
+ */
+#define S3C24XX_CHANSEL_WIDTH		4
+#define S3C24XX_CHANSEL_VALID		BIT(3)
+#define S3C24XX_CHANSEL_REQ_MASK	7
+
+/*
+ * struct soc_data - vendor-specific config parameters for individual SoCs
+ * @stride: spacing between the registers of each channel
+ * @has_reqsel: does the controller use the newer requestselection mechanism
+ * @has_clocks: are controllable dma-clocks present
+ */
+struct soc_data {
+	int stride;
+	bool has_reqsel;
+	bool has_clocks;
+};
+
+/*
+ * enum s3c24xx_dma_chan_state - holds the virtual channel states
+ * @S3C24XX_DMA_CHAN_IDLE: the channel is idle
+ * @S3C24XX_DMA_CHAN_RUNNING: the channel has allocated a physical transport
+ * channel and is running a transfer on it
+ * @S3C24XX_DMA_CHAN_WAITING: the channel is waiting for a physical transport
+ * channel to become available (only pertains to memcpy channels)
+ */
+enum s3c24xx_dma_chan_state {
+	S3C24XX_DMA_CHAN_IDLE,
+	S3C24XX_DMA_CHAN_RUNNING,
+	S3C24XX_DMA_CHAN_WAITING,
+};
+
+/*
+ * struct s3c24xx_sg - structure containing data per sg
+ * @src_addr: src address of sg
+ * @dst_addr: dst address of sg
+ * @len: transfer len in bytes
+ * @node: node for txd's dsg_list
+ */
+struct s3c24xx_sg {
+	dma_addr_t src_addr;
+	dma_addr_t dst_addr;
+	size_t len;
+	struct list_head node;
+};
+
+/*
+ * struct s3c24xx_txd - wrapper for struct dma_async_tx_descriptor
+ * @vd: virtual DMA descriptor
+ * @dsg_list: list of children sg's
+ * @at: sg currently being transfered
+ * @width: transfer width
+ * @disrcc: value for source control register
+ * @didstc: value for destination control register
+ * @dcon: base value for dcon register
+ */
+struct s3c24xx_txd {
+	struct virt_dma_desc vd;
+	struct list_head dsg_list;
+	struct list_head *at;
+	u8 width;
+	u32 disrcc;
+	u32 didstc;
+	u32 dcon;
+};
+
+struct s3c24xx_dma_chan;
+
+/*
+ * struct s3c24xx_dma_phy - holder for the physical channels
+ * @id: physical index to this channel
+ * @valid: does the channel have all required elements
+ * @base: virtual memory base (remapped) for the this channel
+ * @irq: interrupt for this channel
+ * @clk: clock for this channel
+ * @lock: a lock to use when altering an instance of this struct
+ * @serving: virtual channel currently being served by this physicalchannel
+ * @host: a pointer to the host (internal use)
+ */
+struct s3c24xx_dma_phy {
+	unsigned int			id;
+	bool				valid;
+	void __iomem			*base;
+	unsigned int			irq;
+	struct clk			*clk;
+	spinlock_t			lock;
+	struct s3c24xx_dma_chan		*serving;
+	struct s3c24xx_dma_engine	*host;
+};
+
+/*
+ * struct s3c24xx_dma_chan - this structure wraps a DMA ENGINE channel
+ * @id: the id of the channel
+ * @name: name of the channel
+ * @vc: wrappped virtual channel
+ * @phy: the physical channel utilized by this channel, if there is one
+ * @runtime_addr: address for RX/TX according to the runtime config
+ * @at: active transaction on this channel
+ * @lock: a lock for this channel data
+ * @host: a pointer to the host (internal use)
+ * @state: whether the channel is idle, running etc
+ * @slave: whether this channel is a device (slave) or for memcpy
+ */
+struct s3c24xx_dma_chan {
+	int id;
+	const char *name;
+	struct virt_dma_chan vc;
+	struct s3c24xx_dma_phy *phy;
+	struct dma_slave_config cfg;
+	struct s3c24xx_txd *at;
+	struct s3c24xx_dma_engine *host;
+	enum s3c24xx_dma_chan_state state;
+	bool slave;
+};
+
+/*
+ * struct s3c24xx_dma_engine - the local state holder for the S3C24XX
+ * @pdev: the corresponding platform device
+ * @pdata: platform data passed in from the platform/machine
+ * @base: virtual memory base (remapped)
+ * @slave: slave engine for this instance
+ * @memcpy: memcpy engine for this instance
+ * @phy_chans: array of data for the physical channels
+ */
+struct s3c24xx_dma_engine {
+	struct platform_device			*pdev;
+	const struct s3c24xx_dma_platdata	*pdata;
+	struct soc_data				*sdata;
+	void __iomem				*base;
+	struct dma_device			slave;
+	struct dma_device			memcpy;
+	struct s3c24xx_dma_phy			*phy_chans;
+};
+
+/*
+ * Physical channel handling
+ */
+
+/*
+ * Check whether a certain channel is busy or not.
+ */
+static int s3c24xx_dma_phy_busy(struct s3c24xx_dma_phy *phy)
+{
+	unsigned int val = readl(phy->base + S3C24XX_DSTAT);
+	return val & S3C24XX_DSTAT_STAT_BUSY;
+}
+
+static bool s3c24xx_dma_phy_valid(struct s3c24xx_dma_chan *s3cchan,
+				  struct s3c24xx_dma_phy *phy)
+{
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
+	struct s3c24xx_dma_channel *cdata = &pdata->channels[s3cchan->id];
+	int phyvalid;
+
+	/* every phy is valid for memcopy channels */
+	if (!s3cchan->slave)
+		return true;
+
+	/* On newer variants all phys can be used for all virtual channels */
+	if (s3cdma->sdata->has_reqsel)
+		return true;
+
+	phyvalid = (cdata->chansel >> (phy->id * S3C24XX_CHANSEL_WIDTH));
+	return (phyvalid & S3C24XX_CHANSEL_VALID) ? true : false;
+}
+
+/*
+ * Allocate a physical channel for a virtual channel
+ *
+ * Try to locate a physical channel to be used for this transfer. If all
+ * are taken return NULL and the requester will have to cope by using
+ * some fallback PIO mode or retrying later.
+ */
+static
+struct s3c24xx_dma_phy *s3c24xx_dma_get_phy(struct s3c24xx_dma_chan *s3cchan)
+{
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
+	struct s3c24xx_dma_channel *cdata;
+	struct s3c24xx_dma_phy *phy = NULL;
+	unsigned long flags;
+	int i;
+	int ret;
+
+	if (s3cchan->slave)
+		cdata = &pdata->channels[s3cchan->id];
+
+	for (i = 0; i < s3cdma->pdata->num_phy_channels; i++) {
+		phy = &s3cdma->phy_chans[i];
+
+		if (!phy->valid)
+			continue;
+
+		if (!s3c24xx_dma_phy_valid(s3cchan, phy))
+			continue;
+
+		spin_lock_irqsave(&phy->lock, flags);
+
+		if (!phy->serving) {
+			phy->serving = s3cchan;
+			spin_unlock_irqrestore(&phy->lock, flags);
+			break;
+		}
+
+		spin_unlock_irqrestore(&phy->lock, flags);
+	}
+
+	/* No physical channel available, cope with it */
+	if (i == s3cdma->pdata->num_phy_channels) {
+		dev_warn(&s3cdma->pdev->dev, "no phy channel available\n");
+		return NULL;
+	}
+
+	/* start the phy clock */
+	if (s3cdma->sdata->has_clocks) {
+		ret = clk_enable(phy->clk);
+		if (ret) {
+			dev_err(&s3cdma->pdev->dev, "could not enable clock for channel %d, err %d\n",
+				phy->id, ret);
+			phy->serving = NULL;
+			return NULL;
+		}
+	}
+
+	return phy;
+}
+
+/*
+ * Mark the physical channel as free.
+ *
+ * This drops the link between the physical and virtual channel.
+ */
+static inline void s3c24xx_dma_put_phy(struct s3c24xx_dma_phy *phy)
+{
+	struct s3c24xx_dma_engine *s3cdma = phy->host;
+
+	if (s3cdma->sdata->has_clocks)
+		clk_disable(phy->clk);
+
+	phy->serving = NULL;
+}
+
+/*
+ * Stops the channel by writing the stop bit.
+ * This should not be used for an on-going transfer, but as a method of
+ * shutting down a channel (eg, when it's no longer used) or terminating a
+ * transfer.
+ */
+static void s3c24xx_dma_terminate_phy(struct s3c24xx_dma_phy *phy)
+{
+	writel(S3C24XX_DMASKTRIG_STOP, phy->base + S3C24XX_DMASKTRIG);
+}
+
+/*
+ * Virtual channel handling
+ */
+
+static inline
+struct s3c24xx_dma_chan *to_s3c24xx_dma_chan(struct dma_chan *chan)
+{
+	return container_of(chan, struct s3c24xx_dma_chan, vc.chan);
+}
+
+static u32 s3c24xx_dma_getbytes_chan(struct s3c24xx_dma_chan *s3cchan)
+{
+	struct s3c24xx_dma_phy *phy = s3cchan->phy;
+	struct s3c24xx_txd *txd = s3cchan->at;
+	u32 tc = readl(phy->base + S3C24XX_DSTAT) & S3C24XX_DSTAT_CURRTC_MASK;
+
+	return tc * txd->width;
+}
+
+static int s3c24xx_dma_set_runtime_config(struct s3c24xx_dma_chan *s3cchan,
+				  struct dma_slave_config *config)
+{
+	if (!s3cchan->slave)
+		return -EINVAL;
+
+	/* Reject definitely invalid configurations */
+	if (config->src_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES ||
+	    config->dst_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES)
+		return -EINVAL;
+
+	s3cchan->cfg = *config;
+
+	return 0;
+}
+
+/*
+ * Transfer handling
+ */
+
+static inline
+struct s3c24xx_txd *to_s3c24xx_txd(struct dma_async_tx_descriptor *tx)
+{
+	return container_of(tx, struct s3c24xx_txd, vd.tx);
+}
+
+static struct s3c24xx_txd *s3c24xx_dma_get_txd(void)
+{
+	struct s3c24xx_txd *txd = kzalloc(sizeof(*txd), GFP_NOWAIT);
+
+	if (txd) {
+		INIT_LIST_HEAD(&txd->dsg_list);
+		txd->dcon = S3C24XX_DCON_INT | S3C24XX_DCON_NORELOAD;
+	}
+
+	return txd;
+}
+
+static void s3c24xx_dma_free_txd(struct s3c24xx_txd *txd)
+{
+	struct s3c24xx_sg *dsg, *_dsg;
+
+	list_for_each_entry_safe(dsg, _dsg, &txd->dsg_list, node) {
+		list_del(&dsg->node);
+		kfree(dsg);
+	}
+
+	kfree(txd);
+}
+
+static void s3c24xx_dma_start_next_sg(struct s3c24xx_dma_chan *s3cchan,
+				       struct s3c24xx_txd *txd)
+{
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+	struct s3c24xx_dma_phy *phy = s3cchan->phy;
+	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
+	struct s3c24xx_sg *dsg = list_entry(txd->at, struct s3c24xx_sg, node);
+	u32 dcon = txd->dcon;
+	u32 val;
+
+	/* transfer-size and -count from len and width */
+	switch (txd->width) {
+	case 1:
+		dcon |= S3C24XX_DCON_DSZ_BYTE | dsg->len;
+		break;
+	case 2:
+		dcon |= S3C24XX_DCON_DSZ_HALFWORD | (dsg->len / 2);
+		break;
+	case 4:
+		dcon |= S3C24XX_DCON_DSZ_WORD | (dsg->len / 4);
+		break;
+	}
+
+	if (s3cchan->slave) {
+		struct s3c24xx_dma_channel *cdata =
+					&pdata->channels[s3cchan->id];
+
+		if (s3cdma->sdata->has_reqsel) {
+			writel_relaxed((cdata->chansel << 1) |
+							S3C24XX_DMAREQSEL_HW,
+					phy->base + S3C24XX_DMAREQSEL);
+		} else {
+			int csel = cdata->chansel >> (phy->id *
+							S3C24XX_CHANSEL_WIDTH);
+
+			csel &= S3C24XX_CHANSEL_REQ_MASK;
+			dcon |= csel << S3C24XX_DCON_HWSRC_SHIFT;
+			dcon |= S3C24XX_DCON_HWTRIG;
+		}
+	} else {
+		if (s3cdma->sdata->has_reqsel)
+			writel_relaxed(0, phy->base + S3C24XX_DMAREQSEL);
+	}
+
+	writel_relaxed(dsg->src_addr, phy->base + S3C24XX_DISRC);
+	writel_relaxed(txd->disrcc, phy->base + S3C24XX_DISRCC);
+	writel_relaxed(dsg->dst_addr, phy->base + S3C24XX_DIDST);
+	writel_relaxed(txd->didstc, phy->base + S3C24XX_DIDSTC);
+	writel_relaxed(dcon, phy->base + S3C24XX_DCON);
+
+	val = readl_relaxed(phy->base + S3C24XX_DMASKTRIG);
+	val &= ~S3C24XX_DMASKTRIG_STOP;
+	val |= S3C24XX_DMASKTRIG_ON;
+
+	/* trigger the dma operation for memcpy transfers */
+	if (!s3cchan->slave)
+		val |= S3C24XX_DMASKTRIG_SWTRIG;
+
+	writel(val, phy->base + S3C24XX_DMASKTRIG);
+}
+
+/*
+ * Set the initial DMA register values and start first sg.
+ */
+static void s3c24xx_dma_start_next_txd(struct s3c24xx_dma_chan *s3cchan)
+{
+	struct s3c24xx_dma_phy *phy = s3cchan->phy;
+	struct virt_dma_desc *vd = vchan_next_desc(&s3cchan->vc);
+	struct s3c24xx_txd *txd = to_s3c24xx_txd(&vd->tx);
+
+	list_del(&txd->vd.node);
+
+	s3cchan->at = txd;
+
+	/* Wait for channel inactive */
+	while (s3c24xx_dma_phy_busy(phy))
+		cpu_relax();
+
+	/* point to the first element of the sg list */
+	txd->at = txd->dsg_list.next;
+	s3c24xx_dma_start_next_sg(s3cchan, txd);
+}
+
+static void s3c24xx_dma_free_txd_list(struct s3c24xx_dma_engine *s3cdma,
+				struct s3c24xx_dma_chan *s3cchan)
+{
+	LIST_HEAD(head);
+
+	vchan_get_all_descriptors(&s3cchan->vc, &head);
+	vchan_dma_desc_free_list(&s3cchan->vc, &head);
+}
+
+/*
+ * Try to allocate a physical channel.  When successful, assign it to
+ * this virtual channel, and initiate the next descriptor.  The
+ * virtual channel lock must be held at this point.
+ */
+static void s3c24xx_dma_phy_alloc_and_start(struct s3c24xx_dma_chan *s3cchan)
+{
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+	struct s3c24xx_dma_phy *phy;
+
+	phy = s3c24xx_dma_get_phy(s3cchan);
+	if (!phy) {
+		dev_dbg(&s3cdma->pdev->dev, "no physical channel available for xfer on %s\n",
+			s3cchan->name);
+		s3cchan->state = S3C24XX_DMA_CHAN_WAITING;
+		return;
+	}
+
+	dev_dbg(&s3cdma->pdev->dev, "allocated physical channel %d for xfer on %s\n",
+		phy->id, s3cchan->name);
+
+	s3cchan->phy = phy;
+	s3cchan->state = S3C24XX_DMA_CHAN_RUNNING;
+
+	s3c24xx_dma_start_next_txd(s3cchan);
+}
+
+static void s3c24xx_dma_phy_reassign_start(struct s3c24xx_dma_phy *phy,
+	struct s3c24xx_dma_chan *s3cchan)
+{
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+
+	dev_dbg(&s3cdma->pdev->dev, "reassigned physical channel %d for xfer on %s\n",
+		phy->id, s3cchan->name);
+
+	/*
+	 * We do this without taking the lock; we're really only concerned
+	 * about whether this pointer is NULL or not, and we're guaranteed
+	 * that this will only be called when it _already_ is non-NULL.
+	 */
+	phy->serving = s3cchan;
+	s3cchan->phy = phy;
+	s3cchan->state = S3C24XX_DMA_CHAN_RUNNING;
+	s3c24xx_dma_start_next_txd(s3cchan);
+}
+
+/*
+ * Free a physical DMA channel, potentially reallocating it to another
+ * virtual channel if we have any pending.
+ */
+static void s3c24xx_dma_phy_free(struct s3c24xx_dma_chan *s3cchan)
+{
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+	struct s3c24xx_dma_chan *p, *next;
+
+retry:
+	next = NULL;
+
+	/* Find a waiting virtual channel for the next transfer. */
+	list_for_each_entry(p, &s3cdma->memcpy.channels, vc.chan.device_node)
+		if (p->state == S3C24XX_DMA_CHAN_WAITING) {
+			next = p;
+			break;
+		}
+
+	if (!next) {
+		list_for_each_entry(p, &s3cdma->slave.channels,
+				    vc.chan.device_node)
+			if (p->state == S3C24XX_DMA_CHAN_WAITING &&
+				      s3c24xx_dma_phy_valid(p, s3cchan->phy)) {
+				next = p;
+				break;
+			}
+	}
+
+	/* Ensure that the physical channel is stopped */
+	s3c24xx_dma_terminate_phy(s3cchan->phy);
+
+	if (next) {
+		bool success;
+
+		/*
+		 * Eww.  We know this isn't going to deadlock
+		 * but lockdep probably doesn't.
+		 */
+		spin_lock(&next->vc.lock);
+		/* Re-check the state now that we have the lock */
+		success = next->state == S3C24XX_DMA_CHAN_WAITING;
+		if (success)
+			s3c24xx_dma_phy_reassign_start(s3cchan->phy, next);
+		spin_unlock(&next->vc.lock);
+
+		/* If the state changed, try to find another channel */
+		if (!success)
+			goto retry;
+	} else {
+		/* No more jobs, so free up the physical channel */
+		s3c24xx_dma_put_phy(s3cchan->phy);
+	}
+
+	s3cchan->phy = NULL;
+	s3cchan->state = S3C24XX_DMA_CHAN_IDLE;
+}
+
+static void s3c24xx_dma_unmap_buffers(struct s3c24xx_txd *txd)
+{
+	struct device *dev = txd->vd.tx.chan->device->dev;
+	struct s3c24xx_sg *dsg;
+
+	if (!(txd->vd.tx.flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
+		if (txd->vd.tx.flags & DMA_COMPL_SRC_UNMAP_SINGLE)
+			list_for_each_entry(dsg, &txd->dsg_list, node)
+				dma_unmap_single(dev, dsg->src_addr, dsg->len,
+						DMA_TO_DEVICE);
+		else {
+			list_for_each_entry(dsg, &txd->dsg_list, node)
+				dma_unmap_page(dev, dsg->src_addr, dsg->len,
+						DMA_TO_DEVICE);
+		}
+	}
+
+	if (!(txd->vd.tx.flags & DMA_COMPL_SKIP_DEST_UNMAP)) {
+		if (txd->vd.tx.flags & DMA_COMPL_DEST_UNMAP_SINGLE)
+			list_for_each_entry(dsg, &txd->dsg_list, node)
+				dma_unmap_single(dev, dsg->dst_addr, dsg->len,
+						DMA_FROM_DEVICE);
+		else
+			list_for_each_entry(dsg, &txd->dsg_list, node)
+				dma_unmap_page(dev, dsg->dst_addr, dsg->len,
+						DMA_FROM_DEVICE);
+	}
+}
+
+static void s3c24xx_dma_desc_free(struct virt_dma_desc *vd)
+{
+	struct s3c24xx_txd *txd = to_s3c24xx_txd(&vd->tx);
+	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(vd->tx.chan);
+
+	if (!s3cchan->slave)
+		s3c24xx_dma_unmap_buffers(txd);
+
+	s3c24xx_dma_free_txd(txd);
+}
+
+static irqreturn_t s3c24xx_dma_irq(int irq, void *data)
+{
+	struct s3c24xx_dma_phy *phy = data;
+	struct s3c24xx_dma_chan *s3cchan = phy->serving;
+	struct s3c24xx_txd *txd;
+
+	dev_dbg(&phy->host->pdev->dev, "interrupt on channel %d\n", phy->id);
+
+	/*
+	 * Interrupts happen to notify the completion of a transfer and the
+	 * channel should have moved into its stop state already on its own.
+	 * Therefore interrupts on channels not bound to a virtual channel
+	 * should never happen. Nevertheless send a terminate command to the
+	 * channel if the unlikely case happens.
+	 */
+	if (unlikely(!s3cchan)) {
+		dev_err(&phy->host->pdev->dev, "interrupt on unused channel %d\n",
+			phy->id);
+
+		s3c24xx_dma_terminate_phy(phy);
+
+		return IRQ_HANDLED;
+	}
+
+	spin_lock(&s3cchan->vc.lock);
+	txd = s3cchan->at;
+	if (txd) {
+		/* when more sg's are in this txd, start the next one */
+		if (!list_is_last(txd->at, &txd->dsg_list)) {
+			txd->at = txd->at->next;
+			s3c24xx_dma_start_next_sg(s3cchan, txd);
+		} else {
+			s3cchan->at = NULL;
+			vchan_cookie_complete(&txd->vd);
+
+			/*
+			 * And start the next descriptor (if any),
+			 * otherwise free this channel.
+			 */
+			if (vchan_next_desc(&s3cchan->vc))
+				s3c24xx_dma_start_next_txd(s3cchan);
+			else
+				s3c24xx_dma_phy_free(s3cchan);
+		}
+	}
+	spin_unlock(&s3cchan->vc.lock);
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * The DMA ENGINE API
+ */
+
+static int s3c24xx_dma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
+			 unsigned long arg)
+{
+	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+	unsigned long flags;
+	int ret = 0;
+
+	spin_lock_irqsave(&s3cchan->vc.lock, flags);
+
+	switch (cmd) {
+	case DMA_SLAVE_CONFIG:
+		ret = s3c24xx_dma_set_runtime_config(s3cchan,
+					      (struct dma_slave_config *)arg);
+		break;
+	case DMA_TERMINATE_ALL:
+		if (!s3cchan->phy && !s3cchan->at) {
+			dev_err(&s3cdma->pdev->dev, "trying to terminate already stopped channel %d\n",
+				s3cchan->id);
+			ret = -EINVAL;
+			break;
+		}
+
+		s3cchan->state = S3C24XX_DMA_CHAN_IDLE;
+
+		 /* Mark physical channel as free */
+		if (s3cchan->phy)
+			s3c24xx_dma_phy_free(s3cchan);
+
+		/* Dequeue current job */
+		if (s3cchan->at) {
+			s3c24xx_dma_desc_free(&s3cchan->at->vd);
+			s3cchan->at = NULL;
+		}
+
+		/* Dequeue jobs not yet fired as well */
+		s3c24xx_dma_free_txd_list(s3cdma, s3cchan);
+		break;
+	default:
+		/* Unknown command */
+		ret = -ENXIO;
+		break;
+	}
+
+	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
+
+	return ret;
+}
+
+static int s3c24xx_dma_alloc_chan_resources(struct dma_chan *chan)
+{
+	return 0;
+}
+
+static void s3c24xx_dma_free_chan_resources(struct dma_chan *chan)
+{
+	/* Ensure all queued descriptors are freed */
+	vchan_free_chan_resources(to_virt_chan(chan));
+}
+
+static enum dma_status s3c24xx_dma_tx_status(struct dma_chan *chan,
+		dma_cookie_t cookie, struct dma_tx_state *txstate)
+{
+	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
+	struct s3c24xx_txd *txd;
+	struct s3c24xx_sg *dsg;
+	struct virt_dma_desc *vd;
+	unsigned long flags;
+	enum dma_status ret;
+	size_t bytes = 0;
+
+	spin_lock_irqsave(&s3cchan->vc.lock, flags);
+	ret = dma_cookie_status(chan, cookie, txstate);
+	if (ret == DMA_SUCCESS) {
+		spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
+		return ret;
+	}
+
+	/*
+	 * There's no point calculating the residue if there's
+	 * no txstate to store the value.
+	 */
+	if (!txstate) {
+		spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
+		return ret;
+	}
+
+	vd = vchan_find_desc(&s3cchan->vc, cookie);
+	if (vd) {
+		/* On the issued list, so hasn't been processed yet */
+		txd = to_s3c24xx_txd(&vd->tx);
+
+		list_for_each_entry(dsg, &txd->dsg_list, node)
+			bytes += dsg->len;
+	} else {
+		/*
+		 * Currently running, so sum over the pending sg's and
+		 * the currently active one.
+		 */
+		txd = s3cchan->at;
+
+		dsg = list_entry(txd->at, struct s3c24xx_sg, node);
+		list_for_each_entry_from(dsg, &txd->dsg_list, node)
+			bytes += dsg->len;
+
+		bytes += s3c24xx_dma_getbytes_chan(s3cchan);
+	}
+	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
+
+	/*
+	 * This cookie not complete yet
+	 * Get number of bytes left in the active transactions and queue
+	 */
+	dma_set_residue(txstate, bytes);
+
+	/* Whether waiting or running, we're in progress */
+	return ret;
+}
+
+/*
+ * Initialize a descriptor to be used by memcpy submit
+ */
+static struct dma_async_tx_descriptor *s3c24xx_dma_prep_memcpy(
+		struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
+		size_t len, unsigned long flags)
+{
+	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+	struct s3c24xx_txd *txd;
+	struct s3c24xx_sg *dsg;
+	int src_mod, dest_mod;
+
+	dev_dbg(&s3cdma->pdev->dev, "prepare memcpy of %d bytes from %s\n",
+			len, s3cchan->name);
+
+	if ((len & S3C24XX_DCON_TC_MASK) != len) {
+		dev_err(&s3cdma->pdev->dev, "memcpy size %d to large\n", len);
+		return NULL;
+	}
+
+	txd = s3c24xx_dma_get_txd();
+	if (!txd)
+		return NULL;
+
+	dsg = kzalloc(sizeof(*dsg), GFP_NOWAIT);
+	if (!dsg) {
+		s3c24xx_dma_free_txd(txd);
+		return NULL;
+	}
+	list_add_tail(&dsg->node, &txd->dsg_list);
+
+	dsg->src_addr = src;
+	dsg->dst_addr = dest;
+	dsg->len = len;
+
+	/*
+	 * Determine a suitable transfer width.
+	 * The DMA controller cannot fetch/store information which is not
+	 * naturally aligned on the bus, i.e., a 4 byte fetch must start at
+	 * an address divisible by 4 - more generally addr % width must be 0.
+	 */
+	src_mod = src % 4;
+	dest_mod = dest % 4;
+	switch (len % 4) {
+	case 0:
+		txd->width = (src_mod == 0 && dest_mod == 0) ? 4 : 1;
+		break;
+	case 2:
+		txd->width = ((src_mod == 2 || src_mod == 0) &&
+			      (dest_mod == 2 || dest_mod == 0)) ? 2 : 1;
+		break;
+	default:
+		txd->width = 1;
+		break;
+	}
+
+	txd->disrcc = S3C24XX_DISRCC_LOC_AHB | S3C24XX_DISRCC_INC_INCREMENT;
+	txd->didstc = S3C24XX_DIDSTC_LOC_AHB | S3C24XX_DIDSTC_INC_INCREMENT;
+	txd->dcon |= S3C24XX_DCON_DEMAND | S3C24XX_DCON_SYNC_HCLK |
+		     S3C24XX_DCON_SERV_WHOLE;
+
+	return vchan_tx_prep(&s3cchan->vc, &txd->vd, flags);
+}
+
+static struct dma_async_tx_descriptor *s3c24xx_dma_prep_slave_sg(
+		struct dma_chan *chan, struct scatterlist *sgl,
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
+{
+	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
+	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
+	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
+	struct s3c24xx_dma_channel *cdata = &pdata->channels[s3cchan->id];
+	struct s3c24xx_txd *txd;
+	struct s3c24xx_sg *dsg;
+	struct scatterlist *sg;
+	dma_addr_t slave_addr;
+	u32 hwcfg = 0;
+	int tmp;
+
+	dev_dbg(&s3cdma->pdev->dev, "prepare transaction of %d bytes from %s\n",
+			sg_dma_len(sgl), s3cchan->name);
+
+	txd = s3c24xx_dma_get_txd();
+	if (!txd)
+		return NULL;
+
+	if (cdata->handshake)
+		txd->dcon |= S3C24XX_DCON_HANDSHAKE;
+
+	switch (cdata->bus) {
+	case S3C24XX_DMA_APB:
+		txd->dcon |= S3C24XX_DCON_SYNC_PCLK;
+		hwcfg |= S3C24XX_DISRCC_LOC_APB;
+		break;
+	case S3C24XX_DMA_AHB:
+		txd->dcon |= S3C24XX_DCON_SYNC_HCLK;
+		hwcfg |= S3C24XX_DISRCC_LOC_AHB;
+		break;
+	}
+
+	/*
+	 * Always assume our peripheral desintation is a fixed
+	 * address in memory.
+	 */
+	hwcfg |= S3C24XX_DISRCC_INC_FIXED;
+
+	/*
+	 * Individual dma operations are requested by the slave,
+	 * so serve only single atomic operations (S3C24XX_DCON_SERV_SINGLE).
+	 */
+	txd->dcon |= S3C24XX_DCON_SERV_SINGLE;
+
+	if (direction == DMA_MEM_TO_DEV) {
+		txd->disrcc = S3C24XX_DISRCC_LOC_AHB |
+			      S3C24XX_DISRCC_INC_INCREMENT;
+		txd->didstc = hwcfg;
+		slave_addr = s3cchan->cfg.dst_addr;
+		txd->width = s3cchan->cfg.dst_addr_width;
+	} else if (direction == DMA_DEV_TO_MEM) {
+		txd->disrcc = hwcfg;
+		txd->didstc = S3C24XX_DIDSTC_LOC_AHB |
+			      S3C24XX_DIDSTC_INC_INCREMENT;
+		slave_addr = s3cchan->cfg.src_addr;
+		txd->width = s3cchan->cfg.src_addr_width;
+	} else {
+		s3c24xx_dma_free_txd(txd);
+		dev_err(&s3cdma->pdev->dev,
+			"direction %d unsupported\n", direction);
+		return NULL;
+	}
+
+	for_each_sg(sgl, sg, sg_len, tmp) {
+		dsg = kzalloc(sizeof(*dsg), GFP_NOWAIT);
+		if (!dsg) {
+			s3c24xx_dma_free_txd(txd);
+			return NULL;
+		}
+		list_add_tail(&dsg->node, &txd->dsg_list);
+
+		dsg->len = sg_dma_len(sg);
+		if (direction == DMA_MEM_TO_DEV) {
+			dsg->src_addr = sg_dma_address(sg);
+			dsg->dst_addr = slave_addr;
+		} else { /* DMA_DEV_TO_MEM */
+			dsg->src_addr = slave_addr;
+			dsg->dst_addr = sg_dma_address(sg);
+		}
+		break;
+	}
+
+	return vchan_tx_prep(&s3cchan->vc, &txd->vd, flags);
+}
+
+/*
+ * Slave transactions callback to the slave device to allow
+ * synchronization of slave DMA signals with the DMAC enable
+ */
+static void s3c24xx_dma_issue_pending(struct dma_chan *chan)
+{
+	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&s3cchan->vc.lock, flags);
+	if (vchan_issue_pending(&s3cchan->vc)) {
+		if (!s3cchan->phy && s3cchan->state != S3C24XX_DMA_CHAN_WAITING)
+			s3c24xx_dma_phy_alloc_and_start(s3cchan);
+	}
+	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
+}
+
+/*
+ * Bringup and teardown
+ */
+
+/*
+ * Initialise the DMAC memcpy/slave channels.
+ * Make a local wrapper to hold required data
+ */
+static int s3c24xx_dma_init_virtual_channels(struct s3c24xx_dma_engine *s3cdma,
+		struct dma_device *dmadev, unsigned int channels, bool slave)
+{
+	struct s3c24xx_dma_chan *chan;
+	int i;
+
+	INIT_LIST_HEAD(&dmadev->channels);
+
+	/*
+	 * Register as many many memcpy as we have physical channels,
+	 * we won't always be able to use all but the code will have
+	 * to cope with that situation.
+	 */
+	for (i = 0; i < channels; i++) {
+		chan = devm_kzalloc(dmadev->dev, sizeof(*chan), GFP_KERNEL);
+		if (!chan) {
+			dev_err(dmadev->dev,
+				"%s no memory for channel\n", __func__);
+			return -ENOMEM;
+		}
+
+		chan->id = i;
+		chan->host = s3cdma;
+		chan->state = S3C24XX_DMA_CHAN_IDLE;
+
+		if (slave) {
+			chan->slave = true;
+			chan->name = kasprintf(GFP_KERNEL, "slave%d", i);
+			if (!chan->name)
+				return -ENOMEM;
+		} else {
+			chan->name = kasprintf(GFP_KERNEL, "memcpy%d", i);
+			if (!chan->name)
+				return -ENOMEM;
+		}
+		dev_dbg(dmadev->dev,
+			 "initialize virtual channel \"%s\"\n",
+			 chan->name);
+
+		chan->vc.desc_free = s3c24xx_dma_desc_free;
+		vchan_init(&chan->vc, dmadev);
+	}
+	dev_info(dmadev->dev, "initialized %d virtual %s channels\n",
+		 i, slave ? "slave" : "memcpy");
+	return i;
+}
+
+static void s3c24xx_dma_free_virtual_channels(struct dma_device *dmadev)
+{
+	struct s3c24xx_dma_chan *chan = NULL;
+	struct s3c24xx_dma_chan *next;
+
+	list_for_each_entry_safe(chan,
+				 next, &dmadev->channels, vc.chan.device_node)
+		list_del(&chan->vc.chan.device_node);
+}
+
+/* s3c2412 and s3c2413 have a 0x40 stride and dmareqsel mechanism */
+static struct soc_data soc_s3c2412 = {
+	.stride = 0x40,
+	.has_reqsel = true,
+	.has_clocks = true,
+};
+
+/* s3c2443 and following have a 0x100 stride and dmareqsel mechanism */
+static struct soc_data soc_s3c2443 = {
+	.stride = 0x100,
+	.has_reqsel = true,
+	.has_clocks = true,
+};
+
+static struct platform_device_id s3c24xx_dma_driver_ids[] = {
+	{
+		.name		= "s3c2412-dma",
+		.driver_data	= (kernel_ulong_t)&soc_s3c2412,
+	}, {
+		.name		= "s3c2443-dma",
+		.driver_data	= (kernel_ulong_t)&soc_s3c2443,
+	},
+	{ },
+};
+
+static struct soc_data *s3c24xx_dma_get_soc_data(struct platform_device *pdev)
+{
+	return (struct soc_data *)
+			 platform_get_device_id(pdev)->driver_data;
+}
+
+static int s3c24xx_dma_probe(struct platform_device *pdev)
+{
+	const struct s3c24xx_dma_platdata *pdata = dev_get_platdata(&pdev->dev);
+	struct s3c24xx_dma_engine *s3cdma;
+	struct soc_data *sdata;
+	struct resource *res;
+	int ret;
+	int i;
+
+	if (!pdata) {
+		dev_err(&pdev->dev, "platform data missing\n");
+		return -ENODEV;
+	}
+
+	/* Basic sanity check */
+	if (pdata->num_phy_channels > MAX_DMA_CHANNELS) {
+		dev_err(&pdev->dev, "to many dma channels %d, max %d\n",
+			pdata->num_phy_channels, MAX_DMA_CHANNELS);
+		return -EINVAL;
+	}
+
+	sdata = s3c24xx_dma_get_soc_data(pdev);
+	if (!sdata)
+		return -EINVAL;
+
+	s3cdma = devm_kzalloc(&pdev->dev, sizeof(*s3cdma), GFP_KERNEL);
+	if (!s3cdma)
+		return -ENOMEM;
+
+	s3cdma->pdev = pdev;
+	s3cdma->pdata = pdata;
+	s3cdma->sdata = sdata;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	s3cdma->base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(s3cdma->base))
+		return PTR_ERR(s3cdma->base);
+
+	s3cdma->phy_chans = devm_kzalloc(&pdev->dev,
+					      sizeof(struct s3c24xx_dma_phy) *
+							pdata->num_phy_channels,
+					      GFP_KERNEL);
+	if (!s3cdma->phy_chans)
+		return -ENOMEM;
+
+	/* aquire irqs and clocks for all physical channels */
+	for (i = 0; i < pdata->num_phy_channels; i++) {
+		struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
+		char clk_name[6];
+
+		phy->id = i;
+		phy->base = s3cdma->base + (i * sdata->stride);
+		phy->host = s3cdma;
+
+		phy->irq = platform_get_irq(pdev, i);
+		if (phy->irq < 0) {
+			dev_err(&pdev->dev, "failed to get irq %d, err %d\n",
+				i, phy->irq);
+			continue;
+		}
+
+		ret = devm_request_irq(&pdev->dev, phy->irq, s3c24xx_dma_irq,
+				       0, pdev->name, phy);
+		if (ret) {
+			dev_err(&pdev->dev, "Unable to request irq for channel %d, error %d\n",
+				i, ret);
+			continue;
+		}
+
+		if (sdata->has_clocks) {
+			sprintf(clk_name, "dma.%d", i);
+			phy->clk = devm_clk_get(&pdev->dev, clk_name);
+			if (IS_ERR(phy->clk) && sdata->has_clocks) {
+				dev_err(&pdev->dev, "unable to aquire clock for channel %d, error %lu",
+					i, PTR_ERR(phy->clk));
+				continue;
+			}
+
+			ret = clk_prepare(phy->clk);
+			if (ret) {
+				dev_err(&pdev->dev, "clock for phy %d failed, error %d\n",
+					i, ret);
+				continue;
+			}
+		}
+
+		spin_lock_init(&phy->lock);
+		phy->valid = true;
+
+		dev_dbg(&pdev->dev, "physical channel %d is %s\n",
+			i, s3c24xx_dma_phy_busy(phy) ? "BUSY" : "FREE");
+	}
+
+	/* Initialize memcpy engine */
+	dma_cap_set(DMA_MEMCPY, s3cdma->memcpy.cap_mask);
+	dma_cap_set(DMA_PRIVATE, s3cdma->memcpy.cap_mask);
+	s3cdma->memcpy.dev = &pdev->dev;
+	s3cdma->memcpy.device_alloc_chan_resources =
+					s3c24xx_dma_alloc_chan_resources;
+	s3cdma->memcpy.device_free_chan_resources =
+					s3c24xx_dma_free_chan_resources;
+	s3cdma->memcpy.device_prep_dma_memcpy = s3c24xx_dma_prep_memcpy;
+	s3cdma->memcpy.device_tx_status = s3c24xx_dma_tx_status;
+	s3cdma->memcpy.device_issue_pending = s3c24xx_dma_issue_pending;
+	s3cdma->memcpy.device_control = s3c24xx_dma_control;
+
+	/* Initialize slave engine for SoC internal dedicated peripherals */
+	dma_cap_set(DMA_SLAVE, s3cdma->slave.cap_mask);
+	dma_cap_set(DMA_PRIVATE, s3cdma->slave.cap_mask);
+	s3cdma->slave.dev = &pdev->dev;
+	s3cdma->slave.device_alloc_chan_resources =
+					s3c24xx_dma_alloc_chan_resources;
+	s3cdma->slave.device_free_chan_resources =
+					s3c24xx_dma_free_chan_resources;
+	s3cdma->slave.device_tx_status = s3c24xx_dma_tx_status;
+	s3cdma->slave.device_issue_pending = s3c24xx_dma_issue_pending;
+	s3cdma->slave.device_prep_slave_sg = s3c24xx_dma_prep_slave_sg;
+	s3cdma->slave.device_control = s3c24xx_dma_control;
+
+	/* Register as many memcpy channels as there are physical channels */
+	ret = s3c24xx_dma_init_virtual_channels(s3cdma, &s3cdma->memcpy,
+						pdata->num_phy_channels, false);
+	if (ret <= 0) {
+		dev_warn(&pdev->dev,
+			 "%s failed to enumerate memcpy channels - %d\n",
+			 __func__, ret);
+		goto err_memcpy;
+	}
+
+	/* Register slave channels */
+	ret = s3c24xx_dma_init_virtual_channels(s3cdma, &s3cdma->slave,
+				pdata->num_channels, true);
+	if (ret <= 0) {
+		dev_warn(&pdev->dev,
+			"%s failed to enumerate slave channels - %d\n",
+				__func__, ret);
+		goto err_slave;
+	}
+
+	ret = dma_async_device_register(&s3cdma->memcpy);
+	if (ret) {
+		dev_warn(&pdev->dev,
+			"%s failed to register memcpy as an async device - %d\n",
+			__func__, ret);
+		goto err_memcpy_reg;
+	}
+
+	ret = dma_async_device_register(&s3cdma->slave);
+	if (ret) {
+		dev_warn(&pdev->dev,
+			"%s failed to register slave as an async device - %d\n",
+			__func__, ret);
+		goto err_slave_reg;
+	}
+
+	platform_set_drvdata(pdev, s3cdma);
+	dev_info(&pdev->dev, "Loaded dma driver with %d physical channels\n",
+		 pdata->num_phy_channels);
+
+	return 0;
+
+err_slave_reg:
+	dma_async_device_unregister(&s3cdma->memcpy);
+err_memcpy_reg:
+	s3c24xx_dma_free_virtual_channels(&s3cdma->slave);
+err_slave:
+	s3c24xx_dma_free_virtual_channels(&s3cdma->memcpy);
+err_memcpy:
+	if (sdata->has_clocks)
+		for (i = 0; i < pdata->num_phy_channels; i++) {
+			struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
+			if (phy->valid)
+				clk_unprepare(phy->clk);
+		}
+
+	return ret;
+}
+
+static int s3c24xx_dma_remove(struct platform_device *pdev)
+{
+	const struct s3c24xx_dma_platdata *pdata = dev_get_platdata(&pdev->dev);
+	struct s3c24xx_dma_engine *s3cdma = platform_get_drvdata(pdev);
+	struct soc_data *sdata = s3c24xx_dma_get_soc_data(pdev);
+	int i;
+
+	dma_async_device_unregister(&s3cdma->slave);
+	dma_async_device_unregister(&s3cdma->memcpy);
+
+	s3c24xx_dma_free_virtual_channels(&s3cdma->slave);
+	s3c24xx_dma_free_virtual_channels(&s3cdma->memcpy);
+
+	if (sdata->has_clocks)
+		for (i = 0; i < pdata->num_phy_channels; i++) {
+			struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
+			if (phy->valid)
+				clk_unprepare(phy->clk);
+		}
+
+	return 0;
+}
+
+static struct platform_driver s3c24xx_dma_driver = {
+	.driver		= {
+		.name	= "s3c24xx-dma",
+		.owner	= THIS_MODULE,
+	},
+	.id_table	= s3c24xx_dma_driver_ids,
+	.probe		= s3c24xx_dma_probe,
+	.remove		= s3c24xx_dma_remove,
+};
+
+module_platform_driver(s3c24xx_dma_driver);
+
+bool s3c24xx_dma_filter(struct dma_chan *chan, void *param)
+{
+	struct s3c24xx_dma_chan *s3cchan;
+
+	if (chan->device->dev->driver != &s3c24xx_dma_driver.driver)
+		return false;
+
+	s3cchan = to_s3c24xx_dma_chan(chan);
+
+	return s3cchan->id == (int)param;
+}
+EXPORT_SYMBOL(s3c24xx_dma_filter);
+
+MODULE_DESCRIPTION("S3C24XX DMA Driver");
+MODULE_AUTHOR("Heiko Stuebner");
+MODULE_LICENSE("GPL v2");
diff --git a/include/linux/platform_data/dma-s3c24xx.h b/include/linux/platform_data/dma-s3c24xx.h
new file mode 100644
index 0000000..5a0cfff
--- /dev/null
+++ b/include/linux/platform_data/dma-s3c24xx.h
@@ -0,0 +1,43 @@
+/*
+ * S3C24XX DMA handling
+ *
+ * Copyright (c) 2013 Heiko Stuebner <heiko@sntech.de>
+ *
+ * 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.
+ */
+
+enum s3c24xx_dma_bus {
+	S3C24XX_DMA_APB,
+	S3C24XX_DMA_AHB,
+};
+
+/**
+ * @bus: on which bus does the peripheral reside - AHB or APB.
+ * @handshake: is a handshake with the peripheral necessary
+ * @chansel: channel selection information, depending on variant; reqsel for
+ *	     s3c2443 and later and channel-selection map for earlier SoCs
+ *	     see CHANSEL doc in s3c2443-dma.c
+ */
+struct s3c24xx_dma_channel {
+	enum s3c24xx_dma_bus bus;
+	bool handshake;
+	u16 chansel;
+};
+
+/**
+ * struct s3c24xx_dma_platdata - platform specific settings
+ * @num_phy_channels: number of physical channels
+ * @channels: array of virtual channel descriptions
+ * @num_channels: number of virtual channels
+ */
+struct s3c24xx_dma_platdata {
+	int num_phy_channels;
+	struct s3c24xx_dma_channel *channels;
+	int num_channels;
+};
+
+struct dma_chan;
+bool s3c24xx_dma_filter(struct dma_chan *chan, void *param);
-- 
1.7.10.4

[PATCH v5 3/4] ARM: S3C24XX: add platform-devices for new dma driver for s3c2412 and s3c2443

[Heiko Stübner]

This includes defining the mapping for the request sources.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
---
changes since v1:
- follow new pdata definition

 arch/arm/mach-s3c24xx/common.c        |  106 +++++++++++++++++++++++++++++++++
 arch/arm/mach-s3c24xx/common.h        |    3 +
 arch/arm/mach-s3c24xx/mach-jive.c     |    1 +
 arch/arm/mach-s3c24xx/mach-smdk2413.c |    1 +
 arch/arm/mach-s3c24xx/mach-smdk2416.c |    1 +
 arch/arm/mach-s3c24xx/mach-smdk2443.c |    1 +
 arch/arm/mach-s3c24xx/mach-vstms.c    |    1 +
 7 files changed, 114 insertions(+)

diff --git a/arch/arm/mach-s3c24xx/common.c b/arch/arm/mach-s3c24xx/common.c
index 457261c..16ac669 100644
--- a/arch/arm/mach-s3c24xx/common.c
+++ b/arch/arm/mach-s3c24xx/common.c
@@ -31,6 +31,7 @@
 #include <linux/platform_device.h>
 #include <linux/delay.h>
 #include <linux/io.h>
+#include <linux/platform_data/dma-s3c24xx.h>
 
 #include <mach/hardware.h>
 #include <mach/regs-clock.h>
@@ -44,6 +45,7 @@
 
 #include <mach/regs-gpio.h>
 #include <plat/regs-serial.h>
+#include <mach/dma.h>
 
 #include <plat/cpu.h>
 #include <plat/devs.h>
@@ -329,3 +331,107 @@ void __init_or_cpufreq s3c24xx_setup_clocks(unsigned long fclk,
 	clk_p.rate = pclk;
 	clk_f.rate = fclk;
 }
+
+#if defined(CONFIG_CPU_S3C2410) || defined(CONFIG_CPU_S3C2412) || \
+    defined(CONFIG_CPU_S3C2440) || defined(CONFIG_CPU_S3C2442)
+static struct resource s3c2410_dma_resource[] = {
+	[0] = DEFINE_RES_MEM(S3C24XX_PA_DMA, S3C24XX_SZ_DMA),
+	[1] = DEFINE_RES_IRQ(IRQ_DMA0),
+	[2] = DEFINE_RES_IRQ(IRQ_DMA1),
+	[3] = DEFINE_RES_IRQ(IRQ_DMA2),
+	[4] = DEFINE_RES_IRQ(IRQ_DMA3),
+};
+#endif
+
+#ifdef CONFIG_CPU_S3C2412
+static struct s3c24xx_dma_channel s3c2412_dma_channels[DMACH_MAX] = {
+	[DMACH_XD0] = { S3C24XX_DMA_AHB, true, 17 },
+	[DMACH_XD1] = { S3C24XX_DMA_AHB, true, 18 },
+	[DMACH_SDI] = { S3C24XX_DMA_APB, false, 10 },
+	[DMACH_SPI0_RX] = { S3C24XX_DMA_APB, true, 1 },
+	[DMACH_SPI0_TX] = { S3C24XX_DMA_APB, true, 0 },
+	[DMACH_SPI1_RX] = { S3C24XX_DMA_APB, true, 3 },
+	[DMACH_SPI1_TX] = { S3C24XX_DMA_APB, true, 2 },
+	[DMACH_UART0] = { S3C24XX_DMA_APB, true, 19 },
+	[DMACH_UART1] = { S3C24XX_DMA_APB, true, 21 },
+	[DMACH_UART2] = { S3C24XX_DMA_APB, true, 23 },
+	[DMACH_UART0_SRC2] = { S3C24XX_DMA_APB, true, 20 },
+	[DMACH_UART1_SRC2] = { S3C24XX_DMA_APB, true, 22 },
+	[DMACH_UART2_SRC2] = { S3C24XX_DMA_APB, true, 24 },
+	[DMACH_TIMER] = { S3C24XX_DMA_APB, true, 9 },
+	[DMACH_I2S_IN] = { S3C24XX_DMA_APB, true, 5 },
+	[DMACH_I2S_OUT] = { S3C24XX_DMA_APB, true, 4 },
+	[DMACH_USB_EP1] = { S3C24XX_DMA_APB, true, 13 },
+	[DMACH_USB_EP2] = { S3C24XX_DMA_APB, true, 14 },
+	[DMACH_USB_EP3] = { S3C24XX_DMA_APB, true, 15 },
+	[DMACH_USB_EP4] = { S3C24XX_DMA_APB, true, 16 },
+};
+
+static struct s3c24xx_dma_platdata s3c2412_dma_platdata = {
+	.num_phy_channels = 4,
+	.channels = s3c2412_dma_channels,
+	.num_channels = DMACH_MAX,
+};
+
+struct platform_device s3c2412_device_dma = {
+	.name		= "s3c2412-dma",
+	.id		= 0,
+	.num_resources	= ARRAY_SIZE(s3c2410_dma_resource),
+	.resource	= s3c2410_dma_resource,
+	.dev	= {
+		.platform_data	= &s3c2412_dma_platdata,
+	},
+};
+#endif
+
+#if defined(CONFIG_CPUS_3C2443) || defined(CONFIG_CPU_S3C2416)
+static struct resource s3c2443_dma_resource[] = {
+	[0] = DEFINE_RES_MEM(S3C24XX_PA_DMA, S3C24XX_SZ_DMA),
+	[1] = DEFINE_RES_IRQ(IRQ_S3C2443_DMA0),
+	[2] = DEFINE_RES_IRQ(IRQ_S3C2443_DMA1),
+	[3] = DEFINE_RES_IRQ(IRQ_S3C2443_DMA2),
+	[4] = DEFINE_RES_IRQ(IRQ_S3C2443_DMA3),
+	[5] = DEFINE_RES_IRQ(IRQ_S3C2443_DMA4),
+	[6] = DEFINE_RES_IRQ(IRQ_S3C2443_DMA5),
+};
+
+static struct s3c24xx_dma_channel s3c2443_dma_channels[DMACH_MAX] = {
+	[DMACH_XD0] = { S3C24XX_DMA_AHB, true, 17 },
+	[DMACH_XD1] = { S3C24XX_DMA_AHB, true, 18 },
+	[DMACH_SDI] = { S3C24XX_DMA_APB, false, 10 },
+	[DMACH_SPI0_RX] = { S3C24XX_DMA_APB, true, 1 },
+	[DMACH_SPI0_TX] = { S3C24XX_DMA_APB, true, 0 },
+	[DMACH_SPI1_RX] = { S3C24XX_DMA_APB, true, 3 },
+	[DMACH_SPI1_TX] = { S3C24XX_DMA_APB, true, 2 },
+	[DMACH_UART0] = { S3C24XX_DMA_APB, true, 19 },
+	[DMACH_UART1] = { S3C24XX_DMA_APB, true, 21 },
+	[DMACH_UART2] = { S3C24XX_DMA_APB, true, 23 },
+	[DMACH_UART3] = { S3C24XX_DMA_APB, true, 25 },
+	[DMACH_UART0_SRC2] = { S3C24XX_DMA_APB, true, 20 },
+	[DMACH_UART1_SRC2] = { S3C24XX_DMA_APB, true, 22 },
+	[DMACH_UART2_SRC2] = { S3C24XX_DMA_APB, true, 24 },
+	[DMACH_UART3_SRC2] = { S3C24XX_DMA_APB, true, 26 },
+	[DMACH_TIMER] = { S3C24XX_DMA_APB, true, 9 },
+	[DMACH_I2S_IN] = { S3C24XX_DMA_APB, true, 5 },
+	[DMACH_I2S_OUT] = { S3C24XX_DMA_APB, true, 4 },
+	[DMACH_PCM_IN] = { S3C24XX_DMA_APB, true, 28 },
+	[DMACH_PCM_OUT] = { S3C24XX_DMA_APB, true, 27 },
+	[DMACH_MIC_IN] = { S3C24XX_DMA_APB, true, 29 },
+};
+
+static struct s3c24xx_dma_platdata s3c2443_dma_platdata = {
+	.num_phy_channels = 6,
+	.channels = s3c2443_dma_channels,
+	.num_channels = DMACH_MAX,
+};
+
+struct platform_device s3c2443_device_dma = {
+	.name		= "s3c2443-dma",
+	.id		= 0,
+	.num_resources	= ARRAY_SIZE(s3c2443_dma_resource),
+	.resource	= s3c2443_dma_resource,
+	.dev	= {
+		.platform_data	= &s3c2443_dma_platdata,
+	},
+};
+#endif
diff --git a/arch/arm/mach-s3c24xx/common.h b/arch/arm/mach-s3c24xx/common.h
index 84b2806..fe07189 100644
--- a/arch/arm/mach-s3c24xx/common.h
+++ b/arch/arm/mach-s3c24xx/common.h
@@ -109,4 +109,7 @@ extern void s3c2443_init_irq(void);
 
 extern struct syscore_ops s3c24xx_irq_syscore_ops;
 
+extern struct platform_device s3c2412_device_dma;
+extern struct platform_device s3c2443_device_dma;
+
 #endif /* __ARCH_ARM_MACH_S3C24XX_COMMON_H */
diff --git a/arch/arm/mach-s3c24xx/mach-jive.c b/arch/arm/mach-s3c24xx/mach-jive.c
index a45fcd8..43c23e2 100644
--- a/arch/arm/mach-s3c24xx/mach-jive.c
+++ b/arch/arm/mach-s3c24xx/mach-jive.c
@@ -466,6 +466,7 @@ static struct platform_device *jive_devices[] __initdata = {
 	&jive_device_wm8750,
 	&s3c_device_nand,
 	&s3c_device_usbgadget,
+	&s3c2412_device_dma,
 };
 
 static struct s3c2410_udc_mach_info jive_udc_cfg __initdata = {
diff --git a/arch/arm/mach-s3c24xx/mach-smdk2413.c b/arch/arm/mach-s3c24xx/mach-smdk2413.c
index 8146e92..c9d31ef 100644
--- a/arch/arm/mach-s3c24xx/mach-smdk2413.c
+++ b/arch/arm/mach-s3c24xx/mach-smdk2413.c
@@ -89,6 +89,7 @@ static struct platform_device *smdk2413_devices[] __initdata = {
 	&s3c_device_i2c0,
 	&s3c_device_iis,
 	&s3c_device_usbgadget,
+	&s3c2412_device_dma,
 };
 
 static void __init smdk2413_fixup(struct tag *tags, char **cmdline,
diff --git a/arch/arm/mach-s3c24xx/mach-smdk2416.c b/arch/arm/mach-s3c24xx/mach-smdk2416.c
index cb46847..f88e672 100644
--- a/arch/arm/mach-s3c24xx/mach-smdk2416.c
+++ b/arch/arm/mach-s3c24xx/mach-smdk2416.c
@@ -215,6 +215,7 @@ static struct platform_device *smdk2416_devices[] __initdata = {
 	&s3c_device_hsmmc0,
 	&s3c_device_hsmmc1,
 	&s3c_device_usb_hsudc,
+	&s3c2443_device_dma,
 };
 
 static void __init smdk2416_map_io(void)
diff --git a/arch/arm/mach-s3c24xx/mach-smdk2443.c b/arch/arm/mach-s3c24xx/mach-smdk2443.c
index 9435c3b..d9933fc 100644
--- a/arch/arm/mach-s3c24xx/mach-smdk2443.c
+++ b/arch/arm/mach-s3c24xx/mach-smdk2443.c
@@ -115,6 +115,7 @@ static struct platform_device *smdk2443_devices[] __initdata = {
 #ifdef CONFIG_SND_SOC_SMDK2443_WM9710
 	&s3c_device_ac97,
 #endif
+	&s3c2443_device_dma,
 };
 
 static void __init smdk2443_map_io(void)
diff --git a/arch/arm/mach-s3c24xx/mach-vstms.c b/arch/arm/mach-s3c24xx/mach-vstms.c
index b665884..f7ec9c5 100644
--- a/arch/arm/mach-s3c24xx/mach-vstms.c
+++ b/arch/arm/mach-s3c24xx/mach-vstms.c
@@ -126,6 +126,7 @@ static struct platform_device *vstms_devices[] __initdata = {
 	&s3c_device_iis,
 	&s3c_device_rtc,
 	&s3c_device_nand,
+	&s3c2412_device_dma,
 };
 
 static void __init vstms_fixup(struct tag *tags, char **cmdline,
-- 
1.7.10.4

[PATCH v5 4/4] ARM: SAMSUNG: set s3c24xx_dma_filter for s3c64xx-spi0 device

[Heiko Stübner]

The spi-s3c64xx device is also used on the s3c2416 and s3c2443 SoCs.
The driver also already uses only generic dma-engine operations.

Therefore add another elif to set the s3c24xx filter.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/plat-samsung/devs.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/arch/arm/plat-samsung/devs.c b/arch/arm/plat-samsung/devs.c
index 8ce0ac0..a7b9b9e 100644
--- a/arch/arm/plat-samsung/devs.c
+++ b/arch/arm/plat-samsung/devs.c
@@ -32,6 +32,7 @@
 #include <linux/ioport.h>
 #include <linux/platform_data/s3c-hsudc.h>
 #include <linux/platform_data/s3c-hsotg.h>
+#include <linux/platform_data/dma-s3c24xx.h>
 
 #include <media/s5p_hdmi.h>
 
@@ -1499,8 +1500,10 @@ void __init s3c64xx_spi0_set_platdata(int (*cfg_gpio)(void), int src_clk_nr,
 	pd.num_cs = num_cs;
 	pd.src_clk_nr = src_clk_nr;
 	pd.cfg_gpio = (cfg_gpio) ? cfg_gpio : s3c64xx_spi0_cfg_gpio;
-#ifdef CONFIG_PL330_DMA
+#if defined(CONFIG_PL330_DMA)
 	pd.filter = pl330_filter;
+#elif defined(CONFIG_S3C24XX_DMAC)
+	pd.filter = s3c24xx_dma_filter;
 #endif
 
 	s3c_set_platdata(&pd, sizeof(pd), &s3c64xx_device_spi0);
-- 
1.7.10.4

Re: [PATCH v5 2/4] dmaengine: add driver for Samsung s3c24xx SoCs

[Heiko Stübner]

Hi Vinod,

with the removed double-retrival of the cookie state, is the driver ok now?
Because I think kgene needs an ack on it.


Thanks
Heiko


Am Mittwoch, 18. September 2013, 21:19:54 schrieb Heiko Stübner:
> This adds a new driver to support the s3c24xx dma using the dmaengine
> and makes the old one in mach-s3c24xx obsolete in the long run.
> 
> Conceptually the s3c24xx-dma feels like a distant relative of the pl08x
> with numerous virtual channels being mapped to a lot less physical ones.
> The driver therefore borrows a lot from the amba-pl08x driver in this
> regard. Functionality-wise the driver gains a memcpy ability in addition
> to the slave_sg one.
> 
> The driver supports both the method for requesting the peripheral used
> by SoCs before the S3C2443 and the different method for S3C2443 and later.
> 
> On earlier SoCs the hardware channels usable for specific peripherals is
> constrainted while on later SoCs all channels can be used for any
> peripheral.
> 
> Tested on a s3c2416-based board, memcpy using the dmatest module and
> slave_sg partially using the spi-s3c64xx driver.
> 
> Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> Acked-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> changes since v4:
> - fix the unnecessary double retrieval of the cookie status
>   pointed out by Vinod Koul
> changes since v3:
> - address comments from Tomasz Figa:
>   - fixed a lot of nitpics
>   - and added a rudimentary mechanism to select higher bus-widths for
> memcpy - address more comments from Vinod Koul - reorder dma_control
> contents
> 
> changes since v2:
> - add missing channel validation in s3c24xx_dma_phy_free when repurposing
>   a physical channel for another virtual channel
> - address comments from Vinod Koul:
>   - fix naming of constants to prevent conflicts
>   - better handle the unlikely case of an interrupt happening on an
>     unused channel
> 
> changes since v1:
> - address comments from Linus Walleij
> - support sg-lists with more than 1 element
> - add support for earlier s3c24xx SoCs
> - redo the channeldata definitions:
>   - no need for static ids for the virtual channels
>   - provide a way to encode constraints for earlier variants
>   - better base for later dt bindings
> 
>  drivers/dma/Kconfig                       |   12 +
>  drivers/dma/Makefile                      |    1 +
>  drivers/dma/s3c24xx-dma.c                 | 1340
> +++++++++++++++++++++++++++++ include/linux/platform_data/dma-s3c24xx.h | 
>  43 +
>  4 files changed, 1396 insertions(+)
>  create mode 100644 drivers/dma/s3c24xx-dma.c
>  create mode 100644 include/linux/platform_data/dma-s3c24xx.h
> 
> diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
> index 526ec77..d639115 100644
> --- a/drivers/dma/Kconfig
> +++ b/drivers/dma/Kconfig
> @@ -154,6 +154,18 @@ config TEGRA20_APB_DMA
>  	  This DMA controller transfers data from memory to peripheral fifo
>  	  or vice versa. It does not support memory to memory data transfer.
> 
> +config S3C24XX_DMAC
> +	tristate "Samsung S3C24XX DMA support"
> +	depends on ARCH_S3C24XX && !S3C24XX_DMA
> +	select DMA_ENGINE
> +	select DMA_VIRTUAL_CHANNELS
> +	help
> +	  Support for the Samsung S3C24XX DMA controller driver. The
> +	  DMA controller is having multiple DMA channels which can be
> +	  configured for different peripherals like audio, UART, SPI.
> +	  The DMA controller can transfer data from memory to peripheral,
> +	  periphal to memory, periphal to periphal and memory to memory.
> +
>  source "drivers/dma/sh/Kconfig"
> 
>  config COH901318
> diff --git a/drivers/dma/Makefile b/drivers/dma/Makefile
> index db89035..0ce2da9 100644
> --- a/drivers/dma/Makefile
> +++ b/drivers/dma/Makefile
> @@ -30,6 +30,7 @@ obj-$(CONFIG_SIRF_DMA) += sirf-dma.o
>  obj-$(CONFIG_TI_EDMA) += edma.o
>  obj-$(CONFIG_STE_DMA40) += ste_dma40.o ste_dma40_ll.o
>  obj-$(CONFIG_TEGRA20_APB_DMA) += tegra20-apb-dma.o
> +obj-$(CONFIG_S3C24XX_DMAC) += s3c24xx-dma.o
>  obj-$(CONFIG_PL330_DMA) += pl330.o
>  obj-$(CONFIG_PCH_DMA) += pch_dma.o
>  obj-$(CONFIG_AMBA_PL08X) += amba-pl08x.o
> diff --git a/drivers/dma/s3c24xx-dma.c b/drivers/dma/s3c24xx-dma.c
> new file mode 100644
> index 0000000..56c9253
> --- /dev/null
> +++ b/drivers/dma/s3c24xx-dma.c
> @@ -0,0 +1,1340 @@
> +/*
> + * S3C24XX DMA handling
> + *
> + * Copyright (c) 2013 Heiko Stuebner <heiko@sntech.de>
> + *
> + * based on amba-pl08x.c
> + *
> + * Copyright (c) 2006 ARM Ltd.
> + * Copyright (c) 2010 ST-Ericsson SA
> + *
> + * Author: Peter Pearse <peter.pearse@arm.com>
> + * Author: Linus Walleij <linus.walleij@stericsson.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.
> + *
> + * The DMA controllers in S3C24XX SoCs have a varying number of DMA
> signals + * that can be routed to any of the 4 to 8 hardware-channels.
> + *
> + * Therefore on these DMA controllers the number of channels
> + * and the number of incoming DMA signals are two totally different
> things. + * It is usually not possible to theoretically handle all
> physical signals, + * so a multiplexing scheme with possible denial of use
> is necessary. + *
> + * Open items:
> + * - bursts
> + */
> +
> +#include <linux/platform_device.h>
> +#include <linux/types.h>
> +#include <linux/dmaengine.h>
> +#include <linux/dma-mapping.h>
> +#include <linux/interrupt.h>
> +#include <linux/clk.h>
> +#include <linux/module.h>
> +#include <linux/slab.h>
> +#include <linux/platform_data/dma-s3c24xx.h>
> +
> +#include "dmaengine.h"
> +#include "virt-dma.h"
> +
> +#define MAX_DMA_CHANNELS	8
> +
> +#define S3C24XX_DISRC			0x00
> +#define S3C24XX_DISRCC			0x04
> +#define S3C24XX_DISRCC_INC_INCREMENT	0
> +#define S3C24XX_DISRCC_INC_FIXED	BIT(0)
> +#define S3C24XX_DISRCC_LOC_AHB		0
> +#define S3C24XX_DISRCC_LOC_APB		BIT(1)
> +
> +#define S3C24XX_DIDST			0x08
> +#define S3C24XX_DIDSTC			0x0c
> +#define S3C24XX_DIDSTC_INC_INCREMENT	0
> +#define S3C24XX_DIDSTC_INC_FIXED	BIT(0)
> +#define S3C24XX_DIDSTC_LOC_AHB		0
> +#define S3C24XX_DIDSTC_LOC_APB		BIT(1)
> +#define S3C24XX_DIDSTC_INT_TC0		0
> +#define S3C24XX_DIDSTC_INT_RELOAD	BIT(2)
> +
> +#define S3C24XX_DCON			0x10
> +
> +#define S3C24XX_DCON_TC_MASK		0xfffff
> +#define S3C24XX_DCON_DSZ_BYTE		(0 << 20)
> +#define S3C24XX_DCON_DSZ_HALFWORD	(1 << 20)
> +#define S3C24XX_DCON_DSZ_WORD		(2 << 20)
> +#define S3C24XX_DCON_DSZ_MASK		(3 << 20)
> +#define S3C24XX_DCON_DSZ_SHIFT		20
> +#define S3C24XX_DCON_AUTORELOAD		0
> +#define S3C24XX_DCON_NORELOAD		BIT(22)
> +#define S3C24XX_DCON_HWTRIG		BIT(23)
> +#define S3C24XX_DCON_HWSRC_SHIFT	24
> +#define S3C24XX_DCON_SERV_SINGLE	0
> +#define S3C24XX_DCON_SERV_WHOLE		BIT(27)
> +#define S3C24XX_DCON_TSZ_UNIT		0
> +#define S3C24XX_DCON_TSZ_BURST4		BIT(28)
> +#define S3C24XX_DCON_INT		BIT(29)
> +#define S3C24XX_DCON_SYNC_PCLK		0
> +#define S3C24XX_DCON_SYNC_HCLK		BIT(30)
> +#define S3C24XX_DCON_DEMAND		0
> +#define S3C24XX_DCON_HANDSHAKE		BIT(31)
> +
> +#define S3C24XX_DSTAT			0x14
> +#define S3C24XX_DSTAT_STAT_BUSY		BIT(20)
> +#define S3C24XX_DSTAT_CURRTC_MASK	0xfffff
> +
> +#define S3C24XX_DMASKTRIG		0x20
> +#define S3C24XX_DMASKTRIG_SWTRIG	BIT(0)
> +#define S3C24XX_DMASKTRIG_ON		BIT(1)
> +#define S3C24XX_DMASKTRIG_STOP		BIT(2)
> +
> +#define S3C24XX_DMAREQSEL		0x24
> +#define S3C24XX_DMAREQSEL_HW		BIT(0)
> +
> +/*
> + * S3C2410, S3C2440 and S3C2442 SoCs cannot select any physical channel
> + * for a DMA source. Instead only specific channels are valid.
> + * All of these SoCs have 4 physical channels and the number of request
> + * source bits is 3. Additionally we also need 1 bit to mark the channel
> + * as valid.
> + * Therefore we separate the chansel element of the channel data into 4
> + * parts of 4 bits each, to hold the information if the channel is valid
> + * and the hw request source to use.
> + *
> + * Example:
> + * SDI is valid on channels 0, 2 and 3 - with varying hw request sources.
> + * For it the chansel field would look like
> + *
> + * ((BIT(3) | 1) << 3 * 4) | // channel 3, with request source 1
> + * ((BIT(3) | 2) << 2 * 4) | // channel 2, with request source 2
> + * ((BIT(3) | 2) << 0 * 4)   // channel 0, with request source 2
> + */
> +#define S3C24XX_CHANSEL_WIDTH		4
> +#define S3C24XX_CHANSEL_VALID		BIT(3)
> +#define S3C24XX_CHANSEL_REQ_MASK	7
> +
> +/*
> + * struct soc_data - vendor-specific config parameters for individual SoCs
> + * @stride: spacing between the registers of each channel
> + * @has_reqsel: does the controller use the newer requestselection
> mechanism + * @has_clocks: are controllable dma-clocks present
> + */
> +struct soc_data {
> +	int stride;
> +	bool has_reqsel;
> +	bool has_clocks;
> +};
> +
> +/*
> + * enum s3c24xx_dma_chan_state - holds the virtual channel states
> + * @S3C24XX_DMA_CHAN_IDLE: the channel is idle
> + * @S3C24XX_DMA_CHAN_RUNNING: the channel has allocated a physical
> transport + * channel and is running a transfer on it
> + * @S3C24XX_DMA_CHAN_WAITING: the channel is waiting for a physical
> transport + * channel to become available (only pertains to memcpy
> channels) + */
> +enum s3c24xx_dma_chan_state {
> +	S3C24XX_DMA_CHAN_IDLE,
> +	S3C24XX_DMA_CHAN_RUNNING,
> +	S3C24XX_DMA_CHAN_WAITING,
> +};
> +
> +/*
> + * struct s3c24xx_sg - structure containing data per sg
> + * @src_addr: src address of sg
> + * @dst_addr: dst address of sg
> + * @len: transfer len in bytes
> + * @node: node for txd's dsg_list
> + */
> +struct s3c24xx_sg {
> +	dma_addr_t src_addr;
> +	dma_addr_t dst_addr;
> +	size_t len;
> +	struct list_head node;
> +};
> +
> +/*
> + * struct s3c24xx_txd - wrapper for struct dma_async_tx_descriptor
> + * @vd: virtual DMA descriptor
> + * @dsg_list: list of children sg's
> + * @at: sg currently being transfered
> + * @width: transfer width
> + * @disrcc: value for source control register
> + * @didstc: value for destination control register
> + * @dcon: base value for dcon register
> + */
> +struct s3c24xx_txd {
> +	struct virt_dma_desc vd;
> +	struct list_head dsg_list;
> +	struct list_head *at;
> +	u8 width;
> +	u32 disrcc;
> +	u32 didstc;
> +	u32 dcon;
> +};
> +
> +struct s3c24xx_dma_chan;
> +
> +/*
> + * struct s3c24xx_dma_phy - holder for the physical channels
> + * @id: physical index to this channel
> + * @valid: does the channel have all required elements
> + * @base: virtual memory base (remapped) for the this channel
> + * @irq: interrupt for this channel
> + * @clk: clock for this channel
> + * @lock: a lock to use when altering an instance of this struct
> + * @serving: virtual channel currently being served by this
> physicalchannel + * @host: a pointer to the host (internal use)
> + */
> +struct s3c24xx_dma_phy {
> +	unsigned int			id;
> +	bool				valid;
> +	void __iomem			*base;
> +	unsigned int			irq;
> +	struct clk			*clk;
> +	spinlock_t			lock;
> +	struct s3c24xx_dma_chan		*serving;
> +	struct s3c24xx_dma_engine	*host;
> +};
> +
> +/*
> + * struct s3c24xx_dma_chan - this structure wraps a DMA ENGINE channel
> + * @id: the id of the channel
> + * @name: name of the channel
> + * @vc: wrappped virtual channel
> + * @phy: the physical channel utilized by this channel, if there is one
> + * @runtime_addr: address for RX/TX according to the runtime config
> + * @at: active transaction on this channel
> + * @lock: a lock for this channel data
> + * @host: a pointer to the host (internal use)
> + * @state: whether the channel is idle, running etc
> + * @slave: whether this channel is a device (slave) or for memcpy
> + */
> +struct s3c24xx_dma_chan {
> +	int id;
> +	const char *name;
> +	struct virt_dma_chan vc;
> +	struct s3c24xx_dma_phy *phy;
> +	struct dma_slave_config cfg;
> +	struct s3c24xx_txd *at;
> +	struct s3c24xx_dma_engine *host;
> +	enum s3c24xx_dma_chan_state state;
> +	bool slave;
> +};
> +
> +/*
> + * struct s3c24xx_dma_engine - the local state holder for the S3C24XX
> + * @pdev: the corresponding platform device
> + * @pdata: platform data passed in from the platform/machine
> + * @base: virtual memory base (remapped)
> + * @slave: slave engine for this instance
> + * @memcpy: memcpy engine for this instance
> + * @phy_chans: array of data for the physical channels
> + */
> +struct s3c24xx_dma_engine {
> +	struct platform_device			*pdev;
> +	const struct s3c24xx_dma_platdata	*pdata;
> +	struct soc_data				*sdata;
> +	void __iomem				*base;
> +	struct dma_device			slave;
> +	struct dma_device			memcpy;
> +	struct s3c24xx_dma_phy			*phy_chans;
> +};
> +
> +/*
> + * Physical channel handling
> + */
> +
> +/*
> + * Check whether a certain channel is busy or not.
> + */
> +static int s3c24xx_dma_phy_busy(struct s3c24xx_dma_phy *phy)
> +{
> +	unsigned int val = readl(phy->base + S3C24XX_DSTAT);
> +	return val & S3C24XX_DSTAT_STAT_BUSY;
> +}
> +
> +static bool s3c24xx_dma_phy_valid(struct s3c24xx_dma_chan *s3cchan,
> +				  struct s3c24xx_dma_phy *phy)
> +{
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
> +	struct s3c24xx_dma_channel *cdata = &pdata->channels[s3cchan->id];
> +	int phyvalid;
> +
> +	/* every phy is valid for memcopy channels */
> +	if (!s3cchan->slave)
> +		return true;
> +
> +	/* On newer variants all phys can be used for all virtual channels */
> +	if (s3cdma->sdata->has_reqsel)
> +		return true;
> +
> +	phyvalid = (cdata->chansel >> (phy->id * S3C24XX_CHANSEL_WIDTH));
> +	return (phyvalid & S3C24XX_CHANSEL_VALID) ? true : false;
> +}
> +
> +/*
> + * Allocate a physical channel for a virtual channel
> + *
> + * Try to locate a physical channel to be used for this transfer. If all
> + * are taken return NULL and the requester will have to cope by using
> + * some fallback PIO mode or retrying later.
> + */
> +static
> +struct s3c24xx_dma_phy *s3c24xx_dma_get_phy(struct s3c24xx_dma_chan
> *s3cchan) +{
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
> +	struct s3c24xx_dma_channel *cdata;
> +	struct s3c24xx_dma_phy *phy = NULL;
> +	unsigned long flags;
> +	int i;
> +	int ret;
> +
> +	if (s3cchan->slave)
> +		cdata = &pdata->channels[s3cchan->id];
> +
> +	for (i = 0; i < s3cdma->pdata->num_phy_channels; i++) {
> +		phy = &s3cdma->phy_chans[i];
> +
> +		if (!phy->valid)
> +			continue;
> +
> +		if (!s3c24xx_dma_phy_valid(s3cchan, phy))
> +			continue;
> +
> +		spin_lock_irqsave(&phy->lock, flags);
> +
> +		if (!phy->serving) {
> +			phy->serving = s3cchan;
> +			spin_unlock_irqrestore(&phy->lock, flags);
> +			break;
> +		}
> +
> +		spin_unlock_irqrestore(&phy->lock, flags);
> +	}
> +
> +	/* No physical channel available, cope with it */
> +	if (i == s3cdma->pdata->num_phy_channels) {
> +		dev_warn(&s3cdma->pdev->dev, "no phy channel available\n");
> +		return NULL;
> +	}
> +
> +	/* start the phy clock */
> +	if (s3cdma->sdata->has_clocks) {
> +		ret = clk_enable(phy->clk);
> +		if (ret) {
> +			dev_err(&s3cdma->pdev->dev, "could not enable clock for channel 
%d, err
> %d\n", +				phy->id, ret);
> +			phy->serving = NULL;
> +			return NULL;
> +		}
> +	}
> +
> +	return phy;
> +}
> +
> +/*
> + * Mark the physical channel as free.
> + *
> + * This drops the link between the physical and virtual channel.
> + */
> +static inline void s3c24xx_dma_put_phy(struct s3c24xx_dma_phy *phy)
> +{
> +	struct s3c24xx_dma_engine *s3cdma = phy->host;
> +
> +	if (s3cdma->sdata->has_clocks)
> +		clk_disable(phy->clk);
> +
> +	phy->serving = NULL;
> +}
> +
> +/*
> + * Stops the channel by writing the stop bit.
> + * This should not be used for an on-going transfer, but as a method of
> + * shutting down a channel (eg, when it's no longer used) or terminating a
> + * transfer.
> + */
> +static void s3c24xx_dma_terminate_phy(struct s3c24xx_dma_phy *phy)
> +{
> +	writel(S3C24XX_DMASKTRIG_STOP, phy->base + S3C24XX_DMASKTRIG);
> +}
> +
> +/*
> + * Virtual channel handling
> + */
> +
> +static inline
> +struct s3c24xx_dma_chan *to_s3c24xx_dma_chan(struct dma_chan *chan)
> +{
> +	return container_of(chan, struct s3c24xx_dma_chan, vc.chan);
> +}
> +
> +static u32 s3c24xx_dma_getbytes_chan(struct s3c24xx_dma_chan *s3cchan)
> +{
> +	struct s3c24xx_dma_phy *phy = s3cchan->phy;
> +	struct s3c24xx_txd *txd = s3cchan->at;
> +	u32 tc = readl(phy->base + S3C24XX_DSTAT) & S3C24XX_DSTAT_CURRTC_MASK;
> +
> +	return tc * txd->width;
> +}
> +
> +static int s3c24xx_dma_set_runtime_config(struct s3c24xx_dma_chan
> *s3cchan, +				  struct dma_slave_config *config)
> +{
> +	if (!s3cchan->slave)
> +		return -EINVAL;
> +
> +	/* Reject definitely invalid configurations */
> +	if (config->src_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES ||
> +	    config->dst_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES)
> +		return -EINVAL;
> +
> +	s3cchan->cfg = *config;
> +
> +	return 0;
> +}
> +
> +/*
> + * Transfer handling
> + */
> +
> +static inline
> +struct s3c24xx_txd *to_s3c24xx_txd(struct dma_async_tx_descriptor *tx)
> +{
> +	return container_of(tx, struct s3c24xx_txd, vd.tx);
> +}
> +
> +static struct s3c24xx_txd *s3c24xx_dma_get_txd(void)
> +{
> +	struct s3c24xx_txd *txd = kzalloc(sizeof(*txd), GFP_NOWAIT);
> +
> +	if (txd) {
> +		INIT_LIST_HEAD(&txd->dsg_list);
> +		txd->dcon = S3C24XX_DCON_INT | S3C24XX_DCON_NORELOAD;
> +	}
> +
> +	return txd;
> +}
> +
> +static void s3c24xx_dma_free_txd(struct s3c24xx_txd *txd)
> +{
> +	struct s3c24xx_sg *dsg, *_dsg;
> +
> +	list_for_each_entry_safe(dsg, _dsg, &txd->dsg_list, node) {
> +		list_del(&dsg->node);
> +		kfree(dsg);
> +	}
> +
> +	kfree(txd);
> +}
> +
> +static void s3c24xx_dma_start_next_sg(struct s3c24xx_dma_chan *s3cchan,
> +				       struct s3c24xx_txd *txd)
> +{
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +	struct s3c24xx_dma_phy *phy = s3cchan->phy;
> +	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
> +	struct s3c24xx_sg *dsg = list_entry(txd->at, struct s3c24xx_sg, node);
> +	u32 dcon = txd->dcon;
> +	u32 val;
> +
> +	/* transfer-size and -count from len and width */
> +	switch (txd->width) {
> +	case 1:
> +		dcon |= S3C24XX_DCON_DSZ_BYTE | dsg->len;
> +		break;
> +	case 2:
> +		dcon |= S3C24XX_DCON_DSZ_HALFWORD | (dsg->len / 2);
> +		break;
> +	case 4:
> +		dcon |= S3C24XX_DCON_DSZ_WORD | (dsg->len / 4);
> +		break;
> +	}
> +
> +	if (s3cchan->slave) {
> +		struct s3c24xx_dma_channel *cdata =
> +					&pdata->channels[s3cchan->id];
> +
> +		if (s3cdma->sdata->has_reqsel) {
> +			writel_relaxed((cdata->chansel << 1) |
> +							S3C24XX_DMAREQSEL_HW,
> +					phy->base + S3C24XX_DMAREQSEL);
> +		} else {
> +			int csel = cdata->chansel >> (phy->id *
> +							S3C24XX_CHANSEL_WIDTH);
> +
> +			csel &= S3C24XX_CHANSEL_REQ_MASK;
> +			dcon |= csel << S3C24XX_DCON_HWSRC_SHIFT;
> +			dcon |= S3C24XX_DCON_HWTRIG;
> +		}
> +	} else {
> +		if (s3cdma->sdata->has_reqsel)
> +			writel_relaxed(0, phy->base + S3C24XX_DMAREQSEL);
> +	}
> +
> +	writel_relaxed(dsg->src_addr, phy->base + S3C24XX_DISRC);
> +	writel_relaxed(txd->disrcc, phy->base + S3C24XX_DISRCC);
> +	writel_relaxed(dsg->dst_addr, phy->base + S3C24XX_DIDST);
> +	writel_relaxed(txd->didstc, phy->base + S3C24XX_DIDSTC);
> +	writel_relaxed(dcon, phy->base + S3C24XX_DCON);
> +
> +	val = readl_relaxed(phy->base + S3C24XX_DMASKTRIG);
> +	val &= ~S3C24XX_DMASKTRIG_STOP;
> +	val |= S3C24XX_DMASKTRIG_ON;
> +
> +	/* trigger the dma operation for memcpy transfers */
> +	if (!s3cchan->slave)
> +		val |= S3C24XX_DMASKTRIG_SWTRIG;
> +
> +	writel(val, phy->base + S3C24XX_DMASKTRIG);
> +}
> +
> +/*
> + * Set the initial DMA register values and start first sg.
> + */
> +static void s3c24xx_dma_start_next_txd(struct s3c24xx_dma_chan *s3cchan)
> +{
> +	struct s3c24xx_dma_phy *phy = s3cchan->phy;
> +	struct virt_dma_desc *vd = vchan_next_desc(&s3cchan->vc);
> +	struct s3c24xx_txd *txd = to_s3c24xx_txd(&vd->tx);
> +
> +	list_del(&txd->vd.node);
> +
> +	s3cchan->at = txd;
> +
> +	/* Wait for channel inactive */
> +	while (s3c24xx_dma_phy_busy(phy))
> +		cpu_relax();
> +
> +	/* point to the first element of the sg list */
> +	txd->at = txd->dsg_list.next;
> +	s3c24xx_dma_start_next_sg(s3cchan, txd);
> +}
> +
> +static void s3c24xx_dma_free_txd_list(struct s3c24xx_dma_engine *s3cdma,
> +				struct s3c24xx_dma_chan *s3cchan)
> +{
> +	LIST_HEAD(head);
> +
> +	vchan_get_all_descriptors(&s3cchan->vc, &head);
> +	vchan_dma_desc_free_list(&s3cchan->vc, &head);
> +}
> +
> +/*
> + * Try to allocate a physical channel.  When successful, assign it to
> + * this virtual channel, and initiate the next descriptor.  The
> + * virtual channel lock must be held at this point.
> + */
> +static void s3c24xx_dma_phy_alloc_and_start(struct s3c24xx_dma_chan
> *s3cchan) +{
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +	struct s3c24xx_dma_phy *phy;
> +
> +	phy = s3c24xx_dma_get_phy(s3cchan);
> +	if (!phy) {
> +		dev_dbg(&s3cdma->pdev->dev, "no physical channel available for xfer 
on
> %s\n", +			s3cchan->name);
> +		s3cchan->state = S3C24XX_DMA_CHAN_WAITING;
> +		return;
> +	}
> +
> +	dev_dbg(&s3cdma->pdev->dev, "allocated physical channel %d for xfer on
> %s\n", +		phy->id, s3cchan->name);
> +
> +	s3cchan->phy = phy;
> +	s3cchan->state = S3C24XX_DMA_CHAN_RUNNING;
> +
> +	s3c24xx_dma_start_next_txd(s3cchan);
> +}
> +
> +static void s3c24xx_dma_phy_reassign_start(struct s3c24xx_dma_phy *phy,
> +	struct s3c24xx_dma_chan *s3cchan)
> +{
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +
> +	dev_dbg(&s3cdma->pdev->dev, "reassigned physical channel %d for xfer on
> %s\n", +		phy->id, s3cchan->name);
> +
> +	/*
> +	 * We do this without taking the lock; we're really only concerned
> +	 * about whether this pointer is NULL or not, and we're guaranteed
> +	 * that this will only be called when it _already_ is non-NULL.
> +	 */
> +	phy->serving = s3cchan;
> +	s3cchan->phy = phy;
> +	s3cchan->state = S3C24XX_DMA_CHAN_RUNNING;
> +	s3c24xx_dma_start_next_txd(s3cchan);
> +}
> +
> +/*
> + * Free a physical DMA channel, potentially reallocating it to another
> + * virtual channel if we have any pending.
> + */
> +static void s3c24xx_dma_phy_free(struct s3c24xx_dma_chan *s3cchan)
> +{
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +	struct s3c24xx_dma_chan *p, *next;
> +
> +retry:
> +	next = NULL;
> +
> +	/* Find a waiting virtual channel for the next transfer. */
> +	list_for_each_entry(p, &s3cdma->memcpy.channels, vc.chan.device_node)
> +		if (p->state == S3C24XX_DMA_CHAN_WAITING) {
> +			next = p;
> +			break;
> +		}
> +
> +	if (!next) {
> +		list_for_each_entry(p, &s3cdma->slave.channels,
> +				    vc.chan.device_node)
> +			if (p->state == S3C24XX_DMA_CHAN_WAITING &&
> +				      s3c24xx_dma_phy_valid(p, s3cchan->phy)) {
> +				next = p;
> +				break;
> +			}
> +	}
> +
> +	/* Ensure that the physical channel is stopped */
> +	s3c24xx_dma_terminate_phy(s3cchan->phy);
> +
> +	if (next) {
> +		bool success;
> +
> +		/*
> +		 * Eww.  We know this isn't going to deadlock
> +		 * but lockdep probably doesn't.
> +		 */
> +		spin_lock(&next->vc.lock);
> +		/* Re-check the state now that we have the lock */
> +		success = next->state == S3C24XX_DMA_CHAN_WAITING;
> +		if (success)
> +			s3c24xx_dma_phy_reassign_start(s3cchan->phy, next);
> +		spin_unlock(&next->vc.lock);
> +
> +		/* If the state changed, try to find another channel */
> +		if (!success)
> +			goto retry;
> +	} else {
> +		/* No more jobs, so free up the physical channel */
> +		s3c24xx_dma_put_phy(s3cchan->phy);
> +	}
> +
> +	s3cchan->phy = NULL;
> +	s3cchan->state = S3C24XX_DMA_CHAN_IDLE;
> +}
> +
> +static void s3c24xx_dma_unmap_buffers(struct s3c24xx_txd *txd)
> +{
> +	struct device *dev = txd->vd.tx.chan->device->dev;
> +	struct s3c24xx_sg *dsg;
> +
> +	if (!(txd->vd.tx.flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
> +		if (txd->vd.tx.flags & DMA_COMPL_SRC_UNMAP_SINGLE)
> +			list_for_each_entry(dsg, &txd->dsg_list, node)
> +				dma_unmap_single(dev, dsg->src_addr, dsg->len,
> +						DMA_TO_DEVICE);
> +		else {
> +			list_for_each_entry(dsg, &txd->dsg_list, node)
> +				dma_unmap_page(dev, dsg->src_addr, dsg->len,
> +						DMA_TO_DEVICE);
> +		}
> +	}
> +
> +	if (!(txd->vd.tx.flags & DMA_COMPL_SKIP_DEST_UNMAP)) {
> +		if (txd->vd.tx.flags & DMA_COMPL_DEST_UNMAP_SINGLE)
> +			list_for_each_entry(dsg, &txd->dsg_list, node)
> +				dma_unmap_single(dev, dsg->dst_addr, dsg->len,
> +						DMA_FROM_DEVICE);
> +		else
> +			list_for_each_entry(dsg, &txd->dsg_list, node)
> +				dma_unmap_page(dev, dsg->dst_addr, dsg->len,
> +						DMA_FROM_DEVICE);
> +	}
> +}
> +
> +static void s3c24xx_dma_desc_free(struct virt_dma_desc *vd)
> +{
> +	struct s3c24xx_txd *txd = to_s3c24xx_txd(&vd->tx);
> +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(vd->tx.chan);
> +
> +	if (!s3cchan->slave)
> +		s3c24xx_dma_unmap_buffers(txd);
> +
> +	s3c24xx_dma_free_txd(txd);
> +}
> +
> +static irqreturn_t s3c24xx_dma_irq(int irq, void *data)
> +{
> +	struct s3c24xx_dma_phy *phy = data;
> +	struct s3c24xx_dma_chan *s3cchan = phy->serving;
> +	struct s3c24xx_txd *txd;
> +
> +	dev_dbg(&phy->host->pdev->dev, "interrupt on channel %d\n", phy->id);
> +
> +	/*
> +	 * Interrupts happen to notify the completion of a transfer and the
> +	 * channel should have moved into its stop state already on its own.
> +	 * Therefore interrupts on channels not bound to a virtual channel
> +	 * should never happen. Nevertheless send a terminate command to the
> +	 * channel if the unlikely case happens.
> +	 */
> +	if (unlikely(!s3cchan)) {
> +		dev_err(&phy->host->pdev->dev, "interrupt on unused channel %d\n",
> +			phy->id);
> +
> +		s3c24xx_dma_terminate_phy(phy);
> +
> +		return IRQ_HANDLED;
> +	}
> +
> +	spin_lock(&s3cchan->vc.lock);
> +	txd = s3cchan->at;
> +	if (txd) {
> +		/* when more sg's are in this txd, start the next one */
> +		if (!list_is_last(txd->at, &txd->dsg_list)) {
> +			txd->at = txd->at->next;
> +			s3c24xx_dma_start_next_sg(s3cchan, txd);
> +		} else {
> +			s3cchan->at = NULL;
> +			vchan_cookie_complete(&txd->vd);
> +
> +			/*
> +			 * And start the next descriptor (if any),
> +			 * otherwise free this channel.
> +			 */
> +			if (vchan_next_desc(&s3cchan->vc))
> +				s3c24xx_dma_start_next_txd(s3cchan);
> +			else
> +				s3c24xx_dma_phy_free(s3cchan);
> +		}
> +	}
> +	spin_unlock(&s3cchan->vc.lock);
> +
> +	return IRQ_HANDLED;
> +}
> +
> +/*
> + * The DMA ENGINE API
> + */
> +
> +static int s3c24xx_dma_control(struct dma_chan *chan, enum dma_ctrl_cmd
> cmd, +			 unsigned long arg)
> +{
> +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +	unsigned long flags;
> +	int ret = 0;
> +
> +	spin_lock_irqsave(&s3cchan->vc.lock, flags);
> +
> +	switch (cmd) {
> +	case DMA_SLAVE_CONFIG:
> +		ret = s3c24xx_dma_set_runtime_config(s3cchan,
> +					      (struct dma_slave_config *)arg);
> +		break;
> +	case DMA_TERMINATE_ALL:
> +		if (!s3cchan->phy && !s3cchan->at) {
> +			dev_err(&s3cdma->pdev->dev, "trying to terminate already stopped
> channel %d\n", +				s3cchan->id);
> +			ret = -EINVAL;
> +			break;
> +		}
> +
> +		s3cchan->state = S3C24XX_DMA_CHAN_IDLE;
> +
> +		 /* Mark physical channel as free */
> +		if (s3cchan->phy)
> +			s3c24xx_dma_phy_free(s3cchan);
> +
> +		/* Dequeue current job */
> +		if (s3cchan->at) {
> +			s3c24xx_dma_desc_free(&s3cchan->at->vd);
> +			s3cchan->at = NULL;
> +		}
> +
> +		/* Dequeue jobs not yet fired as well */
> +		s3c24xx_dma_free_txd_list(s3cdma, s3cchan);
> +		break;
> +	default:
> +		/* Unknown command */
> +		ret = -ENXIO;
> +		break;
> +	}
> +
> +	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> +
> +	return ret;
> +}
> +
> +static int s3c24xx_dma_alloc_chan_resources(struct dma_chan *chan)
> +{
> +	return 0;
> +}
> +
> +static void s3c24xx_dma_free_chan_resources(struct dma_chan *chan)
> +{
> +	/* Ensure all queued descriptors are freed */
> +	vchan_free_chan_resources(to_virt_chan(chan));
> +}
> +
> +static enum dma_status s3c24xx_dma_tx_status(struct dma_chan *chan,
> +		dma_cookie_t cookie, struct dma_tx_state *txstate)
> +{
> +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> +	struct s3c24xx_txd *txd;
> +	struct s3c24xx_sg *dsg;
> +	struct virt_dma_desc *vd;
> +	unsigned long flags;
> +	enum dma_status ret;
> +	size_t bytes = 0;
> +
> +	spin_lock_irqsave(&s3cchan->vc.lock, flags);
> +	ret = dma_cookie_status(chan, cookie, txstate);
> +	if (ret == DMA_SUCCESS) {
> +		spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> +		return ret;
> +	}
> +
> +	/*
> +	 * There's no point calculating the residue if there's
> +	 * no txstate to store the value.
> +	 */
> +	if (!txstate) {
> +		spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> +		return ret;
> +	}
> +
> +	vd = vchan_find_desc(&s3cchan->vc, cookie);
> +	if (vd) {
> +		/* On the issued list, so hasn't been processed yet */
> +		txd = to_s3c24xx_txd(&vd->tx);
> +
> +		list_for_each_entry(dsg, &txd->dsg_list, node)
> +			bytes += dsg->len;
> +	} else {
> +		/*
> +		 * Currently running, so sum over the pending sg's and
> +		 * the currently active one.
> +		 */
> +		txd = s3cchan->at;
> +
> +		dsg = list_entry(txd->at, struct s3c24xx_sg, node);
> +		list_for_each_entry_from(dsg, &txd->dsg_list, node)
> +			bytes += dsg->len;
> +
> +		bytes += s3c24xx_dma_getbytes_chan(s3cchan);
> +	}
> +	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> +
> +	/*
> +	 * This cookie not complete yet
> +	 * Get number of bytes left in the active transactions and queue
> +	 */
> +	dma_set_residue(txstate, bytes);
> +
> +	/* Whether waiting or running, we're in progress */
> +	return ret;
> +}
> +
> +/*
> + * Initialize a descriptor to be used by memcpy submit
> + */
> +static struct dma_async_tx_descriptor *s3c24xx_dma_prep_memcpy(
> +		struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
> +		size_t len, unsigned long flags)
> +{
> +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +	struct s3c24xx_txd *txd;
> +	struct s3c24xx_sg *dsg;
> +	int src_mod, dest_mod;
> +
> +	dev_dbg(&s3cdma->pdev->dev, "prepare memcpy of %d bytes from %s\n",
> +			len, s3cchan->name);
> +
> +	if ((len & S3C24XX_DCON_TC_MASK) != len) {
> +		dev_err(&s3cdma->pdev->dev, "memcpy size %d to large\n", len);
> +		return NULL;
> +	}
> +
> +	txd = s3c24xx_dma_get_txd();
> +	if (!txd)
> +		return NULL;
> +
> +	dsg = kzalloc(sizeof(*dsg), GFP_NOWAIT);
> +	if (!dsg) {
> +		s3c24xx_dma_free_txd(txd);
> +		return NULL;
> +	}
> +	list_add_tail(&dsg->node, &txd->dsg_list);
> +
> +	dsg->src_addr = src;
> +	dsg->dst_addr = dest;
> +	dsg->len = len;
> +
> +	/*
> +	 * Determine a suitable transfer width.
> +	 * The DMA controller cannot fetch/store information which is not
> +	 * naturally aligned on the bus, i.e., a 4 byte fetch must start at
> +	 * an address divisible by 4 - more generally addr % width must be 0.
> +	 */
> +	src_mod = src % 4;
> +	dest_mod = dest % 4;
> +	switch (len % 4) {
> +	case 0:
> +		txd->width = (src_mod == 0 && dest_mod == 0) ? 4 : 1;
> +		break;
> +	case 2:
> +		txd->width = ((src_mod == 2 || src_mod == 0) &&
> +			      (dest_mod == 2 || dest_mod == 0)) ? 2 : 1;
> +		break;
> +	default:
> +		txd->width = 1;
> +		break;
> +	}
> +
> +	txd->disrcc = S3C24XX_DISRCC_LOC_AHB | S3C24XX_DISRCC_INC_INCREMENT;
> +	txd->didstc = S3C24XX_DIDSTC_LOC_AHB | S3C24XX_DIDSTC_INC_INCREMENT;
> +	txd->dcon |= S3C24XX_DCON_DEMAND | S3C24XX_DCON_SYNC_HCLK |
> +		     S3C24XX_DCON_SERV_WHOLE;
> +
> +	return vchan_tx_prep(&s3cchan->vc, &txd->vd, flags);
> +}
> +
> +static struct dma_async_tx_descriptor *s3c24xx_dma_prep_slave_sg(
> +		struct dma_chan *chan, struct scatterlist *sgl,
> +		unsigned int sg_len, enum dma_transfer_direction direction,
> +		unsigned long flags, void *context)
> +{
> +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> +	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
> +	struct s3c24xx_dma_channel *cdata = &pdata->channels[s3cchan->id];
> +	struct s3c24xx_txd *txd;
> +	struct s3c24xx_sg *dsg;
> +	struct scatterlist *sg;
> +	dma_addr_t slave_addr;
> +	u32 hwcfg = 0;
> +	int tmp;
> +
> +	dev_dbg(&s3cdma->pdev->dev, "prepare transaction of %d bytes from %s\n",
> +			sg_dma_len(sgl), s3cchan->name);
> +
> +	txd = s3c24xx_dma_get_txd();
> +	if (!txd)
> +		return NULL;
> +
> +	if (cdata->handshake)
> +		txd->dcon |= S3C24XX_DCON_HANDSHAKE;
> +
> +	switch (cdata->bus) {
> +	case S3C24XX_DMA_APB:
> +		txd->dcon |= S3C24XX_DCON_SYNC_PCLK;
> +		hwcfg |= S3C24XX_DISRCC_LOC_APB;
> +		break;
> +	case S3C24XX_DMA_AHB:
> +		txd->dcon |= S3C24XX_DCON_SYNC_HCLK;
> +		hwcfg |= S3C24XX_DISRCC_LOC_AHB;
> +		break;
> +	}
> +
> +	/*
> +	 * Always assume our peripheral desintation is a fixed
> +	 * address in memory.
> +	 */
> +	hwcfg |= S3C24XX_DISRCC_INC_FIXED;
> +
> +	/*
> +	 * Individual dma operations are requested by the slave,
> +	 * so serve only single atomic operations (S3C24XX_DCON_SERV_SINGLE).
> +	 */
> +	txd->dcon |= S3C24XX_DCON_SERV_SINGLE;
> +
> +	if (direction == DMA_MEM_TO_DEV) {
> +		txd->disrcc = S3C24XX_DISRCC_LOC_AHB |
> +			      S3C24XX_DISRCC_INC_INCREMENT;
> +		txd->didstc = hwcfg;
> +		slave_addr = s3cchan->cfg.dst_addr;
> +		txd->width = s3cchan->cfg.dst_addr_width;
> +	} else if (direction == DMA_DEV_TO_MEM) {
> +		txd->disrcc = hwcfg;
> +		txd->didstc = S3C24XX_DIDSTC_LOC_AHB |
> +			      S3C24XX_DIDSTC_INC_INCREMENT;
> +		slave_addr = s3cchan->cfg.src_addr;
> +		txd->width = s3cchan->cfg.src_addr_width;
> +	} else {
> +		s3c24xx_dma_free_txd(txd);
> +		dev_err(&s3cdma->pdev->dev,
> +			"direction %d unsupported\n", direction);
> +		return NULL;
> +	}
> +
> +	for_each_sg(sgl, sg, sg_len, tmp) {
> +		dsg = kzalloc(sizeof(*dsg), GFP_NOWAIT);
> +		if (!dsg) {
> +			s3c24xx_dma_free_txd(txd);
> +			return NULL;
> +		}
> +		list_add_tail(&dsg->node, &txd->dsg_list);
> +
> +		dsg->len = sg_dma_len(sg);
> +		if (direction == DMA_MEM_TO_DEV) {
> +			dsg->src_addr = sg_dma_address(sg);
> +			dsg->dst_addr = slave_addr;
> +		} else { /* DMA_DEV_TO_MEM */
> +			dsg->src_addr = slave_addr;
> +			dsg->dst_addr = sg_dma_address(sg);
> +		}
> +		break;
> +	}
> +
> +	return vchan_tx_prep(&s3cchan->vc, &txd->vd, flags);
> +}
> +
> +/*
> + * Slave transactions callback to the slave device to allow
> + * synchronization of slave DMA signals with the DMAC enable
> + */
> +static void s3c24xx_dma_issue_pending(struct dma_chan *chan)
> +{
> +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> +	unsigned long flags;
> +
> +	spin_lock_irqsave(&s3cchan->vc.lock, flags);
> +	if (vchan_issue_pending(&s3cchan->vc)) {
> +		if (!s3cchan->phy && s3cchan->state != S3C24XX_DMA_CHAN_WAITING)
> +			s3c24xx_dma_phy_alloc_and_start(s3cchan);
> +	}
> +	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> +}
> +
> +/*
> + * Bringup and teardown
> + */
> +
> +/*
> + * Initialise the DMAC memcpy/slave channels.
> + * Make a local wrapper to hold required data
> + */
> +static int s3c24xx_dma_init_virtual_channels(struct s3c24xx_dma_engine
> *s3cdma, +		struct dma_device *dmadev, unsigned int channels, bool 
slave)
> +{
> +	struct s3c24xx_dma_chan *chan;
> +	int i;
> +
> +	INIT_LIST_HEAD(&dmadev->channels);
> +
> +	/*
> +	 * Register as many many memcpy as we have physical channels,
> +	 * we won't always be able to use all but the code will have
> +	 * to cope with that situation.
> +	 */
> +	for (i = 0; i < channels; i++) {
> +		chan = devm_kzalloc(dmadev->dev, sizeof(*chan), GFP_KERNEL);
> +		if (!chan) {
> +			dev_err(dmadev->dev,
> +				"%s no memory for channel\n", __func__);
> +			return -ENOMEM;
> +		}
> +
> +		chan->id = i;
> +		chan->host = s3cdma;
> +		chan->state = S3C24XX_DMA_CHAN_IDLE;
> +
> +		if (slave) {
> +			chan->slave = true;
> +			chan->name = kasprintf(GFP_KERNEL, "slave%d", i);
> +			if (!chan->name)
> +				return -ENOMEM;
> +		} else {
> +			chan->name = kasprintf(GFP_KERNEL, "memcpy%d", i);
> +			if (!chan->name)
> +				return -ENOMEM;
> +		}
> +		dev_dbg(dmadev->dev,
> +			 "initialize virtual channel \"%s\"\n",
> +			 chan->name);
> +
> +		chan->vc.desc_free = s3c24xx_dma_desc_free;
> +		vchan_init(&chan->vc, dmadev);
> +	}
> +	dev_info(dmadev->dev, "initialized %d virtual %s channels\n",
> +		 i, slave ? "slave" : "memcpy");
> +	return i;
> +}
> +
> +static void s3c24xx_dma_free_virtual_channels(struct dma_device *dmadev)
> +{
> +	struct s3c24xx_dma_chan *chan = NULL;
> +	struct s3c24xx_dma_chan *next;
> +
> +	list_for_each_entry_safe(chan,
> +				 next, &dmadev->channels, vc.chan.device_node)
> +		list_del(&chan->vc.chan.device_node);
> +}
> +
> +/* s3c2412 and s3c2413 have a 0x40 stride and dmareqsel mechanism */
> +static struct soc_data soc_s3c2412 = {
> +	.stride = 0x40,
> +	.has_reqsel = true,
> +	.has_clocks = true,
> +};
> +
> +/* s3c2443 and following have a 0x100 stride and dmareqsel mechanism */
> +static struct soc_data soc_s3c2443 = {
> +	.stride = 0x100,
> +	.has_reqsel = true,
> +	.has_clocks = true,
> +};
> +
> +static struct platform_device_id s3c24xx_dma_driver_ids[] = {
> +	{
> +		.name		= "s3c2412-dma",
> +		.driver_data	= (kernel_ulong_t)&soc_s3c2412,
> +	}, {
> +		.name		= "s3c2443-dma",
> +		.driver_data	= (kernel_ulong_t)&soc_s3c2443,
> +	},
> +	{ },
> +};
> +
> +static struct soc_data *s3c24xx_dma_get_soc_data(struct platform_device
> *pdev) +{
> +	return (struct soc_data *)
> +			 platform_get_device_id(pdev)->driver_data;
> +}
> +
> +static int s3c24xx_dma_probe(struct platform_device *pdev)
> +{
> +	const struct s3c24xx_dma_platdata *pdata = dev_get_platdata(&pdev->dev);
> +	struct s3c24xx_dma_engine *s3cdma;
> +	struct soc_data *sdata;
> +	struct resource *res;
> +	int ret;
> +	int i;
> +
> +	if (!pdata) {
> +		dev_err(&pdev->dev, "platform data missing\n");
> +		return -ENODEV;
> +	}
> +
> +	/* Basic sanity check */
> +	if (pdata->num_phy_channels > MAX_DMA_CHANNELS) {
> +		dev_err(&pdev->dev, "to many dma channels %d, max %d\n",
> +			pdata->num_phy_channels, MAX_DMA_CHANNELS);
> +		return -EINVAL;
> +	}
> +
> +	sdata = s3c24xx_dma_get_soc_data(pdev);
> +	if (!sdata)
> +		return -EINVAL;
> +
> +	s3cdma = devm_kzalloc(&pdev->dev, sizeof(*s3cdma), GFP_KERNEL);
> +	if (!s3cdma)
> +		return -ENOMEM;
> +
> +	s3cdma->pdev = pdev;
> +	s3cdma->pdata = pdata;
> +	s3cdma->sdata = sdata;
> +
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	s3cdma->base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(s3cdma->base))
> +		return PTR_ERR(s3cdma->base);
> +
> +	s3cdma->phy_chans = devm_kzalloc(&pdev->dev,
> +					      sizeof(struct s3c24xx_dma_phy) *
> +							pdata->num_phy_channels,
> +					      GFP_KERNEL);
> +	if (!s3cdma->phy_chans)
> +		return -ENOMEM;
> +
> +	/* aquire irqs and clocks for all physical channels */
> +	for (i = 0; i < pdata->num_phy_channels; i++) {
> +		struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
> +		char clk_name[6];
> +
> +		phy->id = i;
> +		phy->base = s3cdma->base + (i * sdata->stride);
> +		phy->host = s3cdma;
> +
> +		phy->irq = platform_get_irq(pdev, i);
> +		if (phy->irq < 0) {
> +			dev_err(&pdev->dev, "failed to get irq %d, err %d\n",
> +				i, phy->irq);
> +			continue;
> +		}
> +
> +		ret = devm_request_irq(&pdev->dev, phy->irq, s3c24xx_dma_irq,
> +				       0, pdev->name, phy);
> +		if (ret) {
> +			dev_err(&pdev->dev, "Unable to request irq for channel %d, error 
%d\n",
> +				i, ret);
> +			continue;
> +		}
> +
> +		if (sdata->has_clocks) {
> +			sprintf(clk_name, "dma.%d", i);
> +			phy->clk = devm_clk_get(&pdev->dev, clk_name);
> +			if (IS_ERR(phy->clk) && sdata->has_clocks) {
> +				dev_err(&pdev->dev, "unable to aquire clock for channel %d, 
error
> %lu", +					i, PTR_ERR(phy->clk));
> +				continue;
> +			}
> +
> +			ret = clk_prepare(phy->clk);
> +			if (ret) {
> +				dev_err(&pdev->dev, "clock for phy %d failed, error %d\n",
> +					i, ret);
> +				continue;
> +			}
> +		}
> +
> +		spin_lock_init(&phy->lock);
> +		phy->valid = true;
> +
> +		dev_dbg(&pdev->dev, "physical channel %d is %s\n",
> +			i, s3c24xx_dma_phy_busy(phy) ? "BUSY" : "FREE");
> +	}
> +
> +	/* Initialize memcpy engine */
> +	dma_cap_set(DMA_MEMCPY, s3cdma->memcpy.cap_mask);
> +	dma_cap_set(DMA_PRIVATE, s3cdma->memcpy.cap_mask);
> +	s3cdma->memcpy.dev = &pdev->dev;
> +	s3cdma->memcpy.device_alloc_chan_resources =
> +					s3c24xx_dma_alloc_chan_resources;
> +	s3cdma->memcpy.device_free_chan_resources =
> +					s3c24xx_dma_free_chan_resources;
> +	s3cdma->memcpy.device_prep_dma_memcpy = s3c24xx_dma_prep_memcpy;
> +	s3cdma->memcpy.device_tx_status = s3c24xx_dma_tx_status;
> +	s3cdma->memcpy.device_issue_pending = s3c24xx_dma_issue_pending;
> +	s3cdma->memcpy.device_control = s3c24xx_dma_control;
> +
> +	/* Initialize slave engine for SoC internal dedicated peripherals */
> +	dma_cap_set(DMA_SLAVE, s3cdma->slave.cap_mask);
> +	dma_cap_set(DMA_PRIVATE, s3cdma->slave.cap_mask);
> +	s3cdma->slave.dev = &pdev->dev;
> +	s3cdma->slave.device_alloc_chan_resources =
> +					s3c24xx_dma_alloc_chan_resources;
> +	s3cdma->slave.device_free_chan_resources =
> +					s3c24xx_dma_free_chan_resources;
> +	s3cdma->slave.device_tx_status = s3c24xx_dma_tx_status;
> +	s3cdma->slave.device_issue_pending = s3c24xx_dma_issue_pending;
> +	s3cdma->slave.device_prep_slave_sg = s3c24xx_dma_prep_slave_sg;
> +	s3cdma->slave.device_control = s3c24xx_dma_control;
> +
> +	/* Register as many memcpy channels as there are physical channels */
> +	ret = s3c24xx_dma_init_virtual_channels(s3cdma, &s3cdma->memcpy,
> +						pdata->num_phy_channels, false);
> +	if (ret <= 0) {
> +		dev_warn(&pdev->dev,
> +			 "%s failed to enumerate memcpy channels - %d\n",
> +			 __func__, ret);
> +		goto err_memcpy;
> +	}
> +
> +	/* Register slave channels */
> +	ret = s3c24xx_dma_init_virtual_channels(s3cdma, &s3cdma->slave,
> +				pdata->num_channels, true);
> +	if (ret <= 0) {
> +		dev_warn(&pdev->dev,
> +			"%s failed to enumerate slave channels - %d\n",
> +				__func__, ret);
> +		goto err_slave;
> +	}
> +
> +	ret = dma_async_device_register(&s3cdma->memcpy);
> +	if (ret) {
> +		dev_warn(&pdev->dev,
> +			"%s failed to register memcpy as an async device - %d\n",
> +			__func__, ret);
> +		goto err_memcpy_reg;
> +	}
> +
> +	ret = dma_async_device_register(&s3cdma->slave);
> +	if (ret) {
> +		dev_warn(&pdev->dev,
> +			"%s failed to register slave as an async device - %d\n",
> +			__func__, ret);
> +		goto err_slave_reg;
> +	}
> +
> +	platform_set_drvdata(pdev, s3cdma);
> +	dev_info(&pdev->dev, "Loaded dma driver with %d physical channels\n",
> +		 pdata->num_phy_channels);
> +
> +	return 0;
> +
> +err_slave_reg:
> +	dma_async_device_unregister(&s3cdma->memcpy);
> +err_memcpy_reg:
> +	s3c24xx_dma_free_virtual_channels(&s3cdma->slave);
> +err_slave:
> +	s3c24xx_dma_free_virtual_channels(&s3cdma->memcpy);
> +err_memcpy:
> +	if (sdata->has_clocks)
> +		for (i = 0; i < pdata->num_phy_channels; i++) {
> +			struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
> +			if (phy->valid)
> +				clk_unprepare(phy->clk);
> +		}
> +
> +	return ret;
> +}
> +
> +static int s3c24xx_dma_remove(struct platform_device *pdev)
> +{
> +	const struct s3c24xx_dma_platdata *pdata = dev_get_platdata(&pdev->dev);
> +	struct s3c24xx_dma_engine *s3cdma = platform_get_drvdata(pdev);
> +	struct soc_data *sdata = s3c24xx_dma_get_soc_data(pdev);
> +	int i;
> +
> +	dma_async_device_unregister(&s3cdma->slave);
> +	dma_async_device_unregister(&s3cdma->memcpy);
> +
> +	s3c24xx_dma_free_virtual_channels(&s3cdma->slave);
> +	s3c24xx_dma_free_virtual_channels(&s3cdma->memcpy);
> +
> +	if (sdata->has_clocks)
> +		for (i = 0; i < pdata->num_phy_channels; i++) {
> +			struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
> +			if (phy->valid)
> +				clk_unprepare(phy->clk);
> +		}
> +
> +	return 0;
> +}
> +
> +static struct platform_driver s3c24xx_dma_driver = {
> +	.driver		= {
> +		.name	= "s3c24xx-dma",
> +		.owner	= THIS_MODULE,
> +	},
> +	.id_table	= s3c24xx_dma_driver_ids,
> +	.probe		= s3c24xx_dma_probe,
> +	.remove		= s3c24xx_dma_remove,
> +};
> +
> +module_platform_driver(s3c24xx_dma_driver);
> +
> +bool s3c24xx_dma_filter(struct dma_chan *chan, void *param)
> +{
> +	struct s3c24xx_dma_chan *s3cchan;
> +
> +	if (chan->device->dev->driver != &s3c24xx_dma_driver.driver)
> +		return false;
> +
> +	s3cchan = to_s3c24xx_dma_chan(chan);
> +
> +	return s3cchan->id == (int)param;
> +}
> +EXPORT_SYMBOL(s3c24xx_dma_filter);
> +
> +MODULE_DESCRIPTION("S3C24XX DMA Driver");
> +MODULE_AUTHOR("Heiko Stuebner");
> +MODULE_LICENSE("GPL v2");
> diff --git a/include/linux/platform_data/dma-s3c24xx.h
> b/include/linux/platform_data/dma-s3c24xx.h new file mode 100644
> index 0000000..5a0cfff
> --- /dev/null
> +++ b/include/linux/platform_data/dma-s3c24xx.h
> @@ -0,0 +1,43 @@
> +/*
> + * S3C24XX DMA handling
> + *
> + * Copyright (c) 2013 Heiko Stuebner <heiko@sntech.de>
> + *
> + * 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.
> + */
> +
> +enum s3c24xx_dma_bus {
> +	S3C24XX_DMA_APB,
> +	S3C24XX_DMA_AHB,
> +};
> +
> +/**
> + * @bus: on which bus does the peripheral reside - AHB or APB.
> + * @handshake: is a handshake with the peripheral necessary
> + * @chansel: channel selection information, depending on variant; reqsel
> for + *	     s3c2443 and later and channel-selection map for earlier SoCs
> + *	     see CHANSEL doc in s3c2443-dma.c
> + */
> +struct s3c24xx_dma_channel {
> +	enum s3c24xx_dma_bus bus;
> +	bool handshake;
> +	u16 chansel;
> +};
> +
> +/**
> + * struct s3c24xx_dma_platdata - platform specific settings
> + * @num_phy_channels: number of physical channels
> + * @channels: array of virtual channel descriptions
> + * @num_channels: number of virtual channels
> + */
> +struct s3c24xx_dma_platdata {
> +	int num_phy_channels;
> +	struct s3c24xx_dma_channel *channels;
> +	int num_channels;
> +};
> +
> +struct dma_chan;
> +bool s3c24xx_dma_filter(struct dma_chan *chan, void *param);

RE: [PATCH v5 2/4] dmaengine: add driver for Samsung s3c24xx SoCs

[Kukjin Kim]

Heiko Stübner wrote:
> 
> Hi Vinod,
> 
> with the removed double-retrival of the cookie state, is the driver ok now?
> Because I think kgene needs an ack on it.
> 
Yes! Vinod, do you OK on this? If you have any concerns, please kindly let us know.

Thanks,
Kukjin

> 
> Thanks
> Heiko
> 
> 
> Am Mittwoch, 18. September 2013, 21:19:54 schrieb Heiko Stübner:
> > This adds a new driver to support the s3c24xx dma using the dmaengine
> > and makes the old one in mach-s3c24xx obsolete in the long run.
> >
> > Conceptually the s3c24xx-dma feels like a distant relative of the pl08x
> > with numerous virtual channels being mapped to a lot less physical ones.
> > The driver therefore borrows a lot from the amba-pl08x driver in this
> > regard. Functionality-wise the driver gains a memcpy ability in addition
> > to the slave_sg one.
> >
> > The driver supports both the method for requesting the peripheral used
> > by SoCs before the S3C2443 and the different method for S3C2443 and
> later.
> >
> > On earlier SoCs the hardware channels usable for specific peripherals is
> > constrainted while on later SoCs all channels can be used for any
> > peripheral.
> >
> > Tested on a s3c2416-based board, memcpy using the dmatest module and
> > slave_sg partially using the spi-s3c64xx driver.
> >
> > Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> > Acked-by: Linus Walleij <linus.walleij@linaro.org>
> > ---
> > changes since v4:
> > - fix the unnecessary double retrieval of the cookie status
> >   pointed out by Vinod Koul
> > changes since v3:
> > - address comments from Tomasz Figa:
> >   - fixed a lot of nitpics
> >   - and added a rudimentary mechanism to select higher bus-widths for
> > memcpy - address more comments from Vinod Koul - reorder dma_control
> > contents
> >
> > changes since v2:
> > - add missing channel validation in s3c24xx_dma_phy_free when
> repurposing
> >   a physical channel for another virtual channel
> > - address comments from Vinod Koul:
> >   - fix naming of constants to prevent conflicts
> >   - better handle the unlikely case of an interrupt happening on an
> >     unused channel
> >
> > changes since v1:
> > - address comments from Linus Walleij
> > - support sg-lists with more than 1 element
> > - add support for earlier s3c24xx SoCs
> > - redo the channeldata definitions:
> >   - no need for static ids for the virtual channels
> >   - provide a way to encode constraints for earlier variants
> >   - better base for later dt bindings
> >
> >  drivers/dma/Kconfig                       |   12 +
> >  drivers/dma/Makefile                      |    1 +
> >  drivers/dma/s3c24xx-dma.c                 | 1340
> > +++++++++++++++++++++++++++++ include/linux/platform_data/dma-s3c24xx.h
> |
> >  43 +
> >  4 files changed, 1396 insertions(+)
> >  create mode 100644 drivers/dma/s3c24xx-dma.c
> >  create mode 100644 include/linux/platform_data/dma-s3c24xx.h
> >
> > diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
> > index 526ec77..d639115 100644
> > --- a/drivers/dma/Kconfig
> > +++ b/drivers/dma/Kconfig
> > @@ -154,6 +154,18 @@ config TEGRA20_APB_DMA
> >  	  This DMA controller transfers data from memory to peripheral fifo
> >  	  or vice versa. It does not support memory to memory data transfer.
> >
> > +config S3C24XX_DMAC
> > +	tristate "Samsung S3C24XX DMA support"
> > +	depends on ARCH_S3C24XX && !S3C24XX_DMA
> > +	select DMA_ENGINE
> > +	select DMA_VIRTUAL_CHANNELS
> > +	help
> > +	  Support for the Samsung S3C24XX DMA controller driver. The
> > +	  DMA controller is having multiple DMA channels which can be
> > +	  configured for different peripherals like audio, UART, SPI.
> > +	  The DMA controller can transfer data from memory to peripheral,
> > +	  periphal to memory, periphal to periphal and memory to memory.
> > +
> >  source "drivers/dma/sh/Kconfig"
> >
> >  config COH901318
> > diff --git a/drivers/dma/Makefile b/drivers/dma/Makefile
> > index db89035..0ce2da9 100644
> > --- a/drivers/dma/Makefile
> > +++ b/drivers/dma/Makefile
> > @@ -30,6 +30,7 @@ obj-$(CONFIG_SIRF_DMA) += sirf-dma.o
> >  obj-$(CONFIG_TI_EDMA) += edma.o
> >  obj-$(CONFIG_STE_DMA40) += ste_dma40.o ste_dma40_ll.o
> >  obj-$(CONFIG_TEGRA20_APB_DMA) += tegra20-apb-dma.o
> > +obj-$(CONFIG_S3C24XX_DMAC) += s3c24xx-dma.o
> >  obj-$(CONFIG_PL330_DMA) += pl330.o
> >  obj-$(CONFIG_PCH_DMA) += pch_dma.o
> >  obj-$(CONFIG_AMBA_PL08X) += amba-pl08x.o
> > diff --git a/drivers/dma/s3c24xx-dma.c b/drivers/dma/s3c24xx-dma.c
> > new file mode 100644
> > index 0000000..56c9253
> > --- /dev/null
> > +++ b/drivers/dma/s3c24xx-dma.c
> > @@ -0,0 +1,1340 @@
> > +/*
> > + * S3C24XX DMA handling
> > + *
> > + * Copyright (c) 2013 Heiko Stuebner <heiko@sntech.de>
> > + *
> > + * based on amba-pl08x.c
> > + *
> > + * Copyright (c) 2006 ARM Ltd.
> > + * Copyright (c) 2010 ST-Ericsson SA
> > + *
> > + * Author: Peter Pearse <peter.pearse@arm.com>
> > + * Author: Linus Walleij <linus.walleij@stericsson.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.
> > + *
> > + * The DMA controllers in S3C24XX SoCs have a varying number of DMA
> > signals + * that can be routed to any of the 4 to 8 hardware-channels.
> > + *
> > + * Therefore on these DMA controllers the number of channels
> > + * and the number of incoming DMA signals are two totally different
> > things. + * It is usually not possible to theoretically handle all
> > physical signals, + * so a multiplexing scheme with possible denial of
> use
> > is necessary. + *
> > + * Open items:
> > + * - bursts
> > + */
> > +
> > +#include <linux/platform_device.h>
> > +#include <linux/types.h>
> > +#include <linux/dmaengine.h>
> > +#include <linux/dma-mapping.h>
> > +#include <linux/interrupt.h>
> > +#include <linux/clk.h>
> > +#include <linux/module.h>
> > +#include <linux/slab.h>
> > +#include <linux/platform_data/dma-s3c24xx.h>
> > +
> > +#include "dmaengine.h"
> > +#include "virt-dma.h"
> > +
> > +#define MAX_DMA_CHANNELS	8
> > +
> > +#define S3C24XX_DISRC			0x00
> > +#define S3C24XX_DISRCC			0x04
> > +#define S3C24XX_DISRCC_INC_INCREMENT	0
> > +#define S3C24XX_DISRCC_INC_FIXED	BIT(0)
> > +#define S3C24XX_DISRCC_LOC_AHB		0
> > +#define S3C24XX_DISRCC_LOC_APB		BIT(1)
> > +
> > +#define S3C24XX_DIDST			0x08
> > +#define S3C24XX_DIDSTC			0x0c
> > +#define S3C24XX_DIDSTC_INC_INCREMENT	0
> > +#define S3C24XX_DIDSTC_INC_FIXED	BIT(0)
> > +#define S3C24XX_DIDSTC_LOC_AHB		0
> > +#define S3C24XX_DIDSTC_LOC_APB		BIT(1)
> > +#define S3C24XX_DIDSTC_INT_TC0		0
> > +#define S3C24XX_DIDSTC_INT_RELOAD	BIT(2)
> > +
> > +#define S3C24XX_DCON			0x10
> > +
> > +#define S3C24XX_DCON_TC_MASK		0xfffff
> > +#define S3C24XX_DCON_DSZ_BYTE		(0 << 20)
> > +#define S3C24XX_DCON_DSZ_HALFWORD	(1 << 20)
> > +#define S3C24XX_DCON_DSZ_WORD		(2 << 20)
> > +#define S3C24XX_DCON_DSZ_MASK		(3 << 20)
> > +#define S3C24XX_DCON_DSZ_SHIFT		20
> > +#define S3C24XX_DCON_AUTORELOAD		0
> > +#define S3C24XX_DCON_NORELOAD		BIT(22)
> > +#define S3C24XX_DCON_HWTRIG		BIT(23)
> > +#define S3C24XX_DCON_HWSRC_SHIFT	24
> > +#define S3C24XX_DCON_SERV_SINGLE	0
> > +#define S3C24XX_DCON_SERV_WHOLE		BIT(27)
> > +#define S3C24XX_DCON_TSZ_UNIT		0
> > +#define S3C24XX_DCON_TSZ_BURST4		BIT(28)
> > +#define S3C24XX_DCON_INT		BIT(29)
> > +#define S3C24XX_DCON_SYNC_PCLK		0
> > +#define S3C24XX_DCON_SYNC_HCLK		BIT(30)
> > +#define S3C24XX_DCON_DEMAND		0
> > +#define S3C24XX_DCON_HANDSHAKE		BIT(31)
> > +
> > +#define S3C24XX_DSTAT			0x14
> > +#define S3C24XX_DSTAT_STAT_BUSY		BIT(20)
> > +#define S3C24XX_DSTAT_CURRTC_MASK	0xfffff
> > +
> > +#define S3C24XX_DMASKTRIG		0x20
> > +#define S3C24XX_DMASKTRIG_SWTRIG	BIT(0)
> > +#define S3C24XX_DMASKTRIG_ON		BIT(1)
> > +#define S3C24XX_DMASKTRIG_STOP		BIT(2)
> > +
> > +#define S3C24XX_DMAREQSEL		0x24
> > +#define S3C24XX_DMAREQSEL_HW		BIT(0)
> > +
> > +/*
> > + * S3C2410, S3C2440 and S3C2442 SoCs cannot select any physical channel
> > + * for a DMA source. Instead only specific channels are valid.
> > + * All of these SoCs have 4 physical channels and the number of request
> > + * source bits is 3. Additionally we also need 1 bit to mark the
> channel
> > + * as valid.
> > + * Therefore we separate the chansel element of the channel data into 4
> > + * parts of 4 bits each, to hold the information if the channel is
> valid
> > + * and the hw request source to use.
> > + *
> > + * Example:
> > + * SDI is valid on channels 0, 2 and 3 - with varying hw request
> sources.
> > + * For it the chansel field would look like
> > + *
> > + * ((BIT(3) | 1) << 3 * 4) | // channel 3, with request source 1
> > + * ((BIT(3) | 2) << 2 * 4) | // channel 2, with request source 2
> > + * ((BIT(3) | 2) << 0 * 4)   // channel 0, with request source 2
> > + */
> > +#define S3C24XX_CHANSEL_WIDTH		4
> > +#define S3C24XX_CHANSEL_VALID		BIT(3)
> > +#define S3C24XX_CHANSEL_REQ_MASK	7
> > +
> > +/*
> > + * struct soc_data - vendor-specific config parameters for individual
> SoCs
> > + * @stride: spacing between the registers of each channel
> > + * @has_reqsel: does the controller use the newer requestselection
> > mechanism + * @has_clocks: are controllable dma-clocks present
> > + */
> > +struct soc_data {
> > +	int stride;
> > +	bool has_reqsel;
> > +	bool has_clocks;
> > +};
> > +
> > +/*
> > + * enum s3c24xx_dma_chan_state - holds the virtual channel states
> > + * @S3C24XX_DMA_CHAN_IDLE: the channel is idle
> > + * @S3C24XX_DMA_CHAN_RUNNING: the channel has allocated a physical
> > transport + * channel and is running a transfer on it
> > + * @S3C24XX_DMA_CHAN_WAITING: the channel is waiting for a physical
> > transport + * channel to become available (only pertains to memcpy
> > channels) + */
> > +enum s3c24xx_dma_chan_state {
> > +	S3C24XX_DMA_CHAN_IDLE,
> > +	S3C24XX_DMA_CHAN_RUNNING,
> > +	S3C24XX_DMA_CHAN_WAITING,
> > +};
> > +
> > +/*
> > + * struct s3c24xx_sg - structure containing data per sg
> > + * @src_addr: src address of sg
> > + * @dst_addr: dst address of sg
> > + * @len: transfer len in bytes
> > + * @node: node for txd's dsg_list
> > + */
> > +struct s3c24xx_sg {
> > +	dma_addr_t src_addr;
> > +	dma_addr_t dst_addr;
> > +	size_t len;
> > +	struct list_head node;
> > +};
> > +
> > +/*
> > + * struct s3c24xx_txd - wrapper for struct dma_async_tx_descriptor
> > + * @vd: virtual DMA descriptor
> > + * @dsg_list: list of children sg's
> > + * @at: sg currently being transfered
> > + * @width: transfer width
> > + * @disrcc: value for source control register
> > + * @didstc: value for destination control register
> > + * @dcon: base value for dcon register
> > + */
> > +struct s3c24xx_txd {
> > +	struct virt_dma_desc vd;
> > +	struct list_head dsg_list;
> > +	struct list_head *at;
> > +	u8 width;
> > +	u32 disrcc;
> > +	u32 didstc;
> > +	u32 dcon;
> > +};
> > +
> > +struct s3c24xx_dma_chan;
> > +
> > +/*
> > + * struct s3c24xx_dma_phy - holder for the physical channels
> > + * @id: physical index to this channel
> > + * @valid: does the channel have all required elements
> > + * @base: virtual memory base (remapped) for the this channel
> > + * @irq: interrupt for this channel
> > + * @clk: clock for this channel
> > + * @lock: a lock to use when altering an instance of this struct
> > + * @serving: virtual channel currently being served by this
> > physicalchannel + * @host: a pointer to the host (internal use)
> > + */
> > +struct s3c24xx_dma_phy {
> > +	unsigned int			id;
> > +	bool				valid;
> > +	void __iomem			*base;
> > +	unsigned int			irq;
> > +	struct clk			*clk;
> > +	spinlock_t			lock;
> > +	struct s3c24xx_dma_chan		*serving;
> > +	struct s3c24xx_dma_engine	*host;
> > +};
> > +
> > +/*
> > + * struct s3c24xx_dma_chan - this structure wraps a DMA ENGINE channel
> > + * @id: the id of the channel
> > + * @name: name of the channel
> > + * @vc: wrappped virtual channel
> > + * @phy: the physical channel utilized by this channel, if there is one
> > + * @runtime_addr: address for RX/TX according to the runtime config
> > + * @at: active transaction on this channel
> > + * @lock: a lock for this channel data
> > + * @host: a pointer to the host (internal use)
> > + * @state: whether the channel is idle, running etc
> > + * @slave: whether this channel is a device (slave) or for memcpy
> > + */
> > +struct s3c24xx_dma_chan {
> > +	int id;
> > +	const char *name;
> > +	struct virt_dma_chan vc;
> > +	struct s3c24xx_dma_phy *phy;
> > +	struct dma_slave_config cfg;
> > +	struct s3c24xx_txd *at;
> > +	struct s3c24xx_dma_engine *host;
> > +	enum s3c24xx_dma_chan_state state;
> > +	bool slave;
> > +};
> > +
> > +/*
> > + * struct s3c24xx_dma_engine - the local state holder for the S3C24XX
> > + * @pdev: the corresponding platform device
> > + * @pdata: platform data passed in from the platform/machine
> > + * @base: virtual memory base (remapped)
> > + * @slave: slave engine for this instance
> > + * @memcpy: memcpy engine for this instance
> > + * @phy_chans: array of data for the physical channels
> > + */
> > +struct s3c24xx_dma_engine {
> > +	struct platform_device			*pdev;
> > +	const struct s3c24xx_dma_platdata	*pdata;
> > +	struct soc_data				*sdata;
> > +	void __iomem				*base;
> > +	struct dma_device			slave;
> > +	struct dma_device			memcpy;
> > +	struct s3c24xx_dma_phy			*phy_chans;
> > +};
> > +
> > +/*
> > + * Physical channel handling
> > + */
> > +
> > +/*
> > + * Check whether a certain channel is busy or not.
> > + */
> > +static int s3c24xx_dma_phy_busy(struct s3c24xx_dma_phy *phy)
> > +{
> > +	unsigned int val = readl(phy->base + S3C24XX_DSTAT);
> > +	return val & S3C24XX_DSTAT_STAT_BUSY;
> > +}
> > +
> > +static bool s3c24xx_dma_phy_valid(struct s3c24xx_dma_chan *s3cchan,
> > +				  struct s3c24xx_dma_phy *phy)
> > +{
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
> > +	struct s3c24xx_dma_channel *cdata = &pdata->channels[s3cchan->id];
> > +	int phyvalid;
> > +
> > +	/* every phy is valid for memcopy channels */
> > +	if (!s3cchan->slave)
> > +		return true;
> > +
> > +	/* On newer variants all phys can be used for all virtual channels
> */
> > +	if (s3cdma->sdata->has_reqsel)
> > +		return true;
> > +
> > +	phyvalid = (cdata->chansel >> (phy->id * S3C24XX_CHANSEL_WIDTH));
> > +	return (phyvalid & S3C24XX_CHANSEL_VALID) ? true : false;
> > +}
> > +
> > +/*
> > + * Allocate a physical channel for a virtual channel
> > + *
> > + * Try to locate a physical channel to be used for this transfer. If
> all
> > + * are taken return NULL and the requester will have to cope by using
> > + * some fallback PIO mode or retrying later.
> > + */
> > +static
> > +struct s3c24xx_dma_phy *s3c24xx_dma_get_phy(struct s3c24xx_dma_chan
> > *s3cchan) +{
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
> > +	struct s3c24xx_dma_channel *cdata;
> > +	struct s3c24xx_dma_phy *phy = NULL;
> > +	unsigned long flags;
> > +	int i;
> > +	int ret;
> > +
> > +	if (s3cchan->slave)
> > +		cdata = &pdata->channels[s3cchan->id];
> > +
> > +	for (i = 0; i < s3cdma->pdata->num_phy_channels; i++) {
> > +		phy = &s3cdma->phy_chans[i];
> > +
> > +		if (!phy->valid)
> > +			continue;
> > +
> > +		if (!s3c24xx_dma_phy_valid(s3cchan, phy))
> > +			continue;
> > +
> > +		spin_lock_irqsave(&phy->lock, flags);
> > +
> > +		if (!phy->serving) {
> > +			phy->serving = s3cchan;
> > +			spin_unlock_irqrestore(&phy->lock, flags);
> > +			break;
> > +		}
> > +
> > +		spin_unlock_irqrestore(&phy->lock, flags);
> > +	}
> > +
> > +	/* No physical channel available, cope with it */
> > +	if (i == s3cdma->pdata->num_phy_channels) {
> > +		dev_warn(&s3cdma->pdev->dev, "no phy channel available\n");
> > +		return NULL;
> > +	}
> > +
> > +	/* start the phy clock */
> > +	if (s3cdma->sdata->has_clocks) {
> > +		ret = clk_enable(phy->clk);
> > +		if (ret) {
> > +			dev_err(&s3cdma->pdev->dev, "could not enable clock
> for channel
> %d, err
> > %d\n", +				phy->id, ret);
> > +			phy->serving = NULL;
> > +			return NULL;
> > +		}
> > +	}
> > +
> > +	return phy;
> > +}
> > +
> > +/*
> > + * Mark the physical channel as free.
> > + *
> > + * This drops the link between the physical and virtual channel.
> > + */
> > +static inline void s3c24xx_dma_put_phy(struct s3c24xx_dma_phy *phy)
> > +{
> > +	struct s3c24xx_dma_engine *s3cdma = phy->host;
> > +
> > +	if (s3cdma->sdata->has_clocks)
> > +		clk_disable(phy->clk);
> > +
> > +	phy->serving = NULL;
> > +}
> > +
> > +/*
> > + * Stops the channel by writing the stop bit.
> > + * This should not be used for an on-going transfer, but as a method of
> > + * shutting down a channel (eg, when it's no longer used) or
> terminating a
> > + * transfer.
> > + */
> > +static void s3c24xx_dma_terminate_phy(struct s3c24xx_dma_phy *phy)
> > +{
> > +	writel(S3C24XX_DMASKTRIG_STOP, phy->base + S3C24XX_DMASKTRIG);
> > +}
> > +
> > +/*
> > + * Virtual channel handling
> > + */
> > +
> > +static inline
> > +struct s3c24xx_dma_chan *to_s3c24xx_dma_chan(struct dma_chan *chan)
> > +{
> > +	return container_of(chan, struct s3c24xx_dma_chan, vc.chan);
> > +}
> > +
> > +static u32 s3c24xx_dma_getbytes_chan(struct s3c24xx_dma_chan *s3cchan)
> > +{
> > +	struct s3c24xx_dma_phy *phy = s3cchan->phy;
> > +	struct s3c24xx_txd *txd = s3cchan->at;
> > +	u32 tc = readl(phy->base + S3C24XX_DSTAT) &
> S3C24XX_DSTAT_CURRTC_MASK;
> > +
> > +	return tc * txd->width;
> > +}
> > +
> > +static int s3c24xx_dma_set_runtime_config(struct s3c24xx_dma_chan
> > *s3cchan, +				  struct dma_slave_config *config)
> > +{
> > +	if (!s3cchan->slave)
> > +		return -EINVAL;
> > +
> > +	/* Reject definitely invalid configurations */
> > +	if (config->src_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES ||
> > +	    config->dst_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES)
> > +		return -EINVAL;
> > +
> > +	s3cchan->cfg = *config;
> > +
> > +	return 0;
> > +}
> > +
> > +/*
> > + * Transfer handling
> > + */
> > +
> > +static inline
> > +struct s3c24xx_txd *to_s3c24xx_txd(struct dma_async_tx_descriptor *tx)
> > +{
> > +	return container_of(tx, struct s3c24xx_txd, vd.tx);
> > +}
> > +
> > +static struct s3c24xx_txd *s3c24xx_dma_get_txd(void)
> > +{
> > +	struct s3c24xx_txd *txd = kzalloc(sizeof(*txd), GFP_NOWAIT);
> > +
> > +	if (txd) {
> > +		INIT_LIST_HEAD(&txd->dsg_list);
> > +		txd->dcon = S3C24XX_DCON_INT | S3C24XX_DCON_NORELOAD;
> > +	}
> > +
> > +	return txd;
> > +}
> > +
> > +static void s3c24xx_dma_free_txd(struct s3c24xx_txd *txd)
> > +{
> > +	struct s3c24xx_sg *dsg, *_dsg;
> > +
> > +	list_for_each_entry_safe(dsg, _dsg, &txd->dsg_list, node) {
> > +		list_del(&dsg->node);
> > +		kfree(dsg);
> > +	}
> > +
> > +	kfree(txd);
> > +}
> > +
> > +static void s3c24xx_dma_start_next_sg(struct s3c24xx_dma_chan *s3cchan,
> > +				       struct s3c24xx_txd *txd)
> > +{
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +	struct s3c24xx_dma_phy *phy = s3cchan->phy;
> > +	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
> > +	struct s3c24xx_sg *dsg = list_entry(txd->at, struct s3c24xx_sg,
> node);
> > +	u32 dcon = txd->dcon;
> > +	u32 val;
> > +
> > +	/* transfer-size and -count from len and width */
> > +	switch (txd->width) {
> > +	case 1:
> > +		dcon |= S3C24XX_DCON_DSZ_BYTE | dsg->len;
> > +		break;
> > +	case 2:
> > +		dcon |= S3C24XX_DCON_DSZ_HALFWORD | (dsg->len / 2);
> > +		break;
> > +	case 4:
> > +		dcon |= S3C24XX_DCON_DSZ_WORD | (dsg->len / 4);
> > +		break;
> > +	}
> > +
> > +	if (s3cchan->slave) {
> > +		struct s3c24xx_dma_channel *cdata =
> > +					&pdata->channels[s3cchan->id];
> > +
> > +		if (s3cdma->sdata->has_reqsel) {
> > +			writel_relaxed((cdata->chansel << 1) |
> > +							S3C24XX_DMAREQSEL_HW,
> > +					phy->base + S3C24XX_DMAREQSEL);
> > +		} else {
> > +			int csel = cdata->chansel >> (phy->id *
> > +							S3C24XX_CHANSEL_WIDTH);
> > +
> > +			csel &= S3C24XX_CHANSEL_REQ_MASK;
> > +			dcon |= csel << S3C24XX_DCON_HWSRC_SHIFT;
> > +			dcon |= S3C24XX_DCON_HWTRIG;
> > +		}
> > +	} else {
> > +		if (s3cdma->sdata->has_reqsel)
> > +			writel_relaxed(0, phy->base + S3C24XX_DMAREQSEL);
> > +	}
> > +
> > +	writel_relaxed(dsg->src_addr, phy->base + S3C24XX_DISRC);
> > +	writel_relaxed(txd->disrcc, phy->base + S3C24XX_DISRCC);
> > +	writel_relaxed(dsg->dst_addr, phy->base + S3C24XX_DIDST);
> > +	writel_relaxed(txd->didstc, phy->base + S3C24XX_DIDSTC);
> > +	writel_relaxed(dcon, phy->base + S3C24XX_DCON);
> > +
> > +	val = readl_relaxed(phy->base + S3C24XX_DMASKTRIG);
> > +	val &= ~S3C24XX_DMASKTRIG_STOP;
> > +	val |= S3C24XX_DMASKTRIG_ON;
> > +
> > +	/* trigger the dma operation for memcpy transfers */
> > +	if (!s3cchan->slave)
> > +		val |= S3C24XX_DMASKTRIG_SWTRIG;
> > +
> > +	writel(val, phy->base + S3C24XX_DMASKTRIG);
> > +}
> > +
> > +/*
> > + * Set the initial DMA register values and start first sg.
> > + */
> > +static void s3c24xx_dma_start_next_txd(struct s3c24xx_dma_chan *s3cchan)
> > +{
> > +	struct s3c24xx_dma_phy *phy = s3cchan->phy;
> > +	struct virt_dma_desc *vd = vchan_next_desc(&s3cchan->vc);
> > +	struct s3c24xx_txd *txd = to_s3c24xx_txd(&vd->tx);
> > +
> > +	list_del(&txd->vd.node);
> > +
> > +	s3cchan->at = txd;
> > +
> > +	/* Wait for channel inactive */
> > +	while (s3c24xx_dma_phy_busy(phy))
> > +		cpu_relax();
> > +
> > +	/* point to the first element of the sg list */
> > +	txd->at = txd->dsg_list.next;
> > +	s3c24xx_dma_start_next_sg(s3cchan, txd);
> > +}
> > +
> > +static void s3c24xx_dma_free_txd_list(struct s3c24xx_dma_engine *s3cdma,
> > +				struct s3c24xx_dma_chan *s3cchan)
> > +{
> > +	LIST_HEAD(head);
> > +
> > +	vchan_get_all_descriptors(&s3cchan->vc, &head);
> > +	vchan_dma_desc_free_list(&s3cchan->vc, &head);
> > +}
> > +
> > +/*
> > + * Try to allocate a physical channel.  When successful, assign it to
> > + * this virtual channel, and initiate the next descriptor.  The
> > + * virtual channel lock must be held at this point.
> > + */
> > +static void s3c24xx_dma_phy_alloc_and_start(struct s3c24xx_dma_chan
> > *s3cchan) +{
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +	struct s3c24xx_dma_phy *phy;
> > +
> > +	phy = s3c24xx_dma_get_phy(s3cchan);
> > +	if (!phy) {
> > +		dev_dbg(&s3cdma->pdev->dev, "no physical channel available
> for xfer
> on
> > %s\n", +			s3cchan->name);
> > +		s3cchan->state = S3C24XX_DMA_CHAN_WAITING;
> > +		return;
> > +	}
> > +
> > +	dev_dbg(&s3cdma->pdev->dev, "allocated physical channel %d for xfer
> on
> > %s\n", +		phy->id, s3cchan->name);
> > +
> > +	s3cchan->phy = phy;
> > +	s3cchan->state = S3C24XX_DMA_CHAN_RUNNING;
> > +
> > +	s3c24xx_dma_start_next_txd(s3cchan);
> > +}
> > +
> > +static void s3c24xx_dma_phy_reassign_start(struct s3c24xx_dma_phy *phy,
> > +	struct s3c24xx_dma_chan *s3cchan)
> > +{
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +
> > +	dev_dbg(&s3cdma->pdev->dev, "reassigned physical channel %d for
> xfer on
> > %s\n", +		phy->id, s3cchan->name);
> > +
> > +	/*
> > +	 * We do this without taking the lock; we're really only concerned
> > +	 * about whether this pointer is NULL or not, and we're guaranteed
> > +	 * that this will only be called when it _already_ is non-NULL.
> > +	 */
> > +	phy->serving = s3cchan;
> > +	s3cchan->phy = phy;
> > +	s3cchan->state = S3C24XX_DMA_CHAN_RUNNING;
> > +	s3c24xx_dma_start_next_txd(s3cchan);
> > +}
> > +
> > +/*
> > + * Free a physical DMA channel, potentially reallocating it to another
> > + * virtual channel if we have any pending.
> > + */
> > +static void s3c24xx_dma_phy_free(struct s3c24xx_dma_chan *s3cchan)
> > +{
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +	struct s3c24xx_dma_chan *p, *next;
> > +
> > +retry:
> > +	next = NULL;
> > +
> > +	/* Find a waiting virtual channel for the next transfer. */
> > +	list_for_each_entry(p, &s3cdma->memcpy.channels,
> vc.chan.device_node)
> > +		if (p->state == S3C24XX_DMA_CHAN_WAITING) {
> > +			next = p;
> > +			break;
> > +		}
> > +
> > +	if (!next) {
> > +		list_for_each_entry(p, &s3cdma->slave.channels,
> > +				    vc.chan.device_node)
> > +			if (p->state == S3C24XX_DMA_CHAN_WAITING &&
> > +				      s3c24xx_dma_phy_valid(p, s3cchan->phy)) {
> > +				next = p;
> > +				break;
> > +			}
> > +	}
> > +
> > +	/* Ensure that the physical channel is stopped */
> > +	s3c24xx_dma_terminate_phy(s3cchan->phy);
> > +
> > +	if (next) {
> > +		bool success;
> > +
> > +		/*
> > +		 * Eww.  We know this isn't going to deadlock
> > +		 * but lockdep probably doesn't.
> > +		 */
> > +		spin_lock(&next->vc.lock);
> > +		/* Re-check the state now that we have the lock */
> > +		success = next->state == S3C24XX_DMA_CHAN_WAITING;
> > +		if (success)
> > +			s3c24xx_dma_phy_reassign_start(s3cchan->phy, next);
> > +		spin_unlock(&next->vc.lock);
> > +
> > +		/* If the state changed, try to find another channel */
> > +		if (!success)
> > +			goto retry;
> > +	} else {
> > +		/* No more jobs, so free up the physical channel */
> > +		s3c24xx_dma_put_phy(s3cchan->phy);
> > +	}
> > +
> > +	s3cchan->phy = NULL;
> > +	s3cchan->state = S3C24XX_DMA_CHAN_IDLE;
> > +}
> > +
> > +static void s3c24xx_dma_unmap_buffers(struct s3c24xx_txd *txd)
> > +{
> > +	struct device *dev = txd->vd.tx.chan->device->dev;
> > +	struct s3c24xx_sg *dsg;
> > +
> > +	if (!(txd->vd.tx.flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
> > +		if (txd->vd.tx.flags & DMA_COMPL_SRC_UNMAP_SINGLE)
> > +			list_for_each_entry(dsg, &txd->dsg_list, node)
> > +				dma_unmap_single(dev, dsg->src_addr, dsg->len,
> > +						DMA_TO_DEVICE);
> > +		else {
> > +			list_for_each_entry(dsg, &txd->dsg_list, node)
> > +				dma_unmap_page(dev, dsg->src_addr, dsg->len,
> > +						DMA_TO_DEVICE);
> > +		}
> > +	}
> > +
> > +	if (!(txd->vd.tx.flags & DMA_COMPL_SKIP_DEST_UNMAP)) {
> > +		if (txd->vd.tx.flags & DMA_COMPL_DEST_UNMAP_SINGLE)
> > +			list_for_each_entry(dsg, &txd->dsg_list, node)
> > +				dma_unmap_single(dev, dsg->dst_addr, dsg->len,
> > +						DMA_FROM_DEVICE);
> > +		else
> > +			list_for_each_entry(dsg, &txd->dsg_list, node)
> > +				dma_unmap_page(dev, dsg->dst_addr, dsg->len,
> > +						DMA_FROM_DEVICE);
> > +	}
> > +}
> > +
> > +static void s3c24xx_dma_desc_free(struct virt_dma_desc *vd)
> > +{
> > +	struct s3c24xx_txd *txd = to_s3c24xx_txd(&vd->tx);
> > +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(vd->tx.chan);
> > +
> > +	if (!s3cchan->slave)
> > +		s3c24xx_dma_unmap_buffers(txd);
> > +
> > +	s3c24xx_dma_free_txd(txd);
> > +}
> > +
> > +static irqreturn_t s3c24xx_dma_irq(int irq, void *data)
> > +{
> > +	struct s3c24xx_dma_phy *phy = data;
> > +	struct s3c24xx_dma_chan *s3cchan = phy->serving;
> > +	struct s3c24xx_txd *txd;
> > +
> > +	dev_dbg(&phy->host->pdev->dev, "interrupt on channel %d\n", phy-
> >id);
> > +
> > +	/*
> > +	 * Interrupts happen to notify the completion of a transfer and the
> > +	 * channel should have moved into its stop state already on its own.
> > +	 * Therefore interrupts on channels not bound to a virtual channel
> > +	 * should never happen. Nevertheless send a terminate command to
> the
> > +	 * channel if the unlikely case happens.
> > +	 */
> > +	if (unlikely(!s3cchan)) {
> > +		dev_err(&phy->host->pdev->dev, "interrupt on unused
> channel %d\n",
> > +			phy->id);
> > +
> > +		s3c24xx_dma_terminate_phy(phy);
> > +
> > +		return IRQ_HANDLED;
> > +	}
> > +
> > +	spin_lock(&s3cchan->vc.lock);
> > +	txd = s3cchan->at;
> > +	if (txd) {
> > +		/* when more sg's are in this txd, start the next one */
> > +		if (!list_is_last(txd->at, &txd->dsg_list)) {
> > +			txd->at = txd->at->next;
> > +			s3c24xx_dma_start_next_sg(s3cchan, txd);
> > +		} else {
> > +			s3cchan->at = NULL;
> > +			vchan_cookie_complete(&txd->vd);
> > +
> > +			/*
> > +			 * And start the next descriptor (if any),
> > +			 * otherwise free this channel.
> > +			 */
> > +			if (vchan_next_desc(&s3cchan->vc))
> > +				s3c24xx_dma_start_next_txd(s3cchan);
> > +			else
> > +				s3c24xx_dma_phy_free(s3cchan);
> > +		}
> > +	}
> > +	spin_unlock(&s3cchan->vc.lock);
> > +
> > +	return IRQ_HANDLED;
> > +}
> > +
> > +/*
> > + * The DMA ENGINE API
> > + */
> > +
> > +static int s3c24xx_dma_control(struct dma_chan *chan, enum dma_ctrl_cmd
> > cmd, +			 unsigned long arg)
> > +{
> > +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +	unsigned long flags;
> > +	int ret = 0;
> > +
> > +	spin_lock_irqsave(&s3cchan->vc.lock, flags);
> > +
> > +	switch (cmd) {
> > +	case DMA_SLAVE_CONFIG:
> > +		ret = s3c24xx_dma_set_runtime_config(s3cchan,
> > +					      (struct dma_slave_config *)arg);
> > +		break;
> > +	case DMA_TERMINATE_ALL:
> > +		if (!s3cchan->phy && !s3cchan->at) {
> > +			dev_err(&s3cdma->pdev->dev, "trying to terminate
> already stopped
> > channel %d\n", +				s3cchan->id);
> > +			ret = -EINVAL;
> > +			break;
> > +		}
> > +
> > +		s3cchan->state = S3C24XX_DMA_CHAN_IDLE;
> > +
> > +		 /* Mark physical channel as free */
> > +		if (s3cchan->phy)
> > +			s3c24xx_dma_phy_free(s3cchan);
> > +
> > +		/* Dequeue current job */
> > +		if (s3cchan->at) {
> > +			s3c24xx_dma_desc_free(&s3cchan->at->vd);
> > +			s3cchan->at = NULL;
> > +		}
> > +
> > +		/* Dequeue jobs not yet fired as well */
> > +		s3c24xx_dma_free_txd_list(s3cdma, s3cchan);
> > +		break;
> > +	default:
> > +		/* Unknown command */
> > +		ret = -ENXIO;
> > +		break;
> > +	}
> > +
> > +	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> > +
> > +	return ret;
> > +}
> > +
> > +static int s3c24xx_dma_alloc_chan_resources(struct dma_chan *chan)
> > +{
> > +	return 0;
> > +}
> > +
> > +static void s3c24xx_dma_free_chan_resources(struct dma_chan *chan)
> > +{
> > +	/* Ensure all queued descriptors are freed */
> > +	vchan_free_chan_resources(to_virt_chan(chan));
> > +}
> > +
> > +static enum dma_status s3c24xx_dma_tx_status(struct dma_chan *chan,
> > +		dma_cookie_t cookie, struct dma_tx_state *txstate)
> > +{
> > +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> > +	struct s3c24xx_txd *txd;
> > +	struct s3c24xx_sg *dsg;
> > +	struct virt_dma_desc *vd;
> > +	unsigned long flags;
> > +	enum dma_status ret;
> > +	size_t bytes = 0;
> > +
> > +	spin_lock_irqsave(&s3cchan->vc.lock, flags);
> > +	ret = dma_cookie_status(chan, cookie, txstate);
> > +	if (ret == DMA_SUCCESS) {
> > +		spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> > +		return ret;
> > +	}
> > +
> > +	/*
> > +	 * There's no point calculating the residue if there's
> > +	 * no txstate to store the value.
> > +	 */
> > +	if (!txstate) {
> > +		spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> > +		return ret;
> > +	}
> > +
> > +	vd = vchan_find_desc(&s3cchan->vc, cookie);
> > +	if (vd) {
> > +		/* On the issued list, so hasn't been processed yet */
> > +		txd = to_s3c24xx_txd(&vd->tx);
> > +
> > +		list_for_each_entry(dsg, &txd->dsg_list, node)
> > +			bytes += dsg->len;
> > +	} else {
> > +		/*
> > +		 * Currently running, so sum over the pending sg's and
> > +		 * the currently active one.
> > +		 */
> > +		txd = s3cchan->at;
> > +
> > +		dsg = list_entry(txd->at, struct s3c24xx_sg, node);
> > +		list_for_each_entry_from(dsg, &txd->dsg_list, node)
> > +			bytes += dsg->len;
> > +
> > +		bytes += s3c24xx_dma_getbytes_chan(s3cchan);
> > +	}
> > +	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> > +
> > +	/*
> > +	 * This cookie not complete yet
> > +	 * Get number of bytes left in the active transactions and queue
> > +	 */
> > +	dma_set_residue(txstate, bytes);
> > +
> > +	/* Whether waiting or running, we're in progress */
> > +	return ret;
> > +}
> > +
> > +/*
> > + * Initialize a descriptor to be used by memcpy submit
> > + */
> > +static struct dma_async_tx_descriptor *s3c24xx_dma_prep_memcpy(
> > +		struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
> > +		size_t len, unsigned long flags)
> > +{
> > +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +	struct s3c24xx_txd *txd;
> > +	struct s3c24xx_sg *dsg;
> > +	int src_mod, dest_mod;
> > +
> > +	dev_dbg(&s3cdma->pdev->dev, "prepare memcpy of %d bytes from %s\n",
> > +			len, s3cchan->name);
> > +
> > +	if ((len & S3C24XX_DCON_TC_MASK) != len) {
> > +		dev_err(&s3cdma->pdev->dev, "memcpy size %d to large\n",
> len);
> > +		return NULL;
> > +	}
> > +
> > +	txd = s3c24xx_dma_get_txd();
> > +	if (!txd)
> > +		return NULL;
> > +
> > +	dsg = kzalloc(sizeof(*dsg), GFP_NOWAIT);
> > +	if (!dsg) {
> > +		s3c24xx_dma_free_txd(txd);
> > +		return NULL;
> > +	}
> > +	list_add_tail(&dsg->node, &txd->dsg_list);
> > +
> > +	dsg->src_addr = src;
> > +	dsg->dst_addr = dest;
> > +	dsg->len = len;
> > +
> > +	/*
> > +	 * Determine a suitable transfer width.
> > +	 * The DMA controller cannot fetch/store information which is not
> > +	 * naturally aligned on the bus, i.e., a 4 byte fetch must start at
> > +	 * an address divisible by 4 - more generally addr % width must be
> 0.
> > +	 */
> > +	src_mod = src % 4;
> > +	dest_mod = dest % 4;
> > +	switch (len % 4) {
> > +	case 0:
> > +		txd->width = (src_mod == 0 && dest_mod == 0) ? 4 : 1;
> > +		break;
> > +	case 2:
> > +		txd->width = ((src_mod == 2 || src_mod == 0) &&
> > +			      (dest_mod == 2 || dest_mod == 0)) ? 2 : 1;
> > +		break;
> > +	default:
> > +		txd->width = 1;
> > +		break;
> > +	}
> > +
> > +	txd->disrcc = S3C24XX_DISRCC_LOC_AHB | S3C24XX_DISRCC_INC_INCREMENT;
> > +	txd->didstc = S3C24XX_DIDSTC_LOC_AHB | S3C24XX_DIDSTC_INC_INCREMENT;
> > +	txd->dcon |= S3C24XX_DCON_DEMAND | S3C24XX_DCON_SYNC_HCLK |
> > +		     S3C24XX_DCON_SERV_WHOLE;
> > +
> > +	return vchan_tx_prep(&s3cchan->vc, &txd->vd, flags);
> > +}
> > +
> > +static struct dma_async_tx_descriptor *s3c24xx_dma_prep_slave_sg(
> > +		struct dma_chan *chan, struct scatterlist *sgl,
> > +		unsigned int sg_len, enum dma_transfer_direction direction,
> > +		unsigned long flags, void *context)
> > +{
> > +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> > +	struct s3c24xx_dma_engine *s3cdma = s3cchan->host;
> > +	const struct s3c24xx_dma_platdata *pdata = s3cdma->pdata;
> > +	struct s3c24xx_dma_channel *cdata = &pdata->channels[s3cchan->id];
> > +	struct s3c24xx_txd *txd;
> > +	struct s3c24xx_sg *dsg;
> > +	struct scatterlist *sg;
> > +	dma_addr_t slave_addr;
> > +	u32 hwcfg = 0;
> > +	int tmp;
> > +
> > +	dev_dbg(&s3cdma->pdev->dev, "prepare transaction of %d bytes
> from %s\n",
> > +			sg_dma_len(sgl), s3cchan->name);
> > +
> > +	txd = s3c24xx_dma_get_txd();
> > +	if (!txd)
> > +		return NULL;
> > +
> > +	if (cdata->handshake)
> > +		txd->dcon |= S3C24XX_DCON_HANDSHAKE;
> > +
> > +	switch (cdata->bus) {
> > +	case S3C24XX_DMA_APB:
> > +		txd->dcon |= S3C24XX_DCON_SYNC_PCLK;
> > +		hwcfg |= S3C24XX_DISRCC_LOC_APB;
> > +		break;
> > +	case S3C24XX_DMA_AHB:
> > +		txd->dcon |= S3C24XX_DCON_SYNC_HCLK;
> > +		hwcfg |= S3C24XX_DISRCC_LOC_AHB;
> > +		break;
> > +	}
> > +
> > +	/*
> > +	 * Always assume our peripheral desintation is a fixed
> > +	 * address in memory.
> > +	 */
> > +	hwcfg |= S3C24XX_DISRCC_INC_FIXED;
> > +
> > +	/*
> > +	 * Individual dma operations are requested by the slave,
> > +	 * so serve only single atomic operations
> (S3C24XX_DCON_SERV_SINGLE).
> > +	 */
> > +	txd->dcon |= S3C24XX_DCON_SERV_SINGLE;
> > +
> > +	if (direction == DMA_MEM_TO_DEV) {
> > +		txd->disrcc = S3C24XX_DISRCC_LOC_AHB |
> > +			      S3C24XX_DISRCC_INC_INCREMENT;
> > +		txd->didstc = hwcfg;
> > +		slave_addr = s3cchan->cfg.dst_addr;
> > +		txd->width = s3cchan->cfg.dst_addr_width;
> > +	} else if (direction == DMA_DEV_TO_MEM) {
> > +		txd->disrcc = hwcfg;
> > +		txd->didstc = S3C24XX_DIDSTC_LOC_AHB |
> > +			      S3C24XX_DIDSTC_INC_INCREMENT;
> > +		slave_addr = s3cchan->cfg.src_addr;
> > +		txd->width = s3cchan->cfg.src_addr_width;
> > +	} else {
> > +		s3c24xx_dma_free_txd(txd);
> > +		dev_err(&s3cdma->pdev->dev,
> > +			"direction %d unsupported\n", direction);
> > +		return NULL;
> > +	}
> > +
> > +	for_each_sg(sgl, sg, sg_len, tmp) {
> > +		dsg = kzalloc(sizeof(*dsg), GFP_NOWAIT);
> > +		if (!dsg) {
> > +			s3c24xx_dma_free_txd(txd);
> > +			return NULL;
> > +		}
> > +		list_add_tail(&dsg->node, &txd->dsg_list);
> > +
> > +		dsg->len = sg_dma_len(sg);
> > +		if (direction == DMA_MEM_TO_DEV) {
> > +			dsg->src_addr = sg_dma_address(sg);
> > +			dsg->dst_addr = slave_addr;
> > +		} else { /* DMA_DEV_TO_MEM */
> > +			dsg->src_addr = slave_addr;
> > +			dsg->dst_addr = sg_dma_address(sg);
> > +		}
> > +		break;
> > +	}
> > +
> > +	return vchan_tx_prep(&s3cchan->vc, &txd->vd, flags);
> > +}
> > +
> > +/*
> > + * Slave transactions callback to the slave device to allow
> > + * synchronization of slave DMA signals with the DMAC enable
> > + */
> > +static void s3c24xx_dma_issue_pending(struct dma_chan *chan)
> > +{
> > +	struct s3c24xx_dma_chan *s3cchan = to_s3c24xx_dma_chan(chan);
> > +	unsigned long flags;
> > +
> > +	spin_lock_irqsave(&s3cchan->vc.lock, flags);
> > +	if (vchan_issue_pending(&s3cchan->vc)) {
> > +		if (!s3cchan->phy && s3cchan->state !=
> S3C24XX_DMA_CHAN_WAITING)
> > +			s3c24xx_dma_phy_alloc_and_start(s3cchan);
> > +	}
> > +	spin_unlock_irqrestore(&s3cchan->vc.lock, flags);
> > +}
> > +
> > +/*
> > + * Bringup and teardown
> > + */
> > +
> > +/*
> > + * Initialise the DMAC memcpy/slave channels.
> > + * Make a local wrapper to hold required data
> > + */
> > +static int s3c24xx_dma_init_virtual_channels(struct s3c24xx_dma_engine
> > *s3cdma, +		struct dma_device *dmadev, unsigned int channels, bool
> slave)
> > +{
> > +	struct s3c24xx_dma_chan *chan;
> > +	int i;
> > +
> > +	INIT_LIST_HEAD(&dmadev->channels);
> > +
> > +	/*
> > +	 * Register as many many memcpy as we have physical channels,
> > +	 * we won't always be able to use all but the code will have
> > +	 * to cope with that situation.
> > +	 */
> > +	for (i = 0; i < channels; i++) {
> > +		chan = devm_kzalloc(dmadev->dev, sizeof(*chan), GFP_KERNEL);
> > +		if (!chan) {
> > +			dev_err(dmadev->dev,
> > +				"%s no memory for channel\n", __func__);
> > +			return -ENOMEM;
> > +		}
> > +
> > +		chan->id = i;
> > +		chan->host = s3cdma;
> > +		chan->state = S3C24XX_DMA_CHAN_IDLE;
> > +
> > +		if (slave) {
> > +			chan->slave = true;
> > +			chan->name = kasprintf(GFP_KERNEL, "slave%d", i);
> > +			if (!chan->name)
> > +				return -ENOMEM;
> > +		} else {
> > +			chan->name = kasprintf(GFP_KERNEL, "memcpy%d", i);
> > +			if (!chan->name)
> > +				return -ENOMEM;
> > +		}
> > +		dev_dbg(dmadev->dev,
> > +			 "initialize virtual channel \"%s\"\n",
> > +			 chan->name);
> > +
> > +		chan->vc.desc_free = s3c24xx_dma_desc_free;
> > +		vchan_init(&chan->vc, dmadev);
> > +	}
> > +	dev_info(dmadev->dev, "initialized %d virtual %s channels\n",
> > +		 i, slave ? "slave" : "memcpy");
> > +	return i;
> > +}
> > +
> > +static void s3c24xx_dma_free_virtual_channels(struct dma_device *dmadev)
> > +{
> > +	struct s3c24xx_dma_chan *chan = NULL;
> > +	struct s3c24xx_dma_chan *next;
> > +
> > +	list_for_each_entry_safe(chan,
> > +				 next, &dmadev->channels, vc.chan.device_node)
> > +		list_del(&chan->vc.chan.device_node);
> > +}
> > +
> > +/* s3c2412 and s3c2413 have a 0x40 stride and dmareqsel mechanism */
> > +static struct soc_data soc_s3c2412 = {
> > +	.stride = 0x40,
> > +	.has_reqsel = true,
> > +	.has_clocks = true,
> > +};
> > +
> > +/* s3c2443 and following have a 0x100 stride and dmareqsel mechanism */
> > +static struct soc_data soc_s3c2443 = {
> > +	.stride = 0x100,
> > +	.has_reqsel = true,
> > +	.has_clocks = true,
> > +};
> > +
> > +static struct platform_device_id s3c24xx_dma_driver_ids[] = {
> > +	{
> > +		.name		= "s3c2412-dma",
> > +		.driver_data	= (kernel_ulong_t)&soc_s3c2412,
> > +	}, {
> > +		.name		= "s3c2443-dma",
> > +		.driver_data	= (kernel_ulong_t)&soc_s3c2443,
> > +	},
> > +	{ },
> > +};
> > +
> > +static struct soc_data *s3c24xx_dma_get_soc_data(struct platform_device
> > *pdev) +{
> > +	return (struct soc_data *)
> > +			 platform_get_device_id(pdev)->driver_data;
> > +}
> > +
> > +static int s3c24xx_dma_probe(struct platform_device *pdev)
> > +{
> > +	const struct s3c24xx_dma_platdata *pdata = dev_get_platdata(&pdev-
> >dev);
> > +	struct s3c24xx_dma_engine *s3cdma;
> > +	struct soc_data *sdata;
> > +	struct resource *res;
> > +	int ret;
> > +	int i;
> > +
> > +	if (!pdata) {
> > +		dev_err(&pdev->dev, "platform data missing\n");
> > +		return -ENODEV;
> > +	}
> > +
> > +	/* Basic sanity check */
> > +	if (pdata->num_phy_channels > MAX_DMA_CHANNELS) {
> > +		dev_err(&pdev->dev, "to many dma channels %d, max %d\n",
> > +			pdata->num_phy_channels, MAX_DMA_CHANNELS);
> > +		return -EINVAL;
> > +	}
> > +
> > +	sdata = s3c24xx_dma_get_soc_data(pdev);
> > +	if (!sdata)
> > +		return -EINVAL;
> > +
> > +	s3cdma = devm_kzalloc(&pdev->dev, sizeof(*s3cdma), GFP_KERNEL);
> > +	if (!s3cdma)
> > +		return -ENOMEM;
> > +
> > +	s3cdma->pdev = pdev;
> > +	s3cdma->pdata = pdata;
> > +	s3cdma->sdata = sdata;
> > +
> > +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > +	s3cdma->base = devm_ioremap_resource(&pdev->dev, res);
> > +	if (IS_ERR(s3cdma->base))
> > +		return PTR_ERR(s3cdma->base);
> > +
> > +	s3cdma->phy_chans = devm_kzalloc(&pdev->dev,
> > +					      sizeof(struct s3c24xx_dma_phy) *
> > +							pdata->num_phy_channels,
> > +					      GFP_KERNEL);
> > +	if (!s3cdma->phy_chans)
> > +		return -ENOMEM;
> > +
> > +	/* aquire irqs and clocks for all physical channels */
> > +	for (i = 0; i < pdata->num_phy_channels; i++) {
> > +		struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
> > +		char clk_name[6];
> > +
> > +		phy->id = i;
> > +		phy->base = s3cdma->base + (i * sdata->stride);
> > +		phy->host = s3cdma;
> > +
> > +		phy->irq = platform_get_irq(pdev, i);
> > +		if (phy->irq < 0) {
> > +			dev_err(&pdev->dev, "failed to get irq %d, err %d\n",
> > +				i, phy->irq);
> > +			continue;
> > +		}
> > +
> > +		ret = devm_request_irq(&pdev->dev, phy->irq, s3c24xx_dma_irq,
> > +				       0, pdev->name, phy);
> > +		if (ret) {
> > +			dev_err(&pdev->dev, "Unable to request irq for
> channel %d, error
> %d\n",
> > +				i, ret);
> > +			continue;
> > +		}
> > +
> > +		if (sdata->has_clocks) {
> > +			sprintf(clk_name, "dma.%d", i);
> > +			phy->clk = devm_clk_get(&pdev->dev, clk_name);
> > +			if (IS_ERR(phy->clk) && sdata->has_clocks) {
> > +				dev_err(&pdev->dev, "unable to aquire clock for
> channel %d,
> error
> > %lu", +					i, PTR_ERR(phy->clk));
> > +				continue;
> > +			}
> > +
> > +			ret = clk_prepare(phy->clk);
> > +			if (ret) {
> > +				dev_err(&pdev->dev, "clock for phy %d failed,
> error %d\n",
> > +					i, ret);
> > +				continue;
> > +			}
> > +		}
> > +
> > +		spin_lock_init(&phy->lock);
> > +		phy->valid = true;
> > +
> > +		dev_dbg(&pdev->dev, "physical channel %d is %s\n",
> > +			i, s3c24xx_dma_phy_busy(phy) ? "BUSY" : "FREE");
> > +	}
> > +
> > +	/* Initialize memcpy engine */
> > +	dma_cap_set(DMA_MEMCPY, s3cdma->memcpy.cap_mask);
> > +	dma_cap_set(DMA_PRIVATE, s3cdma->memcpy.cap_mask);
> > +	s3cdma->memcpy.dev = &pdev->dev;
> > +	s3cdma->memcpy.device_alloc_chan_resources =
> > +					s3c24xx_dma_alloc_chan_resources;
> > +	s3cdma->memcpy.device_free_chan_resources =
> > +					s3c24xx_dma_free_chan_resources;
> > +	s3cdma->memcpy.device_prep_dma_memcpy = s3c24xx_dma_prep_memcpy;
> > +	s3cdma->memcpy.device_tx_status = s3c24xx_dma_tx_status;
> > +	s3cdma->memcpy.device_issue_pending = s3c24xx_dma_issue_pending;
> > +	s3cdma->memcpy.device_control = s3c24xx_dma_control;
> > +
> > +	/* Initialize slave engine for SoC internal dedicated peripherals
> */
> > +	dma_cap_set(DMA_SLAVE, s3cdma->slave.cap_mask);
> > +	dma_cap_set(DMA_PRIVATE, s3cdma->slave.cap_mask);
> > +	s3cdma->slave.dev = &pdev->dev;
> > +	s3cdma->slave.device_alloc_chan_resources =
> > +					s3c24xx_dma_alloc_chan_resources;
> > +	s3cdma->slave.device_free_chan_resources =
> > +					s3c24xx_dma_free_chan_resources;
> > +	s3cdma->slave.device_tx_status = s3c24xx_dma_tx_status;
> > +	s3cdma->slave.device_issue_pending = s3c24xx_dma_issue_pending;
> > +	s3cdma->slave.device_prep_slave_sg = s3c24xx_dma_prep_slave_sg;
> > +	s3cdma->slave.device_control = s3c24xx_dma_control;
> > +
> > +	/* Register as many memcpy channels as there are physical channels
> */
> > +	ret = s3c24xx_dma_init_virtual_channels(s3cdma, &s3cdma->memcpy,
> > +						pdata->num_phy_channels, false);
> > +	if (ret <= 0) {
> > +		dev_warn(&pdev->dev,
> > +			 "%s failed to enumerate memcpy channels - %d\n",
> > +			 __func__, ret);
> > +		goto err_memcpy;
> > +	}
> > +
> > +	/* Register slave channels */
> > +	ret = s3c24xx_dma_init_virtual_channels(s3cdma, &s3cdma->slave,
> > +				pdata->num_channels, true);
> > +	if (ret <= 0) {
> > +		dev_warn(&pdev->dev,
> > +			"%s failed to enumerate slave channels - %d\n",
> > +				__func__, ret);
> > +		goto err_slave;
> > +	}
> > +
> > +	ret = dma_async_device_register(&s3cdma->memcpy);
> > +	if (ret) {
> > +		dev_warn(&pdev->dev,
> > +			"%s failed to register memcpy as an async device -
>  %d\n",
> > +			__func__, ret);
> > +		goto err_memcpy_reg;
> > +	}
> > +
> > +	ret = dma_async_device_register(&s3cdma->slave);
> > +	if (ret) {
> > +		dev_warn(&pdev->dev,
> > +			"%s failed to register slave as an async device -
>  %d\n",
> > +			__func__, ret);
> > +		goto err_slave_reg;
> > +	}
> > +
> > +	platform_set_drvdata(pdev, s3cdma);
> > +	dev_info(&pdev->dev, "Loaded dma driver with %d physical
> channels\n",
> > +		 pdata->num_phy_channels);
> > +
> > +	return 0;
> > +
> > +err_slave_reg:
> > +	dma_async_device_unregister(&s3cdma->memcpy);
> > +err_memcpy_reg:
> > +	s3c24xx_dma_free_virtual_channels(&s3cdma->slave);
> > +err_slave:
> > +	s3c24xx_dma_free_virtual_channels(&s3cdma->memcpy);
> > +err_memcpy:
> > +	if (sdata->has_clocks)
> > +		for (i = 0; i < pdata->num_phy_channels; i++) {
> > +			struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
> > +			if (phy->valid)
> > +				clk_unprepare(phy->clk);
> > +		}
> > +
> > +	return ret;
> > +}
> > +
> > +static int s3c24xx_dma_remove(struct platform_device *pdev)
> > +{
> > +	const struct s3c24xx_dma_platdata *pdata = dev_get_platdata(&pdev-
> >dev);
> > +	struct s3c24xx_dma_engine *s3cdma = platform_get_drvdata(pdev);
> > +	struct soc_data *sdata = s3c24xx_dma_get_soc_data(pdev);
> > +	int i;
> > +
> > +	dma_async_device_unregister(&s3cdma->slave);
> > +	dma_async_device_unregister(&s3cdma->memcpy);
> > +
> > +	s3c24xx_dma_free_virtual_channels(&s3cdma->slave);
> > +	s3c24xx_dma_free_virtual_channels(&s3cdma->memcpy);
> > +
> > +	if (sdata->has_clocks)
> > +		for (i = 0; i < pdata->num_phy_channels; i++) {
> > +			struct s3c24xx_dma_phy *phy = &s3cdma->phy_chans[i];
> > +			if (phy->valid)
> > +				clk_unprepare(phy->clk);
> > +		}
> > +
> > +	return 0;
> > +}
> > +
> > +static struct platform_driver s3c24xx_dma_driver = {
> > +	.driver		= {
> > +		.name	= "s3c24xx-dma",
> > +		.owner	= THIS_MODULE,
> > +	},
> > +	.id_table	= s3c24xx_dma_driver_ids,
> > +	.probe		= s3c24xx_dma_probe,
> > +	.remove		= s3c24xx_dma_remove,
> > +};
> > +
> > +module_platform_driver(s3c24xx_dma_driver);
> > +
> > +bool s3c24xx_dma_filter(struct dma_chan *chan, void *param)
> > +{
> > +	struct s3c24xx_dma_chan *s3cchan;
> > +
> > +	if (chan->device->dev->driver != &s3c24xx_dma_driver.driver)
> > +		return false;
> > +
> > +	s3cchan = to_s3c24xx_dma_chan(chan);
> > +
> > +	return s3cchan->id == (int)param;
> > +}
> > +EXPORT_SYMBOL(s3c24xx_dma_filter);
> > +
> > +MODULE_DESCRIPTION("S3C24XX DMA Driver");
> > +MODULE_AUTHOR("Heiko Stuebner");
> > +MODULE_LICENSE("GPL v2");
> > diff --git a/include/linux/platform_data/dma-s3c24xx.h
> > b/include/linux/platform_data/dma-s3c24xx.h new file mode 100644
> > index 0000000..5a0cfff
> > --- /dev/null
> > +++ b/include/linux/platform_data/dma-s3c24xx.h
> > @@ -0,0 +1,43 @@
> > +/*
> > + * S3C24XX DMA handling
> > + *
> > + * Copyright (c) 2013 Heiko Stuebner <heiko@sntech.de>
> > + *
> > + * 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.
> > + */
> > +
> > +enum s3c24xx_dma_bus {
> > +	S3C24XX_DMA_APB,
> > +	S3C24XX_DMA_AHB,
> > +};
> > +
> > +/**
> > + * @bus: on which bus does the peripheral reside - AHB or APB.
> > + * @handshake: is a handshake with the peripheral necessary
> > + * @chansel: channel selection information, depending on variant;
> reqsel
> > for + *	     s3c2443 and later and channel-selection map for earlier
> SoCs
> > + *	     see CHANSEL doc in s3c2443-dma.c
> > + */
> > +struct s3c24xx_dma_channel {
> > +	enum s3c24xx_dma_bus bus;
> > +	bool handshake;
> > +	u16 chansel;
> > +};
> > +
> > +/**
> > + * struct s3c24xx_dma_platdata - platform specific settings
> > + * @num_phy_channels: number of physical channels
> > + * @channels: array of virtual channel descriptions
> > + * @num_channels: number of virtual channels
> > + */
> > +struct s3c24xx_dma_platdata {
> > +	int num_phy_channels;
> > +	struct s3c24xx_dma_channel *channels;
> > +	int num_channels;
> > +};
> > +
> > +struct dma_chan;
> > +bool s3c24xx_dma_filter(struct dma_chan *chan, void *param);

Re: [PATCH v5 2/4] dmaengine: add driver for Samsung s3c24xx SoCs

[Vinod Koul]

On Wed, Sep 18, 2013 at 09:19:54PM +0200, Heiko Stübner wrote:
> This adds a new driver to support the s3c24xx dma using the dmaengine
> and makes the old one in mach-s3c24xx obsolete in the long run.
> 
> Conceptually the s3c24xx-dma feels like a distant relative of the pl08x
> with numerous virtual channels being mapped to a lot less physical ones.
> The driver therefore borrows a lot from the amba-pl08x driver in this
> regard. Functionality-wise the driver gains a memcpy ability in addition
> to the slave_sg one.
> 
> The driver supports both the method for requesting the peripheral used
> by SoCs before the S3C2443 and the different method for S3C2443 and later.
> 
> On earlier SoCs the hardware channels usable for specific peripherals is
> constrainted while on later SoCs all channels can be used for any peripheral.
> 
> Tested on a s3c2416-based board, memcpy using the dmatest module and
> slave_sg partially using the spi-s3c64xx driver.
> 
> Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Vinod Koul <vinod.koul@intel.com>

--
~Vinod

RE: [PATCH v5 2/4] dmaengine: add driver for Samsung s3c24xx SoCs

[Kukjin Kim]

Vinod Koul wrote:
> 
> On Wed, Sep 18, 2013 at 09:19:54PM +0200, Heiko Stübner wrote:
> > This adds a new driver to support the s3c24xx dma using the dmaengine
> > and makes the old one in mach-s3c24xx obsolete in the long run.
> >
> > Conceptually the s3c24xx-dma feels like a distant relative of the pl08x
> > with numerous virtual channels being mapped to a lot less physical ones.
> > The driver therefore borrows a lot from the amba-pl08x driver in this
> > regard. Functionality-wise the driver gains a memcpy ability in addition
> > to the slave_sg one.
> >
> > The driver supports both the method for requesting the peripheral used
> > by SoCs before the S3C2443 and the different method for S3C2443 and
> later.
> >
> > On earlier SoCs the hardware channels usable for specific peripherals is
> > constrainted while on later SoCs all channels can be used for any
> peripheral.
> >
> > Tested on a s3c2416-based board, memcpy using the dmatest module and
> > slave_sg partially using the spi-s3c64xx driver.
> >
> > Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> > Acked-by: Linus Walleij <linus.walleij@linaro.org>
> Acked-by: Vinod Koul <vinod.koul@intel.com>
> 
Vinod, thanks for your ack :-)

Applied this whole series.
Kukjin

Re: [PATCH v5 3/4] ARM: S3C24XX: add platform-devices for new dma driver for s3c2412 and s3c2443

[Olof Johansson]

Hi,

On Wed, Sep 18, 2013 at 12:20 PM, Heiko Stübner <heiko@sntech.de> wrote:
> This includes defining the mapping for the request sources.
>
> Signed-off-by: Heiko Stuebner <heiko@sntech.de>
> Acked-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> changes since v1:
> - follow new pdata definition

This showed up in -next today, and introduced:

arch/arm/mach-s3c24xx/common.c:337:24: warning: 's3c2410_dma_resource'
defined but not used [-Wunused-variable]

> +#if defined(CONFIG_CPU_S3C2410) || defined(CONFIG_CPU_S3C2412) || \
> +    defined(CONFIG_CPU_S3C2440) || defined(CONFIG_CPU_S3C2442)
> +static struct resource s3c2410_dma_resource[] = {
> +       [0] = DEFINE_RES_MEM(S3C24XX_PA_DMA, S3C24XX_SZ_DMA),
> +       [1] = DEFINE_RES_IRQ(IRQ_DMA0),
> +       [2] = DEFINE_RES_IRQ(IRQ_DMA1),
> +       [3] = DEFINE_RES_IRQ(IRQ_DMA2),
> +       [4] = DEFINE_RES_IRQ(IRQ_DMA3),
> +};
> +#endif
> +

Log at:

http://arm-soc.lixom.net/buildlogs/next-thierry/v3.12-rc4-4340-g56a7390/buildall.arm.mini2440_defconfig.log.passed


-Olof