Skip to contentSkip to conversation

tnetv107x ssp drivers

PATCH Superseded Cyril Chemparathy 6 participants 28 messages Dec 7, 2010 – Dec 31, 2010 2 messages not in the archive

Subsystems (9)

v7 of 8 Superseded Latest is v8 →

lore
Share & tools
lore
#0 lore raw

v7 cover tnetv107x ssp drivers

TI's sequencer serial port (TI-SSP) is a jack-of-all-trades type of serial port device. It has a built-in programmable execution engine that can be programmed to operate as almost any serial bus (I2C, SPI, EasyScale, and others).

This patch series implements a driver stack that looks like the following:

                             +--------+
                             | eeprom | . . .
                             +--------+
   +-----------+             +--------------+ +---------+
   | regulator | . . .       |   i2c-gpio   | | 1-wire  | . . .
   +-----------+             +--------------+ +---------+
   +----------------------+  +--------------------------------+
   |       ssp-spi        |  |           ssp-gpio             |
   +----------------------+  +--------------------------------+
   +----------------------------------------------------------+
   |                            ssp                           |
   +----------------------------------------------------------+

Changes between v7 and v6 if this series:
- Workaround for iosel2 register not reading back set bits.
- Update backlight status once probe succeeds.

Changes between v6 and v5 of this series:
- Changed initcalls to module_init() across all drivers. This series now uses a late_initcall() in the board to delay initialization of gpio and regulator dependent devices.

Changes between v5 and v4 of this series:
- Moved drivers from misc/gpio/spi to mfd
- Removed implicit init-time iosel setup
- Minor cleanups in backlight driver

Changes between v3 and v4 of this series:
- Replaced polled wait for sequence termination with interrupt
- Improved locking within SSP driver
- Other minor cleanups

Changes between v2 and v3 of this series:
- Minor cleanups in Kconfig and Makefile ordering

Changes between v1 and v2 of this series:
- Replaced open()/close() semantics with dynamic platform_device registration on SSP probe.
- Removed user-land interface to regulator registers
- More sensible regulator constraints
- Other minor cleanups

Cyril Chemparathy (12):
misc: add driver for sequencer serial port
davinci: add tnetv107x ssp platform device
davinci: add ssp config for tnetv107x evm board
spi: add ti-ssp spi master driver
davinci: add spi devices on tnetv107x evm
regulator: add driver for tps6524x regulator
davinci: add tnetv107x evm regulators
gpio: add ti-ssp gpio driver
davinci: add tnetv107x evm ti-ssp gpio device
backlight: add support for tps6116x controller
davinci: add tnetv107x evm backlight device
davinci: add tnetv107x evm i2c eeprom device

 arch/arm/mach-davinci/board-tnetv107x-evm.c    |  197 +++++++
 arch/arm/mach-davinci/devices-tnetv107x.c      |   25 +
 arch/arm/mach-davinci/include/mach/tnetv107x.h |    2 +
 arch/arm/mach-davinci/tnetv107x.c              |    2 +-
 drivers/gpio/Kconfig                           |   10 +
 drivers/gpio/Makefile                          |    1 +
 drivers/gpio/ti-ssp-gpio.c                     |  206 +++++++
 drivers/mfd/Kconfig                            |   11 +
 drivers/mfd/Makefile                           |    1 +
 drivers/mfd/ti-ssp.c                           |  476 ++++++++++++++++
 drivers/regulator/Kconfig                      |   10 +
 drivers/regulator/Makefile                     |    1 +
 drivers/regulator/tps6524x-regulator.c         |  692 ++++++++++++++++++++++++
 drivers/spi/Kconfig                            |   10 +
 drivers/spi/Makefile                           |    1 +
 drivers/spi/ti-ssp-spi.c                       |  402 ++++++++++++++
 drivers/video/backlight/Kconfig                |    7 +
 drivers/video/backlight/Makefile               |    2 +-
 drivers/video/backlight/tps6116x.c             |  299 ++++++++++
 include/linux/mfd/ti_ssp.h                     |   97 ++++
 20 files changed, 2450 insertions(+), 2 deletions(-)
 create mode 100644 drivers/gpio/ti-ssp-gpio.c
 create mode 100644 drivers/mfd/ti-ssp.c
 create mode 100644 drivers/regulator/tps6524x-regulator.c
 create mode 100644 drivers/spi/ti-ssp-spi.c
 create mode 100644 drivers/video/backlight/tps6116x.c
 create mode 100644 include/linux/mfd/ti_ssp.h
#1 lore raw

PATCH v7 1/12 misc: add driver for sequencer serial port

TI's sequencer serial port (TI-SSP) is a jack-of-all-trades type of serial port device. It has a built-in programmable execution engine that can be programmed to operate as almost any serial bus (I2C, SPI, EasyScale, and others).

This patch adds a driver for this controller device. The driver does not expose a user-land interface. Protocol drivers built on top of this layer are expected to remain in-kernel.

---
 drivers/mfd/Kconfig        |   11 +
 drivers/mfd/Makefile       |    1 +
 drivers/mfd/ti-ssp.c       |  476 ++++++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/ti_ssp.h |   87 ++++++++
 4 files changed, 575 insertions(+), 0 deletions(-)
 create mode 100644 drivers/mfd/ti-ssp.c
 create mode 100644 include/linux/mfd/ti_ssp.h
drivers/mfd/Kconfig +11 -0
index 3a1493b..a4b4b70 100644
@@ -81,6 +81,17 @@ config MFD_DM355EVM_MSP
 	  boards.  MSP430 firmware manages resets and power sequencing,
 	  inputs from buttons and the IR remote, LEDs, an RTC, and more.
 
+config MFD_TI_SSP
+	tristate "TI Sequencer Serial Port support"
+	depends on ARCH_DAVINCI_TNETV107X
+	select MFD_CORE
+	---help---
+	  Say Y here if you want support for the Sequencer Serial Port
+	  in a Texas Instruments TNETV107X SoC.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called ti-ssp.
+
 config HTC_EGPIO
 	bool "HTC EGPIO support"
 	depends on GENERIC_HARDIRQS && GPIOLIB && ARM
drivers/mfd/Makefile +1 -0
index f54b365..f64cf13 100644
@@ -14,6 +14,7 @@ obj-$(CONFIG_HTC_I2CPLD)	+= htc-i2cpld.o
 
 obj-$(CONFIG_MFD_DAVINCI_VOICECODEC)	+= davinci_voicecodec.o
 obj-$(CONFIG_MFD_DM355EVM_MSP)	+= dm355evm_msp.o
+obj-$(CONFIG_MFD_TI_SSP)	+= ti-ssp.o
 
 obj-$(CONFIG_MFD_STMPE)		+= stmpe.o
 obj-$(CONFIG_MFD_TC35892)	+= tc35892.o
drivers/mfd/ti-ssp.c new file +476 -0

476 lines not on this page · raw message

include/linux/mfd/ti_ssp.h new file +87 -0

87 lines not on this page · raw message

Signature · 1 line
1.7.1
#2 lore raw

PATCH v7 1/12 misc: add driver for sequencer serial port

Quoted 9 lines · Cyril Chemparathy

TI's sequencer serial port (TI-SSP) is a jack-of-all-trades type of serial port device. It has a built-in programmable execution engine that can be programmed to operate as almost any serial bus (I2C, SPI, EasyScale, and others).

This patch adds a driver for this controller device. The driver does not expose a user-land interface. Protocol drivers built on top of this layer are expected to remain in-kernel.

Minor nit: subject still says 'misc', but this has now been moved to mfd.

Kevin

Quoted 618 lines
---
 drivers/mfd/Kconfig        |   11 +
 drivers/mfd/Makefile       |    1 +
 drivers/mfd/ti-ssp.c       |  476 ++++++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/ti_ssp.h |   87 ++++++++
 4 files changed, 575 insertions(+), 0 deletions(-)
 create mode 100644 drivers/mfd/ti-ssp.c
 create mode 100644 include/linux/mfd/ti_ssp.h
drivers/mfd/Kconfig +11 -0
index 3a1493b..a4b4b70 100644
@@ -81,6 +81,17 @@ config MFD_DM355EVM_MSP
 	  boards.  MSP430 firmware manages resets and power sequencing,
 	  inputs from buttons and the IR remote, LEDs, an RTC, and more.
 
+config MFD_TI_SSP
+	tristate "TI Sequencer Serial Port support"
+	depends on ARCH_DAVINCI_TNETV107X
+	select MFD_CORE
+	---help---
+	  Say Y here if you want support for the Sequencer Serial Port
+	  in a Texas Instruments TNETV107X SoC.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called ti-ssp.
+
 config HTC_EGPIO
 	bool "HTC EGPIO support"
 	depends on GENERIC_HARDIRQS && GPIOLIB && ARM
drivers/mfd/Makefile +1 -0
index f54b365..f64cf13 100644
@@ -14,6 +14,7 @@ obj-$(CONFIG_HTC_I2CPLD)	+= htc-i2cpld.o
 
 obj-$(CONFIG_MFD_DAVINCI_VOICECODEC)	+= davinci_voicecodec.o
 obj-$(CONFIG_MFD_DM355EVM_MSP)	+= dm355evm_msp.o
+obj-$(CONFIG_MFD_TI_SSP)	+= ti-ssp.o
 
 obj-$(CONFIG_MFD_STMPE)		+= stmpe.o
 obj-$(CONFIG_MFD_TC35892)	+= tc35892.o
drivers/mfd/ti-ssp.c new file +476 -0
new file mode 100644
index 0000000..af9ab0e
@@ -0,0 +1,476 @@
+/*
+ * Sequencer Serial Port (SSP) driver for Texas Instruments' SoCs
+ *
+ * Copyright (C) 2010 Texas Instruments Inc
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/wait.h>
+#include <linux/clk.h>
+#include <linux/interrupt.h>
+#include <linux/device.h>
+#include <linux/spinlock.h>
+#include <linux/platform_device.h>
+#include <linux/delay.h>
+#include <linux/io.h>
+#include <linux/mfd/core.h>
+#include <linux/mfd/ti_ssp.h>
+
+/* Register Offsets */
+#define REG_REV		0x00
+#define REG_IOSEL_1	0x04
+#define REG_IOSEL_2	0x08
+#define REG_PREDIV	0x0c
+#define REG_INTR_ST	0x10
+#define REG_INTR_EN	0x14
+#define REG_TEST_CTRL	0x18
+
+/* Per port registers */
+#define PORT_CFG_2	0x00
+#define PORT_ADDR	0x04
+#define PORT_DATA	0x08
+#define PORT_CFG_1	0x0c
+#define PORT_STATE	0x10
+
+#define SSP_PORT_CONFIG_MASK	(SSP_EARLY_DIN | SSP_DELAY_DOUT)
+#define SSP_PORT_CLKRATE_MASK	0x0f
+
+#define SSP_SEQRAM_WR_EN	BIT(4)
+#define SSP_SEQRAM_RD_EN	BIT(5)
+#define SSP_START		BIT(15)
+#define SSP_BUSY		BIT(10)
+#define SSP_PORT_ASL		BIT(7)
+#define SSP_PORT_CFO1		BIT(6)
+
+#define SSP_PORT_SEQRAM_SIZE	32
+
+static const int ssp_port_base[]   = {0x040, 0x080};
+static const int ssp_port_seqram[] = {0x100, 0x180};
+
+struct ti_ssp {
+	struct resource		*res;
+	struct device		*dev;
+	void __iomem		*regs;
+	spinlock_t		lock;
+	struct clk		*clk;
+	int			irq;
+	wait_queue_head_t	wqh;
+
+	/*
+	 * Some of the iosel2 register bits always read-back as 0, we need to
+	 * remember these values so that we don't clobber previously set
+	 * values.
+	 */
+	u32			iosel2;
+};
+
+static inline struct ti_ssp *dev_to_ssp(struct device *dev)
+{
+	return dev_get_drvdata(dev->parent);
+}
+
+static inline int dev_to_port(struct device *dev)
+{
+	return to_platform_device(dev)->id;
+}
+
+/* Register Access Helpers, rmw() functions need to run locked */
+static inline u32 ssp_read(struct ti_ssp *ssp, int reg)
+{
+	return __raw_readl(ssp->regs + reg);
+}
+
+static inline void ssp_write(struct ti_ssp *ssp, int reg, u32 val)
+{
+	__raw_writel(val, ssp->regs + reg);
+}
+
+static inline void ssp_rmw(struct ti_ssp *ssp, int reg, u32 mask, u32 bits)
+{
+	ssp_write(ssp, reg, (ssp_read(ssp, reg) & ~mask) | bits);
+}
+
+static inline u32 ssp_port_read(struct ti_ssp *ssp, int port, int reg)
+{
+	return ssp_read(ssp, ssp_port_base[port] + reg);
+}
+
+static inline void ssp_port_write(struct ti_ssp *ssp, int port, int reg,
+				  u32 val)
+{
+	ssp_write(ssp, ssp_port_base[port] + reg, val);
+}
+
+static inline void ssp_port_rmw(struct ti_ssp *ssp, int port, int reg,
+				u32 mask, u32 bits)
+{
+	ssp_rmw(ssp, ssp_port_base[port] + reg, mask, bits);
+}
+
+static inline void ssp_port_clr_bits(struct ti_ssp *ssp, int port, int reg,
+				     u32 bits)
+{
+	ssp_port_rmw(ssp, port, reg, bits, 0);
+}
+
+static inline void ssp_port_set_bits(struct ti_ssp *ssp, int port, int reg,
+				     u32 bits)
+{
+	ssp_port_rmw(ssp, port, reg, 0, bits);
+}
+
+/* Called to setup port clock mode, caller must hold ssp->lock */
+static int __set_mode(struct ti_ssp *ssp, int port, int mode)
+{
+	mode &= SSP_PORT_CONFIG_MASK;
+	ssp_port_rmw(ssp, port, PORT_CFG_1, SSP_PORT_CONFIG_MASK, mode);
+
+	return 0;
+}
+
+int ti_ssp_set_mode(struct device *dev, int mode)
+{
+	struct ti_ssp *ssp = dev_to_ssp(dev);
+	int port = dev_to_port(dev);
+	int ret;
+
+	spin_lock(&ssp->lock);
+	ret = __set_mode(ssp, port, mode);
+	spin_unlock(&ssp->lock);
+
+	return ret;
+}
+EXPORT_SYMBOL(ti_ssp_set_mode);
+
+/* Called to setup iosel2, caller must hold ssp->lock */
+static void __set_iosel2(struct ti_ssp *ssp, u32 mask, u32 val)
+{
+	ssp->iosel2 = (ssp->iosel2 & ~mask) | val;
+	ssp_write(ssp, REG_IOSEL_2, ssp->iosel2);
+}
+
+/* Called to setup port iosel, caller must hold ssp->lock */
+static void __set_iosel(struct ti_ssp *ssp, int port, u32 iosel)
+{
+	unsigned val, shift = port ? 16 : 0;
+
+	/* IOSEL1 gets the least significant 16 bits */
+	val = ssp_read(ssp, REG_IOSEL_1);
+	val &= 0xffff << (port ? 0 : 16);
+	val |= (iosel & 0xffff) << (port ? 16 : 0);
+	ssp_write(ssp, REG_IOSEL_1, val);
+
+	/* IOSEL2 gets the most significant 16 bits */
+	val = (iosel >> 16) & 0x7;
+	__set_iosel2(ssp, 0x7 << shift, val << shift);
+}
+
+int ti_ssp_set_iosel(struct device *dev, u32 iosel)
+{
+	struct ti_ssp *ssp = dev_to_ssp(dev);
+	int port = dev_to_port(dev);
+
+	spin_lock(&ssp->lock);
+	__set_iosel(ssp, port, iosel);
+	spin_unlock(&ssp->lock);
+
+	return 0;
+}
+EXPORT_SYMBOL(ti_ssp_set_iosel);
+
+int ti_ssp_load(struct device *dev, int offs, u32* prog, int len)
+{
+	struct ti_ssp *ssp = dev_to_ssp(dev);
+	int port = dev_to_port(dev);
+	int i;
+
+	if (len > SSP_PORT_SEQRAM_SIZE)
+		return -ENOSPC;
+
+	spin_lock(&ssp->lock);
+
+	/* Enable SeqRAM access */
+	ssp_port_set_bits(ssp, port, PORT_CFG_2, SSP_SEQRAM_WR_EN);
+
+	/* Copy code */
+	for (i = 0; i < len; i++) {
+		__raw_writel(prog[i], ssp->regs + offs + 4*i +
+			     ssp_port_seqram[port]);
+	}
+
+	/* Disable SeqRAM access */
+	ssp_port_clr_bits(ssp, port, PORT_CFG_2, SSP_SEQRAM_WR_EN);
+
+	spin_unlock(&ssp->lock);
+
+	return 0;
+}
+EXPORT_SYMBOL(ti_ssp_load);
+
+int ti_ssp_raw_read(struct device *dev)
+{
+	struct ti_ssp *ssp = dev_to_ssp(dev);
+	int port = dev_to_port(dev);
+	int shift = port ? 27 : 11;
+
+	return (ssp_read(ssp, REG_IOSEL_2) >> shift) & 0xf;
+}
+EXPORT_SYMBOL(ti_ssp_raw_read);
+
+int ti_ssp_raw_write(struct device *dev, u32 val)
+{
+	struct ti_ssp *ssp = dev_to_ssp(dev);
+	int port = dev_to_port(dev), shift;
+
+	spin_lock(&ssp->lock);
+
+	shift = port ? 22 : 6;
+	val &= 0xf;
+	__set_iosel2(ssp, 0xf << shift, val << shift);
+
+	spin_unlock(&ssp->lock);
+
+	return 0;
+}
+EXPORT_SYMBOL(ti_ssp_raw_write);
+
+static inline int __xfer_done(struct ti_ssp *ssp, int port)
+{
+	return !(ssp_port_read(ssp, port, PORT_CFG_1) & SSP_BUSY);
+}
+
+int ti_ssp_run(struct device *dev, u32 pc, u32 input, u32 *output)
+{
+	struct ti_ssp *ssp = dev_to_ssp(dev);
+	int port = dev_to_port(dev);
+	int ret;
+
+	if (pc & ~(0x3f))
+		return -EINVAL;
+
+	/* Grab ssp->lock to serialize rmw on ssp registers */
+	spin_lock(&ssp->lock);
+
+	ssp_port_write(ssp, port, PORT_ADDR, input >> 16);
+	ssp_port_write(ssp, port, PORT_DATA, input & 0xffff);
+	ssp_port_rmw(ssp, port, PORT_CFG_1, 0x3f, pc);
+
+	/* grab wait queue head lock to avoid race with the isr */
+	spin_lock_irq(&ssp->wqh.lock);
+
+	/* kick off sequence execution in hardware */
+	ssp_port_set_bits(ssp, port, PORT_CFG_1, SSP_START);
+
+	/* drop ssp lock; no register writes beyond this */
+	spin_unlock(&ssp->lock);
+
+	ret = wait_event_interruptible_locked_irq(ssp->wqh,
+						  __xfer_done(ssp, port));
+	spin_unlock_irq(&ssp->wqh.lock);
+
+	if (ret < 0)
+		return ret;
+
+	if (output) {
+		*output = (ssp_port_read(ssp, port, PORT_ADDR) << 16) |
+			  (ssp_port_read(ssp, port, PORT_DATA) &  0xffff);
+	}
+
+	ret = ssp_port_read(ssp, port, PORT_STATE) & 0x3f; /* stop address */
+
+	return ret;
+}
+EXPORT_SYMBOL(ti_ssp_run);
+
+static irqreturn_t ti_ssp_interrupt(int irq, void *dev_data)
+{
+	struct ti_ssp *ssp = dev_data;
+
+	spin_lock(&ssp->wqh.lock);
+
+	ssp_write(ssp, REG_INTR_ST, 0x3);
+	wake_up_locked(&ssp->wqh);
+
+	spin_unlock(&ssp->wqh.lock);
+
+	return IRQ_HANDLED;
+}
+
+static int __devinit ti_ssp_probe(struct platform_device *pdev)
+{
+	static struct ti_ssp *ssp;
+	const struct ti_ssp_data *pdata = pdev->dev.platform_data;
+	int error = 0, prediv = 0xff, id;
+	unsigned long sysclk;
+	struct device *dev = &pdev->dev;
+	struct mfd_cell cells[2];
+
+	ssp = kzalloc(sizeof(*ssp), GFP_KERNEL);
+	if (!ssp) {
+		dev_err(dev, "cannot allocate device info\n");
+		return -ENOMEM;
+	}
+
+	ssp->dev = dev;
+	dev_set_drvdata(dev, ssp);
+
+	ssp->res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!ssp->res) {
+		error = -ENODEV;
+		dev_err(dev, "cannot determine register area\n");
+		goto error_res;
+	}
+
+	if (!request_mem_region(ssp->res->start, resource_size(ssp->res),
+				pdev->name)) {
+		error = -ENOMEM;
+		dev_err(dev, "cannot claim register memory\n");
+		goto error_res;
+	}
+
+	ssp->regs = ioremap(ssp->res->start, resource_size(ssp->res));
+	if (!ssp->regs) {
+		error = -ENOMEM;
+		dev_err(dev, "cannot map register memory\n");
+		goto error_map;
+	}
+
+	ssp->clk = clk_get(dev, NULL);
+	if (IS_ERR(ssp->clk)) {
+		error = PTR_ERR(ssp->clk);
+		dev_err(dev, "cannot claim device clock\n");
+		goto error_clk;
+	}
+
+	ssp->irq = platform_get_irq(pdev, 0);
+	if (ssp->irq < 0) {
+		error = -ENODEV;
+		dev_err(dev, "unknown irq\n");
+		goto error_irq;
+	}
+
+	error = request_threaded_irq(ssp->irq, NULL, ti_ssp_interrupt, 0,
+				     dev_name(dev), ssp);
+	if (error < 0) {
+		dev_err(dev, "cannot acquire irq\n");
+		goto error_irq;
+	}
+
+	spin_lock_init(&ssp->lock);
+	init_waitqueue_head(&ssp->wqh);
+
+	/* Power on and initialize SSP */
+	error = clk_enable(ssp->clk);
+	if (error) {
+		dev_err(dev, "cannot enable device clock\n");
+		goto error_enable;
+	}
+
+	/* Reset registers to a sensible known state */
+	ssp_write(ssp, REG_IOSEL_1, 0);
+	ssp_write(ssp, REG_IOSEL_2, 0);
+	ssp_write(ssp, REG_INTR_EN, 0x3);
+	ssp_write(ssp, REG_INTR_ST, 0x3);
+	ssp_write(ssp, REG_TEST_CTRL, 0);
+	ssp_port_write(ssp, 0, PORT_CFG_1, SSP_PORT_ASL);
+	ssp_port_write(ssp, 1, PORT_CFG_1, SSP_PORT_ASL);
+	ssp_port_write(ssp, 0, PORT_CFG_2, SSP_PORT_CFO1);
+	ssp_port_write(ssp, 1, PORT_CFG_2, SSP_PORT_CFO1);
+
+	sysclk = clk_get_rate(ssp->clk);
+	if (pdata && pdata->out_clock)
+		prediv = (sysclk / pdata->out_clock) - 1;
+	prediv = clamp(prediv, 0, 0xff);
+	ssp_rmw(ssp, REG_PREDIV, 0xff, prediv);
+
+	memset(cells, 0, sizeof(cells));
+	for (id = 0; id < 2; id++) {
+		const struct ti_ssp_dev_data *data = &pdata->dev_data[id];
+
+		cells[id].id		= id;
+		cells[id].name		= data->dev_name;
+		cells[id].platform_data	= data->pdata;
+		cells[id].data_size	= data->pdata_size;
+	}
+
+	error = mfd_add_devices(dev, 0, cells, 2, NULL, 0);
+	if (error < 0) {
+		dev_err(dev, "cannot add mfd cells\n");
+		goto error_enable;
+	}
+
+	return 0;
+
+error_enable:
+	free_irq(ssp->irq, ssp);
+error_irq:
+	clk_put(ssp->clk);
+error_clk:
+	iounmap(ssp->regs);
+error_map:
+	release_mem_region(ssp->res->start, resource_size(ssp->res));
+error_res:
+	kfree(ssp);
+	return error;
+}
+
+static int __devexit ti_ssp_remove(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct ti_ssp *ssp = dev_get_drvdata(dev);
+
+	mfd_remove_devices(dev);
+	clk_disable(ssp->clk);
+	free_irq(ssp->irq, ssp);
+	clk_put(ssp->clk);
+	iounmap(ssp->regs);
+	release_mem_region(ssp->res->start, resource_size(ssp->res));
+	kfree(ssp);
+	dev_set_drvdata(dev, NULL);
+	return 0;
+}
+
+static struct platform_driver ti_ssp_driver = {
+	.probe		= ti_ssp_probe,
+	.remove		= __devexit_p(ti_ssp_remove),
+	.driver		= {
+		.name	= "ti-ssp",
+		.owner	= THIS_MODULE,
+	}
+};
+
+static int __init ti_ssp_init(void)
+{
+	return platform_driver_register(&ti_ssp_driver);
+}
+module_init(ti_ssp_init);
+
+static void __exit ti_ssp_exit(void)
+{
+	platform_driver_unregister(&ti_ssp_driver);
+}
+module_exit(ti_ssp_exit);
+
+MODULE_DESCRIPTION("Sequencer Serial Port (SSP) Driver");
+MODULE_AUTHOR("Cyril Chemparathy");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:ti-ssp");
include/linux/mfd/ti_ssp.h new file +87 -0
new file mode 100644
index 0000000..021fe09
@@ -0,0 +1,87 @@
+/*
+ * Sequencer Serial Port (SSP) driver for Texas Instruments' SoCs
+ *
+ * Copyright (C) 2010 Texas Instruments Inc
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#ifndef __TI_SSP_H__
+#define __TI_SSP_H__
+
+struct ti_ssp_dev_data {
+	const char	*dev_name;
+	void		*pdata;
+	size_t		pdata_size;
+};
+
+struct ti_ssp_data {
+	unsigned long		out_clock;
+	struct ti_ssp_dev_data	dev_data[2];
+};
+
+/*
+ * Sequencer port IO pin configuration bits.  These do not correlate 1-1 with
+ * the hardware.  The iosel field in the port data combines iosel1 and iosel2,
+ * and is therefore not a direct map to register space.  It is best to use the
+ * macros below to construct iosel values.
+ *
+ * least significant 16 bits --> iosel1
+ * most significant 16 bits  --> iosel2
+ */
+
+#define SSP_IN			0x0000
+#define SSP_DATA		0x0001
+#define SSP_CLOCK		0x0002
+#define SSP_CHIPSEL		0x0003
+#define SSP_OUT			0x0004
+#define SSP_PIN_SEL(pin, v)	((v) << ((pin) * 3))
+#define SSP_PIN_MASK(pin)	SSP_PIN_SEL(pin, 0x7)
+#define SSP_INPUT_SEL(pin)	((pin) << 16)
+
+/* Sequencer port config bits */
+#define SSP_EARLY_DIN		BIT(8)
+#define SSP_DELAY_DOUT		BIT(9)
+
+/* Sequence map definitions */
+#define SSP_CLK_HIGH		BIT(0)
+#define SSP_CLK_LOW		0
+#define SSP_DATA_HIGH		BIT(1)
+#define SSP_DATA_LOW		0
+#define SSP_CS_HIGH		BIT(2)
+#define SSP_CS_LOW		0
+#define SSP_OUT_MODE		BIT(3)
+#define SSP_IN_MODE		0
+#define SSP_DATA_REG		BIT(4)
+#define SSP_ADDR_REG		0
+
+#define SSP_OPCODE_DIRECT	((0x0) << 5)
+#define SSP_OPCODE_TOGGLE	((0x1) << 5)
+#define SSP_OPCODE_SHIFT	((0x2) << 5)
+#define SSP_OPCODE_BRANCH0	((0x4) << 5)
+#define SSP_OPCODE_BRANCH1	((0x5) << 5)
+#define SSP_OPCODE_BRANCH	((0x6) << 5)
+#define SSP_OPCODE_STOP		((0x7) << 5)
+#define SSP_BRANCH(addr)	((addr) << 8)
+#define SSP_COUNT(cycles)	((cycles) << 8)
+
+int ti_ssp_raw_read(struct device *dev);
+int ti_ssp_raw_write(struct device *dev, u32 val);
+int ti_ssp_load(struct device *dev, int offs, u32* prog, int len);
+int ti_ssp_run(struct device *dev, u32 pc, u32 input, u32 *output);
+int ti_ssp_set_mode(struct device *dev, int mode);
+int ti_ssp_set_iosel(struct device *dev, u32 iosel);
+
+#endif /* __TI_SSP_H__ */
#3 lore raw

PATCH v7 2/12 davinci: add tnetv107x ssp platform device

In reply to Cyril Chemparathy

This patch adds an SSP platform device definition for the tnetv107x soc family. The clock lookup entry has also been updated to match.

---
 arch/arm/mach-davinci/devices-tnetv107x.c      |   25 ++++++++++++++++++++++++
 arch/arm/mach-davinci/include/mach/tnetv107x.h |    2 +
 arch/arm/mach-davinci/tnetv107x.c              |    2 +-
 3 files changed, 28 insertions(+), 1 deletions(-)
arch/arm/mach-davinci/devices-tnetv107x.c +25 -0
index 85503de..6162cae 100644
@@ -35,6 +35,7 @@
 #define TNETV107X_SDIO0_BASE			0x08088700
 #define TNETV107X_SDIO1_BASE			0x08088800
 #define TNETV107X_KEYPAD_BASE			0x08088a00
+#define TNETV107X_SSP_BASE			0x08088c00
 #define TNETV107X_ASYNC_EMIF_CNTRL_BASE		0x08200000
 #define TNETV107X_ASYNC_EMIF_DATA_CE0_BASE	0x30000000
 #define TNETV107X_ASYNC_EMIF_DATA_CE1_BASE	0x40000000
@@ -342,6 +343,25 @@ static struct platform_device tsc_device = {
 	.resource	= tsc_resources,
 };
 
+static struct resource ssp_resources[] = {
+	{
+		.start	= TNETV107X_SSP_BASE,
+		.end	= TNETV107X_SSP_BASE + 0x1ff,
+		.flags	= IORESOURCE_MEM,
+	},
+	{
+		.start	= IRQ_TNETV107X_SSP,
+		.flags	= IORESOURCE_IRQ,
+	},
+};
+
+static struct platform_device ssp_device = {
+	.name		= "ti-ssp",
+	.id		= -1,
+	.num_resources	= ARRAY_SIZE(ssp_resources),
+	.resource	= ssp_resources,
+};
+
 void __init tnetv107x_devices_init(struct tnetv107x_device_info *info)
 {
 	int i, error;
@@ -380,4 +400,9 @@ void __init tnetv107x_devices_init(struct tnetv107x_device_info *info)
 		keypad_device.dev.platform_data = info->keypad_config;
 		platform_device_register(&keypad_device);
 	}
+
+	if (info->ssp_config) {
+		ssp_device.dev.platform_data = info->ssp_config;
+		platform_device_register(&ssp_device);
+	}
 }
arch/arm/mach-davinci/include/mach/tnetv107x.h +2 -0
index 5a681d8..89c1fdc 100644
@@ -34,6 +34,7 @@
 
 #include <linux/serial_8250.h>
 #include <linux/input/matrix_keypad.h>
+#include <linux/mfd/ti_ssp.h>
 
 #include <mach/mmc.h>
 #include <mach/nand.h>
@@ -44,6 +45,7 @@ struct tnetv107x_device_info {
 	struct davinci_mmc_config	*mmc_config[2];  /* 2 controllers */
 	struct davinci_nand_pdata	*nand_config[4]; /* 4 chipsels */
 	struct matrix_keypad_platform_data *keypad_config;
+	struct ti_ssp_data		*ssp_config;
 };
 
 extern struct platform_device tnetv107x_wdt_device;
arch/arm/mach-davinci/tnetv107x.c +1 -1
index 6fcdece..1b28fdd 100644
@@ -278,7 +278,7 @@ static struct clk_lookup clks[] = {
 	CLK(NULL,		"timer1",		&clk_timer1),
 	CLK("tnetv107x_wdt.0",	NULL,			&clk_wdt_arm),
 	CLK(NULL,		"clk_wdt_dsp",		&clk_wdt_dsp),
-	CLK("ti-ssp.0",		NULL,			&clk_ssp),
+	CLK("ti-ssp",		NULL,			&clk_ssp),
 	CLK(NULL,		"clk_tdm0",		&clk_tdm0),
 	CLK(NULL,		"clk_vlynq",		&clk_vlynq),
 	CLK(NULL,		"clk_mcdma",		&clk_mcdma),
Signature · 1 line
1.7.1
#4 lore raw

PATCH v7 3/12 davinci: add ssp config for tnetv107x evm board

In reply to Cyril Chemparathy

This patch adds SSP configuration and pin muxing info for tnetv107x evm boards.

---
 arch/arm/mach-davinci/board-tnetv107x-evm.c |   14 ++++++++++++++
 1 files changed, 14 insertions(+), 0 deletions(-)
arch/arm/mach-davinci/board-tnetv107x-evm.c +14 -0
index a6db854..ef526b1 100644
@@ -99,6 +99,12 @@ static const short uart1_pins[] __initdata = {
 	-1
 };
 
+static const short ssp_pins[] __initdata = {
+	TNETV107X_SSP0_0, TNETV107X_SSP0_1, TNETV107X_SSP0_2,
+	TNETV107X_SSP1_0, TNETV107X_SSP1_1, TNETV107X_SSP1_2,
+	TNETV107X_SSP1_3, -1
+};
+
 static struct mtd_partition nand_partitions[] = {
 	/* bootloader (U-Boot, etc) in first 12 sectors */
 	{
@@ -196,17 +202,25 @@ static struct matrix_keypad_platform_data keypad_config = {
 	.no_autorepeat	= 0,
 };
 
+static struct ti_ssp_data ssp_config = {
+	.out_clock	= 250 * 1000,
+	.dev_data	= {
+	},
+};
+
 static struct tnetv107x_device_info evm_device_info __initconst = {
 	.serial_config		= &serial_config,
 	.mmc_config[1]		= &mmc_config,	/* controller 1 */
 	.nand_config[0]		= &nand_config,	/* chip select 0 */
 	.keypad_config		= &keypad_config,
+	.ssp_config		= &ssp_config,
 };
 
 static __init void tnetv107x_evm_board_init(void)
 {
 	davinci_cfg_reg_list(sdio1_pins);
 	davinci_cfg_reg_list(uart1_pins);
+	davinci_cfg_reg_list(ssp_pins);
 
 	tnetv107x_devices_init(&evm_device_info);
 }
Signature · 1 line
1.7.1
#5 lore raw

PATCH v7 4/12 spi: add ti-ssp spi master driver

In reply to Cyril Chemparathy

This patch adds an SPI master implementation that operates on top of an underlying TI-SSP port.

---
 drivers/spi/Kconfig        |   10 +
 drivers/spi/Makefile       |    1 +
 drivers/spi/ti-ssp-spi.c   |  402 ++++++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/ti_ssp.h |    6 +
 4 files changed, 419 insertions(+), 0 deletions(-)
 create mode 100644 drivers/spi/ti-ssp-spi.c
drivers/spi/Kconfig +10 -0
index 78f9fd0..7f0ed2a 100644
@@ -336,6 +336,16 @@ config SPI_TEGRA
 	help
 	  SPI driver for NVidia Tegra SoCs
 
+config SPI_TI_SSP
+	tristate "TI Sequencer Serial Port - SPI Support"
+	depends on MFD_TI_SSP
+	help
+	  This selects an SPI master implementation using a TI sequencer
+	  serial port.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called ti-ssp-spi.
+
 config SPI_TOPCLIFF_PCH
 	tristate "Topcliff PCH SPI Controller"
 	depends on PCI
drivers/spi/Makefile +1 -0
index 8bc1a5a..595e5b8 100644
@@ -40,6 +40,7 @@ obj-$(CONFIG_SPI_S3C24XX_GPIO)		+= spi_s3c24xx_gpio.o
 obj-$(CONFIG_SPI_S3C24XX)		+= spi_s3c24xx_hw.o
 obj-$(CONFIG_SPI_S3C64XX)		+= spi_s3c64xx.o
 obj-$(CONFIG_SPI_TEGRA)			+= spi_tegra.o
+obj-$(CONFIG_SPI_TI_SSP)		+= ti-ssp-spi.o
 obj-$(CONFIG_SPI_TOPCLIFF_PCH)		+= spi_topcliff_pch.o
 obj-$(CONFIG_SPI_TXX9)			+= spi_txx9.o
 obj-$(CONFIG_SPI_XILINX)		+= xilinx_spi.o
drivers/spi/ti-ssp-spi.c new file +402 -0

402 lines not on this page · raw message

include/linux/mfd/ti_ssp.h +6 -0
index 021fe09..dbb4b43 100644
@@ -32,6 +32,12 @@ struct ti_ssp_data {
 	struct ti_ssp_dev_data	dev_data[2];
 };
 
+struct ti_ssp_spi_data {
+	unsigned long	iosel;
+	int		num_cs;
+	void		(*select)(int cs);
+};
+
 /*
  * Sequencer port IO pin configuration bits.  These do not correlate 1-1 with
  * the hardware.  The iosel field in the port data combines iosel1 and iosel2,
Signature · 1 line
1.7.1
#6 lore raw

PATCH v7 4/12 spi: add ti-ssp spi master driver

Quoted 4 lines · Cyril Chemparathy

This patch adds an SPI master implementation that operates on top of an underlying TI-SSP port.

Mostly looks okay, but I'm not thrilled with the ss scheme (see comments below).

g.

Quoted 69 lines
---
 drivers/spi/Kconfig        |   10 +
 drivers/spi/Makefile       |    1 +
 drivers/spi/ti-ssp-spi.c   |  402 ++++++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/ti_ssp.h |    6 +
 4 files changed, 419 insertions(+), 0 deletions(-)
 create mode 100644 drivers/spi/ti-ssp-spi.c
drivers/spi/Kconfig +10 -0
index 78f9fd0..7f0ed2a 100644
@@ -336,6 +336,16 @@ config SPI_TEGRA
 	help
 	  SPI driver for NVidia Tegra SoCs
 
+config SPI_TI_SSP
+	tristate "TI Sequencer Serial Port - SPI Support"
+	depends on MFD_TI_SSP
+	help
+	  This selects an SPI master implementation using a TI sequencer
+	  serial port.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called ti-ssp-spi.
+
 config SPI_TOPCLIFF_PCH
 	tristate "Topcliff PCH SPI Controller"
 	depends on PCI
drivers/spi/Makefile +1 -0
index 8bc1a5a..595e5b8 100644
@@ -40,6 +40,7 @@ obj-$(CONFIG_SPI_S3C24XX_GPIO)		+= spi_s3c24xx_gpio.o
 obj-$(CONFIG_SPI_S3C24XX)		+= spi_s3c24xx_hw.o
 obj-$(CONFIG_SPI_S3C64XX)		+= spi_s3c64xx.o
 obj-$(CONFIG_SPI_TEGRA)			+= spi_tegra.o
+obj-$(CONFIG_SPI_TI_SSP)		+= ti-ssp-spi.o
 obj-$(CONFIG_SPI_TOPCLIFF_PCH)		+= spi_topcliff_pch.o
 obj-$(CONFIG_SPI_TXX9)			+= spi_txx9.o
 obj-$(CONFIG_SPI_XILINX)		+= xilinx_spi.o
drivers/spi/ti-ssp-spi.c new file +22 -0
new file mode 100644
index 0000000..30fc0e2
@@ -0,0 +1,402 @@
+/*
+ * Sequencer Serial Port (SSP) based SPI master driver
+ *
+ * Copyright (C) 2010 Texas Instruments Inc
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include <linux/kernel.h>
+#include <linux/err.h>

drivers/spi/ti-ssp-spi.c line 32

Quoted 10 lines
drivers/spi/ti-ssp-spi.c new file +10 -0
+#include <linux/completion.h>
+#include <linux/delay.h>
+#include <linux/platform_device.h>
+#include <linux/spi/spi.h>
+#include <linux/mfd/ti_ssp.h>
+
+#define MODE_BITS	(SPI_CPHA | SPI_CPOL | SPI_CS_HIGH)
+
+struct ti_ssp_spi {
+	const struct ti_ssp_spi_data	*pdata;

Ideally the driver wouldn't keep the pdata pointer around at all. Instead, the relevant fields would be copied out of pdata and into ti_ssp_spi in the .probe() hook. From what I can tell, only the select function is being used for this purpose.

drivers/spi/ti-ssp-spi.c line 38

Quoted 6 lines
drivers/spi/ti-ssp-spi.c new file +6 -0
+	struct spi_master		*master;
+	struct device			*dev;
+	spinlock_t			lock;
+	struct list_head		msg_queue;
+	struct completion		complete;
+	int				shutdown:1;

Instead of a bitfield, Just make this a 'bool'.

drivers/spi/ti-ssp-spi.c line 46

Quoted 8 lines
drivers/spi/ti-ssp-spi.c new file +8 -0
+	struct workqueue_struct		*workqueue;
+	struct work_struct		work;
+	u8				mode, bpw;
+	int				cs_active;
+	u32				pc_en, pc_dis, pc_wr, pc_rd;
+};
+
+static u32 do_read_data(struct ti_ssp_spi *hw)

As commented in my other reply. All the static private functions still need the ti_ssp_spi_ prefix.

Quoted 145 lines
drivers/spi/ti-ssp-spi.c new file +145 -0
+{
+	u32 ret;
+
+	ti_ssp_run(hw->dev, hw->pc_rd, 0, &ret);
+	return ret;
+}
+
+static void do_write_data(struct ti_ssp_spi *hw, u32 data)
+{
+	ti_ssp_run(hw->dev, hw->pc_wr, data << (32 - hw->bpw), NULL);
+}
+
+static int do_transfer(struct ti_ssp_spi *hw, struct spi_message *msg,
+		       struct spi_transfer *t)
+{
+	int count;
+
+	if (hw->bpw <= 8) {
+		u8		*rx = t->rx_buf;
+		const u8	*tx = t->tx_buf;
+
+		for (count = 0; count < t->len; count += 1) {
+			if (t->tx_buf)
+				do_write_data(hw, *tx++);
+			if (t->rx_buf)
+				*rx++ = do_read_data(hw);
+		}
+	} else if (hw->bpw <= 16) {
+		u16		*rx = t->rx_buf;
+		const u16	*tx = t->tx_buf;
+
+		for (count = 0; count < t->len; count += 2) {
+			if (t->tx_buf)
+				do_write_data(hw, *tx++);
+			if (t->rx_buf)
+				*rx++ = do_read_data(hw);
+		}
+	} else {
+		u32		*rx = t->rx_buf;
+		const u32	*tx = t->tx_buf;
+
+		for (count = 0; count < t->len; count += 4) {
+			if (t->tx_buf)
+				do_write_data(hw, *tx++);
+			if (t->rx_buf)
+				*rx++ = do_read_data(hw);
+		}
+	}
+
+	msg->actual_length += count; /* bytes transferred */
+
+	dev_dbg(&msg->spi->dev, "xfer %s%s, %d bytes, %d bpw, count %d%s\n",
+		t->tx_buf ? "tx" : "", t->rx_buf ? "rx" : "", t->len,
+		hw->bpw, count, (count < t->len) ? " (under)" : "");
+
+	return (count < t->len) ? -EIO : 0; /* left over data */
+}
+
+static void chip_select(struct ti_ssp_spi *hw, int cs_active)
+{
+	cs_active = !!cs_active;
+	if (cs_active == hw->cs_active)
+		return;
+	ti_ssp_run(hw->dev, cs_active ? hw->pc_en : hw->pc_dis, 0, NULL);
+	hw->cs_active = cs_active;
+}
+
+#define __SHIFT_OUT(bits)	(SSP_OPCODE_SHIFT | SSP_OUT_MODE | \
+				 cs_en | clk | SSP_COUNT((bits) * 2 - 1))
+#define __SHIFT_IN(bits)	(SSP_OPCODE_SHIFT | SSP_IN_MODE  | \
+				 cs_en | clk | SSP_COUNT((bits) * 2 - 1))
+
+static int setup_xfer(struct ti_ssp_spi *hw, u8 bpw, u8 mode)
+{
+	int error, idx = 0;
+	u32 seqram[16];
+	u32 cs_en, cs_dis, clk;
+	u32 topbits, botbits;
+
+	mode &= MODE_BITS;
+	if (mode == hw->mode && bpw == hw->bpw)
+		return 0;
+
+	cs_en  = (mode & SPI_CS_HIGH) ? SSP_CS_HIGH : SSP_CS_LOW;
+	cs_dis = (mode & SPI_CS_HIGH) ? SSP_CS_LOW  : SSP_CS_HIGH;
+	clk    = (mode & SPI_CPOL)    ? SSP_CLK_HIGH : SSP_CLK_LOW;
+
+	/* Construct instructions */
+
+	/* Disable Chip Select */
+	hw->pc_dis = idx;
+	seqram[idx++] = SSP_OPCODE_DIRECT | SSP_OUT_MODE | cs_dis | clk;
+	seqram[idx++] = SSP_OPCODE_STOP   | SSP_OUT_MODE | cs_dis | clk;
+
+	/* Enable Chip Select */
+	hw->pc_en = idx;
+	seqram[idx++] = SSP_OPCODE_DIRECT | SSP_OUT_MODE | cs_en | clk;
+	seqram[idx++] = SSP_OPCODE_STOP   | SSP_OUT_MODE | cs_en | clk;
+
+	/* Reads and writes need to be split for bpw > 16 */
+	topbits = (bpw > 16) ? 16 : bpw;
+	botbits = bpw - topbits;
+
+	/* Write */
+	hw->pc_wr = idx;
+	seqram[idx++] = __SHIFT_OUT(topbits) | SSP_ADDR_REG;
+	if (botbits)
+		seqram[idx++] = __SHIFT_OUT(botbits)  | SSP_DATA_REG;
+	seqram[idx++] = SSP_OPCODE_STOP | SSP_OUT_MODE | cs_en | clk;
+
+	/* Read */
+	hw->pc_rd = idx;
+	if (botbits)
+		seqram[idx++] = __SHIFT_IN(botbits) | SSP_ADDR_REG;
+	seqram[idx++] = __SHIFT_IN(topbits) | SSP_DATA_REG;
+	seqram[idx++] = SSP_OPCODE_STOP | SSP_OUT_MODE | cs_en | clk;
+
+	error = ti_ssp_load(hw->dev, 0, seqram, idx);
+	if (error < 0)
+		return error;
+
+	error = ti_ssp_set_mode(hw->dev, ((mode & SPI_CPHA) ?
+					  0 : SSP_EARLY_DIN));
+	if (error < 0)
+		return error;
+
+	hw->bpw = bpw;
+	hw->mode = mode;
+
+	return error;
+}
+
+static void ti_ssp_spi_work(struct work_struct *work)
+{
+	struct ti_ssp_spi *hw = container_of(work, struct ti_ssp_spi, work);
+
+	spin_lock(&hw->lock);
+
+	 while (!list_empty(&hw->msg_queue)) {
+		struct spi_message	*m;
+		struct spi_device	*spi;
+		struct spi_transfer	*t = NULL;
+		int			status = 0;
+
+		m = container_of(hw->msg_queue.next, struct spi_message,

drivers/spi/ti-ssp-spi.c line 201

Quoted 10 lines
drivers/spi/ti-ssp-spi.c new file +10 -0
+				 queue);
+
+		list_del_init(&m->queue);
+
+		spin_unlock(&hw->lock);
+
+		spi = m->spi;
+
+		if (hw->pdata->select)
+			hw->pdata->select(spi->chip_select);

Rather than yet another driver-specific slave select function being defined, I'd be happier to see the driver accept a list of gpio numbers to use for non-native ss lines. Using the gpio infrastructure reduces the amount of custom code needed in each bsp.

Quoted 221 lines
drivers/spi/ti-ssp-spi.c new file +201 -0
+
+		list_for_each_entry(t, &m->transfers, transfer_list) {
+			int bpw = spi->bits_per_word;
+			int xfer_status;
+
+			if (t->bits_per_word)
+				bpw = t->bits_per_word;
+
+			if (setup_xfer(hw, bpw, spi->mode) < 0)
+				break;
+
+			chip_select(hw, 1);
+
+			xfer_status = do_transfer(hw, m, t);
+			if (xfer_status < 0)
+				status = xfer_status;
+
+			if (t->delay_usecs)
+				udelay(t->delay_usecs);
+
+			if (t->cs_change)
+				chip_select(hw, 0);
+		}
+
+		chip_select(hw, 0);
+		m->status = status;
+		m->complete(m->context);
+
+		spin_lock(&hw->lock);
+	}
+
+	if (hw->shutdown)
+		complete(&hw->complete);
+
+	spin_unlock(&hw->lock);
+}
+
+static int ti_ssp_spi_setup(struct spi_device *spi)
+{
+	if (spi->bits_per_word > 32)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int ti_ssp_spi_transfer(struct spi_device *spi, struct spi_message *m)
+{
+	struct ti_ssp_spi	*hw;
+	struct spi_transfer	*t;
+	int			error = 0;
+
+	m->actual_length = 0;
+	m->status = -EINPROGRESS;
+
+	hw = spi_master_get_devdata(spi->master);
+
+	if (list_empty(&m->transfers) || !m->complete)
+		return -EINVAL;
+
+	list_for_each_entry(t, &m->transfers, transfer_list) {
+		if (t->len && !(t->rx_buf || t->tx_buf)) {
+			dev_err(&spi->dev, "invalid xfer, no buffer\n");
+			return -EINVAL;
+		}
+
+		if (t->len && t->rx_buf && t->tx_buf) {
+			dev_err(&spi->dev, "invalid xfer, full duplex\n");
+			return -EINVAL;
+		}
+
+		if (t->bits_per_word > 32) {
+			dev_err(&spi->dev, "invalid xfer width %d\n",
+				t->bits_per_word);
+			return -EINVAL;
+		}
+	}
+
+	spin_lock(&hw->lock);
+	if (hw->shutdown) {
+		error = -ESHUTDOWN;
+		goto error_unlock;
+	}
+	list_add_tail(&m->queue, &hw->msg_queue);
+	queue_work(hw->workqueue, &hw->work);
+error_unlock:
+	spin_unlock(&hw->lock);
+	return error;
+}
+
+static int __devinit ti_ssp_spi_probe(struct platform_device *pdev)
+{
+	const struct ti_ssp_spi_data *pdata;
+	struct ti_ssp_spi *hw;
+	struct spi_master *master;
+	struct device *dev = &pdev->dev;
+	int error = 0;
+
+	pdata = dev->platform_data;
+	if (!pdata) {
+		dev_err(dev, "platform data not found\n");
+		return -EINVAL;
+	}
+
+	master = spi_alloc_master(dev, sizeof(struct ti_ssp_spi));
+	if (!master) {
+		dev_err(dev, "cannot allocate SPI master\n");
+		return -ENOMEM;
+	}
+
+	hw = spi_master_get_devdata(master);
+	platform_set_drvdata(pdev, hw);
+
+	hw->master = master;
+	hw->dev = dev;
+	hw->pdata = pdata;
+
+	spin_lock_init(&hw->lock);
+	init_completion(&hw->complete);
+	INIT_LIST_HEAD(&hw->msg_queue);
+	INIT_WORK(&hw->work, ti_ssp_spi_work);
+
+	hw->workqueue = create_singlethread_workqueue(dev_name(dev));
+	if (!hw->workqueue) {
+		error = -ENOMEM;
+		dev_err(dev, "work queue creation failed\n");
+		goto error_wq;
+	}
+
+	error = ti_ssp_set_iosel(hw->dev, hw->pdata->iosel);
+	if (error < 0) {
+		dev_err(dev, "io setup failed\n");
+		goto error_iosel;
+	}
+
+	master->bus_num		= pdev->id;
+	master->num_chipselect	= hw->pdata->num_cs;
+	master->mode_bits	= MODE_BITS;
+	master->flags		= SPI_MASTER_HALF_DUPLEX;
+	master->setup		= ti_ssp_spi_setup;
+	master->transfer	= ti_ssp_spi_transfer;
+
+	error = spi_register_master(master);
+	if (error) {
+		dev_err(dev, "master registration failed\n");
+		goto error_reg;
+	}
+
+	return 0;
+
+error_reg:
+error_iosel:
+	destroy_workqueue(hw->workqueue);
+error_wq:
+	spi_master_put(master);
+	return error;
+}
+
+static int __devexit ti_ssp_spi_remove(struct platform_device *pdev)
+{
+	struct ti_ssp_spi *hw = platform_get_drvdata(pdev);
+	int error;
+
+	hw->shutdown = 1;
+	while (!list_empty(&hw->msg_queue)) {
+		error = wait_for_completion_interruptible(&hw->complete);
+		if (error < 0) {
+			hw->shutdown = 0;
+			return error;
+		}
+	}
+	destroy_workqueue(hw->workqueue);
+	spi_unregister_master(hw->master);
+
+	return 0;
+}
+
+static struct platform_driver ti_ssp_spi_driver = {
+	.probe		= ti_ssp_spi_probe,
+	.remove		= __devexit_p(ti_ssp_spi_remove),
+	.driver		= {
+		.name	= "ti-ssp-spi",
+		.owner	= THIS_MODULE,
+	},
+};
+
+static int __init ti_ssp_spi_init(void)
+{
+	return platform_driver_register(&ti_ssp_spi_driver);
+}
+module_init(ti_ssp_spi_init);
+
+static void __exit ti_ssp_spi_exit(void)
+{
+	platform_driver_unregister(&ti_ssp_spi_driver);
+}
+module_exit(ti_ssp_spi_exit);
+
+MODULE_DESCRIPTION("SSP SPI Master");
+MODULE_AUTHOR("Cyril Chemparathy");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:ti-ssp-spi");
include/linux/mfd/ti_ssp.h +6 -0
index 021fe09..dbb4b43 100644
@@ -32,6 +32,12 @@ struct ti_ssp_data {
 	struct ti_ssp_dev_data	dev_data[2];
 };
 
+struct ti_ssp_spi_data {
+	unsigned long	iosel;
+	int		num_cs;
+	void		(*select)(int cs);
+};
+
 /*
  * Sequencer port IO pin configuration bits.  These do not correlate 1-1 with
  * the hardware.  The iosel field in the port data combines iosel1 and iosel2,
Signature · 1 line
1.7.1
#7 lore raw

PATCH v7 5/12 davinci: add spi devices on tnetv107x evm

In reply to Cyril Chemparathy

This patch adds definitions for spi devices on the tnetv107x evm platform.

---
 arch/arm/mach-davinci/board-tnetv107x-evm.c |   43 +++++++++++++++++++++++++++
 1 files changed, 43 insertions(+), 0 deletions(-)
arch/arm/mach-davinci/board-tnetv107x-evm.c +43 -0
index ef526b1..1a656e8 100644
@@ -25,6 +25,7 @@
 #include <linux/mtd/partitions.h>
 #include <linux/input.h>
 #include <linux/input/matrix_keypad.h>
+#include <linux/spi/spi.h>
 
 #include <asm/mach/arch.h>
 #include <asm/mach-types.h>
@@ -37,6 +38,7 @@
 
 #define EVM_MMC_WP_GPIO		21
 #define EVM_MMC_CD_GPIO		24
+#define EVM_SPI_CS_GPIO		54
 
 static int initialize_gpio(int gpio, char *desc)
 {
@@ -202,9 +204,45 @@ static struct matrix_keypad_platform_data keypad_config = {
 	.no_autorepeat	= 0,
 };
 
+static void spi_select_device(int cs)
+{
+	static int gpio;
+
+	if (!gpio) {
+		int ret;
+		ret = gpio_request(EVM_SPI_CS_GPIO, "spi chipsel");
+		if (ret < 0) {
+			pr_err("cannot open spi chipsel gpio\n");
+			gpio = -ENOSYS;
+			return;
+		} else {
+			gpio = EVM_SPI_CS_GPIO;
+			gpio_direction_output(gpio, 0);
+		}
+	}
+
+	if (gpio < 0)
+		return;
+
+	return gpio_set_value(gpio, cs ? 1 : 0);
+}
+
+static struct ti_ssp_spi_data spi_master_data = {
+	.num_cs	= 2,
+	.select	= spi_select_device,
+	.iosel	= SSP_PIN_SEL(0, SSP_CLOCK)	| SSP_PIN_SEL(1, SSP_DATA) |
+		  SSP_PIN_SEL(2, SSP_CHIPSEL)	| SSP_PIN_SEL(3, SSP_IN)   |
+		  SSP_INPUT_SEL(3),
+};
+
 static struct ti_ssp_data ssp_config = {
 	.out_clock	= 250 * 1000,
 	.dev_data	= {
+		[1] = {
+			.dev_name = "ti-ssp-spi",
+			.pdata = &spi_master_data,
+			.pdata_size = sizeof(spi_master_data),
+		},
 	},
 };
 
@@ -216,6 +254,9 @@ static struct tnetv107x_device_info evm_device_info __initconst = {
 	.ssp_config		= &ssp_config,
 };
 
+static struct spi_board_info spi_info[] __initconst = {
+};
+
 static __init void tnetv107x_evm_board_init(void)
 {
 	davinci_cfg_reg_list(sdio1_pins);
@@ -223,6 +264,8 @@ static __init void tnetv107x_evm_board_init(void)
 	davinci_cfg_reg_list(ssp_pins);
 
 	tnetv107x_devices_init(&evm_device_info);
+
+	spi_register_board_info(spi_info, ARRAY_SIZE(spi_info));
 }
 
 #ifdef CONFIG_SERIAL_8250_CONSOLE
Signature · 1 line
1.7.1
#8 lore raw

PATCH v7 6/12 regulator: add driver for tps6524x regulator

In reply to Cyril Chemparathy

TPS6524X provides three step-down converters and two general-purpose LDO voltage regulators. This device is interfaced using SPI.

---
 drivers/regulator/Kconfig              |   10 +
 drivers/regulator/Makefile             |    1 +
 drivers/regulator/tps6524x-regulator.c |  692 ++++++++++++++++++++++++++++++++
 3 files changed, 703 insertions(+), 0 deletions(-)
 create mode 100644 drivers/regulator/tps6524x-regulator.c
drivers/regulator/Kconfig +10 -0
index dd30e88..da34981 100644
@@ -250,5 +250,15 @@ config REGULATOR_TPS6586X
 	help
 	  This driver supports TPS6586X voltage regulator chips.
 
+config REGULATOR_TPS6524X
+	tristate "TI TPS6524X Power regulators"
+	depends on SPI
+	help
+	  This driver supports TPS6524X voltage regulator chips. TPS6524X
+	  provides three step-down converters and two general-purpose LDO
+	  voltage regulators.  This device is interfaced using a customized
+	  serial interface currently supported on the sequencer serial
+	  port controller.
+
 endif
 
drivers/regulator/Makefile +1 -0
index bff8157..cf71df7 100644
@@ -34,6 +34,7 @@ obj-$(CONFIG_REGULATOR_AB3100) += ab3100.o
 
 obj-$(CONFIG_REGULATOR_TPS65023) += tps65023-regulator.o
 obj-$(CONFIG_REGULATOR_TPS6507X) += tps6507x-regulator.o
+obj-$(CONFIG_REGULATOR_TPS6524X) += tps6524x-regulator.o
 obj-$(CONFIG_REGULATOR_88PM8607) += 88pm8607.o
 obj-$(CONFIG_REGULATOR_ISL6271A) += isl6271a-regulator.o
 obj-$(CONFIG_REGULATOR_AB8500)	+= ab8500.o
drivers/regulator/tps6524x-regulator.c new file +692 -0

692 lines not on this page · raw message

Signature · 1 line
1.7.1
#9 lore raw

PATCH v7 6/12 regulator: add driver for tps6524x regulator

Review

Quoted 2 lines · Cyril Chemparathy
(quoted diff) +2 -0
+static int set_voltage(struct regulator_dev *rdev, int min_uV, int max_uV)
+{

Updates in the regulator tree mean that this will no longer build in -next. Given the difficulty that this series seems to be having perhaps it would make sense to split out drivers such as this which have nothing to do with SSP itself and get them into the next merge window?

#10 lore raw

PATCH v7 6/12 regulator: add driver for tps6524x regulator

Quoted 9 lines · Mark Brown

Review

Quoted 2 lines · Cyril Chemparathy
(quoted diff) +2 -0
+static int set_voltage(struct regulator_dev *rdev, int min_uV, int max_uV)
+{

Updates in the regulator tree mean that this will no longer build in -next. Given the difficulty that this series seems to be having perhaps it would make sense to split out drivers such as this which have nothing to do with SSP itself and get them into the next merge window?

Agreed. I will split this series up and repost.

#11 lore raw

PATCH v7 7/12 davinci: add tnetv107x evm regulators

In reply to Cyril Chemparathy

This patch adds regulator and spi board info definitions for the tps6524x power management IC found on tnetv107x evm boards.

---
 arch/arm/mach-davinci/board-tnetv107x-evm.c |   85 +++++++++++++++++++++++++++
 1 files changed, 85 insertions(+), 0 deletions(-)
arch/arm/mach-davinci/board-tnetv107x-evm.c +85 -0
index 1a656e8..ca23516 100644
@@ -26,6 +26,9 @@
 #include <linux/input.h>
 #include <linux/input/matrix_keypad.h>
 #include <linux/spi/spi.h>
+#include <linux/regulator/machine.h>
+#include <linux/regulator/consumer.h>
+#include <linux/regulator/driver.h>
 
 #include <asm/mach/arch.h>
 #include <asm/mach-types.h>
@@ -254,7 +257,89 @@ static struct tnetv107x_device_info evm_device_info __initconst = {
 	.ssp_config		= &ssp_config,
 };
 
+static struct regulator_consumer_supply usb_consumers[] = {
+	REGULATOR_SUPPLY("vbus", "musb_hdrc.1"),
+};
+
+static struct regulator_consumer_supply lcd_consumers[] = {
+	REGULATOR_SUPPLY("vlcd", "tps6116x"),
+};
+
+static struct regulator_init_data regulators[] = {
+	{
+		.constraints		= {
+			.name		= "DCDC1",
+			.min_uV		= 1000000,
+			.max_uV		= 1000000,
+			.always_on	= 1,
+			.boot_on	= 1,
+		},
+	},
+	{
+		.constraints		= {
+			.name		= "DCDC2",
+			.min_uV		= 1800000,
+			.max_uV		= 1800000,
+			.always_on	= 1,
+			.boot_on	= 1,
+		},
+	},
+	{
+		.constraints		= {
+			.name		= "DCDC3",
+			.min_uV		= 3300000,
+			.max_uV		= 3300000,
+			.always_on	= 1,
+			.boot_on	= 1,
+		},
+	},
+	{
+		.constraints		= {
+			.name		= "LDO1",
+			.min_uV		= 4800000,
+			.max_uV		= 4800000,
+			.always_on	= 1,
+			.boot_on	= 1,
+		},
+	},
+	{
+		.constraints		= {
+			.name		= "LDO1",
+			.min_uV		= 3300000,
+			.max_uV		= 3300000,
+			.always_on	= 1,
+			.boot_on	= 1,
+		},
+	},
+	{
+		.num_consumer_supplies	= ARRAY_SIZE(usb_consumers),
+		.consumer_supplies	= usb_consumers,
+		.constraints		= {
+			.name		= "USB",
+			.min_uA		= 200000,
+			.max_uA		= 1000000,
+			.valid_ops_mask	= REGULATOR_CHANGE_CURRENT |
+					  REGULATOR_CHANGE_STATUS,
+		},
+	},
+	{
+		.num_consumer_supplies	= ARRAY_SIZE(lcd_consumers),
+		.consumer_supplies	= lcd_consumers,
+		.constraints		= {
+			.name		= "LCD",
+			.valid_ops_mask	= REGULATOR_CHANGE_STATUS,
+		},
+	},
+};
+
 static struct spi_board_info spi_info[] __initconst = {
+	{
+		.modalias	= "tps6524x",
+		.bus_num	= 1,
+		.chip_select	= 0,
+		.mode		= SPI_MODE_0,
+		.platform_data	= regulators,
+	},
 };
 
 static __init void tnetv107x_evm_board_init(void)
Signature · 1 line
1.7.1
#12 lore raw

PATCH v7 8/12 gpio: add ti-ssp gpio driver

In reply to Cyril Chemparathy

TI's SSP controller pins can be directly read and written to behave like a GPIO. This patch adds a GPIO driver that exposes such functionality.

---
 drivers/gpio/Kconfig       |   10 ++
 drivers/gpio/Makefile      |    1 +
 drivers/gpio/ti-ssp-gpio.c |  206 ++++++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/ti_ssp.h |    4 +
 4 files changed, 221 insertions(+), 0 deletions(-)
 create mode 100644 drivers/gpio/ti-ssp-gpio.c
drivers/gpio/Kconfig +10 -0
index 3143ac7..05bbe4c 100644
@@ -128,6 +128,16 @@ config GPIO_VX855
 	  additional drivers must be enabled in order to use the
 	  functionality of the device.
 
+config GPIO_TI_SSP
+	tristate "TI Sequencer Serial Port - GPIO Support"
+	depends on MFD_TI_SSP
+	help
+	  Say yes here to support a GPIO interface on TI SSP port pins.
+	  Each SSP port translates into 4 GPIOs.
+
+	  This driver can also be built as a module. If so, the module
+	  will be called ti-ssp-gpio.
+
 comment "I2C GPIO expanders:"
 
 config GPIO_MAX7300
drivers/gpio/Makefile +1 -0
index bdf3dde..0e2a844 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_GPIO_PL061)	+= pl061.o
 obj-$(CONFIG_GPIO_STMPE)	+= stmpe-gpio.o
 obj-$(CONFIG_GPIO_TC35892)	+= tc35892-gpio.o
 obj-$(CONFIG_GPIO_TIMBERDALE)	+= timbgpio.o
+obj-$(CONFIG_GPIO_TI_SSP)	+= ti-ssp-gpio.o
 obj-$(CONFIG_GPIO_TWL4030)	+= twl4030-gpio.o
 obj-$(CONFIG_GPIO_UCB1400)	+= ucb1400_gpio.o
 obj-$(CONFIG_GPIO_XILINX)	+= xilinx_gpio.o
drivers/gpio/ti-ssp-gpio.c new file +206 -0

206 lines not on this page · raw message

include/linux/mfd/ti_ssp.h +4 -0
index dbb4b43..10c65bb 100644
@@ -38,6 +38,10 @@ struct ti_ssp_spi_data {
 	void		(*select)(int cs);
 };
 
+struct ti_ssp_gpio_data {
+	int		start;
+};
+
 /*
  * Sequencer port IO pin configuration bits.  These do not correlate 1-1 with
  * the hardware.  The iosel field in the port data combines iosel1 and iosel2,
Signature · 1 line
1.7.1
#13 lore raw

PATCH v7 8/12 gpio: add ti-ssp gpio driver

Quoted 99 lines · Cyril Chemparathy

TI's SSP controller pins can be directly read and written to behave like a GPIO. This patch adds a GPIO driver that exposes such functionality.

---
 drivers/gpio/Kconfig       |   10 ++
 drivers/gpio/Makefile      |    1 +
 drivers/gpio/ti-ssp-gpio.c |  206 ++++++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/ti_ssp.h |    4 +
 4 files changed, 221 insertions(+), 0 deletions(-)
 create mode 100644 drivers/gpio/ti-ssp-gpio.c
drivers/gpio/Kconfig +10 -0
index 3143ac7..05bbe4c 100644
@@ -128,6 +128,16 @@ config GPIO_VX855
 	  additional drivers must be enabled in order to use the
 	  functionality of the device.
 
+config GPIO_TI_SSP
+	tristate "TI Sequencer Serial Port - GPIO Support"
+	depends on MFD_TI_SSP
+	help
+	  Say yes here to support a GPIO interface on TI SSP port pins.
+	  Each SSP port translates into 4 GPIOs.
+
+	  This driver can also be built as a module. If so, the module
+	  will be called ti-ssp-gpio.
+
 comment "I2C GPIO expanders:"
 
 config GPIO_MAX7300
drivers/gpio/Makefile +1 -0
index bdf3dde..0e2a844 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_GPIO_PL061)	+= pl061.o
 obj-$(CONFIG_GPIO_STMPE)	+= stmpe-gpio.o
 obj-$(CONFIG_GPIO_TC35892)	+= tc35892-gpio.o
 obj-$(CONFIG_GPIO_TIMBERDALE)	+= timbgpio.o
+obj-$(CONFIG_GPIO_TI_SSP)	+= ti-ssp-gpio.o
 obj-$(CONFIG_GPIO_TWL4030)	+= twl4030-gpio.o
 obj-$(CONFIG_GPIO_UCB1400)	+= ucb1400_gpio.o
 obj-$(CONFIG_GPIO_XILINX)	+= xilinx_gpio.o
drivers/gpio/ti-ssp-gpio.c new file +48 -0
new file mode 100644
index 0000000..387970e
@@ -0,0 +1,206 @@
+/*
+ * Sequencer Serial Port (SSP) based virtual GPIO driver
+ *
+ * Copyright (C) 2010 Texas Instruments Inc
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/err.h>
+#include <linux/slab.h>
+#include <linux/gpio.h>
+#include <linux/platform_device.h>
+#include <linux/mfd/ti_ssp.h>
+
+struct ti_ssp_gpio_chip {
+	struct gpio_chip		chip;
+	struct device			*dev;
+	spinlock_t			lock;
+	u8				out;
+	u32				iosel;
+};
+
+#define to_ssp_gpio_chip(c) container_of(c, struct ti_ssp_gpio_chip, chip)
+
+static int direction_in(struct gpio_chip *chip, unsigned gpio_num)
+{
+	struct ti_ssp_gpio_chip *gpio = to_ssp_gpio_chip(chip);
+	int error = 0;
+
+	spin_lock(&gpio->lock);
+
+	gpio->iosel &= ~SSP_PIN_MASK(gpio_num);

drivers/gpio/ti-ssp-gpio.c line 58

Quoted 10 lines
drivers/gpio/ti-ssp-gpio.c new file +10 -0
+	gpio->iosel |=  SSP_PIN_SEL(gpio_num, SSP_IN);
+
+	error = ti_ssp_set_iosel(gpio->dev, gpio->iosel);
+
+	spin_unlock(&gpio->lock);
+
+	return error;
+}
+
+static int direction_out(struct gpio_chip *chip, unsigned gpio_num, int val)

All of these static functions still need to have the driver-specific prefix to avoid any possibility of collisions in the global namespace. I haven't checked the rest of the patches, but this comment does stand for the whole series.

#14 lore raw

PATCH v7 9/12 davinci: add tnetv107x evm ti-ssp gpio device

In reply to Cyril Chemparathy

This patch adds definitions to hook up one of the ti-ssp ports to the SSP GPIO driver.

---
 arch/arm/mach-davinci/board-tnetv107x-evm.c |   11 +++++++++++
 1 files changed, 11 insertions(+), 0 deletions(-)
arch/arm/mach-davinci/board-tnetv107x-evm.c +11 -0
index ca23516..e3863dd 100644
@@ -39,6 +39,8 @@
 #include <mach/cp_intc.h>
 #include <mach/tnetv107x.h>
 
+#define SSP_GPIO_START		128
+
 #define EVM_MMC_WP_GPIO		21
 #define EVM_MMC_CD_GPIO		24
 #define EVM_SPI_CS_GPIO		54
@@ -238,9 +240,18 @@ static struct ti_ssp_spi_data spi_master_data = {
 		  SSP_INPUT_SEL(3),
 };
 
+static struct ti_ssp_gpio_data ssp_gpio_data = {
+	.start		= SSP_GPIO_START,
+};
+
 static struct ti_ssp_data ssp_config = {
 	.out_clock	= 250 * 1000,
 	.dev_data	= {
+		[0] = {
+			.dev_name = "ti-ssp-gpio",
+			.pdata = &ssp_gpio_data,
+			.pdata_size = sizeof(ssp_gpio_data),
+		},
 		[1] = {
 			.dev_name = "ti-ssp-spi",
 			.pdata = &spi_master_data,
Signature · 1 line
1.7.1
#15 lore raw

PATCH v7 10/12 backlight: add support for tps6116x controller

In reply to Cyril Chemparathy

TPS6116x is an EasyScale backlight controller device. This driver supports TPS6116x devices connected on a single GPIO.

---
 drivers/video/backlight/Kconfig    |    7 +
 drivers/video/backlight/Makefile   |    2 +-
 drivers/video/backlight/tps6116x.c |  299 ++++++++++++++++++++++++++++++++++++
 3 files changed, 307 insertions(+), 1 deletions(-)
 create mode 100644 drivers/video/backlight/tps6116x.c
drivers/video/backlight/Kconfig +7 -0
index e54a337..06e868e 100644
@@ -307,6 +307,13 @@ config BACKLIGHT_PCF50633
 	  If you have a backlight driven by a NXP PCF50633 MFD, say Y here to
 	  enable its driver.
 
+config BACKLIGHT_TPS6116X
+	tristate "TPS6116X LCD Backlight"
+	depends on GENERIC_GPIO
+	help
+	  This driver controls the LCD backlight level for EasyScale capable
+	  SSP connected backlight controllers.
+
 endif # BACKLIGHT_CLASS_DEVICE
 
 endif # BACKLIGHT_LCD_SUPPORT
drivers/video/backlight/Makefile +1 -1
index 44c0f81..5d407c8 100644
@@ -35,4 +35,4 @@ obj-$(CONFIG_BACKLIGHT_ADP5520)	+= adp5520_bl.o
 obj-$(CONFIG_BACKLIGHT_ADP8860)	+= adp8860_bl.o
 obj-$(CONFIG_BACKLIGHT_88PM860X) += 88pm860x_bl.o
 obj-$(CONFIG_BACKLIGHT_PCF50633)	+= pcf50633-backlight.o
-
+obj-$(CONFIG_BACKLIGHT_TPS6116X)+= tps6116x.o
drivers/video/backlight/tps6116x.c new file +299 -0

299 lines not on this page · raw message

Signature · 1 line
1.7.1
#16 lore raw

PATCH v7 11/12 davinci: add tnetv107x evm backlight device

In reply to Cyril Chemparathy

The tnetv107x evm board has a backlight device that is connected on one of the SSP ports. This patch adds the board definitions necessary to plug the backlight driver to the GPIO corresponding to this SSP pin.

---
 arch/arm/mach-davinci/board-tnetv107x-evm.c |   14 ++++++++++++++
 1 files changed, 14 insertions(+), 0 deletions(-)
arch/arm/mach-davinci/board-tnetv107x-evm.c +14 -0
index e3863dd..ac62de2 100644
@@ -44,6 +44,7 @@
 #define EVM_MMC_WP_GPIO		21
 #define EVM_MMC_CD_GPIO		24
 #define EVM_SPI_CS_GPIO		54
+#define EVM_BACKLIGHT_GPIO	(SSP_GPIO_START + 2)
 
 static int initialize_gpio(int gpio, char *desc)
 {
@@ -353,6 +354,12 @@ static struct spi_board_info spi_info[] __initconst = {
 	},
 };
 
+static struct platform_device backlight_device = {
+	.name		= "tps6116x",
+	.id		= -1,
+	.dev.platform_data = (void *)EVM_BACKLIGHT_GPIO,
+};
+
 static __init void tnetv107x_evm_board_init(void)
 {
 	davinci_cfg_reg_list(sdio1_pins);
@@ -364,6 +371,13 @@ static __init void tnetv107x_evm_board_init(void)
 	spi_register_board_info(spi_info, ARRAY_SIZE(spi_info));
 }
 
+static int __init tnetv107x_evm_late_init(void)
+{
+	platform_device_register(&backlight_device);
+	return 0;
+}
+late_initcall(tnetv107x_evm_late_init);
+
 #ifdef CONFIG_SERIAL_8250_CONSOLE
 static int __init tnetv107x_evm_console_init(void)
 {
Signature · 1 line
1.7.1
#17 lore raw

PATCH v7 11/12 davinci: add tnetv107x evm backlight device

Hi Cyril,

Quoted 35 lines · Cyril

The tnetv107x evm board has a backlight device that is connected on one of the SSP ports. This patch adds the board definitions necessary to plug the backlight driver to the GPIO corresponding to this SSP pin.

---
 arch/arm/mach-davinci/board-tnetv107x-evm.c |   14 ++++++++++++++
 1 files changed, 14 insertions(+), 0 deletions(-)
arch/arm/mach-davinci/board-tnetv107x-evm.c +7 -0
index e3863dd..ac62de2 100644
@@ -44,6 +44,7 @@
 #define EVM_MMC_WP_GPIO              21
 #define EVM_MMC_CD_GPIO              24
 #define EVM_SPI_CS_GPIO              54
+#define EVM_BACKLIGHT_GPIO   (SSP_GPIO_START + 2)
 
 static int initialize_gpio(int gpio, char *desc)
 {
@@ -353,6 +354,12 @@ static struct spi_board_info spi_info[] __initconst = {
      },
 };
 
+static struct platform_device backlight_device = {
+     .name           = "tps6116x",
+     .id             = -1,
+     .dev.platform_data = (void *)EVM_BACKLIGHT_GPIO,
+};
+
 static __init void tnetv107x_evm_board_init(void)
 {
      davinci_cfg_reg_list(sdio1_pins);

arch/arm/mach-davinci/board-tnetv107x-evm.c static __init void tnetv107x_evm_board_init(void) · line 379

Quoted 9 lines
arch/arm/mach-davinci/board-tnetv107x-evm.c +6 -0
@@ -364,6 +371,13 @@ static __init void tnetv107x_evm_board_init(void)
      spi_register_board_info(spi_info, ARRAY_SIZE(spi_info));
 }
 
+static int __init tnetv107x_evm_late_init(void)
+{
+     platform_device_register(&backlight_device);
+     return 0;
+}
+late_initcall(tnetv107x_evm_late_init);

This call should simply return if machine is not tnetv107x EVM.

I didn't follow the entire series but wondering why platform device registration should be a late init call. Typically the driver probe can be made a late init call in case of init sequence dependencies.

Thanks,
Sekhar

#18 lore raw

PATCH v7 11/12 davinci: add tnetv107x evm backlight device

Quoted 6 lines · Sekhar

This call should simply return if machine is not tnetv107x EVM.

I didn't follow the entire series but wondering why platform device registration should be a late init call. Typically the driver probe can be made a late init call in case of init sequence dependencies.

I meant driver init, not probe.

Thanks,
Sekhar

#19 lore raw

PATCH v7 11/12 davinci: add tnetv107x evm backlight device

Quoted 10 lines · Sekhar
Quoted 6 lines · Sekhar

This call should simply return if machine is not tnetv107x EVM.

I didn't follow the entire series but wondering why platform device registration should be a late init call. Typically the driver probe can be made a late init call in case of init sequence dependencies.

I meant driver init, not probe.

This was done to handle driver dependencies, and is meant to be temporary (until a real driver dependency framework comes in).

Please see [1] for a bit more of background on changing drivers to use module_init().

Regards
- Cyril.

[1]
http://linux.davincidsp.com/pipermail/davinci-linux-open-source/2010-November/020987.html

#20 lore raw

PATCH v7 11/12 davinci: add tnetv107x evm backlight device

Quoted 10 lines · Cyril
Quoted 9 lines · Sekhar
Quoted 6 lines · Sekhar

This call should simply return if machine is not tnetv107x EVM.

I didn't follow the entire series but wondering why platform device registration should be a late init call. Typically the driver probe can be made a late init call in case of init sequence dependencies.

Quoted 7 lines
Quoted 1 line

I meant driver init, not probe.

This was done to handle driver dependencies, and is meant to be temporary (until a real driver dependency framework comes in).

Please see [1] for a bit more of background on changing drivers to use module_init().

Hmm, okay. Please include a check for machine in that case.

Thanks,
Sekhar

Quoted 7 lines

Regards
- Cyril.

[1]
http://linux.davincidsp.com/pipermail/davinci-linux-open-source/2010-November/020987.html

#21 lore raw

PATCH v7 11/12 davinci: add tnetv107x evm backlight device

Quoted 13 lines · Sekhar
Quoted 12 lines · Cyril
Quoted 10 lines · Sekhar
Quoted 6 lines · Sekhar

This call should simply return if machine is not tnetv107x EVM.

I didn't follow the entire series but wondering why platform device registration should be a late init call. Typically the driver probe can be made a late init call in case of init sequence dependencies.

I meant driver init, not probe.

Quoted 8 lines
Quoted 5 lines

This was done to handle driver dependencies, and is meant to be temporary (until a real driver dependency framework comes in).

Please see [1] for a bit more of background on changing drivers to use module_init().

Hmm, okay. Please include a check for machine in that case.

Sekhar is right.

Today, we don't (can't) build this machine with others for a couple reasons, but we're moving towards removing those restrictions, so when this machine is included in a binary with others, it needs to check.

Thanks,

Kevin

A message that is not in the archive (sent off-list or not yet received): B85A65D85D7EB246BE421B3FB0FBB5930247CD83E1-/tLxBxkBPtCIQmiDNMet8wC/[email protected]

#23 lore raw

RE: [PATCH v7 11/12] davinci: add tnetv107x evm backlight device

Quoted 6 lines · Sekhar

This call should simply return if machine is not tnetv107x EVM.

I didn't follow the entire series but wondering why platform device registration should be a late init call. Typically the driver probe can be made a late init call in case of init sequence dependencies.

I meant driver init, not probe.

Thanks,
Sekhar

------------------------------------------------------------------------------
This SF Dev2Dev email is sponsored by:

WikiLeaks The End of the Free Internet
http://p.sf.net/sfu/therealnews-com

A message that is not in the archive (sent off-list or not yet received): 1291733522-3626-12-git-send-email-cyril-l0cyMroinI0@public.gmane.org

#25 lore raw

RE: [PATCH v7 11/12] davinci: add tnetv107x evm backlight device

Hi Cyril,

Quoted 35 lines · Cyril

The tnetv107x evm board has a backlight device that is connected on one of the SSP ports. This patch adds the board definitions necessary to plug the backlight driver to the GPIO corresponding to this SSP pin.

---
 arch/arm/mach-davinci/board-tnetv107x-evm.c |   14 ++++++++++++++
 1 files changed, 14 insertions(+), 0 deletions(-)
arch/arm/mach-davinci/board-tnetv107x-evm.c +7 -0
index e3863dd..ac62de2 100644
@@ -44,6 +44,7 @@
 #define EVM_MMC_WP_GPIO              21
 #define EVM_MMC_CD_GPIO              24
 #define EVM_SPI_CS_GPIO              54
+#define EVM_BACKLIGHT_GPIO   (SSP_GPIO_START + 2)
 
 static int initialize_gpio(int gpio, char *desc)
 {
@@ -353,6 +354,12 @@ static struct spi_board_info spi_info[] __initconst = {
      },
 };
 
+static struct platform_device backlight_device = {
+     .name           = "tps6116x",
+     .id             = -1,
+     .dev.platform_data = (void *)EVM_BACKLIGHT_GPIO,
+};
+
 static __init void tnetv107x_evm_board_init(void)
 {
      davinci_cfg_reg_list(sdio1_pins);

arch/arm/mach-davinci/board-tnetv107x-evm.c static __init void tnetv107x_evm_board_init(void) · line 379

Quoted 9 lines
arch/arm/mach-davinci/board-tnetv107x-evm.c +6 -0
@@ -364,6 +371,13 @@ static __init void tnetv107x_evm_board_init(void)
      spi_register_board_info(spi_info, ARRAY_SIZE(spi_info));
 }
 
+static int __init tnetv107x_evm_late_init(void)
+{
+     platform_device_register(&backlight_device);
+     return 0;
+}
+late_initcall(tnetv107x_evm_late_init);

This call should simply return if machine is not tnetv107x EVM.

I didn't follow the entire series but wondering why platform device registration should be a late init call. Typically the driver probe can be made a late init call in case of init sequence dependencies.

Thanks,
Sekhar

------------------------------------------------------------------------------
This SF Dev2Dev email is sponsored by:

WikiLeaks The End of the Free Internet
http://p.sf.net/sfu/therealnews-com

#26 lore raw

PATCH v7 12/12 davinci: add tnetv107x evm i2c eeprom device

In reply to Cyril Chemparathy

The tnetv107x evm board has an I2C device connected on one of the SSP ports. This patch adds board definitions for a GPIO based I2C master, as well as definitions for the eeprom device on these boards.

---
 arch/arm/mach-davinci/board-tnetv107x-evm.c |   30 +++++++++++++++++++++++++++
 1 files changed, 30 insertions(+), 0 deletions(-)
arch/arm/mach-davinci/board-tnetv107x-evm.c +30 -0
index ac62de2..869af15 100644
@@ -29,6 +29,9 @@
 #include <linux/regulator/machine.h>
 #include <linux/regulator/consumer.h>
 #include <linux/regulator/driver.h>
+#include <linux/i2c.h>
+#include <linux/i2c/at24.h>
+#include <linux/i2c-gpio.h>
 
 #include <asm/mach/arch.h>
 #include <asm/mach-types.h>
@@ -44,6 +47,8 @@
 #define EVM_MMC_WP_GPIO		21
 #define EVM_MMC_CD_GPIO		24
 #define EVM_SPI_CS_GPIO		54
+#define EVM_I2C_SDA_GPIO	(SSP_GPIO_START + 0)
+#define EVM_I2C_SCL_GPIO	(SSP_GPIO_START + 1)
 #define EVM_BACKLIGHT_GPIO	(SSP_GPIO_START + 2)
 
 static int initialize_gpio(int gpio, char *desc)
@@ -360,6 +365,29 @@ static struct platform_device backlight_device = {
 	.dev.platform_data = (void *)EVM_BACKLIGHT_GPIO,
 };
 
+struct i2c_gpio_platform_data i2c_data = {
+	.sda_pin	= EVM_I2C_SDA_GPIO,
+	.scl_pin	= EVM_I2C_SCL_GPIO,
+};
+
+static struct platform_device i2c_device = {
+	.name		= "i2c-gpio",
+	.id		= 0,
+	.dev.platform_data = &i2c_data,
+};
+
+static struct at24_platform_data at24_config = {
+	.byte_len	= SZ_16K / 8,
+	.page_size	= 16,
+};
+
+static struct i2c_board_info i2c_info[] __initconst =  {
+	{
+		I2C_BOARD_INFO("24c16", 0x50),
+		.platform_data	= &at24_config,
+	},
+};
+
 static __init void tnetv107x_evm_board_init(void)
 {
 	davinci_cfg_reg_list(sdio1_pins);
@@ -369,11 +397,13 @@ static __init void tnetv107x_evm_board_init(void)
 	tnetv107x_devices_init(&evm_device_info);
 
 	spi_register_board_info(spi_info, ARRAY_SIZE(spi_info));
+	i2c_register_board_info(0, i2c_info, ARRAY_SIZE(i2c_info));
 }
 
 static int __init tnetv107x_evm_late_init(void)
 {
 	platform_device_register(&backlight_device);
+	platform_device_register(&i2c_device);
 	return 0;
 }
 late_initcall(tnetv107x_evm_late_init);
Signature · 1 line
1.7.1
#27 lore raw

v7 cover tnetv107x ssp drivers

In reply to Cyril Chemparathy
Quoted 3 lines · Cyril Chemparathy

TI's sequencer serial port (TI-SSP) is a jack-of-all-trades type of serial port device. It has a built-in programmable execution engine that can be programmed to operate as almost any serial bus (I2C, SPI, EasyScale, and others).

OK, this series seems to be stabilizing. The question now is how to merge it.

I'm happy to merge it as a single series via the davinci tree as long as I get some acks on the drivers. For me to merge this, I'm missing some acks for the following:

[PATCH v7 04/12] spi: add ti-ssp spi master driver Grant Likely, although I thought he gave an ack earlier)

[PATCH v7 10/12] backlight: add support for tps6116x controller Richard Purdie

[PATCH v7 08/12] gpio: add ti-ssp gpio driver
David Brownell? (MAINTAINERS isn't clear here)

If I get acks soon, we can get this in for 2.6.38.

Thanks,

Kevin

#28 lore raw

v7 cover tnetv107x ssp drivers

Quoted 2 lines · Kevin Hilman

[PATCH v7 08/12] gpio: add ti-ssp gpio driver
David Brownell? (MAINTAINERS isn't clear here)

It's mostly abandoned at the minute. Andrew Morton has been taking patches. Grant Likely said he might take it over so sending stuff his way might prod him into action.

#29 lore raw

v7 cover tnetv107x ssp drivers

Quoted 8 lines · Mark Brown
Quoted 2 lines · Kevin Hilman

[PATCH v7 08/12] gpio: add ti-ssp gpio driver
David Brownell? (MAINTAINERS isn't clear here)

It's mostly abandoned at the minute. Andrew Morton has been taking patches. Grant Likely said he might take it over so sending stuff his way might prod him into action.

I probably will pick up GPIO, but not just yet. I've got to get my dt and spi house in order first (making good progress here. I've got an autotest instance set up so that I can apply and test patches on many architectures quickly and easily). Once I've taken care of the backlog I'll make a decision on whether or not I can pick up gpio too. Probably sometime mid-January.

g.