tnetv107x ssp drivers
Subsystems (9)
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
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
@@ -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
@@ -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
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.
Signed-off-by Cyril Chemparathy <[email protected]>
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.hdrivers/mfd/Kconfig +11
-0@@ -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 && ARMdrivers/mfd/Makefile +1
-0@@ -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.odrivers/mfd/ti-ssp.c new file +476
-0@@ -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@@ -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__ */
PATCH v7 2/12 davinci: add tnetv107x ssp platform device
In reply to Cyril ChemparathyThis 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
@@ -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
@@ -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
@@ -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
PATCH v7 3/12 davinci: add ssp config for tnetv107x evm board
In reply to Cyril ChemparathyThis 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
@@ -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
PATCH v7 4/12 spi: add ti-ssp spi master driver
In reply to Cyril ChemparathyThis 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
@@ -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
@@ -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
@@ -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
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.
Signed-off-by Cyril Chemparathy <[email protected]>
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.cdrivers/spi/Kconfig +10
-0@@ -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 PCIdrivers/spi/Makefile +1
-0@@ -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.odrivers/spi/ti-ssp-spi.c new file +22
-0@@ -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@@ -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
PATCH v7 5/12 davinci: add spi devices on tnetv107x evm
In reply to Cyril ChemparathyThis 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
@@ -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 54static 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
PATCH v7 6/12 regulator: add driver for tps6524x regulator
In reply to Cyril ChemparathyTPS6524X 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
@@ -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
@@ -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
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?
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.
PATCH v7 7/12 davinci: add tnetv107x evm regulators
In reply to Cyril ChemparathyThis 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
@@ -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
PATCH v7 8/12 gpio: add ti-ssp gpio driver
In reply to Cyril ChemparathyTI'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
@@ -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
@@ -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-$(CONFIG_GPIO_XILINX) += xilinx_gpio.o
drivers/gpio/ti-ssp-gpio.c new file +206 -0
206 lines not on this page · raw message
include/linux/mfd/ti_ssp.h +4 -0
@@ -38,6 +38,10 @@ struct ti_ssp_spi_data {void (*select)(int cs); }; +struct ti_ssp_gpio_data { +int start; +}; +/* * Sequencer port IO pin configuration bits. These do not correlate 1-1 with * the hardware. The iosel field in the port data combines iosel1 and iosel2,
Signature · 1 line
1.7.1
PATCH v7 8/12 gpio: add ti-ssp gpio driver
Quoted 99 lines · Cyril Chemparathy
TI's SSP controller pins can be directly read and written to behave like a GPIO. This patch adds a GPIO driver that exposes such functionality.
Signed-off-by Cyril Chemparathy <[email protected]>--- drivers/gpio/Kconfig | 10 ++ drivers/gpio/Makefile | 1 + drivers/gpio/ti-ssp-gpio.c | 206 ++++++++++++++++++++++++++++++++++++++++++++ include/linux/mfd/ti_ssp.h | 4 + 4 files changed, 221 insertions(+), 0 deletions(-) create mode 100644 drivers/gpio/ti-ssp-gpio.cdrivers/gpio/Kconfig +10
-0@@ -128,6 +128,16 @@ config GPIO_VX855 additional drivers must be enabled in order to use the functionality of the device. +config GPIO_TI_SSP + tristate "TI Sequencer Serial Port - GPIO Support" + depends on MFD_TI_SSP + help + Say yes here to support a GPIO interface on TI SSP port pins. + Each SSP port translates into 4 GPIOs. + + This driver can also be built as a module. If so, the module + will be called ti-ssp-gpio. + comment "I2C GPIO expanders:" config GPIO_MAX7300drivers/gpio/Makefile +1
-0@@ -26,6 +26,7 @@ obj-$(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.odrivers/gpio/ti-ssp-gpio.c new file +48
-0@@ -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.
PATCH v7 9/12 davinci: add tnetv107x evm ti-ssp gpio device
In reply to Cyril ChemparathyThis 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
@@ -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
PATCH v7 10/12 backlight: add support for tps6116x controller
In reply to Cyril ChemparathyTPS6116x 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
@@ -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
@@ -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
PATCH v7 11/12 davinci: add tnetv107x evm backlight device
In reply to Cyril ChemparathyThe 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
@@ -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
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.
Signed-off-by Cyril Chemparathy <[email protected]>--- 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@@ -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
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
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
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
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]
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
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.
Signed-off-by Cyril Chemparathy <[email protected]>--- 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@@ -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
PATCH v7 12/12 davinci: add tnetv107x evm i2c eeprom device
In reply to Cyril ChemparathyThe 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
@@ -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
v7 cover tnetv107x ssp drivers
In reply to Cyril ChemparathyQuoted 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
v7 cover tnetv107x ssp drivers
Quoted 2 lines · Kevin Hilman
[PATCH v7 08/12] gpio: add ti-ssp gpio driver
David Brownell? (MAINTAINERSisn'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.
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? (MAINTAINERSisn'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.