[PATCH RESEND v4] spi/pl022: add devicetree support

[PATCH RESEND v4] spi/pl022: add devicetree support

[Alexandre Pereira da Silva]

Add the chipselect array and cur_cs properties to pl022 main structure

Add a wrapper function to decide if the cs should be controlled by the
cs_control callback or the chipselect gpio

Populate chipselect property from cs-gpios

Populate master->dev.of_node, so the spi bus can find child spi
devices and register them

At pl022 setup, fill chip_data structure from dt nodes, if not provided
by platform.

Signed-off-by: Alexandre Pereira da Silva <aletes.xgr@gmail.com>
Acked-by: Roland Stigge <stigge@antcom.de>
---
Applies to v3.6-rc1

Changes since v3:
* Proper use of IS_ENABLED

Changes since v2:
* Use IS_ENABLED instead of #ifdef
* Remove bogus const change

Changes since v1:
* return EPROBE_DEFFER if gpios are not initialized yet

Thanks Thierry Reding for reviewing and improvements suggestions

 .../devicetree/bindings/spi/spi_pl022.txt          |   15 +++
 drivers/spi/spi-pl022.c                            |  114 ++++++++++++++++----
 2 files changed, 110 insertions(+), 19 deletions(-)

diff --git a/Documentation/devicetree/bindings/spi/spi_pl022.txt b/Documentation/devicetree/bindings/spi/spi_pl022.txt
index 306ec3f..b089ec7 100644
--- a/Documentation/devicetree/bindings/spi/spi_pl022.txt
+++ b/Documentation/devicetree/bindings/spi/spi_pl022.txt
@@ -6,7 +6,22 @@ Required properties:
 - interrupts : Should contain SPI controller interrupt
 
 Optional properties:
+- pl022,num-chipselects : total number of chipselects
 - cs-gpios : should specify GPIOs used for chipselects.
   The gpios will be referred to as reg = <index> in the SPI child nodes.
   If unspecified, a single SPI device without a chip select can be used.
 
+SPI slave nodes must be children of the SPI master node and can
+contain the following properties.
+See include/linux/amba/pl022.h for more details
+
+- pl022,hierarchy : master or slave interface
+- pl022,interface : interface type
+- pl022,slave-tx-disable : disconnect tx line in slave mode
+- pl022,com-mode : polling, interrupt or dma
+- pl022,rx-level-trig : Rx FIFO watermark level
+- pl022,tx-level-trig : Tx FIFO watermark level
+- pl022,ctrl-len : Microwire interface: Control length
+- pl022,wait-state : Microwire interface: Wait state
+- pl022,duplex : Microwire interface: Full/Half duplex
+
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index 400ae21..d8d7d21 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -40,6 +40,8 @@
 #include <linux/dma-mapping.h>
 #include <linux/scatterlist.h>
 #include <linux/pm_runtime.h>
+#include <linux/gpio.h>
+#include <linux/of_gpio.h>
 
 /*
  * This macro is used to define some register default values.
@@ -389,6 +391,8 @@ struct pl022 {
 	char				*dummypage;
 	bool				dma_running;
 #endif
+	int cur_cs;
+	int chipselect[0];
 };
 
 /**
@@ -433,6 +437,14 @@ static void null_cs_control(u32 command)
 	pr_debug("pl022: dummy chip select control, CS=0x%x\n", command);
 }
 
+static void pl022_cs_control(struct pl022 *pl022, u32 command)
+{
+	if (gpio_is_valid(pl022->cur_cs))
+		gpio_set_value(pl022->cur_cs, command);
+	else
+		pl022->cur_chip->cs_control(command);
+}
+
 /**
  * giveback - current spi_message is over, schedule next message and call
  * callback of this message. Assumes that caller already
@@ -479,7 +491,7 @@ static void giveback(struct pl022 *pl022)
 		if (next_msg && next_msg->spi != pl022->cur_msg->spi)
 			next_msg = NULL;
 		if (!next_msg || pl022->cur_msg->state == STATE_ERROR)
-			pl022->cur_chip->cs_control(SSP_CHIP_DESELECT);
+			pl022_cs_control(pl022, SSP_CHIP_DESELECT);
 		else
 			pl022->next_msg_cs_active = true;
 
@@ -813,8 +825,7 @@ static void dma_callback(void *data)
 	/* Update total bytes transferred */
 	msg->actual_length += pl022->cur_transfer->len;
 	if (pl022->cur_transfer->cs_change)
-		pl022->cur_chip->
-			cs_control(SSP_CHIP_DESELECT);
+		pl022_cs_control(pl022, SSP_CHIP_DESELECT);
 
 	/* Move to next transfer */
 	msg->state = next_transfer(pl022);
@@ -1247,8 +1258,7 @@ static irqreturn_t pl022_interrupt_handler(int irq, void *dev_id)
 		/* Update total bytes transferred */
 		msg->actual_length += pl022->cur_transfer->len;
 		if (pl022->cur_transfer->cs_change)
-			pl022->cur_chip->
-				cs_control(SSP_CHIP_DESELECT);
+			pl022_cs_control(pl022, SSP_CHIP_DESELECT);
 		/* Move to next transfer */
 		msg->state = next_transfer(pl022);
 		tasklet_schedule(&pl022->pump_transfers);
@@ -1333,7 +1343,7 @@ static void pump_transfers(unsigned long data)
 
 		/* Reselect chip select only if cs_change was requested */
 		if (previous->cs_change)
-			pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
+			pl022_cs_control(pl022, SSP_CHIP_SELECT);
 	} else {
 		/* STATE_START */
 		message->state = STATE_RUNNING;
@@ -1372,7 +1382,7 @@ static void do_interrupt_dma_transfer(struct pl022 *pl022)
 
 	/* Enable target chip, if not already active */
 	if (!pl022->next_msg_cs_active)
-		pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
+		pl022_cs_control(pl022, SSP_CHIP_SELECT);
 
 	if (set_up_next_transfer(pl022, pl022->cur_transfer)) {
 		/* Error path */
@@ -1424,12 +1434,12 @@ static void do_polling_transfer(struct pl022 *pl022)
 			if (previous->delay_usecs)
 				udelay(previous->delay_usecs);
 			if (previous->cs_change)
-				pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
+				pl022_cs_control(pl022, SSP_CHIP_SELECT);
 		} else {
 			/* STATE_START */
 			message->state = STATE_RUNNING;
 			if (!pl022->next_msg_cs_active)
-				pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
+				pl022_cs_control(pl022, SSP_CHIP_SELECT);
 		}
 
 		/* Configuration Changing Per Transfer */
@@ -1461,7 +1471,7 @@ static void do_polling_transfer(struct pl022 *pl022)
 		/* Update total byte transferred */
 		message->actual_length += pl022->cur_transfer->len;
 		if (pl022->cur_transfer->cs_change)
-			pl022->cur_chip->cs_control(SSP_CHIP_DESELECT);
+			pl022_cs_control(pl022, SSP_CHIP_DESELECT);
 		/* Move to next transfer */
 		message->state = next_transfer(pl022);
 	}
@@ -1490,6 +1500,7 @@ static int pl022_transfer_one_message(struct spi_master *master,
 
 	/* Setup the SPI using the per chip configuration */
 	pl022->cur_chip = spi_get_ctldata(msg->spi);
+	pl022->cur_cs = pl022->chipselect[msg->spi->chip_select];
 
 	restore_state(pl022);
 	flush(pl022);
@@ -1761,12 +1772,14 @@ static const struct pl022_config_chip pl022_default_chip_info = {
 static int pl022_setup(struct spi_device *spi)
 {
 	struct pl022_config_chip const *chip_info;
+	struct pl022_config_chip chip_info_dt;
 	struct chip_data *chip;
 	struct ssp_clock_params clk_freq = { .cpsdvsr = 0, .scr = 0};
 	int status = 0;
 	struct pl022 *pl022 = spi_master_get_devdata(spi->master);
 	unsigned int bits = spi->bits_per_word;
 	u32 tmp;
+	struct device_node *np = spi->dev.of_node;
 
 	if (!spi->max_speed_hz)
 		return -EINVAL;
@@ -1789,10 +1802,36 @@ static int pl022_setup(struct spi_device *spi)
 	chip_info = spi->controller_data;
 
 	if (chip_info == NULL) {
-		chip_info = &pl022_default_chip_info;
-		/* spi_board_info.controller_data not is supplied */
-		dev_dbg(&spi->dev,
-			"using default controller_data settings\n");
+		if (np) {
+			chip_info_dt = pl022_default_chip_info;
+
+			of_property_read_u32(np, "pl022,hierarchy",
+				&chip_info_dt.hierarchy);
+			of_property_read_u32(np, "pl022,interface",
+				&chip_info_dt.iface);
+			chip_info_dt.slave_tx_disable =
+				of_property_read_bool(np,
+					"pl022,slave-tx-disable");
+			of_property_read_u32(np, "pl022,com-mode",
+				&chip_info_dt.com_mode);
+			of_property_read_u32(np, "pl022,rx-level-trig",
+				&chip_info_dt.rx_lev_trig);
+			of_property_read_u32(np, "pl022,tx-level-trig",
+				&chip_info_dt.tx_lev_trig);
+			of_property_read_u32(np, "pl022,ctrl-len",
+				&chip_info_dt.ctrl_len);
+			of_property_read_u32(np, "pl022,wait-state",
+				&chip_info_dt.wait_state);
+			of_property_read_u32(np, "pl022,duplex",
+				&chip_info_dt.duplex);
+
+			chip_info = &chip_info_dt;
+		} else {
+			chip_info = &pl022_default_chip_info;
+			/* spi_board_info.controller_data not is supplied */
+			dev_dbg(&spi->dev,
+				"using default controller_data settings\n");
+		}
 	} else
 		dev_dbg(&spi->dev,
 			"using user supplied controller_data settings\n");
@@ -1835,8 +1874,9 @@ static int pl022_setup(struct spi_device *spi)
 	chip->xfer_type = chip_info->com_mode;
 	if (!chip_info->cs_control) {
 		chip->cs_control = null_cs_control;
-		dev_warn(&spi->dev,
-			 "chip select function is NULL for this chip\n");
+		if (!gpio_is_valid(pl022->chipselect[spi->chip_select]))
+			dev_warn(&spi->dev,
+				 "chip select function is NULL for this chip\n");
 	} else
 		chip->cs_control = chip_info->cs_control;
 
@@ -1988,7 +2028,8 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
 	struct pl022_ssp_controller *platform_info = adev->dev.platform_data;
 	struct spi_master *master;
 	struct pl022 *pl022 = NULL;	/*Data for this driver */
-	int status = 0;
+	struct device_node *np = adev->dev.of_node;
+	int status = 0, i, num_cs;
 
 	dev_info(&adev->dev,
 		 "ARM PL022 driver, device ID: 0x%08x\n", adev->periphid);
@@ -1998,8 +2039,14 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
 		goto err_no_pdata;
 	}
 
+	num_cs = platform_info->num_chipselect;
+	if (IS_ENABLED(CONFIG_OF))
+		of_property_read_u32(np, "pl022,num-chipselects", &num_cs);
+
+
 	/* Allocate master with space for data */
-	master = spi_alloc_master(dev, sizeof(struct pl022));
+	master = spi_alloc_master(dev,
+		sizeof(struct pl022) + sizeof(int) * num_cs);
 	if (master == NULL) {
 		dev_err(&adev->dev, "probe - cannot alloc SPI master\n");
 		status = -ENOMEM;
@@ -2017,13 +2064,41 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
 	 * on this board
 	 */
 	master->bus_num = platform_info->bus_id;
-	master->num_chipselect = platform_info->num_chipselect;
+	master->num_chipselect = num_cs;
 	master->cleanup = pl022_cleanup;
 	master->setup = pl022_setup;
 	master->prepare_transfer_hardware = pl022_prepare_transfer_hardware;
 	master->transfer_one_message = pl022_transfer_one_message;
 	master->unprepare_transfer_hardware = pl022_unprepare_transfer_hardware;
 	master->rt = platform_info->rt;
+	master->dev.of_node = dev->of_node;
+
+	if (IS_ENABLED(CONFIG_OF)) {
+		for (i = 0; i < num_cs; i++) {
+			int cs_gpio = of_get_named_gpio(np, "cs-gpios", i);
+
+			if (cs_gpio == -EPROBE_DEFER) {
+				status = -EPROBE_DEFER;
+				goto err_no_gpio;
+			}
+
+			pl022->chipselect[i] = cs_gpio;
+
+			if (gpio_is_valid(cs_gpio)) {
+				if (gpio_request(cs_gpio, "ssp-pl022"))
+					dev_err(&adev->dev,
+						"could not request %d gpio\n",
+						cs_gpio);
+				else if (gpio_direction_output(cs_gpio, 1))
+					dev_err(&adev->dev,
+						"could set gpio %d as output\n",
+						cs_gpio);
+			}
+		}
+	} else {
+		for (i = 0; i < num_cs; i++)
+			pl022->chipselect[i] = -EINVAL;
+	}
 
 	/*
 	 * Supports mode 0-3, loopback, and active low CS. Transfers are
@@ -2130,6 +2205,7 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
  err_no_ioremap:
 	amba_release_regions(adev);
  err_no_ioregion:
+ err_no_gpio:
 	spi_master_put(master);
  err_no_master:
  err_no_pdata:
-- 
1.7.10

Re: [PATCH RESEND v4] spi/pl022: add devicetree support

[Roland Stigge]

Hi all,

On 08/20/2012 01:39 PM, Alexandre Pereira da Silva wrote:
> Add the chipselect array and cur_cs properties to pl022 main structure
>
> Add a wrapper function to decide if the cs should be controlled by the
> cs_control callback or the chipselect gpio
>
> Populate chipselect property from cs-gpios
>
> Populate master->dev.of_node, so the spi bus can find child spi
> devices and register them
>
> At pl022 setup, fill chip_data structure from dt nodes, if not provided
> by platform.
>
> Signed-off-by: Alexandre Pereira da Silva <aletes.xgr@gmail.com>
> Acked-by: Roland Stigge <stigge@antcom.de>

We somehow missed 3.6 with this one. Would be nice to have it in 3.7 now
since it's useful for mach-lpc32xx when it's available.

If there are any open issues, I can assist Alexandre on this patch.

Thanks in advance,

Roland

Re: [PATCH RESEND v4] spi/pl022: add devicetree support

[Rob Herring]

On 08/20/2012 07:58 AM, Roland Stigge wrote:
> Hi all,
> 
> On 08/20/2012 01:39 PM, Alexandre Pereira da Silva wrote:
>> Add the chipselect array and cur_cs properties to pl022 main structure
>>
>> Add a wrapper function to decide if the cs should be controlled by the
>> cs_control callback or the chipselect gpio
>>
>> Populate chipselect property from cs-gpios
>>
>> Populate master->dev.of_node, so the spi bus can find child spi
>> devices and register them
>>
>> At pl022 setup, fill chip_data structure from dt nodes, if not provided
>> by platform.
>>
>> Signed-off-by: Alexandre Pereira da Silva <aletes.xgr@gmail.com>
>> Acked-by: Roland Stigge <stigge@antcom.de>
> 
> We somehow missed 3.6 with this one. Would be nice to have it in 3.7 now
> since it's useful for mach-lpc32xx when it's available.

Mark Brown is handling SPI right now for Grant, so please send it to
him. I have some comments on the patch though.

Rob

> 
> If there are any open issues, I can assist Alexandre on this patch.
> 
> Thanks in advance,
> 
> Roland
> 

Re: [PATCH RESEND v4] spi/pl022: add devicetree support

[Rob Herring]

On 08/20/2012 06:39 AM, Alexandre Pereira da Silva wrote:
> Add the chipselect array and cur_cs properties to pl022 main structure
> 
> Add a wrapper function to decide if the cs should be controlled by the
> cs_control callback or the chipselect gpio
> 
> Populate chipselect property from cs-gpios
> 
> Populate master->dev.of_node, so the spi bus can find child spi
> devices and register them
> 
> At pl022 setup, fill chip_data structure from dt nodes, if not provided
> by platform.
> 
> Signed-off-by: Alexandre Pereira da Silva <aletes.xgr@gmail.com>
> Acked-by: Roland Stigge <stigge@antcom.de>
> ---
> Applies to v3.6-rc1
> 
> Changes since v3:
> * Proper use of IS_ENABLED
> 
> Changes since v2:
> * Use IS_ENABLED instead of #ifdef
> * Remove bogus const change
> 
> Changes since v1:
> * return EPROBE_DEFFER if gpios are not initialized yet
> 
> Thanks Thierry Reding for reviewing and improvements suggestions
> 
>  .../devicetree/bindings/spi/spi_pl022.txt          |   15 +++
>  drivers/spi/spi-pl022.c                            |  114 ++++++++++++++++----
>  2 files changed, 110 insertions(+), 19 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/spi/spi_pl022.txt b/Documentation/devicetree/bindings/spi/spi_pl022.txt
> index 306ec3f..b089ec7 100644
> --- a/Documentation/devicetree/bindings/spi/spi_pl022.txt
> +++ b/Documentation/devicetree/bindings/spi/spi_pl022.txt
> @@ -6,7 +6,22 @@ Required properties:
>  - interrupts : Should contain SPI controller interrupt
>  
>  Optional properties:
> +- pl022,num-chipselects : total number of chipselects

We have this for other spi controllers too. Define a generic property.

>  - cs-gpios : should specify GPIOs used for chipselects.
>    The gpios will be referred to as reg = <index> in the SPI child nodes.
>    If unspecified, a single SPI device without a chip select can be used.
>  
> +SPI slave nodes must be children of the SPI master node and can
> +contain the following properties.
> +See include/linux/amba/pl022.h for more details

DT binding is independent of Linux, so you should not refer to Linux docs.

> +
> +- pl022,hierarchy : master or slave interface

Is attaching a master and using the pl022 as a slave really a supported
use case? The DT spi bindings are really designed for a master node with
N slave nodes. So there is more general question of how to describe a
spi controller as a slave. It's not really one I care to answer as the
Linux spi framework isn't designed to act as a slave.

> +- pl022,interface : interface type

And valid values are?

> +- pl022,slave-tx-disable : disconnect tx line in slave mode
> +- pl022,com-mode : polling, interrupt or dma
> +- pl022,rx-level-trig : Rx FIFO watermark level
> +- pl022,tx-level-trig : Tx FIFO watermark level
> +- pl022,ctrl-len : Microwire interface: Control length
> +- pl022,wait-state : Microwire interface: Wait state
> +- pl022,duplex : Microwire interface: Full/Half duplex

Most of these properties aren't used anywhere in the kernel other than
u300 and I'm not sure what to purpose of dummy_chip_info is. Perhaps
Linus has some input? I think either they can be decided by the spi
controller (com-mode, fifo watermark) or should be standard properties
(microwire settings). I would argue they should be removed if the spi
framework doesn't have support in a standard way.

> +
> diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
> index 400ae21..d8d7d21 100644
> --- a/drivers/spi/spi-pl022.c
> +++ b/drivers/spi/spi-pl022.c
> @@ -40,6 +40,8 @@
>  #include <linux/dma-mapping.h>
>  #include <linux/scatterlist.h>
>  #include <linux/pm_runtime.h>
> +#include <linux/gpio.h>
> +#include <linux/of_gpio.h>
>  
>  /*
>   * This macro is used to define some register default values.
> @@ -389,6 +391,8 @@ struct pl022 {
>  	char				*dummypage;
>  	bool				dma_running;
>  #endif
> +	int cur_cs;
> +	int chipselect[0];
>  };
>  
>  /**
> @@ -433,6 +437,14 @@ static void null_cs_control(u32 command)
>  	pr_debug("pl022: dummy chip select control, CS=0x%x\n", command);
>  }
>  
> +static void pl022_cs_control(struct pl022 *pl022, u32 command)
> +{
> +	if (gpio_is_valid(pl022->cur_cs))
> +		gpio_set_value(pl022->cur_cs, command);
> +	else
> +		pl022->cur_chip->cs_control(command);
> +}
> +
>  /**
>   * giveback - current spi_message is over, schedule next message and call
>   * callback of this message. Assumes that caller already
> @@ -479,7 +491,7 @@ static void giveback(struct pl022 *pl022)
>  		if (next_msg && next_msg->spi != pl022->cur_msg->spi)
>  			next_msg = NULL;
>  		if (!next_msg || pl022->cur_msg->state == STATE_ERROR)
> -			pl022->cur_chip->cs_control(SSP_CHIP_DESELECT);
> +			pl022_cs_control(pl022, SSP_CHIP_DESELECT);
>  		else
>  			pl022->next_msg_cs_active = true;
>  
> @@ -813,8 +825,7 @@ static void dma_callback(void *data)
>  	/* Update total bytes transferred */
>  	msg->actual_length += pl022->cur_transfer->len;
>  	if (pl022->cur_transfer->cs_change)
> -		pl022->cur_chip->
> -			cs_control(SSP_CHIP_DESELECT);
> +		pl022_cs_control(pl022, SSP_CHIP_DESELECT);
>  
>  	/* Move to next transfer */
>  	msg->state = next_transfer(pl022);
> @@ -1247,8 +1258,7 @@ static irqreturn_t pl022_interrupt_handler(int irq, void *dev_id)
>  		/* Update total bytes transferred */
>  		msg->actual_length += pl022->cur_transfer->len;
>  		if (pl022->cur_transfer->cs_change)
> -			pl022->cur_chip->
> -				cs_control(SSP_CHIP_DESELECT);
> +			pl022_cs_control(pl022, SSP_CHIP_DESELECT);
>  		/* Move to next transfer */
>  		msg->state = next_transfer(pl022);
>  		tasklet_schedule(&pl022->pump_transfers);
> @@ -1333,7 +1343,7 @@ static void pump_transfers(unsigned long data)
>  
>  		/* Reselect chip select only if cs_change was requested */
>  		if (previous->cs_change)
> -			pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> +			pl022_cs_control(pl022, SSP_CHIP_SELECT);
>  	} else {
>  		/* STATE_START */
>  		message->state = STATE_RUNNING;
> @@ -1372,7 +1382,7 @@ static void do_interrupt_dma_transfer(struct pl022 *pl022)
>  
>  	/* Enable target chip, if not already active */
>  	if (!pl022->next_msg_cs_active)
> -		pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> +		pl022_cs_control(pl022, SSP_CHIP_SELECT);
>  
>  	if (set_up_next_transfer(pl022, pl022->cur_transfer)) {
>  		/* Error path */
> @@ -1424,12 +1434,12 @@ static void do_polling_transfer(struct pl022 *pl022)
>  			if (previous->delay_usecs)
>  				udelay(previous->delay_usecs);
>  			if (previous->cs_change)
> -				pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> +				pl022_cs_control(pl022, SSP_CHIP_SELECT);
>  		} else {
>  			/* STATE_START */
>  			message->state = STATE_RUNNING;
>  			if (!pl022->next_msg_cs_active)
> -				pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> +				pl022_cs_control(pl022, SSP_CHIP_SELECT);
>  		}
>  
>  		/* Configuration Changing Per Transfer */
> @@ -1461,7 +1471,7 @@ static void do_polling_transfer(struct pl022 *pl022)
>  		/* Update total byte transferred */
>  		message->actual_length += pl022->cur_transfer->len;
>  		if (pl022->cur_transfer->cs_change)
> -			pl022->cur_chip->cs_control(SSP_CHIP_DESELECT);
> +			pl022_cs_control(pl022, SSP_CHIP_DESELECT);
>  		/* Move to next transfer */
>  		message->state = next_transfer(pl022);
>  	}
> @@ -1490,6 +1500,7 @@ static int pl022_transfer_one_message(struct spi_master *master,
>  
>  	/* Setup the SPI using the per chip configuration */
>  	pl022->cur_chip = spi_get_ctldata(msg->spi);
> +	pl022->cur_cs = pl022->chipselect[msg->spi->chip_select];
>  
>  	restore_state(pl022);
>  	flush(pl022);
> @@ -1761,12 +1772,14 @@ static const struct pl022_config_chip pl022_default_chip_info = {
>  static int pl022_setup(struct spi_device *spi)
>  {
>  	struct pl022_config_chip const *chip_info;
> +	struct pl022_config_chip chip_info_dt;
>  	struct chip_data *chip;
>  	struct ssp_clock_params clk_freq = { .cpsdvsr = 0, .scr = 0};
>  	int status = 0;
>  	struct pl022 *pl022 = spi_master_get_devdata(spi->master);
>  	unsigned int bits = spi->bits_per_word;
>  	u32 tmp;
> +	struct device_node *np = spi->dev.of_node;
>  
>  	if (!spi->max_speed_hz)
>  		return -EINVAL;
> @@ -1789,10 +1802,36 @@ static int pl022_setup(struct spi_device *spi)
>  	chip_info = spi->controller_data;
>  
>  	if (chip_info == NULL) {
> -		chip_info = &pl022_default_chip_info;
> -		/* spi_board_info.controller_data not is supplied */
> -		dev_dbg(&spi->dev,
> -			"using default controller_data settings\n");
> +		if (np) {
> +			chip_info_dt = pl022_default_chip_info;
> +
> +			of_property_read_u32(np, "pl022,hierarchy",
> +				&chip_info_dt.hierarchy);
> +			of_property_read_u32(np, "pl022,interface",
> +				&chip_info_dt.iface);
> +			chip_info_dt.slave_tx_disable =
> +				of_property_read_bool(np,
> +					"pl022,slave-tx-disable");
> +			of_property_read_u32(np, "pl022,com-mode",
> +				&chip_info_dt.com_mode);
> +			of_property_read_u32(np, "pl022,rx-level-trig",
> +				&chip_info_dt.rx_lev_trig);
> +			of_property_read_u32(np, "pl022,tx-level-trig",
> +				&chip_info_dt.tx_lev_trig);
> +			of_property_read_u32(np, "pl022,ctrl-len",
> +				&chip_info_dt.ctrl_len);
> +			of_property_read_u32(np, "pl022,wait-state",
> +				&chip_info_dt.wait_state);
> +			of_property_read_u32(np, "pl022,duplex",
> +				&chip_info_dt.duplex);
> +
> +			chip_info = &chip_info_dt;
> +		} else {
> +			chip_info = &pl022_default_chip_info;
> +			/* spi_board_info.controller_data not is supplied */
> +			dev_dbg(&spi->dev,
> +				"using default controller_data settings\n");
> +		}
>  	} else
>  		dev_dbg(&spi->dev,
>  			"using user supplied controller_data settings\n");
> @@ -1835,8 +1874,9 @@ static int pl022_setup(struct spi_device *spi)
>  	chip->xfer_type = chip_info->com_mode;
>  	if (!chip_info->cs_control) {
>  		chip->cs_control = null_cs_control;
> -		dev_warn(&spi->dev,
> -			 "chip select function is NULL for this chip\n");
> +		if (!gpio_is_valid(pl022->chipselect[spi->chip_select]))
> +			dev_warn(&spi->dev,
> +				 "chip select function is NULL for this chip\n");
>  	} else
>  		chip->cs_control = chip_info->cs_control;
>  
> @@ -1988,7 +2028,8 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
>  	struct pl022_ssp_controller *platform_info = adev->dev.platform_data;
>  	struct spi_master *master;
>  	struct pl022 *pl022 = NULL;	/*Data for this driver */
> -	int status = 0;
> +	struct device_node *np = adev->dev.of_node;
> +	int status = 0, i, num_cs;
>  
>  	dev_info(&adev->dev,
>  		 "ARM PL022 driver, device ID: 0x%08x\n", adev->periphid);
> @@ -1998,8 +2039,14 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
>  		goto err_no_pdata;
>  	}
>  
> +	num_cs = platform_info->num_chipselect;
> +	if (IS_ENABLED(CONFIG_OF))
> +		of_property_read_u32(np, "pl022,num-chipselects", &num_cs);
> +
> +
>  	/* Allocate master with space for data */
> -	master = spi_alloc_master(dev, sizeof(struct pl022));
> +	master = spi_alloc_master(dev,
> +		sizeof(struct pl022) + sizeof(int) * num_cs);
>  	if (master == NULL) {
>  		dev_err(&adev->dev, "probe - cannot alloc SPI master\n");
>  		status = -ENOMEM;
> @@ -2017,13 +2064,41 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
>  	 * on this board
>  	 */
>  	master->bus_num = platform_info->bus_id;
> -	master->num_chipselect = platform_info->num_chipselect;
> +	master->num_chipselect = num_cs;
>  	master->cleanup = pl022_cleanup;
>  	master->setup = pl022_setup;
>  	master->prepare_transfer_hardware = pl022_prepare_transfer_hardware;
>  	master->transfer_one_message = pl022_transfer_one_message;
>  	master->unprepare_transfer_hardware = pl022_unprepare_transfer_hardware;
>  	master->rt = platform_info->rt;
> +	master->dev.of_node = dev->of_node;
> +
> +	if (IS_ENABLED(CONFIG_OF)) {
> +		for (i = 0; i < num_cs; i++) {
> +			int cs_gpio = of_get_named_gpio(np, "cs-gpios", i);
> +
> +			if (cs_gpio == -EPROBE_DEFER) {
> +				status = -EPROBE_DEFER;
> +				goto err_no_gpio;
> +			}
> +
> +			pl022->chipselect[i] = cs_gpio;
> +
> +			if (gpio_is_valid(cs_gpio)) {
> +				if (gpio_request(cs_gpio, "ssp-pl022"))
> +					dev_err(&adev->dev,
> +						"could not request %d gpio\n",
> +						cs_gpio);
> +				else if (gpio_direction_output(cs_gpio, 1))
> +					dev_err(&adev->dev,
> +						"could set gpio %d as output\n",
> +						cs_gpio);
> +			}
> +		}
> +	} else {
> +		for (i = 0; i < num_cs; i++)
> +			pl022->chipselect[i] = -EINVAL;
> +	}
>  
>  	/*
>  	 * Supports mode 0-3, loopback, and active low CS. Transfers are
> @@ -2130,6 +2205,7 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
>   err_no_ioremap:
>  	amba_release_regions(adev);
>   err_no_ioregion:
> + err_no_gpio:
>  	spi_master_put(master);
>   err_no_master:
>   err_no_pdata:
> 

Re: [PATCH RESEND v4] spi/pl022: add devicetree support

[Linus Walleij]

(Currently make sure to include Mark Brown (see To: line) on all SPI patches,
as he's collecting them.)

On Mon, Aug 20, 2012 at 1:39 PM, Alexandre Pereira da Silva
<aletes.xgr@gmail.com> wrote:

> Add the chipselect array and cur_cs properties to pl022 main structure

OK...

> diff --git a/Documentation/devicetree/bindings/spi/spi_pl022.txt b/Documentation/devicetree/bindings/spi/spi_pl022.txt

This looks fine to me.

> +#include <linux/gpio.h>
> +#include <linux/of_gpio.h>

Hm, it looks like two changes are embedded in this patch.

>
>  /*
>   * This macro is used to define some register default values.
> @@ -389,6 +391,8 @@ struct pl022 {
>         char                            *dummypage;
>         bool                            dma_running;
>  #endif
> +       int cur_cs;
> +       int chipselect[0];
>  };

You forgot to add kerneldoc for these two. Please add!

int chipselect[0] really? isn't int *chipselect what you really want
to store in there?

> +       if (gpio_is_valid(pl022->cur_cs))
> +               gpio_set_value(pl022->cur_cs, command);
> +       else
> +               pl022->cur_chip->cs_control(command);
> +}

So it seems like this should be two patches:

- One adding pl022->cur_cs and the ability for the driver to control
the chip select directly from msg->spi->chip_select
Make sure it is possible to populate pl022->chipselect also
from platform data since many people are using that
still.

- Another patch adding device tree and that stuff.

So please split it in two.

> -                       pl022->cur_chip->cs_control(SSP_CHIP_DESELECT);
> +                       pl022_cs_control(pl022, SSP_CHIP_DESELECT);

All these are OK and go in the first patch.

> +       pl022->cur_cs = pl022->chipselect[msg->spi->chip_select];

Goes into the first patch...

> @@ -1761,12 +1772,14 @@ static const struct pl022_config_chip pl022_default_chip_info = {
>  static int pl022_setup(struct spi_device *spi)
>  {
>         struct pl022_config_chip const *chip_info;
> +       struct pl022_config_chip chip_info_dt;

Goes into second patch.

>         struct chip_data *chip;
>         struct ssp_clock_params clk_freq = { .cpsdvsr = 0, .scr = 0};
>         int status = 0;
>         struct pl022 *pl022 = spi_master_get_devdata(spi->master);
>         unsigned int bits = spi->bits_per_word;
>         u32 tmp;
> +       struct device_node *np = spi->dev.of_node;

Does this compile if CONFIG_OF is not selected?

I have seen other drivers have #ifdef CONFIG_OF around these
things.

Make sure you test-compile without DT.

> @@ -1789,10 +1802,36 @@ static int pl022_setup(struct spi_device *spi)
>         chip_info = spi->controller_data;
>
>         if (chip_info == NULL) {
> -               chip_info = &pl022_default_chip_info;
> -               /* spi_board_info.controller_data not is supplied */
> -               dev_dbg(&spi->dev,
> -                       "using default controller_data settings\n");
> +               if (np) {
> +                       chip_info_dt = pl022_default_chip_info;
> +
> +                       of_property_read_u32(np, "pl022,hierarchy",
> +                               &chip_info_dt.hierarchy);
> +                       of_property_read_u32(np, "pl022,interface",
> +                               &chip_info_dt.iface);
> +                       chip_info_dt.slave_tx_disable =
> +                               of_property_read_bool(np,
> +                                       "pl022,slave-tx-disable");
> +                       of_property_read_u32(np, "pl022,com-mode",
> +                               &chip_info_dt.com_mode);
> +                       of_property_read_u32(np, "pl022,rx-level-trig",
> +                               &chip_info_dt.rx_lev_trig);
> +                       of_property_read_u32(np, "pl022,tx-level-trig",
> +                               &chip_info_dt.tx_lev_trig);
> +                       of_property_read_u32(np, "pl022,ctrl-len",
> +                               &chip_info_dt.ctrl_len);
> +                       of_property_read_u32(np, "pl022,wait-state",
> +                               &chip_info_dt.wait_state);
> +                       of_property_read_u32(np, "pl022,duplex",
> +                               &chip_info_dt.duplex);
> +
> +                       chip_info = &chip_info_dt;
> +               } else {
> +                       chip_info = &pl022_default_chip_info;
> +                       /* spi_board_info.controller_data not is supplied */
> +                       dev_dbg(&spi->dev,
> +                               "using default controller_data settings\n");
> +               }

Goes into second patch.

> @@ -1835,8 +1874,9 @@ static int pl022_setup(struct spi_device *spi)
>         chip->xfer_type = chip_info->com_mode;
>         if (!chip_info->cs_control) {
>                 chip->cs_control = null_cs_control;
> -               dev_warn(&spi->dev,
> -                        "chip select function is NULL for this chip\n");
> +               if (!gpio_is_valid(pl022->chipselect[spi->chip_select]))
> +                       dev_warn(&spi->dev,
> +                                "chip select function is NULL for this chip\n");
>         } else
>                 chip->cs_control = chip_info->cs_control;

Goes into first patch.

> @@ -1988,7 +2028,8 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
>         struct pl022_ssp_controller *platform_info = adev->dev.platform_data;
>         struct spi_master *master;
>         struct pl022 *pl022 = NULL;     /*Data for this driver */
> -       int status = 0;
> +       struct device_node *np = adev->dev.of_node;

Does this compile without CONFIG_OF?
(Second patch)

> +       int status = 0, i, num_cs;
>
>         dev_info(&adev->dev,
>                  "ARM PL022 driver, device ID: 0x%08x\n", adev->periphid);
> @@ -1998,8 +2039,14 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
>                 goto err_no_pdata;
>         }
>
> +       num_cs = platform_info->num_chipselect;
> +       if (IS_ENABLED(CONFIG_OF))
> +               of_property_read_u32(np, "pl022,num-chipselects", &num_cs);
> +
> +

Shouldn't it be the other way around: platform data overrides
DT data. Attempt DT and if platform_info->num_chipselect
!= 0 let it override.

(Second patch.)

>         /* Allocate master with space for data */
> -       master = spi_alloc_master(dev, sizeof(struct pl022));
> +       master = spi_alloc_master(dev,
> +               sizeof(struct pl022) + sizeof(int) * num_cs);

First patch.

>         if (master == NULL) {
>                 dev_err(&adev->dev, "probe - cannot alloc SPI master\n");
>                 status = -ENOMEM;
> @@ -2017,13 +2064,41 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id)
>          * on this board
>          */
>         master->bus_num = platform_info->bus_id;
> -       master->num_chipselect = platform_info->num_chipselect;
> +       master->num_chipselect = num_cs;

OK first patch.

>         master->cleanup = pl022_cleanup;
>         master->setup = pl022_setup;
>         master->prepare_transfer_hardware = pl022_prepare_transfer_hardware;
>         master->transfer_one_message = pl022_transfer_one_message;
>         master->unprepare_transfer_hardware = pl022_unprepare_transfer_hardware;
>         master->rt = platform_info->rt;
> +       master->dev.of_node = dev->of_node;

Does this compile without CONFIG_OF?
Second patch.

> +       if (IS_ENABLED(CONFIG_OF)) {
> +               for (i = 0; i < num_cs; i++) {
> +                       int cs_gpio = of_get_named_gpio(np, "cs-gpios", i);
> +
> +                       if (cs_gpio == -EPROBE_DEFER) {
> +                               status = -EPROBE_DEFER;
> +                               goto err_no_gpio;
> +                       }
> +
> +                       pl022->chipselect[i] = cs_gpio;
> +
> +                       if (gpio_is_valid(cs_gpio)) {
> +                               if (gpio_request(cs_gpio, "ssp-pl022"))
> +                                       dev_err(&adev->dev,
> +                                               "could not request %d gpio\n",
> +                                               cs_gpio);
> +                               else if (gpio_direction_output(cs_gpio, 1))
> +                                       dev_err(&adev->dev,
> +                                               "could set gpio %d as output\n",
> +                                               cs_gpio);
> +                       }
> +               }
> +       } else {
> +               for (i = 0; i < num_cs; i++)
> +                       pl022->chipselect[i] = -EINVAL;

Why? Instead, add a  int *chipselects; field to the struct pl022_ssp_controller
platform data in include/linux/amba/pl022.h and copy it from there if
num_chipselects in the same data != 0.

(First patch.)

Yours,
Linus Walleij

Re: [PATCH RESEND v4] spi/pl022: add devicetree support

[Linus Walleij]

On Mon, Aug 20, 2012 at 4:22 PM, Rob Herring <robherring2@gmail.com> wrote:
> On 08/20/2012 06:39 AM, Alexandre Pereira da Silva wrote:

>> +- pl022,hierarchy : master or slave interface
>
> Is attaching a master and using the pl022 as a slave really a supported
> use case?

No not currently. People doing this crazy stuff need to have
the kernel with the hard real-time patches and lightening response
first. But it's in the platform data because it's a feature supported by
the hardware.

> The DT spi bindings are really designed for a master node with
> N slave nodes. So there is more general question of how to describe a
> spi controller as a slave. It's not really one I care to answer as the
> Linux spi framework isn't designed to act as a slave.

Currently the code should hardwire this to master and not define
a binding for it I think.

>> +- pl022,slave-tx-disable : disconnect tx line in slave mode

Applies also to this, then.

>> +- pl022,com-mode : polling, interrupt or dma
>> +- pl022,rx-level-trig : Rx FIFO watermark level
>> +- pl022,tx-level-trig : Tx FIFO watermark level
>> +- pl022,ctrl-len : Microwire interface: Control length
>> +- pl022,wait-state : Microwire interface: Wait state
>> +- pl022,duplex : Microwire interface: Full/Half duplex
>
> Most of these properties aren't used anywhere in the kernel other than
> u300 and I'm not sure what to purpose of dummy_chip_info is.

It is used for loopback-testing of the PL022. They're probably
not necessary there either.

> Perhaps
> Linus has some input? I think either they can be decided by the spi
> controller (com-mode, fifo watermark) or should be standard properties
> (microwire settings). I would argue they should be removed if the spi
> framework doesn't have support in a standard way.

Uncertain about this, others would need to comment.
At some point there has been a chip needing these to be set
to some magic values.

Yours,
Linus Walleij

Re: [PATCH RESEND v4] spi/pl022: add devicetree support

[Roland Stigge]

On 08/20/2012 04:22 PM, Rob Herring wrote:
>>  .../devicetree/bindings/spi/spi_pl022.txt          |   15 +++
>>  drivers/spi/spi-pl022.c                            |  114 ++++++++++++++++----
>>  2 files changed, 110 insertions(+), 19 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/spi/spi_pl022.txt b/Documentation/devicetree/bindings/spi/spi_pl022.txt
>> index 306ec3f..b089ec7 100644
>> --- a/Documentation/devicetree/bindings/spi/spi_pl022.txt
>> +++ b/Documentation/devicetree/bindings/spi/spi_pl022.txt
>> @@ -6,7 +6,22 @@ Required properties:
>>  - interrupts : Should contain SPI controller interrupt
>>  
>>  Optional properties:
>> +- pl022,num-chipselects : total number of chipselects
> 
> We have this for other spi controllers too. Define a generic property.

Currently, we have:

num-cs
ti,spi-num-cs
fsl,spi-num-chipselects
fsl,espi-num-chipselects
(pl022,num-chipselects)

Would "num-cs" be the right one to use?

Thanks,

Roland

Re: [PATCH RESEND v4] spi/pl022: add devicetree support

[Rob Herring]

On 08/20/2012 10:33 AM, Roland Stigge wrote:
> On 08/20/2012 04:22 PM, Rob Herring wrote:
>>>  .../devicetree/bindings/spi/spi_pl022.txt          |   15 +++
>>>  drivers/spi/spi-pl022.c                            |  114 ++++++++++++++++----
>>>  2 files changed, 110 insertions(+), 19 deletions(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/spi/spi_pl022.txt b/Documentation/devicetree/bindings/spi/spi_pl022.txt
>>> index 306ec3f..b089ec7 100644
>>> --- a/Documentation/devicetree/bindings/spi/spi_pl022.txt
>>> +++ b/Documentation/devicetree/bindings/spi/spi_pl022.txt
>>> @@ -6,7 +6,22 @@ Required properties:
>>>  - interrupts : Should contain SPI controller interrupt
>>>  
>>>  Optional properties:
>>> +- pl022,num-chipselects : total number of chipselects
>>
>> We have this for other spi controllers too. Define a generic property.
> 
> Currently, we have:
> 
> num-cs
> ti,spi-num-cs
> fsl,spi-num-chipselects
> fsl,espi-num-chipselects
> (pl022,num-chipselects)
> 
> Would "num-cs" be the right one to use?

Well, it is either that or invent something new, so I think num-cs is fine.

Rob