All of lore.kernel.org
 help / color / mirror / Atom feed
From: Linus Walleij <linus.walleij@linaro.org>
To: Mark Brown <broonie@kernel.org>, linux-spi@vger.kernel.org
Cc: linux-gpio@vger.kernel.org,
	Bartosz Golaszewski <bgolaszewski@baylibre.com>,
	Linus Walleij <linus.walleij@linaro.org>,
	Atsushi Nemoto <atsushi.nemoto@sord.co.jp>
Subject: [PATCH] spi: txx9: Convert to use GPIO descriptors
Date: Wed, 30 Oct 2019 08:38:32 +0100	[thread overview]
Message-ID: <20191030073832.24038-1-linus.walleij@linaro.org> (raw)

This converts the TXX9 SPI driver to use GPIO descriptors
to control the GPIO chip selects.

As the driver was clearly (ab)using the device tree "reg"
property to offset into the global GPIO chip we have to
add a hack to counter the hack: add a 1-to-1 chip select
to GPIO offset mapping for all 16 lines on the TXX9 GPIO
chip. The details are described in a largeish comment
in the patch.

We do not need to set up the GPIO as output any more since
the core will take care of this, as well as it will handle
the polarity inversion semantics.

Cc: Atsushi Nemoto <atsushi.nemoto@sord.co.jp>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 drivers/spi/spi-txx9.c | 75 ++++++++++++++++++++++++++++++++----------
 1 file changed, 58 insertions(+), 17 deletions(-)

diff --git a/drivers/spi/spi-txx9.c b/drivers/spi/spi-txx9.c
index 51759d3fd45f..22b19d960d3b 100644
--- a/drivers/spi/spi-txx9.c
+++ b/drivers/spi/spi-txx9.c
@@ -26,7 +26,8 @@
 #include <linux/clk.h>
 #include <linux/io.h>
 #include <linux/module.h>
-#include <linux/gpio.h>
+#include <linux/gpio/machine.h>
+#include <linux/gpio/consumer.h>
 
 
 #define SPI_FIFO_SIZE 4
@@ -79,7 +80,7 @@ struct txx9spi {
 	void __iomem *membase;
 	int baseclk;
 	struct clk *clk;
-	int last_chipselect;
+	struct gpio_desc *last_chipselect;
 	int last_chipselect_val;
 };
 
@@ -95,20 +96,22 @@ static void txx9spi_wr(struct txx9spi *c, u32 val, int reg)
 static void txx9spi_cs_func(struct spi_device *spi, struct txx9spi *c,
 		int on, unsigned int cs_delay)
 {
-	int val = (spi->mode & SPI_CS_HIGH) ? on : !on;
-
+	/*
+	 * The GPIO descriptor will track polarity inversion inside
+	 * gpiolib.
+	 */
 	if (on) {
 		/* deselect the chip with cs_change hint in last transfer */
-		if (c->last_chipselect >= 0)
-			gpio_set_value(c->last_chipselect,
+		if (c->last_chipselect)
+			gpiod_set_value(c->last_chipselect,
 					!c->last_chipselect_val);
-		c->last_chipselect = spi->chip_select;
-		c->last_chipselect_val = val;
+		c->last_chipselect = spi->cs_gpiod;
+		c->last_chipselect_val = on;
 	} else {
-		c->last_chipselect = -1;
+		c->last_chipselect = NULL;
 		ndelay(cs_delay);	/* CS Hold Time */
 	}
-	gpio_set_value(spi->chip_select, val);
+	gpiod_set_value(spi->cs_gpiod, on);
 	ndelay(cs_delay);	/* CS Setup Time / CS Recovery Time */
 }
 
@@ -119,12 +122,6 @@ static int txx9spi_setup(struct spi_device *spi)
 	if (!spi->max_speed_hz)
 		return -EINVAL;
 
-	if (gpio_direction_output(spi->chip_select,
-			!(spi->mode & SPI_CS_HIGH))) {
-		dev_err(&spi->dev, "Cannot setup GPIO for chipselect.\n");
-		return -EINVAL;
-	}
-
 	/* deselect chip */
 	spin_lock(&c->lock);
 	txx9spi_cs_func(spi, c, 0, (NSEC_PER_SEC / 2) / spi->max_speed_hz);
@@ -320,6 +317,47 @@ static int txx9spi_transfer(struct spi_device *spi, struct spi_message *m)
 	return 0;
 }
 
+/*
+ * Chip select uses GPIO only, further the driver is using the chip select
+ * numer (from the device tree "reg" property, and this can only come from
+ * device tree since this i MIPS and there is no way to pass platform data) as
+ * the GPIO number. As the platform has only one GPIO controller (the txx9 GPIO
+ * chip) it is thus using the chip select number as an offset into that chip.
+ * This chip has a maximum of 16 GPIOs 0..15 and this is what all platforms
+ * register.
+ *
+ * We modernized this behaviour by explicitly converting that offset to an
+ * offset on the GPIO chip using a GPIO descriptor machine table of the same
+ * size as the txx9 GPIO chip with a 1-to-1 mapping of chip select to GPIO
+ * offset.
+ *
+ * This is admittedly a hack, but it is countering the hack of using "reg" to
+ * contain a GPIO offset when it should be using "cs-gpios" as the SPI bindings
+ * state.
+ */
+static struct gpiod_lookup_table txx9spi_cs_gpio_table = {
+	.dev_id = "spi0",
+	.table = {
+		GPIO_LOOKUP_IDX("TXx9", 0, "cs", 0, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 1, "cs", 1, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 2, "cs", 2, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 3, "cs", 3, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 4, "cs", 4, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 5, "cs", 5, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 6, "cs", 6, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 7, "cs", 7, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 8, "cs", 8, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 9, "cs", 9, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 10, "cs", 10, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 11, "cs", 11, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 12, "cs", 12, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 13, "cs", 13, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 14, "cs", 14, GPIO_ACTIVE_LOW),
+		GPIO_LOOKUP_IDX("TXx9", 15, "cs", 15, GPIO_ACTIVE_LOW),
+		{ },
+	},
+};
+
 static int txx9spi_probe(struct platform_device *dev)
 {
 	struct spi_master *master;
@@ -373,12 +411,14 @@ static int txx9spi_probe(struct platform_device *dev)
 	if (ret)
 		goto exit;
 
-	c->last_chipselect = -1;
+	c->last_chipselect = NULL;
 
 	dev_info(&dev->dev, "at %#llx, irq %d, %dMHz\n",
 		 (unsigned long long)res->start, irq,
 		 (c->baseclk + 500000) / 1000000);
 
+	gpiod_add_lookup_table(&txx9spi_cs_gpio_table);
+
 	/* the spi->mode bits understood by this driver: */
 	master->mode_bits = SPI_CS_HIGH | SPI_CPOL | SPI_CPHA;
 
@@ -387,6 +427,7 @@ static int txx9spi_probe(struct platform_device *dev)
 	master->transfer = txx9spi_transfer;
 	master->num_chipselect = (u16)UINT_MAX; /* any GPIO numbers */
 	master->bits_per_word_mask = SPI_BPW_MASK(8) | SPI_BPW_MASK(16);
+	master->use_gpio_descriptors = true;
 
 	ret = devm_spi_register_master(&dev->dev, master);
 	if (ret)
-- 
2.21.0


             reply	other threads:[~2019-10-30  7:40 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-30  7:38 Linus Walleij [this message]
2019-10-30 14:16 ` Applied "spi: txx9: Convert to use GPIO descriptors" to the spi tree Mark Brown

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191030073832.24038-1-linus.walleij@linaro.org \
    --to=linus.walleij@linaro.org \
    --cc=atsushi.nemoto@sord.co.jp \
    --cc=bgolaszewski@baylibre.com \
    --cc=broonie@kernel.org \
    --cc=linux-gpio@vger.kernel.org \
    --cc=linux-spi@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.