[PATCH 0/5] Add STM32 DMAMUX support

[PATCH 0/5] Add STM32 DMAMUX support

[M'boumba Cedric Madianga]

This patchset adds support for the STM32 DMA multiplexer.
It allows to map any peripheral DMA request to any channel of the product
DMAs.
This IP has been introduced with STM32H7 SoC.

M'boumba Cedric Madianga (5):
  dt-bindings: Document the STM32 DMAMUX bindings
  dmaengine: Add STM32 DMAMUX driver
  dt-bindings: stm32-dma: Add property to handle STM32 DMAMUX
  dmaengine: stm32-dma: Add support for STM32 DMAMUX
  ARM: configs: stm32: Add DMAMUX support in STM32 defconfig

 .../devicetree/bindings/dma/stm32-dma.txt          |   3 +
 .../devicetree/bindings/dma/stm32-dmamux.txt       |  57 +++++
 arch/arm/configs/stm32_defconfig                   |   1 +
 drivers/dma/Kconfig                                |   9 +
 drivers/dma/Makefile                               |   1 +
 drivers/dma/stm32-dma.c                            |  13 +-
 drivers/dma/stm32-dmamux.c                         | 231 +++++++++++++++++++++
 7 files changed, 312 insertions(+), 3 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/dma/stm32-dmamux.txt
 create mode 100644 drivers/dma/stm32-dmamux.c

-- 
1.9.1

[PATCH 1/5] dt-bindings: Document the STM32 DMAMUX bindings

[M'boumba Cedric Madianga]

This patch adds the documentation of device tree bindings for the STM32
DMAMUX.

Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
---
 .../devicetree/bindings/dma/stm32-dmamux.txt       | 57 ++++++++++++++++++++++
 1 file changed, 57 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/dma/stm32-dmamux.txt

diff --git a/Documentation/devicetree/bindings/dma/stm32-dmamux.txt b/Documentation/devicetree/bindings/dma/stm32-dmamux.txt
new file mode 100644
index 0000000..1039420
--- /dev/null
+++ b/Documentation/devicetree/bindings/dma/stm32-dmamux.txt
@@ -0,0 +1,57 @@
+STM32 DMA MUX (DMA request router)
+
+Required properties:
+- compatible:	"st,stm32-dmamux"
+- reg:		Memory map for accessing module
+- #dma-cells:	Should be set to <4>.
+		For more details about the four cells, please see stm32-dma.txt
+		documentation binding file
+- dma-masters:	Phandle pointing to the DMA controller
+- clocks: Input clock of the DMAMUX instance.
+
+Optional properties:
+- dma-channels : Number of DMA channels supported.
+- dma-requests : Number of DMA requests supported.
+- resets: Reference to a reset controller asserting the DMA controller
+
+Example:
+
+/* DMA controller */
+dma2: dma-controller@40026400 {
+	compatible = "st,stm32-dma";
+	reg = <0x40026400 0x400>;
+	interrupts = <56>,
+		     <57>,
+		     <58>,
+		     <59>,
+		     <60>,
+		     <68>,
+		     <69>,
+		     <70>;
+	clocks = <&clk_hclk>;
+	#dma-cells = <4>;
+	st,mem2mem;
+	resets = <&rcc 150>;
+	st,dmamux;
+	dma-channels = <8>;
+};
+
+/* DMA mux */
+dmamux2: dma-router@40020820 {
+	compatible = "st,stm32-dmamux";
+	reg = <0x40020800 0x1c>;
+	#dma-cells = <4>;
+	dma-requests = <128>;
+	dma-masters = <&dma2>;
+};
+
+/* DMA client */
+usart1: serial@40011000 {
+	compatible = "st,stm32-usart", "st,stm32-uart";
+	reg = <0x40011000 0x400>;
+	interrupts = <37>;
+	clocks = <&clk_pclk2>;
+	dmas = <&dmamux2 0 41 0x400 0x00>,
+	       <&dmamux2 1 42 0x400 0x00>;
+	dma-names = "rx", "tx";
+};
-- 
1.9.1

[PATCH 2/5] dmaengine: Add STM32 DMAMUX driver

[M'boumba Cedric Madianga]

This patch implements the STM32 DMAMUX driver

Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
---
 drivers/dma/Kconfig        |   9 ++
 drivers/dma/Makefile       |   1 +
 drivers/dma/stm32-dmamux.c | 231 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 241 insertions(+)
 create mode 100644 drivers/dma/stm32-dmamux.c

diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
index fc3435c..6ab80c9 100644
--- a/drivers/dma/Kconfig
+++ b/drivers/dma/Kconfig
@@ -461,6 +461,15 @@ config STM32_DMA
 	  If you have a board based on such a MCU and wish to use DMA say Y
 	  here.
 
+config STM32_DMAMUX
+	bool "STMicroelectronics STM32 dma multiplexer support"
+	depends on STM32_DMA
+	help
+	  Enable support for the on-chip DMA multiplexer on STMicroelectronics
+	  STM32 MCUs.
+	  If you have a board based on such a MCU and wish to use DMAMUX say Y
+	  here.
+
 config S3C24XX_DMAC
 	bool "Samsung S3C24XX DMA support"
 	depends on ARCH_S3C24XX || COMPILE_TEST
diff --git a/drivers/dma/Makefile b/drivers/dma/Makefile
index 0b723e9..449c7a3 100644
--- a/drivers/dma/Makefile
+++ b/drivers/dma/Makefile
@@ -57,6 +57,7 @@ obj-$(CONFIG_RENESAS_DMA) += sh/
 obj-$(CONFIG_SIRF_DMA) += sirf-dma.o
 obj-$(CONFIG_STE_DMA40) += ste_dma40.o ste_dma40_ll.o
 obj-$(CONFIG_STM32_DMA) += stm32-dma.o
+obj-$(CONFIG_STM32_DMAMUX) += stm32-dmamux.o
 obj-$(CONFIG_S3C24XX_DMAC) += s3c24xx-dma.o
 obj-$(CONFIG_TXX9_DMAC) += txx9dmac.o
 obj-$(CONFIG_TEGRA20_APB_DMA) += tegra20-apb-dma.o
diff --git a/drivers/dma/stm32-dmamux.c b/drivers/dma/stm32-dmamux.c
new file mode 100644
index 0000000..3003546
--- /dev/null
+++ b/drivers/dma/stm32-dmamux.c
@@ -0,0 +1,231 @@
+/*
+ * DMA Router driver for STM32 DMA MUX
+ *
+ * Copyright (C) 2015 M'Boumba Cedric Madianga <cedric.madianga@gmail.com>
+ *
+ * Based on LPC18xx/43xx DMA MUX and TI DMA XBAR
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/of_device.h>
+#include <linux/of_dma.h>
+#include <linux/reset.h>
+#include <linux/slab.h>
+
+#define STM32_DMAMUX_CCR(x)		(0x4 * (x))
+#define STM32_DMAMUX_MAX_CHANNELS	32
+#define STM32_DMAMUX_MAX_REQUESTS	255
+
+struct stm32_dmamux {
+	u32 chan_id;
+	u32 request;
+	bool busy;
+};
+
+struct stm32_dmamux_data {
+	struct dma_router dmarouter;
+	struct stm32_dmamux *muxes;
+	struct clk *clk;
+	void __iomem *iomem;
+	u32 dmamux_requests; /* number of DMA requests connected to DMAMUX */
+	u32 dmamux_channels; /* Number of DMA channels supported */
+};
+
+static inline u32 stm32_dmamux_read(void __iomem *iomem, u32 reg)
+{
+	return readl_relaxed(iomem + reg);
+}
+
+static inline void stm32_dmamux_write(void __iomem *iomem, u32 reg, u32 val)
+{
+	writel_relaxed(val, iomem + reg);
+}
+
+static void stm32_dmamux_free(struct device *dev, void *route_data)
+{
+	struct stm32_dmamux_data *dmamux = dev_get_drvdata(dev);
+	struct stm32_dmamux *mux = route_data;
+
+	/* Clear dma request for the right channel */
+	stm32_dmamux_write(dmamux->iomem, STM32_DMAMUX_CCR(mux->chan_id), 0);
+	clk_disable(dmamux->clk);
+	mux->busy = false;
+
+	dev_dbg(dev, "Unmapping dma-router%dchan%d (was routed to request%d)\n",
+		dev->id, mux->chan_id, mux->request);
+}
+
+static void *stm32_dmamux_route_allocate(struct of_phandle_args *dma_spec,
+					 struct of_dma *ofdma)
+{
+	struct platform_device *pdev = of_find_device_by_node(ofdma->of_node);
+	struct stm32_dmamux_data *dmamux = platform_get_drvdata(pdev);
+	struct stm32_dmamux *mux;
+	u32 chan_id;
+	int ret;
+
+	if (dma_spec->args_count != 4) {
+		dev_err(&pdev->dev, "invalid number of dma mux args\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (dma_spec->args[0] >= dmamux->dmamux_channels) {
+		dev_err(&pdev->dev, "invalid channel id: %d\n",
+			dma_spec->args[0]);
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (dma_spec->args[1] > dmamux->dmamux_requests) {
+		dev_err(&pdev->dev, "invalid mux request number: %d\n",
+			dma_spec->args[1]);
+		return ERR_PTR(-EINVAL);
+	}
+
+	/* The of_node_put() will be done in of_dma_router_xlate function */
+	dma_spec->np = of_parse_phandle(ofdma->of_node, "dma-masters", 0);
+	if (!dma_spec->np) {
+		dev_err(&pdev->dev, "can't get dma master\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	chan_id = dma_spec->args[0];
+	mux = &dmamux->muxes[chan_id];
+	mux->chan_id = chan_id;
+	mux->request = dma_spec->args[1];
+
+	if (mux->busy) {
+		dev_err(&pdev->dev, "dma channel %d busy with request %d\n",
+			chan_id, mux->request);
+		return ERR_PTR(-EBUSY);
+	}
+
+	ret = clk_enable(dmamux->clk);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "clk_enable failed: %d\n", ret);
+		return ERR_PTR(ret);
+	}
+
+	stm32_dmamux_write(dmamux->iomem, STM32_DMAMUX_CCR(mux->chan_id),
+			   mux->request);
+	mux->busy = true;
+
+	dev_dbg(&pdev->dev, "Mapping dma-router%dchan%d to request%d\n",
+		pdev->dev.id, mux->chan_id, mux->request);
+
+	return mux;
+}
+
+static int stm32_dmamux_probe(struct platform_device *pdev)
+{
+	struct device_node *dma_node, *node = pdev->dev.of_node;
+	struct stm32_dmamux_data *dmamux;
+	struct reset_control *rst;
+	struct resource *res;
+	int ret;
+
+	if (!node)
+		return -ENODEV;
+
+	dmamux = devm_kzalloc(&pdev->dev, sizeof(struct stm32_dmamux_data),
+			      GFP_KERNEL);
+	if (!dmamux)
+		return -ENOMEM;
+
+	dma_node = of_parse_phandle(node, "dma-masters", 0);
+	if (!dma_node) {
+		dev_err(&pdev->dev, "Can't get DMA master node\n");
+		return -ENODEV;
+	}
+	of_node_put(dma_node);
+
+	ret = of_property_read_u32(node, "dma-channels",
+				   &dmamux->dmamux_channels);
+	if (ret)
+		dmamux->dmamux_channels = STM32_DMAMUX_MAX_CHANNELS;
+
+	ret = of_property_read_u32(node, "dma-requests",
+				   &dmamux->dmamux_requests);
+	if (ret)
+		dmamux->dmamux_requests = STM32_DMAMUX_MAX_REQUESTS;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -ENODEV;
+
+	dmamux->iomem = devm_ioremap_resource(&pdev->dev, res);
+	if (!dmamux->iomem)
+		return -ENOMEM;
+
+	dmamux->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(dmamux->clk)) {
+		dev_err(&pdev->dev, "Missing controller clock\n");
+		return PTR_ERR(dmamux->clk);
+	}
+	ret = clk_prepare(dmamux->clk);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "clk_prep failed: %d\n", ret);
+		return ret;
+	}
+
+	dmamux->muxes = devm_kcalloc(&pdev->dev, dmamux->dmamux_channels,
+				     sizeof(struct stm32_dmamux),
+				     GFP_KERNEL);
+	if (!dmamux->muxes)
+		return -ENOMEM;
+
+	rst = devm_reset_control_get(&pdev->dev, NULL);
+	if (!IS_ERR(rst)) {
+		ret = clk_enable(dmamux->clk);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "clk_enable failed: %d\n", ret);
+			return ret;
+		}
+		reset_control_assert(rst);
+		udelay(2);
+		reset_control_deassert(rst);
+		clk_disable(dmamux->clk);
+	}
+
+	dmamux->dmarouter.dev = &pdev->dev;
+	dmamux->dmarouter.route_free = stm32_dmamux_free;
+	platform_set_drvdata(pdev, dmamux);
+
+	ret = of_dma_router_register(node, stm32_dmamux_route_allocate,
+				     &dmamux->dmarouter);
+	if (ret < 0) {
+		dev_err(&pdev->dev,
+			"STM32 DMAMUX DMA OF registration failed %d\n", ret);
+		return ret;
+	}
+
+	dev_info(&pdev->dev, "STM32 DMAMUX driver registered\n");
+
+	return 0;
+}
+
+static const struct of_device_id stm32_dmamux_match[] = {
+	{ .compatible = "st,stm32-dmamux" },
+	{},
+};
+
+static struct platform_driver stm32_dmamux_driver = {
+	.probe	= stm32_dmamux_probe,
+	.driver = {
+		.name = "stm32-dmamux",
+		.of_match_table = stm32_dmamux_match,
+	},
+};
+
+static int __init stm32_dmamux_init(void)
+{
+	return platform_driver_register(&stm32_dmamux_driver);
+}
+arch_initcall(stm32_dmamux_init);
-- 
1.9.1

[PATCH 3/5] dt-bindings: stm32-dma: Add property to handle STM32 DMAMUX

[M'boumba Cedric Madianga]

This patch adds an optional property needed for STM32 DMA controller
addressed via STM32 DMAMUX.

Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
---
 Documentation/devicetree/bindings/dma/stm32-dma.txt | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/Documentation/devicetree/bindings/dma/stm32-dma.txt b/Documentation/devicetree/bindings/dma/stm32-dma.txt
index 4408af6..7b5e91a 100644
--- a/Documentation/devicetree/bindings/dma/stm32-dma.txt
+++ b/Documentation/devicetree/bindings/dma/stm32-dma.txt
@@ -16,6 +16,9 @@ Optional properties:
 - resets: Reference to a reset controller asserting the DMA controller
 - st,mem2mem: boolean; if defined, it indicates that the controller supports
   memory-to-memory transfer
+- st,dmamux: boolean; if defined, it indicates that the controller is behind a
+  DMA multiplexer. In that case, using dma instances doesn't work for DMA
+  clients. They have to use dma-router instances.
 
 Example:
 
-- 
1.9.1

[PATCH 4/5] dmaengine: stm32-dma: Add support for STM32 DMAMUX

[M'boumba Cedric Madianga]

This patch adds support for STM32 DMAMUX.
When the STM32 DMA controller is behind a STM32 DMAMUX the request line
number has not to be handled by DMA but DMAMUX.

Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
---
 drivers/dma/stm32-dma.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

diff --git a/drivers/dma/stm32-dma.c b/drivers/dma/stm32-dma.c
index 786fc8f..3bf2893 100644
--- a/drivers/dma/stm32-dma.c
+++ b/drivers/dma/stm32-dma.c
@@ -179,6 +179,7 @@ struct stm32_dma_device {
 	struct clk *clk;
 	struct reset_control *rst;
 	bool mem2mem;
+	bool dmamux;
 	struct stm32_dma_chan chan[STM32_DMA_MAX_CHANNELS];
 };
 
@@ -968,10 +969,14 @@ static void stm32_dma_desc_free(struct virt_dma_desc *vdesc)
 static void stm32_dma_set_config(struct stm32_dma_chan *chan,
 			  struct stm32_dma_cfg *cfg)
 {
+	struct stm32_dma_device *dmadev = stm32_dma_get_dev(chan);
+
 	stm32_dma_clear_reg(&chan->chan_reg);
 
 	chan->chan_reg.dma_scr = cfg->stream_config & STM32_DMA_SCR_CFG_MASK;
-	chan->chan_reg.dma_scr |= STM32_DMA_SCR_REQ(cfg->request_line);
+
+	if (!dmadev->dmamux)
+		chan->chan_reg.dma_scr |= STM32_DMA_SCR_REQ(cfg->request_line);
 
 	/* Enable Interrupts  */
 	chan->chan_reg.dma_scr |= STM32_DMA_SCR_TEIE | STM32_DMA_SCR_TCIE;
@@ -998,8 +1003,8 @@ static struct dma_chan *stm32_dma_of_xlate(struct of_phandle_args *dma_spec,
 	cfg.stream_config = dma_spec->args[2];
 	cfg.threshold = dma_spec->args[3];
 
-	if ((cfg.channel_id >= STM32_DMA_MAX_CHANNELS) ||
-	    (cfg.request_line >= STM32_DMA_MAX_REQUEST_ID)) {
+	if ((!dmadev->dmamux && cfg.request_line >= STM32_DMA_MAX_REQUEST_ID) ||
+	    (cfg.channel_id >= STM32_DMA_MAX_CHANNELS)) {
 		dev_err(dev, "Bad channel and/or request id\n");
 		return NULL;
 	}
@@ -1058,6 +1063,8 @@ static int stm32_dma_probe(struct platform_device *pdev)
 	dmadev->mem2mem = of_property_read_bool(pdev->dev.of_node,
 						"st,mem2mem");
 
+	dmadev->dmamux = of_property_read_bool(pdev->dev.of_node, "st,dmamux");
+
 	dmadev->rst = devm_reset_control_get(&pdev->dev, NULL);
 	if (!IS_ERR(dmadev->rst)) {
 		reset_control_assert(dmadev->rst);
-- 
1.9.1

[PATCH 5/5] ARM: configs: stm32: Add DMAMUX support in STM32 defconfig

[M'boumba Cedric Madianga]

This patch adds DMAMUX support in STM32 defconfig file

Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
---
 arch/arm/configs/stm32_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/configs/stm32_defconfig b/arch/arm/configs/stm32_defconfig
index a9d8e3c..afb6832 100644
--- a/arch/arm/configs/stm32_defconfig
+++ b/arch/arm/configs/stm32_defconfig
@@ -60,6 +60,7 @@ CONFIG_RTC_CLASS=y
 CONFIG_RTC_DRV_STM32=y
 CONFIG_DMADEVICES=y
 CONFIG_STM32_DMA=y
+CONFIG_STM32_DMAMUX=y
 CONFIG_IIO=y
 CONFIG_STM32_ADC_CORE=y
 CONFIG_STM32_ADC=y
-- 
1.9.1

Re: [PATCH 1/5] dt-bindings: Document the STM32 DMAMUX bindings

[Rob Herring]

On Mon, Mar 13, 2017 at 03:15:57PM +0100, M'boumba Cedric Madianga wrote:
> This patch adds the documentation of device tree bindings for the STM32
> DMAMUX.
> 
> Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
> ---
>  .../devicetree/bindings/dma/stm32-dmamux.txt       | 57 ++++++++++++++++++++++
>  1 file changed, 57 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/dma/stm32-dmamux.txt
> 
> diff --git a/Documentation/devicetree/bindings/dma/stm32-dmamux.txt b/Documentation/devicetree/bindings/dma/stm32-dmamux.txt
> new file mode 100644
> index 0000000..1039420
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/dma/stm32-dmamux.txt
> @@ -0,0 +1,57 @@
> +STM32 DMA MUX (DMA request router)
> +
> +Required properties:
> +- compatible:	"st,stm32-dmamux"
> +- reg:		Memory map for accessing module
> +- #dma-cells:	Should be set to <4>.
> +		For more details about the four cells, please see stm32-dma.txt
> +		documentation binding file
> +- dma-masters:	Phandle pointing to the DMA controller
> +- clocks: Input clock of the DMAMUX instance.
> +
> +Optional properties:
> +- dma-channels : Number of DMA channels supported.
> +- dma-requests : Number of DMA requests supported.
> +- resets: Reference to a reset controller asserting the DMA controller
> +
> +Example:
> +
> +/* DMA controller */
> +dma2: dma-controller@40026400 {
> +	compatible = "st,stm32-dma";
> +	reg = <0x40026400 0x400>;
> +	interrupts = <56>,
> +		     <57>,
> +		     <58>,
> +		     <59>,
> +		     <60>,
> +		     <68>,
> +		     <69>,
> +		     <70>;
> +	clocks = <&clk_hclk>;
> +	#dma-cells = <4>;
> +	st,mem2mem;
> +	resets = <&rcc 150>;
> +	st,dmamux;
> +	dma-channels = <8>;
> +};
> +
> +/* DMA mux */
> +dmamux2: dma-router@40020820 {
> +	compatible = "st,stm32-dmamux";
> +	reg = <0x40020800 0x1c>;
> +	#dma-cells = <4>;
> +	dma-requests = <128>;
> +	dma-masters = <&dma2>;

I think this should be modeled after the interrupt-map property (or 
Stephen Boyd's gpio-map support which additionally allows pass thru of 
cell values). Something like this:

dma-map = <0 41 0 0 &dma2 0 <req> 0 0>,
	<1 42 0 0 &dma2 1 <req> 0 0>;
dma-map-mask = <0xffffffff 0xffffffff 0 0>;

<req> is the request number on dma2 controller.

This is more generic and would work if you have a single mux with 
multiple DMA controllers.

> +};
> +
> +/* DMA client */
> +usart1: serial@40011000 {
> +	compatible = "st,stm32-usart", "st,stm32-uart";
> +	reg = <0x40011000 0x400>;
> +	interrupts = <37>;
> +	clocks = <&clk_pclk2>;
> +	dmas = <&dmamux2 0 41 0x400 0x00>,
> +	       <&dmamux2 1 42 0x400 0x00>;
> +	dma-names = "rx", "tx";
> +};
> -- 
> 1.9.1
> 

Re: [PATCH 3/5] dt-bindings: stm32-dma: Add property to handle STM32 DMAMUX

[Rob Herring]

On Mon, Mar 13, 2017 at 03:15:59PM +0100, M'boumba Cedric Madianga wrote:
> This patch adds an optional property needed for STM32 DMA controller
> addressed via STM32 DMAMUX.
> 
> Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
> ---
>  Documentation/devicetree/bindings/dma/stm32-dma.txt | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/dma/stm32-dma.txt b/Documentation/devicetree/bindings/dma/stm32-dma.txt
> index 4408af6..7b5e91a 100644
> --- a/Documentation/devicetree/bindings/dma/stm32-dma.txt
> +++ b/Documentation/devicetree/bindings/dma/stm32-dma.txt
> @@ -16,6 +16,9 @@ Optional properties:
>  - resets: Reference to a reset controller asserting the DMA controller
>  - st,mem2mem: boolean; if defined, it indicates that the controller supports
>    memory-to-memory transfer
> +- st,dmamux: boolean; if defined, it indicates that the controller is behind a
> +  DMA multiplexer. In that case, using dma instances doesn't work for DMA
> +  clients. They have to use dma-router instances.

This should not be needed for the same reason we don't need anything 
like this for chained interrupt controllers.

Also, the compatible string should be specific enough to provide this 
information.

Rob

Re: [PATCH 2/5] dmaengine: Add STM32 DMAMUX driver

[Vinod Koul]

On Mon, Mar 13, 2017 at 03:15:58PM +0100, M'boumba Cedric Madianga wrote:
> This patch implements the STM32 DMAMUX driver

Can you describe the controller here pls

> 
> Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
> ---
>  drivers/dma/Kconfig        |   9 ++
>  drivers/dma/Makefile       |   1 +
>  drivers/dma/stm32-dmamux.c | 231 +++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 241 insertions(+)
>  create mode 100644 drivers/dma/stm32-dmamux.c
> 
> diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
> index fc3435c..6ab80c9 100644
> --- a/drivers/dma/Kconfig
> +++ b/drivers/dma/Kconfig
> @@ -461,6 +461,15 @@ config STM32_DMA
>  	  If you have a board based on such a MCU and wish to use DMA say Y
>  	  here.
>  
> +config STM32_DMAMUX
> +	bool "STMicroelectronics STM32 dma multiplexer support"
> +	depends on STM32_DMA

can you add compile test here so that we can get better compile coverage

> +	help
> +	  Enable support for the on-chip DMA multiplexer on STMicroelectronics
> +	  STM32 MCUs.
> +	  If you have a board based on such a MCU and wish to use DMAMUX say Y
> +	  here.
> +
>  config S3C24XX_DMAC
>  	bool "Samsung S3C24XX DMA support"
>  	depends on ARCH_S3C24XX || COMPILE_TEST
> diff --git a/drivers/dma/Makefile b/drivers/dma/Makefile
> index 0b723e9..449c7a3 100644
> --- a/drivers/dma/Makefile
> +++ b/drivers/dma/Makefile
> @@ -57,6 +57,7 @@ obj-$(CONFIG_RENESAS_DMA) += sh/
>  obj-$(CONFIG_SIRF_DMA) += sirf-dma.o
>  obj-$(CONFIG_STE_DMA40) += ste_dma40.o ste_dma40_ll.o
>  obj-$(CONFIG_STM32_DMA) += stm32-dma.o
> +obj-$(CONFIG_STM32_DMAMUX) += stm32-dmamux.o
>  obj-$(CONFIG_S3C24XX_DMAC) += s3c24xx-dma.o
>  obj-$(CONFIG_TXX9_DMAC) += txx9dmac.o
>  obj-$(CONFIG_TEGRA20_APB_DMA) += tegra20-apb-dma.o
> diff --git a/drivers/dma/stm32-dmamux.c b/drivers/dma/stm32-dmamux.c
> new file mode 100644
> index 0000000..3003546
> --- /dev/null
> +++ b/drivers/dma/stm32-dmamux.c
> @@ -0,0 +1,231 @@
> +/*
> + * DMA Router driver for STM32 DMA MUX
> + *
> + * Copyright (C) 2015 M'Boumba Cedric Madianga <cedric.madianga@gmail.com>

we are in '17 now :)

> + *
> + * Based on LPC18xx/43xx DMA MUX and TI DMA XBAR
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + *
> + */
> +
> +#include <linux/clk.h>
> +#include <linux/delay.h>
> +#include <linux/err.h>
> +#include <linux/init.h>
> +#include <linux/of_device.h>
> +#include <linux/of_dma.h>
> +#include <linux/reset.h>
> +#include <linux/slab.h>
> +
> +#define STM32_DMAMUX_CCR(x)		(0x4 * (x))
> +#define STM32_DMAMUX_MAX_CHANNELS	32
> +#define STM32_DMAMUX_MAX_REQUESTS	255
> +
> +struct stm32_dmamux {
> +	u32 chan_id;
> +	u32 request;
> +	bool busy;
> +};
> +
> +struct stm32_dmamux_data {
> +	struct dma_router dmarouter;
> +	struct stm32_dmamux *muxes;
> +	struct clk *clk;
> +	void __iomem *iomem;
> +	u32 dmamux_requests; /* number of DMA requests connected to DMAMUX */
> +	u32 dmamux_channels; /* Number of DMA channels supported */
> +};
> +
> +static inline u32 stm32_dmamux_read(void __iomem *iomem, u32 reg)
> +{
> +	return readl_relaxed(iomem + reg);
> +}
> +
> +static inline void stm32_dmamux_write(void __iomem *iomem, u32 reg, u32 val)
> +{
> +	writel_relaxed(val, iomem + reg);
> +}
> +
> +static void stm32_dmamux_free(struct device *dev, void *route_data)
> +{
> +	struct stm32_dmamux_data *dmamux = dev_get_drvdata(dev);
> +	struct stm32_dmamux *mux = route_data;
> +
> +	/* Clear dma request for the right channel */
> +	stm32_dmamux_write(dmamux->iomem, STM32_DMAMUX_CCR(mux->chan_id), 0);
> +	clk_disable(dmamux->clk);
> +	mux->busy = false;
> +
> +	dev_dbg(dev, "Unmapping dma-router%dchan%d (was routed to request%d)\n",
> +		dev->id, mux->chan_id, mux->request);
> +}
> +
> +static void *stm32_dmamux_route_allocate(struct of_phandle_args *dma_spec,
> +					 struct of_dma *ofdma)
> +{
> +	struct platform_device *pdev = of_find_device_by_node(ofdma->of_node);
> +	struct stm32_dmamux_data *dmamux = platform_get_drvdata(pdev);
> +	struct stm32_dmamux *mux;
> +	u32 chan_id;
> +	int ret;
> +
> +	if (dma_spec->args_count != 4) {
> +		dev_err(&pdev->dev, "invalid number of dma mux args\n");
> +		return ERR_PTR(-EINVAL);
> +	}
> +
> +	if (dma_spec->args[0] >= dmamux->dmamux_channels) {
> +		dev_err(&pdev->dev, "invalid channel id: %d\n",
> +			dma_spec->args[0]);
> +		return ERR_PTR(-EINVAL);
> +	}
> +
> +	if (dma_spec->args[1] > dmamux->dmamux_requests) {
> +		dev_err(&pdev->dev, "invalid mux request number: %d\n",
> +			dma_spec->args[1]);
> +		return ERR_PTR(-EINVAL);
> +	}
> +
> +	/* The of_node_put() will be done in of_dma_router_xlate function */
> +	dma_spec->np = of_parse_phandle(ofdma->of_node, "dma-masters", 0);
> +	if (!dma_spec->np) {
> +		dev_err(&pdev->dev, "can't get dma master\n");
> +		return ERR_PTR(-EINVAL);
> +	}
> +
> +	chan_id = dma_spec->args[0];
> +	mux = &dmamux->muxes[chan_id];
> +	mux->chan_id = chan_id;
> +	mux->request = dma_spec->args[1];
> +
> +	if (mux->busy) {
> +		dev_err(&pdev->dev, "dma channel %d busy with request %d\n",
> +			chan_id, mux->request);
> +		return ERR_PTR(-EBUSY);
> +	}
> +
> +	ret = clk_enable(dmamux->clk);
> +	if (ret < 0) {
> +		dev_err(&pdev->dev, "clk_enable failed: %d\n", ret);
> +		return ERR_PTR(ret);
> +	}
> +
> +	stm32_dmamux_write(dmamux->iomem, STM32_DMAMUX_CCR(mux->chan_id),
> +			   mux->request);
> +	mux->busy = true;
> +
> +	dev_dbg(&pdev->dev, "Mapping dma-router%dchan%d to request%d\n",
> +		pdev->dev.id, mux->chan_id, mux->request);
> +
> +	return mux;
> +}
> +
> +static int stm32_dmamux_probe(struct platform_device *pdev)
> +{
> +	struct device_node *dma_node, *node = pdev->dev.of_node;
> +	struct stm32_dmamux_data *dmamux;
> +	struct reset_control *rst;
> +	struct resource *res;
> +	int ret;
> +
> +	if (!node)
> +		return -ENODEV;
> +
> +	dmamux = devm_kzalloc(&pdev->dev, sizeof(struct stm32_dmamux_data),
> +			      GFP_KERNEL);
> +	if (!dmamux)
> +		return -ENOMEM;
> +
> +	dma_node = of_parse_phandle(node, "dma-masters", 0);
> +	if (!dma_node) {
> +		dev_err(&pdev->dev, "Can't get DMA master node\n");
> +		return -ENODEV;
> +	}
> +	of_node_put(dma_node);
> +
> +	ret = of_property_read_u32(node, "dma-channels",
> +				   &dmamux->dmamux_channels);

can we have property_xxx calls alone, that way driver is not strictly
dependent on of

> +	if (ret)
> +		dmamux->dmamux_channels = STM32_DMAMUX_MAX_CHANNELS;
> +
> +	ret = of_property_read_u32(node, "dma-requests",
> +				   &dmamux->dmamux_requests);
> +	if (ret)
> +		dmamux->dmamux_requests = STM32_DMAMUX_MAX_REQUESTS;
> +
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	if (!res)
> +		return -ENODEV;
> +
> +	dmamux->iomem = devm_ioremap_resource(&pdev->dev, res);
> +	if (!dmamux->iomem)
> +		return -ENOMEM;
> +
> +	dmamux->clk = devm_clk_get(&pdev->dev, NULL);
> +	if (IS_ERR(dmamux->clk)) {
> +		dev_err(&pdev->dev, "Missing controller clock\n");
> +		return PTR_ERR(dmamux->clk);
> +	}
> +	ret = clk_prepare(dmamux->clk);
> +	if (ret < 0) {
> +		dev_err(&pdev->dev, "clk_prep failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	dmamux->muxes = devm_kcalloc(&pdev->dev, dmamux->dmamux_channels,
> +				     sizeof(struct stm32_dmamux),
> +				     GFP_KERNEL);
> +	if (!dmamux->muxes)
> +		return -ENOMEM;
> +
> +	rst = devm_reset_control_get(&pdev->dev, NULL);
> +	if (!IS_ERR(rst)) {
> +		ret = clk_enable(dmamux->clk);
> +		if (ret < 0) {
> +			dev_err(&pdev->dev, "clk_enable failed: %d\n", ret);
> +			return ret;
> +		}
> +		reset_control_assert(rst);
> +		udelay(2);
> +		reset_control_deassert(rst);
> +		clk_disable(dmamux->clk);
> +	}
> +
> +	dmamux->dmarouter.dev = &pdev->dev;
> +	dmamux->dmarouter.route_free = stm32_dmamux_free;
> +	platform_set_drvdata(pdev, dmamux);
> +
> +	ret = of_dma_router_register(node, stm32_dmamux_route_allocate,
> +				     &dmamux->dmarouter);
> +	if (ret < 0) {
> +		dev_err(&pdev->dev,
> +			"STM32 DMAMUX DMA OF registration failed %d\n", ret);
> +		return ret;
> +	}
> +
> +	dev_info(&pdev->dev, "STM32 DMAMUX driver registered\n");
> +
> +	return 0;
> +}
> +
> +static const struct of_device_id stm32_dmamux_match[] = {
> +	{ .compatible = "st,stm32-dmamux" },
> +	{},
> +};
> +
> +static struct platform_driver stm32_dmamux_driver = {
> +	.probe	= stm32_dmamux_probe,
> +	.driver = {
> +		.name = "stm32-dmamux",
> +		.of_match_table = stm32_dmamux_match,
> +	},
> +};
> +
> +static int __init stm32_dmamux_init(void)
> +{
> +	return platform_driver_register(&stm32_dmamux_driver);
> +}
> +arch_initcall(stm32_dmamux_init);

why not module init, wouldnt defer probe solve the dependencies

-- 
~Vinod

Re: [PATCH 1/5] dt-bindings: Document the STM32 DMAMUX bindings

[Pierre Yves MORDRET]

Hi Rob

On 03/20/2017 10:34 PM, Rob Herring wrote:
> On Mon, Mar 13, 2017 at 03:15:57PM +0100, M'boumba Cedric Madianga wrote:
>> This patch adds the documentation of device tree bindings for the STM32
>> DMAMUX.
>>
>> Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
>> ---
>>  .../devicetree/bindings/dma/stm32-dmamux.txt       | 57 ++++++++++++++++++++++
>>  1 file changed, 57 insertions(+)
>>  create mode 100644 Documentation/devicetree/bindings/dma/stm32-dmamux.txt
>>
>> diff --git a/Documentation/devicetree/bindings/dma/stm32-dmamux.txt b/Documentation/devicetree/bindings/dma/stm32-dmamux.txt
>> new file mode 100644
>> index 0000000..1039420
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/dma/stm32-dmamux.txt
>> @@ -0,0 +1,57 @@
>> +STM32 DMA MUX (DMA request router)
>> +
>> +Required properties:
>> +- compatible:	"st,stm32-dmamux"
>> +- reg:		Memory map for accessing module
>> +- #dma-cells:	Should be set to <4>.
>> +		For more details about the four cells, please see stm32-dma.txt
>> +		documentation binding file
>> +- dma-masters:	Phandle pointing to the DMA controller
>> +- clocks: Input clock of the DMAMUX instance.
>> +
>> +Optional properties:
>> +- dma-channels : Number of DMA channels supported.
>> +- dma-requests : Number of DMA requests supported.
>> +- resets: Reference to a reset controller asserting the DMA controller
>> +
>> +Example:
>> +
>> +/* DMA controller */
>> +dma2: dma-controller@40026400 {
>> +	compatible = "st,stm32-dma";
>> +	reg = <0x40026400 0x400>;
>> +	interrupts = <56>,
>> +		     <57>,
>> +		     <58>,
>> +		     <59>,
>> +		     <60>,
>> +		     <68>,
>> +		     <69>,
>> +		     <70>;
>> +	clocks = <&clk_hclk>;
>> +	#dma-cells = <4>;
>> +	st,mem2mem;
>> +	resets = <&rcc 150>;
>> +	st,dmamux;
>> +	dma-channels = <8>;
>> +};
>> +
>> +/* DMA mux */
>> +dmamux2: dma-router@40020820 {
>> +	compatible = "st,stm32-dmamux";
>> +	reg = <0x40020800 0x1c>;
>> +	#dma-cells = <4>;
>> +	dma-requests = <128>;
>> +	dma-masters = <&dma2>;
>
> I think this should be modeled after the interrupt-map property (or
> Stephen Boyd's gpio-map support which additionally allows pass thru of
> cell values). Something like this:
>
> dma-map = <0 41 0 0 &dma2 0 <req> 0 0>,
> 	<1 42 0 0 &dma2 1 <req> 0 0>;
> dma-map-mask = <0xffffffff 0xffffffff 0 0>;
>
> <req> is the request number on dma2 controller.
>
> This is more generic and would work if you have a single mux with
> multiple DMA controllers.
>

Would you mind to detail a little bit more your thoughts please ?
Are dma-map & dma-map-mask part of an existing bindings ? or need to be 
developed ?
I'm a little bit confused where come from values you used.

>> +};
>> +
>> +/* DMA client */
>> +usart1: serial@40011000 {
>> +	compatible = "st,stm32-usart", "st,stm32-uart";
>> +	reg = <0x40011000 0x400>;
>> +	interrupts = <37>;
>> +	clocks = <&clk_pclk2>;
>> +	dmas = <&dmamux2 0 41 0x400 0x00>,
>> +	       <&dmamux2 1 42 0x400 0x00>;
>> +	dma-names = "rx", "tx";
>> +};
>> --
>> 1.9.1
>>
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>
>
>

Re: [PATCH 2/5] dmaengine: Add STM32 DMAMUX driver

[Pierre Yves MORDRET]

On 04/06/2017 08:40 AM, Vinod Koul wrote:
> On Mon, Mar 13, 2017 at 03:15:58PM +0100, M'boumba Cedric Madianga wrote:
>> This patch implements the STM32 DMAMUX driver
>
> Can you describe the controller here pls

OK. Will be done in V2

>
>>
>> Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
>> ---
>>  drivers/dma/Kconfig        |   9 ++
>>  drivers/dma/Makefile       |   1 +
>>  drivers/dma/stm32-dmamux.c | 231 +++++++++++++++++++++++++++++++++++++++++++++
>>  3 files changed, 241 insertions(+)
>>  create mode 100644 drivers/dma/stm32-dmamux.c
>>
>> diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
>> index fc3435c..6ab80c9 100644
>> --- a/drivers/dma/Kconfig
>> +++ b/drivers/dma/Kconfig
>> @@ -461,6 +461,15 @@ config STM32_DMA
>>  	  If you have a board based on such a MCU and wish to use DMA say Y
>>  	  here.
>>
>> +config STM32_DMAMUX
>> +	bool "STMicroelectronics STM32 dma multiplexer support"
>> +	depends on STM32_DMA
>
> can you add compile test here so that we can get better compile coverage

OK. Will be done in V2

>
>> +	help
>> +	  Enable support for the on-chip DMA multiplexer on STMicroelectronics
>> +	  STM32 MCUs.
>> +	  If you have a board based on such a MCU and wish to use DMAMUX say Y
>> +	  here.
>> +
>>  config S3C24XX_DMAC
>>  	bool "Samsung S3C24XX DMA support"
>>  	depends on ARCH_S3C24XX || COMPILE_TEST
>> diff --git a/drivers/dma/Makefile b/drivers/dma/Makefile
>> index 0b723e9..449c7a3 100644
>> --- a/drivers/dma/Makefile
>> +++ b/drivers/dma/Makefile
>> @@ -57,6 +57,7 @@ obj-$(CONFIG_RENESAS_DMA) += sh/
>>  obj-$(CONFIG_SIRF_DMA) += sirf-dma.o
>>  obj-$(CONFIG_STE_DMA40) += ste_dma40.o ste_dma40_ll.o
>>  obj-$(CONFIG_STM32_DMA) += stm32-dma.o
>> +obj-$(CONFIG_STM32_DMAMUX) += stm32-dmamux.o
>>  obj-$(CONFIG_S3C24XX_DMAC) += s3c24xx-dma.o
>>  obj-$(CONFIG_TXX9_DMAC) += txx9dmac.o
>>  obj-$(CONFIG_TEGRA20_APB_DMA) += tegra20-apb-dma.o
>> diff --git a/drivers/dma/stm32-dmamux.c b/drivers/dma/stm32-dmamux.c
>> new file mode 100644
>> index 0000000..3003546
>> --- /dev/null
>> +++ b/drivers/dma/stm32-dmamux.c
>> @@ -0,0 +1,231 @@
>> +/*
>> + * DMA Router driver for STM32 DMA MUX
>> + *
>> + * Copyright (C) 2015 M'Boumba Cedric Madianga <cedric.madianga@gmail.com>
>
> we are in '17 now :)

OK. Will be done in V2

>
>> + *
>> + * Based on LPC18xx/43xx DMA MUX and TI DMA XBAR
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + */
>> +
>> +#include <linux/clk.h>
>> +#include <linux/delay.h>
>> +#include <linux/err.h>
>> +#include <linux/init.h>
>> +#include <linux/of_device.h>
>> +#include <linux/of_dma.h>
>> +#include <linux/reset.h>
>> +#include <linux/slab.h>
>> +
>> +#define STM32_DMAMUX_CCR(x)		(0x4 * (x))
>> +#define STM32_DMAMUX_MAX_CHANNELS	32
>> +#define STM32_DMAMUX_MAX_REQUESTS	255
>> +
>> +struct stm32_dmamux {
>> +	u32 chan_id;
>> +	u32 request;
>> +	bool busy;
>> +};
>> +
>> +struct stm32_dmamux_data {
>> +	struct dma_router dmarouter;
>> +	struct stm32_dmamux *muxes;
>> +	struct clk *clk;
>> +	void __iomem *iomem;
>> +	u32 dmamux_requests; /* number of DMA requests connected to DMAMUX */
>> +	u32 dmamux_channels; /* Number of DMA channels supported */
>> +};
>> +
>> +static inline u32 stm32_dmamux_read(void __iomem *iomem, u32 reg)
>> +{
>> +	return readl_relaxed(iomem + reg);
>> +}
>> +
>> +static inline void stm32_dmamux_write(void __iomem *iomem, u32 reg, u32 val)
>> +{
>> +	writel_relaxed(val, iomem + reg);
>> +}
>> +
>> +static void stm32_dmamux_free(struct device *dev, void *route_data)
>> +{
>> +	struct stm32_dmamux_data *dmamux = dev_get_drvdata(dev);
>> +	struct stm32_dmamux *mux = route_data;
>> +
>> +	/* Clear dma request for the right channel */
>> +	stm32_dmamux_write(dmamux->iomem, STM32_DMAMUX_CCR(mux->chan_id), 0);
>> +	clk_disable(dmamux->clk);
>> +	mux->busy = false;
>> +
>> +	dev_dbg(dev, "Unmapping dma-router%dchan%d (was routed to request%d)\n",
>> +		dev->id, mux->chan_id, mux->request);
>> +}
>> +
>> +static void *stm32_dmamux_route_allocate(struct of_phandle_args *dma_spec,
>> +					 struct of_dma *ofdma)
>> +{
>> +	struct platform_device *pdev = of_find_device_by_node(ofdma->of_node);
>> +	struct stm32_dmamux_data *dmamux = platform_get_drvdata(pdev);
>> +	struct stm32_dmamux *mux;
>> +	u32 chan_id;
>> +	int ret;
>> +
>> +	if (dma_spec->args_count != 4) {
>> +		dev_err(&pdev->dev, "invalid number of dma mux args\n");
>> +		return ERR_PTR(-EINVAL);
>> +	}
>> +
>> +	if (dma_spec->args[0] >= dmamux->dmamux_channels) {
>> +		dev_err(&pdev->dev, "invalid channel id: %d\n",
>> +			dma_spec->args[0]);
>> +		return ERR_PTR(-EINVAL);
>> +	}
>> +
>> +	if (dma_spec->args[1] > dmamux->dmamux_requests) {
>> +		dev_err(&pdev->dev, "invalid mux request number: %d\n",
>> +			dma_spec->args[1]);
>> +		return ERR_PTR(-EINVAL);
>> +	}
>> +
>> +	/* The of_node_put() will be done in of_dma_router_xlate function */
>> +	dma_spec->np = of_parse_phandle(ofdma->of_node, "dma-masters", 0);
>> +	if (!dma_spec->np) {
>> +		dev_err(&pdev->dev, "can't get dma master\n");
>> +		return ERR_PTR(-EINVAL);
>> +	}
>> +
>> +	chan_id = dma_spec->args[0];
>> +	mux = &dmamux->muxes[chan_id];
>> +	mux->chan_id = chan_id;
>> +	mux->request = dma_spec->args[1];
>> +
>> +	if (mux->busy) {
>> +		dev_err(&pdev->dev, "dma channel %d busy with request %d\n",
>> +			chan_id, mux->request);
>> +		return ERR_PTR(-EBUSY);
>> +	}
>> +
>> +	ret = clk_enable(dmamux->clk);
>> +	if (ret < 0) {
>> +		dev_err(&pdev->dev, "clk_enable failed: %d\n", ret);
>> +		return ERR_PTR(ret);
>> +	}
>> +
>> +	stm32_dmamux_write(dmamux->iomem, STM32_DMAMUX_CCR(mux->chan_id),
>> +			   mux->request);
>> +	mux->busy = true;
>> +
>> +	dev_dbg(&pdev->dev, "Mapping dma-router%dchan%d to request%d\n",
>> +		pdev->dev.id, mux->chan_id, mux->request);
>> +
>> +	return mux;
>> +}
>> +
>> +static int stm32_dmamux_probe(struct platform_device *pdev)
>> +{
>> +	struct device_node *dma_node, *node = pdev->dev.of_node;
>> +	struct stm32_dmamux_data *dmamux;
>> +	struct reset_control *rst;
>> +	struct resource *res;
>> +	int ret;
>> +
>> +	if (!node)
>> +		return -ENODEV;
>> +
>> +	dmamux = devm_kzalloc(&pdev->dev, sizeof(struct stm32_dmamux_data),
>> +			      GFP_KERNEL);
>> +	if (!dmamux)
>> +		return -ENOMEM;
>> +
>> +	dma_node = of_parse_phandle(node, "dma-masters", 0);
>> +	if (!dma_node) {
>> +		dev_err(&pdev->dev, "Can't get DMA master node\n");
>> +		return -ENODEV;
>> +	}
>> +	of_node_put(dma_node);
>> +
>> +	ret = of_property_read_u32(node, "dma-channels",
>> +				   &dmamux->dmamux_channels);
>
> can we have property_xxx calls alone, that way driver is not strictly
> dependent on of

Can you please explain what you are asking for ? Not sure to understand 
correctly.

>
>> +	if (ret)
>> +		dmamux->dmamux_channels = STM32_DMAMUX_MAX_CHANNELS;
>> +
>> +	ret = of_property_read_u32(node, "dma-requests",
>> +				   &dmamux->dmamux_requests);
>> +	if (ret)
>> +		dmamux->dmamux_requests = STM32_DMAMUX_MAX_REQUESTS;
>> +
>> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> +	if (!res)
>> +		return -ENODEV;
>> +
>> +	dmamux->iomem = devm_ioremap_resource(&pdev->dev, res);
>> +	if (!dmamux->iomem)
>> +		return -ENOMEM;
>> +
>> +	dmamux->clk = devm_clk_get(&pdev->dev, NULL);
>> +	if (IS_ERR(dmamux->clk)) {
>> +		dev_err(&pdev->dev, "Missing controller clock\n");
>> +		return PTR_ERR(dmamux->clk);
>> +	}
>> +	ret = clk_prepare(dmamux->clk);
>> +	if (ret < 0) {
>> +		dev_err(&pdev->dev, "clk_prep failed: %d\n", ret);
>> +		return ret;
>> +	}
>> +
>> +	dmamux->muxes = devm_kcalloc(&pdev->dev, dmamux->dmamux_channels,
>> +				     sizeof(struct stm32_dmamux),
>> +				     GFP_KERNEL);
>> +	if (!dmamux->muxes)
>> +		return -ENOMEM;
>> +
>> +	rst = devm_reset_control_get(&pdev->dev, NULL);
>> +	if (!IS_ERR(rst)) {
>> +		ret = clk_enable(dmamux->clk);
>> +		if (ret < 0) {
>> +			dev_err(&pdev->dev, "clk_enable failed: %d\n", ret);
>> +			return ret;
>> +		}
>> +		reset_control_assert(rst);
>> +		udelay(2);
>> +		reset_control_deassert(rst);
>> +		clk_disable(dmamux->clk);
>> +	}
>> +
>> +	dmamux->dmarouter.dev = &pdev->dev;
>> +	dmamux->dmarouter.route_free = stm32_dmamux_free;
>> +	platform_set_drvdata(pdev, dmamux);
>> +
>> +	ret = of_dma_router_register(node, stm32_dmamux_route_allocate,
>> +				     &dmamux->dmarouter);
>> +	if (ret < 0) {
>> +		dev_err(&pdev->dev,
>> +			"STM32 DMAMUX DMA OF registration failed %d\n", ret);
>> +		return ret;
>> +	}
>> +
>> +	dev_info(&pdev->dev, "STM32 DMAMUX driver registered\n");
>> +
>> +	return 0;
>> +}
>> +
>> +static const struct of_device_id stm32_dmamux_match[] = {
>> +	{ .compatible = "st,stm32-dmamux" },
>> +	{},
>> +};
>> +
>> +static struct platform_driver stm32_dmamux_driver = {
>> +	.probe	= stm32_dmamux_probe,
>> +	.driver = {
>> +		.name = "stm32-dmamux",
>> +		.of_match_table = stm32_dmamux_match,
>> +	},
>> +};
>> +
>> +static int __init stm32_dmamux_init(void)
>> +{
>> +	return platform_driver_register(&stm32_dmamux_driver);
>> +}
>> +arch_initcall(stm32_dmamux_init);
>
> why not module init, wouldnt defer probe solve the dependencies
>

The reason behind many devices (device_initcall level) rely on DMAs. If 
init is deferred DMAMUX driver will be probed twice if dependents rely 
on it. This sounds not a good call. This explains arch_initcall level.

Re: [PATCH 3/5] dt-bindings: stm32-dma: Add property to handle STM32 DMAMUX

[Pierre Yves MORDRET]

On 03/20/2017 10:37 PM, Rob Herring wrote:
> On Mon, Mar 13, 2017 at 03:15:59PM +0100, M'boumba Cedric Madianga wrote:
>> This patch adds an optional property needed for STM32 DMA controller
>> addressed via STM32 DMAMUX.
>>
>> Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
>> ---
>>  Documentation/devicetree/bindings/dma/stm32-dma.txt | 3 +++
>>  1 file changed, 3 insertions(+)
>>
>> diff --git a/Documentation/devicetree/bindings/dma/stm32-dma.txt b/Documentation/devicetree/bindings/dma/stm32-dma.txt
>> index 4408af6..7b5e91a 100644
>> --- a/Documentation/devicetree/bindings/dma/stm32-dma.txt
>> +++ b/Documentation/devicetree/bindings/dma/stm32-dma.txt
>> @@ -16,6 +16,9 @@ Optional properties:
>>  - resets: Reference to a reset controller asserting the DMA controller
>>  - st,mem2mem: boolean; if defined, it indicates that the controller supports
>>    memory-to-memory transfer
>> +- st,dmamux: boolean; if defined, it indicates that the controller is behind a
>> +  DMA multiplexer. In that case, using dma instances doesn't work for DMA
>> +  clients. They have to use dma-router instances.
>
> This should not be needed for the same reason we don't need anything
> like this for chained interrupt controllers.
>
> Also, the compatible string should be specific enough to provide this
> information.
>
> Rob

I don't understand what you're meant here. Our DMAs can work with or 
without DMAMUX driver. Driver uses this binding to use it or not.

Py

>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>
>
>

Re: [PATCH 2/5] dmaengine: Add STM32 DMAMUX driver

[Vinod Koul]

On Wed, Apr 26, 2017 at 09:17:37AM +0000, Pierre Yves MORDRET wrote:
> >> +
> >> +	ret = of_property_read_u32(node, "dma-channels",
> >> +				   &dmamux->dmamux_channels);
> >
> > can we have property_xxx calls alone, that way driver is not strictly
> > dependent on of
> 
> Can you please explain what you are asking for ? Not sure to understand 
> correctly.

Use device_property_read_u32() which is a generic property API.


> >> +static int __init stm32_dmamux_init(void)
> >> +{
> >> +	return platform_driver_register(&stm32_dmamux_driver);
> >> +}
> >> +arch_initcall(stm32_dmamux_init);
> >
> > why not module init, wouldnt defer probe solve the dependencies
> >
> 
> The reason behind many devices (device_initcall level) rely on DMAs. If 
> init is deferred DMAMUX driver will be probed twice if dependents rely 
> on it. This sounds not a good call. This explains arch_initcall level.

Why not deferred probe was introduced to help with dependencies...

-- 
~Vinod

Re: [PATCH 2/5] dmaengine: Add STM32 DMAMUX driver

[Pierre Yves MORDRET]

On 05/01/2017 08:03 AM, Vinod Koul wrote:
> On Wed, Apr 26, 2017 at 09:17:37AM +0000, Pierre Yves MORDRET wrote:
>>>> +
>>>> +	ret = of_property_read_u32(node, "dma-channels",
>>>> +				   &dmamux->dmamux_channels);
>>>
>>> can we have property_xxx calls alone, that way driver is not strictly
>>> dependent on of
>>
>> Can you please explain what you are asking for ? Not sure to understand
>> correctly.
>
> Use device_property_read_u32() which is a generic property API.

OK Will be done in V2

>
>
>>>> +static int __init stm32_dmamux_init(void)
>>>> +{
>>>> +	return platform_driver_register(&stm32_dmamux_driver);
>>>> +}
>>>> +arch_initcall(stm32_dmamux_init);
>>>
>>> why not module init, wouldnt defer probe solve the dependencies
>>>
>>
>> The reason behind many devices (device_initcall level) rely on DMAs. If
>> init is deferred DMAMUX driver will be probed twice if dependents rely
>> on it. This sounds not a good call. This explains arch_initcall level.
>
> Why not deferred probe was introduced to help with dependencies...
>

Most DMAs devices are already in subsys_initcall to avoid multiple 
useless probe when clients are deferred. This DMAMUx is itself on top of 
DMAs.
DMAMUX ---> DMA#1 ---> Device Client #1
        |          |--> Device Client #2
        |               [...]
        |          |--> Device Client #N
        |    [...]
        |--> DMA#N ---> Device Client #1
                   |--> Device Client #2
                        [...]
                   |--> Device Client #N
arch_initcall might a good call in this case for DMAMUX.