[PATCH v4 0/4] Add STM32 DMAMUX support

[PATCH v4 0/4] Add STM32 DMAMUX support

[Pierre-Yves MORDRET]

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.

---
 Version history:
    v4:
        * Add multi-master ability for STM32 DMAMUX
        * Get rid of st,dmamux property and custom API between STM32
          DMAMUX and DMA. Bindings has changed.
          DMAMUX will read DMA masters from Device Tree from now on.
          Merely one DMAMUX node is needed now.
        * Only STM32 DMA are allowed to be connected onto DMAMUX
        * channelID is computed locally within the driver and crafted in
          dma_psec to be passed toward DMA master.
          DMAMUX router sorts out which DMA master will serve the
          request automatically.
        * This version forbids the use of DMA in standalone and DMAMUX at
          the same time : all clients need to be connected either on DMA
          or DMAMUX ; no mix up
    v3:
        * change compatible to st,stm32h7-dmamux to be mode Soc specific
        * add verbosity in dma-cells
---

Pierre-Yves MORDRET (4):
  dt-bindings: Document the STM32 DMAMUX bindings
  dmaengine: Add STM32 DMAMUX driver
  dt-bindings: stm32-dma: add a property to handle STM32 DMAMUX
  ARM: configs: stm32: Add MDMA support in STM32 defconfig

 .../devicetree/bindings/dma/stm32-dma.txt          |   4 +-
 .../devicetree/bindings/dma/stm32-dmamux.txt       |  84 ++++++
 arch/arm/configs/stm32_defconfig                   |   1 +
 drivers/dma/Kconfig                                |   9 +
 drivers/dma/Makefile                               |   1 +
 drivers/dma/stm32-dmamux.c                         | 321 +++++++++++++++++++++
 6 files changed, 419 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/devicetree/bindings/dma/stm32-dmamux.txt
 create mode 100644 drivers/dma/stm32-dmamux.c

-- 
2.7.4

[PATCH v4 1/4] dt-bindings: Document the STM32 DMAMUX bindings

[Pierre-Yves MORDRET]

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

Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
---
 Version history:
    v4:
        * Add multi-master ability for STM32 DMAMUX
        * Get rid of st,dmamux properties
    v3:
        * Change compatible to st,stm32h7-dmamux to be mode Soc specific
        * Add verbosity in dma-cells
    v2:
        * Move clock bindings from optional to mandatory one
        * Drop channelID bindings as managed dynamically from now on by
          DMAMUX driver.
---
---
 .../devicetree/bindings/dma/stm32-dmamux.txt       | 84 ++++++++++++++++++++++
 1 file changed, 84 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..1b893b2
--- /dev/null
+++ b/Documentation/devicetree/bindings/dma/stm32-dmamux.txt
@@ -0,0 +1,84 @@
+STM32 DMA MUX (DMA request router)
+
+Required properties:
+- compatible:	"st,stm32h7-dmamux"
+- reg:		Memory map for accessing module
+- #dma-cells:	Should be set to <3>.
+		First parameter is request line number.
+		Second is DMA channel configuration
+		Third is Fifo threshold
+		For more details about the three cells, please see
+		stm32-dma.txt documentation binding file
+- dma-masters:	Phandle pointing to the DMA controllers.
+		Several controllers are allowed. Only "st,stm32-dma" DMA
+		compatible are supported.
+
+Optional properties:
+- dma-channels : Number of DMA requests supported.
+- dma-requests : Number of DMAMUX requests supported.
+- resets: Reference to a reset controller asserting the DMA controller
+- clocks: Input clock of the DMAMUX instance.
+
+Example:
+
+/* DMA controller 1 */
+dma1: dma-controller@40020000 {
+	compatible = "st,stm32-dma";
+	reg = <0x40020000 0x400>;
+	interrupts = <11>,
+		     <12>,
+		     <13>,
+		     <14>,
+		     <15>,
+		     <16>,
+		     <17>,
+		     <47>;
+	clocks = <&timer_clk>;
+	#dma-cells = <4>;
+	st,mem2mem;
+	resets = <&rcc 150>;
+	dma-channels = <8>;
+	dma-requests = <8>;
+};
+
+/* DMA controller 1 */
+dma2: dma@40020400 {
+	compatible = "st,stm32-dma";
+	reg = <0x40020400 0x400>;
+	interrupts = <56>,
+		     <57>,
+		     <58>,
+		     <59>,
+		     <60>,
+		     <68>,
+		     <69>,
+		     <70>;
+	clocks = <&timer_clk>;
+	#dma-cells = <4>;
+	st,mem2mem;
+	resets = <&rcc 150>;
+	dma-channels = <8>;
+	dma-requests = <8>;
+};
+
+/* DMA mux */
+dmamux1: dma-router@40020800 {
+	compatible = "st,stm32h7-dmamux";
+	reg = <0x40020800 0x3c>;
+	#dma-cells = <3>;
+	dma-requests = <128>;
+	dma-channels = <16>;
+	dma-masters = <&dma1 &dma2>;
+	clocks = <&timer_clk>;
+};
+
+/* DMA client */
+usart1: serial@40011000 {
+	compatible = "st,stm32-usart", "st,stm32-uart";
+	reg = <0x40011000 0x400>;
+	interrupts = <37>;
+	clocks = <&timer_clk>;
+	dmas = <&dmamux1 41 0x414 0>,
+	       <&dmamux1 42 0x414 0>;
+	dma-names = "rx", "tx";
+};
-- 
2.7.4

[PATCH v4 2/4] dmaengine: Add STM32 DMAMUX driver

[Pierre-Yves MORDRET]

This patch implements the STM32 DMAMUX driver.

The DMAMUX request multiplexer allows routing a DMA request line between
the peripherals and the DMA controllers of the product. The routing
function is ensured by a programmable multi-channel DMA request line
multiplexer. Each channel selects a unique DMA request line,
unconditionally or synchronously with events from its DMAMUX
synchronization inputs. The DMAMUX may also be used as a DMA request
generator from programmable events on its input trigger signals

Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
---
 Version history:
    v4:
        * Get rid of st,dmamux property and custom API between STM32
          DMAMUX and DMA.
          DMAMUX will read DMA masters from Device Tree from now on.
          Merely one DMAMUX node is needed now.
        * Only STM32 DMA are allowed to be connected onto DMAMUX
        * channelID is computed locally within the driver and crafted in
          dma_psec to be passed toward DMA master.
          DMAMUX router sorts out which DMA master will serve the
          request automatically.
        * This version forbids the use of DMA in standalone and DMAMUX at
          the same time : all clients need to be connected either on DMA
          or DMAMUX ; no mix up
    v3:
        * change compatible to st,stm32h7-dmamux to be mode SoC specific
    v2:
        * Dynamic channelID allocation.
        * Change of_property_... by device_property.
        * New clock management.
        * DMAMUX Configuration API.
---
---
 drivers/dma/Kconfig        |   9 ++
 drivers/dma/Makefile       |   1 +
 drivers/dma/stm32-dmamux.c | 321 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 331 insertions(+)
 create mode 100644 drivers/dma/stm32-dmamux.c

diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
index fadc4d8..04e381b 100644
--- a/drivers/dma/Kconfig
+++ b/drivers/dma/Kconfig
@@ -483,6 +483,15 @@ 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 || COMPILE_TEST
+	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 f08f8de..a145ad1 100644
--- a/drivers/dma/Makefile
+++ b/drivers/dma/Makefile
@@ -59,6 +59,7 @@ 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..d90da2ca
--- /dev/null
+++ b/drivers/dma/stm32-dmamux.c
@@ -0,0 +1,321 @@
+/*
+ *
+ * Copyright (C) STMicroelectronics SA 2017
+ * Author(s): M'boumba Cedric Madianga <cedric.madianga@gmail.com>
+ *            Pierre-Yves Mordret <pierre-yves.mordret@st.com>
+ *
+ * License terms: GPL V2.0.
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
+ * details.
+ *
+ * DMA Router driver for STM32 DMA MUX
+ *
+ * Based on TI DMA Crossbar driver
+ *
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/of_dma.h>
+#include <linux/reset.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+
+#define STM32_DMAMUX_CCR(x)		(0x4 * (x))
+#define STM32_DMAMUX_MAX_DMA_REQUESTS	32
+#define STM32_DMAMUX_MAX_REQUESTS	255
+
+struct stm32_dmamux {
+	u32 master;
+	u32 request;
+	u32 chan_id;
+};
+
+struct stm32_dmamux_data {
+	struct dma_router dmarouter;
+	struct clk *clk;
+	struct reset_control *rst;
+	void __iomem *iomem;
+	u32 dma_requests; /* Number of DMA requests connected to DMAMUX */
+	u32 dmamux_requests; /* Number of DMA requests routed toward DMAs */
+	spinlock_t lock; /* Protects register access */
+	unsigned long *dma_inuse; /* Used DMA channel */
+	u32 dma_reqs[]; /* Number of DMA Request per DMA masters.
+			 *  [0] holds number of DMA Masters.
+			 *  To be kept at very end end of this structure
+			 */
+};
+
+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;
+	unsigned long flags;
+
+	/* Clear dma request */
+	spin_lock_irqsave(&dmamux->lock, flags);
+
+	stm32_dmamux_write(dmamux->iomem, STM32_DMAMUX_CCR(mux->chan_id), 0);
+	clear_bit(mux->chan_id, dmamux->dma_inuse);
+
+	if (!IS_ERR(dmamux->clk))
+		clk_disable(dmamux->clk);
+
+	spin_unlock_irqrestore(&dmamux->lock, flags);
+
+	dev_dbg(dev, "Unmapping DMAMUX(%u) to DMA%u(%u)\n",
+		mux->request, mux->master, mux->chan_id);
+
+	kfree(mux);
+}
+
+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 i, min, max, ret;
+	unsigned long flags;
+
+	if (dma_spec->args_count != 3) {
+		dev_err(&pdev->dev, "invalid number of dma mux args\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (dma_spec->args[0] > dmamux->dmamux_requests) {
+		dev_err(&pdev->dev, "invalid mux request number: %d\n",
+			dma_spec->args[0]);
+		return ERR_PTR(-EINVAL);
+	}
+
+	mux = kzalloc(sizeof(*mux), GFP_KERNEL);
+	if (!mux)
+		return ERR_PTR(-ENOMEM);
+
+	spin_lock_irqsave(&dmamux->lock, flags);
+	mux->chan_id = find_first_zero_bit(dmamux->dma_inuse,
+					   dmamux->dma_requests);
+	spin_unlock_irqrestore(&dmamux->lock, flags);
+	if (mux->chan_id == dmamux->dma_requests) {
+		dev_err(&pdev->dev, "Run out of free DMA requests\n");
+		kfree(mux);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	/* Look for DMA Master */
+	for (i = 1, min = 0, max = dmamux->dma_reqs[i];
+	     i <= dmamux->dma_reqs[0];
+	     min += dmamux->dma_reqs[i], max += dmamux->dma_reqs[++i])
+		if (mux->chan_id < max)
+			break;
+	mux->master = i - 1;
+
+	/* The of_node_put() will be done in of_dma_router_xlate function */
+	dma_spec->np = of_parse_phandle(ofdma->of_node, "dma-masters", i - 1);
+	if (!dma_spec->np) {
+		dev_err(&pdev->dev, "can't get dma master\n");
+		kfree(mux);
+		return ERR_PTR(-EINVAL);
+	}
+
+	/* Set dma request */
+	spin_lock_irqsave(&dmamux->lock, flags);
+	if (!IS_ERR(dmamux->clk)) {
+		ret = clk_enable(dmamux->clk);
+		if (ret < 0) {
+			spin_unlock_irqrestore(&dmamux->lock, flags);
+			kfree(mux);
+			dev_err(&pdev->dev, "clk_prep_enable issue: %d\n", ret);
+			return ERR_PTR(ret);
+		}
+	}
+	spin_unlock_irqrestore(&dmamux->lock, flags);
+
+	set_bit(mux->chan_id, dmamux->dma_inuse);
+	mux->request = dma_spec->args[0];
+
+	/*  craft DMA spec */
+	dma_spec->args[3] = dma_spec->args[2];
+	dma_spec->args[2] = dma_spec->args[1];
+	dma_spec->args[1] = 0;
+	dma_spec->args[0] = mux->chan_id - min;
+	dma_spec->args_count = 4;
+
+	stm32_dmamux_write(dmamux->iomem, STM32_DMAMUX_CCR(mux->chan_id),
+			   mux->request);
+	dev_dbg(&pdev->dev, "Mapping DMAMUX(%u) to DMA%u(%u)\n",
+		mux->request, mux->master, mux->chan_id);
+
+	return mux;
+}
+
+static const struct of_device_id stm32_stm32dma_master_match[] = {
+	{ .compatible = "st,stm32-dma", },
+	{},
+};
+
+static int stm32_dmamux_probe(struct platform_device *pdev)
+{
+	struct device_node *node = pdev->dev.of_node;
+	const struct of_device_id *match;
+	struct device_node *dma_node;
+	struct stm32_dmamux_data *stm32_dmamux;
+	struct resource *res;
+	void __iomem *iomem;
+	int i, count, ret;
+	u32 dma_req;
+
+	if (!node)
+		return -ENODEV;
+
+	count = device_property_read_u32_array(&pdev->dev, "dma-masters",
+					       NULL, 0);
+	if (count < 0) {
+		dev_err(&pdev->dev, "Can't get DMA master(s) node\n");
+		return -ENODEV;
+	}
+
+	stm32_dmamux = devm_kzalloc(&pdev->dev, sizeof(*stm32_dmamux) +
+				    sizeof(u32) * (count + 1), GFP_KERNEL);
+	if (!stm32_dmamux)
+		return -ENOMEM;
+
+	dma_req = 0;
+	for (i = 1; i <= count; i++) {
+		dma_node = of_parse_phandle(node, "dma-masters", i - 1);
+
+		match = of_match_node(stm32_stm32dma_master_match, dma_node);
+		if (!match) {
+			dev_err(&pdev->dev, "DMA master is not supported\n");
+			of_node_put(dma_node);
+			return -EINVAL;
+		}
+
+		if (of_property_read_u32(dma_node, "dma-requests",
+					 &stm32_dmamux->dma_reqs[i])) {
+			dev_info(&pdev->dev,
+				 "Missing MUX output information, using %u.\n",
+				 STM32_DMAMUX_MAX_DMA_REQUESTS);
+			stm32_dmamux->dma_reqs[i] =
+				STM32_DMAMUX_MAX_DMA_REQUESTS;
+		}
+		dma_req += stm32_dmamux->dma_reqs[i];
+		of_node_put(dma_node);
+	}
+
+	if (dma_req > STM32_DMAMUX_MAX_DMA_REQUESTS) {
+		dev_err(&pdev->dev, "Too many DMA Master Requests to manage\n");
+		return -ENODEV;
+	}
+
+	stm32_dmamux->dma_requests = dma_req;
+	stm32_dmamux->dma_reqs[0] = count;
+	stm32_dmamux->dma_inuse = devm_kcalloc(&pdev->dev,
+					       BITS_TO_LONGS(dma_req),
+					       sizeof(unsigned long),
+					       GFP_KERNEL);
+	if (!stm32_dmamux->dma_inuse)
+		return -ENOMEM;
+
+	if (device_property_read_u32(&pdev->dev, "dma-requests",
+				     &stm32_dmamux->dmamux_requests)) {
+		stm32_dmamux->dmamux_requests = STM32_DMAMUX_MAX_REQUESTS;
+		dev_warn(&pdev->dev, "DMAMUX defaulting on %u requests\n",
+			 stm32_dmamux->dmamux_requests);
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -ENODEV;
+
+	iomem = devm_ioremap_resource(&pdev->dev, res);
+	if (!iomem)
+		return -ENOMEM;
+
+	spin_lock_init(&stm32_dmamux->lock);
+
+	stm32_dmamux->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(stm32_dmamux->clk)) {
+		ret = PTR_ERR(stm32_dmamux->clk);
+		if (ret == -EPROBE_DEFER)
+			dev_info(&pdev->dev, "Missing controller clock\n");
+		return ret;
+	}
+
+	stm32_dmamux->rst = devm_reset_control_get(&pdev->dev, NULL);
+	if (!IS_ERR(stm32_dmamux->rst)) {
+		reset_control_assert(stm32_dmamux->rst);
+		udelay(2);
+		reset_control_deassert(stm32_dmamux->rst);
+	}
+
+	stm32_dmamux->iomem = iomem;
+	stm32_dmamux->dmarouter.dev = &pdev->dev;
+	stm32_dmamux->dmarouter.route_free = stm32_dmamux_free;
+
+	platform_set_drvdata(pdev, stm32_dmamux);
+
+	if (!IS_ERR(stm32_dmamux->clk)) {
+		ret = clk_prepare_enable(stm32_dmamux->clk);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "clk_prep_enable error: %d\n", ret);
+			return ret;
+		}
+	}
+
+	/* Reset the dmamux */
+	for (i = 0; i < stm32_dmamux->dma_requests; i++)
+		stm32_dmamux_write(stm32_dmamux->iomem, STM32_DMAMUX_CCR(i), 0);
+
+	if (!IS_ERR(stm32_dmamux->clk))
+		clk_disable(stm32_dmamux->clk);
+
+	return of_dma_router_register(node, stm32_dmamux_route_allocate,
+				     &stm32_dmamux->dmarouter);
+}
+
+static const struct of_device_id stm32_dmamux_match[] = {
+	{ .compatible = "st,stm32h7-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);
+
+MODULE_DESCRIPTION("DMA Router driver for STM32 DMA MUX");
+MODULE_AUTHOR("M'boumba Cedric Madianga <cedric.madianga@gmail.com>");
+MODULE_AUTHOR("Pierre-Yves Mordret <pierre-yves.mordret@st.com>");
+MODULE_LICENSE("GPL v2");
-- 
2.7.4

[PATCH v4 3/4] dt-bindings: stm32-dma: add a property to handle STM32 DMAMUX

[Pierre-Yves MORDRET]

STM32 DMA controller has to exposed its number of request line to be
addressed via STM32 DMAMUX.

Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
---
 Version history:
    v4:
        * get rid of st,dmamux property
        * number of DMA requests is exposed for DMAMUX
    v3:
        * None
    v2:
        * Typo fix
---
---
 Documentation/devicetree/bindings/dma/stm32-dma.txt | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/Documentation/devicetree/bindings/dma/stm32-dma.txt b/Documentation/devicetree/bindings/dma/stm32-dma.txt
index 4408af6..77542e1 100644
--- a/Documentation/devicetree/bindings/dma/stm32-dma.txt
+++ b/Documentation/devicetree/bindings/dma/stm32-dma.txt
@@ -13,6 +13,7 @@ Required properties:
 - #dma-cells : Must be <4>. See DMA client paragraph for more details.
 
 Optional properties:
+- dma-requests : Number of DMA requests supported.
 - 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
@@ -34,12 +35,13 @@ Example:
 		#dma-cells = <4>;
 		st,mem2mem;
 		resets = <&rcc 150>;
+		dma-requests = <8>;
 	};
 
 * DMA client
 
 DMA clients connected to the STM32 DMA controller must use the format
-described in the dma.txt file, using a five-cell specifier for each
+described in the dma.txt file, using a four-cell specifier for each
 channel: a phandle to the DMA controller plus the following four integer cells:
 
 1. The channel id
-- 
2.7.4

[PATCH v4 4/4] ARM: configs: stm32: Add MDMA support in STM32 defconfig

[Pierre-Yves MORDRET]

This patch adds MDMA support in STM32 defconfig file

Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
---
 Version history:
    v4:
    v3:
        * None
    v2:
        * Add MDMA support in STM32 defconfig
---
---
 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 90e5c46..988f395 100644
--- a/arch/arm/configs/stm32_defconfig
+++ b/arch/arm/configs/stm32_defconfig
@@ -67,6 +67,7 @@ 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
-- 
2.7.4

Re: [PATCH v4 1/4] dt-bindings: Document the STM32 DMAMUX bindings

[Rob Herring]

On Thu, Sep 07, 2017 at 01:52:45PM +0200, Pierre-Yves MORDRET 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>
> Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
> ---
>  Version history:
>     v4:
>         * Add multi-master ability for STM32 DMAMUX
>         * Get rid of st,dmamux properties
>     v3:
>         * Change compatible to st,stm32h7-dmamux to be mode Soc specific
>         * Add verbosity in dma-cells
>     v2:
>         * Move clock bindings from optional to mandatory one
>         * Drop channelID bindings as managed dynamically from now on by
>           DMAMUX driver.
> ---
> ---
>  .../devicetree/bindings/dma/stm32-dmamux.txt       | 84 ++++++++++++++++++++++
>  1 file changed, 84 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/dma/stm32-dmamux.txt

Acked-by: Rob Herring <robh@kernel.org>

Re: [PATCH v4 3/4] dt-bindings: stm32-dma: add a property to handle STM32 DMAMUX

[Rob Herring]

On Thu, Sep 07, 2017 at 01:52:47PM +0200, Pierre-Yves MORDRET wrote:
> STM32 DMA controller has to exposed its number of request line to be
> addressed via STM32 DMAMUX.
> 
> Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
> Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
> ---
>  Version history:
>     v4:
>         * get rid of st,dmamux property
>         * number of DMA requests is exposed for DMAMUX
>     v3:
>         * None
>     v2:
>         * Typo fix
> ---
> ---
>  Documentation/devicetree/bindings/dma/stm32-dma.txt | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)

Acked-by: Rob Herring <robh@kernel.org>

Re: [PATCH v4 0/4] Add STM32 DMAMUX support

[Pierre Yves MORDRET]

Hello

Gentle ping for driver review since DT Bindings have been acked by Rob Herring.

Thanks
Py

On 09/07/2017 01:52 PM, Pierre-Yves MORDRET wrote:
> 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.
> 
> ---
>  Version history:
>     v4:
>         * Add multi-master ability for STM32 DMAMUX
>         * Get rid of st,dmamux property and custom API between STM32
>           DMAMUX and DMA. Bindings has changed.
>           DMAMUX will read DMA masters from Device Tree from now on.
>           Merely one DMAMUX node is needed now.
>         * Only STM32 DMA are allowed to be connected onto DMAMUX
>         * channelID is computed locally within the driver and crafted in
>           dma_psec to be passed toward DMA master.
>           DMAMUX router sorts out which DMA master will serve the
>           request automatically.
>         * This version forbids the use of DMA in standalone and DMAMUX at
>           the same time : all clients need to be connected either on DMA
>           or DMAMUX ; no mix up
>     v3:
>         * change compatible to st,stm32h7-dmamux to be mode Soc specific
>         * add verbosity in dma-cells
> ---
> 
> Pierre-Yves MORDRET (4):
>   dt-bindings: Document the STM32 DMAMUX bindings
>   dmaengine: Add STM32 DMAMUX driver
>   dt-bindings: stm32-dma: add a property to handle STM32 DMAMUX
>   ARM: configs: stm32: Add MDMA support in STM32 defconfig
> 
>  .../devicetree/bindings/dma/stm32-dma.txt          |   4 +-
>  .../devicetree/bindings/dma/stm32-dmamux.txt       |  84 ++++++
>  arch/arm/configs/stm32_defconfig                   |   1 +
>  drivers/dma/Kconfig                                |   9 +
>  drivers/dma/Makefile                               |   1 +
>  drivers/dma/stm32-dmamux.c                         | 321 +++++++++++++++++++++
>  6 files changed, 419 insertions(+), 1 deletion(-)
>  create mode 100644 Documentation/devicetree/bindings/dma/stm32-dmamux.txt
>  create mode 100644 drivers/dma/stm32-dmamux.c
> 

Re: [PATCH v4 2/4] dmaengine: Add STM32 DMAMUX driver

[Peter Ujfalusi]


Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

On 2017-09-07 14:52, Pierre-Yves MORDRET wrote:
> This patch implements the STM32 DMAMUX driver.
> 
> The DMAMUX request multiplexer allows routing a DMA request line between
> the peripherals and the DMA controllers of the product. The routing
> function is ensured by a programmable multi-channel DMA request line
> multiplexer. Each channel selects a unique DMA request line,
> unconditionally or synchronously with events from its DMAMUX
> synchronization inputs. The DMAMUX may also be used as a DMA request
> generator from programmable events on its input trigger signals
> 
> Signed-off-by: M'boumba Cedric Madianga <cedric.madianga@gmail.com>
> Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
> ---
>  Version history:
>     v4:
>         * Get rid of st,dmamux property and custom API between STM32
>           DMAMUX and DMA.
>           DMAMUX will read DMA masters from Device Tree from now on.
>           Merely one DMAMUX node is needed now.
>         * Only STM32 DMA are allowed to be connected onto DMAMUX
>         * channelID is computed locally within the driver and crafted in
>           dma_psec to be passed toward DMA master.
>           DMAMUX router sorts out which DMA master will serve the
>           request automatically.
>         * This version forbids the use of DMA in standalone and DMAMUX at
>           the same time : all clients need to be connected either on DMA
>           or DMAMUX ; no mix up

Great that you got it working w/o a custom API!
I have one comment, which actually valid for the ti-dma-crossbar driver
as well...

> +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 i, min, max, ret;
> +	unsigned long flags;
> +
> +	if (dma_spec->args_count != 3) {
> +		dev_err(&pdev->dev, "invalid number of dma mux args\n");
> +		return ERR_PTR(-EINVAL);
> +	}
> +
> +	if (dma_spec->args[0] > dmamux->dmamux_requests) {
> +		dev_err(&pdev->dev, "invalid mux request number: %d\n",
> +			dma_spec->args[0]);
> +		return ERR_PTR(-EINVAL);
> +	}
> +
> +	mux = kzalloc(sizeof(*mux), GFP_KERNEL);
> +	if (!mux)
> +		return ERR_PTR(-ENOMEM);
> +
> +	spin_lock_irqsave(&dmamux->lock, flags);
> +	mux->chan_id = find_first_zero_bit(dmamux->dma_inuse,
> +					   dmamux->dma_requests);

you pick the first available chan_id here under the lock.

> +	spin_unlock_irqrestore(&dmamux->lock, flags);
> +	if (mux->chan_id == dmamux->dma_requests) {
> +		dev_err(&pdev->dev, "Run out of free DMA requests\n");
> +		kfree(mux);
> +		return ERR_PTR(-ENOMEM);
> +	}
> +
> +	/* Look for DMA Master */
> +	for (i = 1, min = 0, max = dmamux->dma_reqs[i];
> +	     i <= dmamux->dma_reqs[0];
> +	     min += dmamux->dma_reqs[i], max += dmamux->dma_reqs[++i])
> +		if (mux->chan_id < max)
> +			break;
> +	mux->master = i - 1;
> +
> +	/* The of_node_put() will be done in of_dma_router_xlate function */
> +	dma_spec->np = of_parse_phandle(ofdma->of_node, "dma-masters", i - 1);
> +	if (!dma_spec->np) {
> +		dev_err(&pdev->dev, "can't get dma master\n");
> +		kfree(mux);
> +		return ERR_PTR(-EINVAL);
> +	}
> +
> +	/* Set dma request */
> +	spin_lock_irqsave(&dmamux->lock, flags);
> +	if (!IS_ERR(dmamux->clk)) {
> +		ret = clk_enable(dmamux->clk);
> +		if (ret < 0) {
> +			spin_unlock_irqrestore(&dmamux->lock, flags);
> +			kfree(mux);
> +			dev_err(&pdev->dev, "clk_prep_enable issue: %d\n", ret);
> +			return ERR_PTR(ret);
> +		}
> +	}
> +	spin_unlock_irqrestore(&dmamux->lock, flags);
> +
> +	set_bit(mux->chan_id, dmamux->dma_inuse);

But nothing stops other parallel threads to pick the same chan_id since
you have released the lock (released, got the lock to protect the set
dma request and released it again). imho the find_first_zero_bit() and
the set_bit() should be done within the same lock to avoid race conditions.

- Péter

Re: [PATCH v4 2/4] dmaengine: Add STM32 DMAMUX driver

[Pierre Yves MORDRET]

On 09/21/2017 01:25 PM, Peter Ujfalusi wrote:
> 
> Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
> 
> 
> Great that you got it working w/o a custom API!
> I have one comment, which actually valid for the ti-dma-crossbar driver
> as well...

Yes. That cleans up a little bit the sw architecture. But still this custom API
allowed both DMAMUX and DMA at the same time since using the same channel ID
allocator.
Ok this is another story to be addressed out of this thread ;)

> 
>> +static void *stm32_dmamux_route_allocate(struct of_phandle_args *dma_spec,
>> +					 struct of_dma *ofdma)
>> +
>> +	spin_lock_irqsave(&dmamux->lock, flags);
>> +	mux->chan_id = find_first_zero_bit(dmamux->dma_inuse,
>> +					   dmamux->dma_requests);
> 
> you pick the first available chan_id here under the lock.
> 
>> +	spin_unlock_irqrestore(&dmamux->lock, flags);
>> +	if (mux->chan_id == dmamux->dma_requests) {
>> ...
>> +	/* Set dma request */
>> +	spin_lock_irqsave(&dmamux->lock, flags);
>> +	if (!IS_ERR(dmamux->clk)) {
>> ...
>> +	spin_unlock_irqrestore(&dmamux->lock, flags);
>> +
>> +	set_bit(mux->chan_id, dmamux->dma_inuse);
> 
> But nothing stops other parallel threads to pick the same chan_id since
> you have released the lock (released, got the lock to protect the set
> dma request and released it again). imho the find_first_zero_bit() and
> the set_bit() should be done within the same lock to avoid race conditions.
> 
> - Péter
> 

Yep good catch : That's correct. Even if probability to happen is rather low, it
may happen.
Will solve that.

Py