On 16:09 Wed 14 Mar     , Dong Aisheng wrote:

On Wed, Mar 14, 2012 at 03:23:43PM +0800, Jean-Christophe PLAGNIOL-VILLARD wrote:

quoted

On 16:47 Tue 13 Mar     , Dong Aisheng wrote:

quoted

From: Dong Aisheng <redacted>

Signed-off-by: Dong Aisheng <redacted>

---
The patch is still using a private way for dma part binding
since the common dma binding is still under discussion.
http://www.spinics.net/lists/linux-omap/msg65528.html

Will update to use common dma binding when it hits mainline.
---
 .../devicetree/bindings/mmc/fsl-mxs-mmc.txt        |   23 ++++++
 drivers/mmc/host/mxs-mmc.c                         |   82 +++++++++++++++++++-
 2 files changed, 102 insertions(+), 3 deletions(-)

diff --git a/Documentation/devicetree/bindings/mmc/fsl-mxs-mmc.txt b/Documentation/devicetree/bindings/mmc/fsl-mxs-mmc.txt
new file mode 100644
index 0000000..adc1142
--- /dev/null
+++ b/Documentation/devicetree/bindings/mmc/fsl-mxs-mmc.txt

@@ -0,0 +1,23 @@
+* FREESCALE MXS MMC peripheral
+
+Required properties:
+- compatible : Should be "fsl,<chip>-mmc"
+- reg : Should contain registers location and length
+- interrupts : Should contain interrupt.
+  The format is <irq_err irq_dma>.
+- dma_channel: Should contain the dma channel it uses
+
+Optional properties:
+- wp-gpios : Specify GPIOs for write protection
+- slot-4bit: Specify 4 bit mode support
+- slot-8bit: Specify 8 bit and 4 bit mode support
+
+Examples:
+mmc1: ssp@80010000 {
+	compatible = "fsl,imx28-mmc";
+	reg = <0x80010000 2000>;
+	/* <irq_err irq_dma> */
+	interrupts = <96 82>;
+	dma_channel = <0>;
+	slot-8bit;
+};

diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c
index 382c835..6cf2d17 100644
--- a/drivers/mmc/host/mxs-mmc.c
+++ b/drivers/mmc/host/mxs-mmc.c

@@ -38,6 +38,10 @@
 #include <linux/gpio.h>
 #include <linux/regulator/consumer.h>
 #include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_gpio.h>
+#include <linux/slab.h>
 
 #include <mach/mxs.h>
 #include <mach/common.h>

@@ -673,17 +677,79 @@ static bool mxs_mmc_dma_filter(struct dma_chan *chan, void *param)
 	return true;
 }
 
+#ifdef CONFIG_OF
+static struct resource * __devinit mxs_mmc_get_of_dmares(
+				struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct resource *dmares;
+	int ret;
+
+	if (!np)
+		return NULL;
+
+	dmares = kzalloc(sizeof(*dmares), GFP_KERNEL);
+	dmares->flags = IORESOURCE_DMA;
+	ret = of_property_read_u32(np, "dma_channel", &dmares->start);
+	if (ret) {
+		dev_err(&pdev->dev, "unable to get dmares from dt\n");
+		return NULL;
+	}
+	dmares->end = dmares->start;
+
+	return dmares;
+}
+
+static int __devinit mxs_mmc_get_of_property(struct platform_device *pdev,
+				struct mxs_mmc_platform_data **ppdata)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct mxs_mmc_platform_data *pdata = *ppdata;
+
+	if (!np)
+		return -ENODEV;
+
+	pdata = kzalloc(sizeof(*pdata), GFP_KERNEL);
+
+	if (of_get_property(np, "slot-8bit", NULL))
+		pdata->flags |= SLOTF_8_BIT_CAPABLE;
+
+	if (of_get_property(np, "slot-4bit", NULL))
+		pdata->flags |= SLOTF_4_BIT_CAPABLE;

it will conflit if both binding are set use a number instead

Hmm, i did not see conflict, can you explain more?
The "slot-8bit" includes the support for 4bit.(see binding doc)
Even user define them two property in dt by mistake, it does not cause conflict.
See:
if (pdata) {
	if (pdata->flags & SLOTF_8_BIT_CAPABLE)
		mmc->caps |= MMC_CAP_4_BIT_DATA | MMC_CAP_8_BIT_DATA;
	if (pdata->flags & SLOTF_4_BIT_CAPABLE)
		mmc->caps |= MMC_CAP_4_BIT_DATA;
}

this is mmc specifc sohould have a generic binding

Best Regards,
J.