Re: [PATCHv2] mtd: nand: atmel_nand: retrieve NFC clock

[Josh Wu <josh.wu@atmel.com>]

Hi, Alexandre

On 9/12/2014 1:52 AM, Alexandre Belloni wrote:
> From: Boris BREZILLON <boris.brezillon@free-electrons.com>
>
> Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
> Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
> ---
>
> Changes in v2:
>   - reworked the error path to really make the clock optional
>   - Documented the new optional property
>
>   .../devicetree/bindings/mtd/atmel-nand.txt         |  1 +
>   drivers/mtd/nand/atmel_nand.c                      | 25 ++++++++++++++++++++++
>   2 files changed, 26 insertions(+)
>
> diff --git a/Documentation/devicetree/bindings/mtd/atmel-nand.txt b/Documentation/devicetree/bindings/mtd/atmel-nand.txt
> index c4728839d0c1..f71e2ebab15b 100644
> --- a/Documentation/devicetree/bindings/mtd/atmel-nand.txt
> +++ b/Documentation/devicetree/bindings/mtd/atmel-nand.txt
> @@ -38,6 +38,7 @@ Optional properties:
>               if don't want to use it.
>     - Optional properties:
>       - atmel,write-by-sram: boolean to enable NFC write by sram.
> +    - clocks: phandle to the peripheral clock if it exists
>   
>   Examples:
>   nand0: nand@40000000,0 {
> diff --git a/drivers/mtd/nand/atmel_nand.c b/drivers/mtd/nand/atmel_nand.c
> index 9c5f717bda54..69e0eb1ace54 100644
> --- a/drivers/mtd/nand/atmel_nand.c
> +++ b/drivers/mtd/nand/atmel_nand.c
> @@ -27,6 +27,7 @@
>    *
>    */
>   
> +#include <linux/clk.h>
>   #include <linux/dma-mapping.h>
>   #include <linux/slab.h>
>   #include <linux/module.h>
> @@ -96,6 +97,8 @@ struct atmel_nfc {
>   	bool			use_nfc_sram;
>   	bool			write_by_sram;
>   
> +	struct clk		*clk;
> +
>   	bool			is_initialized;
>   	struct completion	comp_ready;
>   	struct completion	comp_cmd_done;
> @@ -2248,6 +2251,7 @@ static int atmel_nand_nfc_probe(struct platform_device *pdev)
>   {
>   	struct atmel_nfc *nfc = &nand_nfc;
>   	struct resource *nfc_cmd_regs, *nfc_hsmc_regs, *nfc_sram;
> +	int ret;
>   
>   	nfc_cmd_regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>   	nfc->base_cmd_regs = devm_ioremap_resource(&pdev->dev, nfc_cmd_regs);
> @@ -2281,6 +2285,26 @@ static int atmel_nand_nfc_probe(struct platform_device *pdev)
>   
>   	nfc->is_initialized = true;
>   	dev_info(&pdev->dev, "NFC is probed.\n");
> +
> +	nfc->clk = devm_clk_get(&pdev->dev, NULL);
> +	if (!IS_ERR(nfc->clk)) {
> +		ret = clk_prepare_enable(nfc->clk);
> +		if (ret)
> +			return ret;

In this case, the NFC clock is not enabled, so I think 
nfc->is_initialized should set to false.
Otherwise, the nand driver will try to use the NFC without enabled the 
clock.

Is it better to move the code block

    nfc->is_initialized = true;
    dev_info(&pdev->dev, "NFC is probed.\n");

to just before the:

    return 0;


Best Regards,
Josh Wu
> +	} else {
> +		dev_warn(&pdev->dev, "NFC clock is missing");
> +	}
> +
> +	return 0;
> +}
> +
> +static int atmel_nand_nfc_remove(struct platform_device *pdev)
> +{
> +	struct atmel_nfc *nfc = &nand_nfc;
> +
> +	if (!IS_ERR(nfc->clk))
> +		clk_disable_unprepare(nfc->clk);
> +
>   	return 0;
>   }
>   
> @@ -2297,6 +2321,7 @@ static struct platform_driver atmel_nand_nfc_driver = {
>   		.of_match_table = of_match_ptr(atmel_nand_nfc_match),
>   	},
>   	.probe = atmel_nand_nfc_probe,
> +	.remove = atmel_nand_nfc_remove,
>   };
>   
>   static struct platform_driver atmel_nand_driver = {