Re: [PATCH] firmware: coreboot: Let OF core populate platform device

[Brian Norris]

Hi Stephen,

Although I already 'Reviewed' this, I noticed another problem when I
went to look again:

On Mon, Aug 06, 2018 at 10:10:47AM -0700, Stephen Boyd wrote:
> Now that the /firmware/coreboot node in DT is populated by the core DT
> platform code with commit 3aa0582fdb82 ("of: platform: populate
> /firmware/ node from of_platform_default_populate_init()") we should and
> can remove the platform device creation here. Otherwise, the
> of_platform_device_create() call will fail, the coreboot of driver won't
> be registered, and this driver will never bind. At the same time, we can
> move this driver to use platform device APIs instead of DT specific ones
> and drop the of_node handling that was presumably placed there to hold a
> reference to the DT node created during module init.
> 
> Cc: Wei-Ning Huang <wnhuang@xxxxxxxxxxxx>
> Cc: Julius Werner <jwerner@xxxxxxxxxxxx>
> Cc: Brian Norris <briannorris@xxxxxxxxxxxx>
> Cc: Samuel Holland <samuel@xxxxxxxxxxxx>
> Cc: Thierry Escande <thierry.escande@xxxxxxxxxxxxx>
> Signed-off-by: Stephen Boyd <swboyd@xxxxxxxxxxxx>
> ---
>  drivers/firmware/google/coreboot_table-of.c | 39 +++++----------------
>  1 file changed, 9 insertions(+), 30 deletions(-)
> 
> diff --git a/drivers/firmware/google/coreboot_table-of.c b/drivers/firmware/google/coreboot_table-of.c
> index f15bf404c579..270f112bdc54 100644
> --- a/drivers/firmware/google/coreboot_table-of.c
> +++ b/drivers/firmware/google/coreboot_table-of.c
> @@ -18,21 +18,20 @@
>  #include <linux/device.h>
>  #include <linux/io.h>
>  #include <linux/module.h>
> -#include <linux/of_address.h>
> -#include <linux/of_platform.h>
> +#include <linux/mod_devicetable.h>
>  #include <linux/platform_device.h>
>  
>  #include "coreboot_table.h"
>  
>  static int coreboot_table_of_probe(struct platform_device *pdev)
>  {
> -	struct device_node *fw_dn = pdev->dev.of_node;
>  	void __iomem *ptr;
> +	struct resource *res;
>  
> -	ptr = of_iomap(fw_dn, 0);
> -	of_node_put(fw_dn);
> -	if (!ptr)
> -		return -ENOMEM;
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	ptr = devm_ioremap_resource(&pdev->dev, res);

You're making this a devm_* ioremap, but coreboot_table_exit() will
handle unmapping this for you already. So you're introducing a potential
double-unmap.

> +	if (IS_ERR(ptr))
> +		return PTR_ERR(ptr);
>  
>  	return coreboot_table_init(&pdev->dev, ptr);
>  }
> @@ -44,8 +43,9 @@ static int coreboot_table_of_remove(struct platform_device *pdev)
>  
>  static const struct of_device_id coreboot_of_match[] = {
>  	{ .compatible = "coreboot" },
> -	{},
> +	{}

...and since I had a *real* comment, I'll mention this bit I noticed
previously...this diff seems a bit superfluous. I understand we don't
really want to extend this list beyond the NULL terminator, but
normally, I'd expect we leave it alone.

Brian

>  };
> +MODULE_DEVICE_TABLE(of, coreboot_of_match);
>  
>  static struct platform_driver coreboot_table_of_driver = {
>  	.probe = coreboot_table_of_probe,
> @@ -55,28 +55,7 @@ static struct platform_driver coreboot_table_of_driver = {
>  		.of_match_table = coreboot_of_match,
>  	},
>  };
> -
> -static int __init platform_coreboot_table_of_init(void)
> -{
> -	struct platform_device *pdev;
> -	struct device_node *of_node;
> -
> -	/* Limit device creation to the presence of /firmware/coreboot node */
> -	of_node = of_find_node_by_path("/firmware/coreboot");
> -	if (!of_node)
> -		return -ENODEV;
> -
> -	if (!of_match_node(coreboot_of_match, of_node))
> -		return -ENODEV;
> -
> -	pdev = of_platform_device_create(of_node, "coreboot_table_of", NULL);
> -	if (!pdev)
> -		return -ENODEV;
> -
> -	return platform_driver_register(&coreboot_table_of_driver);
> -}
> -
> -module_init(platform_coreboot_table_of_init);
> +module_platform_driver(coreboot_table_of_driver);
>  
>  MODULE_AUTHOR("Google, Inc.");
>  MODULE_LICENSE("GPL");
> -- 
> Sent by a computer through tubes
>