Re: [PATCH v4 2/5] Input: atkbd: Expose function row physical map to userspace

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

Hi Rajat,

On Sat, Mar 28, 2020 at 11:59:13AM -0700, Rajat Jain wrote:
> Certain keyboards have their top-row keys intended
> for actions such as "Browser back", "Browser Refresh", "Fullscreen"
> etc as their primary mode, thus they will send physical codes for those
> actions. Further, they don't have a dedicated "Fn" key so don't have
> the capability to generate function key codes (e.g. F1, F2 etc..).
> However in this case, if userspace still wants to "synthesize" those
> function keys using the top row action keys, it needs to know the
> physical position of the top row keys. (Essentially a mapping between
> usage codes and a physical location in the top row).
> 
> This patch enhances the atkbd driver to receive such a mapping from the
> firmware / device tree, and expose it to userspace in the form of
> a function-row-physmap attribute. The attribute would be a space
> separated ordered list of physical codes, for the keys in the function
> row, in left-to-right order.
> 
> The attribute will only be present if the kernel knows about such
> mapping, otherwise the attribute shall not be visible.
> 
> Signed-off-by: Rajat Jain <rajatja@google.com>
> ---
> v4: Same as v3
> v3: Change to dev_dbg and remove unecessary error check
> v2: Remove the Change-Id from the commit log
> 
>  drivers/input/keyboard/atkbd.c | 56 ++++++++++++++++++++++++++++++++++
>  1 file changed, 56 insertions(+)
> 
> diff --git a/drivers/input/keyboard/atkbd.c b/drivers/input/keyboard/atkbd.c
> index 7e3eae54c1926..3b20aba1861cd 100644
> --- a/drivers/input/keyboard/atkbd.c
> +++ b/drivers/input/keyboard/atkbd.c
> @@ -24,6 +24,7 @@
>  #include <linux/libps2.h>
>  #include <linux/mutex.h>
>  #include <linux/dmi.h>
> +#include <linux/property.h>
>  
>  #define DRIVER_DESC	"AT and PS/2 keyboard driver"
>  
> @@ -63,6 +64,8 @@ static bool atkbd_terminal;
>  module_param_named(terminal, atkbd_terminal, bool, 0);
>  MODULE_PARM_DESC(terminal, "Enable break codes on an IBM Terminal keyboard connected via AT/PS2");
>  
> +#define MAX_FUNCTION_ROW_KEYS	24
> +
>  /*
>   * Scancode to keycode tables. These are just the default setting, and
>   * are loadable via a userland utility.
> @@ -230,6 +233,9 @@ struct atkbd {
>  
>  	/* Serializes reconnect(), attr->set() and event work */
>  	struct mutex mutex;
> +
> +	u16 function_row_physmap[MAX_FUNCTION_ROW_KEYS];
> +	int num_function_row_keys;
>  };
>  
>  /*
> @@ -283,6 +289,7 @@ static struct device_attribute atkbd_attr_##_name =				\
>  	__ATTR(_name, S_IRUGO, atkbd_do_show_##_name, NULL);
>  
>  ATKBD_DEFINE_RO_ATTR(err_count);
> +ATKBD_DEFINE_RO_ATTR(function_row_physmap);
>  
>  static struct attribute *atkbd_attributes[] = {
>  	&atkbd_attr_extra.attr,
> @@ -292,11 +299,42 @@ static struct attribute *atkbd_attributes[] = {
>  	&atkbd_attr_softrepeat.attr,
>  	&atkbd_attr_softraw.attr,
>  	&atkbd_attr_err_count.attr,
> +	&atkbd_attr_function_row_physmap.attr,
>  	NULL
>  };
>  
> +static ssize_t atkbd_show_function_row_physmap(struct atkbd *atkbd, char *buf)
> +{
> +	ssize_t size = 0;
> +	int i;
> +
> +	if (!atkbd->num_function_row_keys)
> +		return 0;
> +
> +	for (i = 0; i < atkbd->num_function_row_keys; i++)
> +		size += sprintf(buf + size, "%02X ",
> +				atkbd->function_row_physmap[i]);
> +	size += sprintf(buf + size, "\n");

Can we change this to scnprintf() with PAGE_SIZE as the initial limit?

> +	return size;
> +}
> +
> +static umode_t atkbd_attr_is_visible(struct kobject *kobj,
> +				struct attribute *attr, int i)
> +{
> +	struct device *dev = container_of(kobj, struct device, kobj);
> +	struct serio *serio = to_serio_port(dev);
> +	struct atkbd *atkbd = serio_get_drvdata(serio);
> +
> +	if (attr == &atkbd_attr_function_row_physmap.attr &&
> +	    !atkbd->num_function_row_keys)
> +		return 0;
> +
> +	return attr->mode;
> +}
> +
>  static struct attribute_group atkbd_attribute_group = {
>  	.attrs	= atkbd_attributes,
> +	.is_visible = atkbd_attr_is_visible,
>  };
>  
>  static const unsigned int xl_table[] = {
> @@ -1121,6 +1159,22 @@ static void atkbd_set_device_attrs(struct atkbd *atkbd)
>  	}
>  }
>  
> +static void atkbd_parse_fwnode_data(struct serio *serio)
> +{
> +	struct atkbd *atkbd = serio_get_drvdata(serio);
> +	struct device *dev = &serio->dev;
> +	int n;
> +
> +	/* Parse "function-row-physmap" property */
> +	n = device_property_count_u16(dev, "function-row-physmap");

I think for ACPI case it does not matter, but if we ever have device
tree using it, it will lead to less errors if we use u32 as the element
size. Do you mind switching to it?

> +	if (n > 0 && n <= MAX_FUNCTION_ROW_KEYS &&
> +	    !device_property_read_u16_array(dev, "function-row-physmap",
> +					    atkbd->function_row_physmap, n)) {
> +		atkbd->num_function_row_keys = n;
> +		dev_dbg(dev, "FW reported %d function-row key locations\n", n);
> +	}
> +}
> +
>  /*
>   * atkbd_connect() is called when the serio module finds an interface
>   * that isn't handled yet by an appropriate device driver. We check if
> @@ -1184,6 +1238,8 @@ static int atkbd_connect(struct serio *serio, struct serio_driver *drv)
>  		atkbd->id = 0xab00;
>  	}
>  
> +	atkbd_parse_fwnode_data(serio);
> +
>  	atkbd_set_keycode_table(atkbd);
>  	atkbd_set_device_attrs(atkbd);
>  

Thanks.

-- 
Dmitry