Message-ID: <16ffd1e5-bb9d-03f9-9bfc-5f5e712f6957@gmail.com>
Date:   Wed, 1 Feb 2023 03:24:36 +0530
From:   Rishit Bansal <rishitbansal0@...il.com>
To:     Hans de Goede <hdegoede@...hat.com>,
        Rishit Bansal <rishitbansal0@...il.com>,
        Mark Gross <markgross@...nel.org>,
        platform-driver-x86@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] platform/x86: hp-wmi: Support omen backlight control
 wmi-acpi methods


On 30/01/23 19:41, Hans de Goede wrote:

> Hi,
>
> On 1/20/23 23:39, Rishit Bansal wrote:
>> The HP Omen Command Studio application includes a Light Studio feature
>> which can be used to control various features related to the keyboard
>> backlight via the 0x20009 command.
>>
>> The command supports the following queries:
>>
>> - 0x1: Checks if keyboard lighting is supported
>> - 0x2: Get the zone colors of each of the 4 zones on the keyboard
>> - 0x3: Set the zone colors of each of the 4 zones on the keyboard
>> - 0x4: Gets the state (on/off) of the backlight
>> - 0x5: Sets the state (on/off) of the backlight
>>
>> This patch introduces a new sysfs led class called
>> "hp_omen::kbd_backlight" which can be used to control the state of the
>> backlight. It also includes a sysfs RW attribute called "kbd_rgb"
>> which can be used to get/set the current color of each zone.
>>
>> Additionally, it also maps the backlight event to the KEY_KBDILLUMTOGGLE
>> key so it shows the correct notification on userspace.
>>
>> The patch has been tested on an HP Omen 15-en0037AX (AMD) laptop.
>>
>> Signed-off-by: Rishit Bansal <rishitbansal0@...il.com>
>> ---
>> Changes since v1:
>>   - Map backlight key to KEY_KBDILLUMTOGGLE
>> ---
>>   drivers/platform/x86/hp/hp-wmi.c | 113 +++++++++++++++++++++++++++++++
>>   1 file changed, 113 insertions(+)
>>
>> diff --git a/drivers/platform/x86/hp/hp-wmi.c b/drivers/platform/x86/hp/hp-wmi.c
>> index 0a99058be813..a9e2634a9d46 100644
>> --- a/drivers/platform/x86/hp/hp-wmi.c
>> +++ b/drivers/platform/x86/hp/hp-wmi.c
>> @@ -27,6 +27,7 @@
>>   #include <linux/rfkill.h>
>>   #include <linux/string.h>
>>   #include <linux/dmi.h>
>> +#include <linux/leds.h>
>>   
>>   MODULE_AUTHOR("Matthew Garrett <mjg59@...f.ucam.org>");
>>   MODULE_DESCRIPTION("HP laptop WMI hotkeys driver");
>> @@ -136,6 +137,7 @@ enum hp_wmi_command {
>>   	HPWMI_WRITE	= 0x02,
>>   	HPWMI_ODM	= 0x03,
>>   	HPWMI_GM	= 0x20008,
>> +	HPWMI_KB    = 0x20009,
>>   };
>>   
>>   enum hp_wmi_hardware_mask {
>> @@ -219,6 +221,7 @@ static const struct key_entry hp_wmi_keymap[] = {
>>   	{ KE_KEY, 0x21a9,  { KEY_TOUCHPAD_OFF } },
>>   	{ KE_KEY, 0x121a9, { KEY_TOUCHPAD_ON } },
>>   	{ KE_KEY, 0x231b,  { KEY_HELP } },
>> +	{ KE_KEY, KEY_KBDILLUMTOGGLE, { KEY_KBDILLUMTOGGLE }},
> Please drop this entry (also see the comment about this below).
>
>>   	{ KE_END, 0 }
>>   };
>>   
>> @@ -734,12 +737,56 @@ static ssize_t postcode_store(struct device *dev, struct device_attribute *attr,
>>   	return count;
>>   }
>>   
>> +static ssize_t kbd_rgb_show(struct device *dev,
>> +				    struct device_attribute *attr, char *buf)
>> +{
>> +	u8 val[128];
>> +
>> +	int ret = hp_wmi_perform_query(HPWMI_HDDTEMP_QUERY, HPWMI_KB, &val,
>> +				       zero_if_sup(val), sizeof(val));
>> +
>> +	if (ret)
>> +		return ret;
>> +
>> +	strncat(buf, &val[25], 12);
>> +
>> +	return strlen(buf);
> strncat requires that the buffer is pre-zeroed which I'm not
> sure is always the case. Instead you should problably just do something
> like this:
>
> 	strscpy(buf, 12, &val[25]);


I think I'll go with a memcpy() instead (more details in a comment 
below), because it turns out \x00 is a valid hex color code, so none of 
the inputs/outputs should be handled with string functions, as they'll 
terminate early.


>
> 	return strlen(buf);
>
> Alos may I ask what the output of reading the new kbd_rgb file
> actually looks like ?


The output of the kbd_rgb file looks something like this (this is when I 
have set

some random colors on each zone):

rishit@...hit-OMEN-Laptop-15-en0xxx:~$ xxd 
/sys/devices/platform/hp-wmi/kbd_rgb
00000000: 0f84 fa71 0ffa f935 0ffa ac0f            ...q...5....

The keyboard has "4 zones" of lighting.

Each zone takes 'R', 'G', and 'B' components of the color for that zone.

Each of these components can take a value from 1-255 (i.e, 1 byte).

This gives a total of 1 (byte) * 3 (R,G,B) * 4 (zones) = 12 bytes of 
information.

Although this is the most compact way to represent this information, I 
feel that it may not be the most user friendly way, so I am open for 
suggestions on better to represent the output.


To summarize, it looks something like this currently:


0                   1                   2                   3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|             Zone 1                            |  Zone 2...    |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ...                           |           Zone 3...           |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ...           |            Zone 4                             |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+


Also on hindsight, this was clearly not documented well on the initial 
commit message. I'll try to include a short description of this format 
in the next patch description.


>
> And can you please start a new Documentation file:
>
> Documentation/ABI/testing/sysfs-platform-hp-wmi
>
> and document this new kbd_rgb file there?
>
> See for example:
>
> Documentation/ABI/testing/sysfs-platform-asus-wmi
>
> for what this file should look like.
>
> Bonus points if you also do a follow-up patch documenting more
> of the sysfs attributes used by this driver, but at a minimum
> lets start documenting any new files we add.
>
>
>
>> +}
>> +
>> +static ssize_t kbd_rgb_store(struct device *dev,
>> +				     struct device_attribute *attr,
>> +				     const char *buf, size_t count)
>> +{
>> +	u8 val[128];
>> +	int ret;
>> +
>> +	ret = hp_wmi_perform_query(HPWMI_HDDTEMP_QUERY, HPWMI_KB, &val,
>> +				   zero_if_sup(val), sizeof(val));
>> +
>> +	if (ret)
>> +		return ret;
>> +
>> +	if (count != 12)
>> +		return -1;
> So what is the exact format here?

See previous comment  where I describe the format ^


> and should we not do more error
> checking?


I tried to mess with the input a bit and try to reproduce some edge 
cases. One thing I definitely realized from your comment about the null 
byte (0), is that if a single null byte is present in the input string, 
strncpy() will terminate early and not write the entire bytes as we want 
to. I'll fix this to be a memcpy() instead. Apart from that, as each 
byte can take the full range of inputs (0-255), I'm not sure if any more 
error checking can be done on the input itself, though I am open for 
more suggestions to improve this.


> Also we want 12 chars, then the show code above should use 13,
> so that the buffer gets 12 chars + a terminating 0.


The reason I didn't null terminate the output string is because I 
intended to keep the output to exactly 12 bytes. It was unintentional to 
use str_*() functions in the patch, I will correct those to memcpy()'s 
instead.


>
> Also maybe add defines for the offset value of 25 and the buflen of 12 ?
>
>> +
>> +	strncpy(&val[25], buf, count);
>> +
>> +	ret = hp_wmi_perform_query(HPWMI_ALS_QUERY, HPWMI_KB, &val, sizeof(val),
>> +				   0);
>> +
>> +	if (ret)
>> +		return ret;
>> +
>> +	return count;
>> +}
>> +
>>   static DEVICE_ATTR_RO(display);
>>   static DEVICE_ATTR_RO(hddtemp);
>>   static DEVICE_ATTR_RW(als);
>>   static DEVICE_ATTR_RO(dock);
>>   static DEVICE_ATTR_RO(tablet);
>>   static DEVICE_ATTR_RW(postcode);
>> +static DEVICE_ATTR_RW(kbd_rgb);
>>   
>>   static struct attribute *hp_wmi_attrs[] = {
>>   	&dev_attr_display.attr,
>> @@ -748,6 +795,7 @@ static struct attribute *hp_wmi_attrs[] = {
>>   	&dev_attr_dock.attr,
>>   	&dev_attr_tablet.attr,
>>   	&dev_attr_postcode.attr,
>> +	&dev_attr_kbd_rgb.attr,
>>   	NULL,
>>   };
>>   ATTRIBUTE_GROUPS(hp_wmi);
>> @@ -853,6 +901,8 @@ static void hp_wmi_notify(u32 value, void *context)
>>   	case HPWMI_PROXIMITY_SENSOR:
>>   		break;
>>   	case HPWMI_BACKLIT_KB_BRIGHTNESS:
>> +		sparse_keymap_report_event(hp_wmi_input_dev,
>> +						KEY_KBDILLUMTOGGLE, 1, true);
> Please just directly report the key instead of inserting a fake scancode into
> the parse-map:
>
> 		input_report_key(hp_wmi_input_dev, KEY_KBDILLUMTOGGLE, true);
> 		input_sync(hp_wmi_input_dev);
> 		input_report_key(hp_wmi_input_dev, KEY_KBDILLUMTOGGLE, false);
> 		input_sync(hp_wmi_input_dev);
>>   		break;
>>   	case HPWMI_PEAKSHIFT_PERIOD:
>>   		break;
>> @@ -1294,6 +1344,63 @@ static int thermal_profile_setup(void)
>>   
>>   static int hp_wmi_hwmon_init(void);
>>   
>> +static struct led_classdev omen_kbd_led;
>> +
>> +static enum led_brightness get_omen_backlight_brightness(struct led_classdev *cdev)
>> +{
>> +	u8 val;
>> +
>> +	int ret = hp_wmi_perform_query(HPWMI_HARDWARE_QUERY, HPWMI_KB, &val, zero_if_sup(val), sizeof(val));
>> +
>> +	if (ret)
>> +		return ret;
>> +
>> +	return (val & 0x80) ? LED_ON : LED_OFF;
>> +}
>> +
>> +static void set_omen_backlight_brightness(struct led_classdev *cdev, enum led_brightness value)
>> +{
>> +	char buffer[4] = { (value == LED_OFF) ? 0x64 : 0xe4, 0, 0, 0 };
>> +
>> +	hp_wmi_perform_query(HPWMI_WIRELESS_QUERY, HPWMI_KB, &buffer,
>> +				       sizeof(buffer), 0);
>> +}
>> +
>> +
>> +static bool is_omen_lighting_supported(void)
>> +{
>> +	u8 val;
>> +
>> +	int ret = hp_wmi_perform_query(HPWMI_DISPLAY_QUERY, HPWMI_KB, &val, zero_if_sup(val), sizeof(val));
>> +
>> +	if (ret)
>> +		return false;
>> +
>> +	return (val & 1) == 1;
>> +}
>> +
>> +static int omen_backlight_init(struct device *dev)
>> +{
>> +	int ret;
>> +
>> +	omen_kbd_led.name = "hp_omen::kbd_backlight";
>> +	omen_kbd_led.brightness_set = set_omen_backlight_brightness;
>> +	omen_kbd_led.brightness_get = get_omen_backlight_brightness;
>> +	omen_kbd_led.max_brightness = 1;
>> +
>> +	ret = devm_led_classdev_register(dev, &omen_kbd_led);
>> +
>> +	if (ret < 0)
>> +		return -1;
>> +
>> +	return 0;
>> +}
>> +
>> +static void omen_backlight_exit(struct device *dev)
>> +{
>> +	devm_led_classdev_unregister(dev, &omen_kbd_led);
>> +}
>> +
> This is not necessary, the whole idea behin devm is that this
> will automatically get unegistered on driver unbind.
>
>>   static int __init hp_wmi_bios_setup(struct platform_device *device)
>>   {
>>   	int err;
>> @@ -1321,6 +1428,9 @@ static int __init hp_wmi_bios_setup(struct platform_device *device)
>>   
>>   	thermal_profile_setup();
>>   
>> +	if (is_omen_lighting_supported())
>> +		omen_backlight_init(&device->dev);
>> +
>>   	return 0;
>>   }
>>   
>> @@ -1349,6 +1459,9 @@ static int __exit hp_wmi_bios_remove(struct platform_device *device)
>>   	if (platform_profile_support)
>>   		platform_profile_remove();
>>   
>> +	if (is_omen_lighting_supported())
>> +		omen_backlight_exit(&device->dev);
>> +
> And this likewise thus is not necessary.
>
>>   	return 0;
>>   }
>>   


Thank you for reviewing this! I'll make a follow up v2 patch with the 
requested changes.


> Regards,
>
> Hans
>

Powered by blists - more mailing lists