HP Dragonfly Elite - Nobara Linux HP Privacy Screen Button

ncc74656k · August 29, 2024, 7:27pm

Greetings fellow nerds,

I was gifted a Dragonfly Elite Chromebook (redrix) from a previous job after an eval I ran, and I wanted more than ChromeOS. I first tried Windows, but couldn’t get the drivers to register from coolstar. I moved onto Linux and got to the point of getting Nobara Linux on it, and almost everything except audio and USB C worked out of the box. They’re now solved thanks to the Linux QOL Fixes page.

My last white whale is the Privacy Screen button HP builds into their Wolf Security Edition devices. I’d genuinely love to get that function enabled because this would be my travel device. Has anyone tried this before? Thanks for any help!

boredsquirrel · September 3, 2024, 2:24pm

can you elaborate on that feature? i am sure this is not common knowledge.

ncc74656k · September 3, 2024, 9:04pm

It’s a unique feature to HP (near as I can tell) that has a built in privacy screen feature which dramatically restricts viewing angles. It has a dedicated function key to activate it.

boredsquirrel · September 10, 2024, 4:40pm

This sounds totally cool but no idea how it works, and what toggles it. Changing the arrangement of the display internally?? Thats crazy

the keyword is “hp sureview”

ncc74656k · September 10, 2024, 6:51pm

I’m assuming it operates similarly to the electronically activated privacy glass? But yeah, it’s pretty cool. It has a distinct effect on the display’s viewing angles and clarity even when it’s off, so it’s far from perfect, but it’s still pretty awesome especially for a business class device.

elly · September 13, 2024, 1:06am

This sounds interesting, though I’ve never seen that before so it’d be challenging to implement remotely.

I’ll take a look at devicetree when I’ll have more time on my hands and let you know, been busy recently moving to a different country (after which I caught COVID…)

ncc74656k · September 13, 2024, 2:04am

Oh wow, that’d be awesome! If I can poke at anything to have a look on my end let me know. I’m pretty much a noob with Linux, despite fucking up a gentoo box “worse than anything I’ve seen before” according to my friend who helped me after trying to sort it out, lol.

Still, he said that must mean I’ve learned something along the way, lol.

elly · September 15, 2024, 11:37pm

Well, we all had to learn somehow. Failure is a part of learning process

Could you run debug script from our docs? I will need to take a look at ACPI tables and cbmem.

I believe we’ll need to make a driver for this, but we should check if this GPIO (GPP_H21 configured as output) is controlled by EC itself (it should, per GOOG0010) - in which case, you could control it with ectool and simply write a script/keybind to turn it on/off.

github.com

MrChromebox/coreboot/blob/MrChromebox-2408/src/mainboard/google/brya/variants/redrix/gpio.c#L62-L63


      
          	/* H21 : IMGCLKOUT2 ==> Privacy screen */
          	PAD_CFG_GPO(GPP_H21, 0, DEEP),

github.com

MrChromebox/coreboot/blob/MrChromebox-2408/src/mainboard/google/brya/variants/redrix/overridetree.cb#L108-L113


      
          				# Use ChromeOS privacy screen _HID
          				register "device[0].hid" = ""GOOG0010""
          				# Internal panel on the first port of the graphics chip
          				register "device[0].type" = "panel"
          				register "device[0].privacy.enabled" = "1"
          				register "device[0].privacy.gpio" = "ACPI_GPIO_OUTPUT_ACTIVE_HIGH(GPP_H21)"

ncc74656k · September 16, 2024, 2:29am

In fairness, he said “Actually to me it means you must be absorbing something because you found your way pretty deep into things and if this is the first time you’re touching things something must be sticking.” Also, in my defense, it was a Gentoo Stage 1 install, so I was begging for it, lmao.

I am afraid in this case that I don’t see where to go for the debug script. I don’t see a docs listing anywhere. Sorry if it’s obvious and I’m just looking right past it.

elly · September 19, 2024, 10:07am

Sure, you can find it here

github.com

chrultrabook/linux-tools/blob/main/debugging.sh

#!/usr/bin/env bash
printf 'Welcome.\nThis script will ask you for sudo password, which is necessary to extract required information.\n'

# Get current username before escalating permissions
user=$USER

# Get board name
board="$(cat /sys/class/dmi/id/product_name)"

# Get current date and time
date="$(date +"%Y-%m-%d_%Hh%Mm")"

# Specify where to store logs
logdir="debug-logs-$board-$date"
logarchive="debug-logs-$board-$date.tar.gz"

# Create directory to store logs
mkdir -p $logdir
cd $logdir

This file has been truncated. show original

ncc74656k · September 19, 2024, 8:06pm

cbmem.log:

ncc74656k · September 19, 2024, 8:13pm

Looks like the ACPI tables aren’t friendly to being uploaded as text, so I went ahead and posted them up here. Let me know if you’re able to access them and when you’re done so I can remove them.

Thank you so much!!!

https://drive.google.com/drive/folders/1GQydYbRmStk4JtBFh35ytU2MPrUU1rPA?usp=sharing

ncc74656k · October 11, 2024, 2:11am

Hey Elly! No rush at all, just curious if there’s anything else I can provide on this for you. Thanks again!

elly · November 1, 2024, 4:35pm

Hi, sorry it took so long!

First of all, it’s interesting that your device has IPU6 cameras:

[INFO ]  \_SB.PCI0.IPU0.IPU0: Intel MIPI Camera Device I2C address 00h
[INFO ]  \_SB.PCI0.I2C2.CAM0: Intel MIPI Camera Device I2C address 020h
[INFO ]  \_SB.PCI0.I2C2.NVM0: Intel MIPI Camera Device I2C address 050h

I’ve heard IPU6 should be in mainline 6.12, so you should have a chance to have working webcam in few weeks

As for the privacy screen, I don’t see any i2c devices exposed to the host. I suppose that it is controlled by the EC (just like gyroscope etc).

github.com

coreboot/chrome-ec/blob/16210b446aaea9483f78f3b34d9c41d334810e94/zephyr/subsys/usb_dc/hid_vivaldi.c#L38


      
          	[TK_FORWARD] = { .mask = BIT(1), .usage = 0x000C0225 },
          	[TK_REFRESH] = { .mask = BIT(2), .usage = 0x000C0227 },
          	[TK_FULLSCREEN] = { .mask = BIT(3), .usage = 0x000C0232 },
          	[TK_OVERVIEW] = { .mask = BIT(4), .usage = 0x000C029F },
          	[TK_BRIGHTNESS_DOWN] = { .mask = BIT(5), .usage = 0x000C0070 },
          	[TK_BRIGHTNESS_UP] = { .mask = BIT(6), .usage = 0x000C006F },
          	[TK_VOL_MUTE] = { .mask = BIT(7), .usage = 0x000C00E2 },
          	[TK_VOL_DOWN] = { .mask = BIT(8), .usage = 0x000C00EA },
          	[TK_VOL_UP] = { .mask = BIT(9), .usage = 0x000C00E9 },
          	[TK_SNAPSHOT] = { .mask = BIT(10), .usage = 0x00070046 },
          	[TK_PRIVACY_SCRN_TOGGLE] = { .mask = BIT(11), .usage = 0x000C02D0 },
          	[TK_KBD_BKLIGHT_DOWN] = { .mask = BIT(12), .usage = 0x000C007A },
          	[TK_KBD_BKLIGHT_UP] = { .mask = BIT(13), .usage = 0x000C0079 },
          	[TK_PLAY_PAUSE] = { .mask = BIT(14), .usage = 0x000C00CD },
          	[TK_NEXT_TRACK] = { .mask = BIT(15), .usage = 0x000C00B5 },
          	[TK_PREV_TRACK] = { .mask = BIT(16), .usage = 0x000C00B6 },
          	[TK_KBD_BKLIGHT_TOGGLE] = { .mask = BIT(17), .usage = 0x000C007C },
          	[TK_MICMUTE] = { .mask = BIT(18), .usage = 0x000B002F },
          };
          
          /* TK_* is 1-indexed, so the next bit is at ARRAY_SIZE(action_key) - 1 */

Unfortunately I don’t know much about vivaldi. Do you have any idea what’s the sanest way to implement this functionality, @MrChromebox?

ChocolateLoverRaj · November 1, 2024, 8:39pm

I thought vivaldi was just about the keyboard. THe real challenge would be finding out how to tun on/off the privacy feature.