FAQ

If you have a 2021 G15 and use a second NVMe drive you might experience a problem while using s0ix (s2idle) resulting in a delayed system suspend.

The problem lies within a faulty DSDT table and the secondary NVMe drive does not enter suspend properly.

With an updated DSDT table you can apply the necessary patches to make it suspend properly.

https://gitlab.com/smbruce/GA503QR-StorageD3Enable-DSDT-Patch

Important: This is not necessary if you want to use S3 (see above). It applies only if you want to further use the standard suspend method s0ix.

Fixed in kernels versions 6.1.x and up.

Archived. All patches that were used to make those modules are now upstreamed. For the best ASUS experience please ensure you are using the minimum of 5.13 kernel.

After BIOS version 313, ASUS fix ACPI support for Linux, which is crucial if you want Performance mode to work properly.

And ASUS optimized power distribution between CPU and GPU, which before caused stuttering/frame drops in performance mode that confuse many users for a long time.

Only Fedora (current version) and OpenSUSE Tumbleweed officially. Other distros may have packaging, or you will need to compile and install it yourself.

BIOS versions 408 & 409 cannot boot a Linux kernel newer than 5.15.x so you will need to upgrade to the 410 bios here.

Supergfxctl v5.0.0+ now supports switching between Hybrid, Integrated and Vfio modes on the Flow X13.

Be sure to consider the following:

• disable fast boot within the BIOS
• disable fast boot in Windows
• always fully shutdown after using and switchting to another OS so the hardware gets correctly initialized

If you still experience an issue, hold down the power button while on battery for a few seconds to force a shutdown. This often helps to reset some things.

These steps are not needed if you are running Linux exclusively.

This function is now integrated into the ROG Control Center, so long as you run it in the background you will get the notifications.

You can find all notify settings in the "System Settings" in the ROG Control Center.

Depending on your kernel version, you may occasioinally experience issues with the 2021/2022 versions of the Zephyrus G14/G15 which affects the proper use or newer suspend methods, like s0ix.

A potential fix is to patch your DSDT tables so your machine uses the older suspend method, called S3. In our tests this works great on the 2021 / 2022 G14 and G15. Those patches are not part of the main repo and can't be. It will always be a manual matter and cannot be integrated into the kernel.

Important: If you update the BIOS be sure you disable your DSDT table and create a new one. DSDT tables could change with newer BIOS versions!

You can find the script here: https://gitlab.com/marcaux/g14-2021-s3-dsdt

Please ensure you are using a recent kernel. As of now the minimum is 5.11, but please use at least 5.13 so that you get all the most recent patches and fixes for ASUS laptops.

You need to map the key-combo to an action in your desktop, like this:

Check the logs with sudo journalctl -b -u asusd and look for errors.

The daemon isn't running, check the logs with sudo journalctl -b -u asusd and look for errors.

Ensure you are using kernel 5.11 at minimum. We recommend to use at least 5.13 so that you get all the most recent patches and fixes for ASUS laptops.

It's also possible that your laptop doesn't support this so if the kernel update doesn't solve this feel free to make a :sadface: (sorry).

Custom fan curves (not speaking of the built in power profiles) are currently only supported on Ryzen ROG laptops.
The necessary kernel patches are merged since 5.17.

The format is shown here.

There are three fan profiles namely Quiet, Balanced and Performance to choose from. Each profile is linked to power profile and gets applied when the power profile is set. You can enable/disable the fan profiles using the following command:

asusctl fan-curve -m <profile_name> -e true/false

All three fan profiles can be activated at once. If no profile is activated manually then the fan curve from the BIOS is used.

To change the fan curve data for a specific profile use the following command:

asusctl fan-curve -m <profile_name> -D <fan_curve_data>

There's a few steps to troubleshoot.

1. If you are dual-booting Windows 10 / 11 (or think Windows may have been in iGPU Mode in Armoury Crate) then try this,

journalctl -b -u supergfxd

If there are lines which end with ERROR: Could not find dGPU, then reboot into Windows and make sure that Armoury Crate has the dGPU set to Auto Mode. If you have set it to iGPU only mode, then the dGPU will be forcibly removed from the system and Linux will no longer be able to see it.

If you no longer have Windows installed, and you suspect that it was set to iGPU only mode when you installed Linux, then you can try to forcibly enable the dGPU with the following.

echo 0 | sudo tee /sys/devices/platform/asus-nb-wmi/dgpu_disable
echo 1 > sudo /sys/bus/pci/rescan
echo 0 | sudo tee /sys/devices/platform/asus-nb-wmi/dgpu_disable

It may take a few tries to enable it, and a reboot is recommended afterwards.

2. Are your Nvidia drivers installed correctly?

In hybrid or nvidia mode they should be loaded, check with lsmod |grep nvidia for similar to:

[luke@datatron]$ lsmod |grep nvidia
nvidia_drm             65536  11
nvidia_modeset       1232896  18 nvidia_drm
nvidia_uvm           1150976  0
nvidia              34185216  906 nvidia_uvm,nvidia_modeset
drm_kms_helper        274432  2 nvidia_drm,i915
drm                   618496  14 drm_kms_helper,nvidia_drm,i915

if that doesn't show similar, then you need to reinstal your Nvidia drivers.

3. Is Nouveau blocklisted correctly? It's common to require the following in /etc/default/grub:

GRUB_CMDLINE_LINUX="rd.driver.blacklist=nouveau modprobe.blacklist=nouveau"

that line may not be exact, and do not delete existing entry to match it, just add those two items to the end of the line before the last "

4. What does journalctl -b -u asusd say after a failed switch?

If it contains lines like the following then we need to look at active sessions

asusd[1373]: INFO: GFX: Switching gfx mode to integrated
asusd[1373]: INFO: GFX: display-manager thread started
asusd[1373]: WARN: GFX: thread time exceeded 3 minutes, exiting

Do: loginctl list-sessions, if it shows more than one session lets have a look with:

loginctl show-session --property Type --property Class --property Active <num>

where <num> is the number or char+num from the SESSION column of the output above. If there is more than one active session of type Wayland | X11 | Mir then somehow you've ended up with an extra session.

Save your work, Ctrl+Alt+F4 (or any F<num>) key to a TTY terminal session and run

asusctl graphics -m <desired mode>
sudo loginctl kill-user <username>

your display-manager should then restart.

5. If none of the above helps, and/or you use KDE, and/or have a G14 with GTX graphics, you may need to this,

sudo sed -i 's/#KillUserProcesses=no/KillUserProcesses=yes/' /etc/systemd/logind.conf

then sudo systemctl restart display-manager.

This happens on AMD machines typically, if your machine is AMD then you need to add the following to run on graphical login:

xrandr --setprovideroutputsource modesetting NVIDIA-0
xrandr --auto

there are many ways to do this so you will need to search for the right way for your login manager.

KDE/SDDM NOTES The location of the startup script for KDE/SDDM is wrong on most resources found on the internet: To get the login screen to show up on the laptop screen when in Nvidia mode, the xrandr commands must be added to /etc/sddm/Xsetup (NOT /usr/share/sddm/scripts/Xsetup which most resources mention). (thanks motoridersd )

As of version 510.60.02, NVIDIA still doesn't support dynamic boost on AMD CPUs. This is a known issue might get fixed in some future release.

Since version 525.53, NVIDIA added official Dynamic Boost support for AMD laptop that uses Ryzen 6000 Series (or newer) CPU.

To enable it, follow the next steps:

1. Start nvidia-powerd.service

sudo systemctl start nvidia-powerd.service

If you don't want to start it manually each time, then:

sudo systemctl enable nvidia-powerd.service

Notice: There is a known bug that nvidia-powerd process uses excessive CPU after you offload the prime run application. You can stop by:

sudo systemctl stop nvidia-powerd.service

Update: version 530.41.03/525.105.17 and newer driver fixed this issue. Now you can enable nvidia-powered without worrying about high CPU usage.

2. Set your power profile to "Performance" Mode

You can do it in many ways. If you already installed asusctl, you can switch to it using:

asusctl profile -P Performance

or ROG Control Center (GUI) to set it.

Besides, you can enable it using your DE control panel (such as KDE/GNOME).

If you didn't see any option in DE panel, then you need to check if the "power-profiles-daemon" package was properly installed.

3. Test it

Using tools like "nvtop" and "mangohud", you can monitor your CPU and GPU power in realtime.

To check if the Dynamic boost works, you need to identify what is the MAX TGP your model support. Usually, it can be found on manufacturers' websites.

Here are some power data collected from Zephyrus G15 (2022), which rated 120 watt MAX TGP.

Quiete Mode: 25 Watt (CPU) + 60 Watt (GPU)

Balanced Mode: 30 Watt (CPU) + 80 Watt (GPU)

Performance Mode: 25 Watt (CPU) + 115 Watt (GPU)

This problem mostly appears when connect your external display to the dGPU in Hybrid mode (rendering screens on separate GPUs). There are various reasons that can cause this problem and we don't have a perfect solution for it yet. But here are some possible ways to solve it.

1. Use X11 instead of Wayland.

2. If your model supports video via USB-C and one of the USB-C ports takes the signal from iGPU, then buy a USB-C to HDMI/DP cable or dongle. this way, you can avoid rendering two screens on separate GPUs. (tested on 2022 Zephyrus G15)

Yes! Similar to remapping the Arrow-Keys above, you can remap - say the menu (fn+RightCtrl) key to SysRq.
Just add another line to /etc/udev/hwdb.d/90-nkey.hwdb with the following, including the leading whitespaces:

  KEYBOARD_KEY_<ScanCode>=sysrq        # force remap sysrq to Fn+RightCtrl

You can get the <ScanCode> by running

evtest /dev/input/by-id/usb-ASUSTeK_Computer_Inc._N-KEY_Device-*-kbd

and pressing the RightCtrl key.
In this case, it is 70065

Testing ... (interrupt to exit)
Event: time 1662839073.640933, type 4 (EV_MSC), code 4 (MSC_SCAN), value 70065 <--------- Substitute this as <ScanCode>
Event: time 1662839073.640933, type 1 (EV_KEY), code 127 (KEY_COMPOSE), value 1
Event: time 1662839073.640933, -------------- SYN_REPORT ------------

Then update hwdb with:

sudo systemd-hwdb update
sudo udevadm trigger

The issue is that the current keycode being emitted by the driver is only recognised by Wayland. X11 and desktops using X11 require F20 to be emitted.

Create a file named /etc/udev/hwdb.d/90-nkey.hwdb with:

# Format evdev:input:b<bus_id>v<vendor_id>p<product_id>

# ** Note **
# The line evdev:input:b0003v0B05p1866* may vary on your ASUS Laptop.  
# Modify the <vendor_id> and <product_id> based on the output of this command to ensure remaps work:
# $ lsusb | grep 'ASUSTek Computer, Inc. N-KEY Device' | awk -F'[: ]' '{print $7" "$8}' | tr '[:lower:]' '[:upper:]'

evdev:input:b0003v0B05p1866*
  KEYBOARD_KEY_ff31007c=f20 # x11 mic-mute

then update hwdb with:

sudo systemd-hwdb update
sudo udevadm trigger

Yes, create a file named /etc/udev/hwdb.d/90-nkey.hwdb with:

# Format evdev:input:b<bus_id>v<vendor_id>p<product_id>

# ** Note **
# The line evdev:input:b0003v0B05p1866* may vary on your ASUS Laptop.  
# Modify the <vendor_id> and <product_id> based on the output of this command to ensure remaps work:
# $ lsusb | grep 'ASUSTek Computer, Inc. N-KEY Device' | awk -F'[: ]' '{print $7" "$8}' | tr '[:lower:]' '[:upper:]'

evdev:input:b0003v0B05p1866*
  KEYBOARD_KEY_c00b6=kbdillumdown # Fn+F2 (music prev)
  KEYBOARD_KEY_c00b5=kbdillumup   # Fn+F4 (music skip)
  KEYBOARD_KEY_ff3100c5=pagedown  # Fn+Down
  KEYBOARD_KEY_ff3100c4=pageup    # Fn+Up
  KEYBOARD_KEY_ff3100b2=home      # Fn+Left
  KEYBOARD_KEY_ff3100b3=end       # Fn+Right

then update hwdb with:

sudo systemd-hwdb update
sudo udevadm trigger

You can see a list of keycodes here.

Yes, you can.

You may have a different keyboard. Please request support in one of the related projects on gitlab, or in the discord server.

No, the key is on a physically different circuit and used to physically signal the keyboard EC to switch key circuits.

There are three different circuits for the 0x8166 keyboard.