FreeBSD as my new desktop, frequent total freeze

Freedom · Thursday at 10:56 PM

Hi,

I made the big jump and setup FreeBSD as my desktop
First I have to say the installation and setup of Wayland, XWayland, the login manager and Sway was a clear path and flawless. Thanks to the excellent FreeBSD handbook.

Now the big problem:
I have one of those very common Intel N100 with an Alder Lake-N UHD graphics

vgapci0@pci0:0:2:0: class=0x030000 rev=0x00 hdr=0x00 vendor=0x8086 device=0x46d1 subvendor=0x0000 subdevice=0x0000
vendor = 'Intel Corporation'
device = 'Alder Lake-N [UHD Graphics]'
class = display
subclass = VGA

And I knew there might be some issues:
https://bsd-hardware.info/?id=pci:8086-46d1

20241212 - needs drm-610 package, min FreeBSD 14, still kernel panics occasionally, hardware decoding via vaapi/vdupa crashes. Open GL Works

I have installed drm-kmod gpu-firmware-kmod and drm-61-kmod

Problems:
1/ When suspend with
acpiconf -s 3
I see a lot of error at wake up and get about 2min30 of those messages:

Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Controller timeout
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ============== REGISTER DUMP ==============
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Sys addr: 0x00000000 | Version: 0x00001002
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Blk size: 0x00000000 | Blk cnt: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Argument: 0x00000000 | Trn mode: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Present: 0x1fff0000 | Host ctl: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Power: 0x0000000b | Blk gap: 0x00000080
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Wake-up: 0x00000000 | Clock: 0x0000fa07
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Timeout: 0x00000000 | Int stat: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Int enab: 0x01ff003b | Sig enab: 0x01ff003b
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: AC12 err: 0x00000000 | Host ctl2:0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Caps: 0x546ec881 | Caps2: 0x80000807
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Max curr: 0x00000000 | ADMA err: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ADMA addr:0x00000000 | Slot int: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ===========================================
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Controller timeout
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ============== REGISTER DUMP ==============
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Sys addr: 0x00000000 | Version: 0x00001002
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Blk size: 0x00000000 | Blk cnt: 0x00000000

I can live with that... no problem.

2/ I get a lot of total freeze with the only GUI application I use:
- `chromium` (I tried both `--ozone-platform=wayland` and `--ozone-platform=x11`)
- but I got them with librewolf too
- qutebrowser just throw a `Trace/BPT trap (core dumped)` at startup

When it freeze I can't even get to a tty.
Sometimes only one of my monitor freeze.

Some things I collected:

> doas kldstat
Password:
Id Refs Address Size Name
1 90 0xffffffff80200000 1f41500 kernel
2 1 0xffffffff82143000 5e9328 zfs.ko
3 1 0xffffffff8272d000 7808 cryptodev.ko
4 1 0xffffffff82735000 1c5e8 geom_eli.ko
5 1 0xffffffff831e5000 1e0228 i915kms.ko
6 2 0xffffffff833c6000 85090 drm.ko
7 1 0xffffffff8344c000 22b8 iic.ko
8 2 0xffffffff8344f000 4120 linuxkpi_video.ko
9 3 0xffffffff83454000 7350 dmabuf.ko
10 4 0xffffffff8345c000 3378 lindebugfs.ko
11 1 0xffffffff83460000 c338 ttm.ko
12 1 0xffffffff8346d000 4250 ichsmb.ko
13 1 0xffffffff83472000 2178 smbus.ko
14 1 0xffffffff83475000 b1270 if_iwlwifi.ko
15 1 0xffffffff83527000 6020 ig4.ko
16 1 0xffffffff8352e000 3360 uhid.ko
17 1 0xffffffff83532000 4364 ums.ko
18 1 0xffffffff83537000 5640 ng_ubt.ko
19 4 0xffffffff8353d000 abb8 netgraph.ko
20 3 0xffffffff83548000 a250 ng_hci.ko
21 2 0xffffffff83553000 2670 ng_bluetooth.ko
22 1 0xffffffff83556000 2260 pflog.ko
23 1 0xffffffff83559000 4f978 pf.ko
24 1 0xffffffff835a9000 2a80 mac_ntpd.ko

> dmesg
...
drmn0: [drm] *ERROR* [CRTC:80ipe A] flip_done timed out
WARNING new_crtc_state->event failed at /wrkdir/usr/ports/graphics/drm-61-kmod/work/drm-kmod-drm_v6.1.128_6/drivers/gpu/drm/drm_atomic_helper.c:2485
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [CRTC:80ipe A] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [CONNECTOR:195:HDMI-A-1] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [PLANE:31lane 1A] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [PLANE:76:cursor A] commit wait timed out
drmn0: [drm] *ERROR* [CRTC:80ipe A] flip_done timed out
WARNING new_crtc_state->event failed at /wrkdir/usr/ports/graphics/drm-61-kmod/work/drm-kmod-drm_v6.1.128_6/drivers/gpu/drm/drm_atomic_helper.c:2485
drmn0: [drm] *ERROR* [CRTC:131ipe B] flip_done timed out
WARNING new_crtc_state->event failed at /wrkdir/usr/ports/graphics/drm-61-kmod/work/drm-kmod-drm_v6.1.128_6/drivers/gpu/drm/drm_atomic_helper.c:2485
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [CRTC:80ipe A] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [CRTC:131ipe B] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [CONNECTOR:195:HDMI-A-1] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [CONNECTOR:202:HDMI-A-2] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [PLANE:31lane 1A] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [PLANE:76:cursor A] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [PLANE:82lane 1B] commit wait timed out
...

cat /var/log/messages
...
Nov 6 12:58:50 lander2 pulseaudio[31035]: [] cli-command.c: stat('/usr/local/etc/pulse/default.pa.d'): No such file or directory
Nov 6 13:03:50 lander2 devd[72439]: check_clients: dropping disconnected client
Nov 6 15:09:53 lander2 kernel: drmn0: [drm] *ERROR* [CRTC:80ipe A] flip_done timed out
Nov 6 15:09:53 lander2 kernel: WARNING new_crtc_state->event failed at /wrkdir/usr/ports/graphics/drm-61-kmod/work/drm-kmod-drm_v6.1.128_6/drivers/gpu/drm/drm_atomic_helper.c:2485
Nov 6 15:21:32 lander2 kernel: drmn0: [drm] *ERROR* [CRTC:131ipe B] flip_done timed out
Nov 6 15:21:32 lander2 kernel: WARNING new_crtc_state->event failed at /wrkdir/usr/ports/graphics/drm-61-kmod/work/drm-kmod-drm_v6.1.128_6/drivers/gpu/drm/drm_atomic_helper.c:2485
Nov 6 15:21:45 lander2 seatd[13433]: 05:21:03.850 [INFO] [seatd/seat.c:673] Switching from VT 2 to VT 3
Nov 6 15:21:45 lander2 seatd[13433]: 05:21:03.851 [INFO] [seatd/seat.c:606] Disabling client 2 on seat0
Nov 6 15:22:00 lander2 kernel: drmn0: [drm] *ERROR* flip_done timed out
Nov 6 15:22:00 lander2 kernel: drmn0: [drm] *ERROR* [CRTC:80ipe A] commit wait timed out
Nov 6 15:22:07 lander2 kernel: drmn0: [drm] *ERROR* flip_done timed out
Nov 6 15:22:07 lander2 kernel: drmn0: [drm] *ERROR* [CRTC:131ipe B] commit wait timed out
Nov 6 15:22:37 lander2 kernel: drmn0: [drm] *ERROR* flip_done timed out
Nov 6 15:22:37 lander2 kernel: drmn0: [drm] *ERROR* [CONNECTOR:195:HDMI-A-1] commit wait timed out
Nov 6 15:22:37 lander2 kernel: drmn0: [drm] *ERROR* flip_done timed out
Nov 6 15:22:37 lander2 kernel: drmn0: [drm] *ERROR* [CONNECTOR:202:HDMI-A-2] commit wait timed out
Nov 6 15:22:37 lander2 kernel: drmn0: [drm] *ERROR* flip_done timed out
Nov 6 15:22:37 lander2 kernel: drmn0: [drm] *ERROR* [PLANE:31lane 1A] commit wait timed out
Nov 6 15:22:56 lander2 seatd[13433]: 05:22:14.071 [INFO] [seatd/seat.c:625] Disabled client 2 on seat0
Nov 6 15:22:56 lander2 kernel: drmn0: [drm] *ERROR* flip_done timed out
Nov 6 15:22:56 lander2 kernel: drmn0: [drm] *ERROR* [PLANE:76:cursor A] commit wait timed out
Nov 6 15:22:56 lander2 kernel: drmn0: [drm] *ERROR* flip_done timed out
Nov 6 15:22:56 lander2 kernel: drmn0: [drm] *ERROR* [PLANE:82lane 1B] commit wait timed out
Nov 6 15:23:40 lander2 devd[72439]: check_clients: dropping disconnected client
Nov 6 15:37:34 lander2 seatd[13433]: 05:36:51.961 [INFO] [seatd/seat.c:563] Opened client 2 on seat0
...

Also related to the suspend problem:

Nov 6 16:44:21 lander2 kernel: vgapci0: child drmn0 requested pci_set_powerstate
Nov 6 16:44:21 lander2 kernel: pci0: failed to set ACPI power state D3 on \_SB_.PC00.GFX0: AE_BAD_PARAMETER
Nov 6 16:44:21 lander2 kernel: acpi0: cleared fixed power button status
Nov 6 16:44:21 lander2 kernel: vgapci0: child drmn0 requested pci_set_powerstate
Nov 6 16:44:21 lander2 kernel: vgapci0: child drmn0 requested pci_enable_io
Nov 6 16:44:21 lander2 syslogd: last message repeated 1 times
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Controller timeout
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ============== REGISTER DUMP ==============
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Sys addr: 0x00000000 | Version: 0x00001002
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Blk size: 0x00000000 | Blk cnt: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Argument: 0x00000000 | Trn mode: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Present: 0x1fff0000 | Host ctl: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Power: 0x0000000b | Blk gap: 0x00000080
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Wake-up: 0x00000000 | Clock: 0x0000fa07
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Timeout: 0x00000000 | Int stat: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Int enab: 0x01ff003b | Sig enab: 0x01ff003b
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: AC12 err: 0x00000000 | Host ctl2:0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Caps: 0x546ec881 | Caps2: 0x80000807
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Max curr: 0x00000000 | ADMA err: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ADMA addr:0x00000000 | Slot int: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ===========================================
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Controller timeout
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ============== REGISTER DUMP ==============
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Sys addr: 0x00000000 | Version: 0x00001002
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Blk size: 0x00000000 | Blk cnt: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Argument: 0x00000000 | Trn mode: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Present: 0x1fff0000 | Host ctl: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Power: 0x0000000b | Blk gap: 0x00000080
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Wake-up: 0x00000000 | Clock: 0x0000fa07
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Timeout: 0x00000000 | Int stat: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Int enab: 0x01ff003b | Sig enab: 0x01ff003b
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: AC12 err: 0x00000000 | Host ctl2:0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Caps: 0x546ec881 | Caps2: 0x80000807
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Max curr: 0x00000000 | ADMA err: 0x00000000
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ADMA addr:0x00000000 | Slot int: 0x00000001
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: ===========================================
Nov 6 16:44:21 lander2 kernel: sdhci_pci0-slot0: Controller timeout

I will make a bug report once I have more insight in this thread.

My questions are:
Am I right assuming this is related to the Alder Lake-N UHD graphics driver WIP ?
Do you think there is something I could do to make FreeBSD usable on this hardware while things progress? (like disabling HW acceleration in chrome, GPU, etc., directly going for FreeBSD 15 BETA ).
The most fancy thing I do is maybe watch youtube video in HD, so I could live with some feature disable.
If not I will have to format and go back to Linux for the time being because this is just unusable, I have to hard reboot every 10 minutes.

Many thanks for your help !

By the way, just sharing an opinion:
I always hear Wifi drivers are a big barrier to the adoption of FreeBSD on the desktop and never understood why.
If I need Wifi I have one of those $30 GL iNet mini access point running OpenWRT and an Ethernet cable, no big deal. But finding a board and GPU available on the market and that is well supported is much harder.

scottro · Thursday at 11:10 PM

I'll just answer the comment on wireless. I have an AX200 on a Beelink SER5 and it runs quite well, comparable to Linux speeds. Wireless is getting better and better, at present some cards only run at relatively slow speeds, but some, such as my AX200 are as functional as they are on Linux. (I don't have Windows, so I don't know if Linux wireless speed compares to Linux.

Espionage724 · Thursday at 11:18 PM

Freedom said:
Do you think there is something I could do to make FreeBSD usable on this hardware while things progress? (like disabling HW acceleration in chrome, GPU, etc., directly going for FreeBSD 15 BETA ).

FreeBSD 15.0 (possibly other versions) has drm-latest-kmod that might be better on newer UHD (works for UHD 630)

I have this in loader.conf:

Code:

hw.i915kms.modeset=1
hw.i915kms.nuclear_pageflip=1
hw.i915kms.fastboot=1
hw.i915kms.enable_psr=0
hw.i915kms.enable_psr2_sel_fetch=0
hw.i915kms.panel_use_ssc=0
hw.i915kms.enable_ips=0
hw.i915kms.enable_fbc=0
hw.i915kms.enable_dc=0
hw.i915kms.disable_power_well=0
hw.i915kms.enable_guc=3
hw.i915kms.guc_log_level=0
hw.i915kms.verbose_state_checks=0
hw.i915kms.error_capture=0

And in /usr/local/etc/profile.d/general-tweaks.sh (custom-created) I have this to disable graphics pipeline/GPL shaders (caused Dota 2 Vulkan to randomly crash Linux and FreeBSD):

Code:

# Intel
export ANV_NO_GPL='1'

I'm thinking atomic errors might be resolved with hw.i915kms.nuclear_pageflip=0. Flip if I understand right is usually related to Vsync, so theoretically the less apps/things trying to respect Vsync might have less-errors reported; I'm not sure how to do that with Wayland, but Xorg I force Intel's DDX (not default modesetting) and disable a bunch of Vsync stuff with a xorg.conf snippet:

Code:

Section "OutputClass"
    Identifier "Intel (intel custom)"
    MatchDriver "i915"
    Driver "intel"
    Option "AccelMethod" "SNA"
    Option "DRI" "3"
    Option "TearFree" "0"
    Option "VSync" "0"
    Option "PageFlip" "0"
    Option "SwapbuffersWait" "0"
    Option "TripleBuffer" "0"
    Option "LinearFramebuffer" "1"
    Option "Throttle" "0"
EndSection

Putting this into a similar profile.d file might work:

Code:

# Vsync
export vblank_mode=0
export MESA_VK_WSI_PRESENT_MODE=immediate

Freedom said:
If I need Wifi I have one of those $30 GL iNet mini access point running OpenWRT and an Ethernet cable, no big deal. But finding a board and GPU available on the market and that is well supported is much harder.

I used to do that with a full-sized router

(it worked pretty good!). I bought a AX210 card later specifically for FreeBSD, and it seemingly works good (faster speeds than previous AC 9560 card). I bought this card on Amazon for $17.

Freedom · Friday at 12:06 AM

Thanks ! that gives me a lot of path to explore and test

Of course I first tried the quick and dumb way, copied your full loader.conf What gave me an opportunity to learn about system rescue on FreeBSD...
So time to do it the slow/smart way and try to understand what I am doing

Charlie Brown · Friday at 3:26 AM

Welcome to FreeBSD, hope you'll have fun

If I'm not mistaken, suspend and resume is not working well on FreeBSD, for my part I don't use it because it will crash my session. Is it a laptop or a desktop that you install the bestie on ? which model ?

blackbird9 · Friday at 7:12 AM

I've run a desktop on an N100 mini pc with 14.3 release, but using X11, not wayland. Didn't have any problems. Although I use windowmaker and firefox, not chromium.

I suggest checking that you've installed the cpu microcode, as described here

Post in thread 'Intel N100 generates data corruption ?'

Oct 20, 2025

For anyone else reading this, running:-

# pkg install cpu-microcode-intel

will install the latest microcode level.

And in /boot/loader.conf you need:-

# update cpu microcode
cpu_microcode_load="YES"
cpu_microcode_name="/boot/firmware/intel-ucode.bin"

pebkac · Friday at 8:01 AM

I also at times had problems with Wayland (in my case: Hyprland) locking up from time to time.

Xorg has been stable for me though.

Maybe you want to check if Xorg behaves differently for you, too.

Freedom · Friday at 9:49 AM

Charlie Brown said:
Welcome to FreeBSD, hope you'll have fun
If I'm not mistaken, suspend and resume is not working well on FreeBSD, for my part I don't use it because it will crash my session. Is it a laptop or a desktop that you install the bestie on ? which model ?

Thanks !
I have 2 of those Chuwi Larkbox X 2023 https://www.chuwi.com/product/items/chuwi-larkbox-x-2023/specs.html
It was recommended on various forums and I have been using one of them as a FreeBSD home server for a year now to test.
I have been very happy about it (Linux Desktop and FreeBSD server)

bsd-hardware.info is down right now but I saw a lot of people run OPNsense on it.

Freedom · Friday at 9:50 AM

blackbird9 said:
I've run a desktop on an N100 mini pc with 14.3 release, but using X11, not wayland. Didn't have any problems. Although I use windowmaker and firefox, not chromium.

I suggest checking that you've installed the cpu microcode, as described here

Post in thread 'Intel N100 generates data corruption ?'

Oct 20, 2025

For anyone else reading this, running:-

# pkg install cpu-microcode-intel

will install the latest microcode level.

And in /boot/loader.conf you need:-

# update cpu microcode
cpu_microcode_load="YES"
cpu_microcode_name="/boot/firmware/intel-ucode.bin"

blackbird9

Yes

I saw this post while searching and while my microcode was a few months old I updated it yesterday.

Freedom · Friday at 10:29 AM

pebkac said:
I also at times had problems with Wayland (in my case: Hyprland) locking up from time to time.

Yes I started with Hyprland and when the freeze happened I immediately switched to Sway to be sure it wasn't compositor related.

pebkac said:
Xorg has been stable for me though.

Maybe you want to check if Xorg behaves differently for you, too.

This could be an option I just need Kitty and Chromium to feel comfortable. But looking at this forum I saw mostly people complaining of crash with this setup on X (probably because there are more Xorg users...)

Right now I am just trying to understand how the whole drm-kmod, modules and settings works, read dmesg and /var/log/messages so I can make changes and test methodically.
I am still confused about all the hw.i915kms.* options but I guess it is normal

blackbird9 · Saturday at 8:55 AM

I haven't tried a larkbox, I've got a couple of different generic mini-pc's. I think the chuwi kit usually has a good reputation though, from what I've read online.

blackbird9 · Saturday at 8:57 AM

Robonuggie has a nice introductory video here on setting up a freebsd desktop with X11, if that helps.

View: https://www.youtube.com/watch?v=UP1185OfCco

blackbird9 · Saturday at 9:00 AM

Not sure if this is the same machine you have but there are a couple of people reporting larkbox works with freebsd 13 and 14. That's a "larkbox X", not sure if that's the same as yours though. But that one has an N100 CPU.

Meet the LarkBox X 2023 - Revolutionizing Mini PCs

Enter the realm of future computing with the LarkBox X 2023, the compact powerhouse that is revolutionizing the world of mini PCs.

www.chuwi.com

https://bsd-hardware.info/?id=board:chuwi-innovation-and-technologyco-larkbox-x

blackbird9 · Saturday at 9:05 AM

I would suggest trying setting it up with X11 as per robonuggie's video, good luck!

Espionage724 · Saturday at 10:14 AM

Freedom said:
I am still confused about all the hw.i915kms.* options but I guess it is normal

A lot of it is disabling power-saving stuff, but hw.i915kms.enable_guc=3 might be different for newer GPUs (I heard something had problems loading GuC firmware); it's usually disabled/0.

I'm not sure if hw.i915kms.modeset=1 or hw.i915kms.fastboot=1 really do anything, but the rest looks relatively safe

Freedom · Saturday at 12:50 PM

blackbird9 said:
Not sure if this is the same machine you have but there are a couple of people reporting larkbox works with freebsd 13 and 14. That's a "larkbox X", not sure if that's the same as yours though. But that one has an N100 CPU.

Meet the LarkBox X 2023 - Revolutionizing Mini PCs

Enter the realm of future computing with the LarkBox X 2023, the compact powerhouse that is revolutionizing the world of mini PCs.

www.chuwi.com

https://bsd-hardware.info/?id=board:chuwi-innovation-and-technologyco-larkbox-x

Yes that is the one. Now be careful because the new version they sell got the N150

I set it up about a year ago with my first FreeBSD install as a home server. I just had to install the driver for the second realtek 2.5G NIC and never had any problem.
So I would say I would recommend it.
Good to know I never heard the fan spinning up on FreeBSD or Linux

Freedom · Saturday at 1:03 PM

Espionage724 said:
A lot of it is disabling power-saving stuff, but hw.i915kms.enable_guc=3 might be different for newer GPUs (I heard something had problems loading GuC firmware); it's usually disabled/0.

I'm not sure if hw.i915kms.modeset=1 or hw.i915kms.fastboot=1 really do anything, but the rest looks relatively safe

I understood the drm-61-kmod vs drm-66-kmod on FreeBSD 15 thing and installed FreeBSD BETA4 today since it is no use to debug an old version.

I still have the S3 wake up bug and freeze.

The acpi wake up :

vgapci0: child drmn0 requested pci_set_powerstate
acpi0: cleared fixed power button status
vgapci0: child drmn0 requested pci_set_powerstate
vgapci0: child drmn0 requested pci_enable_io
vgapci0: child drmn0 requested pci_enable_io
sdhci_pci0-slot0: Controller timeout
sdhci_pci0-slot0: ============== REGISTER DUMP ==============
sdhci_pci0-slot0: Sys addr: 0x00000000 | Version: 0x00001002
sdhci_pci0-slot0: Blk size: 0x00000000 | Blk cnt: 0x00000000
sdhci_pci0-slot0: Argument: 0x00000000 | Trn mode: 0x00000000
sdhci_pci0-slot0: Present: 0x1fff0000 | Host ctl: 0x00000001
sdhci_pci0-slot0: Power: 0x0000000b | Blk gap: 0x00000080
sdhci_pci0-slot0: Wake-up: 0x00000000 | Clock: 0x0000fa07
sdhci_pci0-slot0: Timeout: 0x00000000 | Int stat: 0x00000001
sdhci_pci0-slot0: Int enab: 0x01ff003b | Sig enab: 0x01ff003b
sdhci_pci0-slot0: AC12 err: 0x00000000 | Host ctl2:0x00000000
sdhci_pci0-slot0: Caps: 0x546ec881 | Caps2: 0x80000807
sdhci_pci0-slot0: Max curr: 0x00000000 | ADMA err: 0x00000000
sdhci_pci0-slot0: ADMA addr:0x00000000 | Slot int: 0x00000001
sdhci_pci0-slot0: ===========================================

...
repeat 14 times...
...

sdhci_pci0-slot0: ===========================================
sdhci_pci0-slot0: Controller timeout
sdhci_pci0-slot0: ============== REGISTER DUMP ==============
sdhci_pci0-slot0: Sys addr: 0x00000000 | Version: 0x00001002
sdhci_pci0-slot0: Blk size: 0x00000000 | Blk cnt: 0x00000000
sdhci_pci0-slot0: Argument: 0x00000000 | Trn mode: 0x00000000
sdhci_pci0-slot0: Present: 0x1fff0000 | Host ctl: 0x00000001
sdhci_pci0-slot0: Power: 0x0000000b | Blk gap: 0x00000080
sdhci_pci0-slot0: Wake-up: 0x00000000 | Clock: 0x0000fa07
sdhci_pci0-slot0: Timeout: 0x00000000 | Int stat: 0x00018000
sdhci_pci0-slot0: Int enab: 0x01ff003b | Sig enab: 0x01ff003b
sdhci_pci0-slot0: AC12 err: 0x00000000 | Host ctl2:0x00000000
sdhci_pci0-slot0: Caps: 0x546ec881 | Caps2: 0x80000807
sdhci_pci0-slot0: Max curr: 0x00000000 | ADMA err: 0x00000000
sdhci_pci0-slot0: ADMA addr:0x00000000 | Slot int: 0x00000001
sdhci_pci0-slot0: ===========================================
mmc0: No compatible cards found on bus

Freeze with qutebrowser

drmn0: [drm] *ERROR* CPU pipe A FIFO underrun: port,transcoder,
drmn0: [drm] *ERROR* [CRTC:80ipe A] flip_done timed out
WARNING new_crtc_state->event failed at /wrkdir/usr/ports/graphics/drm-66-kmod/work/drm-kmod-drm_v6.6.25_7/drivers/gpu/drm/drm_atomic_helper.c:2512
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [CRTC:80ipe A] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [CONNECTOR:185:HDMI-A-1] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [PLANE:31lane 1A] commit wait timed out
drmn0: [drm] *ERROR* flip_done timed out
drmn0: [drm] *ERROR* [PLANE:76:cursor A] commit wait timed out
drmn0: [drm] *ERROR* [CRTC:131ipe B] flip_done timed out

Fatal trap 12: page fault while in kernel mode
cpuid = 0; apic id = 00
fault virtual address = 0x2f
fault code = supervisor write data, page not present
instruction pointer = 0x20:0xffffffff80be37d5
stack pointer = 0x28:0xfffffe00af2c0c10
frame pointer = 0x28:0xfffffe00af2c0c90
code segment = base 0x0, limit 0xfffff, type 0x1b
= DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 82207 (python3.11)
rdi: fffff80004d0d000 rsi: 0000000000000004 rdx: ffffffff813a02d0
rcx: 0000000000000000 r8: fffffe00af2c0c50 r9: fffff802a854ea50
rax: 0000000000000007 rbx: fffff8019bd39580 rbp: fffffe00af2c0c90
r10: fffffe00eb0ac860 r11: 00000000000000c0 r12: 0000000000000001
r13: fffffe00af2c0ca0 r14: 0000000000000000 r15: 0000000000000000
trap number = 12
panic: page fault
cpuid = 0
time = 1762595509
KDB: stack backtrace:
#0 0xffffffff80bbeaad at kdb_backtrace+0x5d
#1 0xffffffff80b71e36 at vpanic+0x136
#2 0xffffffff80b71cf3 at panic+0x43
#3 0xffffffff8107af6f at trap_pfault+0x3cf
#4 0xffffffff81050e78 at calltrap+0x8
#5 0xffffffff80be333d at kern_poll+0xfd
#6 0xffffffff80be31d0 at sys_poll+0x50
#7 0xffffffff8107b8a6 at amd64_syscall+0x126
#8 0xffffffff8105176b at fast_syscall_common+0xf8
Uptime: 1h53m48s
Dumping 956 out of 11985 MB:..2%..11%..21%..31%..41%..51%..61%..71%..81%..91%---<<BOOT>>---

also some interesting lines in dmesg

acpi0: <ALASKA A M I >
acpi0: Power Button (fixed)
ACPI Error: No handler for Region [ECF2] (0xfffff800017a3000) [EmbeddedControl] (20250807/evregion-292)
ACPI Error: Region EmbeddedControl (ID=3) has no handler (20250807/exfldio-428)
ACPI Error: Aborting method \134_SB.PC00.LPCB.H_EC.BAT0._STA due to previous error (AE_NOT_EXIST) (20250807/psparse-689)

sdhci_pci0-slot0: ADMA addr:0x00000000 | Slot int: 0x00000001
drmn0: [drm] [ENCODER:184DI B/PHY B] is disabled/in DSI mode with an ungated DDI clock, gate it
i915/adlp_dmc.bin: could not load binary firmware /boot/firmware/i915/adlp_dmc.bin either
adlp_dmc.bin: could not load binary firmware /boot/firmware/adlp_dmc.bin either
i915_adlp_dmc.bin: could not load binary firmware /boot/firmware/i915_adlp_dmc.bin either
drmn0: successfully loaded firmware image 'i915/adlp_dmc.bin'
drmn0: [drm] Finished loading DMC firmware i915/adlp_dmc.bin (v2.20)
sysctl_add_oid: can't re-use a leaf (hw.dri.debug)!

So I will go back to testing every options you mentioned.
It is slow to test because the freezes happen a bit randomly: it can be ok for hours and then it freezes every 10 minutes.

Hopefully I get a solution or a good bug report out of this

freezr · Saturday at 4:16 PM

A lot of people here consider awkward using FreeBSD as desktop machine, I consider awkward using Wayland and all the other Linux pieces instead. If you need to use Wayland, and perhaps the latest mesa driver, you had better stick with GNU/Linux.

I would say that FreeBSD, especially for a desktop use, is good for everybody that thinks Debian is not conservative enough...

blackbird9 · Saturday at 5:55 PM

Freedom said:
Yes that is the one. Now be careful because the new version they sell got the N150

Yes I've noticed that with the mini pc's, they keep the same "brand / model name" but put a whole different motherboard and CPU inside. If you're lucky they call it version 2. Pretty much standard for shenzhen, still, that's how they get you the great selling price.

blackbird9 · Saturday at 6:00 PM

Freedom said:
Good to know I never heard the fan spinning up on FreeBSD or Linux

The fan probably is spinning, you can see the tacho reading in the bios, but at a slow speed and inaudible. My box works the same. I can only get audible fan noise if I make it do some heavy lifting like burning all cores at 100% at the same time.

A quick and easy cpu stress test is to run:-

$ pbzip2 -c </dev/urandom >/dev/null

- which will drive all available cores to 100% until you stop it with ctrl-c.
Of course that only burns cpu and doesn't test anything else (storage or graphics) but it's quick and easy to do. You can monitor the core temps while it's under load.

blackbird9 · Saturday at 6:05 PM

Another tip is to consider changing the heatsink paste. I've done that on a couple of them, for whatever reason they seem to like putting a huge lump of crappy cheap paste on in the factory, so you get poor heat transfer because it ends up too thick. I guess on a production line it's all about application speed, they're not going to spend minutes smoothing it out and getting it perfect. I've re-done the h/s paste on about 3 of them so far, I used arctic MX-4 because I had some lying around but any good quality paste should be fine. I think the factory fit stuff is the cheapest stuff they can get. The last time I did that I got temperature under load dropped by around 10-15 degrees C so it was well worth doing, and improves stability. Excess heat can have a big impact on these small boxes, and the cooling hardware is probably marginal already in such a small form factor. If you're getting random hardware freezes you may find upgrading the heatsink paste helps. Remember when you put the new paste on, you only need a thin layer on the back of the chip (use a plastic card), not a big dollop.

blackbird9 · Saturday at 6:18 PM

freezr said:
A lot of people here consider awkward using FreeBSD as desktop machine, I consider awkward using Wayland and all the other Linux pieces instead. If you need to use Wayland, and perhaps the latest mesa driver, you had better stick with GNU/Linux.

I would say that FreeBSD, especially for a desktop use, is good for everybody that thinks Debian is not conservative enough...

I think freebsd is fine for a desktop machine, it has everything I need available, with the possible exception of chrome (not chromium). Been using it with X11 for years.

Yeah don't know about wayland yet. Maybe I'll try it some time.

blackbird9 · Saturday at 6:24 PM

This video is a pretty good guide. I can't find one specifically on the larkbox X. I did find a couple of teardown videos for the larkbox X but they don't show you disassembling the heatsink from the motherboard (although, it should be pretty easy tbh).

View: https://www.youtube.com/watch?v=fXe4rGqEk1s

I would recommend giving it a try, for the sake of a few dimes worth of paste, you might find it makes the box a lot more stable. Software isn't always the problem. Random system freezes sounds like a possible thermal problem to me. If it doesn't work, well, it hasn't cost much to try it out.

blackbird9 · Saturday at 7:16 PM

Not really directly related but I found this video on beelink minipc production, very interesting. They have an impressive production process, it looks well thought out. The heatpipe assembly is placed at 15:10, it looks like they are using a pre-applied thermal tape (pad) on the heatpipe face rather than paste in this build, and fixture is done in a plastic jig. It would be good to check what's actually been used inside the larkbox X before taking it apart if possible, if they used tape too then changing it to paste may not help so much. The couple of machines I've taken apart so far have all used paste though.

View: https://www.youtube.com/watch?v=ohwI3V207Ts

Erichans · Saturday at 7:31 PM

Freedom said:

Rich (BB code):

...
drmn0: [drm] *ERROR* [CRTC:80:pipe A] flip_done timed out
WARNING new_crtc_state->event failed at /wrkdir/usr/ports/graphics/drm-61-kmod/work/drm-kmod-drm_v6.1.128_6/drivers/gpu/drm/drm_atomic_helper.c:2485

This suggests you might not be using the latest appropriate version of drm-61-kmod.
Please post the output of:
pkg -vv | sed -nE -e '/(OSVERSION|ABI)/ p' -e '/^Repositories:/,$ p'
pkg query -x '[%R] %o %n %v' '^drm-(61-|66-)?kmod'

P.S. I have no experience with anything Wayland.

EDIT:

Freedom said:
My questions are:
Am I right assuming this is related to the Alder Lake-N UHD graphics driver WIP ?

Not to my knowledge, rather the contrary for the N100. For the N150 the 'recognition' has been added: N150. Note that I'm no graphics specialist/developer; with that caveat, I think it is not necessary just for this issue to switch to FreeBSD 15.0.