Hi,
am using an RockPi 4 (one HDMI version) and wanted to try the actual version of DietPi.
But already on boot I got a problem. Console output is starting, then vertical stripes appears and the screen get black.
OS is booting up (SSH server is running). Tried to install Xcfe with graphical login but it changes nothing.
At least I can logon via ssh.
I already using Armbian with Bullseye (but they dont support RockPi 4 officially anymore) without problems.
Here some output from the system
root@DietPi:~# uname -a
Linux DietPi 5.10.63-rockchip64 #21.08.2 SMP PREEMPT Wed Sep 8 10:57:23 UTC 2021 aarch64 GNU/Linux
root@DietPi:~# dmesg | grep hdmi
[ 2.826708] vcc1v8_hdmi: supplied by vcc5v0_sys
[ 2.832148] vcc0v9_hdmi: supplied by vcc5v0_sys
[ 3.851982] dwhdmi-rockchip ff940000.hdmi: Detected HDMI TX controller v2.11a with HDCP (DWC HDMI 2.0 TX PHY)
[ 3.853658] rockchip-drm display-subsystem: bound ff940000.hdmi (ops dw_hdmi_rockchip_ops [rockchipdrm])
[ 3.929222] dwhdmi-rockchip ff940000.hdmi: PHY configuration failed (clock 148501000)
[ 7.987857] rc rc0: dw_hdmi as /devices/platform/ff940000.hdmi/rc/rc0
[ 7.987939] input: dw_hdmi as /devices/platform/ff940000.hdmi/rc/rc0/input10
root@DietPi:~#
An monitor I am using a Dell WQHD Monitor.
So whats the problem here ?
Thanks in advance.
Regards
MichaIng
can you have a look
Which Armbian image did you use? We use Armbian bootloader, kernel, dtb and firmware as well, so it should work the same way. But the mainline kernel version, not legacy. Also, can you check:
dmesg -l emerg,alert,crit,err
ls -l /dev/fb* /dev/dri
Hi,
with Armbian its a pretty good question ^^ since they stopped the official support.
I think it should be the
Armbian 21.08.1 Rockpi4a bullseye current with kernel 5.10.63
it also has some flaws with unhandled page fault in pamfrost. This is the reason I wanted to try DietPi.
Here the outputs
root@DietPi:~# dmesg -l emerg,alert,crit,err
[ 1.252553] reg-fixed-voltage vcc-0v9: Failed to register regulator: -517
[ 2.833279] vcc_cam: failed to get the current voltage: -EPROBE_DEFER
[ 2.834199] vcc_mipi: failed to get the current voltage: -EPROBE_DEFER
[ 3.127347] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.128062] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.128763] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.129462] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.130162] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.130890] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.131593] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.132293] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.132992] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.133691] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.134389] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.135336] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.136040] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.136743] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.137443] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.138145] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.138863] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.139659] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.140363] debugfs: Directory 'power-domain' with parent 'pm_genpd' already present!
[ 3.779414] rk_gmac-dwmac fe300000.ethernet: cannot get clock clk_mac_speed
[ 3.927584] dwhdmi-rockchip ff940000.hdmi: PHY configuration failed (clock 148501000)
root@DietPi:~# ls -l /dev/fb* /dev/dri
crw-rw---- 1 root video 29, 0 Feb 17 06:59 /dev/fb0
/dev/dri:
total 0
drwxr-xr-x 2 root root 100 Feb 17 06:59 by-path
crw-rw----+ 1 root video 226, 0 Feb 17 06:59 card0
crw-rw----+ 1 root video 226, 1 Feb 17 06:59 card1
crw-rw----+ 1 root render 226, 128 Feb 17 06:59 renderD128
Maybe a small hint: I have the old/first version of the rockpi4 with only one HDMI output.
Hope it helps.
Regards.
I see Armbian does not offer Bullseye images for ROCK Pi 4 anymore (A/B/C). I remember we made a distro upgrade, but the final result indeed should be the same. I’ll however have a look whether there are any non-default scripts or steps done by Armbian userland, which may change something.
This error caught my eyes, seems related:
[ 3.927584] dwhdmi-rockchip ff940000.hdmi: PHY configuration failed (clock 148501000)
We used the ROCK Pi 4 B image, btw, hence for 1 HDMI port, the C model has a micro HDMI and an additional display port (DP). I cannot imagine that regarding the HDMI port it makes any difference, but probably the display port does not work when using A/B variant image, at least when the bootloader does not pick the right device tree OOTB. I’ll have a look at these and the boot configs as well.
KMS/DRM is supported and there, as well as a framebuffer device, so that doesn’t seem to be the problem.
In case due to different CPU scheduler, have you tried another power supply. Alternatively could you try to change the CPU scheduler to “powersave” in dietpi-config > Performance Options to reduce power consumption?
Hi,
I checked what a RockPi4 I have and its a model A revision 1.4
That point with the HDMI ports was just an idea because I saw 2 “card” devices which confused me a little.
I changed the scheduler to “powersave” but with no effect.
Regards.
Hi,
I played around with another Armbian image from its archive.
Armbian_21.08.1_Rockpi-4a_bullseye_current_5.10.60
and it shows the same problem like the Dietpi image. I don’t know what Armbian image I am using now. Which runs but with some flaws.
I think I took the image for the Rockpi 4B by accident some months ago. But it has some page faults within panfrost but at least has HDMI output.
Regards.
Ah, I’ve overseen the Bullseye images from archive. I could at least generate a few one of ours based on ROCK Pi 4B Bullseye image. The panfrost driver is not required for console/HDMI output in general, and I see it often printing error messages on Armbian kernel, even when hardware accelerated GUI and video outputs work fine.
2 DRI devices are btw common, even if there is only a single HDMI.
Hi,
OK thanks for the information.
I think I got the reason. I removed the KVM switch between the RockPi and the monitor. With that after the yellow stripes the normal screen appears.
Don’t know why this is a problem began showing with bullseye.
I will setup the system to my normal setup and will check if I will get the same panfrost errors here. Problem is when the panfrost page faults occur my running VoiP session (via Twinkle oder Linphone) is broken and I have to redial.
But I think this ticket/issue here can be closed.
Thanks.
Regards.
Okay, great that it generally works again. No idea who the GPU panfrost driver, which is not really in use unless you use GUI applications, a desktop or such, can affect a VoIP session 
. Aren’t those errors showing up on early boot only?
Hi,
I had the “VoiP crash” also on DietPi now.
The interesting thing is that this occurs reliable after 30 or 45 minutes. The VoiP connection remains active but I get a loud noise and my companions cant here me any more.
When this occurs I get in journal
Mär 01 07:58:11 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46680
Mär 01 07:58:11 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46680
Mär 01 07:58:11 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46680
Mär 01 07:58:11 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46680
Mär 01 07:58:11 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46680
Mär 01 07:58:11 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46580
Mär 01 07:58:11 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46680
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46680
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46580
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46580
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46580
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D46680
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=0, config=0x3300, status=0x8, head=0x3198200, tail=0x3198200, sched_job=0000000016e30f12
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000104D465C0
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=0, config=0x3300, status=0x8, head=0x3149300, tail=0x3149300, sched_job=0000000011a4fa31
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1100
Mär 01 07:58:12 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1040
Mär 01 07:58:13 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:13 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:13 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:13 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:13 DietPi kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000106AD1300
Mär 01 07:58:13 DietPi kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=0, config=0x3300, status=0x8, head=0x311df00, tail=0x311df00, sched_job=00000000fac77f2c
...
I never got this with a Buster based image. Maybe I will switch back to a Buster image for work.
Pretty strange.
Regards.
Hi,
forget my last post. I tested VoiP on a Notebook and had the same problem after 45 minutes . So I think its not related to “panfrost” or DietPi.
Thanks and Regards