**tps://mega.nz/file/g89zCabQ#4OJxRG6tCn4hnPRPaYi_HTJlYnL2BQV6ppCHrvX10zY
**=ht
It sems that I am not permitted to upload links??
I also changed dr_mode to “otg” at line 2983 , but with no difference as to getting a functioning mouse/keyboard.
New mesa-git and latest linux in unstable is looking good for G52.
Do test it out and share the feedback
Installed the latest linux-5.10-16-1 (replacing linux-vim-5.10.8-1). No more GPU Sched timeout error and GPU: Unhandled Page Fault but glmark2-es2 score is much lower lower now.
Good news is no visual glitches.
LibreOffice Impress 7.0.4.2 still crashes on Slide Show mode with linux-5.10.16-1.
Edit: With CPU performance set to “Performance”, glmark2-es2 improved significantly.
Testing linux-5.10.16-1 and latest mesa-git from Unstable Branch. I noticed these errors in dmesg. Is there anything of concern?
[ 520.014616] panfrost ffe40000.gpu: AS_ACTIVE bit stuck
[ 643.023693] panfrost ffe40000.gpu: AS_ACTIVE bit stuck
[ 1073.109205] ieee80211 phy0: brcmf_proto_bcdc_query_dcmd: brcmf_proto_bcdc_msg failed w/status -110
[ 1073.109224] ieee80211 phy0: brcmf_cfg80211_get_station: GET STA INFO failed, -110
@spikerguy
I just tried “for_linux-vim-5101.tar u-boot.ext” use it on SD Card, with SD Card indeed it can boot up with another USB Flash drive or USB HDD attached to GT King Pro USB port.
Just to give a complete picture of the “for_linux-vim-5101.tar u-boot.ext” ability to boot up on Beelink GT King Pro.
-
Boot image on SD Card, it can boot up successfully with USB Flash or USB HDD connected to other USB 3.0 port.
-
Boot image on USB Flash drive, it can only boot up successfully if USB Flash Boot image inserted on USB OTG port but without any other USB drive connected to any of the USB 3.0 ports.
-
Boot image on USB Flash inserted on a USB 3.0 Hub and the USB 3.0 Hub connected to any of the USB 3.0 ports but No other USB drive connected any of the other USB ports. With this method, the USB Flash image drive will heat up fast on the USB 3.0 Hub.
-
Boot image on USB Flash drive, it will NOT boot up if it is inserted into any of the USB 3.0 directly.
-
Boot image on USB Flash drive, it will not boot up even if inserted on USB 3.0 Hub and USB 3.0 Hub connected to USB OTG port. Infinite Boot Loop.
2 Likes
Hi @TheMojoMan nice job on getting rtl8822cs up. I have installed manjaro on my x96 max plus. I have compiled module and insmoded it, now in ip a i have wlan0, but it do not see any nwtworkes, any idea?
EDIT: Sorry it was my bad, wrong dtb
1 Like
Burned a fresh Manjaro-VIM3-XFCE 20201220 on a SD Card with linux-vim 5.9.8.2, to avoid kernel panic on GT King Pro, first I install the new boot-vim3 20.10-1 to use the new u-boot.ext which also convert the boot up using extlinux.conf instead of uEnv.ini.
Then “sudo pacman -Syu”, system update to linux-vim 5.10.5-1, just for your info the sound is loss after update to linux-vim 5.10.5-1, the issue is due to the g12_sound.sh Amlogic New Conf being used. Edit to revert to Amlogic Old Conf, and system enable sound and reboot sound return.
Subsequently install "sudo pacman -U http://mirror.easyname.at/manjaro/arm-unstable/core/aarch64/linux-5.11.0-3-aarch64.pkg.tar.zst http://mirror.easyname.at/manjaro/arm-unstable/extra/aarch64/mesa-git-21.1.0_devel.135076.a740318e425-1-aarch64.pkg.tar.zst to install the latest linux 5.11.0-3 and mesa-git from Unstable Branch.
Had errors:
error: failed to commit transaction (conflicting files)
linux-firmware: /usr/lib/firmware/brcm/brcmfmac4356-pcie.bin exists in filesystem
linux-firmware: /usr/lib/firmware/brcm/brcmfmac4356-pcie.gpd-win-pocket.txt exists in filesystem
linux-firmware: /usr/lib/firmware/brcm/brcmfmac4356-sdio.bin exists in filesystem
Errors occurred, no packages were upgraded.
But resolved by renaming the existing conflicting brcmfmac4356 files and kernel update to linux-5.11.0-3 and mesa-git without further issue.
Performance improved by following the tip (‘xfconf-query -c xfwm4 -p /general/vblank_mode -s xpresent’) from Panfrost Status on Odroid N2 - Page 4 - ODROID
- XFCE without compositing seems to have regressed and also while doing the test noticed that with the recent mesa build the default compositing of xfwm4 which uses glx doesn’t works properly so to fix it I used xpresent
xfconf-query -c xfwm4 -p /general/vblank_mode -s xpresent, the positive news is that now xfce with xpresent compositing is faster than running without compositing.
Good news is since linux-5.10-16-1 and now linux-5.11-0-3, the Panfrost visual glitches are gone.
LibreOffice Impress 7.0.4-3 still crashes on Slide Show Mode with Xfce Compositor “Enabled”.
1 Like
Let me first say that I am by no means an expert of device trees. I have just gained some limited knowledge by reading the DeviceTree Specification Release v0.3 and by experimenting with .dtbs.
If you want to fix the problem with the USB ports of your box be aware that you have to invest a lot of time and do the experiments yourself. You might gain a lot of knowledge on that journey which can be fun and very rewarding if you succeed…
I had a look in the Android .dts and tried to compare to the ugoos.dts. The following image might get you a jumpstart how to approach the problem:
Just check whether is there a way to include CPU or GPU Temperature on “HTOP” similar to the HTOP in Armbian ?
First of all I really appreciate your detailed answer and the tips you gave me. Also I would like to thank you for the time you invested to do such a helpfull post.
I agree with you that sometimes what is worthwhile is: “Its not the Destination, It’s the journey.”
Thanks for your nice feedback! I hope my answer will help you.
I had another look in the ‘Device Tree Spec’ document and the numbers after ‘interrupts’ keyword are indeed values and not pointers. Have corrected the image accordingly and reuploaded.
I tried to “blindly” make some changes according to your tips. I don’t know if I did something completely wrong but the test “dtb” file doesn’t boot at all. I am attaching the relevant dtb&dts files for you here.
I am eager to test any file you think would be in the right direction, even based on “wild” guesses as it is obvious that you are more experienced at this subject than me. So if you think it is not much trouble for you to create such file (you don’t even have to convert them to dtbs-even in dts format is enough) I will be glad to try them.
Thank you again for your time.
**tps://mega.nz/file/l48W2ZIQ#LnouvC4Cdqai_hqvQMMou5AOa-64tjmlTQG8fdE-JbE
**tps://mega.nz/file/h91GALbb#acYD17h6gR8DcmeTM-_qZwDVp1xl5RLpH4zjNKbowxo
**=ht
Obviously, you should only change one value at a time, compile a new .dtb and test if it’s still working. I did the mistake myself to change too many things at once , then .dtb was not booting and I did not know what was wrong. 
Obviously you are correct at this. 
Do you have a tip what you think to change first?(most needed changes/changes in priority?). Also I think that if there is a way to enable only 1 USB port it will be enough, at least for debugging purposes(or it is both or nothing?).
Sorry, no general tips because the cause why USB is not working is unknown.
It is possible to delete non critical parts of the dtb. That way you can try to reduce the complexity of the problem and then later add parts again to the .dtb.
Also try the following:
Delete lines 1915 - 2219 in ugoos.dts. These are directly following USB definitions and are responsible for ethernet and audio (obviously ethernet and audio will not work after boot). It is good to check that there is no interplay between these components (I once had these problems which I described in one of my posts in this thread).
If box then does not boot anymore:
- Only delete lines 1949 - 1955 and see if box boots and check if USB works.
- Try also deleting lines 1915 - 1957. Again: see if box boots and check if USB works.
No, armbian maintains a custom version of htop while we use vanilla version which is coming from arch linux arm.
Deleting lines 1915-1957 gives me an error when trying to convert from dts to dtb. Also it doesn’t seem to me that this is logical? (meaning that it doesn’t seem to be a complete part/portion/branch of what you say we are trying to delete). Maybe we have different versions of ugoos.dts file and therefore different line numbers???
If you wish you could give me 2-3 different versions of ugoos.dts (with deleted parts)and I can convert them to dtbs and test them.