Comparing two dtbs, I see one potential issue:
ethernet-phy@0 in 20250117 lacks reg=<0> property although mdio@0 node explicitly indicates 1 address-cell.
But I currently cannot tell if the dtb I am comparing against had a working eth.
It’s auto detected as some devices don’t have 0.
Hmm, adding the reg=<0> manually to the ethernet phy node does not help. I will go through the CoreELEC Amlogic versions I tried before to get a confirmation where eth was working.
Sigh, I double checked the PCB - it also lacks the external eth phy. Will try the 4g (100M) dtb
I have just try latest build 20250120 and i must say its much better, i do see some small statters here and there, but its like 98% better then before. Thanks Portisch and entire team for hard work fixing all bugs
1 Like
A95X Max S905x2 amlogic chip
4gb ram, 1gb ethernet
This box runs Omega 21.2 amlogic-ng fine, no issues
Tried to run latest nightly (Piers 22 alpha), but cannot get device to boot from sdcard. Just get the A95X logo on a boot loop. This same box boots all previous versions of CoreELEC just fine, but cannot get the amlogic-no v22 to boot at all.
I looked at page coreelec:ce_dev_cycle [CoreELEC Wiki] and does suggest that S905x2 SOC is supported for amlogic-no.
Used the same device tree name that is used for Omega 21.2. Is there anything else I can provide?
Is this a known issue?
Many thanks
Yes, UART log.
No, there are more S905X2 devices working with g12a_s905x2_4g.
https://paste.coreelec.org/MorrisonFarts
Some errors in the log
2025-01-22 15:19:57.713 T:1423 info <general>: ## CoreELEC Addon ## _service_::run ## MESSAGE:openConfigurationWindow
2025-01-22 15:19:57.717 T:1355 debug <general>: PushCecKeypress - received key b duration 70
2025-01-22 15:19:57.733 T:1429 debug <general>: Python interpreter stopped
2025-01-22 15:19:57.733 T:1429 debug <general>: Thread LanguageInvoker 545819455040 terminating
2025-01-22 15:19:57.746 T:1267 debug <general>: ------ Window Init (/usr/share/kodi/addons/service.coreelec.settings/resources/skins/skin.estuary/xml/service-CoreELEC-Settings-mainWindow.xml) ------
2025-01-22 15:19:57.746 T:1267 info <general>: Loading skin file: /usr/share/kodi/addons/service.coreelec.settings/resources/skins/skin.estuary/xml/service-CoreELEC-Settings-mainWindow.xml, load type: LOAD_ON_GUI_INIT
2025-01-22 15:19:57.768 T:1267 error <general>: Control 1000 in window 13001 has been asked to focus, but it can't
2025-01-22 15:19:57.768 T:1423 info <general>: ## CoreELEC Addon ## _service_::run ## WAITING:
2025-01-22 15:19:58.203 T:1284 debug <general>: Sink changed
...
...
...
2025-01-22 15:20:08.050 T:1355 debug <general>: PushCecKeypress - received key b duration 70
2025-01-22 15:20:08.272 T:1267 debug <general>: ------ Window Deinit (DialogSelect.xml) ------
2025-01-22 15:20:08.326 T:1267 debug <general>: ------ Window Init (/usr/share/kodi/addons/service.coreelec.settings/resources/skins/skin.estuary/xml/service-CoreELEC-Settings-mainWindow.xml) ------
2025-01-22 15:20:08.327 T:1267 info <general>: Loading skin file: /usr/share/kodi/addons/service.coreelec.settings/resources/skins/skin.estuary/xml/service-CoreELEC-Settings-mainWindow.xml, load type: LOAD_ON_GUI_INIT
2025-01-22 15:20:08.339 T:1267 error <general>: Control 1000 in window 13000 has been asked to focus, but it can't
2025-01-22 15:20:08.566 T:1267 debug <general>: ------ Window Init (/usr/share/kodi/addons/service.coreelec.settings/resources/skins/skin.estuary/xml/service-CoreELEC-Settings-mainWindow.xml) ------
2025-01-22 15:20:08.598 T:1284 debug <general>: Sink changed
Do you know if this device coud do the job of collecting UART log?
CH343P USB To Serial Port Module 6Mbps 5V3.3V TTL UART CH343 Debugging Burning And Downloading Type-C Mini Switch Instead CH340 - AliExpress 502
Are there any instructions on how to set this up and collect the log?
Thx
Yes, some tool like this.
And what’s the issue?
No issue.
Just for your information.
Will be visible again in next nightly.
4 Likes
Latest release from 23.01 have problem that after some time of video playback, video gets stuck (freeze) audio is playing in background but whole interface lockup and only reboot helps
Edit:
Build from 24.01 bring back clock in upper right corner, and videos dont freeze when watching content, i just watched entire movie without problem.
I am having an issue with my Remote (CE default) on my Harmony with NO.
Win media player remote I think as that works out of the box also OK on NG/NE but similar issues on NO.
Not always acting on button presses.
Same remote works faultlessly in NG/NE.
No remote files used in Remote Config.
Is there a way to copy default remote from NG/NE to NO?
Is it embedded in the program or in the DTB?
I was reading that the CoreElec Amlogic-ne would be coming to an end and I wanted to get Retroarch working in CoreElec for my generic S905W2 box so thought I would test the nightly build (CoreELEC-Amlogic-no.aarch64-22.0-Piers_nightly_20250128-Generic) of the Amlogic-no development.
After installing on micro SD and copying over the correct dtb (s4_s905w2_2g.dtb) and renaming dtb.img I power on the box get the boot screen that always appears for the box. The screen then goes black and I get Unsupported Video Signal on the TV.
I have also been trying to get EmuElec (RETROFENIX-Amlogic-ne.aarch64-4.8-EmuELEC_devel_20250122163755-Generic) to boot on this box as well and I get the same thing.
The dmesg for the nightly build with Unsupported Video Signal is: Amlogic-no nightly build dmesg
The dmesg from the working CoreElec (CoreELEC-Amlogic-ne.aarch64-21.1.1-Omega-Generic) is: Amlogic-ne working system dmesg
Edit: Turns out the new CoreElec and EmuElec didn’t like my TV. I tested on a different, newer TV and both came up. The offending TV is an older Toshiba 32" and the newer one is an Insignia 43" (not sure the exact models of them at this moment).
20250128 is buggy, my videos gets constant freezes and interuptions, im going back to older release
Unfortunately archive.coreelec.org is not working so cant download older release
Thx, i tried many version where was first success booting on am6plus with newer kernel 5.15.137 (from 20241231). So in all cases was bad hw acceleration of late identificated problematic video formats at CE22NO. HW acceleration for am6plus with all supported video formats (also streams) is good only on older kernel 5.15.119 (to 20241220) at CE22NO. Maybe kernel 5.15.137 need compatible/newer/better driver from Mali Mali Drivers | Mali GPU User-Space Binary Drivers – Arm Developer . Easy solution is maybe using identical mali drivers in 5.15.137 from 5.15.119 (for confirmation of reason of current issues, if was used now different version).
Libreelec have too old driver r16 with fix for kernel 5.5, 5.8, 5.9 and 5.11: GitHub - cdu13a/mali-bifrost: ARM Mali kernel drivers used with Bifrost GPU's in LibreELEC or r23: GitHub - chewitt/mali-bifrost at BX301A01B-SW-99002-r23p0-01rel0-amlogic Last opensource kernel driver is from august 2024 as r51: Mali Drivers | Open Source Bifrost Mali 3rd Gen GPU Architecture Kernel Drivers – Arm Developer
There is supporting to kernel 6.7.
Also for 5.15 are changes in last drivers mali according to file version_compat_defs.h:
#if (KERNEL_VERSION(5, 15, 0) > LINUX_VERSION_CODE)
uint uint_val;
int ret;
if (!val)
return -EINVAL;
ret = kstrtouint(val, 0, &uint_val);
if (ret == 0) {
if (uint_val < min || uint_val > max)
return -EINVAL;
*((uint *)kp->arg) = uint_val;
}
return ret;
#else
return param_set_uint_minmax(val, kp, min, max);
#endif
1 Like