Amlogic-NO discussion

I use this one and no problem up to date.

dovi.ko (429.9 KB)

1 Like
[    0.277422@2]- ERROR: could not get clock /cvbsout:venci_top_gate(0)
[    0.277442@2]- cvbs_out: error: cvbsout_clktree_probe: clk venci_top_gate
[    0.277453@2]- ERROR: could not get clock /cvbsout:venci_0_gate(1)
[    0.277462@2]- cvbs_out: error: cvbsout_clktree_probe: clk venci_0_gate
[    0.277473@2]- ERROR: could not get clock /cvbsout:venci_1_gate(2)
[    0.277481@2]- cvbs_out: error: cvbsout_clktree_probe: clk venci_1_gate
[    0.277491@2]- ERROR: could not get clock /cvbsout:vdac_clk_gate(3)
[    0.277500@2]- cvbs_out: error: cvbsout_clktree_probe: clk vdac_clk_gate
[    0.443003@2]- ERROR: could not get clock /amhdmitx:hdmi_vapb_clk(3)
[    0.443026@2]- ERROR: could not get clock /amhdmitx:hdmi_vpu_clk(4)
[    0.443039@2]- ERROR: could not get clock /amhdmitx:venci_top_gate(0)
[    0.443050@2]- ERROR: could not get clock /amhdmitx:venci_0_gate(1)
[    0.443060@2]- ERROR: could not get clock /amhdmitx:venci_1_gate(2)
[    0.615175@2]- aml_iomap: write cbus reg f2 error -1
[    0.615230@2]- vlock clk not cfg
[    0.615599@2]- aml_iomap: read cbus reg 20 error -1
[    0.616230@2]- aml_iomap: read cbus reg 21 error -1
[    0.627672@2]- aml_iomap: read cbus reg d1 error -1
[    0.627703@2]- aml_iomap: write cbus reg d1 error -1
[    0.628209@2]- aml_iomap: read cbus reg d1 error -1
[    0.628816@2]- aml_iomap: write cbus reg d1 error -1
[    0.791787@2]- request_irq error ret=-22
[    0.791842@2]- dev_pm_set_wake_irq failed: -22
[    0.792227@2]- input: input_btrcu as /devices/platform/bt-dev/input/input2
[    0.792752@2]- meson-saradc fe026000.saradc: set delay per tick to <1ms> by default.
[    0.792756@2]- meson-saradc fe026000.saradc: set ticks per period to <1> by default.
[    0.793597@2]- dmc_monitor_probe
[    0.794094@2]- meson-mmc: emmc: adj_win: < 0 1 2 3 4 >
[    0.794097@2]- meson-mmc: _aml_sd_emmc_execute_tuning() d1_dly 0, window start 0, size 5
[    0.794101@2]- meson-mmc: emmc: clk 199999997 tuning start
[    0.798894@2]- aml_aucpu aucpu: Direct firmware load for aucpu_fw.bin failed with error -2
[    0.798899@2]- load aucpu_fw.bin fail
[    0.798900@2]- load start_aucpu_fw fail
[    0.799139@2]- socdata fe010000.soc_info: socdata_devno:1f200000
[    0.799423@2]- bl301_manager: driver init
[    0.799770@2]- usbcore: registered new interface driver snd-usb-audio
[    0.800170@2]- es8388_i2c_probe() spk_ctl_gpio < 0
[    0.800173@2]- es8388 pdn pin(4294967295) is invalid
[    0.800317@2]- es8388 probe error
[    0.798894@2]- aml_aucpu aucpu: Direct firmware load for aucpu_fw.bin failed with error -2
[    0.798899@2]- load aucpu_fw.bin fail
[    0.798900@2]- load start_aucpu_fw fail
[    0.799139@2]- socdata fe010000.soc_info: socdata_devno:1f200000
[    0.799423@2]- bl301_manager: driver init
[    0.799770@2]- usbcore: registered new interface driver snd-usb-audio
[    0.800170@2]- es8388_i2c_probe() spk_ctl_gpio < 0
[    0.800173@2]- es8388 pdn pin(4294967295) is invalid
[    0.800317@2]- es8388 probe error
[    0.800903@2]- aml_codec_T9015 fe01a000.t9015: aml_T9015_audio_codec_probe
[    0.800922@2]- T9015 acodec used by auge, tdmout:1
[    0.802498@2]- asoc debug: aml_audio_controller_probe-129
[    0.802974@2]- aml_tdm_platform_probe, tdm ID = 1, lane_cnt = 4
[    0.802985@2]- TDM id 1 samesource_sel:3
[    0.802994@2]- snd_tdm fe330000.audiobus:tdm@1: lane_mask_out = 1, lane_oe_mask_out = 0
[    0.803176@2]- TDM id 1 output clk enable:1
[    0.803196@2]- No channel mask node Channel_Mask
[    0.803208@2]- TDM id 1 tuning clk enable:1
[    0.803258@2]- aml_tdm_platform_probe, tdm ID = 2, lane_cnt = 4
[    0.803267@2]- TDM id 2 i2s2hdmi:1
[    0.803273@2]- snd_tdm fe330000.audiobus:i2s2hdmi: lane_mask_out = f, lane_oe_mask_out = 0
[    0.803340@2]- snd_tdm fe330000.audiobus:i2s2hdmi: neither mclk_pad nor mclk2pad set
[    0.803349@2]- snd_tdm fe330000.audiobus:i2s2hdmi: aml_tdm_get_pins error!
[    0.803352@2]- No channel mask node Channel_Mask
[    0.803837@2]- aml_spdif_platform_probe, spdif ID = 0
[    0.804204@2]- Spdif id 0 tuning clk enable:1
[    0.804245@2]- aml_spdif_platform_probe, spdif ID = 1
[    0.804260@2]- snd_spdif fe330000.audiobus:spdif@1: spdif 1 has no pinctrl!
[    0.804816@2]- audio-ddr-manager fe330000.audiobus:ddr_manager: 0, irqs toddr 24, frddr 28
[    0.804828@2]- audio-ddr-manager fe330000.audiobus:ddr_manager: 1, irqs toddr 25, frddr 29
[    0.804839@2]- audio-ddr-manager fe330000.audiobus:ddr_manager: 2, irqs toddr 26, frddr 30
[    0.804849@2]- audio-ddr-manager fe330000.audiobus:ddr_manager: 3, irqs toddr 27, frddr 31
[    0.805122@2]- 	datain_src:4, datain_chnum:2, datain_chumask:3
[    0.805124@2]- 	datalb_src:1, datalb_chnum:2, datalb_chmask:3
[    0.805125@2]- 	datain_lane_mask:0x1, datalb_lane_mask:0x1
[    0.805127@2]- datalb_format: 1, chmask for lanes: 0x3
[    0.805354@2]- loopback_platform_probe, p_loopback->id:0 register soc platform
[    0.805631@2]- audiolocker_platform_probe
[    0.805636@2]- ERROR: could not get clock /locker:lock_in

Post full logs & description of the problem

1 Like

How you think we should check it without a sample?
Ask the magic glas ball?

1 Like

There are FBO HDR GUI changes ongoing in nigthlies.
Just try every nightly again if it get better or not and post logs!

@Portisch Was anything ever found regarding this odd crash which clearly started occurring after the kernel was upgraded? I haven’t tried any recent builds since that.

The HDMI DRM handling was reworked and system should be more stable again.

4 Likes

I will try this dovi.ko file again with the latest NO build tonight. If it still doesn’t work, I will post the logs along with a sample file.

37 posts were merged into an existing topic: CoreELEC 22.0-Piers_alpha3 Discussion

Amlogic S905X4 Android 11.0 HK1 Box X4 4GB RAM 64GB ROM Dual-WiFi 2.4GHz/5GHz BT Quad Core 64 Bits 3D/8K 1000M

Updating NO when Nightly available for update.

Currently running
CoreELEC-Amlogic-no.aarch64-22.0-Piers_nightly_20260418-Generic.img.gz

Maybe it is me but typically when I enable reboot after download of nightly; it never reboots, it freezes and I just unplug device and plug it back in.

Am I doing something wrong here and not waiting long enough for the reboot?

Same here. I have had reboot take 10mins. I now have my Tox3’s on smart switches.

“Alexa Reebot CE”

I use it for crashes and box freezes too :slight_smile: Baz

1 Like

Hi,

I have a Ugoos AM9 Pro → Sony X90K with the nightly build on it. I’m having a problem getting any screensavers that play the ATV Aerial video to work. I initially tried the Aerial screensaver directly and have also now tried Video Screensaver and both have the same problem. If I try to play the Aerial videos (which are available over smb on my network and locally) by browsing to them and playing them, they work flawlessly. If I set them playing from a screensaver (by hitting the preview button in the screensaver settings) the TV shows either a blue lined corrupt display or a bright red corrupt display.

I’ve tried all sorts of combinations of switching display/refresh/resolution settings on and off but I don’t seem to be able to find any combination that works. It definitely feels refresh rate switching related.

I have noticed that last night’s build changed the behaviour of screensaver preview from “crash the box, needs a cold start” to “recovers back to the UI resolution and displays OK when I hit back” - if that’s any help.

Thanks for your time.

There is still something wrong about HDMI mode change. Let’s see when it get fixed…

3 Likes

[BUG REPORT] Dolby Vision black screen when subtitle track has ID 0 in MKV container

Device: Ugoos SK1
CoreELEC version: CE-NO 22.0-Piers nightly 20260423
Kernel: 5.15.196


Bug description:

Some DV files show black screen with audio only. HDR10 files play correctly regardless of track order.

After investigation, the root cause was identified:

When a subtitle track has ID 0 in the MKV container (appearing before the video track), the Dolby Vision pipeline fails to initialize and shows black screen with audio only.

Files with video track at ID 0 play DV correctly.
Files with subtitle track at ID 0 (pushing video to ID 1 or higher) show black screen.
HDR10 files are NOT affected by this issue, regardless of track order.


How to reproduce:

  1. Take any DV MKV file where subtitle track has ID 0 (before video track)
  2. Play in Kodi on CE-NO 22 nightly → black screen with audio only
  3. Remux the same file with MKVToolNix to reorder tracks so video is first (ID 0)
  4. Play again → DV works correctly

Workaround:
Remux affected files with MKVToolNix to place video track first.

Root cause theory:
The DV pipeline appears to read only track ID 0 for DV metadata initialization. When track ID 0 is a subtitle instead of video, DV init fails silently.


Hope this helps the team fix it.

Did you try remuxing an affected file WITHOUT reordering the tracks?

Was it the track order or simply the remux that fixed it?

Update: I tested this further using the original .bak files.

Confirmed: it is specifically the track order that causes the issue.

  • Remuxed WITH reordering (video at lowest ID) → DV works correctly
  • Remuxed WITHOUT reordering (video NOT at lowest ID) → DV black screen

To be more precise: DV fails whenever the video stream does not have the lowest track ID in the container. It does not matter what track type has ID 0 specifically — as long as video is not at the lowest ID, DV init fails.

For reference, dovi.ko loads with version warnings but DV init reports success:

[    0.750286] dovi disable in uboot
[   14.235329] dovi_gen_5_15_stb26: disagrees about version of symbol module_layout
[   14.235345] please rebuild ko: dovi_gen_5_15_stb26
[   14.235354] dovi_gen_5_15_stb26: module license 'AML' taints kernel.
[   14.236052] dovi_gen_5_15_stb26: disagrees about version of symbol __stack_chk_fail
[   14.236062] please rebuild ko: dovi_gen_5_15_stb26
[   14.236748] dovi_gen_5_15_stb26: using CFI init symbol __cfi_jt_init_module
[   14.236754] dovi_gen_5_15_stb26: using CFI exit symbol __cfi_jt_cleanup_module
[   14.237175] [dovi_gen_5_15_stb26]: *** amlogic_dolby_vision_init dv: gen ***
[   14.237383] [dovi_gen_5_15_stb26]: Creating DV mp success
[   14.237604] [dovi_gen_5_15_stb26]: Creating DV mp success

Hope this helps narrow down the fix.

And still no sample…

Narrows down only some potential causes, I can’t think of anything in Kodi that would look specifically for track id 0 for video since everything passes from ffmpeg.

I also can’t think of anything in CoreELEC that would be bypassing Kodi to get video information directly from the file.

Time for a sample, create a test clip cut from an affected file, verify it still exhibits the described behavior for you then upload it to mega.nz

(Other file sharing would be acceptable but much more convenient from megz.nz)

ofcouse here is the sample 298.87 MB folder on MEGA

Crashing seems to have got much worse on the latest nightly.

1 Like