Official LE12 Test Images for Amlogic (Kodi-21)

diagus · March 6, 2024 at 5:01 PM

Quote from chewitt

Thanks, that points fingers to somethig changing in u-boot. The question is what ??

I don't know if this can help... Below, the diffs in the console booting LE11 (left) and LE12 (right):

Code

1c1
< =~=~=~=~=~=~=~=~=~=~=~= PuTTY log 2024.03.06 17:45:32 =~=~=~=~=~=~=~=~=~=~=~=
---
> =~=~=~=~=~=~=~=~=~=~=~= PuTTY log 2024.03.06 17:43:32 =~=~=~=~=~=~=~=~=~=~=~=
4c4
< TE: 300981
---
> TE: 563649
26c26
< 82 23 84 47 b5 54 7c 27 cb 26 69 12 [0.400291 Inits done]
---
> 82 23 84 47 b5 54 7c 27 cb 26 69 12 [0.660656 Inits done]
30c30
< Load bl33 from SD, src: 0x0003c000, des: 0x01000000, size: 0x0008eff0
---
> Load bl33 from SD, src: 0x0003c000, des: 0x01000000, size: 0x0009c550
43c43
< U-Boot 2022.10 (Jan 26 2024 - 19:01:04 +0000) wetek-play2
---
> U-Boot 2024.01 (Feb 25 2024 - 17:07:40 +0000) wetek-play2
47,49c47,49
< DRAM:  2 GiB
< Core:  164 devices, 26 uclasses, devicetree: separate
< MMC:   mmc@70000: 0, mmc@72000: 1, mmc@74000: 2
---
> DRAM:  1 GiB (effective 2 GiB)
> Core:  170 devices, 28 uclasses, devicetree: separate
> MMC:   mmc@70000: 2, mmc@72000: 0, mmc@74000: 1
51c51
< In:    serial
---
> In:    usbkbd,serial
55c55,58
< Net:   eth0: ethernet@c9410000
---
> [BL31]: tee size: 0
> Net:   Could not get PHY for ethernet@c9410000: addr -1
> No ethernet found.
> 
58,59c61,62
< mmc1 is current device
< Scanning mmc 1:1...
---
> mmc0 is current device
> Scanning mmc 0:1...
69c72
<    Data Size:    11576344 Bytes = 11 MiB
---
>    Data Size:    12580402 Bytes = 12 MiB
74a78
> Working FDT set to 8008000
76c80,81
<    Loading Device Tree to 000000003fff3000, end 000000003ffff639 ... OK
---
>    Loading Device Tree to 000000003fff3000, end 000000003ffff638 ... OK
> Working FDT set to 3fff3000
78c83
< Starting kerne ...
---
> Starting kernel ...
80a86
> [    0.585740] mdio_bus stmmac-0: MDIO device at address 0 is missing.

Display More

**chewitt** · March 7, 2024 at 10:09 AM

Not helpful

I need to bisect the upstream changes between u-boot 2022.10 and 2023.10 (already done 2023.10 to 2024.01). It's a slow process as beyond a certain point you need to add patches to create WP2 support.

**chewitt** · March 7, 2024 at 5:30 PM

diagus Ethernet issue resolved https://github.com/chewitt/u-boot…b9285bae8061bff

I'll send some patches and an apology to the mailing list and LE master branch soon but https://chewitt.libreelec.tv/testing/LibreE…ek-play2.img.gz contains the fix.

diagus · March 7, 2024 at 7:57 PM

Quote from chewitt

diagus Ethernet issue resolved https://github.com/chewitt/u-boot…b9285bae8061bff
I'll send some patches and an apology to the mailing list and LE master branch soon but https://chewitt.libreelec.tv/testing/LibreE…ek-play2.img.gz contains the fix.

Thanks chewitt. I will try it.

_emanuel_ · March 8, 2024 at 8:08 PM

current build works well on dreambox

diagus · March 9, 2024 at 8:45 AM

chewitt, tested your img on WP2 and the ethernet works.

Code

[    9.938572] meson8b-dwmac c9410000.ethernet eth0: Register MEM_TYPE_PAGE_POOL RxQ-0
[   10.127492] meson8b-dwmac c9410000.ethernet eth0: PHY [stmmac-0:00] driver [RTL8211F Gigabit Ethernet] (irq=29)
[   10.138801] meson8b-dwmac c9410000.ethernet eth0: No Safety Features support found
[   10.138870] meson8b-dwmac c9410000.ethernet eth0: PTP not supported by HW
[   10.141584] meson8b-dwmac c9410000.ethernet eth0: configuring for phy/rgmii link mode
[   13.671718] meson8b-dwmac c9410000.ethernet eth0: Link is Up - 1Gbps/Full - flow control off

**chewitt** · March 21, 2024 at 7:21 AM

I'm going to break some rules and push a kernel bump to Linux 6.8.y for LE12 b2 .. so there are 11.95.2 images in my test share.

dmladenov the S905L content is now p271 (all mention of p261 is gone) so you will need to update then edit uEnv.ini to change the dtb name. If you confirm it all still works (at least the essentials) I will send the device-tree upstream.

WP2 fixes are in the process of being upstreamed. I found some other minor things and am figuring out how best to do thing before sending a revised patchset.

I've added an Amlogic version of the "getedid" script to support devices turn on before TV's, for those who need it.

I've abandoned interest in supporting overlays. I'm not sure they're really needed and it complicates things.

U-Boot 2024.04 removes all Amlogic device-tree files (so we only maintain the kernel ones, which is great) but the change assumes that all dtb files live in an /amlogic folder in the root of boot media so I've changed the layout in our .img files. This drops a couple of patches and aligns us with Allwinner and Rockchip (also good). Updates work the same as always; the update script will check for dtb files in existing locations in addition to the new one and update them accordingly. No need for new installs or panic.

I've added support for the Vero 4K (again). It now has S/PDIF audio support and whoever it was who needed a p212 device with S/PDIF audio should be able to us it to get their no-name box working (can't find the thread.. never mind).

tomaszc · March 22, 2024 at 4:33 PM

chewitt Do you know anything new about the progress of the hardware decoder for Amlogic? (It was mentioned in the LE 11 thread and at the beginning of this thread)

I'm currently using the official test releases, if I upgrade to your version on the new U-Boot is it also possible to easily revert to the official version in case of problems?

**chewitt** · March 22, 2024 at 5:44 PM

There's nothing to report on hardware decoders. Development funded by Libre.Computer last year pursued a stateless video codec implementation (as it would eliminate closed-source firmware) but this ran into issues that need technical info from Amlogic (this is being negotiated) and as things took time; it also needs more funds to continue.

If you're already running LE12 nightlies there should be no issue to switch to the images in my fileshare and back.

dmladenov · March 22, 2024 at 6:44 PM

Quote from chewitt

dmladenov the S905L content is now p271 (all mention of p261 is gone) so you will need to update then edit uEnv.ini to change the dtb name. If you confirm it all still works (at least the essentials) I will send the device-tree upstream.

I updated with https://chewitt.libreelec.tv/testing/LibreE…h64-11.95.2.tar , dtb_name=/dtb/meson-gxl-s905l-p271.dtb and no sound like before.

**chewitt** · March 23, 2024 at 10:43 AM

dmladenov please update to https://chewitt.libreelec.tv/testing/LibreE…h64-11.95.2.tar .. is audio working?

dmladenov · March 23, 2024 at 11:07 AM

Quote from chewitt

dmladenov please update to https://chewitt.libreelec.tv/testing/LibreE…h64-11.95.2.tar .. is audio working?

no sound again

**chewitt** · March 23, 2024 at 11:53 AM

dmladenov can you run "cat /sys/kernel/debug/regmap/*adc*/registers | paste" and share the URL pls.

Then update to https://chewitt.libreelec.tv/testing/LibreE…h64-11.95.2.tar (again) and test; and if audio is working please run the ^ cat command again.

It's important that you dump register details first before updating the image.

dmladenov · March 23, 2024 at 1:55 PM

Quote from chewitt

dmladenov can you run "cat /sys/kernel/debug/regmap/*adc*/registers | paste" and share the URL pls.
Then update to https://chewitt.libreelec.tv/testing/LibreE…h64-11.95.2.tar (again) and test; and if audio is working please run the ^ cat command again.
It's important that you dump register details first before updating the image.

Before update: https://paste.libreelec.tv/together-piranha.log

After update: The audio is OK now https://paste.libreelec.tv/sweeping-possum.log

**chewitt** · March 23, 2024 at 2:39 PM

So my patching approach/logic is bad and I have to read "the idiots guide to programming in c" some more. You can expect some further testing requests

**chewitt** · March 23, 2024 at 3:03 PM

dmladenov same again, update and see if audio works, then dump registers: https://chewitt.libreelec.tv/testing/LibreE…h64-11.95.2.tar

dmladenov · March 23, 2024 at 3:11 PM

Quote from chewitt

dmladenov same again, update and see if audio works, then dump registers: https://chewitt.libreelec.tv/testing/LibreE…h64-11.95.2.tar

https://paste.libreelec.tv/deciding-aardvark.log

and the audio is OK

**chewitt** · March 23, 2024 at 3:32 PM

Quote from dmladenov

and the audio is OK..

It appears I've learned a new copy/paste skill

Can you run "pastekodi" so I get the whole log.