PiotrO

Hmm - interesting: in my case on 5.3-rc8 it looks I need to look in: /sys/devices/system/cpu/cpufreq/policy0 And indeed i.e.: cat /sys/devices/system/cpu/cpufreq/policy0/stats/trans_table From : To : 480000 720000 816000 888000 1080000 1320000 1488000 1800000 480000: 0 0 0 0 0 0 0 0 720000: 0 0 0 0 0 0 0 0 816000: 0 0 0 0 0 0 0 0 888000: 0 0 0 0 0 0 0 0 1080000: 0 0 0 0 0 0 0 1 1320000: 0 0 0 0 0 0 0 0 1488000: 0 0 0 0 0 0 0 0 1800000: 0 0 0 0 0 0 0 0 but cat /sys/devices/system/cpu/cpufreq/policy0/scaling_setspeed <unsupported> echo 1080000 > /sys/devices/system/cpu/cpufreq/policy0/scaling_setspeed ash: write error: Invalid argument is it setspeed working for You on 5.2?

@hexdump, Yes. I added these entries as indeed adding cpu supply is a must to get my GS1 and eachlink working. I understand You have working DVFS on Yours H6 hw. Do You have any entries in /sys/class/cpufreq ? Also - what is best way to see what is current cpu freq/voltage & also to see how system walks on dvfs trips?

@jernej, thx for hint! for sure i'll test it. rc7+patches You mention gives me working screen on eachlink h6 mini. Last days I'm playing a bit with DVFS on h6 (beelink gs1). I'm using 5.3-rc7 mainline + and megous patches (also verified with Yours libreelec patches for h6). I got ths working (as I see entries in /sys/class/thermal) So next I tried to play with cpufreq. I added clk to cpu cores (like in https://github.com/megous/linux/commit/19a1b4942b8f040f2b8f2d322368518c5776025b), OPP tables (like in https://github.com/megous/linux/commit/b4aca20e5ce774f6f91bc79317d232ead36e8ede) and trip points/cooling maps (like in https://github.com/megous/linux/commit/b1a1c764c50b6d2d86d1c9124d01ca1726836e9b ). Next I enabled cpu0 power from dcdca reg and my gs1 boots. but here is no any entries in /sys/class/cpufreq What intriguing me: in megous & Yours patches there is no any touch in cpufreq-dt-platdev for H6 ( https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/drivers/cpufreq/cpufreq-dt-platdev.c?id=089cf7f6ecb266b6a4164919a2e69bd2f938374a#n28 so cpufreq-dt will not load for h6. By curiosity I added 'sun50i-h6' entry in cpufreq-dt-platdev.c and now kernel tries to load cpufreq-dt module but fails with error -2 (cant get cpu0 clock). interesting. maybe clock framework not provides required method clock_get to cpufreq?

Guys, Just quick Q: on 5.3-rc7 I still have blackscreen on Eachlink H6 (even with video=HDMI-A-1:e). Is this 5.3 kernel issue or rather something else? btw: I switch to beelink Gs1 and got almost fully working with mainline 5.3-rc7 (so far except: thermal, cpu freq. scaling and t720 gpu - but this is different story...)

Well - thx for pointing this! Unfortunately this probably not explains why I still don't have anything on hdmi output with 5.3 kernel while exactly the same patch set + kernel config gives screen with 5.2 kernel. I change H6 related patch set from Your git to https://github.com/megous/linux/commits/opi3-5.3 but this not helps. It looks like 5.3 needs some more work to get hdmi working in each link h6. boot log for 5.3-rc2 with megous patches attached. H6_5.3-rc2_boot.log

Oh- sorry. I miss Your request :-p Here are list of patches I'm using to get 5.3 mainline working ok (with gfx & sound) on my hw test set: x86 intel/amd, s905, s912, rk3328, rpi3 and very initially h6 (forget some of h6 patches. this is for pure experiments): https://github.com/warpme/minimyth2/blob/master/script/kernel/linux-5.3/Makefile Patches are here: https://github.com/warpme/minimyth2/tree/master/script/kernel/linux-5.3/files DT for eachlink H6 i'm using is here: https://github.com/warpme/minimyth2/blob/master/script/kernel/linux-5.3/files/0812-arm64-dts-allwinner-add-h6-eachlink-h6mini.patch

I'm using mainline 5.3-rc2. Sill can't get HDMI screen working on 5.3. Exactly the same build but with 5.2 mainline gives HDMI - so it looks 5.3 has something hanged... (5.3-rc2 boot log attached) H6 kernel 5.3-rc2.log

sorry for late replay. With 5.2 kernel adding video=HDMI-A-1:e is required to get anything on screen. With 5.3-rc3 screen not works. Regarding board ver. - it is v1.1. Did You see anywhere on Internet android DT for this board?

Oh - I just was extrapolating from other H6 designs :-p. Is anybody looking on GPU support in eachlink h6? If not - may You pls provide me (or point to) Android DT for this device for experimenting? I can play with this hw for gpu support - but probably will need some assistance..... BTW: regarding audio support: I already experimenting with 5.2 mainline + Your's LibreELEC H6 patches from https://github.com/jernejsk/LibreELEC.tv/tree/tanix-tx6/projects/Allwinner/devices/H6/patches/linux As I mentioned in my post: sound modules seems to be loaded - but no any audio reported by alsa. May You hint me where to further investigate why there is no any sound card?

Thx for replay! It looks eachlink mini box isn't god investment for anybody who wants to play with H6.... This box is not using AXP 805 PMIC (typically bundled with the H6) but unknown IC which cryptic label "A8038 8099330k". Unless we will have support for this PMIC - no GPU support nor any other HW parts supplied by this PMIC :-(

@hexdump, Thx for help! My strategy is to relay on mainline + minimal required set of patches because my goal (minimyth2 appliance) is to have common kernel for rpi3, s905, rk33xx and H6. This is reason why I want to avoid to go with heavily patched forks like i.e. https://github.com/megous/linux). Do You know exact set of patches applied to mainline within magus tree to get freq.scalling? I already applied dtb changes based on diff for your patch vs. mainline+jernej's patches (see attached patch) but I'm getting: [ 1.843275] core: _opp_supported_by_regulators: OPP minuV: 880000 maxuV: 880000, not supported by regulator [ 1.854018] cpu cpu0: _opp_add: OPP not supported by regulators (480000000) [ 1.861803] core: _opp_supported_by_regulators: OPP minuV: 880000 maxuV: 880000, not supported by regulator [ 1.872509] cpu cpu0: _opp_add: OPP not supported by regulators (720000000) [ 1.880284] core: _opp_supported_by_regulators: OPP minuV: 880000 maxuV: 880000, not supported by regulator [ 1.890985] cpu cpu0: _opp_add: OPP not supported by regulators (816000000) [ 1.898733] core: _opp_supported_by_regulators: OPP minuV: 880000 maxuV: 880000, not supported by regulator [ 1.909440] cpu cpu0: _opp_add: OPP not supported by regulators (888000000) [ 1.917209] core: _opp_supported_by_regulators: OPP minuV: 940000 maxuV: 940000, not supported by regulator [ 1.927912] cpu cpu0: _opp_add: OPP not supported by regulators (1080000000) [ 1.935754] core: _opp_supported_by_regulators: OPP minuV: 1000000 maxuV: 1000000, not supported by regulator [ 1.946645] cpu cpu0: _opp_add: OPP not supported by regulators (1320000000) [ 1.954510] core: _opp_supported_by_regulators: OPP minuV: 1060000 maxuV: 1060000, not supported by regulator [ 1.965404] cpu cpu0: _opp_add: OPP not supported by regulators (1488000000) [ 1.973336] cpufreq: cpufreq_online: CPU0: Running at unlisted freq: 912000 KHz [ 1.981430] cpufreq: cpufreq_online: CPU0: Unlisted initial frequency changed to: 1800000 KHz it looks something additional needs to be patched in mainline... . 0814-H6-add-thermal-zones.patch

@jernej Thx so much for help here. Got networking working :-) Now I want to get general view what is currently possible to get as supported on H6 Eachlink with mainline 5.2.3 (+ patches from Your repo). (strategy here is to have minimal required set of patches because my goal (minimyth2 appliance) is to have common kernel for rpi3, s905, rk33xx and H6. This is reason why I want to avoid to go with heavily patched forks like i.e. https://github.com/megous/linux). May You (and @hexdump, others) pls give me some hints about following aspects regarding current H6 support status: 1. to get anything on my HDMI connected monitor I need add video=HDMI-A-1:e at kernel boot cmd. Is it still required? 2. at boot, my screen has only 4 tux pictures at upper left cornet. no any dmesg output. Is it normal at this stage of H6 support? 3. I don't have any audio cards reported by kernel. But I see snd modules loaded: Module Size Used by sunxi_cedrus 28672 0 snd_soc_simple_card 24576 0 v4l2_mem2mem 32768 1 sunxi_cedrus snd_soc_simple_card_utils 28672 1 snd_soc_simple_card videobuf2_v4l2 28672 2 sunxi_cedrus,v4l2_mem2mem dwmac_sun8i 28672 0 videobuf2_dma_contig 24576 1 sunxi_cedrus stmmac_platform 24576 1 dwmac_sun8i videobuf2_memops 20480 1 videobuf2_dma_contig stmmac 172032 2 dwmac_sun8i,stmmac_platform videobuf2_common 61440 3 sunxi_cedrus,videobuf2_v4l2,v4l2_mem2mem joydev 28672 0 panfrost 45056 0 videodev 217088 4 sunxi_cedrus,videobuf2_v4l2,videobuf2_common,v4l2_mem2mem gpu_sched 32768 1 panfrost media 53248 5 sunxi_cedrus,videodev,videobuf2_v4l2,videobuf2_common,v4l2_mem2mem nvmem_sunxi_sid 16384 1 uinput 20480 0 loop 32768 2 overlay 118784 2 Is it possible to get audio on H6 Eachlink? 4. 3d block (panfrost) module load reports: [ 8.109566] panfrost 1800000.gpu: clock rate = 432000000 [ 8.134402] panfrost 1800000.gpu: gpu soft reset timed out [ 8.140037] panfrost 1800000.gpu: Fatal error during GPU init Is this expected at this stage of H6 support? 5. May You pls briefly look on my boot.log and see - is there anything worth to improve (by further patching kernel)? thx in advance! boot.log

Guys, May anybody hint how to solve issue: ac200 0-0010: Can't obtain the clock: -517 I’m using mainline 5.2.2 with patches from jernejsk repo https://github.com/jernejsk/LibreELEC.tv/tree/tanix-tx6/projects/Allwinner/devices/H6/patches/linux and sun50i-h6-eachlink-h6mini.dtb from patch included on this thread msg: https://forum.armbian.com/applications/core/interface/file/attachment.php?id=4955 My HW is Eachlink H6 Mini. System boots but no eth due ac200 can't obtain clock error. (pls see attached dmesg.log) For reference: I'm also attaching sun50i-h6.dtsi include file which is used by my kernel tree to compile sun50i-h6-eachlink-h6mini.dtb and also my kernel config. I suspect something is still missing to get ac200 working... sun50i-h6-eachlink-h6mini.dts sun50i-h6.dtsi dmesg.log linux-5.2-arm64-armv8.config

Hi, I'm trying to get working sources compiled mainline 5.0.5 on beelink a1 rk3328. Kernel boots fine but I have following issue with eth (see attached png). Kernel config is like this: https://github.com/warpme/minimyth2/blob/master/script/kernel/linux-5.0/files/linux-5.0-arm64-armv8.config Can somebody hint me: is this issue of 5.0.5 mainline or rather something is missing in kernel config or i.e. lacking of required kernel patch?

By accident I discover that issue events_freezable mmc_rescan is 4.19 GA regression compared to 4.19-rc4 as I don't have this in 4.19-rc4. Need to look at kernel sources. Any pointers to narrow such sources inspection?