Orangepi 3 h6 allwiner chip

[sadnblueish]

I witness changes at Nightly Downloads.  List is empty.  Package base states "Buster Stable".  Anyone can help me understand as I see "WIP" at OPi3. Has it entered into stable release or is it just the web page changes?

[Igor]

24 minutes ago, sadnblueish said:

Has it entered into stable release or is it just the web page changes?

Does WIP represent a stable release for you?

 

24 minutes ago, sadnblueish said:

List is empty.

Will be back online when this will make sense. When DEV moves to 5.4. For now, use test 5.3.y release which is virtually identical to "yesterdays" nightly.

 

24 minutes ago, sadnblueish said:

changes at 

 

[dolphs]

 "yous guys" are bad :-), thanks meanwhile upgraded current rock4a succesfully.

 

so building fram scratch should go like this " ./compile.sh BOARD=orangepioneplus BRANCH=current RELEASE=buster BUILD_DESKTOP=no KERNEL_ONLY=no KERNEL_CONFIGURE=no BUILD_MINIMAL=yes " or still DEV?

 

[Igor]

1 hour ago, dolphs said:

so building fram scratch should go like this " ./compile.sh BOARD=orangepioneplus BRANCH=current RELEASE=buster BUILD_DESKTOP=no KERNEL_ONLY=no KERNEL_CONFIGURE=no BUILD_MINIMAL=yes "

yes

[dolphs]

OK - pity though built from scratch buster minimal as well downloaded image, however image does not boot up (OpiOnePlus) from sdcards tested.

[ UPDATE: slight correction I do see orangepioneplus in my DHCP leases window, however unable to SSH to it ]

 

Rgd WIP , the compile tool shows H6 boards are mature now ( screenshot attachment )

Also noticed kernel is still on 5.3.9 while others ( OdroidC2, Rockpi4, etc ) are on 5.3.12/11 currently so guess I just wait another two weeks aiming for 5.4.x ;-)

cheers

[sadnblueish]

On last build, random reboots happening when under load of greater than 30%. I tried EMMC boot and SDCard boot. EXT4 Rootfs and btrfs.  Applications requiring "mono" and LXD, definitely crashes the system.  Changing CPU governor also definitely crashes the system. just to clarify no issues on power input. I also ran "stress" test to generate 100% load, it crashes instantly. If this is not known, happy to provide other details if required.

[dolphs]

Barrel connected or "USB OTG" power supply, also is it 5v 2 or 3ampere that is connected?

I tried a 5v 2ampere USB OTG as well the barrel ( 5v 3a) but unable to get it stable, perhaps 5.3.12 ( looks like "next" is being compiled with that version ) 

 

[Igor]

2 hours ago, sadnblueish said:

If this is not known

It is known that some boards are crashing. Cause is explained few posts back ... will be fixed in 5.4.y, while bring fix to 5.3.y is simply too expensive.

[sadnblueish]

25 minutes ago, Igor said:

It is known that some boards are crashing. Cause is explained few posts back ... will be fixed in 5.4.y, while bring fix to 5.3.y is simply too expensive.

Agreed.

[sadnblueish]

39 minutes ago, dolphs said:

Barrel connected or "USB OTG" power supply, also is it 5v 2 or 3ampere that is connected?

I tried a 5v 2ampere USB OTG as well the barrel ( 5v 3a) but unable to get it stable, perhaps 5.3.12 ( looks like "next" is being compiled with that version ) 

 

Tried both barrel and Micro USB. Tried few power adapters too. same result.  Tried 2/2.5/3/4 AMP adapters. I also notice this board runs much cooler when booted from SDcard. I believe @Igor explained in above post. Have to wait for 5.4 to bring the dead out of grave.

[guidol]

5 hours ago, sadnblueish said:

Changing CPU governor also definitely crashes the system. just to clarify no issues on power input. I also ran "stress" test to generate 100% load, it crashes instantly. If this is not known, happy to provide other details if required.

on nanoPi systems where such problems when using higer cpu-frequencys without installing the right voltage-regulator.
Ok your system wouldnt have that special problem, but how about lowering (for test purposes) the max. cpu-frequency to a max of about 1Ghz (1008000 Mhz on a NanoPi). Which governor is working? ondemand? Did you try conservative while the system does react unstable?

[Igor]

@jernej I was playing with H6 Librelec today... Is HDMI audio broken or simply not there yet?

[jernej]

6 minutes ago, Igor said:

Is HDMI audio broken or simply not there yet?

HDMI audio should work just fine on H6. How did you test it? Which board? Which image? What is your audio format?

[Igor]

2 hours ago, jernej said:

HDMI audio should work just fine on H6. How did you test it? Which board? Which image? What is your audio format?

OK, thanks.  What about the rest? Analogue and mic support?

Orangepi 3, latest build from https://test.libreelec.tv/ - I tried few videos over Youtube plugin. Also no UI Kodi sounds ... well. Will investigate further. I will see if it can be brought up on Armbian as well.

[jernej]

1 hour ago, Igor said:

What about the rest? Analogue and mic support?

Nope, that goes through AC200 chip and there is no driver for it. Unfortunately it's not easy to reuse BSP driver as the API changed too much. If we would have AC100 driver, this would be easy... I might look into it in future...

1 hour ago, Igor said:

I tried few videos over Youtube plugin. Also no UI Kodi sounds ... well. Will investigate further. I will see if it can be brought up on Armbian as well.

I'll test that a bit later. Anyway, I mostly test using some video files, not with plugins. Note that plugins may have troubles due to recent switch to Python 3 and development version of Kodi.

[jernej]

Sorry for late reply. I tested with https://test.libreelec.tv/LibreELEC-H6.arm-9.80-nightly-20191123-6353c73-orangepi-3.img.gz and everything works fine. There is GUI sound and also YT videos play fine (with sound).

 

Can you give me output of:

edid-decode /sys/class/drm/card0-HDMI-A-1/edid

I suspect your monitor/TV isn't recognized correctly.

[Igor]

52 minutes ago, jernej said:

I suspect your monitor/TV isn't recognized correctly.

Possible. It's attached via HDMI switch and HDMI recorder  but in most cases, with other boards, everything works well.

I can/will re-try later in the evening. Tnx.

[Flole]

Unfortunately I am suffering from stability issues in this (unstable) build aswell. I am booting from a USB HDD and I have connected the entire thing to a 5V 3A power supply that also really delivers the 3A (actually verified that). Also it's slightly above 5V at the Micro USB Plug so I think the power should be enough. Could this be related to the issue mentioned on the page before (increased voltage to the CPU)? How are chances that 5.4 kernel will be more stable for me? Is there anything I can do now to make it more reliable or at least to have it automatically reboot? I have set panic=1 already but that doesn't help, maybe the watchdog can be enabled as a workaround for now somehow?

[forever_ noob]

I have bought orange pi lite 2 board and I had a lot of trouble with it, barely ever it booted to login prompt and when it did it crashed with kernel panic.

I dont have screen capture device so i cant show you any screenshots.

This is what i got running:

root@orangepilite2:~# uname -a

Linux orangepilite2 5.3.9-sunxi64 #19.11.3 SMP Mon Nov 18 19:12:54 CET 2019 aarch64 aarch64 aarch64 GNU/Linux

 

I lurked through forum and found some useful information in this post :

I modified with editor /etc/default/cpufrequtils file that is on sd card with my ubuntu linux machine to look like this:

 

ENABLE=true

MIN_SPEED=480000

MAX_SPEED=1080000

GOVERNOR=performance

 

After modification of the file /etc/default/cpufrequtils and than putting sdcard back to Orange Pi Lite2 system boots and works.

MAX_SPEED if set over 1080Mhz crashes system again, also if GOVERNOR is set to ondemand also crashes system if pushed hard.

Sometimes it does two boot loops when restarted and sometimes it does not, it is still buggy but it works most of the time.

 

Hope someone finds this useful and dont quit on his board.

[Igor]

39 minutes ago, forever_ noob said:

MAX_SPEED if set over 1080Mhz crashes system again, also if GOVERNOR is set to ondemand also crashes system if pushed hard.

Sometimes it does two boot loops when restarted and sometimes it does not, it is still buggy but it works most of the time.

This simple workaround does not apply on this chip. Try nightly version image with a kernel 5.4.y and read few posts/pages back in this thread for more information.

[forever_ noob]

Hey Igor,

I have switched to Nightly Builds in armbian-config but now it seems my kernel is downgraded from 5.3.9 insted raised to 5.4.y

 

root@orangepilite2:~# uname -a

Linux orangepilite2 5.3.13-sunxi64 #19.11.3.332 SMP Thu Nov 28 19:42:41 CET 2019 aarch64 aarch64 aarch64 GNU/Linux

 

Hmmmmm...

 

Where can I find info how to put 5.4.y onto orange pi lite 2 ?

[Igor]

Ah, you have to switch also to latest DEV kernel in alternative kernels menu.

 

2 hours ago, forever_ noob said:

I have switched to Nightly Builds

I told you to use nightly image which came with correct DEV branch by default.

[forever_ noob]

You speak alien to me. I wanted to use sdcard image but nightly folder is empty... https://dl.armbian.com/orangepilite2/

I have found alternative kernels menu in armbian-config, kernel is being upgraded to 5.4.0 .

My fingers are crossed.

...

And I am in.

New kernel booted nicely without any boot loops.

root@orangepilite2:~# uname -a

Linux orangepilite2 5.4.0-rc8-sunxi64 #19.11.3.332 SMP Thu Nov 28 19:52:00 CET 2019 aarch64 aarch64 aarch64 GNU/Linux

 

Thanks Igor.

 

[martinayotte]

15 minutes ago, forever_ noob said:

I wanted to use sdcard image but nightly folder is empty...

Older builds are there : https://dl.armbian.com/_old/orangepilite2/nightly/

[Flole]

In order to solve my issues with the lockup I tried to enable the watchdog but I found the board locked up again this morning. Either the reboot is not working cleanly or the watchdog is not resetting the board correctly or my USB HDD is not initialized properly.

I upgraded to the new kernel now (5.4) and hope that it's more stable with this one. I'll report back on the next lockup.

As I'm facing those lockups on an almost daily basis I thought about adding a timer to the power supply that cuts the power once per day to do a clean restart. That would definitely work but if there are other ways to make it more stable I'd prefer those. I'm running munin on the device aswell but couldn't find anything that happend right before all of the crashes, CPU utilization and frequency was below maximum, there was still enough RAM so absolutely no indication of any issue.

[GeorgeP]

32 minutes ago, Flole said:

In order to solve my issues with the lockup I tried to enable the watchdog but I found the board locked up again this morning. Either the reboot is not working cleanly or the watchdog is not resetting the board correctly or my USB HDD is not initialized properly.

 

I can't help thinking that something isn't right with your hardware. Have you tried booting the bare board "normally" from an SD card?

 

I'm running my OPi3 on what is quite an old build but it seems rock solid.

If it ain't broke.....

  ___  ____  _   _____ 
 / _ \|  _ \(_) |___ / 
| | | | |_) | |   |_ \ 
| |_| |  __/| |  ___) |
 \___/|_|   |_| |____/ 
                       
Welcome to Ubuntu Bionic with Armbian Linux 5.3.0-sunxi64
System load:   0.15 0.16 0.14  	Up time:       69 days		
Memory usage:  38 % of 1993MB 	Zram usage:    5 % of 996Mb  	IP: 192.168.0.5
CPU temp:      51°C           	
Usage of /:    50% of 7.1G

 

[Flole]

Unfortunately my SD cards are for some reason wearing out extremely fast, so I prefer other methods. Also my CPU temperature is constantly around 80-85°C. My board is one of the first ones by the way. Bought shortly after they launched it.

 

The new kernel seems to have broken CPU frequency scaling, it's running now on 0Mhz according to htop and the CPU is on the limit.....

 

 

# cpufreq-info -ecpufrequtils 008: cpufreq-info (C) Dominik Brodowski 2004-2009Report errors and bugs to cpufreq@vger.kernel.org, please.analyzing CPU 0: no or unknown cpufreq driver is active on this CPU maximum transition latency: 4294.55 ms.analyzing CPU 1: no or unknown cpufreq driver is active on this CPU maximum transition latency: 4294.55 ms.analyzing CPU 2: no or unknown cpufreq driver is active on this CPU maximum transition latency: 4294.55 ms.analyzing CPU 3: no or unknown cpufreq driver is active on this CPU maximum transition latency: 4294.55 ms.

 

 

[dolphs]

WIP - similar here and has been issued to Jira AR-86

 

If you open .config ( update KERNEL ), this part needs to be updated:

 

  *
  * CPU frequency scaling drivers
  *
  Generic DT based cpufreq driver (CPUFREQ_DT) [M/n/y/?] m
  Allwinner nvmem based SUN50I CPUFreq driver (ARM_ALLWINNER_SUN50I_CPUFREQ_NVMEM) [N/m/y/?] (NEW) y
  SCPI based CPUfreq driver (ARM_SCPI_CPUFREQ) [M/n/y/?] m
  SCMI based CPUfreq driver (ARM_SCMI_CPUFREQ) [N/m/y/?] n
  CPU frequency scaling driver for Freescale QorIQ SoCs (QORIQ_CPUFREQ) [N/m/y/?] n

 

Sadly that brings partly joy as 1,5GHz is currently max speed

[Belgrado]

I don't know if I am the only person who has the problem with sound with my Orange Pi 3. I've installed Armbian Buster and Bionic, different versions but both sound output doesn't work.

 

I tested the android (7.0) that comes with Orange Pi 3 and there sound works perfectly. I tried Debian and Ubuntu from Orange Pi website but I had the same problem with Armbian.

 

So, are there user here who has or had the same problem? Except the matter or OS in WIP what do you do? So, I would like to know if I am the person with this problem, and who knows, solve it.

[Igor]

1 hour ago, Belgrado said:

I don't know if I am the only person who has the problem with sound with my Orange Pi 3. I've installed Armbian Buster and Bionic, different versions but both sound output doesn't work.

Don't open new topics and read the topic which yours was merged to. You will find all answers here.