Orange Pi3 lts

[jernej]

4 minutes ago, Ukhellfire said:

Should I go back to what was in the original from orange pi? Or is rsb better to use? I've read rsb is better? 

Both works, but RSB is faster. I suggest to use RSB, just to be in line with OPi 3 DT.

[Ukhellfire]

6 minutes ago, jernej said:

Both works, but RSB is faster. I suggest to use RSB, just to be in line with OPi 3 DT.

Yeah that's what I've just adjusted back to. Question is, does this need to be reflected in uboot? Or is only i2c available for uboot?

[jernej]

8 minutes ago, Ukhellfire said:

Would it not just be plugged into mains as normal? Then just the tx/rx/gnd as normal?

It's a long time ago since I worked with Arduino, but they usually run at 5 V. This is too much for H6 serial port, which expects 3.3 V signals. I may be wrong about Arduino, though. I highly recommend measuring idle state voltage level on TX pin. If it is above 3.3 V, then don't use it.

[Ukhellfire]

Just now, jernej said:

It's a long time ago since I worked with Arduino, but they usually run at 5 V. This is too much for H6 serial port, which expects 3.3 V signals. I may be wrong about Arduino, though. I highly recommend measuring idle state voltage level on TX pin. If it is above 3.3 V, then don't use it.

Ahhh OK. I get you. I've never really used arduino for anything other than mounting a ramps shield on! Never had the need to use it!

I'll double check that in the morning! 

[jernej]

1 minute ago, Ukhellfire said:

Question is, does this need to be reflected in uboot?

No, U-Boot (currently) doesn't set any voltage regulators for H6 boards. However, ATF does, but it works with both busses (at least new versions). But even this is not really necessary. Most important regulators already run at power on. You can boot to Linux with them.

[Ukhellfire]

55 minutes ago, jernej said:

It's a long time ago since I worked with Arduino, but they usually run at 5 V

It's seems your right. Tx is 5v. So I guess I won't be using my arduino! 

 

Can I input using the keyboard and hdmi, and then the output through the rx? That way I'm not sending any signal from the arduino, only too it? 

Does uboot work like that? 

[jernej]

7 hours ago, Ukhellfire said:

Can I input using the keyboard and hdmi, and then the output through the rx?

No, output method is always the same as input. Anyway, RX on Arduino still operates at 5 V internally, so it might not reliably decode 3.3 V signal (it can be even slightly lower than that).

 

Why don't you just fix regulator address first and see if this is it? As I mentioned, boot reliability issues can come from incorrectly set voltage regulator.

[Ukhellfire]

1 hour ago, jernej said:

No, output method is always the same as input. Anyway, RX on Arduino still operates at 5 V internally, so it might not reliably decode 3.3 V signal (it can be even slightly lower than that).

 

Why don't you just fix regulator address first and see if this is it? As I mentioned, boot reliability issues can come from incorrectly set voltage regulator.

I'm just impatient! Haha! If I can do something without buying something to, I'm all for it. Well I read the rx is 3.3v but the tx is 5v, so thought maybe! But I guess I'm outta luck! Haha! Oh well!

 

I started off a build before I went to bed last night, with the regulator set back to @745 in kernel dts. Will be testing very shortly. So fingers crossed! I'll report back! 

I'm also going to tidy my repos up in github, so they make a bit more sense, and aren't just me messing with settings haha! 

[jernej]

4 minutes ago, Ukhellfire said:

with the regulator set back to @745 in kernel dts.

Note that actual address is in reg property, it's just convention that address is in node name too. So make sure you fixed address in both places.

[Ukhellfire]

On 3/20/2022 at 8:57 AM, jernej said:

Note that actual address is in reg property, it's just convention that address is in node name too. So make sure you fixed address in both places.

I just reverted the commit that changed it in the first place. Both places are back to @745 now 😀

 

@jernej

So it seems that it's booted! Straight away! No messing about! 

Obviously there a fair bit that isn't working. 

Ethernet, wifi, aswell as probably a few other things! Haha!

I've managed to grab a load of logs. I'm still getting the modules error too? 

Link to logs...

https://drive.google.com/drive/folders/1VSl0DS_GFbRpt1rp8PBDnV6QKI8KPOjY

 

On another note, the img is up on google drive:

https://drive.google.com/drive/folders/16UHmZsGD4qv7HqY5pPtZcQZ1WeEBDFA8

If anyone else is up for testing it for me? 

@Kamileon @Rockford the Roe @orangepifan

If possible, if things don't boot etc, any chance of some logs back too? 

Pretty please? Haha!

 

And bow I sit here and anxiously read through the plethora of logs, and await some testing! 

[jernej]

14 minutes ago, Ukhellfire said:

I'm still getting the modules error too? 

I have zero idea what is that about... As I said, it looks like some image building issue, but don't hold me on that. BTW, can you try to boot ordinary OPi3 image and see if there is same issue? it should mostly work.

[Ukhellfire]

20 hours ago, jernej said:

I have zero idea what is that about... As I said, it looks like some image building issue, but don't hold me on that. BTW, can you try to boot ordinary OPi3 image and see if there is same issue? it should mostly work.

I've started digging into it a bit. It seems (from what I can see) that it's not every module. Just some. This could purely be down to the fact I haven't gone through and check absolutely everything. Just the booting sides of things...

Yeah I'll build an OPi3 img and boot that. See what happens. I know the ethernet worked last time I tried that? Just had to plug it in from boot, as it wouldn't connect if I plugged in once booted....

Still, it's all a step closer! 

Tha k you for what you've done so far!

Never know, I may get to the point of attempting to debug your WiFi driver too! Haha!

 

@jernej

So. I've built and booted the OPi3 image. And everything seems to be good at a slight glance. No red errors as it's booting saying about failing to load kernel modules etc... Will dig deeper shortly to double check! 

 

Although, digging a bit deeper with the OPi3-LTS build, I've found a few modules are are failing to load.

I tried starting systemd-modules-load.service manually, and then run journalctl -xe as suggested by the prompt. 

The output is as follows:

https://drive.google.com/file/d/1rdit5SVwaYIggGXfRa4NXQQxIwlnashm/view?usp=drivesdk

 

There's 3 main modules. That's the 3 from the WiFi driver. So I've disabled them from the build for now. And am building again without them. Hopefully this allows systemd-modules-load.service to start properly...

[Kamileon]

7 hours ago, Ukhellfire said:

On another note, the img is up on google drive:

https://drive.google.com/drive/folders/16UHmZsGD4qv7HqY5pPtZcQZ1WeEBDFA8

If anyone else is up for testing it for me? 

@Kamileon @Rockford the Roe @orangepifan

If possible, if things don't boot etc, any chance of some logs back too? 

Pretty please? Haha!

Tested new image, still same result on my SBC, dropping to shell (initramfs)

[Ukhellfire]

1 hour ago, Kamileon said:

Tested new image, still same result on my SBC, dropping to shell (initramfs)

Ok this is strange. I really don't get why I can boot. But you can't?

Are you able to grab a dmesg?

[Kamileon]

10 hours ago, Ukhellfire said:

Are you able to grab a dmesg?

Unfortunately no, my USB keyboard doesn't work on this image

[Ukhellfire]

27 minutes ago, Kamileon said:

Unfortunately no, my USB keyboard doesn't work on this image

I did enable it in uboot? Have you tried all ports? Think mine worked in the top one of the 2. The USB3 port? 

 

Also, I've uploaded a new image. In the ethernet folder.

Ethernet works perfectly, and I've removed the WiFi driver for now (it wasn't loading properly).

The image now boots without any errors. 

Maybe this will help you to boot it? 

I've also got one more thing I want to try in regards to u reliable booting... I'll keep you posted in that one! @Kamileon

[Kamileon]

Yeah, tried ethernet image and same result dropping to shell but i did manage to do dmesg, USB was working on top port

https://nextcloud.lodz.webredirect.org/s/T8dMn3gk2YHe64p

[Ukhellfire]

1 hour ago, Kamileon said:

Yeah, tried ethernet image and same result dropping to shell but i did manage to do dmesg, USB was working on top port

https://nextcloud.lodz.webredirect.org/s/T8dMn3gk2YHe64p

Ah brilliant! I'll have a read through! Did anything stand out to you?

[Kamileon]

13 minutes ago, Ukhellfire said:

Did anything stand out to you?

"Couldn't get bank PC/PD regulator" and series of pin numbers that followed but don't know if it's related to not booting properly

[Ukhellfire]

9 minutes ago, Kamileon said:

"Couldn't get bank PC/PD regulator" and series of pin numbers that followed but don't know if it's related to not booting properly

They are the exact same errors that I got rid of, that allowed me to boot. Somethings different in out setups I think. Maybe power supply rating? I don't know.

Let me do another build.. 

[mumrau]

Power ratings over USB-C might be the issues, see this guys problem: 

On my hand, I also had issues, my phone powercable did work only on a "dumb" USB output, but not on its own powerdelivery plug, while a dumb cable worked with the powerdelivery plug but not on the "dumb" USB output, so...

[Ukhellfire]

16 minutes ago, mumrau said:

Power ratings over USB-C might be the issues, see this guys problem: 

On my hand, I also had issues, my phone powercable did work only on a "dumb" USB output, but not on its own powerdelivery plug, while a dumb cable worked with the powerdelivery plug but not on the "dumb" USB output, so...

Well I was thinking along those lines. As the builds aren't that different at all. Mine and orange pis. 

@Kamileon @orangepifan @Rockford the Roe

This is the power supply I'm using...

 https://www.amazon.co.uk/dp/B09CPGFT3H/ref=cm_sw_r_apan_glt_i_F969Q5FGJYJHWW0QMBDP?_encoding=UTF8&psc=1

 

I'm booting very reliably now using this. 

this being said, I am going to try that one last ditch attempt to get it running on your boards with code change. The difference being, orangepi use I2C to manage the pmic, whereas armbian is using RSB.

I'm going g to switch to I2C as a test and let you guys test that out. May not work, but may. They regulate things differently so maybe!

[jernej]

40 minutes ago, Ukhellfire said:

The difference being, orangepi use I2C to manage the pmic, whereas armbian is using RSB.

I'm going g to switch to I2C as a test and let you guys test that out. May not work, but may. They regulate things differently so maybe!

There is absolutely no difference, otherwise there is a bug somewhere. Reason why OrangePi image has I2C and Armbian RSB is that switch from I2C to RSB happened with kernel 5.13 and OrangePi uses kernel 5.10, which predates this. If OrangePi releases image with newer kernel, it will most likely use RSB too. The only important differences could be in regulator settings (subnodes to axp), those are important settings to copy.

[Ukhellfire]

5 minutes ago, jernej said:

There is absolutely no difference, otherwise there is a bug somewhere. Reason why OrangePi image has I2C and Armbian RSB is that switch from I2C to RSB happened with kernel 5.13 and OrangePi uses kernel 5.10, which predates this. If OrangePi releases image with newer kernel, it will most likely use RSB too. The only important differences could be in regulator settings (subnodes to axp), those are important settings to copy.

Ah OK. I was reading it wrong then. From what I understood it was the way it dished out addresses or whatever it is. 

I shall look into those settings. Otherwise, I haven't got a clue why it wouldn't be booting on others devices, unless it is power supply related like suggested? But Durley that would be down to the regulator settings too? So this can be sorted too? Either that or I'm u derstanding things wrong? Guess im kinda learning tidbits on the fly here! 

But I love it, its all good! I get a buzz out of getting things going! Haha! 

[orangepifan]

1 hour ago, Ukhellfire said:

Well I was thinking along those lines. As the builds aren't that different at all. Mine and orange pis. 

@Kamileon @orangepifan @Rockford the Roe

This is the power supply I'm using...

 https://www.amazon.co.uk/dp/B09CPGFT3H/ref=cm_sw_r_apan_glt_i_F969Q5FGJYJHWW0QMBDP?_encoding=UTF8&psc=1

 

I'm booting very reliably now using this. 

this being said, I am going to try that one last ditch attempt to get it running on your boards with code change. The difference being, orangepi use I2C to manage the pmic, whereas armbian is using RSB.

I'm going g to switch to I2C as a test and let you guys test that out. May not work, but may. They regulate things differently so maybe!

I should have time to flash & try myself tonight. I will report in. 

[jernej]

5 minutes ago, Ukhellfire said:

From what I understood it was the way it dished out addresses or whatever it is.

Well, you have to be careful which address you use with which bus type. As long as bus and address pair match, there should be no observable difference in behaviour.

 

I have no clue why it wouldn't work on another board. BTW, error -517 means that driver can't be loaded due to dependency issues. In this case, where axp address was wrong, this happened because many devices tried to enable or set up their regulator, but regulator couldn't be located on bus due to wrong address.

[Kamileon]

1 hour ago, Ukhellfire said:

Well I was thinking along those lines. As the builds aren't that different at all. Mine and orange pis. 

@Kamileon @orangepifan @Rockford the Roe

This is the power supply I'm using...

 https://www.amazon.co.uk/dp/B09CPGFT3H/ref=cm_sw_r_apan_glt_i_F969Q5FGJYJHWW0QMBDP?_encoding=UTF8&psc=1

 

I'm booting very reliably now using this. 

this being said, I am going to try that one last ditch attempt to get it running on your boards with code change. The difference being, orangepi use I2C to manage the pmic, whereas armbian is using RSB.

I'm going g to switch to I2C as a test and let you guys test that out. May not work, but may. They regulate things differently so maybe!

I have this power supply

https://www.conrad.com/p/xiaomi-mdy-10-ef-mobile-phone-charger-type-quick-charge-mode-usb-2355049

And it works with official debian image from OrangePi 100% of the time so there must be some differences in regard to power settings between images

[amaethon]

I tested again too. Still dropping to the initramfs prompt. My power supply is a cheap 5v 3amp from aliexpress so I doubt it is particularly good, but I am able to boot and run the orange pi image without any trouble at all.

 

dmesg at the initramfs prompt doesn't seem to work for me unless I am doing something wrong?

[Ukhellfire]

23 hours ago, jernej said:

Well, you have to be careful which address you use with which bus type. As long as bus and address pair match, there should be no observable difference in behaviour.

 

I have no clue why it wouldn't work on another board. BTW, error -517 means that driver can't be loaded due to dependency issues. In this case, where axp address was wrong, this happened because many devices tried to enable or set up their regulator, but regulator couldn't be located on bus due to wrong address.

At the moment I'm stumped. Not th sure where to find it all. Thought I did, but everything looked the same, OK. So maybe I didn't find it? 

[jernej]

34 minutes ago, Ukhellfire said:

Not th sure where to find it all.

Find what?