Jump to content

Recommended Posts

Posted

Preliminary Armbian images with mainline kernel are available (nightly builds that means untested and without any end user support): https://dl.armbian.com/orangepipc2/

 

This is neither an announcement of Armbian support for this device nor anything a typical end user should play with. If you want to watch video go buy an Android TV box, if you want to help improving Linux support and are able and willing to accept that this is just a test image to get stability reports for DVFS then please read through:

It would be great if we get a lot of feedback on @ErwinH's new dvfs table that means playing around with different workloads, cramp the board in tiny enclosures (read as: eg. cardboard boxes) to let throttling start early, changing max cpufreq by adjusting /etc/defaults/cpufreq-utils and test through different clockspeeds since we need results from a lot of boards to get an idea whether the safety headroom regarding DVFS operating points is ok or needs to be extended.

 

BTW: This is also not a call for stupid requests ('When will Armbian be ready...'). Those will be either deleted or moved to a special thread 'The user requests that prevent any dev doing any more work'.

Posted

Note for the nightly images: Docker support will be added in the next nightly - after 170127

Note for the desktop nightly images: No sound - only USB sound cards are supported; video playback in MPV may work with software decoding (read as - with proper cooling and no other processes eating the CPU), at least short 720p videos should play fine

 

Edit: Also I deleted the previous broken nightly image

Posted
  On 1/26/2017 at 12:40 PM, hojnikb said:

Can resolution be changed from 720p ?

The situation is the same as with H3 mainline images - video mode is selected in u-boot using EDID (and it should default to the monitor's native resolution) and it can't be changed right now.

Posted (edited)

Is the build repository able to build PC2 images?

Tried a PC2 build (Debian jessie desktop) a few hours ago (up to commit a0a3bbe) and got this problem, repeated reboots

 

  Reveal hidden contents

 

Edited by zador.blood.stained
Added spoiler tags
Posted

@Christos

 

@martinayotte has the same problem with his board, while for other people (including me) these images work fine. I have no idea if this is a HW issue or it can be solved on software level.

 

Edit: It is possible to boot from USB, but it is tricky if you don't have u-boot in SPI flash since having the same image on USB storage and SD card will cause a UUID collision, and I'm not sure if this u-boot can be loaded in FEL.

Posted

I used this u-boot with FEL, but you have to get the files by hand:
 

You need the sunxi-spl.bin from the 32 bit version of the u-boot and the u-boot-dtb.bin from the 64 bit. sunxi-spl.bin is in the u-boot-sun50i/master/ folder if it isn't cleaned.

 

This is my script to boot through FEL:

#!/bin/sh
UBOOT=u-boot-dtb.bin
DTB=sun50i-h5-orangepi-pc2.dtb
UIMAGE=Image
DTBADDR=0x4FA00000
KERNELADDR=0x40080000
BOOTSCRADDR=0x43100000
BOOTSCR=boot.scr

sunxi-fel -v -p spl sunxi-spl.bin \
        write 0x44000 bl31.bin \
        write 0x4a000000 $UBOOT \
        write $DTBADDR $DTB \
        write $KERNELADDR $UIMAGE \
        write $BOOTSCRADDR $BOOTSCR \
        reset64 0x44000 \
Posted

@zador.blood.stained

 

Thanks.

That was the final blow on this board.

I gave up on it quite some time ago due to the fact that its expansion pinheader is not compatible to any audio add-on (really bad design from orangepi..) but just made a compilation since I happened to have the board.

Should left it in the drawer, its a waste.

 

It looks orangepi has put all of us into a really big mess with the PC2.

 

 

For what is worth it though, you guys are the best, regardless of my opinion on the specific board, your work is exceptional.

Posted

@zador.blood.stained

 

Although this board is of no use to me, it might be to others so as soon as I see aditional commits on it I'll give it another try, just to give you some feedback.

 

 

 

 

p.s.

Added some board photos just in case you need some resistor inspection

 

  Reveal hidden contents

 

Posted
  On 1/26/2017 at 1:03 PM, zador.blood.stained said:

The situation is the same as with H3 mainline images - video mode is selected in u-boot using EDID (and it should default to the monitor's native resolution) and it can't be changed right now.

Great stuff.. It pretty annoying having fixed 720p resolution on a 1080p display.

 

Will give it a go tonight.

Posted
  On 1/26/2017 at 2:14 PM, hojnikb said:

Question... Is there any kind of 2D accelaration for desktop or is just a pure framebuffer done by cpu ?

Pure FB. No DRM display driver, no Mali, no HW video decoding. Still it can play 720p video in mpv, but you'll need a USB sound card to get any kind of audio output from it.

Posted
  On 1/26/2017 at 2:17 PM, zador.blood.stained said:

Pure FB. No DRM display driver, no Mali, no HW video decoding. Still it can play 720p video in mpv, but you'll need a USB sound card to get any kind of audio output from it.

 

Makes sense. I guess mali is a hard nut to crack with mainline kernels.

As for video accelaration; Cedrus doesn't work with H5 ?

 

Audio is bust for hdmi and analog out ?

Posted (edited)

Ok, it boots now.

 

  Reveal hidden contents

 

Edited by zador.blood.stained
Add spoiler tags
Posted
  On 1/26/2017 at 2:21 PM, hojnikb said:

Makes sense. I guess mali is a hard nut to crack with mainline kernels.

Framebuffer Mali can be added, it's not that hard (but it's mostly useless). For a X11 Mali you would need a DRM display driver first (if I understand things correctly), and current H3 one is too limited and experimental ATM.

 

  On 1/26/2017 at 2:21 PM, hojnikb said:

As for video accelaration; Cedrus doesn't work with H5 ?

H3 cedrus may work if you are talking about mainline one (well, a fork of this one that supports H3), but again - it's early WIP that depends on another WIP feature (Requests API), so I wouldn't expect it to be fully ready and operational this year.

 

  On 1/26/2017 at 2:21 PM, hojnikb said:

Audio is bust for hdmi and analog out ?

Analog audio will be present in 4.10 if anyone finds time and motivation to rebase this branch on top of 4.10. I also saw SPDIF patches for H3, so they may appear in 4.11-4.12 too.

For the HDMI audio - I don't think any sunxi board has it supported in Linux kernel, and I don't think anybody is working on it.

Posted
  On 1/26/2017 at 2:29 PM, zador.blood.stained said:

if anyone finds time and motivation to rebase this branch on top of 4.10

I can rebase these commits on top of 4.10 if you like. 

Posted
  On 1/26/2017 at 3:04 PM, ErwinH said:

I can rebase these commits on top of 4.10 if you like. 

Let's wait at least for the official 4.10.0 release, no need to rush only for the analog audio yet.

Also please note that we changed and fixed some of yours and apritzel's commits and patches in order to support Pine64 and booting from SPI.

Posted
  On 1/26/2017 at 3:06 PM, zador.blood.stained said:

no need to rush only for the analog audio yet.

 

Yep, let's focus on basic things first: Stress testing to get stability reports from a much larger sample size than currently.

 

I still fear that it ends with this H5 mainline image like with dev vanilla images for H3 boards. Zero useful feedback from any user but them starting to use 4.x images productive and flooding the forum with complaints about stuff that is known to not work (ATM).

Posted

Any chance for a script that would simplify stress and temperature testing for noobs ?

 

Maybe even include one in the image and let it run on the first boot, then submit results to armbian server.

Posted
  On 1/26/2017 at 3:25 PM, tkaiser said:

I still fear that it ends with this H5 mainline image like with dev vanilla images for H3 boards. Zero useful feedback from any user but them starting to use 4.x images productive and flooding the forum with complaints about stuff that is known to not work (ATM).

Exactly why I didn't want to announce it yet. Link to the images and to this thread already appeared in some other PC2 discussions (hi 4pda.ru forum :) ), but other people may run into the same MMC pull-up issue with current nightlies.

Edit: as expected.

Posted

For noobs - some Device Tree definitions were wrong, it will be fixed in the tomorrows nightly (or there is a test CLI image linked above that has this issue fixed).

 

Not for noobs - most SD cards require external (relative to the card) pull-ups on its data lines. Previously kernel relied on pull-ups enabled by u-boot, but after the pinctrl rework same definition (SUN4I_PINCTRL_NO_PULL) caused kernel to actively disable pull-ups instead of not touching previous value.

Posted
  On 1/26/2017 at 3:29 PM, hojnikb said:

Any chance for a script that would simplify stress and temperature testing for noobs ?

 

Maybe even include one in the image and let it run on the first boot, then submit results to armbian server.

 

Nope, fire and forget doesn't work (based on experiences already made wasting hours for automated testing images and receiving not that much feedback), it requires some time and at least the will to get some knowledge. There's a reason why I referenced 3 github issues/threads above.

 

  On 1/26/2017 at 3:30 PM, zador.blood.stained said:

Exactly why I didn't want to announce it yet.

 

Hmm... please feel free to pro-actively prevent me then from leaking such stuff the next time (already when I asked the first time regarding your commit from the day before yesterday) :)

Posted
  On 1/26/2017 at 2:10 PM, hojnikb said:

Great stuff.. It pretty annoying having fixed 720p resolution on a 1080p display.

 

If I understand you correctly, you didn't try it yet? Because the driver should select native screen resolution, in this case 1080p. If it is not, then it is bug.

 

 

  On 1/26/2017 at 2:14 PM, hojnikb said:

Question... Is there any kind of 2D accelaration for desktop or is just a pure framebuffer done by cpu ?

 

H5 doesn't have any HW 2D acceleration, only 3D through mali.

 

  On 1/26/2017 at 2:29 PM, zador.blood.stained said:

For the HDMI audio - I don't think any sunxi board has it supported in Linux kernel, and I don't think anybody is working on it.

 

For this you need DRM driver. Maybe something can be done with H3 experimental one, but let's focus on basics first.

Posted
  On 1/26/2017 at 1:21 PM, zador.blood.stained said:
@martinayotte has the same problem with his board, while for other people (including me) these images work fine. I have no idea if this is a HW issue or it can be solved on software level.
 

I had this issue with my own build done before your official commit, so, maybe I missed something, since when rebuilded with your changes, the problem disappeared.

 

  On 1/26/2017 at 1:33 PM, zador.blood.stained said:

Hm. I stil wonder if this is the same MMC pins pull-up issue that affects 4.10-rc kernels. In this case it can be easily fixed...

 

I though that too, and was thinking that you've included that fix. Since you didn't, the problem was maybe elsewhere.

Posted

Good work team, congratulations. I have already tried it, installed on hdd with nand-sata-install, but I am thinking if it's possible to remove sd card, and have boot files on SPI flash. Is that possible?

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

Terms of Use - Privacy Policy - Guidelines