Linux Images (ROCK64)

[gene83]

ayufan has been building some that work. Jessie at 5.15.130 works ok,

His dl address is somewhere in this form, probably more than once.

Cheers, Gene83

[Trash_Can_Man]

https://github.com/ayufan-rock64/linux-build
Look for the "Releases" section,
latest stable = 0.5.15 actually, stable (mostly)
pre-release = 0.6.20, in heavy testing, with several bugs, some are nasty atm, like not having hdmi at all, you get ssh tho (but im not sure if this one is solved, i saw some commit recently about hdmi)
and btw im using his ubuntu minimal 0.5.15 and working pretty stable with several dockers

[gene83]

I've built up on the -136 stretch release but am having trouble making a shrunken copy like ayufan and raspian supply.

I killed udev by editing its rules to ignore /dev/sdd, so gparted could shrink the last partition without it getting remounted under gparted.

I thought, but I just dd'd 8Gb of it to a file, gparted shows about 4Gb used, and wrote that to another sd-card. As far as gparted is concerned, the copy is a blank card. This is getting frustrating. WTF am I doing wrong?

Not that I am impressed with gparted's error messages. It does not tell you on screen why it failed, you have to save the details, and wade thru kilobytes of html to find out the partition you unmounted, has been automounted even if gparted supposedly has a lock on it.

udev strikes again. My file copy has some data in it early on, but there is no way to scan thru an 8Gb file with khexedit yet this year. Or even the next. Where the heck is a scroll bar when you need it?

Is there another hex editor that does have a functioning scroll bar, or at least a working down arrow?

[thewonderer]

Testing xenial minimal 0.6.23 and the network connection fails after an hour or so. When i log in from console ifconfig and ethtool show everything is working fine and is up, but on the network i can't ping it. A restart fixes it until it fails again. When i use the sudo ifdown eth0; sudo ifup eth0 when the connection doesn't work from an external LAN PC, it does a massive page fault. Not sure how to capture it as rock64 doesn't respond.

[prof7bit]

updating from 4.4.112-rockchip-ayufan-188 to -189 does not seem to install it properly, after reboot uname still reports the 188 build and in /boot/efi/ are still the old files (file dates on all files in that folder did not change).

[Ricky]

II haven't been in Linux in over 15 years, so I'm re-learning everything.  I loaded artful minimal on Emmc. All boots fine except I either have no nameserver or one that is not working.  I can ping 8.8.8.8 but not any named server such as google.com.  I can't run apt-get update because of it.  I have tried other sources for info but files I'm  told I need to edit are not there. Can anyone help?

[pfeerick]

@gene83 Don't chop the image up by only reading part of it! dd the whole card, and then use xz or pxz to compress the resulting image. That is the main reason the images are so small - the initial partition size is small (and resized to expand to the full card on first boot) and empty space is compressed. ayufan uses xz to compress, raspberry pi folks use zip.

@Ricky Firstly just note that Artful is probably not the prefered image if you want the most reliable and tested image... xenial is the prefered one (LTS release so will receive updates for a while / has been most rigorously tested). I haven't used that particular one yet (0.5.15?), but I would expect something may need changing or tweaking in /etc/network/interfaces or /etc/network/interfaces/interfaces.d/ to suit your network setup.

[Ricky]

@gene83 Don't chop the image up by only reading part of it! dd the whole card, and then use xz or pxz to compress the resulting image. That is the main reason the images are so small - the initial partition size is small (and resized to expand to the full card on first boot) and empty space is compressed. ayufan uses xz to compress, raspberry pi folks use zip.

@Ricky Firstly just note that Artful is probably not the prefered image if you want the most reliable and tested image... xenial is the prefered one (LTS release so will receive updates for a while / has been most rigorously tested). I haven't used that particular one yet (0.5.15?), but I would expect something may need changing or tweaking in /etc/network/interfaces or /etc/network/interfaces/interfaces.d/ to suit your network setup.

Pfeerick,  Thanks for the feedback.  I'll try playing with /etc/network/interfaces/interfaces.d/  i found what should be in this file as a default to use DhCPcd.  If I can't get it to work I 'll reflash with xenial.

[t4_4t]

There seems to have been some reports already,
 "0.6.26: jenkins-linux-build-rock-64-195"
Crashes while booting.

boot.log
[    1.561006] Console: switching to colour frame buffer device 240x67
[    1.629487] rockchip-drm display-subsystem: fb0:  frame buffer device
[    1.636248] usbcore: registered new interface driver udl
[    1.640066] Unable to detect cache hierarchy for CPU 0
[    1.642053] brd: module loaded
[    1.651803] loop: module loaded
[    1.652439] lkdtm: No crash points registered, enable through debugfs
[--> 1.655143] rockchip-spi ff190000.spi: no high_speed pinctrl state
[    1.656897] Unable to handle kernel NULL pointer dereference at virtual address 00000005
[    1.657974] pgd = ffffff80093b9000
[    1.658437] [00000005] *pgd=000000007fffe003, *pud=000000007fffe003, *pmd=0000000000000000
[    1.659563] Internal error: Oops: 96000005 [#1] SMP
[    1.660187] Modules linked in:
[    1.660607] CPU: 2 PID: 1 Comm: swapper/0 Not tainted 4.4.120-rockchip-ayufan-195 #1
[    1.661585] Hardware name: Pine64 Rock64 (DT)
[    1.662146] task: ffffffc079118000 task.stack: ffffffc079114000
[    1.662909] PC is at pinctrl_select_state+0x3c/0x12c
[    1.663548] LR is at pinctrl_select_state+0x7c/0x12c
[    1.664186] pc : [<ffffff8008596840>] lr : [<ffffff8008596880>] pstate: 60000045
[    1.665118] sp : ffffffc079117420


The cause is the following commit
spi: rockchip: set higher io driver when sclk higher than 24MHz
https://github.com/ayufan-rock64/linux-k...85539ea1bd

[Malmgren]

Anyone knows what this is about?

Code:

root@rock64:/etc/network# systemctl status -- -.mount
● -.mount - Root Mount
  Loaded: error (Reason: No such file or directory)
  Active: active (mounted) since Mon 2018-04-02 12:30:52 CEST; 55min ago
   Where: /
    What: /dev/mmcblk0p7
  CGroup: /system.slice/-.mount

Warning: Journal has been rotated since unit was started. Log output is incomplete or unavailable.

Everything is working as intented and /dev/mmcblk07p is definitely mounted at /, but I really don't like having this error that I don't understand here...

I'm running the latest ayufan (currently 0.6.30) stretch minimal.