Ubuntu 22.04 Jammy LTS for Libre Computer Boards

Start
,
1

The images are located on our distro server . If you encounter any problems, please review the general troubleshoot guide and reply to the post with your issue related to this image.

  1. Use the image with a suffix that matches your board! eg. +aml-s905x-cc.img.xz
    1.a. See Libre Comptuer Board Naming and Conventions
  2. Decompress/extract the xz compressed image .img.xz with 7-zip.
  3. Write to a MicroSD card with a bit-accurate flashing tool!

For the desktop images, we recommend using 2GB+ boards and having a heatsink and fan for best experience. It will run very poorly on 1GB boards due to swapping.

We will be continuously releasing images to address the issues discovered.
Continuous updates are delivered through our repo servers via apt.

Please note that some issues may be related to Ubuntu and not are specific to our boards. You can leave a comment for those issues and we will raise it in upstream in Ubuntu proper.

1 Like
2
3

On the AML-S905x-CC, I find the current image locks up and reboots roughly every 2 out of 3 boots when reaching

ā€œBegin: Running /scripts/init-bottom ā€¦ done.ā€

On occasion it doesnā€™t reboot, just locks up, and it is then necessary to power-cycle the board.

As well, on the very first boot, I get the following:
Begin: Running /scripts/init-bottom ā€¦[ 5.760783] Unable to handle kernel paging request at virtual address 49e35000a99a11be
[ 5.763275] Mem abort info:
[ 5.766003] ESR = 0x0000000096000021
[ 5.769606] EC = 0x25: DABT (current EL), IL = 32 bits
[ 5.774879] SET = 0, FnV = 0
[ 5.777872] EA = 0, S1PTW = 0
[ 5.780985] FSC = 0x21: alignment fault
[ 5.784945] Data abort info:
[ 5.787774] ISV = 0, ISS = 0x00000021
[ 5.791585] CM = 0, WnR = 0
[ 5.794509] [49e35000a99a11be] address between user and kernel address ranges
[ 5.801623] Internal error: Oops: 96000021 [#1] PREEMPT SMP
[ 5.807120] Modules linked in: btrfs blake2b_generic xor xor_neon raid6_pq zstd_compress libcrc32c dw_hdmi_i2s_audio meson_gxl dwmac_generic crct10dif_ce meson_dw_hdmi dw_hdmi meson_drm lima drm_display_helper drm_cma_helper gpu_sched drm_shmem_helper dwmac_meson8b meson_canvas cec stmmac_platform rc_core stmmac pcs_xpcs display_connector drm_kms_helper drm
[ 5.838762] CPU: 3 PID: 329 Comm: plymouth Not tainted 5.19.0-gd26a03ff5568 #1
[ 5.845690] Hardware name: Unknown Unknown Product/Unknown Product, BIOS 2022.04-00026-g1cf922c4a8 04/01/2022
[ 5.855547] pstate: 80000005 (Nzcv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=ā€“)
[ 5.862418] pc : page_readlink+0x90/0x120
[ 5.866358] lr : path_openat+0x43c/0xfd0
[ 5.870238] sp : ffff800008fa3b90
[ 5.873510] x29: ffff800008fa3b90 x28: ffff800008fa3c88 x27: ffff08130143c9c0
[ 5.880618] x26: 0000000000000000 x25: ffff800008fa3ce8 x24: 0000000000000001
[ 5.887691] x23: ffff800008fa3ce8 x22: 0000000000000000 x21: ffff08135a0c1600
[ 5.894763] x20: ffff800008fa3da4 x19: 0000000000000000 x18: 0000000000000000
[ 5.901836] x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
[ 5.908908] x14: 0000ffffffffffff x13: 0000000000000000 x12: fefefefefefefeff
[ 5.915981] x11: 0000fffffffffff8 x10: ffff081301428000 x9 : 0000000000000003
[ 5.923053] x8 : 637261612f62696c x7 : d0e0fee0fee5febf x6 : f2921e6b9faf5f44
[ 5.930126] x5 : 0000000000000007 x4 : 49e35000a99a11be x3 : 0000000000000000
[ 5.937198] x2 : 0000000000000001 x1 : 49e35000a99a118b x0 : 49e35000a99a118a
[ 5.944271] Call trace:
[ 5.946642] page_readlink+0x90/0x120
[ 5.950276] do_filp_open+0x80/0x130
[ 5.953812] do_sys_openat2+0xb4/0x170
[ 5.957522] __arm64_sys_openat+0x64/0xb0
[ 5.961492] invoke_syscall+0x48/0x114
[ 5.965198] el0_svc_common.constprop.0+0x44/0xfc
[ 5.969865] do_el0_svc+0x30/0xc0
[ 5.973128] el0_svc+0x34/0xb0
[ 5.976145] el0t_64_sync_handler+0xbc/0x140
[ 5.980382] el0t_64_sync+0x18c/0x190
[ 5.984004] Code: d50323bf d65f03c0 9429397e f9800091 (885f7c81)
[ 5.990062] ā€”[ end trace 0000000000000000 ]ā€”
Segmentation fault
done.

1 Like
4

Thanks for the report. We have fixed this and it will be resolved when the CI generates the new images today.

1 Like
5

For a different topic, and I think that was already a known issue, but when sending a shutdown command through the console the board reboots instead of halting. Was that by design or just an oversight? Overwriting the systemd shutdown.target with halt.target might be a solution but I havenā€™t tested it yet.

Also, if I can make two small suggestions:

  • changing the grub timeout on the server version so it shows the grub menu for 3-5 seconds (currently itā€™s 0 and hidden), to either select recovery mode or a previous kernel build if something goes pear-shaped with a kernel upgrade. I can already do that change by hand and it works quite well even over the serial debugging link.

  • adding a ā€˜defaultā€™ minimal X.org/X11 configuration, otherwise ā€œX --configureā€ canā€™t identify the display driver to use if someone wants to run a different, lighter GUI thatā€™s not Wayland compatible (such as LXDE/LXQT/XFCE). This would apply to both the server and desktop builds.

The following two files, placed in /etc/X11/xorg.conf.d/ work just fine after some testing.

01-aml-s905x-cc-defaults.conf

Section "Monitor"
    Identifier	"Monitor0"
    Option	"DPMS" "false"
EndSection

Section "ServerFlags"
    Option	"BlankTime" "0"
    Option	"StandbyTime" "0"
    Option	"SuspendTime" "0"
    Option	"Offtime" "0"
EndSection

02-aml-s905x-cc-driver.conf

Section "ServerFlags"
    Option	"AutoAddGPU" "off"
EndSection

Section "OutputClass"
    Identifier	"Meson"
    MatchDriver	"meson"
    Driver	"modesetting"
    Option	"PrimaryGPU" "true"
EndSection

Finally, thatā€™s more of a minor hardware documentation issue on the AML-S905X-CC. The jumper to select between HDMI-CEC and GPIO11, near the HDMI connector and TTY debugging header, has no silkscreen marking whatsoever. Which position on that jumper gives which option?

6

Good recommendations. We will add 3 second delay and default X11 configurations.

@Maes The jumper configuration by default for HDMI towards the inside of the board. To move it to control the GPIO on the header, move the jumper towards the edge of the board. Updating the GPIO header spreadsheet.

1 Like
7

For the ALL-H3-CC-H5 Iā€™m getting to a grey background with mouse cursor, but no further. Iā€™ll check more tomorrow, I realized I have no idea where my Potato has gotten toā€¦

8

Is that with the default desktop image?
Also, is it after GUI login or do you not get a login prompt?

AFAIK if thereā€™s no login prompt it would point to something weird going on with the display manager, if itā€™s after the login prompt it points to something with the shell itself.

9

ALL-H3-CC-H5 might take a bit if the SoC gets hot or the MicroSD card is slow. Will double check after pushing the images with the update for Le Potato and X11 configurations today.

1 Like
10

Replicated the hang. It was working last week. Will fix this today.

11

If anything is ever constant, itā€™s that Linux will fight you. :smiley:

1 Like
12

Iā€™m having issues with the network connection with the s805x image on La Frite. I get the following:
meson8b-dwmac c9410000.ethernet eth0: Reset adapter
This was an issue with previous revisions of the image and would work after a couple reboots, but it doesnā€™t seem to got back up again anymore

1 Like
13

This needs a firmware update. Will generate a firmware update shortly.

1 Like
14

The 2022-08-07 server image for the 905X seems to work fine and boots without freezing or crashing at /scripts/init-bottom as long as the emmc isnā€™t plugged in, but I might have an old bootloader on it that I will wipe before doing more testing.

I also tested my idea of replacing systemdā€™s shutdown.target with halt.target and thatā€™s a no-go, not only does it not work but it outright makes it impossible to shut down the system since it creates a circular dependency. Renaming the files back to their original and power-cycling the system fixes that issue.

The 3-second delay to the grub menu works great and doesnā€™t add a noticeable delay to boot time, that will definitely be useful in case of issues.

Even after installing pulseaudio on the server image, the HDMI audio doesnā€™t seem to be detected correctly (or at all)? I will do more testing on that.

I will test the desktop image later today and tomorrow.

15

Thereā€™s a lot of fixes in flight. Best to test the images tomorrow. The changes will be in the top-post. HDMI audio should be working on the images. Double check thereā€™s an /etc/alsa/asound.conf in your image.

With regards to shutdown, try shutdown -H instead of shutdown -P. Halting the machine. Normal shutdown triggers the reset pin on the SoC which does a full reboot.

For audio, test using speaker-test -c 2 to test ALSA. If pulse doesnā€™t work, then it is somewhere between pulse and alsa. You can also try pipewire.

1 Like
16

Thereā€™s a bug in upstream Ubuntuā€™s chromium-browser snap. It does not read from .config/chromium-flags.conf.

17

Just uploaded new images. We tested the server and the desktop images and were able to get both on Le Potato.

2 Likes
18

It would be just a minor thing, but during the debugging and development phase would it be possible to add the build or upload date, or a revision number, to the image filenames? Just to make it easier to sort out which is which with multiple versions on the same machine.

1 Like
19

We have a lot of scripts tied to the naming. We have prettyā€™d the download site with better looking indexes so you can see when new images changed.

http://distro.libre.computer/ci/ubuntu/22.04/

1 Like
20

For the latest S905x server image uploaded on August 10th, at least, there isnā€™t an asound.conf in /etc/alsa, only a ā€˜conf.dā€™ folder containing the file ā€˜50-pipewire.confā€™

speaker-text -c 2, however, does work just fine.

Here is the current output of aplay -l

**** List of PLAYBACK Hardware Devices ****
card 0: LIBRETECHCC [LIBRETECH-CC], device 0: fe.dai-link-0 (*) []
  Subdevices: 1/1
  Subdevice #0: subdevice #0

and here is the current output of aplay -L

null
    Discard all samples (playback) or generate zero samples (capture)
pipewire
    PipeWire Sound Server
hw:CARD=LIBRETECHCC,DEV=0
    LIBRETECH-CC,
    Direct hardware device without any conversions
plughw:CARD=LIBRETECHCC,DEV=0
    LIBRETECH-CC,
    Hardware device with all software conversions
default:CARD=LIBRETECHCC
    LIBRETECH-CC,
    Default Audio Device
sysdefault:CARD=LIBRETECHCC
    LIBRETECH-CC,
    Default Audio Device
dmix:CARD=LIBRETECHCC,DEV=0
    LIBRETECH-CC,
    Direct sample mixing device

Here is the content of /proc/asound/cards

0 [LIBRETECHCC    ]: gx-sound-card - LIBRETECH-CC
                      libre_computer-aml_s905x_cc-

and here is the content of /proc/asound/devices

  1:        : sequencer
  2: [ 0- 0]: digital audio playback
  3: [ 0]   : control
 33:        : timer

What should be the content of /etc/alsa/asound.conf? Iā€™ll type it in the current system and see how it goes.

For the next image would it be ok for me to request a few additional kernel modules and overlays, or would that be scheduled for later on in the debugging process?
For the intended use I would need the modules for

  • spi-meson-spicc.ko (already present in the current image, but not sure how to activate it)
  • whichever would be the one appropriate to enable the 1st i2c bus (that goes to 7J1 pins 3 & 5)
  • whichever would be the one appropriate to enable uarta (uart on 7J1 pins 8&10)
  • module sc16is7xx.ko (for sc16is750/752 uart-over-i2c or spi)
  • pps-gpio

On a different note, the X11 configuration files mentioned previously werenā€™t added in /etc/X11/xorg.conf.d on the server image at least