[OS] Debian Bullseye

BaldNerd · Post by **BaldNerd** » Sat Feb 04, 2023 1:27 am

Hi @meveric -- hope you've been well! Thanks for the continued work on this.

I'm seeing very odd filesystem issues with this image. I've checked the MD5 and it matches.

Working with Debian-Bullseye-1.1-20221003-XU4.

Upon first boot I'm seeing this on the HDMI output of my XU4:

Code: Select all

EXT4-fs (mmcblk0p2) bad geometry: block count exceeds size of device

I've tried both a 32 GB mSD and 128 GB. Same result.

After a bit, it starts spewing out journal write failures on the HDMI output due to being out of space on that same partition (which I understand from your post should have been auto-resized upon boot, but was not).

I flashed using RPi Imager first, and wanted to rule that out so then flashed with BalenaEtcher. Same result.

I can get the image to boot on an HC1, but it locks up during filesystem sync and behaves strangely.

It doesn't boot at all on XU4 most of the time (sometimes it gets there).

I've even tried manually resizing mmcblk0p2 on a separate system and expanding its filesystem, just in case there was an issue with the partition size or the first-boot resize script not working, but this didn't change anything.

I've tried a different microSD card.

Haha - I can't think of what else I might try... Any ideas? Any thoughts on what's going on here?

Thanks!!

meveric · Post by **meveric** » Sat Feb 04, 2023 5:14 am

I think I've seen similar issues in the past I will check the image again and see if I can improve it.
The journal errors I've seen a couple of times before but can normally be ignored, happens when it does extensive logging during the first boot for example, and even after resize it complains often that the file system was damaged or something like that, but it refers to old journals that I normally clean up before I publish the image.
But it could well be that back then I did it incorrectly which causes the "old" files to be invalid, but it doesn't affect new logs, so it can be ignored.

But yes resize should work and I've had instances where it didn't work for me either and the next time I tried it worked fine again, so I probably did not check it as thorough as I should have, and since my image is running for months now I wasn't checking back on the image since then but only did updates.

I'll see what I can find out, I probably shrunk the image to excessive to conserve image space which can cause issues.
But yes let's see what I can find out

BaldNerd · Post by **BaldNerd** » Sat Feb 04, 2023 5:55 am

Thanks so much.

Looking forward to hearing what you find.

meveric · Post by **meveric** » Sat Feb 04, 2023 8:31 pm

narf and of cause I try and it works fine on the first try

But yes the rootfs is very tiny apparently I shrunk it to the bare minimum of what the device supported

Tried it a second time worked as well.
I've slightly increased the rootfs of the default image and uploaded the new image, no changes besides that.

Maybe it solves your problem, also I remember the XU3/XU4 had some issues with some of the SD card brands.. like i've seen that booting from an SanDisk Extreme damaged the rootfs, simply by booting, and second boot was already finding tons of filesystem errors.

BaldNerd · Post by **BaldNerd** » Sat Feb 04, 2023 11:04 pm

Okay thanks merveric.

As mentioned, I'd resized p2 as well, without success. I wonder if perhaps there is some data that got truncated when it got scaled down (I've done that when packaging my images with dd before, so I've seen the behavior). Funny that yours boots every time - well, not funny for me, but odd. Maybe my board or mSD card has something different that fills up the journal faster or something like that.

I'll keep playing with it - thanks so much!

meveric · Post by **meveric** » Sun Feb 05, 2023 1:07 am

For the image creation I use a slightly bigger size than the end of the partition, so it should not have any issues with truncated images.

I write my image directly under Linux with:

Code: Select all

xzcat Debian-Bullseye-1.1.1-20230204-XU4.img.xz | dd of=/dev/mmcblk0 bs=1M conv=notrunc,fsync status=progress

I don't use any of the other tools and haven't had issues with this method yet.

ChicoDeGoma · Post by **ChicoDeGoma** » Thu Mar 02, 2023 6:07 pm

meveric wrote: ↑
Tue Dec 13, 2022 5:37 am

eel wrote: ↑
Tue Dec 13, 2022 5:01 am
My XU4 is a headless server, so the old kernel should be fine. I changed all references from buster to bullseye in /etc/apt/sources.list and /etc/apt/sources.list.d/* and did an upgrade. Is that all I need to do to upgrade to Bullseye from your Buster image?
Yes that would be all that is needed.

eel wrote: ↑
Tue Dec 13, 2022 5:01 am
I also tried upgrading to kernel 6.0.12 from your Bullseye image. iperf3 improved to 400Mbps on average with spikes up to 600Mbps. It is still not normal performance, but it is 2x what I was getting on 5.19. Do you know if the network drivers are new, or if the performance difference is just from the kernel? By comparison, kernel 4.14 is a stable 930Mbps.
Yes I updated the drivers: viewtopic.php?p=361499#p361499
But as you said, it's not fully there yet.

Does someone have any news related to the networking drivers? I'm planing to move soon but I would really love to have my full ISP speeds. I also love meveric's image so I wouldn't like to have to search another

meveric · Post by **meveric** » Fri Mar 03, 2023 12:50 am

I saw slight improvements using "irqbalance" package where network traffic is automatically switched to the big cores, which gives overall better performance, but it's still not as good as it it with Kernel 4.14.

ChicoDeGoma · Post by **ChicoDeGoma** » Sat Mar 04, 2023 4:18 pm

I'm really far from even a basic OS developer but how could I help if anything?
I love your images because they fit my setups really nice but some programs I use got deprecated on stretch backports even and it won't let me install the dependencies for bullseye backports.
So I doubt I can help but I offer to help if I could, again, I have almost no experience but the will is there.

meveric · Post by **meveric** » Sat Mar 04, 2023 5:22 pm

unless you're a Kernel developer and understand why the performance of the NIC suddenly dropped that much, I don't think you can help.

But if you're missing some dependencies, let me know I can normally create backports from older or newer applications and drivers.

ChicoDeGoma · Post by **ChicoDeGoma** » Sat Mar 04, 2023 10:09 pm

I guess I will PMIng you soon then.
I can't help at all as you said, sorry, however, I noticed armbian put out special drivers for new kernels and labeled them as a fix?
The only link I could get is this and is just an image so sorry if it is useless:
viewtopic.php?p=336435#p336435

meveric · Post by **meveric** » Sun Mar 05, 2023 2:48 am

I've seen that one, and it's only Kernel 5.4 so not a recent Kernel, probably close to Kernel 4.14.
Hardkernel has a Kernel 5.4 branch for the XU4 as well and I have a 5.4 Kernel as well which is based on that Kernel tree from HardKernel,
but that would not work with this image, as you won't have proper panfrost support which means you're limited to OpenGL ES only and other things.

johannfr · Post by **johannfr** » Thu Mar 23, 2023 1:02 am

Hi!

I'm running the latest Bullseye images and all packages (including kernel) are up-to date on an HC2.

After the board has been running for a while, I always get a kernel-crash, which *seems* to be related to an r8152 watchdog.

The HC2 is headless, but I have a serial-console attached and captured the dump I've included below.

Any ideas? Others having similar issues?

Code: Select all

[ 7372.918158] ------------[ cut here ]------------
[ 7372.923809] WARNING: CPU: 0 PID: 0 at net/sched/sch_generic.c:525 dev_watchdog+0x26c/0x328
[ 737[ 7372.932562] NE DEV WATCHDOG: eth0 (r8952): transmit :ueue 0 timed out
U: 0 PID: 0 at net/sched/sch_generic.c:525 dev_watchdog+0x26c/0[ 7372.939972] Modules linked in: wireguard curve25519_neon libcurve25519_generic libchacha20poly1305 chacha_neon poly1305_arm s5p_mfc exynos_gsc s5p_jpeg videobuf2_dma_contig videobuf2_memops v4l2_mem2mem videobuf2_v4l2 videobuf2_common videodev mc fuse ip_tables x_tables ext4 mbcache
[ 7372.971229] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.2.7-armmp-lpae #exynos5
[ 7372.978000] Hardware name: Samsung Exynos (Flattened Device Tree)
[ 7372.983946]  unwind_backtrace from show_stack+0x10/0x14
[ 7372.989039]  show_stack from dump_stack_lvl+0x58/0x70
[ 7372.994047]  dump_stack_lvl from __warn+0x7c/0x214
[ 7372.998786]  __warn from warn_slowpath_fmt+0x138/0x1c8
[ 7373.003935]  warn_slowpath_fmt from dev_watchdog+0x26c/0x328
[ 7373.009618]  dev_watchdog from call_timer_fn+0xb8/0x39c
[ 7373.014776]  call_timer_fn from expire_timers+0x11c/0x280
[ 7373.020163]  expire_timers from run_timer_softirq+0x114/0x2c0
[ 7373.025916]  run_timer_softirq from __do_softirq+0x14c/0x630
[ 7373.031541]  __do_softirq from __irq_exit_rcu+0x140/0x1ec
[ 7373.036888]  __irq_exit_rcu from irq_exit+0x8/0x34
[ 7373.041612]  irq_exit from __irq_svc+0x8c/0xd0
[ 7373.045983] Exception stack(0xc1801f00 to 0xc1801f48)
[ 7373.051093] 1f00: 00000000 c14acf5c 00000001 c011c620 c1835014 c18391c0 00000000 c1835074
[ 7373.059569] 1f20: c1789aa8 00000000 00000000 00000000 00000000 c1801f50 c010937c c0109380
[ 7373.067702] 1f40: 200f0013 ffffffff
[ 7373.070674]  __irq_svc from arch_cpu_idle+0x40/0x44
[ 7373.075665]  arch_cpu_idle from default_idle_call+0x48/0x84
[ 7373.081282]  default_idle_call from do_idle+0x218/0x2c4
[ 7373.086446]  do_idle from cpu_startup_entry+0x18/0x1c
[ 7373.091451]  cpu_startup_entry from rest_init+0xf0/0x194
[ 7373.096764]  rest_init from arch_post_acpi_subsys_init+0x0/0x8
[ 7373.098061] exynos5-hsi2c 12ca0000.i2c: rx timeout
[ 7373.102473] irq event stamp: 21186618
[ 7373.112934] hardirqs last  enabled at (21186646): [<c01ac214>] __up_console_sem+0x50/0x60
[ 7373.121276] hardirqs last disabled at (21186665): [<c01ac200>] __up_console_sem+0x3c/0x60
[ 7373.129714] softirqs last  enabled at (21185954): [<c0101690>] __do_softirq+0x354/0x630
[ 7373.137884] softirqs last disabled at (21186381): [<c0131064>] __irq_exit_rcu+0x140/0x1ec
[ 7373.146146] ---[ end trace 0000000000000000 ]---
[ 7373.150705] r8152 6-1:1.0 eth0: Tx timeout
x328
[ 7372.932562] NETDEV WATCHDOG: eth0 (r8152): transmit queue 0 timed out
[ 7372.939972] Modules linked in: wireguard curve25519_neon libcurve25519_generic libchacha20poly1305 chacha_neon poly1305_arm s5p_mfc exynos_gsc s5p_jpeg videobuf2_dma_contig videobuf2_memops v4l2_mem2mem videobuf2_v4l2 videobuf2_common videodev mc fuse ip_tables x_tables ext4 mbcache
[ 7372.971229] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.2.7-armmp-lpae #exynos5
[ 7372.978000] Hardware name: Samsung Exynos (Flattened Device Tree)
[ 7372.983946]  unwind_backtrace from show_stack+0x10/0x14
[ 7372.989039]  show_stack from dump_stack_lvl+0x58/0x70
[ 7372.994047]  dump_stack_lvl from __warn+0x7c/0x214
[ 7372.998786]  __warn from warn_slowpath_fmt+0x138/0x1c8
[ 7373.003935]  warn_slowpath_fmt from dev_watchdog+0x26c/0x328
[ 7373.009618]  dev_watchdog from call_timer_fn+0xb8/0x39c
[ 7373.014776]  call_timer_fn from expire_timers+0x11c/0x280
[ 7373.020163]  expire_timers from run_timer_softirq+0x114/0x2c0
[ 7373.025916]  run_timer_softirq from __do_softirq+0x14c/0x630
[ 7373.031541]  __do_softirq from __irq_exit_rcu+0x140/0x1ec
[ 7373.036888]  __irq_exit_rcu from irq_exit+0x8/0x34
[ 7373.041612]  irq_exit from __irq_svc+0x8c/0xd0
[ 7373.045983] Exception stack(0xc1801f00 to 0xc1801f48)
[ 7373.051093] 1f00: 00000000 c14acf5c 00000001 c011c620 c1835014 c18391c0 00000000 c1835074
[ 7373.059569] 1f20: c1789aa8 00000000 00000000 00000000 00000000 c1801f50 c010937c c0109380
[ 7373.067702] 1f40: 200f0013 ffffffff
[ 7373.070674]  __irq_svc from arch_cpu_idle+0x40/0x44
[ 7373.075665]  arch_cpu_idle from default_idle_call+0x48/0x84
[ 7373.081282]  default_idle_call from do_idle+0x218/0x2c4
[ 7373.086446]  do_idle from cpu_startup_entry+0x18/0x1c
[ 7373.091451]  cpu_startup_entry from rest_init+0xf0/0x194
[ 7373.096764]  rest_init from arch_post_acpi_subsys_init+0x0/0x8
[ 7373.098061] exynos5-hsi2c 12ca0000.i2c: rx timeout
[ 7373.102473] irq event stamp: 21186618
[ 7373.112934] hardirqs last  enabled at (21186646): [<c01ac214>] __up_console_sem+0x50/0x60
[ 7373.121276] hardirqs last disabled at (21186665): [<c01ac200>] __up_console_sem+0x3c/0x60
[ 7373.129714] softirqs last  enabled at (21185954): [<c0101690>] __do_softirq+0x354/0x630
[ 7373.137884] softirqs last disabled at (21186381): [<c0131064>] __irq_exit_rcu+0x140/0x1ec
[ 7373.146146] ---[ end trace 0000000000000000 ]---
[ 7373.150705] r8152 6-1:1.0 eth0: Tx timeout
[ 7378.206014] xhci-hcd xhci-hcd.8.auto: xHCI host not responding to stop endpoint command
[ 7378.234783] xhci-hcd xhci-hcd.8.auto: xHCI host controller not responding, assume dead
[ 7378.245012] xhci-hcd xhci-hcd.8.auto: HC died; cleaning up
[ 7378.206014] xhci-hcd xhci-hcd[ 7378.255124] r8152 6-1:1.0 eth0: Tx status -108
.8.aut[ 7378.n60869] r8152 6-1:1.0 ech0m Tx stat3s -108
d
[ 7u78.2[ 7378h266411] r8152 6-1:1.  ethr: Tx status -108
er not responding[ 7378.272247] r8152 6-1:1.0-etd0: Tx st ti hch.8.auto:aHus -108
378.28 191]1r015r206-1:1.0 eth0: get
                                    re7i7ters -19
seihe: Tx statu disco nect, 0evtc0 number 2
                -108
[ 7378.272247] r8152 6-1:1.0 eth0: Tx status -108
[ 7378.281198] r8152 6-1:1.0 eth0: get_registers -19
[ 7378.288316] usb 6-1: USB disconnect, device number 2
[ 7378.342088] r8152 6-1:1.0 eth0: Get ether addr fail
[ 7378.342088] r8152 6-1:1.0 eth0: Get ether addr fail

meveric · Post by **meveric** » Thu Mar 23, 2023 3:14 am

hmm can't say I have seen similar issues, but I'm also not using network for extended periods of times.

Did the issue happen with Kernel 6.2.1 as well or did it only start with 6.2.7?

johannfr · Post by **johannfr** » Thu Mar 23, 2023 4:03 am

meveric wrote: ↑
Thu Mar 23, 2023 3:14 am
Did the issue happen with Kernel 6.2.1 as well or did it only start with 6.2.7?

It happened with 6.2.1 as well. However, I'm starting to suspect issues with either the SD-card itself or the SD-card interface/controller.

I'm going to experiment with that a bit, and I'll let you know.

johannfr · Post by **johannfr** » Thu Mar 23, 2023 4:29 am

Strike my last remark.

I'm able to reproduce it *fairly* quickly by inducing a lot of network-traffic on the machine, e.g.
On one machine:

Code: Select all

cat /dev/urandom | nc -l 5890

And on the HC2:

Code: Select all

nc <ip-address-of-other-machine> 5890 > /dev/null

After a short while, I get the kernel-error message I previously posted on my serial-terminal. However, I previously made the error of assuming the HC2 had completely hanged, which is not the case, but the Ethernet interface is completely down, and won't even show up in ip link show.

meveric · Post by **meveric** » Thu Mar 23, 2023 5:02 am

how long is fairly quickly?

I'm currently running the LAN <-> WLAN which is "only" 22 MiB/s so roughly 175 Mbit/s and running the test for several minutes already and have not seen the error.
Is there a specific time frame I should be waiting?
What is the network speed you can see during your test?

I will switch to a LAN <-> LAN test later to compare if the speed is also a factor, but up to now it's running stable.

Edit:
30mins of LAN <-> WLAN showed no issue.
Switched to LAN <-> LAN maxing out at about 675 Mbit/s (~84MiB/s)

Edit 2:
40mins and 2TB network traffic later.. still nothing new:

johannfr · Post by **johannfr** » Thu Mar 23, 2023 6:22 am

Happens after a few minutes. It's running on a 1000Mbit network (don't know what the actual throughput is).
I've tried flashing the official Ubuntu-minimal 22.04 image and booted from it. Doesn't have the same problem.
TBH, I think I'll just run that official Ubuntu image on my board - problem "solved".

But thank you so much for responding so quickly!

catzilla · Post by **catzilla** » Tue Mar 28, 2023 1:34 am

Hey all,

I can't seem to get wireless working on xu4, I was hoping someone could help me.

I added wlan0 to /etc/network/interfaces, but it looks like wlan0 doesn't exist!
I also tried running setup-odroid and installing alternative wifi kernel modules, which seemingly had no effect whatsoever. The output below did not change before and after.

Code: Select all

root@odroid:/sys/class# iwconfig
lo        no wireless extensions.

sit0      no wireless extensions.

ip6tnl0   no wireless extensions.

eth0      no wireless extensions.

root@odroid:/sys/class# ifconfig
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.1.233  netmask 255.255.255.0  broadcast 192.168.1.255
        inet6 fe80::21e:6ff:fe30:7155  prefixlen 64  scopeid 0x20<link>
        ether 00:1e:06:30:71:55  txqueuelen 1000  (Ethernet)
        RX packets 3580  bytes 622343 (607.7 KiB)
        RX errors 0  dropped 2  overruns 0  frame 0
        TX packets 2618  bytes 1760011 (1.6 MiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 1000  (Local Loopback)
        RX packets 1467  bytes 1709317 (1.6 MiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 1467  bytes 1709317 (1.6 MiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

Note I am running the image directly from an sdcard instead of flashing uboot to emmc, would that potentially be the issue?

meveric · Post by **meveric** » Tue Mar 28, 2023 3:38 am

It makes no difference if the Image is on SD or eMMC.
But it does make a difference what WLAN adapter you use.
The ODROID XU4 does not have a build on WiFi so it depends on what adapter you use if it's supported or not.

catzilla · Post by **catzilla** » Tue Mar 28, 2023 4:17 am

Thank you for responding!
Well, I do have a USB wifi card that's supposedly supported:

Code: Select all

root@odroid:~# lsusb | grep "Wi"
Bus 003 Device 004: ID 148f:5372 Ralink Technology, Corp. RT5372 Wireless Adapter
root@odroid:~# inxi -N
Network:   Device-1: Ralink RT5372 Wireless Adapter type: USB driver: usb-network
           Device-2: Realtek RTL8153 Gigabit Ethernet Adapter type: USB driver: r8152

Any tips on getting it working? How does one roll back the alternative kernel modules, could that be part of the issue? I wasn't aware that xu4 doesn't have a built-in wifi card when making the original post earlier as it was just given to me by a friend earlier today, and plugged in the wifi card I had laying around before this post.

The output from the first post is unchanged, however, will no wlan0.

------
Edit:
Solved with

Code: Select all

sudo modprobe rt2800usb

meveric · Post by **meveric** » Wed Mar 29, 2023 5:25 am

odd that it doesn't find it automatically.
One thing you should check is if firmware-ralink is installed.. could be that this is the reason it was not finding the first one.. the second one is a Realtek the firmware for that should already be installed.

skull · Post by **skull** » Mon Apr 03, 2023 8:44 pm

Hi meveric!

First of all, many thanks for your work.
Recently I have flashed your bullseye image on emmc without success because I did something wrong.
Now is booting after following this instructions https://github.com/home-assistant/opera ... oid-xu4.md with a temporal sd card.
But I have problems with cloudshell2 lcd, it is off. I have modified /boot/config.ini file adding to the end

Code: Select all

overlay_profile="hktft_cs_ogst"

[overlay_hktft_cs_ogst]
overlays="hktft-cs-ogst i2c0 i2c1"

But after reboot nothing happens

Code: Select all

[root@cloudshell:~# i2cdetect -l
i2c-4	i2c       	exynos5-i2c                     	I2C adapter
i2c-2	i2c       	s3c2410-i2c                     	I2C adapter

I have also run this command that I have read on this thread viewtopic.php?t=40942 without success.

Code: Select all

fdtoverlay -v -i /boot/dtb -o /boot/dtb /boot/dtbs/6.2.8-armmp-lpae/overlays/hktft-cs-ogst.dtbo /boot/dtbs/6.2.8-armmp-lpae/overlays/i2c0.dtbo /boot/dtbs/6.2.8-armmp-lpae/overlays/i2c1.dtbo

What am I doing wrong? Is there any package missing?

Thanks

meveric · Post by **meveric** » Tue Apr 04, 2023 8:39 am

skull wrote: ↑
Mon Apr 03, 2023 8:44 pm
But I have problems with cloudshell2 lcd, it is off. I have modified /boot/config.ini file adding to the end
Code: Select all
overlay_profile="hktft_cs_ogst"

[overlay_hktft_cs_ogst]
overlays="hktft-cs-ogst i2c0 i2c1"
But after reboot nothing happens

That does not surprise me.
Just adding "random lines" to the config.ini will not result in anything happening.
I do neither use "overlay_profile" nor "overlays" as such, so anything you add like this to the config.ini will simply be ignored, as there is no code handling these lines anywhere.

Instead if you check the config.ini you will find, there is a section about device tree overlays, looking something like this:

Code: Select all

; Device Tree Overlay
; some available overlays hktft32 hktft35 pcf8563 spi0 sx865x-i2c1 w1-gpio_p15 w1-gpio_p22 uart0 uart1
; check /usr/lib/linux-image-<Kernel-version>/overlays for full list
; fk_overlays=spi0 pcf8563 sx865x-i2c1
; fk_overlays=hktft35 sx865x-i2c1

Here you can define the overlays you want to load, for example:

Code: Select all

fk_overlays=i2c0 i2c1 hktft-cs-ogst

I can't guarantee it will work though. Apparently there were some known issues in the past with higher kernel version and some of the overlays.
I haven't tested it in quite a while.

skull · Post by **skull** » Tue Apr 04, 2023 3:18 pm

Hi meveric,

Thanks for your reply. I have also tried fk_overlarys before without success. Odroid XU4 does not boot and blue light heartbeat very fast.
So, there is no way to have cloudshell lcd with your bullseye image?

meveric · Post by **meveric** » Wed Apr 05, 2023 6:11 am

I haven't tried it in a while.. let me see if I find a free ODROID that I can use for testing.
I know with Kernel 5.19 it did not work, and it was said to be worked on again in 6.2.
So let's check if something changed.. But I'll need some time to find a setup where I can test this

skull · Post by **skull** » Wed Apr 05, 2023 3:34 pm

Thanks for wasting your time on this!

meveric · Post by **meveric** » Fri Apr 07, 2023 8:17 pm

so here is how far I got:
Overlays itself work fine, I can add i2c0 i2c1 sx865x-i2c1 and even hktft-cs-ogst and they work "fine".
Meaning they show up if I check:

Code: Select all

root@odroid-bullseye:~# i2cdetect -l
i2c-1	i2c       	exynos5-i2c                     	I2C adapter
i2c-4	i2c       	exynos5-i2c                     	I2C adapter
i2c-2	i2c       	s3c2410-i2c                     	I2C adapter
i2c-0	i2c       	s3c2410-i2c                     	I2C adapter

Without Display:

Code: Select all

root@odroid-bullseye:~# gpioinfo | grep "\[used\]"
	line   3:      unnamed  "power key"   input   active-low [used]
	line   4:      unnamed  "interrupt"   input  active-high [used]
	line   0:      unnamed      "reset"  output   active-low [used]
	line   7:      unnamed        "hpd"   input  active-high [used]
root@odroid-bullseye:~# ll /dev/fb*
crw-rw---- 1 root video 29, 0 Apr  7 11:10 /dev/fb0
root@odroid-bullseye:~# lsmod
Module                  Size  Used by
exynos_gsc             40960  0
s5p_mfc               151552  0
s5p_jpeg               49152  0
videobuf2_dma_contig    20480  3 s5p_jpeg,s5p_mfc,exynos_gsc
videobuf2_memops       16384  1 videobuf2_dma_contig
v4l2_mem2mem           36864  2 s5p_jpeg,exynos_gsc
videobuf2_v4l2         32768  4 s5p_jpeg,s5p_mfc,exynos_gsc,v4l2_mem2mem
videobuf2_common       69632  7 s5p_jpeg,videobuf2_dma_contig,s5p_mfc,exynos_gsc,videobuf2_memops,v4l2_mem2mem,videobuf2_v4l2
videodev              258048  5 s5p_jpeg,s5p_mfc,exynos_gsc,v4l2_mem2mem,videobuf2_v4l2
mc                     61440  4 videobuf2_common,videodev,v4l2_mem2mem,videobuf2_v4l2
s5p_cec                16384  0
exynos_gpiomem         16384  0
fuse                  131072  1
ip_tables              24576  0
x_tables               32768  1 ip_tables
ext4                  901120  2
mbcache                16384  1 ext4

With Display "hktft-cs-ogst" activated:

Code: Select all

root@odroid-bullseye:~# gpioinfo | grep "\[used\]"
	line   3:      unnamed  "power key"   input   active-low [used]
	line   4:      unnamed  "interrupt"   input  active-high [used]
	line   0:      unnamed      "reset"  output   active-low [used]
	line   5:      unnamed      "reset"  output  active-high [used]
	line   6:      unnamed         "dc"  output  active-high [used]
	line   5:      unnamed   "spi0 CS0"  output   active-low [used]
	line   1:      unnamed   "spi0 CS1"  output  active-high [used]
	line   7:      unnamed        "hpd"   input  active-high [used]
root@odroid-bullseye:~# ll /dev/fb*
crw-rw---- 1 root video 29, 0 Apr  7 11:12 /dev/fb0
crw-rw---- 1 root video 29, 1 Apr  7 11:12 /dev/fb1
root@odroid-bullseye:~# lsmod
Module                  Size  Used by
ili9341                16384  0                          <-------------------
drm_mipi_dbi           28672  1 ili9341                  <-------------------
drm_dma_helper         16384  1 ili9341                  <-------------------
exynos_gsc             40960  0
s5p_jpeg               49152  0
v4l2_mem2mem           36864  2 s5p_jpeg,exynos_gsc
s5p_mfc               151552  0
videobuf2_dma_contig    20480  3 s5p_jpeg,s5p_mfc,exynos_gsc
videobuf2_memops       16384  1 videobuf2_dma_contig
videobuf2_v4l2         32768  4 s5p_jpeg,s5p_mfc,exynos_gsc,v4l2_mem2mem
videobuf2_common       69632  7 s5p_jpeg,videobuf2_dma_contig,s5p_mfc,exynos_gsc,videobuf2_memops,v4l2_mem2mem,videobuf2_v4l2
videodev              258048  5 s5p_jpeg,s5p_mfc,exynos_gsc,v4l2_mem2mem,videobuf2_v4l2
mc                     61440  4 videobuf2_common,videodev,v4l2_mem2mem,videobuf2_v4l2
s5p_cec                16384  0
exynos_gpiomem         16384  0
fuse                  131072  1
ip_tables              24576  0
x_tables               32768  1 ip_tables
ext4                  901120  2
mbcache                16384  1 ext4

Sadly the display itself stays black

No backlight, no output.
If I remember correctly someone said spi is broken for upstream Kernel.. but I don't know what that means or how to fix it.

I'm also not sure if the GPIOs are set correctly.
Since upstream kernel rework the GPIO stuff quite often, it could be that the mapping in the overlay is incorrect, but I'm not too familiar with that.

alvy · Post by **alvy** » Fri May 26, 2023 6:37 am

meveric wrote: ↑
Wed Jul 13, 2022 6:10 am
Please Note: the image requires a newer version of u-boot!
This means while you can flash the image on an SD card and it will work out of the box, for eMMC the image only works if you flash the u-boot on the eMMC yourself.
For this you need the hardkernel_1mb_uboot flashing tool https://github.com/hardkernel/u-boot/tr ... _1mb_uboot with the u-boot attached to this post.
eMMC was tested and confirmed to be working fine.

meveric, could you please explain how to do this? I don't understand how to use the flashing tool and what files need to be flashed.

meveric · Post by **meveric** » Fri May 26, 2023 4:14 pm

1. download the image in the first post, and download the u-boot.bin.tar.gz
2. Flash the image on your eMMC
3. either use git clone https://github.com/hardkernel/u-boot/ -b odroidxu3-v2012.07
or download the files in the link: https://github.com/hardkernel/u-boot/tr ... _1mb_uboot manually.
4. You need a Linux PC or Virtual Machine where you have connected the eMMC to
5. Copy the files you downloaded from https://github.com/hardkernel/u-boot/tr ... _1mb_uboot to a folder on that Linux machine, extract the u-boot.bin.tar.gz in the same folder
6. run bash ./sd_fusing.sh <device of your eMMC module> to flash the new u-boot on the eMMC.
For example bash ./sd_fusing.sh /dev/sdb

alvy · Post by **alvy** » Sat May 27, 2023 12:30 am

meveric wrote: ↑
Fri May 26, 2023 4:14 pm
1. download the image in the first post, and download the u-boot.bin.tar.gz
2. Flash the image on your eMMC
3. either use git clone https://github.com/hardkernel/u-boot/ -b odroidxu3-v2012.07
or download the files in the link: https://github.com/hardkernel/u-boot/tr ... _1mb_uboot manually.
4. You need a Linux PC or Virtual Machine where you have connected the eMMC to
5. Copy the files you downloaded from https://github.com/hardkernel/u-boot/tr ... _1mb_uboot to a folder on that Linux machine, extract the u-boot.bin.tar.gz in the same folder
6. run bash ./sd_fusing.sh <device of your eMMC module> to flash the new u-boot on the eMMC.
For example bash ./sd_fusing.sh /dev/sdb

Thank you. I booted my xu4 from sd card with your bullseye image and then get an unexpected token error in step 6 when trying to copy the files to the emmc module with the flashing tool:

root@xu4:/home/rene/Downloads# df

Filesystem     1K-blocks     Used Available Use% Mounted on

udev              955424        0    955424   0% /dev

tmpfs             202320     3512    198808   2% /run

/dev/mmcblk1p2  30059552 14384808  14428272  50% /

tmpfs            1011600        0   1011600   0% /dev/shm

tmpfs               5120        4      5116   1% /run/lock

/dev/mmcblk1p1    498900    54740    407468  12% /boot

tmpfs             202320       60    202260   1% /run/user/1000

tmpfs             202320       28    202292   1% /run/user/0



root@xu4:/home/rene/Downloads# ls -l

-rw-r--r-- 1 root root 142868 May 26 16:51 bl1.bin.hardkernel

-rw-r--r-- 1 root root 143128 May 26 16:52 bl2.bin.hardkernel.1mb_uboot

-rw-r--r-- 1 root root 153804 May 26 16:52 movi_partition.patch

-rw-r--r-- 1 root root 174277 May 26 16:52 sd_fusing.1M.sh

-rw-r--r-- 1 root root 142889 May 26 16:52 tzsw.bin.hardkernel

-rw-r--r-- 1 root root 542720 May 26 17:00 u-boot.bin.tar



root@xu4:/home/rene/Downloads# fdisk /dev/mmcblk0



Welcome to fdisk (util-linux 2.36.1).

Changes will remain in memory only, until you decide to write them.

Be careful before using the write command.



Command (m for help): p

Disk /dev/mmcblk0: 14.56 GiB, 15634268160 bytes, 30535680 sectors

Units: sectors of 1 * 512 = 512 bytes

Sector size (logical/physical): 512 bytes / 512 bytes

I/O size (minimum/optimal): 512 bytes / 512 bytes

Disklabel type: dos

Disk identifier: 0x000f1766



Device         Boot   Start     End Sectors  Size Id Type

/dev/mmcblk0p1         8192 1056767 1048576  512M 83 Linux

/dev/mmcblk0p2      1056768 2498559 1441792  704M 83 Linux



Command (m for help): q



root@xu4:/home/rene/Downloads# bash ./sd_fusing.1M.sh /dev/mmcblk0

./sd_fusing.1M.sh: line 8: syntax error near unexpected token `newline'

./sd_fusing.1M.sh: line 8: `<!DOCTYPE html>'

mad_ady · Post by **mad_ady** » Sat May 27, 2023 3:19 am

You didn't download the sd fusing script, but it looks like you're trying to execute a webpage instead. The actual script is about 10-20 lines in length.
Edit: sorry, it's longer. Here's the actual content: https://raw.githubusercontent.com/hardk ... sing.1M.sh

Make sure the other files are valid, too.

meveric · Post by **meveric** » Sat May 27, 2023 3:53 am

if the downloaded flashing script is incorrect, it's most likely that the other files are downloaded incorrectly as well.

To downloads files manually from: https://github.com/hardkernel/u-boot/tr ... _1mb_uboot
You first need to click the file and then either click the download button or the "raw" button before you can save the file.
If you just right click and do "save to" you will only download the website that contains the link to the file, not the file itself.

On top of that you have u-boot.bin.tar while it should be u-boot.bin, extracting the tar.bz2 should be done with tar xvf u-boot.bin.tar.bz2

Also be very careful with the device you're using.
I'm not sure if /dev/mmcblk0 is the eMMC, it could also be the SD instead if your XU4 has SD card and eMMC connected at the same time it could easily be the wrong device.
So make sure you pick the correct one, compare sizes carefully to determine if you have to use /dev/mmcblk0 or /dev/mmcblk1

alvy · Post by **alvy** » Sat May 27, 2023 9:37 am

I'm too stupid to download a bunch of files. Finally this worked for me:

Install git and clone u-boot

Code: Select all

root@xu4:/home/rene/Downloads# apt-get install git
...
root@xu4:/home/rene/Downloads# git clone https://github.com/hardkernel/u-boot/ -b odroidxu3-v2012.07

Navigate to hardkernel_1mb_uboot subfolder:

Code: Select all

root@xu4:/home/rene/Downloads# cd u-boot
root@xu4:/home/rene/Downloads/u-boot# cd sd_fuse
root@xu4:/home/rene/Downloads/u-boot/sd_fuse# cd hardkernel_1mb_uboot
root@xu4:/home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot# ls -l

-rw-r--r-- 1 root root  15616 May 27 01:48 bl1.bin.hardkernel
-rw-r--r-- 1 root root  14592 May 27 01:48 bl2.bin.hardkernel.1mb_uboot
-rw-r--r-- 1 root root    536 May 27 01:48 movi_partition.patch
-rwxr-xr-x 1 root root   1894 May 27 01:48 sd_fusing.1M.sh
-rw-r--r-- 1 root root 262144 May 27 01:48 tzsw.bin.hardkernel

There are my files.

Leftclick meveric's u-boot file, move it to the hardkernel_1mb_uboot folder, extract, change owner to root and remove the compressed file:

Code: Select all

root@xu4:/home/rene/Downloads# mv u-boot.bin.tar.gz /home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot
root@xu4:/home/rene/Downloads# cd /home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot
root@xu4:/home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot# tar xvf u-boot.bin.tar.gz
root@xu4:/home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot# ls -l
-rw-r--r-- 1 root root  15616 May 27 01:48 bl1.bin.hardkernel
-rw-r--r-- 1 root root  14592 May 27 01:48 bl2.bin.hardkernel.1mb_uboot
-rw-r--r-- 1 root root    536 May 27 01:48 movi_partition.patch
-rwxr-xr-x 1 root root   1894 May 27 01:48 sd_fusing.1M.sh
-rw-r--r-- 1 root root 262144 May 27 01:48 tzsw.bin.hardkernel
-rw-r--r-- 1 rene rene 541116 Aug 17  2022 u-boot.bin
-rw-r--r-- 1 rene rene 320329 May 27 02:05 u-boot.bin.tar.gz
root@xu4:/home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot# chown root:root u-boot.bin
root@xu4:/home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot# rm u-boot.bin.tar.gz
root@xu4:/home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot# ls -l
-rw-r--r-- 1 root root  15616 May 27 01:48 bl1.bin.hardkernel
-rw-r--r-- 1 root root  14592 May 27 01:48 bl2.bin.hardkernel.1mb_uboot
-rw-r--r-- 1 root root    536 May 27 01:48 movi_partition.patch
-rwxr-xr-x 1 root root   1894 May 27 01:48 sd_fusing.1M.sh
-rw-r--r-- 1 root root 262144 May 27 01:48 tzsw.bin.hardkernel
-rw-r--r-- 1 root root 541116 Aug 17  2022 u-boot.bin

If both are present the xu4 assigns mmcblk0 to my 16gb emmc module and mmcblk1 to my 32gb sd card which are thus identifiable by size.

Code: Select all

root@xu4:/home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot# bash ./sd_fusing.1M.sh /dev/mmcblk0
+ '[' -z /dev/mmcblk0 ']'
+ '[' -b /dev/mmcblk0 ']'
+ echo '/dev/mmcblk0 reader is identified.'
/dev/mmcblk0 reader is identified.
+ '[' -d /sys/block/mmcblk0boot0 ']'
+ echo '/dev/mmcblk0 is an eMMC card, disabling mmcblk0boot0 ro'
/dev/mmcblk0 is an eMMC card, disabling mmcblk0boot0 ro
+ echo -n 0
+ sudo tee /sys/block/mmcblk0boot0/force_ro
0+ emmc=1
+ '[' -n 1 ']'
+ signed_bl1_position=0
+ bl2_position=30
+ uboot_position=62
+ tzsw_position=2110
+ device=/dev/mmcblk0boot0
+ '[' -f ./u-boot-dtb.bin ']'
+ '[' -f ./u-boot.bin ']'
+ uboot=./u-boot.bin
+ echo 'BL1 fusing'
BL1 fusing
+ sudo dd iflag=dsync oflag=dsync if=./bl1.bin.hardkernel of=/dev/mmcblk0boot0 seek=0
30+1 records in
30+1 records out
15616 bytes (16 kB, 15 KiB) copied, 0.0640457 s, 244 kB/s
+ echo 'BL2 fusing'
BL2 fusing
+ sudo dd iflag=dsync oflag=dsync if=./bl2.bin.hardkernel.1mb_uboot of=/dev/mmcblk0boot0 seek=30
28+1 records in
28+1 records out
14592 bytes (15 kB, 14 KiB) copied, 0.0544177 s, 268 kB/s
+ echo 'u-boot fusing'
u-boot fusing
+ sudo dd iflag=dsync oflag=dsync if=./u-boot.bin of=/dev/mmcblk0boot0 seek=62
1056+1 records in
1056+1 records out
541116 bytes (541 kB, 528 KiB) copied, 2.24008 s, 242 kB/s
+ echo 'TrustZone S/W fusing'
TrustZone S/W fusing
+ sudo dd iflag=dsync oflag=dsync if=./tzsw.bin.hardkernel of=/dev/mmcblk0boot0 seek=2110
512+0 records in
512+0 records out
262144 bytes (262 kB, 256 KiB) copied, 1.1441 s, 229 kB/s
+ echo 'U-boot image is fused successfully.'
U-boot image is fused successfully.
+ echo 'Eject /dev/mmcblk0 and insert it again.'
Eject /dev/mmcblk0 and insert it again.
root@xu4:/home/rene/Downloads/u-boot/sd_fuse/hardkernel_1mb_uboot# shutdown -h now

After shutdown I removed the sd card and switched the selector over to emmc boot. The xu4 then booted from emmc nicely.

meveric · Post by **meveric** » Sat May 27, 2023 6:01 pm

Glad it worked

alvy · Post by **alvy** » Mon May 29, 2023 5:15 am

The fan is constantly starting and stopping and from elevated temperatures the fan never stops again even in idle at <50°C. Hardkernel's fan control tutorial doesn't work for me at all (https://magazine.odroid.com/article/fan ... -settings/).

root@xu4:/sys/devices/platform/pwm-fan/hwmon/hwmon0# cat /sys/devices/virtual/thermal/thermal_zone{0,1,2,3}/trip_point_{0,1,2}_temp

50000

60000

70000

50000

60000

70000

50000

60000

70000

50000

60000

70000



root@xu4:/sys/devices/platform/pwm-fan/hwmon/hwmon0# ls -l



lrwxrwxrwx 1 root root    0 May 28 19:48 device -> ../../../pwm-fan

-r--r--r-- 1 root root 4096 May 28 19:48 name

lrwxrwxrwx 1 root root    0 May 28 19:48 of_node -> ../../../../../firmware/devicetree/base/pwm-fan

drwxr-xr-x 2 root root    0 May 28 19:48 power

-rw-r--r-- 1 root root 4096 May 28 19:48 pwm1

-rw-r--r-- 1 root root 4096 May 28 19:48 pwm1_enable

lrwxrwxrwx 1 root root    0 May 28 19:40 subsystem -> ../../../../../class/hwmon

-rw-r--r-- 1 root root 4096 May 28 19:40 uevent

I get reasonable temperatures and I'm able to switch off the fan manually (echo 0 | sudo tee /sys/devices/platform/pwm-fan/hwmon/hwmon0/pwm1) but it immediately kicks in again so it shouldn't be an hardware issue:

root@xu4:/sys/devices/platform/pwm-fan/hwmon/hwmon0# cat /sys/devices/virtual/thermal/thermal_zone{0,1,2,3}/temp

41000

42000

45000

42000

Is there another way to configure the fan with this image?

meveric · Post by **meveric** » Mon May 29, 2023 6:03 am

The guide was made for older Kernel versions and does not apply for this image.

To be honest I'm not sure if you can manually adjust trip points.
It should work by adjusting the .dtb file, but that's a bit complicated and would have to be done for each Kernel update.

I'll see if I can find out more.

zyssai · Post by **zyssai** » Mon May 29, 2023 6:04 pm

Hi!

I'm new to XU4, don't know what I am doing wrong but I am facing Bootloop.
Tried 3 different SD cards up to 16GB.
Same behavior.

I just see no space left on drive, but as it seems your startup script resizes partition, I don't know where it stucks.

Any help appreciated

Edit: problem seems to be resolved with a freshly brand new 5V adapter

Will buy a new one asap.
Didn't know XU4 was so restreint with adapter voltage.

meveric · Post by **meveric** » Mon May 29, 2023 10:51 pm

More likely it's a current issue. The XU4 can draw up to 4A.. While you may have had the correct voltage, if your PSU only supplies 1 or 2A it could have been to little for the XU4

zyssai · Post by **zyssai** » Mon May 29, 2023 10:56 pm

meveric wrote: ↑
Mon May 29, 2023 10:51 pm
More likely it's a current issue. The XU4 can draw up to 4A.. While you may have had the correct voltage, if your PSU only supplies 1 or 2A it could have been to little for the XU4

Yes I have two 2A adapters, no one was sufficient. Impressive for a such tiny board.
And for the PSU I needed to short all the 5V rails to have a stable power. And even with this, sometimes it even reboots

zyssai · Post by **zyssai** » Tue May 30, 2023 1:09 am

So everything is working perfectly, thanks

Just one question please,
Do you know if VU7A+ touchscreen can work with this debian?

meveric · Post by **meveric** » Wed May 31, 2023 5:04 am

I haven't tested all accessories yet.
The display itself probably runs out of the box as standard HDMI, touch input, if not working yet, may only require some adjustment for the Kernel modules, but I can test this with the VU7+ to see if it works in general.
Audio I'm not sure, I currently don't have the VU7A+ so I can't confirm if audio would work with the upstream Kernel.

zyssai · Post by **zyssai** » Wed May 31, 2023 2:36 pm

Thanks. Screens works perfectly even if I don't manage to get recommended resolution in settings (1024x600). But not a problem.

No need for audio in my case.

The only need is touchscreen. I think VU7+ and VU7A+ have similar touchscreen panel.
I will try with last Ubuntu when will have time and let you know