[OS] Debian Stretch

meveric · Post by **meveric** » Sun Mar 25, 2018 11:49 pm

This is the first version of my Debian Stretch image for the ODROID XU3/XU4-Series.

It is uses the 4.14 LTS Kernel by default, as a Kernel >= 3.16 is require for Debian Stretch, and also for G2D support on the Desktop.

It's a headless server image only with user root. But it has all my repositories included, which allows for easy installation and updates of packages such as Kernel and Headers and other packages.

The image has my usual setup: means on first boot it's resizing the rootfs partition and configures SSH. It will automatically reboot after the initial setup after which this image is ready to use.
Kernel and headers are already installed if you need to build your own drivers.
A few basic tools such as htop, mc, vim and bash-completion are already installed for your convenience.

Download: https://oph.mdrjr.net/meveric/images/St ... XU4.img.xz md5 sha512 sig

Size compressed: 108 MB
Size uncompressed: 584 MB

Default Login: root
Default Password: odroid

It's ready to be used as a server image.

First things first:
Since there's constantly new development for this image and ODROIDs in general the First thing you should do after the image is up and running is to install all updates:

Code: Select all

apt-get update && apt-get upgrade && apt-get dist-upgrade

How to convert the image into a Desktop image:
The image can easily converted into a full Desktop image looking similar to this:

I created a setup tool that's already preinstalled:

Code: Select all

setup-odroid

Which should easily allow you to install a Desktop Environment.

Hints:
Changing language:
If you install a Desktop environment you should already have keybord-configuration installed, it will ask upon the first installation what keyboard layout it should use.
This configuration is for your X11 Desktop so you have your desired keyboard layout.
For console you need to install console-setup to do the same, so just in case:

Code: Select all

apt-get install console-setup keyboard-configuration
dpkg-reconfigure keyboard-configuration

should be all that is needed to change your keyboard layout.
You probably want to set timezone as well

Code: Select all

dpkg-reconfigure tzdata

Installing Chromium Webbrowser (with hardware acceleration):
Chromium is part of the Default Debian Stretch repository and can be installed with:

Code: Select all

apt-get install chromium

You will need to configure 3D acceleration for yourself.
Therefore you need to create a config file under /etc/chromium.d/ and add the following line:

Code: Select all

CHROMIUM_FLAGS="$CHROMIUM_FLAGS --use-gl=egl --ignore-gpu-blacklist --disable-accelerated-2d-canvas --num-raster-threads=4 --audio-buffer-size=4096"

Chromium also does not come with OpenGL ES GPU drivers configured correctly and you need to add them manually for Chromium to work.

Code: Select all

ln -sf /usr/lib/arm-linux-gnueabihf/libGLESv2.so /usr/lib/chromium/libGLESv2.so
ln -sf /usr/lib/arm-linux-gnueabihf/libEGL.so /usr/lib/chromium/libEGL.so

rooted · Post by **rooted** » Mon Mar 26, 2018 2:55 pm

Just in time, I'm tired of running the terribly buggy OMV on my CloudShell 1.

meveric · Post by **meveric** » Mon Mar 26, 2018 3:21 pm

understandable, but Debian Stretch is gonna have it's issues as well for a while.
For example I still can't get Kodi to work properly for me. I've been building different versions left and right.. but something is always off, either sound or video.

rooted · Post by **rooted** » Mon Mar 26, 2018 11:38 pm

I am running headless, mostly just tvheadend and comskip.

It's actually Armbian, not OMV. OMV is on my CloudShell 2 and it's a buggy mess as well, due to Armbian. Not sure why but damn it's terrible, your Jessie minimal image is the most stable but the software libraries are just too old to get some of the stuff I need compiled.

Hopefully Stretch will have updated enough libraries.

meveric · Post by **meveric** » Mon Mar 26, 2018 11:41 pm

should have.. if not let me know what's missing I can add a lot to my repository

SamuraiCrow · Post by **SamuraiCrow** » Wed Mar 28, 2018 5:34 pm

Is your subdomain configured properly? I tried 'apt-get update' but some of the servers weren't found.

meveric · Post by **meveric** » Wed Mar 28, 2018 6:24 pm

Just installed updates a couple of minutes ago (new firefox esr release) and I haven't had any issues.
What servers weren't found?

jeanrob98 · Post by **jeanrob98** » Mon Apr 02, 2018 2:01 am

This is excellent news.
Thanks very much for the work that went into this!
Just one issue:
after using the setup-odroid script and installing the Mate desktop,
the X server was looking for armsoc, so I installed xf86-video-armsoc-odroid.
After that X started up properly.

meveric · Post by **meveric** » Mon Apr 02, 2018 5:04 am

hmm normally the setup-odroid script should install the armsoc driver as well.
Have you seen an error, that it could not be installed or something?
Anyway I'll see what I can find out. Thanks for reporting.

jeanrob98 · Post by **jeanrob98** » Mon Apr 02, 2018 6:00 am

meveric wrote:hmm normally the setup-odroid script should install the armsoc driver as well.
Have you seen an error, that it could not be installed or something?
Anyway I'll see what I can find out. Thanks for reporting.

Nope sorry. I spotted nothing that would have given me a hint.

francisscott · Post by **francisscott** » Mon Apr 02, 2018 6:28 pm

I have some problems.

When I use setup-odroid to install mate, it asks to create a new user. But when the user is created, it returns to main menu. Then I try to install mate again and it asks again to create an user. If I answer no, it returns to the main menu.

If I reboot, I can't connect to repositories anymore: apt-get update returns: Could not resolve host: oph.mdrjr.net.

Thank for your help.

meveric · Post by **meveric** » Mon Apr 02, 2018 6:37 pm

That it returns to the menu is normal you install the desktop then the user and then it needs to reboot to load into MATE.
Installing MATE will not instantly start the Desktop.
The reason why apt-get update does not work anymore is cause /etc/network/interfaces is changed to remove entry for network adapter, as with a Desktop the Network Manager is installed and that one is suppose to take over the network configuration.

Can you give me the output of the following command:

Code: Select all

apt-cache policy xf86-video-armsoc-odroid

francisscott · Post by **francisscott** » Mon Apr 02, 2018 9:39 pm

Thank for your answer.

What I mean is that there is no installation at all.

But I found out that the problem is that the partition was not resized. Now I am installing mate and it seems working.

justin0a0 · Post by **justin0a0** » Tue Apr 03, 2018 9:15 pm

francisscott wrote:But I found out that the problem is that the partition was not resized. Now I am installing mate and it seems working.

This is a thing I have also noticed the first run scripts don't ever get ran

meveric · Post by **meveric** » Tue Apr 03, 2018 9:19 pm

justin0a0 wrote:This is a thing I have also noticed the first run scripts don't ever get ran

Hmm this worked last time I tried, I will try again and see if something changed here.
Maybe I forgot something when I created the .img file.

gom · Post by **gom** » Wed Apr 04, 2018 1:15 am

I just got my xu4 and install stretch using your image. Everything worked except for the repartitioning. I have tried for several hours to get sound functioning. I have the xu4 connected via hdmi. I can get sound from the monitor using "speaker-test -c2 --test=wav -w /usr/share/sounds/alsa/Front_Center.wav" but nothing from firefox. Vlc replies "Unable to find an X11 visual which matches EGL config 9" and "Segmentation fault" so I can not test that. What is the trick to getting this to work? I have a usb sound device if nothing else will work.

justin0a0 · Post by **justin0a0** » Wed Apr 04, 2018 1:35 am

meveric wrote:Maybe I forgot something when I created the .img file.

noticed a line missing in rc.local
/root/scripts/first-boot

meveric · Post by **meveric** » Wed Apr 04, 2018 2:23 am

justin0a0 wrote:
meveric wrote:Maybe I forgot something when I created the .img file.
noticed a line missing in rc.local
/root/scripts/first-boot

Arg, that should be there.. I thought I added that.. Thanks for the report I'm gonna fix that.

meveric · Post by **meveric** » Wed Apr 04, 2018 2:53 am

gom wrote:I just got my xu4 and install stretch using your image. Everything worked except for the repartitioning. I have tried for several hours to get sound functioning. I have the xu4 connected via hdmi. I can get sound from the monitor using "speaker-test -c2 --test=wav -w /usr/share/sounds/alsa/Front_Center.wav" but nothing from firefox.

I just checked on my image and sound is working fine. I had no issues what so ever.
Are you using a desktop environment like MATE or what are you using to run firefox?
Normally desktop environments and firefox run on pulseaudio and as long as this is up and running you shouldn't have any issue.

gom wrote:Vlc replies "Unable to find an X11 visual which matches EGL config 9" and "Segmentation fault" so I can not test that.

VLC sucks on ARM platforms, still it should work if you set the rendering engine to "OpenGL for Embedded Systems 2 video output"
But VLC sucks it's one of the worst players for ARM at the moment.

gom wrote:What is the trick to getting this to work? I have a usb sound device if nothing else will work.

Normally sound should work out of the box.. you can install and run pavucontrol to check if pulseaudio is up and running it should also show if applications use the output.

meveric · Post by **meveric** » Wed Apr 04, 2018 3:13 am

Update:
I've just uploaded a new image which has the first-boot script in /etc/rc.local activated hope this solves most of the issues reported

justin0a0 · Post by **justin0a0** » Wed Apr 04, 2018 4:36 am

in the resize script it runs
resize2fs mmcblk0p2
instead of
resize2fs /dev/mmcblk0p2

fdisk shows the partition resizes just df does not until i run that manually

meveric · Post by **meveric** » Wed Apr 04, 2018 7:22 am

Yes you're right. I've already located the issue and fixed it... -> reuploading

Edit:
All done, hopefully no more surprises

gom · Post by **gom** » Wed Apr 04, 2018 4:54 pm

I was running xfce desktop and was having sound problems. I read in many post that mate worked when nothing else did so I used your new image. The partition resizing works now. Did a fresh install and let setup-odroid install mate. As you said sound works. I have never used mate but it appears not too bad. In recent years I have not run in to these difficulties with different desktop managers. What make mate so odroid friendly?

meveric · Post by **meveric** » Wed Apr 04, 2018 5:04 pm

MATE is based on the "old" Gnome 2 desktop, the resources it uses are rather minimal (in fact it should be the same or even less than recent versions of XFCE). While still offering modern features available on GTK3.
It can be highly modified. In fact on Ubuntu there are many different default layouts for MATE some that mimic the Windows XP style (redmond):

Some that including a dock similar to MacOS (pantheon)

or even a look that aims to act similar to Unity. (munity)

So TL;DR
Low resource requirements
Modern Features
Very customizable

chromodoris · Post by **chromodoris** » Thu Apr 05, 2018 3:24 am

I have installed the stretch image on a HC1 and a HC2. It mostly went well, with just a few issues.

Here are the problems I found:

After the fist series of reboots, I added a few regular Debian repositories and attempted
an apt-get update. It failed with an error "no space left on device". It seemed to be
related to the size of rootfs which was really small. Maybe the initial resize is too
aggressive? I ended up putting th SDcard on my desktop computer and ran
ext4fs and resizefs to have the root file system expand to the full size of the SD card.
After that, the error disappeared and I was able to install a few packages.

Another problem was when I attempted to install cryptsetup. The package (or one of its
dependencies) run initramfs-tools, probably in a post-install script. This fails with
the following error:

Processing triggers for initramfs-tools (0.130) ...
ln: failed to create hard link '/boot/initrd.img-4.14.28+.dpkg-bak' => '/boot/initrd.img-4.14.28+': Operation not permitted

I guess this is due to /boot being a vfat filesystem, which does not seem to support links. Would it be
possible to have /boot as an ext4 filesystem?

Thanks for providing this image!

meveric · Post by **meveric** » Thu Apr 05, 2018 5:43 am

chromodoris wrote:I have installed the stretch image on a HC1 and a HC2. It mostly went well, with just a few issues.

Here are the problems I found:

After the fist series of reboots, I added a few regular Debian repositories and attempted
an apt-get update. It failed with an error "no space left on device". It seemed to be
related to the size of rootfs which was really small. Maybe the initial resize is too
aggressive? I ended up putting th SDcard on my desktop computer and ran
ext4fs and resizefs to have the root file system expand to the full size of the SD card.
After that, the error disappeared and I was able to install a few packages.

This should be fixed since yesterday evening. When did you download the image?

chromodoris wrote:Another problem was when I attempted to install cryptsetup. The package (or one of its
dependencies) run initramfs-tools, probably in a post-install script. This fails with
the following error:

Processing triggers for initramfs-tools (0.130) ...
ln: failed to create hard link '/boot/initrd.img-4.14.28+.dpkg-bak' => '/boot/initrd.img-4.14.28+': Operation not permitted

I guess this is due to /boot being a vfat filesystem, which does not seem to support links. Would it be
possible to have /boot as an ext4 filesystem?

Thanks for providing this image!

It does not fail, it continues just fine, the error can be ignored.
You are right this is due to the the fact that the bootfs is vfat system which does not support links.
Im not sure if ext filesystem works for bootfs, as this is handled in the uboot and if the uboot does not support ext for bootfs it won't boot at all.
I remember that in the past there were some tests about using ext for bootfs but I don't know the state of it.
vfat is uses so you can make "fixes" even under windows and adjust boot.ini or fix some other kernel related issues.

chromodoris · Post by **chromodoris** » Fri Apr 06, 2018 3:44 am

meveric wrote: This should be fixed since yesterday evening. When did you download the image?

I downloaded it on April 3rd, so it was probably before the fix.

meveric wrote: It does not fail, it continues just fine, the error can be ignored.

No, it is a failure. Now each time I run aptitude full-upgrade, I get this:

Code: Select all

root@cerberilla-ambonensis:~# aptitude full-upgrade
The following partially installed packages will be configured:
  initramfs-tools 
No packages will be installed, upgraded, or removed.
0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B of archives. After unpacking 0 B will be used.
Setting up initramfs-tools (0.130) ...   
update-initramfs: deferring update (trigger activated)
Processing triggers for initramfs-tools (0.130) ...
ln: failed to create hard link '/boot/initrd.img-4.14.28+.dpkg-bak' => '/boot/initrd.img-4.14.28+': Operation not permitted
update-initramfs: Generating /boot/initrd.img-4.14.28+
E: busybox or busybox-static, version 1:1.22.0-17~ or later, is required but not installed
update-initramfs: failed for /boot/initrd.img-4.14.28+ with 1.
dpkg: error processing package initramfs-tools (--configure):
 subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
 initramfs-tools
E: Sub-process /usr/bin/dpkg returned an error code (1)
Setting up initramfs-tools (0.130) ...
update-initramfs: deferring update (trigger activated)
Processing triggers for initramfs-tools (0.130) ...
ln: failed to create hard link '/boot/initrd.img-4.14.28+.dpkg-bak' => '/boot/initrd.img-4.14.28+': Operation not permitted
update-initramfs: Generating /boot/initrd.img-4.14.28+
E: busybox or busybox-static, version 1:1.22.0-17~ or later, is required but not installed
update-initramfs: failed for /boot/initrd.img-4.14.28+ with 1.
dpkg: error processing package initramfs-tools (--configure):
 subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
 initramfs-tools
                                         
root@cerberilla-ambonensis:~#

meveric wrote: You are right this is due to the the fact that the bootfs is vfat system which does not support links.
Im not sure if ext filesystem works for bootfs, as this is handled in the uboot and if the uboot does not support ext for bootfs it won't boot at all.
I remember that in the past there were some tests about using ext for bootfs but I don't know the state of it.
vfat is uses so you can make "fixes" even under windows and adjust boot.ini or fix some other kernel related issues.

I tried to convert the boot partition. Boot failed with this message from uboot

Code: Select all

** Unrecognized filesystem type **

However, it seems to me uboot supports ext4 since years, and it seems to work on some Odroid boards,
as per this (old) thread: viewtopic.php?f=112&t=8419.

Not being able to have cryptsetup (which requires initramfs-tools) is a major problem for me. The board will
be used as a remote backup server in an hosting space and I want the data to be safe.

chromodoris · Post by **chromodoris** » Fri Apr 06, 2018 4:08 am

chromodoris wrote: I tried to convert the boot partition. Boot failed with this message from uboot
Code: Select all
** Unrecognized filesystem type **
However, it seems to me uboot supports ext4 since years, and it seems to work on some Odroid boards,
as per this (old) thread: viewtopic.php?f=112&t=8419.

I confirm the uboot embedded with the HC2 does support ext4. Using serial connection and interrupting
boot by pressing [enter] twice, I was able to do

Code: Select all

ext4load mmc 0:1 0x40008000 zImage
ext4load mmc 0:1 0x42000000 uInitrd
ext4load mmc 0:1 0x44000000 exynos5422-odroidxu4.dtb

That is I was able to load the files in the new ext4 boot partition directly from uboot.
I then tried to simply replace all the fatload commands with ext4load commands in boot.ini,
but still failed, as uboot did not access the boot.ini file itself

Perhaps I should try things like setenv/savenv from the uboot prompt.

I definitely think it is possible to have a working ext4 boot partition.

meveric · Post by **meveric** » Fri Apr 06, 2018 6:30 am

chromodoris wrote:

meveric wrote: It does not fail, it continues just fine, the error can be ignored.

No, it is a failure. Now each time I run aptitude full-upgrade, I get this:

Code: Select all

root@cerberilla-ambonensis:~# aptitude full-upgrade
The following partially installed packages will be configured:
  initramfs-tools 
No packages will be installed, upgraded, or removed.
0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B of archives. After unpacking 0 B will be used.
Setting up initramfs-tools (0.130) ...   
update-initramfs: deferring update (trigger activated)
Processing triggers for initramfs-tools (0.130) ...
ln: failed to create hard link '/boot/initrd.img-4.14.28+.dpkg-bak' => '/boot/initrd.img-4.14.28+': Operation not permitted
update-initramfs: Generating /boot/initrd.img-4.14.28+
E: busybox or busybox-static, version 1:1.22.0-17~ or later, is required but not installed
update-initramfs: failed for /boot/initrd.img-4.14.28+ with 1.
dpkg: error processing package initramfs-tools (--configure):
 subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
 initramfs-tools
E: Sub-process /usr/bin/dpkg returned an error code (1)
Setting up initramfs-tools (0.130) ...
update-initramfs: deferring update (trigger activated)
Processing triggers for initramfs-tools (0.130) ...
ln: failed to create hard link '/boot/initrd.img-4.14.28+.dpkg-bak' => '/boot/initrd.img-4.14.28+': Operation not permitted
update-initramfs: Generating /boot/initrd.img-4.14.28+
E: busybox or busybox-static, version 1:1.22.0-17~ or later, is required but not installed
update-initramfs: failed for /boot/initrd.img-4.14.28+ with 1.
dpkg: error processing package initramfs-tools (--configure):
 subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
 initramfs-tools
                                         
root@cerberilla-ambonensis:~#

These are different issues:
The first one is the one I said and can be ignored:

Code: Select all

ln: failed to create hard link '/boot/initrd.img-4.14.28+.dpkg-bak' => '/boot/initrd.img-4.14.28+': Operation not permitted

You can actually see that the process goes further after that and THAT is where the real issue happend:

Code: Select all

update-initramfs: Generating /boot/initrd.img-4.14.28+
E: busybox or busybox-static, version 1:1.22.0-17~ or later, is required but not installed
update-initramfs: failed for /boot/initrd.img-4.14.28+ with 1.
dpkg: error processing package initramfs-tools (--configure):

This happens cause it's required from cryptsetup, but has nothing to do with a regular update-initramfs.
Still as busybox is available in version 1:1.22.0-19+b3 for Debian, this is probably easily solved if you install the package.

chromodoris wrote: Not being able to have cryptsetup (which requires initramfs-tools) is a major problem for me. The board will
be used as a remote backup server in an hosting space and I want the data to be safe.

cryptsetup works just fine on the image but you need to install the "recommends" manually (I disabled this, as they normally are not required)
Cryptsetup fails due to missing recommends:

Code: Select all

root@debian-stretch:~# apt-get install cryptsetup
Reading package lists... Done
Building dependency tree       
Reading state information... Done
The following additional packages will be installed:
  cryptsetup-bin
Suggested packages:
  keyutils
Recommended packages:
  kbd console-setup busybox | busybox-static
The following NEW packages will be installed:
  cryptsetup cryptsetup-bin
0 upgraded, 2 newly installed, 0 to remove and 0 not upgraded.
Need to get 394 kB of archives.
After this operation, 1,377 kB of additional disk space will be used.
Do you want to continue? [Y/n] 
Get:1 http://ftp.debian.org/debian stretch/main armhf cryptsetup-bin armhf 2:1.7.3-4 [221 kB]
Get:2 http://ftp.debian.org/debian stretch/main armhf cryptsetup armhf 2:1.7.3-4 [173 kB]
Fetched 394 kB in 0s (801 kB/s)     
Preconfiguring packages ...
Selecting previously unselected package cryptsetup-bin.
(Reading database ... 201007 files and directories currently installed.)
Preparing to unpack .../cryptsetup-bin_2%3a1.7.3-4_armhf.deb ...
Unpacking cryptsetup-bin (2:1.7.3-4) ...
Selecting previously unselected package cryptsetup.
Preparing to unpack .../cryptsetup_2%3a1.7.3-4_armhf.deb ...
Unpacking cryptsetup (2:1.7.3-4) ...
Processing triggers for systemd (232-25+deb9u3) ...
Setting up cryptsetup-bin (2:1.7.3-4) ...
Processing triggers for man-db (2.7.6.1-2) ...
Setting up cryptsetup (2:1.7.3-4) ...
update-initramfs: deferring update (trigger activated)
Processing triggers for systemd (232-25+deb9u3) ...
Processing triggers for initramfs-tools (0.130) ...
ln: failed to create hard link '/boot/initrd.img-4.14.32+.dpkg-bak' => '/boot/initrd.img-4.14.32+': Operation not permitted
update-initramfs: Generating /boot/initrd.img-4.14.32+
E: busybox or busybox-static, version 1:1.22.0-17~ or later, is required but not installed
update-initramfs: failed for /boot/initrd.img-4.14.32+ with 1.
dpkg: error processing package initramfs-tools (--configure):
 subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
 initramfs-tools
E: Sub-process /usr/bin/dpkg returned an error code (1)

Cryptsetup installs fine with recommends:

Code: Select all

root@debian-stretch:~# apt-get install cryptsetup busybox kbd console-setup
Reading package lists... Done
Building dependency tree       
Reading state information... Done
0 upgraded, 0 newly installed, 4 reinstalled, 0 to remove and 0 not upgraded.
Need to get 0 B/1,001 kB of archives.
After this operation, 0 B of additional disk space will be used.
Preconfiguring packages ...
(Reading database ... 201643 files and directories currently installed.)
Preparing to unpack .../console-setup_1.164_all.deb ...
Unpacking console-setup (1.164) over (1.164) ...
Preparing to unpack .../busybox_1%3a1.22.0-19+b3_armhf.deb ...
Unpacking busybox (1:1.22.0-19+b3) over (1:1.22.0-19+b3) ...
Preparing to unpack .../cryptsetup_2%3a1.7.3-4_armhf.deb ...
Unpacking cryptsetup (2:1.7.3-4) over (2:1.7.3-4) ...
Preparing to unpack .../kbd_2.0.3-2+b1_armhf.deb ...
Unpacking kbd (2.0.3-2+b1) over (2.0.3-2+b1) ...
Setting up console-setup (1.164) ...
Setting up busybox (1:1.22.0-19+b3) ...
Setting up cryptsetup (2:1.7.3-4) ...
update-initramfs: deferring update (trigger activated)
Setting up kbd (2.0.3-2+b1) ...
Processing triggers for systemd (232-25+deb9u3) ...
Processing triggers for man-db (2.7.6.1-2) ...
Processing triggers for initramfs-tools (0.130) ...
ln: failed to create hard link '/boot/initrd.img-4.14.32+.dpkg-bak' => '/boot/initrd.img-4.14.32+': Operation not permitted
update-initramfs: Generating /boot/initrd.img-4.14.32+
root@debian-stretch:~#

as you can see the complains about the linking is still there, but the process completes just fine

chromodoris · Post by **chromodoris** » Fri Apr 06, 2018 5:24 pm

meveric wrote: This happens cause it's required from cryptsetup, but has nothing to do with a regular update-initramfs.
Still as busybox is available in version 1:1.22.0-19+b3 for Debian, this is probably easily solved if you install the package.

Great! Thanks a lot! I tried it on the board for which I did not yet change the filesystem type, and it worked flawlessly.

I may still investigate on having an ext4 /boot, as I find strange that a partition that is intended to be used
as a system partition for a Debian OS is typed for the contingency use case of someone getting the SDcard out of
the board, loading it a Windows host, and despite wanting to install this Debian system refusing to install the
required tools on the Windows side. The nominal use of this partition is on the Debian side, not on the Windows side.
I could understand this use case for the other images that can run on Odroid boards, but for a Debian image, it seems
weird to me.

I was able to completely boot my second board with an ext4 /boot, so I guess I only have to come up with an appropriate
boot_ext4 command and some conditionals to have it selected when the boot partition is ext4, and save this along other
boot commands in the uboot environment. I have yet to find the correct /etc/fw_env.conf file to use so I can use the
fw_printenv/fw_setenv commands to simplify investigation.

Anyway, this ext4 stuff is probably only me being blinkered, so I will not hijack your thread with this, except if you
consider this topic is worth discussing.

Thanks again

meveric · Post by **meveric** » Fri Apr 06, 2018 6:28 pm

chromodoris wrote:I find strange that a partition that is intended to be used
as a system partition for a Debian OS is typed for the contingency use case of someone getting the SDcard out of
the board, loading it a Windows host, and despite wanting to install this Debian system refusing to install the
required tools on the Windows side. The nominal use of this partition is on the Debian side, not on the Windows side.
I could understand this use case for the other images that can run on Odroid boards, but for a Debian image, it seems
weird to me.

This is still from a time around 2012 where ARM boards where rather new and many people did not know how to use Linux.
Imagine the following scenario:
You only have a old crappy monitor, that does not support EDID and only runs up to 1366x768.
The image boots up and tries to set the monitor to 1080p, this is not supported, and the monitor stays black.
The system tries to find supported resolution of the monitor via EDID but it's not supported and the ODROID doesn't know what resolution to take.
The result is, that the system boots up but your screen is black.
Now you're someone that comes from a Windows background, you don't have knowledge about SSH or how to mount an ext file system on a different PC.
What can you do?
Well in most cases this would be the end of the line for many Windows users, if someone tells them they need to adjust the boot.ini to select the right resolution, they have no means to do so, as they don't have access to the boot.ini.

With the vfat partition though, you can even tell a Windows user -> "Put the SD card in your PC, and edit the boot.ini"

And that is the reason why it's vfat.

chromodoris · Post by **chromodoris** » Fri Apr 06, 2018 6:38 pm

I understand. Maybe having only boot.ini is a tiny partition makes sense, but uImage, zImage and *.dtd, would users without
knowledge really mess directly with these specific files?

meveric · Post by **meveric** » Fri Apr 06, 2018 7:12 pm

it's all loaded from the same partition uboot requires boot.ini zImage, dtb and uImage so you can't separate them.
Most user wouldn't know what to do with the other files aside from boot.ini but still you could return to a older Kernel with that as well.. exchanging zImage and uImage with older backups on bootfs is also possible.

rooted · Post by **rooted** » Sat Apr 07, 2018 5:43 pm

Finally got around to swapping Armbian out on my CloudShell 1, so far it's working great.

You left the 3.y modules in /lib/modules and the /boot stuff has been mentioned already. I am so glad to be back to standard Debian, already have all my stuff compiled and had no issues with old libs.

Nice job, thank you.

meveric · Post by **meveric** » Sat Apr 07, 2018 5:47 pm

Thanks for the feedback.. I'll see if I can clean it up a little bit more next time.. after all I plan to use it for the ODROID N1 as well

truckercap · Post by **truckercap** » Sun Apr 08, 2018 7:46 am

Sorry if this may be the wrong spot, I tried searching for similar issues before posting.

After flashing to the eMMC module. I'm having some issues with gaining SSH access.
I am able to ping the address, and can see the orange NIC light flash for each ping. I don't seem to get any response at all for SSH. I have tried this with your Jesse release and got the same results.

Comparing to an install of Debian Stretch on an microSD card on the C2, I also noticed that a hostname is not displaying on my routers DHCP view, the working install on the C2 showed 'Debian-Stretch' or something to that effect.

Is there another step that I am missing for using an eMMC module instead of a microSD? I can not verify a microSD install on my XU4 due to me not have a UHC card. I had Arch running on that eMMC a couple hours ago, and I haven't ran across any errors while flashing.

meveric · Post by **meveric** » Sun Apr 08, 2018 7:57 am

UHC is not required. Any microSD card should do, even a class 4 card should work perfectly fine.

Generally I wouldn't know why SSH shouldn't work. Especially not with the Debian Jessie image, this was done many times and should work fine.
Can you directly use keyboard and monitor on the XU4 to see if it shows any issues?

truckercap · Post by **truckercap** » Sun Apr 08, 2018 10:24 am

I followed armbians install note

```
eMMC install might be broken if you don’t have recent uboot on your eMMC card – you must update it. Boot from SD card, run nand-sata-install and choose "Update bootloader on a special eMMC partition". After that, you can use eMMC as boot device. You only need to do this once.
```
And then I re-flashed your Debian-Stretch. Everything is working great. Thanks for your response.

chromodoris · Post by **chromodoris** » Sun Apr 08, 2018 8:35 pm

I would like to tinker with the uboot environment variables.
In order to do so, I intended to use fw_printenv and fw_setenv, from the u-boot-tools package, which seems
to be installed by default in the image. It failed and I first thought it was only due to a missing /etc/fw_env.config
file, so I tried to create one using information from https://wiki.odroid.com/odroid-xu4/soft ... tion_table.
It also failed.

It seems to me that in fact the mtd (Memory Technology Devices) system is not activated in the kernel, and
is not even available as a module. Both /boot/config-4.14.28+ and /boot/config-4.14.32+ (which was installed
today like a charm) contain a line:

Code: Select all

# CONFIG_MTD is not set

Would it be possible to have a kernel compiled with support for MTD? I think it could be an interesting
addition to this image. Last time I compiled a Linux kernel was probably 15 years ago and I think many
things have changed, so I am a little affraid to do it myself.

meveric · Post by **meveric** » Sun Apr 08, 2018 10:00 pm

maybe would be best to ask HardKernel to add this to their Kernel, as I use their build as a basis.
I can add it myself, but I'm not sure what it does I would have to read into it what MTD is for and if it affects the system, or better in what way.
I don't want to globally activate something for everyone, that I don't understand

rooted · Post by **rooted** » Sun Apr 08, 2018 10:45 pm

AFAIK MTD is used in routers, it's a form of flash memory.

I don't think it would have any affect on those that don't use it.

meveric · Post by **meveric** » Sun Apr 08, 2018 10:56 pm

rooted wrote:AFAIK MTD is used in routers, it's a form of flash memory.

I don't think it would have any affect on those that don't use it.

Thanks for the heads-up

chromodoris · Post by **chromodoris** » Mon Apr 09, 2018 12:53 am

rooted wrote:AFAIK MTD is used in routers, it's a form of flash memory.

I don't think it would have any affect on those that don't use it.

MTD is how the kernel will have access to the persistent flash memory. It is used in many single board computers,
routers or others like all the odroid boards I think. U-boots lives there, as well as its environment variables.
If linux had access to it, we could more simply read and write to these environment variables without having to
rely only on the limited serial console and u-boot command lines.

I also think it would not affect users not wanting to open these devices. Of course, writing garbage there is
a sure way to brick a board...

meveric wrote:maybe would be best to ask HardKernel to add this to their Kernel, as I use their build as a basis.

I didn't find a support page or a contact mail to ask them. Does anybody here know how to post a request for them?

meveric · Post by **meveric** » Mon Apr 09, 2018 4:18 am

chromodoris wrote:I also think it would not affect users not wanting to open these devices. Of course, writing garbage there is
a sure way to brick a board...

Probably something that should be avoided

chromodoris wrote:I didn't find a support page or a contact mail to ask them. Does anybody here know how to post a request for them?

There are different ways to contact HardKernel. Most likely they gonna read this on monday anyway, aside from that, there are several thread for discussions about the Kernel:
viewtopic.php?f=95&t=28895
viewforum.php?f=146

rooted · Post by **rooted** » Mon Apr 09, 2018 7:49 am

chromodoris wrote:
rooted wrote:AFAIK MTD is used in routers, it's a form of flash memory.

I don't think it would have any affect on those that don't use it.
I also think it would not affect users not wanting to open these devices. Of course, writing garbage there is
a sure way to brick a board...

As long as it's a module and not built in shouldn't be a problem, should be blacklisted as well.

I have quite a bit of experience with MTD on routers, I know that SPI flash uses it as well on other embedded devices but I didn't know it was the format used for Uboot's hidden partition.

What are you trying to do? It may be of interest to others as well.

*edit*
Uli has the answer:

viewtopic.php?p=220854#p220854

gom · Post by **gom** » Tue Apr 10, 2018 4:23 pm

Where can I get the gstreamer0.10* packages for stretch?

meveric · Post by **meveric** » Tue Apr 10, 2018 4:47 pm

Debian (Stretch and above) and Ubuntu (16.10 and above) removed GStreamer 0.10 packages, as it's no longer maintained:
https://bugs.debian.org/cgi-bin/bugrepo ... bug=785856

GStreamer 0.10 is no longer maintained and supported by the upstream
project since almost 3 years, and contains many known bugs that are
fixed in the new 1.x release series of GStreamer. Next to many bug
fixes, the new release series also contains many other improvements,
new features and a more streamlined API.

This quote is from 2015, so by now it's about 6 years that GStreamer 0.10 is no longer maintained.

gom · Post by **gom** » Tue Apr 10, 2018 6:05 pm

At my age I understand obsolete but these packages are available for raspbian. I was hoping they might be available for odroid. Can raspbian packages be used on odroid? Or maybe rebuild source package?

meveric · Post by **meveric** » Tue Apr 10, 2018 8:31 pm

Both should work, but make sure you take raspbian packages for Debian Stretch, as there's also Raspbian based on Debian Jessie which still has GStreamer.

nogareth · Post by **nogareth** » Sat Apr 14, 2018 5:19 pm

@Meveric

Are there any news about Kodi running on your image?
My Xu4 Is running now since about 3 Years 24/7 with your Gamestation Turbo image acting as server and always available Media Center for my kids. - Many thanks again for your Work!!
Lately i get strange random crashes of the complete USB system (no response anymore no matter if re-plugged including Core lockups) which makes the Desktop unusuable until rebooted completely.
I dont know if this is a Pre-death symptome of my Xu4 or if it is related on my system setup, but since jessie is slowly getting outdatet i want to setup a new System based on your Stretch image.

Regards , Martin