Jump to content

Prepare v5.15

zador.blood.stained
 Share

Recommended Posts

zador.blood.stained

zador.blood.stained

Posted
Posted (edited)

Short changelog:

  • Removed autodetection for H3 boards, changed DVFS and thermal settings
  • Repackaged tools, hostapd and firmware (may require manual installation of packages)
  • Odroid C2: Fixed headers packaging script, updated kernel configuration, updated boot script
  • Added back watchdog to sun7i-default kernel
  • Purging and reinstalling "unattended-upgrades" package should be recommended in changelog
  • Script "40-updates" update logic was changed; script was switched to python3 since Xenial doesn't use python2 by default. Installing "python3-apt" should be recommended in changelog.
  • Reworked initrd images handling on kernel upgrade
  • ... (to be added)
  • Other fixes and enhancements

 

Release 5.15 closes following issues: #285, #345, #362 (closed), #337 (closed)

 

I would propose some changes in this release:

 

1. Change "auto eth0" to "allow-hotplug eth0" and "auto wlan0" to "allow-hotplug wlan0" (uncommented) for all distributions except Wheezy, pack new interfaces files to board support packages;

 

2. Reduce number of preinstalled packages. In my opinion, default package list is rather big and some software can be removed from prebuilt images, especially if:

  • We don't provide any custom configuration for the package
  • It can be used only with addidional hardware or is rather useless in default CLI images
  • Is not required for any preinstalled scripts and required packages

So I propose removing these packages:

libmtp-runtime
mtp-tools
bluez (but add "bluetooth" metapackage to desktop list?)
ir-keytable
weather-util
weather-util-data
dvb-apps
libbluetooth-dev
libbluetooth3
subversion
iptables
thin-provisioning-tools

Any additions, thoughts or suggestions?

 

Edit: added thin-provisioning-tools

Edited by zador.blood.stained
Amended changelog
2
tkaiser

tkaiser

Posted
Posted

Fine with all of the above. I would like to include 'sysstat' package instead (just to poke users doing an 'iostat 5' before they start complaining about high average load and such things) but am also fine if you say no since asking for installation is pretty reasonable, 99 percent of users will never use it and I'm still stuck at finishing the armbianmonitor-daemon approach so it's ok to postpone inclusion :)

1
Igor

Igor

Posted
Posted

Regarding images, none. I am still working on extending debug mode in build script which is unrelated to release. Feel free to upgrade version number and after that i can rebuild everything.

 

Wrote on mobile phone

1
Tido

Tido

Posted
Posted

I just thought a bit about Bluetooth & InfraRed support, some SBC come with this already onboard.

On the other hand, many do not use it and if they want they can still install it.

 

If I would create a Distro I would like to know how many of them are 'alive' and which Version.

So I would like Ubuntu, ask the user for collecting some data (Opt in) if I were you.

zador.blood.stained

zador.blood.stained

Posted
  • Author
Posted

I just thought a bit about Bluetooth & InfraRed support, some SBC come with this already onboard.

On the other hand, many do not use it and if they want they can still install it.

IR (specifically CIR) is supported on a large number of devices and we provide custom LIRC configuration for some boards, it's not ideal, but it is definitely useful to have by default.

Onboard Bluetooth is present only on few devices, audio support basically doesn't work (it relies on SoC I2S interface and proper configuration which is not tested and may not exist at all) and configuring BT without GUI apps is not that easy, especially with BlueZ 5.x - so BT will be enabled by default only on desktop images.

 

If I would create a Distro I would like to know how many of them are 'alive' and which Version.

So I would like Ubuntu, ask the user for collecting some data (Opt in) if I were you.

Ubuntu and Armbian are slightly different scale projects, I don't think we need this yet, especially since compared to Ubuntu main focus of Armbian is CLI/Server images.

1
Igor

Igor

Posted
Posted (edited)

- added driver for Clearfog switch @mainline kernel

- fixed kernel patches @C1, C2 & Clearfog

- added DEBUG_MODE for simple patch creation

- all images will be / were updated to 5.14 and manually tested for booting + first user creation

- added XU4 vanilla (WIP)

 

This is critical ... the rest could wait.

https://github.com/igorpecovnik/lib/issues/341#issuecomment-227456821

Edited by Igor
xu4
Igor

Igor

Posted
Posted

- all images were updated to 5.14 to have a major reference working point, and that I can clean archive a bit. Most of them are briefly tested. Missing testing: Pcduino's.

- most boards got Xenial CLI image

- fixed on board wifi on Banana M1+ Vanilla kernel

 

Issues to be solved for 5.15 or later:

 

- free memory not working in Xenial motd

- 4k monitor not displaying picuture (A10/A20) (uboot)

- adding extra video modes A10/20/H3 (uboot & kernel)

tkaiser

tkaiser

Posted
Posted

- free memory not working in Xenial motd

 

Not only this (fix/workaround for Xenial or newer procps version posted a few days ago), you also have to rework board name display (I'm currently testing all 4 image variants on a pcDuino3 Nano and get back with results later):

 

 

 

     _      _     _             

  __| | ___| |__ (_) __ _ _ __  

 / _` |/ _ \ '_ \| |/ _` | '_ \ 

| (_| |  __/ |_) | | (_| | | | |

 \__,_|\___|_.__/|_|\__,_|_| |_|

                                

 

Welcome to ARMBIAN Debian GNU/Linux 8 (jessie) 4.6.2-sunxi 

 

System load:   1.36            Up time:       9 min

Memory usage:  16 % of 1000Mb IP:            192.168.83.90 

CPU temp:      30°C           HDD temp:      27°C           

Usage of /:    3% of 59G    

 

tk@pcduino3nano:~$ cat /etc/armbian-release 

# PLEASE DO NOT EDIT THIS FILE

BOARD=pcduino3nano

ID="pcDuino3 Nano"

VERSION=5.14

LINUXFAMILY=sunxi

BRANCH=next

tk@pcduino3nano:~$ cat /var/run/machine.id 

pcDuino3 Nano

 
IMO motd should be fixed and the other issues can be postponed.
zador.blood.stained

zador.blood.stained

Posted
  • Author
Posted

IMO motd should be fixed and the other issues can be postponed.

➜  ~  % cat /etc/os-release
PRETTY_NAME="Debian GNU/Linux stretch/sid"
NAME="Debian GNU/Linux"
ID=debian
HOME_URL="https://www.debian.org/"
SUPPORT_URL="https://www.debian.org/support"
BUG_REPORT_URL="https://bugs.debian.org/"

Looks like we just need to rename ID variable in /etc/armbian-release because it conflicts with /etc/os-release

tkaiser

tkaiser

Posted
Posted

Looks like we just need to rename ID variable in /etc/armbian-release because it conflicts with /etc/os-release

 

Well spotted! :) I would propose changing it to $BOARDNAME in /etc/armbian-release instead (can't do changes because build host is busy doing other stuff).

 

 

macbookpro-tk:lib tk$ grep -r '$ID' *
Binary file bin/NumixHolo.zip matches
scripts/armbianmonitor/armbianmonitor-daemon:web.status.1.name=$ID
scripts/armbianmonitor/armbianmonitor-daemon:web.statistics.1.name=$ID
scripts/armbianmonitor/armbianmonitor-daemon:	sed "s/Raspberry Pi/$ID/" < /etc/rpimonitor/template/uptime.conf >/etc/armbianmonitor/templates/uptime.conf
scripts/armhwinfo:	# Starting with 5.15 read in $BOARD (odroidc2), $ID ("Odroid C2") and $VERSION (5.15)
scripts/armhwinfo:	# IRQ distribution based on $HARDWARE and/or $ID, probably some sort of user feedback (leds)
scripts/armhwinfo:			case $ID in
scripts/armhwinfo:		echo -e "\n\n\n$(date) $HARDWARE $ARCH $KERNELID $MACHINE $ID $VERSION" >>${Log}
scripts/armhwinfo:		echo "$(date) $HARDWARE $ARCH $KERNELID $MACHINE $ID $VERSION" >>${Log}
scripts/armhwinfo:		echo -e "[\e[0;32m ok \x1B[0m] Starting ARM hardware info: $ID (${VERSION})"
scripts/armhwinfo:		echo $ID >/var/run/machine.id # we should leave this, maybe users depend on it
scripts/armhwinfo:		echo "$(date "+%s") $HARDWARE $ARCH $KERNELID $MACHINE $ID $VERSION" >/boot/.verbose
macbookpro-tk:lib tk$ grep -r '${ID}' *
scripts/armbianmonitor/armbianmonitor-daemon:	case ${ID} in
scripts/armbianmonitor/armbianmonitor-daemon:		export GetSoCTemp="${ARCH}-${ID}"
scripts/armbianmonitor/armbianmonitor-daemon:	if [ "X${ID}" = "X" ]; then
scripts/armhwinfo:			case ${ID} in
scripts/nand-sata-install/usr/lib/nand-sata-install/nand-sata-install.sh:		if [[ "${ID}" == "Cubieboard" ]]; then 
scripts/update-motd.d/99-point-to-faq:	case ${ID} in 

 

 

tkaiser

tkaiser

Posted
Posted

Most of them are briefly tested. Missing testing: Pcduino's.

 

PcDuino3 Nano with 3.5" disk connected via USB and 2.5" SATA disk directly connected:

In other words: Everything fine except of the strange network issue when booting from NAND (BTW: I didn't know that this little beast has onboard NAND!). At least I don't care that much about non working network in this mode (most probably related to u-boot and something not bringing up correctly). Since I'm already at it I measure NAND speed now...

Igor

Igor

Posted
Posted

If you connect with WICD than there is one more thing to fix. Use search.

 

Wrote on mobile phone

wildcat_paris

wildcat_paris

Posted
Posted (edited)

based on my experience, NanoPi M1 would need max 960MHz (lesser than 1200MHz)

 

Armbian 5.14 @960MHz idle = 51°C

 

Armbian 5.15 @1200MHz up to 65°C (little load)

now @960MHz = 56°C

 

I have put copper heatsinks on the SoC and DRAM

 

I am using

cpufreq-set -g ondemand -u 960M

 

edit:

the infrared temperature of the copper heatsink is 42°C

 

edit2: full load @960MHz 64°C (copper is 45°C)

 

edit3: full load @1.2GHz 80°C

Edited by wildcat_paris
note
Guest
This topic is now closed to further replies.
 Share
Go to topic listing
×
×
  • Create New...

Important Information

Terms of Use - Privacy Policy - Guidelines

  I accept