December 06, 2019, 03:58:31 pm

News:

Have you visited the Allwinner Chipset wiki? - http://linux-sunxi.org/


Buster and Bionic - updated 21.11.2019

Started by slovenia, December 24, 2013, 03:23:11 pm

Previous topic - Next topic

slovenia

Quote from: Sloop on June 10, 2014, 03:07:40 am
Could you tell me how I can do this best? I downloaded your kernel package on the initial post (Link Kernel Package). When I unpacked the file and changed into directory ./lib/modules I see it is still 3.4.79 ?? And where are the Kernel header files located at, I need to know the exact path. And how do I install the new Kernel into my CT ?

Quote from: slovenia on June 10, 2014, 01:28:20 am
The simplest option is to download my latest kernel package which already have proper headers with it. But it's not .79 ... It's .91



Latest (.91) is on my blog - i forgot to update initial post (doing it right now):
http://cubie.nobrainz.de/downloads/CTDebian_2.0_kernel_mod_head_fw.tgz

After unpack you will find headers in /usr/include
Debian and Ubuntu images with kernel 3.4.110, 4.3.3, 4.4
http://www.armbian.com

Sloop

1.) How can I have the complete kernel, not only the header files? I am asking because I am trying to run the command "module-assistant auto-install mymodulename -k /path/to/kernel/including/headerfiles". Now I have the kernel headers, but the command fails because it says it cannot find a valid Kernel source on this path.

2.)How can I install this newer .91 Kernel into my Cubietruck? Actually I am running 3.4.79+

slovenia

1. If you need the whole source:

git clone https://github.com/igorpecovnik/linux-sunxi

2. If you use my image just do this:
http://www.cubieforums.com/index.php/topic,1275.msg17365.html#msg17365

on others it might me necessary to correct path to kernel in uEnv file, located in /boot or on the root of boot partition.
Debian and Ubuntu images with kernel 3.4.110, 4.3.3, 4.4
http://www.armbian.com

Sloop

why do I get this error, when I try to clone the complete kernel (incl. the header) ???

# git clone https://github.com/igorpecovnik/linux-sunxi
Quote
cloning to 'linux-sunxi'...
remote: Counting objects: 2746855, done.
remote: Compressing objects: 100% (468029/468029), done.
remote: Total 2746855 (delta 2251983), reused 2746854 (delta 2251982)
received objects: 100% (2746855/2746855), 813.98 MiB | 596.00 KiB/s, Done.
error: index-pack died of signal 111/2251983)
fatal: index-pack failed

slovenia

I don't compile kernel on my Cubie so I never encounter this kind of error. After a quick search I found this:

http://www.raspberrypi.org/forums/viewtopic.php?p=128303

Memory related problems ... suggesting enabling swap.
Debian and Ubuntu images with kernel 3.4.110, 4.3.3, 4.4
http://www.armbian.com

mno

Quote from: slovenia on June 10, 2014, 01:19:18 am
Quote from: mno on June 09, 2014, 06:18:55 pm
this Debian Image often connects to these domains:

belgarath.bunkus.org/bunkus.org/www.bunkus.org
ridcully.episode-iv.de
gollum.webersheim.de
solsocog.de

can anyone tell me why?


??? This Debian is quite clean. There are some non-standard enhancements but can't be related to this. I could not find any of mentioned in my logs. What did you install on the top of it?


A lot of software, mainly nginx, mysql, seafile, munin.
You can find a full list here: http://mno.pw/installed-packages.txt

But I don't think the list help. Do you have any tip on how I can log wich package connects to which domain?

soundcheck

Hi Slovenia.

Two comments

1.

In /etc/udev/rules.d/70-persistent-net.rules I find a configured wlan0 interface.

I think the file should not exist on maiden installations.


2.

I found a strange message in my dmesg:
[    1.503986] AHCI is disabled in script.bin


We use ct-hdmi/ct-vga.bin !?!? and not script.bin, which  is the typical default bin file though.

Perhaps not an issue.



Cheers

Sloop

Quote from: slovenia on June 10, 2014, 03:37:11 am
1. If you need the whole source:

git clone https://github.com/igorpecovnik/linux-sunxi

2. If you use my image just do this:
http://www.cubieforums.com/index.php/topic,1275.msg17365.html#msg17365



Hi Slovenia,

I have upgraded my CT image to 3.4.91 as you explained in point 2. and rebooted. The kernel 3.4.91 seems to be active now. Thanks so far

But I still have problems in compiling one of my desired modules. My issue is: I installed DRBD 8.4.4 from Debian Jessie repository, but when I enter "modinfo drbd8" I get the result, that the DRBD8 block module in use is in version 8.3.x (!). See that screenshot: , where the left side shows Debian Jessie on an amd64 arch and the right side shows my Cubietruck running Debian Jessie on your image (on this screenshot I had the older kernel installed, but I am on 3.4.91 as I said before).

I cannot understand, why DRBD 8.4.4 package installed and used the older DRBD8 block device module?? So I tried to re-build the newer and suited DRBD 8.4.4 module for the Cubietruck accordingly that official DRBD8 HowTo

Unfortunately, I am stuck at the stage where I enter "module-assistant auto-install drbd8". It tries to install the linux-headers through the Debian repository which of course will fail because the linux-headers don't exist on the official Debian repository. So I downloaded the kernel header files from your initial post and afterwards I unpacked this archive into /some/dir. When I try to re-run my last command "module-assistant ..." but add the option to use a dir for the kernel dir (the command is "module-assistant auto-install drbd8 -k /some/dir") then I get the error, that no kernel source can be found. It seems to me, that this process is also expecting a valid kernel source on the specified dir.

So, how can I have the kernel source files, too? I tried to download the related and complete 3.4.91 Kernel but when I unzip this file, and use the extracted dir as kernel-base dir in my "module-assistant ..." command it says that the kernel is not configured and the command fails again.

How should I proceed here, any help really appreciated. I just need the correct DRBD8 module in version 8.4.4.


macduck

Quote from: mno on June 10, 2014, 08:53:56 am
Do you have any tip on how I can log wich package connects to which domain?


sudo tcpdump -Xi <YOUR-NETWORK-INTERFACE> tcp[13] == 2 and src host <YOUR-INTERFACE-IP>

Actually I world suggest you to use -n option so tcpdump will not do  reverse lookup before printing dump, but will just print IPs. Do reverse lookup yourself later.

what you'll see is all the outgoing syn packets created by your system (all the connections initiated by your system). Like in this example:

Quotectruck  andrew  ~  sudo tcpdump -Xi eth0 tcp[13] == 2 and src host 192.168.0.5
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), capture size 65535 bytes
09:34:39.142509 IP ctruck.local.33209 > seagatestorage.microsoft-ds: Flags [ S ], seq 1563783514, win 14600, options [mss 1460,sackOK,TS val 305919693 ecr 0,nop,wscale 7], length 0
   0x0000:  4500 003c 8797 4000 4006 31c9 c0a8 0005  E..<..@.@.1.....
   0x0010:  c0a8 0006 81b9 01bd 5d35 715a 0000 0000  ........]5qZ....
   0x0020:  a002 3908 818a 0000 0204 05b4 0402 080a  ..9.............
   0x0030:  123b f6cd 0000 0000 0103 0307            .;..........
09:35:39.271915 IP ctruck.local.33210 > seagatestorage.microsoft-ds: Flags [ S ], seq 2134548069, win 14600, options [mss 1460,sackOK,TS val 305925706 ecr 0,nop,wscale 7], length 0
   0x0000:  4500 003c 902f 4000 4006 2931 c0a8 0005  E..<./@.@.)1....
   0x0010:  c0a8 0006 81ba 01bd 7f3a 9e65 0000 0000  .........:.e....
   0x0020:  a002 3908 818a 0000 0204 05b4 0402 080a  ..9.............
   0x0030:  123c 0e4a 0000 0000 0103 0307            .<.J........


here my Ctruck connects to seagatestorage (microsoft-ds service is NetBIOS over TCP/IP and SMB/CIFS traffic)

tola5

Hi have use you Nice Worning for some time but suddenly it dont World so I take the opportunity for the update to the latest :D but nok my 30gb SD are not ben resize egen it boot so it are only 1gb some who have tried it ? Are it just to take it to a vm machine
and do it

soundcheck

I'd like to share a way of regenerating udev-net-rules.


sudo su
rm -f /etc/udev/rules.d/70-persistent-net.rules
export INTERFACE=eth0
export MATCHADDR=`ip addr show $INTERFACE | grep ether | awk '{print $2}'`
/lib/udev/write_net_rules


(Taken from here )

stuff like e.g. "udevadm trigger --action=add or similar" or just "removing net-rules and rebooting" usually doesn't work.

@ slovenia: Something like above should be run after initial bootup (/etc/init.d/cubian-firstrun) of the
                   image on a new CT.
                   You just need to figure out if eth0 or wlan0 is up.

Cheers


wiesi

I´ve tried this, everything is in place ... bute uname -r still shows me kernel 3.4.79-sun7i+ ... any help? :)

Quote from: slovenia on June 09, 2014, 01:51:37 pm
1. Login as root
2. Execute:

cd /
wget http://cubie.nobrainz.de/downloads/CTDebian_2.0_kernel_mod_head_fw.tgz
tar pxvfz CTDebian_2.0_kernel_mod_head_fw.tgz
reboot


@MrDialUp

Default way yes. I am not bothering to pack a kernel as it should be for normal apt-get update. I don't have energy to bother with that :P

marcelpo

Quote from: slovenia on June 09, 2014, 01:51:37 pm
1. Login as root
2. Execute:

cd /
wget http://cubie.nobrainz.de/downloads/CTDebian_2.0_kernel_mod_head_fw.tgz
tar pxvfz CTDebian_2.0_kernel_mod_head_fw.tgz
reboot


@MrDialUp

Default way yes. I am not bothering to pack a kernel as it should be for normal apt-get update. I don't have energy to bother with that :P



Maybe a good idea to note that to upgrade a NAND or SATA based installation you first need to make sure that /dev/nand1 is mounted to /boot otherwise it looks like everything went ok, but the kernel actually landed on the local drive instead of the NAND boot partition....  I learned this the hard way by doing the kernel upgrade to a remote cubietruck that as a result of this error didn't come back after reboot ;)

wiesi

I am a bit confused because of /dev/nand1 ... where is it mounted to? This is what it looks like at my cubietruck:

Quote
Filesystem     1K-blocks      Used Available Use% Mounted on
rootfs           7369792   2221076   4774344  32% /
/dev/root        7369792   2221076   4774344  32% /
devtmpfs         1023896         0   1023896   0% /dev
tmpfs             131072       196    130876   1% /run
tmpfs               5120         0      5120   0% /run/lock
tmpfs             131072         0    131072   0% /run/shm
/dev/sda1      961433632 400479780 512115776  44% /mnt/sata
tmpfs            1048576       152   1048424   1% /tmp
/dev/root        7369792   2221076   4774344  32% /var/log.hdd
ramlog-tmpfs      262144      4364    257780   2% /var/log


I have a SATA-disc (mounted to /mnt/sata) and no SD-CARD inside ...


Quote from: marcelpo on June 11, 2014, 04:32:43 am
Quote from: slovenia on June 09, 2014, 01:51:37 pm
1. Login as root
2. Execute:

cd /
wget http://cubie.nobrainz.de/downloads/CTDebian_2.0_kernel_mod_head_fw.tgz
tar pxvfz CTDebian_2.0_kernel_mod_head_fw.tgz
reboot


@MrDialUp

Default way yes. I am not bothering to pack a kernel as it should be for normal apt-get update. I don't have energy to bother with that :P



Maybe a good idea to note that to upgrade a NAND or SATA based installation you first need to make sure that /dev/nand1 is mounted to /boot otherwise it looks like everything went ok, but the kernel actually landed on the local drive instead of the NAND boot partition....  I learned this the hard way by doing the kernel upgrade to a remote cubietruck that as a result of this error didn't come back after reboot ;)

schmidtm

Hi Igor,

i have a hard time getting the S/P DIF Interface on my Cubietruck up and running with your v2.0 / 2.6.2014 Image.

1) The cards show up in the proc filesystem:

schmidtm@cubie:~/prj/alsa-test$ cat /proc/asound/cards
0 [sunxicodec     ]: sunxi-CODEC - sunxi-CODEC
                      sunxi-CODEC  Audio Codec
1 [sunxisndspdif  ]: sunxi-sndspdif - sunxi-sndspdif
                      sunxi-sndspdif
2 [sunxisndi2s    ]: sunxi-sndi2s - sunxi-sndi2s
                      sunxi-sndi2s

2) ... but are not recognized by ALSA:

schmidtm@cubie:~/prj/alsa-test$ aplay -l
aplay: device_list:252: no soundcards found...

schmidtm@cubie:~/prj/alsa-test$ aplay -L
null
    Discard all samples (playback) or generate zero samples (capture)


although I have installed alsa-base and alsa-utils. Aplay on the hardware used to work *once* like this:

aplay -c 2 -f cd -t wav --device=hw:1 song.wav

But after using a ~/.asoundrc to set the default to hw:1 everything used to stop working.

Can anyone share some ideas how to track this down?

cheers,

Matthias