572 replies to this topic

[christophecvr]

You don't need the fully bloated sftp for transferring files and such, you can use "scp" (and its windows version) on the built-in dropbear ssh implementation.

Indeed You do not need the full openssh.
The only thing You do need in order to work with nautilus ,gftp or fillezilla as client. Is the sftp-server module which is so small as 28 bytes.
For the rest it's just using the dropbear ssh from openpli. like is done for scp.

For windows yes that's a case apart, windows self does not support sftp standard like does nautilus from ubuntu. Then You can use scp .
But If You install Filezilla in windows it does work ass well.

It's all a mather of choice, what the users prefer. The most of ubuntu users do like the use off nautilus which is a very good file browser.
Also I really like Gftp on linux as there we can even set options to keep and or change file rights , And gftp is really booming, the performance is amazing with it. especially for large video files. On Gb network the performance limitation is not the network anymore but the pc's busses and drives self.

Note with openpli 4.0 , the openpli oe core does not install the sftp-server(module) into /usr/libexec anymore but in /usr/lib/openssh .
As far as I looked up on the net, that module must be in /usr/libexec .
Anyway, As test I once give it a try by installing the full openssh to my box dm8000 running openpli 4.0 . i did not worked out.
Which lead me to think that's it is actually a bug into oe core.

As a work around i just modified the specifig openssh-sftp-server module package by adding a pkg_postinst for that package into the bb file concerned. Which just creates the /usr/libexec directory (if it did not existed already) and then creates a link to the /usr/lib/openssh/sftp-server module into that folder.
Then its' booming again.

[christophecvr]

Sorry the sftp-server module is a bit more then 28 bytes but it's still small 77,3 kB

[gegone]

Help me,I am from china, I compile openpli2.1 firmware to solve the garbage problem on my local cable EPG, I have been added big5.h and gb18030.h source patch file into enigma2 and successfully compiled.But still failed to solve cable EPG display garble.Master ,how to solve this? thank you.

Edited by gegone, 11 March 2013 - 01:40.

[gegone]

The encoding and string files had patched too

Edited by gegone, 11 March 2013 - 01:54.

[Persian Prince]

again

http://openpli.org/download/dm7020/
http://openpli.org/download/dm600pvr/
http://openpli.org/download/dm500plus/

doesn't work

nobody wants to fix this ?

[Frenske]

I will report it to the one that is responsible for this.

Edited by Frenske, 27 March 2013 - 23:11.

[Frenske]

The problem is partly fixed. New builds have been started as from today.
Thanks to PieterG.

[Persian Prince]

i think you can remove "dreambox-kernel-3.2-final" and "dreambox-kernel-3.2" branches of http://openpli.git.s...edded;a=summary

 

it seems that it's impossible to create an image in old oe for dreambox with kernel 3.2

[Persian Prince]

please check http://openpli.org/download/vusolo2/

 

i think something is wrong (there are so many old files on the server)

[maigais]

 

Waiting for 2 active tasks to finish:
0: perl-native-5.10.1-r10 do_compile (pid 18951)
1: m4-native-1.4.14-r0.1 do_configure (pid 1499)
NOTE: package db-5.1.19-r2: task do_fetch: Succeeded
Waiting for 1 active tasks to finish:
0: perl-native-5.10.1-r10 do_compile (pid 18951)
NOTE: package m4-native-1.4.14-r0.1: task do_configure: Succeeded
NOTE: package perl-native-5.10.1-r10: task do_compile: Succeeded
ERROR: '/home/maigonis/openpli/openembedded/recipes/libpng/libpng_1.5.10.bb' failed
make: *** [image] Ошибка 1
maigonis@ubuntu:~/openpli$

[WanWizard]

please check http://openpli.org/download/vusolo2/

 

i think something is wrong (there are so many old files on the server)

 

Solo2 wasn't added to the cleanup script after introducing. Should be cleaned up tonight.

[Persian Prince]

please check http://openpli.org/download/vusolo2/

 

i think something is wrong (there are so many old files on the server)

 

Solo2 wasn't added to the cleanup script after introducing. Should be cleaned up tonight.

 

just a friendly report to inform you

[maigais]

corrected manually libpng_1.5.10.bb and everything went up a problem until the next

 

 

 

NOTE: Running task 3285 of 3540 (ID: 777, /home/maigonis/openpli/openembedded/recipes/gstreamer/gst-plugins-bad_0.10.22.bb, do_build)
NOTE: package gst-plugins-bad-0.10.22-r11.4: task do_build: Started
NOTE: package gst-plugins-bad-0.10.22-r11.4: task do_build: Succeeded
NOTE: Running task 3286 of 3540 (ID: 782, /home/maigonis/openpli/openembedded/recipes/gstreamer/gst-plugins-bad_0.10.22.bb, do_rm_work)
NOTE: package gst-plugins-bad-0.10.22-r11.4: task do_rm_work: Started
NOTE: package gst-plugins-bad-0.10.22-r11.4: task do_rm_work: Succeeded
ERROR: '/home/maigonis/openpli/openembedded/recipes/gnutls/gnutls_2.10.4.bb' failed
ERROR: '/home/maigonis/openpli/openembedded/recipes/linux/linux-dm800_2.6.18.bb' failed
ERROR: '/home/maigonis/openpli/openembedded/recipes/base-passwd/base-passwd_3.5.24.bb' failed
ERROR: '/home/maigonis/openpli/openembedded/recipes/bluez/bluez4_4.84.bb' failed
maigonis@ubuntu:~/openpli/build-dm800$
[6:37:52] maigonis12: пробывал посматреть на  гит
[6:38:41] maigonis12: но там хрен знает как откирваетса
[6:39:45] maigonis12: Oops, looks like something went wrong.

 

git this article

 

Oops, looks like something went wrong.

 

http://sourceforge.n...recipes/gnutls/

 

[betacentauri]

I have compiled the image without problems. Now I want to rebuild the kernel with another configuration. So I have modified ../meta-openpli/recipes-linux/linux/linux-etxx00/et9x00/defconfig.

Afterwards I tried to rebuild the image, but I think that kernel stays old.

I did this:

 

source env.source ; bitbake virtual/kernel -c clean
source env.source ; bitbake -c clean  openpli-enigma2-image
source env.source ; bitbake openpli-enigma2-image

Result:

$ ls -la ./tmp/deploy/images/et9x00/
drwxrwxr-x 2 x x     4096 Apr 27 11:27 .
drwxrwxr-x 3 x x     4096 Apr 27 10:18 ..
-rw-rw-r-- 1 x x  6305428 Apr 20 22:03 modules-3.6.0-r3.11-et9x00.tgz
-rw-r--r-- 1 x x 48036935 Apr 27 11:27 openpli-3.0-et9x00-20130427092435_usb.zip
-rw-r--r-- 1 x x 57147392 Apr 27 11:27 openpli-enigma2-3.0-et9x00.rootfs.ubi
-rw-r--r-- 1 x x 55107584 Apr 27 11:27 openpli-enigma2-3.0-et9x00.rootfs.ubifs
-rw-rw-r-- 1 x x      294 Apr 27 11:25 README_-_DO_NOT_DELETE_FILES_IN_THIS_DIRECTORY.txt
-rw-r--r-- 1 x x  9739091 Apr 20 22:03 vmlinux-3.6.0-r3.11-et9x00-20130420191339.bin
lrwxrwxrwx 1 x x       45 Apr 20 22:03 vmlinux-et9x00.bin -> vmlinux-3.6.0-r3.11-et9x00-20130420191339.bin

For me it looks like old kernel build (from 20.04.) is still used.

 

If I do this:

source env.source ; bitbake virtual/kernel -c clean
source env.source ; bitbake virtual/kernel -c config
source env.source ; bitbake virtual/kernel -c compile

I can see new compiled kernel:

-rwxr-xr-x 1 frank frank 9812603 Apr 27 11:36 ./tmp/work/et9x00-oe-linux/linux-etxx00-3.6.0-r3.11/linux-3.6.0/vmlinux

How do I add the kernel to the image??

[pieterg]

with recent bitbake versions, 'clean' is not enough.
You should use

bitbake -c cleansstate virtual/kernel

that gets rid of everything

[betacentauri]

Thanks! That works.

 

And the last 3 questions for today

 

Can I install kernel via ipk package? If yes, which of the packages do I have to install?

kernel-3.6.0_3.6.0-r3.11_et9x00.ipk
kernel_3.6.0-r3.11_et9x00.ipk
kernel-dev_3.6.0-r3.11_et9x00.ipk
kernel-image-3.6.0_3.6.0-r3.11_et9x00.ipk
kernel-vmlinux_3.6.0-r3.11_et9x00.ipk

And can I do it normally in runlevel 3 or must I do something special, because kernel is use while installation?

Edited by betacentauri, 27 April 2013 - 11:07.

[pieterg]

Can I install kernel via ipk package? If yes, which of the packages do I have to install?

just the kernel-image ipk.

And can I do it normally in runlevel 3 or must I do something special, because kernel is use while installation?

yes, no problem. Nothing will happen until the next reboot.

BTW, I usually do this to quickly test a kernel:

bitbake -f -c compile virtual/kernel
gzip the resulting vmlinux
ftp vmlinux.gz to /tmp
flash_eraseall /dev/mtd1
nandwrite -p /dev/mtd1 /tmp/vmlinux.gz
and reboot

[betacentauri]

BTW, I usually do this to quickly test a kernel:

bitbake -f -c compile virtual/kernel
gzip the resulting vmlinux
ftp vmlinux.gz to /tmp
flash_eraseall /dev/mtd1
nandwrite -p /dev/mtd1 /tmp/vmlinux.gz
and reboot

 

Guys, don't do that without backup!! I thought that I have bricked my box   It didn't boot and after turn off and turn on it didn't read my usb stick (tv and vfd stayed black). Second restart worked and I could flash the box.

From which directory should I take the vmlinux? There were at least 4 different ones with vmlinux.gz, vmlinux-3.6.... and different sizes in tmp and boot directory.

[pieterg]

You cannot brick your box like that, the bootloader is in a different flashchip (inaccessible by the kernel)
So it's impossible to brick your box from a running image, even if you wanted to.

[pieterg]

I usually take the vmlinux from the workdir, and gzip it myself. Avoids having to wait for the lengthy install steps.