834 replies to this topic

[Erik Slagter]

This is a known problem with certain first generation XTrend receivers. We never found the cause. It seems to be a bad interaction between some bug in enigma, a certain compiler version and a bug in the drivers. Because it doesn't happen on other models / brands. Very hard to debug.

[blzr]

@teleradio

 

FWIW:

https://forums.openp...ndpost&p=737522

[wolfx95]

hi why for vu+solo, openpli6 not available?

 

but,Was available last week.

 

please i need to latest openpli6 for vu+solo

[WanWizard]

There are reports that it bricks some of the first generation VU+ boxes (uno, duo, solo, ultimo), so we stopped the builds until we know if this is true, and if so, what causes it.

[Pr2]

Hi,
 
While watching an IPTV channel entering the channel list with the up arrow crash OpenPLi RC6.
 
Here are the relevant info:

OpenPLi Enigma2 crash log
 
crashdate=dim. oct.  1 21:28:39 2017
compiledate=Sep 26 2017
skin=PLi-FullNightHD/skin.xml
sourcedate=2017-09-25
branch=rc-6.0
rev=c2a78fb
component=enigma2, signal
 
stbmodel=solo4k
stbmodel=dm8000
kernelcmdline=root=/dev/mmcblk0p4 rootwait rw rootflags=data=journal debug coherent_pool=2M brcm_cma=504M@0x10000000 brcm_cma=260M@0x2f800000 brcm_cma=1024M@0x80000000
nimsockets=NIM Socket 0:
imageissue=openpli 6.0-rc %h
 
 
[eDVBPESReader] Created. Opening demux
[eDVBPESReader] Created. Opening demux
...

[eDVBPESReader] Created. Opening demux
[eDVBPESReader] ERROR reading PES (fd=49): Value too large for defined data type
[eDVBPESReader] Created. Opening demux
...

[eDVBPESReader] Created. Opening demux
REMOVE Peer et9x00
REMOVE Peer et9x00
ADD Peer et9x00=et9x00.local:8001
ADD Peer et9x00=et9x00.local:8001
[eDVBPESReader] Created. Opening demux
...

[eDVBPESReader] Created. Opening demux
[eDVBPESReader] ERROR reading PES (fd=59): Value too large for defined data type
[eDVBPESReader] Created. Opening demux
...

[eDVBPESReader] Created. Opening demux
REMOVE Peer et9x00
REMOVE Peer et9x00
ADD Peer et9x00=et9x00.local:8001
ADD Peer et9x00=et9x00.local:8001
[eDVBPESReader] Created. Opening demux
...

[eDVBPESReader] Created. Opening demux
[eDVBPESReader] Created. Opening demux
-------FATAL SIGNAL

 
The ... are always:   [eDVBPESReader] Created. Opening demux[/size]

Edited by Pr2, 1 October 2017 - 20:38.

[sikoko]

openpli 6.0-rc formuler4


formuler4 login: root
Last login: Mon Oct  2 15:09:15 CEST 2017 on pts/0
root@formuler4:~# php-cgi
php-cgi: error while loading shared libraries: libmcrypt.so.4: cannot open shared object file: No such file or directory
root@formuler4:~# ls -l /usr/lib | grep libmcry
-rwxrwxrwt    1 root     root        211212 Oct  2 15:08 libmcrypt.so.4
root@formuler4:~# 

 

It works in OpenPli 4.0.

Attached Files

•  php_libmcrypt.tar.gz   1.97MB   13 downloads

Edited by sikoko, 2 October 2017 - 14:21.

[MiLo]

From the limited info I understand that you tried to run a OpenPLi4 binary on OpenPLi6. That usually won't work, it has to be recompiled. Which is never a problem since everything is open source.

[sikoko]

The problem is that the library is there, but the php-сgi does not see it.

In version 4.0 of php-cgi works, and in 6.0 it does not see the file libmcrypt.so.

Compiled "ld" - does not find libdl.so.2, which is also available.

[MiLo]

Did you compile it or did you just put a binary on the box?

[sikoko]

I compiled it (php-cgi). I checked with OpenPli 4.0 (Formuler 4): libcurl.so.4, libmcrypt.so.2 is missing. Added them from the toolchain to /usr/lib and it works.

Installed OpenPli 6.0, added the libraries - "php-cgi: error while loading shared libraries: libmcrypt.so.4: can not open shared object file: No such file or directory"

No dependency reports.

 

PS: crosstool-ns. php-cgi compiled for mipsel.

Edited by sikoko, 2 October 2017 - 18:20.

[athoik]

But OE builds ipk directly for your receiver/arch. Are you compiling php using another build technique?

[sikoko]

Yes, I compile with crosstool-ng for different architectures.

Works successfully on sh4 (CPU sti7111 and sti7162) and mipsel (OpenPli 4.0, but not on 6.0).

Edited by sikoko, 2 October 2017 - 18:50.

[athoik]

OE provides php (recipes is on meta-openembedded/meta-oe/recipes-devtools/php/php_5.6.26.bb).

Better to build php using OE, like we build everything.

If you like to build things using crosstoll-ng, you are alone here.

[sikoko]

root@formuler4:~# ldd /usr/bin/php-cgi
        linux-vdso.so.1 (0x770b6000)
        libmcrypt.so.4 => not found
        libltdl.so.7 => not found
        libcurl.so.4 => not found
        libcrypto.so.1.0.0 => not found
        libssl.so.1.0.0 => not found
        libm.so.6 => not found
        libdl.so.2 => not found
        librt.so.1 => not found
        libc.so.6 => not found
root@formuler4:~# 

libltdl.so.7, libcrypto.so.1.0.0, libm.so.6, libdl.so.2, libc.so.6 files available in /lib and /usr/lib default but "not found"

Edited by sikoko, 2 October 2017 - 19:17.

[sikoko]

OE provides php (recipes is on meta-openembedded/meta-oe/recipes-devtools/php/php_5.6.26.bb).

github.com/OpenPLi/openpli-oe-core.git

meta-openembedded is empty please, help

Edited by sikoko, 2 October 2017 - 22:15.

[WanWizard]

All the other meta directories are empty too, these are external: https://git-scm.com/...ools-Submodules

[MiLo]

github.com/OpenPLi/openpli-oe-core.git
meta-openembedded is empty please, help

git submodule update --recursive --init

[Erik Slagter]

Even better. Use the makefile. Otherwise you will NOT get it working anyway. E.g. you need to add all BBLAYERS by hand.

[greatred]

Hi,

 

in the event of openpli6 images removed for VU+ boxes, should the current users running openpli6 take any caution ?

I mean for example like avoiding to flash backup images created with FullBackup ? 

[WanWizard]

No, a backup made from a working box is not a problem.

 

We suspect (but have not proven it, nor did we receive feedback from VU+) that a boot image in the zip, larger than the boot partition, causes the issue, because the bootloader doesn't seem to check the size before it flashes.

 

We only find it weird that of all the users that have flashed 6.0-rc on their 1st generation VU+, only two users complained sofar. So it is still an option that clones are involved, although they stated that was not the case.