1943 replies to this topic

[Robinson]

So it looks like there is a significant difference between DMM original image and OpenPLi images which prevents doublet's script from doing its job.

Maybe there is still hope and it is not a question of drivers or hardware support then.

If only some OpenPLi developers could look into doublet's script and try to figure out why it is working on the original image only.

[Robinson]

That's right. PAT (has PID 0) is broadcasted unencrypted so it shall be available anytime frontend is locked on some transponder.

 

We are hunting for descrambled PMT. PMT has different PID for each program (channel) - see PAT table.

 

While "viewing" BBC channel and script is not running, you'll get timeout when you try to get PMT with dvbsnoop:

dvbsnoop -timeout 1000 -nph -pd 3 -n 1 -demux /dev/dvb/adapter0/demux0 0x64

(example PMT PID 0x64 is BBC ONE SD)

 

But when script is running, expected behavior is to get PMT with above dvbsnoop command. If you like experimenting, you can run script with different numbers in

lsDemux = glob.glob("/dev/dvb/adapter0/demux*")

... in the place of *

 

And while script is running, to try catch PMT with dvbsnoop....

 

When your script is not running, with this command dvbsnoop -timeout 1000 -nph -pd 3 -n 1 -demux /dev/dvb/adapter0/demux0 0x64 I get timeout on all demuxes except the last one (demux6):

root@et9x00:~# dvbsnoop -timeout 1000 -nph -pd 3 -n 1 -demux /dev/dvb/adapter0/demux6 0x64
dvbsnoop V1.4.51 -- http://dvbsnoop.sourceforge.net/
Error(22): DMX_SET_FILTER failed: : Invalid argument

 

Then I start your script (in this example deb4 version) and I seem to get PMT values on demux5:

root@et9x00:~# available demuxes:
/dev/dvb/adapter0/demux6
/dev/dvb/adapter0/demux5
/dev/dvb/adapter0/demux4
/dev/dvb/adapter0/demux3
/dev/dvb/adapter0/demux2
/dev/dvb/adapter0/demux1
/dev/dvb/adapter0/demux0
currentSid: 0x1044
trying with demux /dev/dvb/adapter0/demux6
Error(22): DMX_SET_FILTER failed: : Invalid argument
trying with demux /dev/dvb/adapter0/demux5
pmtPid: 0x0064
Error(145): read (PID 0x0064): Connection timed out
0x9f 0x80 0x32 0x1e 0x03 0x10 0x44 0x09 0x00 0x13 0x01 0x81 0x08 0x0c 0xfd 0xac
0xe7 0x10 0x00 0x00 0x01 0x82 0x02 0x20 0x05 0x84 0x02 0x00 0x64 0x0d 0x00 0x64
0x00 0x00
Error(145): read (PID 0x0064): Connection timed out

 

When the script is still running, I get back to dvbsnoop -timeout 1000 -nph -pd 3 -n 1 -demux /dev/dvb/adapter0/demux0 0x64 command and I  get timeout on all demuxes except the last one (demux6), that is exactly the same as when your script was not running.

[display name]

So it looks like there is a significant difference between DMM original image and OpenPLi images which prevents doublet's script from doing its job.

Maybe there is still hope and it is not a question of drivers or hardware support then.

If only some OpenPLi developers could look into doublet's script and try to figure out why it is working on the original image only.

Maybe doublet can try with OpenPLi image and find out why it's not working?

(and it's easier for self testing)

[pieterg]

There should be no real difference, possibly the demux allocation is done in a different order. But demux allocation should be considered random anyway.

[doublet]

Tried to put OpenPli to usb stick on my 7020HD (http://openpli.org/wiki/DM800USBBoot) but it didn't finish boot (both 3.0 and 4.0). I will check with console cable tonight what's wrong there....

Edited by doublet, 30 September 2013 - 07:06.

[Robinson]

I think OpenPLi does not like multiboot.  

I understand your reluctance to swap images if everything works OK for you.

[WanWizard]

It's not that OpenPLi doesn't like multiboot.

 

The problem with multiboot is that you boot the box with the kernel of the base image, and then the multiboot software does some wizardry to swap the mounts with those of another image.

 

Since OpenPLi is quite different from OE (DMM) based images, it can cause all sorts of misery due to imcompatibilities.

 

This has caused us a lot of work chasing ghosts in the past (i.e. issues that were related to the multiboot, and didn't appear when booting OpenPLi), so it's simple: if you have multiboot active, we don't provide support. Recreate the issue without multiboot, and if the issue is still there, then report it and we'll have a look...

[doublet]

It's not that OpenPLi doesn't like multiboot.

 

The problem with multiboot is that you boot the box with the kernel of the base image, and then the multiboot software does some wizardry to swap the mounts with those of another image.

From http://openpli.org/wiki/DM800USBBoot I've got impression that there isn't any kind of mix between flash and usb. Once I set in BIOS primary boot source USB with /autoexec.bat and  there is OpenPli linux kernel in first partition, box shall boot it.

[WanWizard]

Ah, but that is not multiboot, that is just changing the boot device. See the remarks under "Final thoughts"...

[Robinson]

Regarding your friend's box:

 

First screen shows that E2 thinks there is only one stream of MPEG format. In reality in PMT are audio 2 streams (0x17d6 and 0x17da) of HE-AAC format.

 

Second screen shows that E2 receives audio PID 0x17da. I don't know why E2 selected this stream (and not 0x17d6), because on 0x17da there is no sound broadcasted...

 

Maybe it's something wrong with PID caching... you could try edit lamedb... maybe to keep just "p:BBC" with no additional PIDs....

 

OK, let me reply regarding my friend's situation on IPBOX.

When he removes PIDs from lamedb, there is no sound or picture even if your script is running.

The PIDs shown in Service Information screen are actually those that he put in lamedb himself. No matter if he puts 0x17da or 0x17d6, there is no sound on HD channels.

Mind you, he also gets 2 timeouts in your script's log, but both sound and picture appear on SD channels.

Also, "timeout reading PMT" appears on his TV screen despite the fact that there is actually picture (and also sound on SD channels).

 

wget: bad address 'localhost'
currentSid: 0x
currentSid: 0x4484
trying with demux /dev/dvb/adapter0/demux0
pmtPid: 0x19c8
Error(110): read (PID 0x19c8): Connection timed out
0x9f 0x80 0x32 0x1e 0x03 0x44 0x84 0x09 0x00 0x13 0x01 0x81 0x08 0x0c 0xfd 0xac 0xe7
0x10 0x00 0x00 0x01 0x82 0x02 0x01 0x00 0x84 0x02 0x19 0xc8 0x0d 0x19 0xc8 0x00 0x00
Error(110): read (PID 0x19c8): Connection timed out

 

It seems that there are some differences between images and, possibly, boxes, which prevent your script for being universal for all.

For the time being it seems that it does its job only on DMM images (7020, 800se).

[doublet]

Robinson,

with your friend's box that's weird. It seems that script doesn't have any added value there.... Isn't he getting picture (and sound on SD) even without running script?

 

Meanwhile I've resolved issue with booting OpenPli from USB... so I'll be able to look what's different.

Edited by doublet, 30 September 2013 - 19:44.

[doublet]

So here are my findings:

- image OpenPLi-4.0-beta-dm7020hd-20130930 (booted from USB)

- same *CAM used as with DMM OE 2.0 image (see my past posts)

- same channel settings used (see my past posts)

 

Demux allocation is different: demux0 is used (on DMM OE 2.0 demux3 is used). Have to force script using demux0, because dvbsnoop with wrong demux stucked whole box (had to power off with HW switch).

Change in bbc_pmt_deb4.py:

lsDemux = glob.glob("/dev/dvb/adapter0/demux0")

 

Then after running script all channels are descrambled including sound on HD.

 

I think I've read somewhere in forum, that OpenPli uses DMM drivers in Dreambox images. So perhaps DMM driver (for CA device) is able to instruct the chip to descramble PMT section packets.

 

Note to E2 behavior:

While script was running, I've zapped trough all channels to check whether they are ok or not. Then I stopped the script. After that I zapped trough channels again. I've got channels descrambled (but no sound on HD), and PMT timeout message appeared.

So it seems this OpenPli E2 is able to generate CAPMT data from cached PMT data. But as cached PMT data doesn't cotain HE-AAC PID, there is no sound on HD.

[Robinson]

Robinson,

with your friend's box that's weird. It seems that script doesn't have any added value there.... Isn't he getting picture (and sound on SD) even without running script?

 

Oh, it does have added value. Without it, there is no sound and no picture.

He needs either your script or the pmt.tmp file.

[Robinson]

Demux allocation is different: demux0 is used (on DMM OE 2.0 demux3 is used). Have to force script using demux0, because dvbsnoop with wrong demux stucked whole box (had to power off with HW switch).

Change in bbc_pmt_deb4.py:

lsDemux = glob.glob("/dev/dvb/adapter0/demux0")

 

Then after running script all channels are descrambled including sound on HD.

 

I think I've read somewhere in forum, that OpenPli uses DMM drivers in Dreambox images. So perhaps DMM driver (for CA device) is able to instruct the chip to descramble PMT section packets.

Thanks doublet.

Unfortunately, lsDemux = glob.glob("/dev/dvb/adapter0/demux0") line does not work for me. Nothing changes.

El Sato, did you try it on OpenPLi? Does it work for you on DM800se?

 

I have done some more tests on Clarke-Tech ET9000 and VU+ Solo2 (both with OpenPLi) and they seem to behave in exactly the same way. None of doublet's script versions opens any channels.

BBC One England (SD) is, however, opened on both using the correct pmt.tmp file, which doublet posted several days ago. When this file is present in /tmp, the channel opens, also with your script running in the background, so it seems that this script indeed does not do much good in my case.

I am disappointed at myself because actually several days ago I managed to open BBC 2 HD (picture only) with another of your pmt.tmp files. Unfortunately, I cannot reproduce it now; neither on ET9000 nor on Solo2. I'm not sure what happened in the meantime.

[Robinson]

I'm just browsing through this subject to see what I could have missed.

 

http://www.streamboa...ig/oscam.dvbapi

 

I presume an A: line will do the trick (!)

This does not seem to work for me. Maybe it just works for Oscam and not for Oscam_ymod, which may not support the A line.

 

 

Backup of ca0 and demux0:

mv /dev/dvb/adapter0/ca0 /dev/dvb/adapter0/ca4
mv /dev/dvb/adapter0/demux0 /dev/dvb/adapter0/demux5

Symlink ca0 -> ca3, demux0 -> demux3:

ln -s /dev/dvb/adapter0/ca3 /dev/dvb/adapter0/ca0
ln -s /dev/dvb/adapter0/demux3 /dev/dvb/adapter0/demux0

I didn't try that but probably this is specific to your box/image.

 

 

Alternatively to making symlinks like I described earlier, you can rename devices higher than 0 and then restart E2. This forces E2 to use demux0 / ca0, so oscam and E2 will use same devices.

I didn't try that either.

 

 

To watch a HD channel with sound (should work for any other channel on TP):

 

- remove "f:4" flag from channel in lamedb to make E2 use PMT being broadcasted, reload settings

- put your receiver into standby (E2 has to be still running)

- create /tmp/pmt.tmp containing just PMT PID. Here is example for BBC 2 HD:

4440_pmt_bin.txt

- now CAM should start to descramble PMT PID

- bring your receiver from standby to operational state

- E2 will receive descrambled PMT and send "capmt" data (includes all the PIDs of the channel) to CAM

- now CAM should start to descramble all PIDs of the channel

- you should get picture and sound

Please note, doublet, that I failed all along starting right from here, from this PMT-only HD pmt.tmp file. Neither this file nor your script worked for me.

 

[doublet]

Robinson,

 

if you would like to experiment some more, I've created new version of script:

- have removed demux detection (as it seems to be not reliable)

- there are 2 modes of operation:

+ short (CAPMT with only PMT PID is sent to CAM - for boxes which are able to descramble PMT PID, like Dreambox)

+ long (CAPMT with VPID and APID is sent to CAM - for other boxes)

 

Of course for long mode, you have to have PIDs in lamedb. For reasons mentioned earlier (no HE-AAC PID possible in lamedb), there will be no sound for HD channels in long mode.

 

 bbc_pmt_v1.zip   1.64KB   129 downloads

 

Usage examples:

./bbc_pmt_v1.py -t short -d /dev/dvb/adapter0/demux3

./bbc_pmt_v1.py -t long -d /dev/dvb/adapter0/demux0

 

[Robinson]

Of course I would like to experiment more.   Thanks.

Unfortunately this new script does not work for me at all.

I'll provide you with some logs in a private message.

[doublet]

Minor correction of wget parameters:

 

 bbc_pmt_v2.zip   1.64KB   100 downloads

 

Works on Robinson's ET9000 (in "long" mode) and my Dreambox (in "short" mode).

[doublet]

There is minor update to script itself.

I've added two shell scripts to be used for automatic (re-)start of bbc_pmt*.py script with E2.

 

 bbc_pmt_v3.zip   2.14KB   302 downloads

 

Content of archive to be copied into /usr/bin, make files executable.

 

bbc_pmt_v3.py - python script - to be started manually or using scripts below

bbc_pmt_starter.sh - starts bbc_pmt_v3.py, keeps it running forever

enigma2_pre_start.sh - kills bbc_pmt_starter.sh and bbc_pmt*.py, then starts bbc_pmt_starter.sh

 

enigma2_pre_start.sh is called by enigma2.sh before enigma2 itself is started. Before copying enigma2_pre_start.sh, double check if file doesn't already exist.

 

customization: adapt lines in bbc_pmt_starter.sh as needed:

 

scriptName="bbc_pmt_v3.py -t short -d /dev/dvb/adapter0/demux3"
logName=/dev/null
#logName=/tmp/bbc_pmt.log

[Robinson]

I understand this could also work in the short mode?