26 replies to this topic

[Pr2]

Hi,

 

I have flashed today my Xtrend ET-9000 with OpenPLi 6 release version (2017-10-15) then it restaure the settings and I perform and online update, so logo now indicates 2017-10-22.

 

Setup is: Xtrend ET-9000 connected to TV in HDMI without any amplifier in between so Multi-channel downmix is ON in the Xtrend

 

- When the sound bar is high (in the red area) I have sound, if I slightly decrease the sound I have no more sound even if I am still in the red, or orange.

 

The only workaround is to open up the audio menu (yellow button) turn off / on the Multi-channel downmix and sound is back to normal and I can lower and increase it without any problem.

 

Can you have a look at this problem?

 

Thanks,

 

Pr2

P.S.: I was in OpenPLi RC6 and perform the latest possible online update and the problem start from there (just for information) so I flash the real OpenPLi 6 to see if the release version has the problem too.

 

 

[WanWizard]

There are more and more indications that the ET9xxx suffers from the same race conditions as the ET8000 and the ET10000.

 

If you have this issue again, could you login with telnet, and issue the command:

cat /proc/stb/avs/0/volume

and let me know what the value is?  If it is outside the range 0-63, it's got the race condition too.

 

You can confirm that by issuing the command

echo 35 > /proc/stb/avs/0/volume

which should restore the volume at about 50%.

[WanWizard]

p.s.: rc-6.0 has been taken offline, as it was only a test release. so that can't be flashed anymore, and it can't be updated anymore.

 

I have aborted the running build of 6.0-release, due to a bug that caused samba to fail. So if you bump into that, know that the build is running again with a fix, so please update again tomorrow.

[Robinson]

There are more and more indications that the ET9xxx suffers from the same race conditions as the ET8000 and the ET10000.

 

If you have this issue again, could you login with telnet, and issue the command:

cat /proc/stb/avs/0/volume

and let me know what the value is?  If it is outside the range 0-63, it's got the race condition too.

 

et9x00 login: root
Password:
root@et9x00:~# cat /proc/stb/avs/0/volume
265
root@et9x00:~# echo 35 > /proc/stb/avs/0/volume
root@et9x00:~#

[Robinson]

I think we can make an *.sh file with the content as suggested by WanWizard above and place the file somewhere in the flash (where exactly?) and then we can use HotKey to assign execution of the file to an unused key such as MARK.

What do you think?

[WanWizard]

It doesn't really help, as once you use the volume up or down, the problem reappears. Only a restart will fix it.

[Robinson]

I think you are right.

BTW, is Enigma restart enough or do we need box restart?

[Pr2]

Hi,

 

I powered on my 2nd ET-9000 which had a former version of OpenPLi 6 (not RC) and had the problem (but it was late and I don't have the PC close to me), I perform an online update and since I no longer have the problem.

 

You don't need to reboot the box if you have the problem simply toggle off / on the Downmix with the Yellow button and sound is back to normal.

 

Pr2

[Erik Slagter]

I think you are right.

BTW, is Enigma restart enough or do we need box restart?

Try it and share it... I don't have such receivers to test on. I think enigma restart will suffice.

[Robinson]

I will let everybody know when this happens again. Luckily, I have not observed it for 2 days.

[Robinson]

Try it and share it... I don't have such receivers to test on. I think enigma restart will suffice.

 

Indeed. Enigma restart is enough.

[Pr2]

And what about my solution just to disable / enable downmix?

[WanWizard]

If a restart of Enigma fixes it, that may suggest the problem is not a driver problem but an Enigma problem? As it won't re-initialize the drivers?

[Erik Slagter]

But it does close/open the device. That may be enough to reset some values.

[WanWizard]

Which device?

 

On the ET's, it only writes to the proc entry, it doesn't use the mixer (well, it tries, but it's gets a "device busy" error back).

[Erik Slagter]

I see your point. But maybe the closing and re-opening of the other DVB devices also initiate a reset to default values.

[Pr2]

Disabling AC downmix changing the volume and then enabling the AC downmix is enough to fix the problem, I just test it yesterday.

 

But in fact the sound is always still working when it is on high level (vol. near to the max red zone of the volume bar), the sound disappeared when you lower it slightly. So the sound in not really totally off, if you go lower no sound but if you go higher you have sound.

 

But in it seems to me that staying on a DVB channel doesn't produce the problem, I had it when launching an IPTV channel which is using the .ts format (perhaps a pure coincidence). And I wonder if the problem is not coming from the time needed to sync audio and video.

[WanWizard]

I never use IPTV, so that seems to be a coincedence.

 

The problem with both issues seem to be that some race condition triggers a change in the value. In case of the volume I haven't checked as the proc value is in decimal, but for the overscan, the problem is an 8-bit left-shift. My assumption is that the compiler does some sort of word alignment on de variable that stores the value, but the code using it accesses it using a pointer which is or becomes wrong due to the alignment.

 

That would also explain why 6.0 triggers the problem, as it is compiled with a much more recent version of gcc (6.2.x vs 4.9.x).

[Robinson]

But in it seems to me that staying on a DVB channel doesn't produce the problem, I had it when launching an IPTV channel which is using the .ts format (perhaps a pure coincidence). And I wonder if the problem is not coming from the time needed to sync audio and video.

 

It happens on DVB channels, I'm afraid.

 

The downmix off-on seems to have done the trick although I'm not sure if it always does.

[kalehrl]

I'm also having troubles with bad, crackling analogue sound on Xtrend ET6000.

I thought it had something to do with bad capacitors.