71 replies to this topic

[orator89]

Ehi, I start saying that this problem is already experienced by various users (you can read on other forums) with the "Openpli" and similar images, but disappears on OpenATV, as you can read from other users. below the problem:

 

RTL2832 usb device installs correctly and is correctly recognized and it appears correctly in tuners list, but when try to scan the openpli icon spins continually and the device doesn't respond to commands and you have to restart. The same happens on OpenSPA, OpenHDF and other with the same Kernel 4.4.35. Instead, on OpenATV 6.4, with the same kernel, everything works. I already tried to use the same drivers and .ko modules used in the working OpenATV, but nothing changes... It seems a problem with "vtuner" as you can see from dmesg on when system initialized:

 

 

user.err kernel: vtunerc0: unknown IOCTL 0x3

user.err kernel: vtunerc0: unknown IOCTL 0x4
user.err kernel: vtunerc0: unknown IOCTL 0x5

user.err kernel: vtunerc0: unknown IOCTL 0x6

user.warn kernel: vtunerc0: usb frontend initialized as type=0

 

and when you scan and system starts spinning, you get error:

 

user.err kernel: vtunerc0: unknown IOCTL 0x1

 

in the working OpenATV 6.4 there aren't those "unknown IOCTL" errors, and, if I remember correctly, dmesg shows "vtunerc0: usb frontend initialized as type=4" 

and not "type=0", but maybe I'm wrong... Any suggestion? Thanks in advance

Edited by orator89, 20 July 2022 - 15:41.

[WanWizard]

What box are we talkng about?

[orator89]

What box are we talkng about?

Uclan Ustym 4k pro (clone of sf8008), but if you read on internet it's the same on official Sf8008 and I don't know if also on other decoders.

Edited by orator89, 20 July 2022 - 21:43.

[orator89]

I can also tell you that the system icon starts spinning and locking only when antenna cable is inserted in the dvb usb and there's signal. If, instead, antenna is not inserted in the usb tuner, the searching goes ok but obviously it doesn't find any channel.

Edited by orator89, 20 July 2022 - 21:47.

[WanWizard]

Could be ATV has driver patches we don't have.

[orator89]

And eventually where I can find those driver patches in openATV image? Because i tried also copying entire folder "4.4.35" from kernel directory of openATV image and copied all the folder and subfolder with drivers kernel on the openli using commands "init 4", FTP transfer, and the "init 3" (all the permissions setted to 777), and then restarted box normally, but nothing changed, same problem also with kernel folder of OpenATV....

Edited by orator89, 20 July 2022 - 22:19.

[WanWizard]

It is not that simple, if there are patches, they are on the kernel source code.

 

The ATV image core ( = everything apart from Enigma) can be found here: https://github.com/o...-alliance-core/

[orator89]

Understood. I downloaded and seen the "meta-uclan" folder in "meta-brands", from openatv core, and also from openpli core, with subfolders and "*.patch" files inside. Now, if it's really a patch as you ipotized, maybe you can include in a test image the same ".patch" files openatv has? If, instead, is a general patch for all decoders to 4.4.35 kernel, I don't know wich patch file to use from "meta-oe" folder...

Edited by orator89, 21 July 2022 - 00:14.

[WanWizard]

I don't know either, haven't checked, and atm don't have the time to do so and make test images, especially when I don't have the hardware to test with.

[orator89]

I don't know either, haven't checked, and atm don't have the time to do so and make test images, especially when I don't have the hardware to test with.

But are you sure is a patch and is not a problem with the configuration of vtunerc in your image?

Because on openATV, I can see the two correct mex of vtunerc:

 

vtunerc0: setting DVB-T tuner type

vtunerc0: usb frontend initialized as type=4

 

Instead on openPLI I don't see the correct "tuner setting" mex:

"vtunerc0: setting DVB-T tuner type"

 

and maybe, I suppose, because of this lack, you see the errors and you see

 

vtunerc0: usb frontend initialized as "type=0"

and not the correct "type=4"

Edited by orator89, 21 July 2022 - 09:35.

[WanWizard]

No, as I wrote I haven't looked at it at all, it was just an assumption that needed checking.

 

Afaik there is no "vtuner" configuration, the driver determines the type.

[orator89]

No, as I wrote I haven't looked at it at all, it was just an assumption that needed checking.

 

Afaik there is no "vtuner" configuration, the driver determines the type.

Ok, i found the solution, but only you guys can apply. I wrote to OpenATV developers, and as I tought, is not a driver/kernel error, is a mistake in configuration of image... One developer replied and told me that mex with links to view. As I tought it will be good

 

 

 

I'll try to give you a hint that might help you.

Since you have provided the appropriate log entries, you can try to start there.

Then the magic word is: VMSG_TYPE

If you look at the configure.ac script of OpenATV's Enigma2, you will find something for the SF8008.

In the configure.ac script, a variable in the building system is defined with VMSG_TYPE2=1.

This variable is in turn evaluated in Enigma2 and you can find the here.

The IOCTL numbers are changed in the constructor of the eDVBUsbAdapter class depending on the variable VMSG_TYPE2.

When I look at the configure.ac script of OpenPLi's Enigma2, I don't find a definition of the variable VMSG_TYPE2 there.
Furthermore, only a variant of the IOCTL numbers appear in the constructor of the eDVBUsbAdapter class.

That may be the reason why it works under OpenATV and not under OpenPLi. So it would only indirectly be a driver problem / kernel problem.

The patch for OpenPLi's Enigma2 would be the differentiation of the IOCTL numbers as in OpenATV. It may be possible to activate the other variant of the IOCTL numbers via a parameter for the corresponding kernel module.

Please, if you can help solving that... Thanks...
 

[4l3x2k]

@orator89

I replied to you at OpenATV.

 

This change might be enough to bring the driver to life.

 

So I think that OpenPLi on the SF8008 (or clone) doesn't work as a SatIP client either.

[orator89]

yes thanks for the reply. 

 

 

@orator89

I replied to you at OpenATV.

 

This change might be enough to bring the driver to life.

 

So I think that OpenPLi on the SF8008 (or clone) doesn't work as a SatIP client either.

 

Yes thanks for the reply. I don't need SatIP client So, changing and adding the variable vmesg type and variable of ioctl can solve the problem definitely

 

The two files to modify are "configure.ac" and "dvb.cpp", if I understood correctly

Edited by orator89, 21 July 2022 - 18:30.

[WanWizard]

It was added in https://github.com/o...aa2f17563f3a16d, with the remark "test".

 

I don't have the skills to judge what exactly this does, what the impact is, and why it is limited to those stb models (as it doesn't look to be hardware dependent?).

[orator89]

It was added in https://github.com/o...aa2f17563f3a16d, with the remark "test".

 

I don't have the skills to judge what exactly this does, what the impact is, and why it is limited to those stb models (as it doesn't look to be hardware dependent?).

 

Yes I see, I see but I think that adding a line with vmsg variable for vtuner and then appli an "if" variable for ioctl can solve the problem on Openpli and others images, because on OpenATV stable image this modification works perfectly 

Edited by orator89, 21 July 2022 - 19:06.

[4l3x2k]

@WanWizard

It might help to look at the lines after the #DEFINES. There they are used directly.

 

In my view there are two reasons for this:

* The kernel version and thus the DVB API are different

* The low IOCTL numbers conflict with other features for this device class

 

This means that these IOCTL numbers depend on the driver / kernel used.

 

The linked commit is marked test, but it is included in current versions of OpenATV (6.4, 7.0, 7.1) as well.

 

Of course, it would be nice if someone with experience in this area could step in.

[4l3x2k]

yes thanks for the reply. 

 

 

@orator89

I replied to you at OpenATV.

 

This change might be enough to bring the driver to life.

 

So I think that OpenPLi on the SF8008 (or clone) doesn't work as a SatIP client either.

 

Yes thanks for the reply. I don't need SatIP client So, changing and adding the variable vmesg type and variable of ioctl can solve the problem definitely

 

The two files to modify are "configure.ac" and "dvb.cpp", if I understood correctly

 

I only mentioned SATip because it might increase the number of testers.

 

SATip as a client does not work under OpenPLi on the SF8008.

 

SATip also requires the vtunerc driver.

[4l3x2k]

I think the problem can simply be solved by installing the right driver from feed.

 

@orator89
Check the installed kernel modules on the one image and on the other image.

Do you have SATip client installed?

 

If that is the case then these dmesg messages can be misleading and are the reason why SATip does not work but it should not affect your usb dongle.

Edited by 4l3x2k, 21 July 2022 - 20:02.

[orator89]

I think the problem can simply be solved by installing the right driver from feed.

 

@orator89
Check the installed kernel modules on the one image and on the other image.

Do you have SATip client installed?

 

If that is the case then these dmesg messages can be misleading and are the reason why SATip does not work but it should not affect your usb dongle.

I don't have SATip client installed..... If you talk about solving vtunerc problem, no. I've already installed the drivers and modules ko and firmware on openPLI using the feeds opkg of openATV 6.4 and also tried using the entire folder "modules" and "firmware" from openATV 6.4 image, copying on Openpli and replacing all, with "Init 4", "ftp transfer" and "reboot" commands. Drivers not solve anything... Same errors with vtunerc IOCTL...
You already found the real solution in files "configure.ac" and "dvb.cpp"...

Edited by orator89, 21 July 2022 - 20:16.