Softcam issue?

[abu baniaz]

Using the ATV/PLi control method, I left box in standby overnight. I woke box up and found oscam had stalled and was not shown in htop.

I'll flash ATV and repeat test.

[abu baniaz]

I'll flash ATV and repeat test.

I flashed OpenATV 6.4 (03 July 2021), used the exact same oscam binary, exact same oscam config files (.server, .dvabapi, .user, .conf etc). Oscam did not freeze/stall over time.

I used same the same lamedb file as what I had used on Vix, and left it on the same 19.2 channel overnight.

[johnhoe]

I think this confirms what many of us have been saying... the issue is very odd and not happening in other Open images. Although (except for the eventual freezing) it does not seem to happen (using Softcam Manager) in the Python 3 version of the image I am using. So this problem may disappear with Py3.
I have been looking at what changed between Vix 5.4.4 and 5.4.5 where the issue started and there are some specific code changes that only applied to MIPS boxes not ARM boxes. Maybe this is where the problem is.

[ccs]

I have been looking at what changed between Vix 5.4.4 and 5.4.5 where the issue started and there are some specific code changes that only applied to MIPS boxes not ARM boxes. Maybe this is where the problem is.

Such as ???

[johnhoe]

In #195 in the file vix_e2.txt there is an entry as follows:-
commit 93f9626b873f501ea9c17c6ec49abd0d0122c5e7
Author: Erik Slagter <erik@openpli.org>
Date: Fri Jan 8 15:27:26 2021 +0100

filepush: work around bug in OE Zeus glibc. Fix recordings that hang enigma.

On MIPS receivers the default signal blocking mask for a thread appears
to be undefined (and it's really random). If we don't set it explicitly,
it will sometimes let SIGUSR1 work and on other times it won't.

So set the signal blocking mask in the recording thread explicitly now,
and that seems to work.

On ARM receivers this issue never showed.

Maybe nothing....

[abu baniaz]

All those filepush commits were reverted and are not in the image. (Maybe I made a mistake including commits a day either side.)

[ccs]

.... I wondered about filepush for a long while, not finding it in history until I realised it had been reverted.

Is there ever a chance that such long winded commits might not be totally reverted?

[twol]

From memory, Huevos rolled everything back to a specific commit removing any in between commit and history from Git.
The python3 branch has exactly the same revert.

[ccs]

From memory, Huevos rolled everything back to a specific commit removing any in between commit and history from Git.
The python3 branch has exactly the same revert.

Yes, it looks absolutely fine to me, I just wondered if there was the slightest chance that revert didn't quite do everything expected of it.

[Huevos]

Yes, it looks absolutely fine to me, I just wondered if there was the slightest chance that revert didn't quite do everything expected of it.

What makes you think that?

[ccs]

What makes you think that?

Because no one has come up with anything else, something very subtle has happened, and "the slightest chance" was just pure speculation.

I've even looked to see if there are any .py files lying around which shouldn't be there, but I haven't found any.

[Huevos]

What exactly is the problem?

[ccs]

What exactly is the problem?

You might get away with starting here... https://www.world-of-satellite.com/s...l=1#post515234

Otherwise, post #1.

[GiP]

Hi,

Has the problem been resolved?
After I reinstalled 5.4.13 on a VU+ Duo2 my Oscam-latest keep shutting dow while watching.

Thanks,
G

[twol]

Hi,

Has the problem been resolved?
After I reinstalled 5.4.13 on a VU+ Duo2 my Oscam-latest keep shutting dow while watching.

Thanks,
G

see post #183