permanent artifacts with commit 410b47f

[nikolauzi11]

Hi @Barracuda09

First of all, I would like to wish you all the best for the New Year. Unfortunately there are permanent artifacts with the last commit(410b47f). The client is still the VDR.

[Barracuda09]

Hi @nikolauzi11

Yes, you too have a great new year,

So the commit before this was beter? or is this stiil as #181

[nikolauzi11]

Hi @Barracuda09

With commit bdab5f2 it still looks good. I always try to stay somewhat up to date.

[Barracuda09]

Hi @nikolauzi11

Thanks for your support! it is appreciated.

Could you try for testing, to modify Filter.h in: https://github.com/Barracuda09/SATPI/blob/410b47f5d9aaed43d41976b7ea6c345f171397e1/src/mpegts/Filter.h#L205-L215

And change Line 211 to:

			_mutex.unlock();
			const bool done = openPid(pid);
			_mutex.tryLock(15000);

---

And at this place: https://github.com/Barracuda09/SATPI/blob/410b47f5d9aaed43d41976b7ea6c345f171397e1/src/mpegts/Filter.h#L220-L230

And change Line 226 to:

			_mutex.unlock();
			const bool done = closePid(pid);
			_mutex.tryLock(15000);

[nikolauzi11]

Hi @Barracuda09

This makes it work again. Thanks!

[Barracuda09]

Hi @nikolauzi11

That is nice, thanks for trying this. Then we found a permanent solution for this problem.

The problem is (probably) the combination of VDR, Sundtek tuner and this locking in SatPI. VDR request PIDs (or scans PMT) very often when tuned to a specific frequency, and open/close PIDs takes more time for the Sundtek tuner. So this lock (during open/close PIDs) was locking the read of TS data from tuner, now by unlocking the mutex during open/close PID the read can then continue.

I do not see this behavior on my Vu+ receiver, there the open/close of PIDs does not take much time. But testing with the Sundtek tuner I saw this situation.