33 replies to this topic

[johmer]

Hi,

 

I have yesterday updated my dreambox 500HD to the latest nightly build and noticed that every time that i try to switch to an IPTV channel (stream), the box crashes. Attached is the log.

 

Can someone look at this please?

 

Thanks.

Attached Files

•  enigma2_crash.log   9.54KB   6 downloads

[athoik]

Issue is generic one.

I am waiting for stack trace with debug symbols that can show where it crashes.


It can be the following that crashes:
 

+					int system;
+					ePtr<iDVBFrontendParameters> parms;
+					it->second.channel->getCurrentFrontendParameters(parms);
+					parms->getSystem(system);
+
+					it->second.timetable = NULL;
+					if (system == iDVBFrontend::feATSC)
+					{
+						it->second.timetable = new STT(it->second.channel);
+					}
+					else
+					{
+						it->second.timetable = new TDT(it->second.channel);
+					}
+					it->second.timetable->start();

The getCurrentFrontendParameters might return NULL parms on streams.

So we need to check parms before accessing getSystem if above is happening.
 

+					if (parms)
+					{
+						parms->getSystem(system);
+					}

Edit1. Or it can be my change the "simplify thing"...

Edited by athoik, 30 October 2016 - 07:46.

[johmer]

Thanks for your reply.

 

Hence, what do you recommend for now? Will a fix be available soon or should I try to rollback to a previous version?

[vaki]

What to do now

[athoik]

First we need to identify the commit that causes the problem, then fix it or revert it.

If that happens today tomorrow will be ok, unless a manual build occurs.

[Meega]

Does this also has anything to do with remote falback tuner ,because i get the same crashlog on my dm8000.

I am streaming from et10000 to the dm8000.

 

Thought it was the update on the dm8000 but back to image from 22-10 result in the same crash.

 

Openwebif is segnificant changed ,maybe ?

[Meega]

After disabling remote fallbacktuner in settings ,the crashloop is gone.

[johmer]

On the DM500, the problem is still present even if fallback tuner is disabled.

[littlesat]

Image from 2 days ago is fine... So whenit is a commit it should be one of the latest two days...

[littlesat]

Could it be this one?
https://github.com/O...116e67fe0cbf758

[athoik]

Could it be this one?
https://github.com/O...116e67fe0cbf758

No, it's the issue I mentioned before.

Patch soon.

Edited by athoik, 30 October 2016 - 16:49.

[arn354]

Problem still exists even if everything is reverted including https://github.com/O...daeed1fedaf5478

 

i needed to revert the following 3 commits to get it back working:
https://github.com/O...116e67fe0cbf758

https://github.com/O...3a0758a070b4d98

https://github.com/O...0353b7c08a12fc3

[athoik]

diff --git a/lib/dvb/dvbtime.cpp b/lib/dvb/dvbtime.cpp
index 997cfbe..9ca8ad4 100644
--- a/lib/dvb/dvbtime.cpp
+++ b/lib/dvb/dvbtime.cpp
@@ -304,7 +304,7 @@ void eDVBLocalTimeHandler::setUseDVBTime(bool b)
                                        int system;
                                        ePtr<iDVBFrontendParameters> parms;
                                        it->second.channel->getCurrentFrontendParameters(parms);
-                                       parms->getSystem(system);
+                                       if (parms) parms->getSystem(system);

                                        it->second.timetable = NULL;
                                        if (system == iDVBFrontend::feATSC)
@@ -334,7 +334,7 @@ void eDVBLocalTimeHandler::syncDVBTime()
                        int system;
                        ePtr<iDVBFrontendParameters> parms;
                        it->second.channel->getCurrentFrontendParameters(parms);
-                       parms->getSystem(system);
+                       if (parms) parms->getSystem(system);

                        it->second.timetable = NULL;
                        if (system == iDVBFrontend::feATSC)
@@ -542,7 +542,7 @@ void eDVBLocalTimeHandler::updateTime( time_t tp_time, eDVBChannel *chan, int up
                        int system;
                        ePtr<iDVBFrontendParameters> parms;
                        chan->getCurrentFrontendParameters(parms);
-                       parms->getSystem(system);
+                       if (parms) parms->getSystem(system);

                        int updateCount = it->second.timetable->getUpdateCount();
                        it->second.timetable = NULL;
@@ -587,7 +587,7 @@ void eDVBLocalTimeHandler::DVBChannelStateChanged(iDVBChannel *chan)
                        int system;
                        ePtr<iDVBFrontendParameters> parms;
                        it->second.channel->getCurrentFrontendParameters(parms);
-                       parms->getSystem(system);
+                       if (parms) parms->getSystem(system);

                        switch (state)
                        {

Quick fix... I will apply soon above using proper styling.
 

if(parms)
{
    parms->getSystem(system);
}

PS. Why nobody read my first comment?

Edited by athoik, 30 October 2016 - 17:35.

[arn354]

Sorry - correction, athoik is right - patches were not correctly applied

[arn354]

IPTV working again using Athoiks suggestion. thx

Edited by arn354, 30 October 2016 - 17:49.

[athoik]

Here is the backtrace...

<   347.609> download failed, no cuesheet interface
<   347.609> PC: 004a8dd4
<   347.609>     00000000 00000001 00000000 00000004
<   347.609>     00000000 02253a24 7fe2dd64 00d716e0
<   347.609>     02253a00 00002710 8fff843c 00000411
<   347.609>     7fe2dd08 766f5020 004a13ec 02253a00
<   347.609>     00d716e0 009bbc40 7fe2dd30 009bbc54
<   347.609>     021edac8 011c958c 7fe2dfb4 004a13ec
<   347.609>     0046e4cc 00492b38 00000000 00000000
<   347.609>     006e04b0 7fe2dd18 02253a00 004a8dd0
<   347.642> Backtrace:
<   347.643> /usr/bin/enigma2(handleFatalSignal(int, siginfo_t*, void*)) [0x4656A0]
<   347.643> linux-vdso.so.1(n/a) [0x77456230]
<   347.643> /usr/bin/enigma2(eDVBLocalTimeHandler::DVBChannelStateChanged(iDVBChannel*)) [0x4A8DD6]
<   347.643> /usr/bin/enigma2(SigC::Signal1<void, iDVBChannel*, SigC::Marshal<void> >::emit_(iDVBChannel* const&, void*)) [0x4A6894]
<   347.644> /usr/bin/enigma2(eDVBChannel::playSource(ePtr<iTsSource>&, char const*)) [0x4A4194]
<   347.644> /usr/bin/enigma2(eDVBServicePMTHandler::tuneExt(eServiceReferenceDVB&, ePtr<iTsSource>&, char const*, eCueSheet*, bool, eDVBService*, eDVBServicePMTHandler::serviceType, bool)) [0x4D806C]
<   347.644> /usr/bin/enigma2(eDVBServicePlay::start()) [0x5BBA10]
<   347.645> /usr/bin/enigma2(eNavigation::playService(eServiceReference const&)) [0x5AA2E8]
<   347.645> /usr/bin/enigma2(n/a) [0x58B2A8]
<   347.645> /usr/lib/libpython2.7.so.1.0(PyEval_EvalFrameEx) [0x76BB1840]
<   347.645> -------FATAL SIGNAL

In 5 min will send the PR.

[athoik]

Please merge PR: https://github.com/O...nigma2/pull/505

[WanWizard]

Done by Littlesat.

[littlesat]

When we just have an if with one line we could consider (not mandatory!) to remove the {}

[betacentauri]

I would propose not to remove the {}.
Several years ago I commented out the one line and accidentally not the if line. So the if referred to the next line in the code.
I searched several hours why the software behaved in a strange way afterwards.