mythtv-setup DVB channel scan seg fault (rev 7534)

[mythtv-trac@…]

Doing a channel scan causes a segmentation fault. Using a HD-3000 ./configure --enable-xvmc --enable-opengl-vsync --enable-dvb --disable-xvmc-vld --compile-type=debug

Attaching log and backtrace.

[mythtv-trac@…]

Log and backtrace of channel scan

[seth.lists@…]

Here is another backtrace, this time with rev. 7466 and an AirStar HD card.

./configure --compile-type=debug --enable-dvb --enable-xvmc --enable-opengl-vsync --disable-joystick-menu

Attached backtrace below.

[anonymous]

backtrace of 2005-10-25 myth cvs, segfault doing qam256 std cable channel scan with airstart2/hd5000 tuner

[danielk]

debugging patch

[danielk]

Can someone try the debugging patch and report back with the output of "mythtv-setup -v siparser,channel" and the backtrace from the crash?

[mythtv-trac@…]

Backtrace with current svn and debug patch

[mythtv-trac@…]

I attached the log and backtrace. With this build, it didn't get as far as it did previously. Here, it seems to have crashed on the first channel lock. With the first backtrace I attached it crashed on about the 4th channel lock. Not sure if that matters. I hope this helps. Thanks.

[danielk]

Can you run this a 3-4 more times and produce different stacktraces? The last one had a corrupted stack.

Also if you know how to run Valgrind that might be even more helpful. This appears to be a corrupted memory problem.

[danielk]

(In [7661]) References #498.

Adds a VERBOSE(VB_SIPARSER, ..) macro to ProgramMapTable::Parse()

[danielk]

(In [7662]) References #498.

Adds some sanity checks on PMT table in SIScan::UpdatePATinDB().

[danielk]

I've committed a few sanity checks and a debugging macro.

Valgrind doesn't report anything for me, but it might for the people experiencing the SEGFAULTS.

Since I can't reproduce the problem and I'm not getting responses from the bug reporters anymore, I'm marking this as "worksforme"...

[matthew.annen@…]

mythtv-setup full cable chan scan using valgrind with memory leak test=yes

[matthew.annen@…]

I ran mythtv-setup using valgrind. Above I have attached the output. Mythtv-setup segfaulted during scan. -matt

[danielk]

Matt, can you create some backtraces too?

If they all look the same, just send me the one. But since this appears to be a memory problem, they might be different.

Unfortunately, I couldn't find anything useful in the valgrind output.

[matthew.annen@…]

I ran mythtv-setup twice using the command line args you mentioned above. I ran a full qam256 cable scan. Both times the scan crashed at what seems to me to be the same point in the scan, at C105 somewhere. I hope my traces are of some use. I copied some of the corresponding console output from mythtv-setup, too. The files are paired above. Thanks.

-matt

[matthew.annen@…]

FYI, I was using rev 7662.

-matt

[matthew.annen@…]

Wow, I was able to get a few of the FTA channels on the scan at least. This looks GREAT!

[mythtv-trac@…]

I'll get you some more backtraces and valgrind output, but it will be a couple more days before I can get to it. Thanks for looking in to this.

[danielk]

It looks like the problem has to do with DVB tables showing up in an unexpected place. I'll add some debugging for this...

Just to satisfy my curiosity, are you in North America?

[danielk]

(In [7690]) Fixes #498.

There was an indexing problem in handling NIT tables. This was being seen unexpectedly during an ATSC (QAM) scan; but this should also fix at least part of the problems seen when disabling the SIParser during channel scanning.

[mythtv-trac@…]

Fixed it for me THANKS!!!! ... and yes, I'm in North America..