[mythtv] DVB 7133 uk full dvb-t scan problems

[John Pullan]

On 26/08/05, Stuart Auchterlonie <stuarta at squashedfrog.net> wrote:
> On Thu, Aug 25, 2005 at 09:35:44PM +0100, John Pullan wrote:
> > Using 7133, Full uk country scan (dvb-t)
> > I'm seeing the scanner stop after it processes the first multiplex. It
> > segv'd the first time I tried to run it with extra debug, so I ran it
> > under gdb where it appeared to freeze. (so I've attached the gdb trace
> > of that as well)
> >
> > On the plus side it only created one multiplex :)
> >
> 
> I'm digging around this at the moment.
> 
> The SEGV comes from the SISCAN debug print attempt at the top of
> ScanWizardScanner::scan(). A nice little #if 0 has temporarily
> solved the problem.... :-)
> 
> The reason it is only creating one multiplex is because it is
> getting stuck (or frozen..)
> 

Yup I know, just pointing out that all the other ones don't get created :P

> It looks like it is waiting for the card to tune.
> Here is the relevant logs
> 
> --- mythtv-setup log ----
> 
> 2005-08-26 11:27:25.426 SIScan(0): UpdatePATinDB(): mplex: 1:1
> 2005-08-26 11:27:25.426 GetBetterMplexID(mplexId 1, tId 8197, netId 1)
> 2005-08-26 11:27:25.485 GetBetterMplexID(): Returning perfect match 1
> 2005-08-26 11:27:25.606 SIScan(0): Tuning to  offset 1 mplexid(1)
> 2005-08-26 11:27:25.636 DVB#0 Frequency: 482000000 BW: 8MHz HP: Auto LP:
> AutoC: Auto TM: 2K H: None GI: 1/32
> 2005-08-26 11:27:25.636 DVB#0 Waiting for event
> 
> --- end log ---
> 
> and from gdb
> 
> --- backtrace ---
> 
> #2  0xb7d1fc88 in get_dvb_event (fd=12, event=@0xb5977ca0, block=true)
> at dvbchannel.cpp:869
> #3  0xb7d1d2b3 in DVBChannel::TuneTransport (this=0x81b9fd8,
> channel=@0x81ba0ec, all=false) at dvbchannel.cpp:620
> #4  0xb7d15d2b in DVBChannel::TuneMultiplex (this=0x81b9fd8, mplexid=1)
> at dvbchannel.cpp:176
> #5  0xb7c2280c in SIScan::Tune (this=0x8152b30, transport={_it = {node =
> 0x818e008}, _offset = 1}) at siscan.cpp:665
> #6  0xb7c2311f in SIScan::ScanTransport (this=0x8152b30, transport={_it
> = {node = 0x818e008}, _offset = 1}) at siscan.cpp:724
> 
> --- end backtrace ---
> 
> so it is sitting in get_dvb_event getting EWOULDBLOCK back from
> the FE_GET_EVENT ioctl.
> 
> 
> Now I remember reading something in the last few days about what
> the DVB API does when you request a tune to the frequency the
> card is already tuned to (I think this was on linux-dvb mailing list)
> 
> I believe the answer that came back was that
> a) No Event was delivered if the card had no change in tuning
> b) You can't issue a FE_GET_STATUS until a FE_SET_STATUS(?) had be done.
> 
> 
> More research to do...
> Stuart
> 
Things change :)
It now no longer locks, just repeats scanning the same mux . I'm at
work now but I was going to have a closer look tonight when I got back
home.
-- 
John