[vdr] Re: dvbdevice optimization

["C.Y.M" <syphir@syphir.sytes.net>]

C.Y.M wrote:
> Klaus Schmidinger wrote:
>
> > "C.Y.M" wrote:
> >
> > > The following optimization was discovered using valgrind on vdr-1.3.12.
> > > Please let me know what you think of it. Maybe it can get into 1.3.13.
> >
> >
> > Maybe I'm missing something here, but I'm afraid I don't
> > quite see what difference that would make.
> >
> > Can you enlighten me?
> >
> > Klaus
> >
> >
> > > --- vdr-1.3.12/dvbdevice.c.original     2004-10-11 09:50:03.000000000 
> > > -0700
> > > +++ vdr-1.3.12/dvbdevice.c      2004-10-11 09:50:44.000000000 -0700
> > > @@ -481,6 +481,7 @@
> > >         frontendType = feinfo.type;
> > >         ciHandler = cCiHandler::CreateCiHandler(DvbName(DEV_DVB_CA, n));
> > >         dvbTuner = new cDvbTuner(fd_frontend, CardIndex(), 
> > > frontendType, ciHandler);
> > > +        aPid1 = aPid2 = 0;
> > >         }
> > >      else
> > >         LOG_ERROR;
> > > @@ -488,8 +489,6 @@
> > >   else
> > >      esyslog("ERROR: can't open DVB device %d", n);
> > >
> > > -  aPid1 = aPid2 = 0;
> > > -
> > >   StartSectionHandler();
> > > }
> Klaus,
>
>     I have been looking at this change for a while as well.  For some 
> reason, valgrind does not like it the original way.  By changing it, 
> valgrind no long complains.  I was just comparing sources today and 
> thought I would mention this change to see what you thought.  To be 
> honest, I am not sure if it is necessary.
Now I remember why valgrind was complaining.. This has to deal with 
vdr-1.3.12 and the AC3overDVB patch.  If you look at this small snip of 
the AC3overDVB code, the following "SetupPES()" function is inserted and 
wants the "aPid1 = aPid2 = 0" before it.

@@ -482,6 +484,8 @@
ciHandler = cCiHandler::CreateCiHandler(DvbName(DEV_DVB_CA, n));
dvbTuner = new cDvbTuner(fd_frontend, CardIndex(), frontendType, ciHandler);
   aPid1 = aPid2 = 0;
+        // setup the PES analyser
+        SetupPES();
         }
      else
         LOG_ERROR;