Andreas U. Trottmann wrote: >On Tue, Nov 22, 2005 at 06:03:17PM +0100, Sigmund Augdal Helberg wrote: > > > >>>What makes the descrambling stop is when your Session Key/Control Word >>>changes. This control word is present in the ECM, and is encrypted and >>>the service keys present in the EMM. Now when these keys do change, you >>>will not see a difference in the CA_PMT message. >>> >>>This change need not always be accompanied by a change in the CA_PMT >>>message, as the message itself is not composed of EMM's and ECM's. >>> >>>So the solution is to send the message continuously, in a timely manner. >>>say for example once in 5 minutes, to the module. >>> >>> > > > >>I've tried various approaces to this: >>1: send a capmt update each time the pmt changes. This is not nearly >>often enough. >>2: send a capmt update each time a pmt is received. Same as 1 >>3: send a capmt update every 5 minutes as suggested. This seem to give >>breaks of up to 5 minutes where the stream is not decrypted >>4: send a capmt very often (every 10 second). Even this does not give >>good results. It seem that when the decryption stops it can take up to >>20 sec before it start again. >> >>Any further suggestions is appreciated. >> >> > >I've tried to do this, too - indeed, "sometimes" sending CA_PMT >does "not work" as well, and it has to wait for the next message. > >But, it appears that using the non-CVS drivers (plain Linux >2.6.14.2) gives significantly better results. Of course it is very >difficult to "prove" that a problem that "sometimes" happens just >went away, but I now have watched about 60 "channel-hours" with >2.6.14.2 drivers without any problems (i.e. without the >descrambling suddenly stopping), while just loading CVS driver >modules makes the problem appear within minutes. > > One more thing, can you please test the RESET patch that i posted a few days back .. I would surely be interested in the results. If that makes a significant difference, i will check it in into CVS If only people can give some rigorous testing on the patches, we can make things straight. The more testing it gets the better, we can be sure of it, and on wider range of hardware too Thanks, Manu
