Tvheadend

I forgot to write, that vdr with sc plugin works fine. (And my dreambox work with mgcamd works fine.)
So, this is not bug oscam server.

Iz it possible take code for irdeto to tvheadend from sc for vdr?

Iam not programmer, so i not understand howto rewrite it

how it shows in oscam:

2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26557: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26558: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26559: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26560: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26561: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26562: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26563: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26564: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26565: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26566: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26567: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26568: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26569: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26570: skipped (0 ms) by via
2012/01/04 21:39:04 931C318 r login emmtype=global, len=180, idx=61, cnt=26571: skipped (0 ms) by via

The same is on server with irdeto.

cnt=26571 - counter. It grow too fast And oscam loads cpu 99%.

What i do:
take tvheadend latest.
Setup one transporder, setup viasat. And via sharing.
Enable AU - emm spam begin.

After i setup tvheadend for one transporder raduga. Viasat deleted.
Sharing via deleted, raduga setup.
Enable AU - emm spam begin.

New information:

I tested latest tvheadend on old system : FC14, kernel 2.6.35.14-106.fc14, 32bit. (PIII 1.13 Ghz)

All work fine EMM works fine!

My system is: FC16, kernel 3.1.6-1.fc16.x86_64 (64 bit, AMD athlon 64 dual core)

How i can test, that there is broblem only in "64 bit" system? Install the same system, FC16, but 32bit?

Ubuntu 10.04.3 LTS

2.6.35-31-generic-pae

32bit system

The same.
AU brocken.

Same here. TVheadend floods the server with global EMM periodically with very short intervals until the server dies (overload) if AU is enabled. Other clients (vdr-sc) makes this job very well.

System is 2.6.32-29-generic #58-Ubuntu SMP Fri Feb 11 19:00:09 UTC 2011 i686 GNU/Linux
Ubuntu 10.04.2 LTS - XBMCLive Dharma

updated with opdenkamp-pvr (~pulse-4) and latest tvheadend within this package (even ~pulse-4)

Same here too. Using the codewordclient to connect to oscam results in an emm flood. EMMs also get sent when not watching any channel! When i use dvbapi to connect tvheadend to oscam EMMs are fine. Also then they get only send to oscam when watching a channel.
Im using latest tvheadend on suse 12.1 64bit.

oo pepe wrote:

When i use dvbapi to connect tvheadend to oscam EMMs are fine. Also then they get only send to oscam when watching a channel.
Im using latest tvheadend on suse 12.1 64bit.

This don't work for me. I've setup oscam on the same machine as TVHeadend and use oscam as a proxy to the cardserver. TVHeadend is connectet via dvbapi. Same flood of EMMs if AU is enabled.

My connection chain:

Cardserver <-> Oscam server <--newcamd--> Oscam Client <-> dvbapi <-> TVHeadend (codewordclient)

I think i fixed it for my card. I changed cwc_emm_nds(cwc_t *cwc, uint8_t *data, int len) in cwc.c in like this:

@
void
cwc_emm_nds(cwc_t *cwc, uint8_t *data, int len) {
int match = 0;
int i;
int serial_count = ((data³ >> 4) & 3) + 1;
unsigned char emmtype = (data³ & 0xC0) >> 6;

if ((data[0] & 0xFF) != 0x82) {
      return;
  }
  if (emmtype  1 || emmtype  2) {  // unique|shared
    for (i = 0; i < serial_count; i++) {
      if (memcmp(&data[i * 4 + 4], &cwc->cwc_ua[4], 5 - emmtype) == 0) {
        match = 1;
        break;
      }
    }
  }
  else if (emmtype == 0) {  // global
    match = 1;
  }

if (match)
    cwc_send_msg(cwc, data, len, 0, 1);
}
@

I got this from videoguard_get_emm_filter in oscam. Maybe you can adapt this to your card systems.
Now i get way less emms. Im not 100% sure if this is correct, but it seems to work for me.

oo pepe wrote:

I think i fixed it for my card. I changed cwc_emm_nds(cwc_t *cwc, uint8_t *data, int len) in cwc.c in like this:

I think you've insert this in cwc_emm_nds

  if ((data[0] & 0xFF) != 0x82) {
      return;
  }

I got this from videoguard_get_emm_filter in oscam. Maybe you can adapt this to your card systems.
Now i get way less emms. Im not 100% sure if this is correct, but it seems to work for me.

I've patched this and recompiled tvheadend. But this doesn't work for me. Same behaviour. What means the bitmask 0x82?

What card systems do you use? This is only for nds cards. Maybe i can apply it for other systems, too.

oo pepe wrote:

What card systems do you use? This is only for nds cards. Maybe i can apply it for other systems, too.

Irdeto tunneled Nagra (Sky Germany)