In order for the 7025 to start displaying video/audio it appears (through dmesg) that the ioctl that calls '[AUDIO] enable sync' needs to be sent to xilleon.ko.

However I have been having a few problems:
1. The /dev/dvb/adapter0/audio0 device is opened & locked by enigma2. Using fcntl to change the lock doesn't seem to help as enigma2 reasserts the lock almost immediately.
2. From the enigma2 source, the only ioctl's called on the audio0 device, are:
- AUDIO_STOP
- AUDIO_SELECT_SOURCE
- AUDIO_PLAY
Logically I expected AUDIO_SET_AV_SYNC, but that is never called in the source.

If someone could please help with any of the following it woudl be greatly appreciated:
- What ioctl (if an ioctl is necessary) is called to cause the enable sync?
- Can enigma2 be forced to cause the sync somehow? Through what device?
- If the select source, or set av sync is needed to be called on the audio device, how can that device be opened with the enigma2 lock in place?

Thanks for the reply

Opening the demux on the correct pid should be:

struct dmx_sct_filter_params dmxp;
memset(&dmxp, 0, sizeof(dmxp));
dmxp.pid = e_pid;
dmxp.filter.filter[0] = 0x80;
dmxp.filter.mask[0] = 0xf0;
dmxp.timeout = 1000;
dmxp.flags = DMX_IMMEDIATE_START;
ioctl(fd, DMX_SET_FILTER, dmxp);

Followed by:

ioctl(fd, DMX_START);

If the CA device is used then it needs to have the pid set as well using

ca_pid_t ca_pid;
ca_pid.pid = c_pid;
ca_pid.index = 0;
ioctl(ca_fd, CA_SET_PID, &ca_pid);

If the audio device hasn't been started, how is it possible to start it if enigma2 has /dev/dvb/adapter0/audio0 open and locked. fcntl doesn't work as enigma almost immediately takes control again.

I figured out my problem

You need to do:

ca_pid_t ca_pid;
ca_pid.pid = c_pid;
ca_pid.index = 0;
ioctl(ca_fd, CA_SET_PID, &ca_pid);

ca_pid_t va_pid;
va_pid.pid = v_pid;
va_pid.index = 0;
ioctl(ca_fd, CA_SET_PID, &va_pid); //Video PID

ca_pid_t aa_pid;
aa_pid.pid = a_pid;
aa_pid.index = 0;
ioctl(ca_fd, CA_SET_PID, &aa_pid); //Audio PID

tmbinc was right, I did need to set the correct PID, I did not know that the c_pid, v_pid and a_pid (all audio pids, if there is more than one stream) all had to SET_PID in the ca device.