The commit fixes simultaneous channel decryption.
We are enumerating services based on the generated seq value.
In fact it should be elementary stream PID, but currently
we can leave it until PMT code rework.
The CA_SET_PID and CA_SET_DESCR from OSCam was always set with
seq=0 leading to setting all keys always to first subscribed
service.
capmt: add simplified OSCam dvbapi protocol as new capmt mode
The new mode uses one domain socket connection for all subscribed
services. The CA_SET_PID and CA_SET_DESCR are received over same
socket connection instead of one UDP socket per adapter.
The minimum OSCam version for use with this mode is svn rev 9063.
Dave Chapman [Thu, 28 Nov 2013 00:50:39 +0000 (00:50 +0000)]
Merge relevant parts of PR#255 to complete the implementation of PAT/PMT passthrough, with the option of rewriting the PAT and PMT so the PAT only contains a reference to the included service, and the PMT only contains references to the included streams. This still needs more work - the PMT PID is being passed to the muxer as zero the first time a channel is streamed after tvh starts, and the PMT rewrite code fails if the PMT spans multiple TS packets
Adam Sutton [Sun, 24 Nov 2013 20:32:15 +0000 (20:32 +0000)]
iptv: added new configuration to allow subs and bandwidth limitations
The bandwidth limit is a soft one, it will stop any new subscriptions if
the peak bandwidth on the current services exceeds the limit. Possibly
this will need a bit of additional averaging in future to stop false
triggers.
Adam Sutton [Mon, 18 Nov 2013 09:21:35 +0000 (09:21 +0000)]
linuxdvb: old code would update tuning params even without auto-discovery
This is important for correcting mistakes in the pre-defined configuration
files, and also picking up potential changes (not sure that will quite work).
The one thing that is left untouched is the frequency, a) to not confuse people
and b) I've seen 2 transponders sending slightly diff freq and this can cause
the config to be constantly updated.
The one thing this will also do, for better or worse, is remove an AUTO
settings. Cons are it could be less flexible an tuners may automatically deal
with small changes in settings, Pro's are if you have some tuners that accept
AUTO and some that don't, once its scanned on something that does it will get
updated. Plus its probably better, where possible, to have the full settings.
This was done for a reason, but unfortunately it wasn't really done properly
and causes some unintended side-effects. The basic principle still needs
some looking at, but properly handled differently.
Adam Sutton [Mon, 11 Nov 2013 17:48:11 +0000 (17:48 +0000)]
webui: some changes to default display for UI
This is based on feedback from linuxstb, I'm not 100% convinced about all
aspects. But will throw it out there and see what others think of the change.
Adam Sutton [Mon, 11 Nov 2013 17:41:05 +0000 (17:41 +0000)]
status: remove HTTP status info for now
This breaks if, like me, you're using a proxy (or browser) which doesn't
support persitent connections, since you just get a constant reloading
of the UI! Will have to re-think, since ultimately what I really want is
to know user auth's within the UI, so possibly that will be easier when
we do away with HTTP Basic Auth, but also probably require restructuring
of this code.
Adam Sutton [Mon, 11 Nov 2013 09:52:29 +0000 (09:52 +0000)]
status: collect (and post) status of all TCP connections
This will indiscrimately include all HTSP and HTTP connections, first pass
I was going to just dump the lot on a UI status tab. However it could be
some filtering might be useful.
Adam Sutton [Sat, 9 Nov 2013 13:15:21 +0000 (13:15 +0000)]
capmt linuxdvb: some minor tweaks to ensure only necessary are enabled
Only adapters that have active frontends will be enabled in capmt. In
addition to this I've also fixed code so it will actually honour the port
configuration (which was being ignored in OSCAM mode).
Adam Sutton [Sat, 2 Nov 2013 21:05:09 +0000 (21:05 +0000)]
linuxdvb: various fixes and simplifications to the satconfig
it's now possible to change hte number of satconfs in advanced mode,
however at the moment the only way to delete something is by reducing the
number of orbital positions and this will always delete the last element.
Not ideal and I'm sure people will complain, but i can't be assed to add
the delete button (at the moment)!
Adam Sutton [Thu, 17 Oct 2013 20:39:44 +0000 (21:39 +0100)]
linuxdvb: started updating to the new hardware API
I actually had this basically working, but then barrelled on with changing
satconf further, without committing anything (got back into old SVN bad habits)
so what's here is a bit of a half finsihed mess
Adam Sutton [Mon, 21 Oct 2013 12:33:02 +0000 (13:33 +0100)]
mpegts: fix processing errors when <188 bytes is received
Also fix raw streaming where data is not %188, previously the code would have
sent duplicate bytes which would have caused (unecessary) errors downstream.
Adam Sutton [Sat, 12 Oct 2013 11:26:22 +0000 (12:26 +0100)]
webui: stop multi select combo selecting the wrong thing
this was because it reverted to using the display string and doing
a reverse mapping. Which is clearly the wrong way to do things when you
can have multiple display fields for the same key value.
Adam Sutton [Fri, 11 Oct 2013 20:50:30 +0000 (21:50 +0100)]
htsp: add delayed unsubscription in response to HTSP request
The reason for this is there isn't a command for "changing" channel, this may
be changed in the future. So to simulate it the subscription weight is change
to a minimum level and then a new one can be created, thus reducing the need
to stop the tuner (especially for in mux zaps).
User Decke [Mon, 3 Jun 2013 09:36:47 +0000 (11:36 +0200)]
tvhpoll: various fixes for tvhpoll on FreeBSD
- kqueue filters are not bitmasks so we need to treat them a bit different
- Properly check timeout because default value is -1
- Use correct variable kevent.filter to detect which event was triggered
- Fix building on FreeBSD i386
Adam Sutton [Fri, 11 Oct 2013 12:41:46 +0000 (13:41 +0100)]
misc: Remove memory leaks and other anomalies
Many of these are somewhat redundant now, as I've suppressed many of the
warnings as they're false positives. However the changes, such as added
a detach flag to tvhthread_create(), have been kept anyway.