oscam (1.20-11439~r11209-vivid) vivid; urgency=medium
* [r11209]
Fix setting client_proto_version for the connection
The client_proto_version was lost when the PMT data was not send in the very first data packet processed.
The trace for the problem:
- new socket connection fd: 14
- dvbapi_handlesockdata, client_proto_ver=0, pfd2[i]: i=0
- Got DVBAPI_CLIENT_INFO packet with size 42
- Client connected: 'vdr-plugin-dvbapi 2.2.1 / VDR 2.0.0' (protocol version = 2)
- dvbapi_handlesockdata AFTER, client_proto_ver=2, pfd2[i]: i=0
Now the connection is added to unassoc_fd, because there was only a greeting and no caPMT.
Then the PMT update is on other pfd2 index (i=1) and the client_proto was set previously on index i=0 !
- dvbapi_handlesockdata, client_proto_ver=0, pfd2[i]: i=1
And OSCam was using wrong protocol leading to problems with communication.
The commit uses the last_client_proto in this case to correctly fill the client_proto_version[] table with correct index.
* [r11208]
- More fix for re-enabled provid matching!
* [r11207]
DVBAPI:
- In CSA mode extended_cw_api should be always 0 regardless what user selected!
* [r11206]
- Reworked the re-enabled provid check in r11204 (Hope working setups dont need to be adopted by the thighter provid check!)
* [r11205]
DVBAPI:
- Fetching realpmt and ecmpid provid was fixed by an ecm request it lead to an additional ecmpid -> fixed!
- Optimized the ecmpid / streampid addition code
* [r11204]
- Re-enable provid check. Some cards report it wrong may not be the reason to disable the check complete.
If a card reports provid wrong it should be replaced with provid 000000
* [r11203]
- Tryfix streampid issues dvbnetapi() (fatal trigger by r11202, but issue was already present for long time!)
* [r11202]
DVBAPI:
- Fixed a streampid bug in case decoding was temporarily stalled (introduced r11190)
- On first streampid enable always disable the streampid on ca device with index -1 (makes 100% sure the index we use is used for decoding!)
- Log cosmetics: dont hide "stop descrambling program number..." (user doesnt have to guess why is my channel not descrambled if its done on purpose!)
* [r11201]
WEBIF:
- Fix for ticket #4388 not cleaning notfound stats using webif->config->loadbalancer->Clear Not Found button
* [r11200]
VIACCESS:
- Fix for broken Nano 0E support regression introduced r11199 (Tnx Pr2 for patch!)
* [r11199]
VIACCESS:
- Improved nano E0 logging: only log if no initCA28 (tnx Pr2 for patch!)
* [r11198]
VIACCESS:
- Removed misleading failed logwarning if user isnt using ca28 init (there are cards that even dont need it anyway and work fine without it!)
* [r11197]
DVBAPI:
- Updating errno isnt mandatory in case a systemcall was successfull!
* [r11196]
DVBAPI:
- revert 1 line too much -> corrected!
* [r11195]
DVBAPI:
- Revert recent changes that might be the cause of fd's that dont get closed on some systems!
* [r11194]
DVBAPI:
- Tryfix some setups suffered from errno=24 Too many open files after some zapping since r11181
* [r11193]
DVBAPI:
- Tryfix STAPI no matching S: line in oscam.dvbapi for pmtfile
* [r11192]
DVBAPI:
- Set PAT & PMT filter only on new caPMT list items, not on all (Tnx Schwa266 for providing patch!)
- Only start SDT filter on new caPMT list items too!
* [r11191]
DVBAPI:
= Tryfix VU+ streampid sensitivity reintroduced in r11168: dont disable streampids on ca device that are not used for decoding!
* [r11190]
DVBAPI:
- Experimental: on cw not found do not directly disable stream on ca since the already written cw is still valid for some time and we have time left to get the fresh cw before old cw is expired!
- Experimental: only channel stop descrambling disables all streams used by this demuxer on the ca device
- Tryfix: Seems an solved bug in r11064 was reintroduced in r11168, modded the involved code so it should work same as in r11064
* [r11189]
small cosmetic fix to hide inactive clients
* [r11188]
fix stapi5 streampids >16
found by hoho
* [r11187]
- Addition: ...Forgot the comment in code for future updates!
* [r11186]
- Proper TommyDS fix for buggy toolchain used for fritz 73xx and 74xx
And made a remark in code to prevent it to happen with future updates of TommyDS!
* [r11185]
forget another point to fix thanks AbrahaM...
* [r11184]
fix build error on fritz 73xx und fritz 74xx
* [r11183]
- (small) update tommyDS lib from https://github.com/amadvance/tommyds/
* [r11182]
fix compile warning
* [r11181]
poll fix patch for TVheadend
patch by perexg
http://www.streamboard.tv/oscam/ticket/4373
* [r11180]
fix proc reading
patch by plnic
http://www.streamboard.tv/oscam/ticket/4372
* [r11179]
fix #4366
patch by hook
* [r11178]
DVBAPI: (see ticket #4369)
- fix unhandled_buf migration from temporary fd range to demuxer
* [r11177]
DVBAPI: (see ticket #4370)
- The unassoc_fd _MUST NOT_ be unset outside the poll loop for the network protocol (and pcnodmx mode)
* [r11176]
DVBAPI:
- Preserve client version for unassoc_fd[] (see ticket #4367)
* [r11175]
DVBAPI:
- Register adapterindex early (ticket #4368)
* [r11174]
DVBAPI:
- Tryfix ticket #4365
* [r11173]
ca_mask should have 32 bits
* [r11172]
fix activeindexers cast
* [r11171]
log stream_type before it gets patched
* [r11170]
raise INDEX_MAX_LOCAL to 15, at least the av7110 supports it
* [r11169]
store client_proto_version for each connected dvbapi client
* [r11168]
fixes for dvbapi pid index handling
dvbapi via tcp now supports up to 64 indexes (0 to 63)
dvbapi non-tcp now supports up to 8 indexes (0 to 7)
fixed possible 100% cpu freeze bug;
fixed index disable bug on stapi
* [r11167]
fix typo
* [r11166]
revert last commit
* [r11165]
rewrote dvbapi pid index handling
dvbapi via tcp now supports up to 0xDFFFFFFF indexes;
dvbapi non-tcp now supports up to 8 indexes (can now be raised to any value <= 0xDFFFFFFF when needed);
fixed possible 100% cpu freeze bug;
* [r11164]
NAGRA:
- Additional EMM filtering for emms starting with 0x84 and 0x87 as used by e.g. MaxTV
Tnx goes to Pr2 for delivering initial patch, I merely reworked it a bit so that not 3 additional filters will be setup!
* [r11163]
VIACCESS:
- Dont return error on softfail 9040
* [r11162]
CCcam:
- Dont set both last send and get on keepalive, just set what is actually done (e.g. time of last send action or time of last get action)
* [r11161]
CCCAM:
- Add keepalive response from CCcam server to keepalive request from cccam client
* [r11160]
DVBAPI:
- Dont try to close and flush filter fd if stop filter already returned bad fd error
* [r11159]
fix oscam spam the dvbapi log with "PMT Update on socket.."
on dvbapi client disconnect
patch by schwa226
http://www.streamboard.tv/oscam/ticket/4349#comment:4
* [r11158]
revert last commit
* [r11157]
tryfix2: zap takes too long
* [r11156]
tryfix: zapping takes too long
* [r11155]
DVBAPI:
- Keep trying to close a filter that reported bad fd is useless -> skip!
* [r11154]
DVBAPI:
- Tryfix ticket #4353 (DVBAPI_AOT_CA_STOP command)
* [r11153]
fix status_loghistory height
* [r11152]
added missing file
* [r11151]
make number of displayed log lines in webif adjustable
* [r11150]
make sure we do not loop too long on on socket
* [r11149]
re-wrote packet receiving code.
problems are expected, please keep calm and open a ticket or post in the forum when you encounter one.
* [r11148]
- Another tryfix for cardreaders that cant handle nagra layer switching!
* [r11147]
DVBAPI:
- Why should oscam convert subtitle type streampids to audio type?
* [r11146]
reduce livelog cpu load
* [r11145]
VIACCESS:
- A9 Nano: not all classes need to be already present in card entitlement -> disabled all classes must match check
- Dont throw error in log for softfail 9040 (card gives this back if all entitlements were already up to date)
* [r11144]
CCcam:
- Attempt to get keepalive, read timeout and inactivity timeout working like expected
Keepalive only works if both sides (client and server) have it enabled
Its now possible to use -1 for inactivity timeout too.
* [r11143]
block commercial libOscam (BOXTYPE_PC_NODMX) for Samsung TVs (the free one with BOXTYPE_SAMYGO still works)
* [r11142]
DVBAPI:
- Do not reset ecmpidcount on fetching realpmt as it will invalidate values like status, tries, checked and influences the running decoding!
* [r11141]
DVBAPI:
- Add missing connfd that is needed to start filters on dvbnetapi
* [r11140]
undo last fix, should be fine on samygo
* [r11139]
DVBAPI:
- Yet another quick & dirty workaround for "Samygo" messing up the dvbapi code.
* [r11138]
DVBAPI:
- Some fixes in communication handler
* [r11137]
DVBAPI:
- Tryfix realpmt doesnt add fake ecmpid (reported here: http://www.streamboard.tv/wbb2/thread.php?postid=554891#post554891 )
* [r11136]
DVBAPI:
- Tryfix ticket #4349 (logspam dvbapi networkmode)
* [r11135]
DVBAPI:
- Fix for ticket #4350 (PAT/PMT could not be started on STAPI devices)
- Some logcosmetics
* [r11134]
DVBAPI:
- Improved the correct srvid pmt fetching using section filtering: let the box do the heavy lifting!
* [r11133]
fix: one pmt stream may be used for multiple channels -> verify srvid
* [r11132]
DVBAPI:
- Output SDT/PAT/CAPMT filterdata it might be handy for debugging purposes!
* [r11131]
revert last change
* [r11130]
tryfix: problems on some channels after fetching real PMT
* [r11129]
default to audio for stream_type 0x06 if no descriptors are available; this breaks when other datastreams come in the way, but better than nothing
* [r11128]
directly filter pmt if pmtpid available
some log cosmetics
* [r11127]
second fix from #4347
* [r11126]
applied fix from #4347
* [r11125]
revert 11122; breaks reading on some stbs
* [r11124]
better handling of invalid pmts
* [r11123]
remove bit-rate info
* [r11122]
do not spam cs_recv() if there is no data
* [r11121]
fixed typo
* [r11120]
updated streamtxt
* [r11119]
fix log output for stream_type
* [r11118]
tryfix: switching back to csa on OE2.2
* [r11117]
fix possible warning
* [r11116]
fix possible endian problem
* [r11115]
tryfix for encrypted PMTS; however we still need a capmt with valid caid descriptor, so this most likely will not be enough
* [r11114]
if we cannot detect the command, skip one byte, or we will be stuck
* [r11113]
fix build
* [r11112]
fix: connection dropped if no incoming data (revert)
fix: do not try to read capmt after killing the connection
* [r11111]
use real PMT if it can be filtered
* [r11110]
re-wrote loghistory
loghistorylines now sets the amount of log history lines (the old loghistorysize has been removed to avoid confusion)
skipping/invisible log lines (especially when using loglevel > 0) should now be fixed
* [r11109]
fixes for extended cw mode
* [r11108]
- Tryfix VDR http://www.streamboard.tv/wbb2/thread.php?postid=554649#post554649
* [r11107]
DVBAPI:
- Fix for openxcas not using prio / non-prio run
-- Andrey Pavlenko <email address hidden> Thu, 21 Jan 2016 00:17:54 +0300