oscam (1.20-10584~r10356-utopic) utopic; urgency=medium
* [r10356]
Webif: (Tnx Ultra47 and oscam community members for their effort!)
- Fix for http://www.streamboard.tv/oscam/ticket/4064
- Patch r9947 enhanced
- Fix show written emms if emm logdir empty
- Optical fixes show written emms
- Show written emms on write emm tab sorted and filtred (oldest first, newest last), idea by schorsch
* [r10355]
stat: Remove useless check.
Checking for 'ea' just tricks static checkers that 'ea' can be NULL
which it can't be.
* [r10354]
smartreader: Prevent NULL pointer derefence.
* [r10353]
Correctly restore errno after nanosleep().
* [r10352]
[gbx] Obey reshare parameter of cccam also in gbx network.
Check for ECMs taking too many hops.
* [r10351]
fix for #4043
* [r10350]
oscam-config: Fix variable types.
If you have set `lb_force_fallback` in your config, because of wrong
variable type in oscam-config-reader.c initializing this variable overwrite
others after it.
This may lead to crashes and unexpected behaviour.
* [r10349]
devtools: Improve messages in check_config_tables.sh
* [r10348]
devtools: Improve config table checking script.
* [r10347]
Prevent "too many open files" issue on OS X.
Patch by powercore posted in ticket #4008.
* [r10346]
log: Third try to fix anticasc log.
This is a fix for ticket #4092.
* [r10345]
cccam: Move cacheex code into separate file.
This cleans up module-cccam.c (which is already very big) and
the ability to compare camd35 and cccam cacheex implementations.
They have a lot of common.
Another nice thing is the removal of all CS_CACHEEX ifdefs in
cccam module.
* [r10344]
cacheex: Remove useless ifdefs.
The whole file is guarded by CS_CACHEEX ifdef, there is no
need for these extra ifdefs.
* [r10343]
log: Fix for anticasc log.
Additional fix for ticket #4092.
* [r10342]
build: Do not warn about using openssl on OS X.
* [r10341]
log: Add missing new line in log lines.
This fixes ticket #4092.
* [r10340]
camd35: Move cacheex related code in module-camd35-cacheex.{c,h}
Over 45% of the code in camd35 was dedicated to cacheex. So much
code deserves it's own file. This cleans camd35 a lot and makes
it ifdef free and a lot easier to follow.
* [r10339]
camd35: Add proper prefix to all functions that didn't have one.
* All camd35 funcs now have camd35_ prefix.
* All camd35 funcs that deal with cacheex now have camd35_cacheex_ prefix.
* [r10338]
camd35: Mark funcs static where needed.
These functions are used only in this module.
* [r10337]
camd35: Prepare cacheex code for moving into separate file.
This commit is just preparation for moving camd35 cacheex code
into separate file.
* [r10336]
camd35: Prepare cacheex code for moving into separate file.
This commit is just preparation for moving camd35 cacheex code
into separate file.
* [r10335]
camd35: Prepare cacheex code for moving into separate file.
This commit is just preparation for moving camd35 cacheex code
into separate file.
* [r10334]
camd35: Prepare cacheex code for moving into separate file.
This commit is just preparation for moving camd35 cacheex code
into separate file.
* [r10333]
camd35: Prepare cacheex code for moving into separate file.
This commit is just preparation for moving camd35 cacheex code
into separate file.
* [r10332]
camd35: Prepare cacheex code for moving into separate file.
This commit is just preparation for moving camd35 cacheex code
into separate file.
* [r10331]
camd35: Remove unneeded CS_CACHEEX ifdefs.
This code is already covered by ifdef CS_CACHEX, no need for
second check inside this ifdef.
* [r10330]
dvbapi experimental: auto filter limiting
Some users experience "ERROR: read error on fd xx"
I guess this due to the fact too much filters (presumable emm type) are started and the box is killing random filters started before.
Even worse, soms boxes accept it all without returning to us no more filters we reached maximum.
On attempt to read the data from a certain filter the "ERROR: read error on fd xx" is returned.
Now this filter is killed and the max ammount filters is lowered by 1.
Once oscam is running for some time the max ammount filters will be settled on a value that works for your box.
Feedback you can give here: http://www.streamboard.tv/wbb2/thread.php?postid=534445#post534445
* [r10329]
WebIf: Right maxlegth for input (set Group) in user_edit.html
Thx uf2k.
* [r10328]
- tryfix: moved the cs_sleepms() to a (hopefully) better place in the handler code
* [r10327]
build: Fix smartreader compilation on OS X.
* [r10326]
webif/pages_gen: Do not leak fds in readfile().
readfile() was not closing the file it opened. If your system
have lowered file open limits (like < 256) pages_gen failed
because there were too many open files.
* [r10325]
Reduce CS_CACHEEX ifdefs.
* Structure definitions do not need to be guarded.
* Funcs that are not used would be dropped by the compiler.
* Func declarations do not need to be guarded.
No binary changes.
* [r10324]
webif: Reduce WITH_LB ifdefs.
We can use config_enabled() instead of ifdef in these cases.
* [r10323]
config-reader: Reduce WITH_LB ifdefs.
Just code movement, no functional changes.
* [r10322]
config-account: Remove duplicated function.
The function already exists in oscam-config-global.c.
No need to copy/paste it.
* [r10321]
camd35: Remove WITH_LB ifdefs.
We can use config_enabled() instead of ifdef in this case.
* [r10320]
chk: Remove WITH_LB ifdefs.
Just code movement, no functional changes.
* [r10319]
reader: Remove WITH_LB ifdefs.
Just code movement, no functional changes.
* [r10318]
ecm: Reduce WITH_LB ifdefs.
Just code movement, no functional changes.
* [r10317]
ecm: Reduce WITH_LB ifdefs.
Just code movement, no functional changes.
* [r10316]
ecm/dvbapi: Reduce WITH_LB ifdefs.
Just code movement, no functional changes.
* [r10315]
ecm: Reduce WITH_LB ifdefs.
Just code movement, no functional changes.
* [r10314]
ecm: Remove dead code guarded by WITH_LB ifdef.
The removed code boils down to:
{{{
if (!provid) {
loop {
if (some_check || some_other_check)
provid = 0;
break;
}
}
}}}
which is obviously bogus since provid is already 0.
* [r10313]
Revert "gbox: Do not use popen() to write into files."
This reverts commit r10286. It turns out that gbox really wanted
to execute shell commands. See commit r9723.
* [r10312]
gbox: Add new lines at the of OSD files.
Without new lines r10286 probably broke the osd functionality.
* [r10311]
pcsc: Do not include PCSC/reader.h on OS X.
This header do not exist on OS X's pcsc.
* [r10310]
build: Check USE_ flags when card reader config is being listed.
This allows the build to display SMART and PCSC readers, when they
are enabled.
* [r10309]
build: Show what config file is being used.
* [r10308]
[gbx] One more it found.
* [r10307]
[gbx] Small fix for r10305.
* [r10306]
- Fix compile warning introduced in r10305
* [r10305]
[gbx] Fix total cards in hello display. Fight against spaghetti code.
* [r10304]
[gbx] Consolidate MSG_HELLO1 handling.
* [r10303]
[gbx] Collapse first hello handling. Do not malloc hostname on every new hello.
* [r10302]
build: When getting include PATHs do not try to compile.
This speeds up include path detection nearly three times (not that
it was slow but anyway).
* [r10301]
build: Allow PCSC headers to be installed in CROSS/.../local/include/PCSC
Without this oscam would try to include PCSC headers from host's
/usr/local/include/PCSC instead of the headers in the toolchain.
* [r10300]
build: Fix cross compilation of PCSC.
We use the compiler's last include PATHs and add PCSC to it.
Basicaly it is /usr/include but in case of cross compilation it
might be something else.
Since using -Iinc_path instructs the compiler to use inc_path
(without add the toolchain system root) we need to have this hack
to get the "real" last include path. Why we needs this?
Well, the PCSC headers are broken and rely on having the directory
that they are installed it to be in the include PATH.
We can't just use -I/usr/include/PCSC because it won't work in
case of cross compilation.
* [r10299]
build: A simpler way of fixing PCSC header idiotism.
Just playing with include paths is enough to work around
broken PCSC headers which are installed in include/PCSC/...
but include themselves without PCSC/
* [r10298]
README.build: Update.
* [r10297]
build: Remove TARGET_SYSROOT.
gcc option -print-sysroot is not always supported. For example
on older toolchains and for toolchains compiled without --with-sysroot.
Also this variable is not used.
* [r10296]
build: Allow PCSC to be built with other than shipped headers in extapi.
If you do not want to use the PCSC headers shipped in extapi,
you may build oscam by specifiyng another include directory.
{{{
make USE_PCSC=1 DEFAULT_PCSC_FLAGS=-I/usr/local/include
}}}
* [r10295]
[gbx] more log cosmetics
* [r10294]
README.build: Update.
* [r10293]
build: Clean pcsc build variables for OS X.
* [r10292]
build: Use local pcsc headers for compilation.
* [r10291]
build: Use pcsc headers in extapi.
* [r10290]
extapi: Add PCSC headers.
PCSC headers as shipped by pcsclite project are buggy. They are
installed in /usr/{local/}include/PCSC but they use #include <wintypes.h>
and #include <pcsclite.h> (Note the missing PCSC/ directory).
The result is that the compilation fails if the compiler is
not instructed to add to its include path /usr/include and/or
/usr/include/local.
Adding these paths comes with it's own set of problems for
cross compilation.
So the "cleanest" thing is to adopt pscs headers in extapi.
Currently extapi/PCSC contains all headers that are needed
for PCSC compilation. The only difference between them and
the ones shipped by pcsclite is that include statements
were fixed to correctly include PCSC/...
Even if you don't have pcsc installed `make pcsc` would
compile oscam, just the final link phase would fail complaining
about missing libpcsclite.
* [r10289]
Make file PCSC
better then hack in 10288
* [r10288]
build: Fix PCSC cross compilation.
This hack is needed because PCSC includes <wintypes.h> instead
of <PCSC/wintypes.h>.
It is ugly but it'll work.
* [r10287]
Revert "build: Remove forced include path in pcsc build."
This reverts commit r10285.
PCSC ships with broken headers which include <wintypes.h> when
they should include <PCSC/wintypes.h>
It worked for me because I have copied wintypes.h into /usr/local/include.
* [r10286]
gbox: Do not use popen() to write into files.
Using popen() to write into files is highly unusual. Let's do
this properly.
* [r10285]
build: Remove forced include path in pcsc build.
Adding /usr/include/PCSC in the include path is wrong in
case of cross compiling. We include pcsc headers using #include <PCSC/...>
which forces the compiler to use default include search path,
which in turn contains /usr/include and /usr/local/include.
So this should fix building pcsc in cross compilation
environment.
* [r10284]
dvbapi:
Add support for mapped A: lines
Example:
A:::10E0:0844::: 2600:000000:1FFF
Receiver wants to demux srvid 10E0 on adapter 0000 camask 0001 index 0000 pmtpid 0844
..
..
Added fake ecmpid 2600:000000:1fff for unencrypted stream on srvid 10E0
* [r10283]
log: Unify logging function names under cs_log prefix.
This commit renames logging functions. Now everything that
logs can be found by searching for cs_log.
Here is what is being done:
`
> Existing logging:
> -> cs_log
>
> Renamed:
> cs_debug_mask -> cs_log_dbg
> rdr_debug_mask -> rdr_log_dbg
> rdr_debug_mask_sensitive -> rdr_log_dbg_sensitive
>
> Existing logging that hexdumps buffer:
> -> cs_log_dump
> -> rdr_log_dump
>
> Renamed:
> cs_ddump_mask -> cs_log_dump_dbg
> rdr_ddump_mask -> rdr_log_dump_dbg
`
Here is the script that was used to make the changes.
{{{
FILES=$(find -name '*.c' -o -name '*.h')
perl -pi -e 's|cs_debug_mask|cs_log_dbg|g' $FILES
perl -pi -e 's|rdr_debug_mask|rdr_log_dbg|g' $FILES
perl -pi -e 's|rdr_debug_mask_sensitive|rdr_log_dbg_sensitive|g' $FILES
perl -pi -e 's|cs_ddump_mask|cs_log_dump_dbg|g' $FILES
perl -pi -e 's|rdr_ddump_mask|rdr_log_dump_dbg|g' $FILES
}}}
The old function names were defined as compatability but don't
use them in new code because these defines would be removed
sometimes in the future.
* [r10282]
log: Unify log headers.
We don't really need 3 types of log headers.
* [r10281]
log: Add log prefix to hex dumps.
* [r10280]
[gbx] Small addition to r10279
* [r10279]
[gbx] Harmonizing log messages after recent changes by gfto. Thx!
* [r10278]
- Revert experimental protection against boxes that seem to close filters without being instructed to do so by dvbapi
- On hypercube su980 always return ok on io_ctl + some code cosmetics in io_ctl to make it more readable
- Some recoding on dvbapi init (boxtype+api) since cs_log() needs these parameters set before otherwise crashing
* [r10277]
Internal readers non-pll (cardmhz < 2000):
Proper fix for regression introduced in r10267
* [r10276]
- Fix for segfaults due to changed logline styles (used vars must be inited before logging now!)
* [r10275]
- Remove unused var .Di from struct cardreadersettings
* [r10274]
- Somehow I managed to get oscam to use too much resources -> give it back!
* [r10273]
dvbapi:
- More rework on resorting ecmpids
- More Logcosmetics
* [r10272]
More dvbapi eyecandy: log how much time sorting the ecmpids takes!
* [r10271]
dvbapi:
Channelcache enhenced once more: deleting whole channelcache on oscam.dvbapi change removed.
Now exact channelcache hit is checked at end of ecmpid prio/ignore matching.
If a channelcache entry matches with highest prio -> start descrambling only that ecmpid otherwise ignore channelcache.
* [r10270]
- Same mixup D <> Di for smargo
* [r10269]
- Correction mixup of D and Di (old code was sending D but smartreader code accepted it as Di...)
* [r10268]
csctapi: Change Fi only on sci card readers.
This fixes regression introduced in r10267.
* [r10267]
csctapi: Merge three write_settings() callbacks.
This patch removes write_settings2 and write_settings3 callbacks
merging them into single write_settings() that receives struct
s_cardreader_settings parameter.
* [r10266]
serial: Don't try to be cute with logging format strings.
The compiler will save duplicated strings only once anyway, so there
is no need for such a trick.
* [r10265]
log: Rename cs_dump() and rdr_dump() to cs_log_dump() and rdr_log_dump().
Now searching for cs_log we'll get places where logging funcs are used.
* [r10264]
Use rdr_dump instead of cs_dump in card readers.
* [r10263]
gbox: Remove duplicated log prefixes.
* [r10262]
constcw: Remove duplicated log prefixes.
* [r10261]
Move logCWtoFile() func to oscam-ecm.c
It is the only place where it is used also the idea of oscam-log.c
is to hold the global logging code.
* [r10260]
log: Move anticasc logging code into module-anticasc.c
And clean an ifdef in module-log.c
* [r10259]
dvbapi/azbox/mca: Clean logging messages.
Now that we have module prefixes these additional prefixes are not needed.
* [r10258]
dvbapi: No need to handle obvious invalid fd!
* [r10257]
dvbapi:
- Small log cosmetic I missed
- Experimental protection against boxes that seem to close filters without being instructed to do so by dvbapi
* [r10256]
dvbapi log cosmetics
* [r10255]
log: Add MODULE_LOG_PREFIX to the files that need it.
* [r10254]
log: Prepare anticascading log for logging prefixes.
* [r10253]
log: Align module prefixes.
* [r10252]
log: Allow each source file to declare it's own log prefix.
Adding `#define MODULE_LOG_PREFIX "mymodule"` before including
globals.h would instruct logging functions to add this prefix
in each log message.
This functionality would allow a lot of cleanup in module logging
messages and it'll make it easier to sort (filter) messages by
module.
* [r10251]
csctapi: Use rdr_debug_mask() instead of cs_debug_mask().
* [r10250]
smartreader: Use rdr_log() instead of cs_log().
* [r10249]
csctapi: Use rdr_log() in azbox and sci card readers.
* [r10248]
readers: Use rdr_log() in cryptoworks and viaccess readers.
* [r10247]
dvbapi: Workaround for CA_SET_DESCR ioctl() on su980.
The kernel on su980 boxes is buggy returns error when CA_SET_DESCR is called.
For more info see: http://www.streamboard.tv/wbb2/thread.php?postid=533940
* [r10246]
Machine info
Extend boxtype info to all internall readers as harmless test. Or at least as check that it's harmless.
* [r10245]
Machine Info
cosmetic removed a not wanted space
* [r10244]
Machine info
added su980 boxtype
* [r10243]
oscam-log: Split cs_log_int().
Split cs_log_int() into separate functions for logging single lines
and logging with buffer dumping. This is made mainly to avoid having
unused parameters in a single function.
As far as the users of the logging API are concerned - the API is
unchanged.
bloatcheck reports 12077 bytes saved from x86_64 oscam binary and
14044 bytes saved on dm500 binary.
* [r10242]
oscam-log: Prepare cs_log_int() for splitting into separate functions.
This is just a preparation - code movement and whitespace changes.
No noticable functional changes.
* [r10241]
oscam-log: Remove lock parameter of cs_log_int().
The comment added in r10240 was wrong. liboscam_stapi.a calls cs_log()
function which is correctly emulated in module-dvbapi-stapi.c so
it is safe to remove lock parameter of cs_log_int().
According to bloatcheck this change saves 9006 bytes from the oscam
binary on x86_64.
* [r10240]
oscam-log: Clarify why we keep 'lock' parameter to cs_log_int().
* [r10239]
oscam-log: Remove XXX_nolock() macroses.
The `lock` parameter of cs_log_int() is even marked as unused
because it really is. There is no point in having XXX_nolock.
* [r10238]
oscam-log: Reorder some declarations to reduce ifdefs.
* [r10237]
Revert "Extra dump functions."
This reverts commit r10016.
Gratiuous copy+paste for no reason at all. These functions were not
even used. Not worth the extra 75 lines of code that does exactly
nothing (luckily the linker knows when a code is not used so these
functions never increased the binary size and after removing them
the binary is exactly the same).
* [r10236]
failban: Restore the behaviour broken by r10232.
* [r10235]
- Refresh nagra entitlement only if last written emm was more than 60 minutes ago
* [r10234]
- fix for pincode (now only 4 digits)
- cosmeticfix for boxkey => now double length
- fix for last channel on status.html
Tnx goes to Ultra47 developing the patch together with some dedicated oscam users!
* [r10233]
Remove ifdef from oscam-config.c
* [r10232]
Remove CS_ANTICASC ifdef in oscam-failban.c
* [r10231]
Clean SO_REUSEPORT ifdefs.
Most (if not all) of these calls to set socket option SO_REUSEPORT are
bogus IMHO, but cleaning this mess is left for another time. For now just
remove create a function and use it instead of sprinkling ifdefs everywhere.
I'm 97.8% sure that all of these calls can be removed...
For more info about SO_REUSEPORT see:
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c617f398edd4db2b8567a28e899a88f8f574798d
http://stackoverflow.com/questions/3261965/so-reuseport-on-linux
* [r10230]
Remove last HAVE_DVBAPI ifdef in oscam-reader.c
* [r10229]
Move function used only in dvbapi module closer to where it is called.
Also cleanup the function a bit.
* [r10228]
Eliminate the last two ifdefs's in oscam-work.c
Move the code where it belongs in module-cacheex.c and remove the
ifdefs.
* [r10227]
config: Some sed implementations do not support -i option.
Patch by kraiv @ satsupreme for improving OpenBSD support.
* [r10226]
- Cosmetic fix for internal readers driving irdeto cards at wrong speed
* [r10225]
- Revert accidental changes to config.h
* [r10224]
- dvbapi fix for boxtype pc using listenport (tnx Manio for reporting and testing)
* [r10223]
[gbx] Remove extra call to gbox_send_good_night
* [r10222]
[gbx] Fix memleak introduced with r10140.
Supposed to fix ticket 4075.
See: http://www.streamboard.tv/wbb2/thread.php?threadid=42906
* [r10221]
OpenBSD fixes from kraiv @ satsupreme.
* [r10220]
- Only allow listenport on boxtype pcnodmx
* [r10219]
- Fix for clear channelcache sporadic crashing
* [r10218]
Machine Info
boxtype with less complicated code done by gf.
* [r10217]
oscam-work: Reduce CS_CACHEEX ifdefs.
* [r10216]
oscam-work: Replace ifdef with config_enabled(MODULE_CCCAM).
* [r10215]
Update README.build.
Somebody (probably me) forgot to run `make README.build` after
changing the help text in the Makefile.
* [r10214]
dvbapi/samygo: Clean the code for future development.
According to the developers SAMYGO dvb code is not currently
working. This patch removes the config symbol and cleans the
code so future development can fix the code easily.
For more info: http://www.streamboard.tv/wbb2/thread.php?postid=533219
In theory if the code was working this transformation should
not break it and fixing only the samygo detection routine
should return it to the state before this patch minus all the
ifdefs.
* [r10213]
Remove more useless WITH_CARDREADER ifdef's.
* [r10212]
Remove some HAVE_DVBAPI ifdefs.
* [r10211]
emm: Remove useless ifdefs.
There is a static inline function defined when WITH_CARDREADER is
disabled, so the removed ifdef's were useless.
* [r10210]
newcamd: Add DVBplug (0x9922) to client list.
Fixes ticket #4076.
* [r10209]
Machine Info : small fix. Free buffer after it's not needed anymore
* [r10208]
Machine Info
added dm500 or dm600pvr boxtypes
* [r10207]
Machine info : fix compile warning mac
* [r10206]
Strings
previous ad of strtonchars was perfect on pc but not for stb's.
now it's ok for both
* [r10205]
- Webif fix for proper handling of 16 and 32 bit boxkeys
* [r10204]
- Experimental attempt to avoid var size issues with time since recently comp_timeb() gives back int64_t
* [r10203]
Strings
Added extra string function.
To read first n or last n chars of a string.
* [r10202]
[gbx] Fix warning.
* [r10201]
[gbx] minor fixes for rebroadcast thread
* [r10200]
Machine Info
forgotten fclose()
* [r10199]
Machine Info
Try out.
Added dm500 and dm600pvr
to go is extra fix to differenciate dm500 and dm6000pvr
* [r10198]
dvbapi: Simpler fix for the bug introduced in r10189 and fixed in r10196.
* [r10197]
Machine Info
cosmetic and some fixes
* [r10196]
- Fix for dvbapi blocker introduced in rev 10189
Perhaps there is a smarter fix possible but this works for me ;)
* [r10195]
dvbapi/stapi: Leave only public functions in module-dvbapi-stapi.h
Along with leaving only public functions in module-dvbapi-stapi.h
add static where it is needed in module-dvbapi-stapi.c
Bloatcheck reports > 610 byte savings after this patch.
* [r10194]
dvbapi/coolapi: Leave only public functions in module-dvbapi-coolapi.h
Along with leaving only public functions in module-dvbapi-coolapi.h
add static where it is needed in module-dvbapi-coolapi.c
Bloatcheck reports > 220 byte savings after this patch.
* [r10193]
dvbapi/mca: Leave only public functions in module-dvbapi-mca.h
Along with leaving only public functions in module-dvbapi-mca.h
add static where it is needed in module-dvbapi-mca.c
Bloatcheck reports > 200 byte savings after this patch.
* [r10192]
dvbapi/azbox: Leave only public functions in module-dvbapi-azbox.h
Also add static where it is needed. Bloatcheck reports minimal (12 bytes)
size saving.
* [r10191]
Add webif/is_defined.txt to the list ignorred files by git.
* [r10190]
dvbapi: Clean powerpc ifdefs.
Now that we have ioctl() wrapper function we can fix dm500 issues
over there.
* [r10189]
dvbapi: Create and use dvbapi_ioctl() instead of ioctl() directly.
This change would allow us to integrate SAMYGO code cleanly and also
remove some ppcold (dm500) ifdefs in the dvbapi module.
* [r10188]
dvbapi/samygo: Reformat samygo's ioctl() function.
Whitespace and comment fixes. No code changes.
* [r10187]
- Channelcache: Cached chid was never used -> Fixed!
- Show in log that channelcache was used to start descrambling a certain pid
* [r10186]
- dvbapi channelcache: only allocate mem if there is valid data that needs it!
* [r10185]
[gbx] Cosmetics.
* [r10184]
Machine Info : No need to allocate bigger mem buffers then required.
* [r10183]
dvbapi/azbox+mca: Clean ifdefs.
azbox and mca use their own handlers for ->s_handler and ->send_dcw.
Rework the code to make this explicit.
* [r10182]
dvbapi: Clean SAMYGO ifdef.
* [r10181]
dvbapi: Clean ifdefs.
According to the comment, ppcold (dm500?) returns wrong error code
on ioctl(CA_SET_DESCR).
There should be no functional changes after this patch.
* [r10180]
Machine Info
added global structure for machine info.
For now only two params, In future there will be some extra.
Added a log output by pll sci readers to see if ok for effectif use.
* [r10179]
Machine Info : boxtype fix
* [r10178]
Machine Info : fixed error handling
* [r10177]
Machine Info : relocated and made function.
* [r10176]
irdeto: Fix warning about unused parameter
* [r10175]
- Fix wrong usage of memset
* [r10174]
dvbapi: don't always reset protocol version when client closes connection
When there is more then one connection and one is closed, the protocol version
was reset, and it badly influences on the other (existing) connections.
Fixed to reset only when last connection is closed.
* [r10173]
seca: Allow each reader to have it's own IDEA state.
* [r10172]
- Init static ideacrypt only once, this will saves cpu resources!
* [r10171]
- Tweak for channelcache: if oscam.dvbapi is changed channelcache is destroyed so new prio/ignore rules are applied!
* [r10170]
seca: Include IDEA library code.
This fixes ticket #4066.
* [r10169]
dvbapi-chancache: Reduce allocations and clean code.
* [r10168]
dvbapi-chancache: Do not set thread name.
This code doesn't need to set thread name!
It is actually a bug because dvbapi_save_channel_cache() is called from main
thread and after set_thread_name() oscam process will be named dvbapi_save_channel_cache
which is unexpected to say the least.
* [r10167]
dvbapi: Move channel cache functions into their own file.
module-dvbapi.c is already too big (5000+ lines, 160k+ size).
* [r10166]
dvbapi: cleanup AZBOX and MCA code (openxcas support).
There should be no functional changes after this patch.
* [r10165]
dvbapi: Prefix channel cache funcs with dvbapi_ and add static where needed.
There should be no functional changes after this patch.
* [r10164]
cccam: Cleanup ifdefs.
Use config_enabled() instead of ifdefs. This provides compile time
checking if the code is OK even if the config option is disabled.
Also make the lb checks code readable.
There should be no functional changes after this patch.
* [r10163]
dvbapi: Cleanup ifdefs.
Use config_enabled() instead of ifdefs. This provides compile time
checking if the code is OK even if the config option is disabled.
There should be no functional changes after this patch.
* [r10162]
dvbapi: Cleanup ifdefs.
Try to minimize ifdefs when they are not needed. It is better to
declare static inline functions when some module is not enable than
to sprinkle ifdefs all over generic code.
There should be no functional changes after this patch.
* [r10161]
cacheex: Cleanup ifdefs.
Try to minimize ifdefs when they are not needed. It is better to
declare static inline functions when some module is not enable than
to sprinkle ifdefs all over generic code.
There should be no functional changes after this patch.
* [r10160]
Remove setvbuf() calls.
In my opinion there is no need to mess with stdio buffers buffers just to
read/write some small config text file. Less code == less things to go wrong,
so following that philosophy this patch removes setvbuf() calls.
* [r10159]
bulcrypt: Fix newcamd support for Bulcrypt.
These changes fixed Oscam ticket 4055, credits to kraiv and chepik from
satsupreme forum.
* [r10158]
log: Remove oscam-log-xtr.h
I fail to see the point of having a header file which defines two variables
and it is included in only one file oscam-log.c
Even if these variables are used in other places other than oscam-log.c (and
they are not) using additional header file is again the wrong thing to do.
There should be no functional changes after this patch.
* [r10157]
- channelcache load/save: some fixes for MCA / AZBOX
- removed some stale testing code
* [r10156]
- Fix for channelcache chid handling
- Fix for requestmode 0 not using exact channelcache match
* [r10155]
- dvbapi load and save of the channel cache to speedup decoding of channels
* [r10154]
Machine info
created extra header file.
Now info can be used in other .c modules if header is included to that file.
* [r10153]
Improve seca readercode
* [r10152]
Machine info
added extra's for xp1000.
added a forgotten fclose()
* [r10151]
- more fix for ticket 4065, all vars are now protected to prevent crashing!
* [r10150]
Machine info
declared some var's in c file self will have to find another way off declarations without braking some uclib compilations
* [r10149]
Machine info
renamed models info vars
* [r10148]
machine info
extra fix
* [r10147]
Machine info
Some small corrections.
* [r10146]
Machine info
extra info about the router , stb or pc .
This info can become very usefull in future to automate settings like for example: sci in function of stb.
* [r10145]
- Proper fix for ticket 4065
* [r10144]
[gbx:] Revert 10143
* [r10143]
[gbx:] Display providing peer id in ecmfmt
* [r10142]
[gbx:] Fix warnings in some build environments. Thx AbrahaM.
* [r10141]
[gbx:] Fix gbox_max_distance setting
* [r10140]
[gbx:] Stop flooding ecms to multiple gbx card peers
See also: http://www.streamboard.tv/wbb2/thread.php?threadid=42843
* [r10139]
- Webif cosmetic fixes for dvbapi FTA channels
- Webif page refresh fixes for side-effects introduced by commit 10134
Tnx Hook and Ultrar47 for providing a combined patch!
-- Andrey Pavlenko <email address hidden> Fri, 30 Jan 2015 16:28:23 +0300