Interestingly, Coverity implies that close takes an unsigned
argument, while my close(2) man page shows it taking a signed
argument. I guess it may be platforms specific.
Coverity 1214602
New approach attempts to be more accurate by measuring
the elapsed time by iteration. Also:
* Use a 10 seconds default timeout and a half a second
polling step. New values should better match the tuning
process on real-life scenarios.
* Correct elapsed_time computation.
* Add _retry_ioctl() to avoid bailing out on temporary
ioctl EINTR failures (no need to check for EAGAIN cause
we are opening the frontend on blocking mode)
* Small corrections to fail condition handling
* Drop remaining sleep() logic in favor of polling
* Use best guess delivery system if none is set
* Make tuning/locking timeout configurable
* Add signals for tuning start, done and fail
* Drop gst_dvbsrc_frontend_status(). It was used only
for signal LOCK checking. This is now part of the
tuning/locking loop
* Break up frontend configuration and tuning
on separate functions
Plus:
* Add some more useful DEBUG/TRACE messages
* Move over misplaced DVB API message
* Fix wrong comment for default DVB buffer size (http://linuxtv.org/downloads/v4l-dvb-apis/dmx_fcalls.html#DMX_SET_BUFFER_SIZE)
This patch builds up on previous work done by
Fabrizio (Misto) Milo <mistobaan@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=641204
c400eef377 introduced some defines to handle
older kernel headers. However, the check is done before the corresponding
kernel header (dvb/frontend.h) is included. As a result the macros are
always defined with results in 'redefined' errors with newer kernel
headers.
Move the check after the include to fix this.
https://bugzilla.gnome.org/show_bug.cgi?id=730570
usecount is unsigned, so too many "unuse" will wrap the counter
around and the >= 0 check will always be fine.
It would be much simpler to just make the counter signed, but
moving the checks where the decrements happen allow a mistake
to be detected earlier, and thus easier to debug.
Coverity 1139791
-add delsys property
-add delivery system capability to the gstreamer adapter structure
-ready for add new delivery systems
Application must ask the adapter structure to know which delivery systems are avaible.
The property delsys must be set.
https://bugzilla.gnome.org/show_bug.cgi?id=709414
Which causes problems when used with _GNU_SOURCE apparently, and it
seems it was only set because of usleep(), which we can just replace
with g_usleep() until we get rid of those entirely.
https://bugzilla.gnome.org/show_bug.cgi?id=705208
Original code to parse satellite delivery descriptors to generate
"satellite" structures appeared to be copy & pasted from cable's code
without amending for satellite.
Also added 8PSK to dvbsrc's enum for modulation.
https://bugzilla.gnome.org/show_bug.cgi?id=654485
Conflicts:
gst/mpegdemux/gstmpegdesc.h
gst/mpegdemux/mpegtspacketizer.c
Make the _read_device function return a GstFlowReturn so that we can propagate
this to the caller. This allows us to differentiate between poll errors and
flushing state.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=674219
This allows to use much higher timeout values because GstPoll
is interruptible and keeps the number of wakeups during signal
loss lower.
Fixes bug #608171.
fixes the following compile errors:
cc1: warnings being treated as errors
camswclient.c: In function 'cam_sw_client_open':
camswclient.c:81: warning: implicit declaration of function 'strncpy'
camswclient.c:81: warning: incompatible implicit declaration of built-in function 'strncpy'
camswclient.c:89: warning: implicit declaration of function 'strerror'
camswclient.c:89: warning: nested extern declaration of 'strerror'
camswclient.c:89: warning: format '%s' expects type 'char *', but argument 9 has type 'int'
camswclient.c: In function 'send_ca_pmt':
camswclient.c:129: warning: implicit declaration of function 'memcpy'
camswclient.c:129: warning: incompatible implicit declaration of built-in
function 'memcpy'
gstdvbsrc.c:48:19: error: error.h: No such file or directory
Signed-off-by: Rob Clark <rob@ti.com>
Move include directives for gst-libs into GST_PLUGINS_BAD_CFLAGS,
and fix all the Makefiles that use it. This is so that all the
include directories are added in the proper order: first the
directories in srcdir/builddir, then gst-plugins-base dirs, then
gstreamer dirs. If the order is wrong, installed headers may be
used instead of local headers and/or uninstalled headers from -base.
Some devices take a while to tune and 500ms was too short to detect
successful tuning. As well as waiting for 5 seconds each 100ms the
status is checked and the loop is broken out of when tuning has suceeded.
Original commit message from CVS:
Patch by: Frederic Crozat <fcrozat@mandriva.org>
* ext/sndfile/gstsf.c: (plugin_init):
* sys/dvb/gstdvbsrc.c: (gst_dvbsrc_plugin_init):
* sys/oss4/oss4-audio.c: (plugin_init):
Make sure gettext returns translations in UTF-8 encoding rather
than in the current locale encoding (#546822).
Original commit message from CVS:
* sys/dvb/gstdvbsrc.c:
* sys/dvb/parsechannels.c:
Add DVB-C support. Special thanks to Christian Schaller
for a testing ground.
Original commit message from CVS:
* sys/dvb/camutils.c:
Don't free the program descriptors, this structure
containing them is stills tored after.
Fixes data corruption.
Original commit message from CVS:
* sys/dvb/gstdvbsrc.c:
Add initial support for multiproto driver (not yet merged into
v4l-dvb mainline yet).
Only works for DVB-S not DVB-S2, DVB-T, DVB-C or other.
Original commit message from CVS:
* sys/dvb/Makefile.am:
* sys/dvb/dvbbasebin.c:
Add URI Handler for dvb.
Re-order pad templates to workaround a bug in playbasebin.
* sys/dvb/parsechannels.c:
* sys/dvb/parsechannels.h:
Add code to parse channels from zap-style channels.conf files.
Original commit message from CVS:
* sys/dvb/gstdvbsrc.c:
Post bus message about adapter type and it's capabilities,
when opening the frontend.
After failing to read from the dvr, post a bus message to
inform the app.
Original commit message from CVS:
* sys/dvb/gstdvbsrc.c:
Do not go on forever if problem with reading from dvr, rather
return NULL.
Handle some cleanup issues of closing filedescriptors when
failing to tune or similar.
Original commit message from CVS:
* ext/ladspa/gstladspa.c: (gst_ladspa_get_property):
* ext/sdl/sdlvideosink.c: (gst_sdlvideosink_show_frame):
* gst/mve/gstmvemux.c: (gst_mve_mux_request_new_pad):
* sys/dvb/dvbbasebin.c: (dvb_base_bin_class_init):
Fix 'xyz may be used uninitialized' compiler warnings caused
by broken g_assert_not_reached() macro in GLib-2.15.x and don't
abort() in any case but properly report the error.
Original commit message from CVS:
* gst/mpegtsparse/mpegtspacketizer.c:
* gst/mpegtsparse/mpegtsparse.c:
A sub table is identified by the pair table_id and
sub_table_identifier, not by pid. So hash with that.
* sys/dvb/dvbbasebin.c:
Make sure initial pids are added properly to filter,
Original commit message from CVS:
* gst/mpegtsparse/Makefile.am:
* gst/mpegtsparse/mpegtspacketizer.c:
* gst/mpegtsparse/mpegtsparse.c:
Remove signals for pat, pmt, nit, eit, sdt. Replace with bus
messages.
* sys/dvb/dvbbasebin.c:
Instead of attaching to signals, use the bus messages.
Also fix up so the dvbsrc starts only outputting the info tables
like PAT, CAT, NIT, SDT, EIT instead of the whole ts.
Original commit message from CVS:
patch by: Alessandro Decina
* gst/mpegtsparse/mpegtspacketizer.c:
* gst/mpegtsparse/mpegtspacketizer.h:
* gst/mpegtsparse/mpegtsparse.c:
* gst/mpegtsparse/mpegtsparse.h:
pat-info is now a signal not a GObject property that
gets notified.
pat-info, pmt-info now instead of passing a GObject as
a parameter, pass a GstStructure.
New signals: nit-info, sdt-info, eit-info for DVB SI information
* sys/dvb/camconditionalaccess.c:
* sys/dvb/camconditionalaccess.h:
* sys/dvb/camdevice.c:
* sys/dvb/camdevice.h:
* sys/dvb/camswclient.c:
* sys/dvb/camswclient.h:
* sys/dvb/camutils.c:
* sys/dvb/camutils.h:
Cam code now uses the pmt GstStructure passed from mpegtsparse
signals rather than the GObject.
* sys/dvb/dvbbasebin.c:
Use new signals in mpegtsparse and use GstStructures as per
mpegtsparse's modified API.
Original commit message from CVS:
* sys/dvb/gstdvbsrc.c:
Actually use the code-rate-hp parameter for DVB-S.
It turns out setting to AUTO does not always work (
especially in diseq situations). Set by default to
FEC_AUTO.
Original commit message from CVS:
patch by: Alessandro Decina
* sys/dvb/Makefile.am:
* sys/dvb/cam.c:
* sys/dvb/cam.h:
* sys/dvb/camapplication.c:
* sys/dvb/camapplication.h:
* sys/dvb/camapplicationinfo.c:
* sys/dvb/camapplicationinfo.h:
* sys/dvb/camconditionalaccess.c:
* sys/dvb/camconditionalaccess.h:
* sys/dvb/camdevice.c:
* sys/dvb/camdevice.h:
* sys/dvb/camresourcemanager.c:
* sys/dvb/camresourcemanager.h:
* sys/dvb/camsession.c:
* sys/dvb/camsession.h:
* sys/dvb/camswclient.c:
* sys/dvb/camswclient.h:
* sys/dvb/camtransport.c:
* sys/dvb/camtransport.h:
* sys/dvb/camutils.c:
* sys/dvb/camutils.h:
* sys/dvb/dvbbasebin.c:
* sys/dvb/dvbbasebin.h:
* sys/dvb/gstdvb.c:
* sys/dvb/gstdvbsrc.c:
* sys/dvb/gstdvbsrc.h:
Integrate SoC work done by Alessandro for the Freevo project.
Adds cam support to the dvb stack in GStreamer and a new
element (actually a bin) called dvbbasebin that integrates
dvbsrc and mpegtsparse to a) handle decryption and b) allow
acquiring multiple channels on same transponder without
knowing pid numbers.
Original commit message from CVS:
2006-11-17 Zaheer Abbas Merali <zaheerabbas at merali dot org>
* sys/dvb/gstdvbsrc.c: (gst_dvbsrc_tune):
Actually set the TONE parameter for DVB-S when we should.
Original commit message from CVS:
2006-11-09 Zaheer Abbas Merali <zaheerabbas at merali dot org>
* sys/dvb/gstdvbsrc.c: (gst_dvbsrc_set_property)
(gst_dvbsrc_create), (gst_dvbsrc_output_frontend_stats),
(gst_dvbsrc_tune), (gst_dvbsrc_set_pes_filters):
Rename set_pes_filter to set_pes_filters for consistency.
Do run-time pid filtering.
Original commit message from CVS:
2006-11-08 Zaheer Abbas Merali <zaheerabbas at merali dot org>
* sys/dvb/gstdvbsrc.c: (read_device):
Do not call output_frontend_stats from a method that does not
have the gstdvbsrc object.
This fixes the build, sorry guys!
Original commit message from CVS:
2006-11-07 Zaheer Abbas Merali <zaheerabbas at merali dot org>
* sys/dvb/gstdvbsrc.c: (read_device), (gst_dvbsrc_create),
(gst_dvbsrc_output_frontend_stats), (gst_dvbsrc_tune),
(gst_dvbsrc_set_pes_filter):
* sys/dvb/gstdvbsrc.h:
Force PAT to always be in the filter.
Try to continue reading after failing 10 times.
Original commit message from CVS:
* po/POTFILES.in:
add dvbsrc
* sys/dvb/gstdvbsrc.c: (gst_dvbsrc_class_init), (gst_dvbsrc_init),
(gst_dvbsrc_set_property), (gst_dvbsrc_get_property),
(gst_dvbsrc_open_frontend), (gst_dvbsrc_open_dvr),
(gst_dvbsrc_finalize), (read_device), (gst_dvbsrc_create),
(gst_dvbsrc_start), (gst_dvbsrc_frontend_status),
(gst_dvbsrc_tune), (gst_dvbsrc_unset_pes_filters),
(gst_dvbsrc_set_pes_filter):
* sys/dvb/gstdvbsrc.h:
adapter-prefix is a terrible name for a device, so replace it.
strerror should be used immediately after the syscall related to it.
Actually throw a reasonable GST_ELEMENT_ERROR if we fail to open
(pretty easy to trigger by opening the same device twice).
Use _OBJECT macros where it makes sense.
This element is going to need some work before it moves.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (extract_initial_length_and_fourcc),
(gst_qtdemux_loop_state_header):
Printf format fixes.
* sys/dvb/gstdvbsrc.c:
Use "_stdint.h".
Original commit message from CVS:
2006-10-10 Zaheer Abbas Merali <zaheerabbas at merali dot org>
* sys/dvb/Makefile.am:
* sys/dvb/dvb-api/Makefile.am:
* sys/dvb/dvb-api/audio.h:
* sys/dvb/dvb-api/ca.h:
* sys/dvb/dvb-api/dmx.h:
* sys/dvb/dvb-api/frontend.h:
* sys/dvb/dvb-api/net.h:
* sys/dvb/dvb-api/osd.h:
* sys/dvb/dvb-api/version.h:
* sys/dvb/dvb-api/video.h:
Remove these files, not needed use system files.
* sys/dvb/gstdvbsrc.c: (gst_dvbsrc_set_property),
(gst_dvbsrc_create), (gst_dvbsrc_output_frontend_stats),
(gst_dvbsrc_tune):
Fix warnings, uint16_t etc. is in stdint.h.
Use system includes not our local ones.
