Re: [vdr] xinelib / vdpau suddenly fails to detect Nvidia ION GPU
Hi, On 03.01.2014 22:01, Harald Milz wrote: On Wed, Jan 01, 2014 at 11:52:11AM +0100, Patrick Boettcher wrote: Hi, On Thursday 26 December 2013 19:13:05 Harald Milz wrote: Any idea where I should look? Why might xinelib or vdr-sxfe fail to correctly detect the ION (MCP79) chip when everything is unchanged? Yes I ran an apt-get upgrade on the underlying Ubuntu 12.04.3 OS but this didn't change anything, apparently. If this upgrade upgraded the xorg-stuff, most likely some of the xorg- libraries, which are normally overwritten by the NVIDIA-driver (assuming you are using the proprietary driver), have been re-overwritten. Hmmm, it's Ubuntu 12.04: ii nvidia-current 295.40-0ubuntu1.2 NVIDIA binary Xorg driver, kernel module and VDPAU library so it's supposed to provide the correct files. What's the output of: debsums -a nvidia-current ? André ___ vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
[vdr] [ANNOUNCE] VDR developer version 2.1.3
VDR developer version 2.1.3 is now available at ftp://ftp.tvdr.de/vdr/Developer/vdr-2.1.3.tar.bz2 A 'diff' against the previous version is available at ftp://ftp.tvdr.de/vdr/Developer/vdr-2.1.2-2.1.3.diff MD5 checksums: 054f80e0045aa6fad118e9285b52f4f2 vdr-2.1.3.tar.bz2 3c5ab05d5c4d0b984b34e84190e80949 vdr-2.1.2-2.1.3.diff WARNING: This is a *developer* version. Even though *I* use it in my productive environment, I strongly recommend that you only use it under controlled conditions and for testing and debugging. Originally I intended to release this version only after the new DiSEqC configuration dialog was finished. But in the meantime quite a few other things have come up, so I decided to postpone that dialog and first release what has piled up so far. The changes since version 2.1.2: - Changed the return value of cPositioner::HorizonLongitude() to 0 in case the latitude of the antenna location is beyond +/-81 degrees. - Updated the Finnish OSD texts (thanks to Rolf Ahrenberg). - Fixed some compiler warnings with gcc-4.6.3 (thanks to Rolf Ahrenberg). - Changed the name of the SVDRP command RENR to MOVR (suggested by Rolf Ahrenberg). - When cutting a recording it is now checked whether there is already an edited version of this recording (with the same name, but starting with '%'), and the user is prompted for confirmation to overwrite it (suggested by Rolf Ahrenberg). - Revoked Added maximum signal strength value for TechniSat SkyStar 2 DVB-S rev 2.3P because it broke things for the TechniSat AirStar 2 DVB-T card. - The LIRC remote control now connects to the socket even if it doesn't yet exist when VDR is started (thanks to Lars Hanisch). - Changed the absolute latitude limit for visible satellites to 81.2 degrees. - Added code for parsing LCN and AVC descriptors to libsi (thanks to Rolf Ahrenberg). - In the Select folder menu pressing Ok now selects the folder, even if this is a folder that contains sub folders (marked with ...). To open such a folder you can press the Red key. - Fixed a possible access to uninitialized data in cEIT::cEIT() (reported by Dominik Strasser). - The new menu category mcRecordingEdit is now used to mark menus that edit recording properties (suggested by Stefan Braun). - Changes in the teletext PID no longer cause retuning (and thus interrupting a recording). - Removed '_' from the FileNameChars and CharMap translations in uk_UA.po. - Updated the Italian OSD texts (thanks to Diego Pierotto). - Fixed a missing initialization in the c'tor of cSkinLCARSDisplayChannel (thanks to Marko Mäkelä). - Simplified some conditional expressions in skinlcars.c and skinsttng.c (suggested by Marko Mäkelä). - Fixed uninitialized item area coordinates in cSkinLCARSDisplayMenu (reported by Marko Mäkelä). - Fixed a possible crash if the recordings list is updated externally while the Recordings menu is open (reported by Lars Hanisch). - Added a missing closing ')' in the help and man page entry of the --vfat option (reported by Lars Hanisch). - Fixed setting the name of the video directory to avoid a crash when using --genindex, and also to use the correct directory with --edit (the latter reported by Marko Mäkelä). - The Recordings menu can now be called with a cRecordingFilter, which allows the caller to have it display only a certain subset of the recordings (thanks to Lars Hanisch). - Added handling UTF-8 'umlaut' characters to cKbdRemote (thanks to Lars Hanisch). - Made it clear that the Data parameter in cDevice::StillPicture() may point to a series of packets, not just a single one (thanks to Thomas Reufer). - cDevice::TrickSpeed() now has an additional parameter named Forward, which indicates the direction in which replay is being done (suggested by Thomas Reufer). This information may be necessary for some output devices in order to properly implement trick modes. Authors of plugins that implement output devices will need to add this parameter to their derived cDevice class, regardless of whether they will make use of it or not. - Added a note to ePlayMode in device.h that VDR itself always uses pmAudioVideo when replaying a recording (suggested by Thomas Reufer). - Fixed some spellings in positioner.h and Doxyfile (thanks to Ville Skyttä). - Changed '%a' to the POSIX compliant '%m' in all scanf() calls (thanks to Ville Skyttä). - The new function cCamSlot::Decrypt() can be used by derived classes to implement a CAM slot that can be freely assigned to any device, without being directly inserted into the full TS data stream in hardware. A derived class that implements Decrypt() will also need to set the new parameter ReceiveCaPids in the call to the cCamSlot base class constructor to true, in order to receive the CA pid TS packets that contain data necessary for decrypting. - Many member functions of cCamSlot have been made virtual to allow for easier
Re: [vdr] [ANNOUNCE] VDR developer version 2.1.3
Am 05.01.2014 12:42, schrieb Klaus Schmidinger: VDR developer version 2.1.3 is now available at ftp://ftp.tvdr.de/vdr/Developer/vdr-2.1.3.tar.bz2 A 'diff' against the previous version is available at ftp://ftp.tvdr.de/vdr/Developer/vdr-2.1.2-2.1.3.diff MD5 checksums: 054f80e0045aa6fad118e9285b52f4f2 vdr-2.1.3.tar.bz2 3c5ab05d5c4d0b984b34e84190e80949 vdr-2.1.2-2.1.3.diff WARNING: This is a *developer* version. Even though *I* use it in my productive environment, I strongly recommend that you only use it under controlled conditions and for testing and debugging. Originally I intended to release this version only after the new DiSEqC configuration dialog was finished. But in the meantime quite a few other things have come up, so I decided to postpone that dialog and first release what has piled up so far. The changes since version 2.1.2: - Changed the return value of cPositioner::HorizonLongitude() to 0 in case the latitude of the antenna location is beyond +/-81 degrees. - Updated the Finnish OSD texts (thanks to Rolf Ahrenberg). - Fixed some compiler warnings with gcc-4.6.3 (thanks to Rolf Ahrenberg). - Changed the name of the SVDRP command RENR to MOVR (suggested by Rolf Ahrenberg). - When cutting a recording it is now checked whether there is already an edited version of this recording (with the same name, but starting with '%'), and the user is prompted for confirmation to overwrite it (suggested by Rolf Ahrenberg). - Revoked Added maximum signal strength value for TechniSat SkyStar 2 DVB-S rev 2.3P because it broke things for the TechniSat AirStar 2 DVB-T card. - The LIRC remote control now connects to the socket even if it doesn't yet exist when VDR is started (thanks to Lars Hanisch). - Changed the absolute latitude limit for visible satellites to 81.2 degrees. - Added code for parsing LCN and AVC descriptors to libsi (thanks to Rolf Ahrenberg). - In the Select folder menu pressing Ok now selects the folder, even if this is a folder that contains sub folders (marked with ...). To open such a folder you can press the Red key. - Fixed a possible access to uninitialized data in cEIT::cEIT() (reported by Dominik Strasser). - The new menu category mcRecordingEdit is now used to mark menus that edit recording properties (suggested by Stefan Braun). - Changes in the teletext PID no longer cause retuning (and thus interrupting a recording). - Removed '_' from the FileNameChars and CharMap translations in uk_UA.po. - Updated the Italian OSD texts (thanks to Diego Pierotto). - Fixed a missing initialization in the c'tor of cSkinLCARSDisplayChannel (thanks to Marko Mäkelä). - Simplified some conditional expressions in skinlcars.c and skinsttng.c (suggested by Marko Mäkelä). - Fixed uninitialized item area coordinates in cSkinLCARSDisplayMenu (reported by Marko Mäkelä). - Fixed a possible crash if the recordings list is updated externally while the Recordings menu is open (reported by Lars Hanisch). - Added a missing closing ')' in the help and man page entry of the --vfat option (reported by Lars Hanisch). - Fixed setting the name of the video directory to avoid a crash when using --genindex, and also to use the correct directory with --edit (the latter reported by Marko Mäkelä). - The Recordings menu can now be called with a cRecordingFilter, which allows the caller to have it display only a certain subset of the recordings (thanks to Lars Hanisch). - Added handling UTF-8 'umlaut' characters to cKbdRemote (thanks to Lars Hanisch). - Made it clear that the Data parameter in cDevice::StillPicture() may point to a series of packets, not just a single one (thanks to Thomas Reufer). - cDevice::TrickSpeed() now has an additional parameter named Forward, which indicates the direction in which replay is being done (suggested by Thomas Reufer). This information may be necessary for some output devices in order to properly implement trick modes. Authors of plugins that implement output devices will need to add this parameter to their derived cDevice class, regardless of whether they will make use of it or not. - Added a note to ePlayMode in device.h that VDR itself always uses pmAudioVideo when replaying a recording (suggested by Thomas Reufer). - Fixed some spellings in positioner.h and Doxyfile (thanks to Ville Skyttä). - Changed '%a' to the POSIX compliant '%m' in all scanf() calls (thanks to Ville Skyttä). - The new function cCamSlot::Decrypt() can be used by derived classes to implement a CAM slot that can be freely assigned to any device, without being directly inserted into the full TS data stream in hardware. A derived class that implements Decrypt() will also need to set the new parameter ReceiveCaPids in the call to the cCamSlot base class constructor to
Re: [vdr] [ANNOUNCE] VDR developer version 2.1.3
I will! :) Who not ... ;-) Thanks for a bunch of little gems in this release like obsolete channels. Yes, indeed ... :tup Have to dig into the new CAM stuff, maybe now it's possible to integrate the CI of Digital Devices... That would be great, christmas has gone recently, but I wouldn't say no for delayed present. === Kind regards fnu -Original Message- From: vdr-boun...@linuxtv.org [mailto:vdr-boun...@linuxtv.org] On Behalf Of Lars Hanisch Sent: Sunday, January 05, 2014 12:54 PM To: vdr@linuxtv.org Subject: Re: [vdr] [ANNOUNCE] VDR developer version 2.1.3 Am 05.01.2014 12:42, schrieb Klaus Schmidinger: VDR developer version 2.1.3 is now available at ftp://ftp.tvdr.de/vdr/Developer/vdr-2.1.3.tar.bz2 A 'diff' against the previous version is available at ftp://ftp.tvdr.de/vdr/Developer/vdr-2.1.2-2.1.3.diff MD5 checksums: 054f80e0045aa6fad118e9285b52f4f2 vdr-2.1.3.tar.bz2 3c5ab05d5c4d0b984b34e84190e80949 vdr-2.1.2-2.1.3.diff WARNING: This is a *developer* version. Even though *I* use it in my productive environment, I strongly recommend that you only use it under controlled conditions and for testing and debugging. Originally I intended to release this version only after the new DiSEqC configuration dialog was finished. But in the meantime quite a few other things have come up, so I decided to postpone that dialog and first release what has piled up so far. The changes since version 2.1.2: - Changed the return value of cPositioner::HorizonLongitude() to 0 in case the latitude of the antenna location is beyond +/-81 degrees. - Updated the Finnish OSD texts (thanks to Rolf Ahrenberg). - Fixed some compiler warnings with gcc-4.6.3 (thanks to Rolf Ahrenberg). - Changed the name of the SVDRP command RENR to MOVR (suggested by Rolf Ahrenberg). - When cutting a recording it is now checked whether there is already an edited version of this recording (with the same name, but starting with '%'), and the user is prompted for confirmation to overwrite it (suggested by Rolf Ahrenberg). - Revoked Added maximum signal strength value for TechniSat SkyStar 2 DVB-S rev 2.3P because it broke things for the TechniSat AirStar 2 DVB-T card. - The LIRC remote control now connects to the socket even if it doesn't yet exist when VDR is started (thanks to Lars Hanisch). - Changed the absolute latitude limit for visible satellites to 81.2 degrees. - Added code for parsing LCN and AVC descriptors to libsi (thanks to Rolf Ahrenberg). - In the Select folder menu pressing Ok now selects the folder, even if this is a folder that contains sub folders (marked with ...). To open such a folder you can press the Red key. - Fixed a possible access to uninitialized data in cEIT::cEIT() (reported by Dominik Strasser). - The new menu category mcRecordingEdit is now used to mark menus that edit recording properties (suggested by Stefan Braun). - Changes in the teletext PID no longer cause retuning (and thus interrupting a recording). - Removed '_' from the FileNameChars and CharMap translations in uk_UA.po. - Updated the Italian OSD texts (thanks to Diego Pierotto). - Fixed a missing initialization in the c'tor of cSkinLCARSDisplayChannel (thanks to Marko Mäkelä). - Simplified some conditional expressions in skinlcars.c and skinsttng.c (suggested by Marko Mäkelä). - Fixed uninitialized item area coordinates in cSkinLCARSDisplayMenu (reported by Marko Mäkelä). - Fixed a possible crash if the recordings list is updated externally while the Recordings menu is open (reported by Lars Hanisch). - Added a missing closing ')' in the help and man page entry of the -- vfat option (reported by Lars Hanisch). - Fixed setting the name of the video directory to avoid a crash when using --genindex, and also to use the correct directory with --edit (the latter reported by Marko Mäkelä). - The Recordings menu can now be called with a cRecordingFilter, which allows the caller to have it display only a certain subset of the recordings (thanks to Lars Hanisch). - Added handling UTF-8 'umlaut' characters to cKbdRemote (thanks to Lars Hanisch). - Made it clear that the Data parameter in cDevice::StillPicture() may point to a series of packets, not just a single one (thanks to Thomas Reufer). - cDevice::TrickSpeed() now has an additional parameter named Forward, which indicates the direction in which replay is being done (suggested by Thomas Reufer). This information may be necessary for some output devices in order to properly implement trick modes. Authors of plugins that implement output devices will need to add this parameter to their derived cDevice class, regardless of whether they will make use of it or not. - Added a note to ePlayMode in device.h that VDR itself always
Re: [vdr] [ANNOUNCE] VDR developer version 2.1.3
Am 05.01.2014 13:01, schrieb fnu: I will! :) Who not ... ;-) Thanks for a bunch of little gems in this release like obsolete channels. Yes, indeed ... :tup Have to dig into the new CAM stuff, maybe now it's possible to integrate the CI of Digital Devices... That would be great, christmas has gone recently, but I wouldn't say no for delayed present. I'll do my very best... :) Lars. === Kind regards fnu -Original Message- From: vdr-boun...@linuxtv.org [mailto:vdr-boun...@linuxtv.org] On Behalf Of Lars Hanisch Sent: Sunday, January 05, 2014 12:54 PM To: vdr@linuxtv.org Subject: Re: [vdr] [ANNOUNCE] VDR developer version 2.1.3 Am 05.01.2014 12:42, schrieb Klaus Schmidinger: VDR developer version 2.1.3 is now available at ftp://ftp.tvdr.de/vdr/Developer/vdr-2.1.3.tar.bz2 A 'diff' against the previous version is available at ftp://ftp.tvdr.de/vdr/Developer/vdr-2.1.2-2.1.3.diff MD5 checksums: 054f80e0045aa6fad118e9285b52f4f2 vdr-2.1.3.tar.bz2 3c5ab05d5c4d0b984b34e84190e80949 vdr-2.1.2-2.1.3.diff WARNING: This is a *developer* version. Even though *I* use it in my productive environment, I strongly recommend that you only use it under controlled conditions and for testing and debugging. Originally I intended to release this version only after the new DiSEqC configuration dialog was finished. But in the meantime quite a few other things have come up, so I decided to postpone that dialog and first release what has piled up so far. The changes since version 2.1.2: - Changed the return value of cPositioner::HorizonLongitude() to 0 in case the latitude of the antenna location is beyond +/-81 degrees. - Updated the Finnish OSD texts (thanks to Rolf Ahrenberg). - Fixed some compiler warnings with gcc-4.6.3 (thanks to Rolf Ahrenberg). - Changed the name of the SVDRP command RENR to MOVR (suggested by Rolf Ahrenberg). - When cutting a recording it is now checked whether there is already an edited version of this recording (with the same name, but starting with '%'), and the user is prompted for confirmation to overwrite it (suggested by Rolf Ahrenberg). - Revoked Added maximum signal strength value for TechniSat SkyStar 2 DVB-S rev 2.3P because it broke things for the TechniSat AirStar 2 DVB-T card. - The LIRC remote control now connects to the socket even if it doesn't yet exist when VDR is started (thanks to Lars Hanisch). - Changed the absolute latitude limit for visible satellites to 81.2 degrees. - Added code for parsing LCN and AVC descriptors to libsi (thanks to Rolf Ahrenberg). - In the Select folder menu pressing Ok now selects the folder, even if this is a folder that contains sub folders (marked with ...). To open such a folder you can press the Red key. - Fixed a possible access to uninitialized data in cEIT::cEIT() (reported by Dominik Strasser). - The new menu category mcRecordingEdit is now used to mark menus that edit recording properties (suggested by Stefan Braun). - Changes in the teletext PID no longer cause retuning (and thus interrupting a recording). - Removed '_' from the FileNameChars and CharMap translations in uk_UA.po. - Updated the Italian OSD texts (thanks to Diego Pierotto). - Fixed a missing initialization in the c'tor of cSkinLCARSDisplayChannel (thanks to Marko Mäkelä). - Simplified some conditional expressions in skinlcars.c and skinsttng.c (suggested by Marko Mäkelä). - Fixed uninitialized item area coordinates in cSkinLCARSDisplayMenu (reported by Marko Mäkelä). - Fixed a possible crash if the recordings list is updated externally while the Recordings menu is open (reported by Lars Hanisch). - Added a missing closing ')' in the help and man page entry of the -- vfat option (reported by Lars Hanisch). - Fixed setting the name of the video directory to avoid a crash when using --genindex, and also to use the correct directory with --edit (the latter reported by Marko Mäkelä). - The Recordings menu can now be called with a cRecordingFilter, which allows the caller to have it display only a certain subset of the recordings (thanks to Lars Hanisch). - Added handling UTF-8 'umlaut' characters to cKbdRemote (thanks to Lars Hanisch). - Made it clear that the Data parameter in cDevice::StillPicture() may point to a series of packets, not just a single one (thanks to Thomas Reufer). - cDevice::TrickSpeed() now has an additional parameter named Forward, which indicates the direction in which replay is being done (suggested by Thomas Reufer). This information may be necessary for some output devices in order to properly implement trick modes. Authors of plugins that implement output devices will need to add this parameter to their derived cDevice class, regardless of whether they will make use of it or not. - Added a note to ePlayMode in device.h that VDR itself
Re: [vdr] [ANNOUNCE] VDR developer version 2.1.3
The changes since version 2.1.2: - The new function cCamSlot::Decrypt() can be used by derived classes to implement a CAM slot that can be freely assigned to any device, without being directly inserted into the full TS data stream in hardware. A derived class that implements Decrypt() will also need to set the new parameter ReceiveCaPids in the call to the cCamSlot base class constructor to true, in order to receive the CA pid TS packets that contain data necessary for decrypting. - Many member functions of cCamSlot have been made virtual to allow for easier implementation of derived classes. - cDvbDevice::GetTSPacket() now calls CamSlot()-Decrypt() in order to allow CAM slots that can be freely assigned to any device access to the TS data stream. This is very exiting to me.I only watch one channel at a time, but I would like to record multiple channels at the same time. I am investigating if it's possible to move the decryption functionality from recording to playback time. This allows me to run VDR on an underpowered platform that is fast enough to record multiple shows, but only has the CPU power to decrypt one channel. (My NAS would be such a platform) Receiving all caPID's is a big step in that direction. Maybe I can alter VDR so that it produces a recording containing the caPID's, the keys from oscam and the audio /video data. Next, during playback I can then feed it through libdvbcsa. Also, when the CPU has some time available, it can decode TV shows in a low-priority process. This will further reduce the realtime CPU load. I wonder if there is a standard-compliant way to store the decryption keys in the resulting .ts file. Thank you for your hard work!. Kind regards, Cedric ___ vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
Re: [vdr] [ANNOUNCE] VDR developer version 2.1.3
Am 05.01.2014 14:40, schrieb Lars Hanisch: I'll do my very best... :) Lars. This is funny! English, but only understood by German. Gerald !DSPAM:52c9992d302581408685597! ___ vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
Re: [vdr] xinelib / vdpau suddenly fails to detect Nvidia ION GPU
On Sun, Jan 05, 2014 at 12:01:50PM +0100, André Weidemann wrote: On 03.01.2014 22:01, Harald Milz wrote: On Wed, Jan 01, 2014 at 11:52:11AM +0100, Patrick Boettcher wrote: Hi, On Thursday 26 December 2013 19:13:05 Harald Milz wrote: Any idea where I should look? Why might xinelib or vdr-sxfe fail to correctly detect the ION (MCP79) chip when everything is unchanged? Yes I ran an apt-get upgrade on the underlying Ubuntu 12.04.3 OS but this didn't change anything, apparently. If this upgrade upgraded the xorg-stuff, most likely some of the xorg- libraries, which are normally overwritten by the NVIDIA-driver (assuming you are using the proprietary driver), have been re-overwritten. Hmmm, it's Ubuntu 12.04: ii nvidia-current 295.40-0ubuntu1.2 NVIDIA binary Xorg driver, kernel module and VDPAU library so it's supposed to provide the correct files. What's the output of: debsums -a nvidia-current ? 132 lines of OK. Ah OK, as of today, Ubuntu upgraded to nvidia-304, testing now ... -- Expect the worst, it's the least you can do. ___ vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr