Hello everyone. This mail is a little bit long, I'm sorry for that. But I have to describe the TV-situation in Germany roughly. Without that knowledge it is definitely not understandable why this DVB-S cards are still very useful here. Reader familiar with this crazy situation may proceed to the driver section below. I'm primary a user of this cards but have profound knowledge about the Hardware. I used to repair the cards for me and other users back then. I own Fullfeatured and Budget-cards and use them daily in my VDR-System. In Germany many channels are free-to-air in the DVB-S version only. I would like to use these cards for a few more years until DVB-S is deactivated or the HD-Versions of the channels become free-to-air. I'm not willing to pay 75 Euro a year for TV-commercials in HD and I'm not the only one with this opinion. 14 million of the 17 million satelite-tv-households in Germany watch this channels in SD-quality only (~82%)![1] In addition uses the encrypting-system a proprietary CAM extension which makes it impossible to watch this channels on a HTPC legally. This situation won't change until 2025 (by a kind of law!). What then happens is currently completely uncertain. The driver topic however is new to me, the cards where always working out of the box. I noticed the upcoming removal right before my first mail. Honestly I was a little shocked that the driver may be removed from Kernel. The card may be old and not produced any more, but they are not rare and easily obtainable second hand. There are always multiple offers for reasonable prices on the common platforms. And the cards are running flawless on current mainboards with PCIe-PCI Bridge. There must be a lot of SAA7146 based cards been sold in Europe. Many brands sold them, mostly rebranded Technotrend cards. Even Hauppauge, the most important brand, sold the TT-Budged as "Nova" and the Fullfeatured as "Nexus" for years. Their own Connexant based designs came pretty late, short before the PCIe-cards. I carefully estimate, at least 50% of all PCI-DVB-cards sold in Europe where SAA7146 based. There must be still a number of users out there. The relevance of SAA7146 for PCI-DVB-cards is almost like the Bttv-Chips for analog TV a few years before. At least in Europe. And the bttv driver not deprecated despite older and using videobuf1 API! SAA7146 driver ============== I've read a lot in the last days and the main issue with the SAA7146 driver seems to be the missing maintainer. All other issues seem to be a result of that. Right? And the driver desperately needs someone with expertise about the cards and the driver. I've spotted a big chunk of unused code just by knowing which cards have been build and which not. Sören Moch offered to maintain the complete SAA7146 driver in the VDRportal and in this list too. This offer includes the videobuf2 conversion too. On condition that the support of the fullfeatured cards stays in the kernel. I understand that. He only owns fullfeatured cards. And I am interested in keeping my fullfeatured cards operational too. I'm convinced Sören could handle the driver well and he is the only one I know who probably could do that. And he is actively offering to do the job. That leads to the DVB-API part for the AV7110 which should be removed. An API conversion for the AV7110 does not make sense any more. 10 years ago maybe, but not now. Working software would be broken and there will be no benefit for the user at all. Converting is however not easy and a driver specific UAPI would be necessary in any case. The ioctl "VIDEO_SELECT_SOURCE" needed and definitely missing in the V4l2-API (see ivtv driver). The OSD of the FF-Cards is more canvas like, not a framebuffer. The OSD ioctl are also needed and I haven't found anything in V4l2-API to replace them. What about putting the 3 API-files into one driver specific UAPI file? The deprecated DVB-API part could be officially removed and the maintainer of the av7110 driver would become responsible for the API. Could that be an acceptable solution for everybody? Or do the ~10Kb of possible redundancy in the header hurt so much? The further deferring of the removal by a few years would be a kind of solution for me too. But I don't think it's a good one. Any other ideas? As a pragmatic user I'm interested in a solution to keep my cards running. Probably finding a compromise all parties can live with. I'm not interested in a lengthy discussion about APIs leading to nowhere. Regards Stefan 1. https://www.dwdl.de/magazin/88023/sdrepublik_deutschland_privatsender_mit_hdstrategie_gescheitert/? In German only, sorry.
