Hi all, Quoting myself: > Is there any objection to me dropping rivatv detection from > sensors-detect? My reasons are: > > * rivatv is still not in mainline after 3 years. I wonder if it will > ever be. > > * rivatv is only one of the several drivers which can provide I2C/SMBus > access to nVidia adapters. nvidiafb and rivafb do it as well (although > I did not check if access to the same busses was provided) and they > are in mainline. And nVidia's proprietary X driver does also in recent > versions, even though it won't work at the moment due to improperly > set i2c functionality flags (and we won't provide support for it, but > that's a different point.) > > * Most PCI drivers will autoload in recent Linux 2.6 kernels (in > particular nvidiafb and rivafb do). So I'd rather let the user pick the > driver he/she wants. sensors-detect will scan all available busses > anyway, whether they were detected or not. > > * There are many many different nVidia graphics adapters. We already > have 22 entries, representing 20% of the total list of known PCI > adapters. Given that the last ones were added over 2 years ago, the list > is certainly incomplete. I don't even know which devices are supported > by rivatv (or the other drivers mentioned above), and I don't want to > spend time tracking this. > > So unless someone finds a good reason I missed, I plan to remove all > references to rivatv from sensors-detect soon. Done. -- Jean Delvare
