On Friday 03 October 2003 14:12, Holger Waechtler wrote: > Oliver Endriss wrote: > > On Friday 03 October 2003 12:25, Holger Waechtler wrote: > > > >>Oliver Endriss wrote: > >> > >>>On Thursday 02 October 2003 15:21, Rene Bartsch wrote: > >>> > >>> > >>>>as it's quite a work to generate a key-code map, I'd like to know if > >>>>it's possible to route/catch the raw IR-codes directly to/from the > >>>>input-device or a socket with the current DVB-driver. > >>> > >>> > >>>Currently, there is no such interface. > >>>The driver forwards the (translated) key codes to the input driver. > >>> > >>>A keymap is required, because > >>>- the protocol type must be specified > >>>- the driver requires a keymap for the translation table > >> > >>as soon you know the protocol you can load a dummy keymap which > >>translates key events 1:1. > > > > > > That's exactly what I'm going to do now. > > There are two protocols (rc5/rcmm) with two inversion inversion settings > > (on/off), i.e. 4 different settings. So I prompt the user to hold down any > > key while cycling through the 4 possibilities. > > :) sounds good.
Unfortunately, this works only if the application is running as *root*, because permissions of /proc/av7110_ir are nailed to 0644. They cannot be modified without patching the driver. :-( IMHO, the application which has opened the dvb devices in r/w mode, should be allowed to write to /proc/av7110_ir, too. While it's not a big security hole, I don't like patching permissions of av7110_ir to 0666. ;-) How could this be solved in a clean way? Are there any plans to change the keymap interface for API-4? CU, Oliver -- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe linux-dvb" as subject.
