Gerd Knorr wrote: > > This is a prof-of-concept port of a lirc infrared remote driver to the > 2.6 kernel input layer. Nice. > Not sure whenever it is a good idea to put the IR keys into the normal > keyboard queue. That means that the application with the keyboard focus > will also receive the IR input. Depending on how you are using lirc you > might call this a bug or a feature. IMHO it is less confusing for the > users if keyboard focus and IR focus are identical. But there might be > situations where it is useful that they are independant ... The DVB driver for WinTV Nexus-S (and similar cards) did that for quite a while now. We use it in conjunction with DirectFB (which reads events directly from /dev/input/eventX). I don't know how well it works for X. Here's a pointer for reference: - driver: linux-2.6.0-test2/drivers/media/dvb/ttpci/av7110_ir.c - keymap load utitlity: http://linuxtv.org/cgi-bin/cvsweb.cgi/DVB/apps/av7110_loadkeys/ AFAIK lirc allows one to control/configure which events get to which app, and how the app should react. This functionality is not replaced by this patch. lircd can die, but the lirc client lib should just read from /dev/input/event* instead from /dev/lircd. Putting IR events in keyboard queue might be asking for trouble. But then, we have some IR keyboards here which work fine ;-) The questions is how you can configure your system to route input events like you want it to. > #if 0 > [ 0x1e ] = KEY_FIXME, // reserved > [ 0x26 ] = KEY_FIXME, // minimize > [ 0x2e ] = KEY_FIXME, // fullscreen > #endif 2.6 has e.g. KEY_VENDOR, KEY_ZOOM, KEY_CYCLEWINDOWS, KEY_SCREEN Johannes
