On Sat, 8 Sep 2001, dman wrote:
> Hmm, nope I don't have /dev/e*. I have /dev/input/mice and
> /dev/input/mouse0 because I am using devfs and a usb mouse.
I've used linux-edev sucessfully with USB mice before.
> According to the linux-evdev section of the doc (just read that) the
> events are from /dev/input/eventX. The only files in /dev/input are
> 'mouse0' and 'mice'. Do I need to change some kernel config to enable
> the event devices?
Yes, this one, under input core support:
qqqqqqqqqqqqqqqqqqqq Event interface support (NEW) qqqqqqqqqqqqqqqqqqqqqk
x CONFIG_INPUT_EVDEV: x
x x
x Say Y here if you want your USB or ADB HID device events be accessible x
x under char device 13:64+ - /dev/input/eventX in a generic way. x
x This is the future ... x
:-)
> I don't know if that protocol is missing (would I get an error
> message?) but it doesn't seem to be working for me :-(.
Hmm, well there is this in the changelog:
* input/mouse/: input.c (1.20), mouse.h (1.6), packetparse.c
(1.19): Restructured packet parser so we can fall back to another
parser when the device can't be initialized. Currently used by
mmanps2 and imps2 parsers. Changed button mask in parse_imps2()
from 0x07 to 0x0f. This allows four buttons as used by the
Linux USB-mouse driver, and hopefully it shouldn't break anything
for three-button mice. Fixed fatal bug in parse_imps2() - We
should return 4, not 3. Added new parser type "lnxusb" for
USB-mice under Linux. It uses the imps2 parser, but doesn't require
any initialization.
--
Brian