Per Brad Hards's recommendation, I've built a newer (2.4.16) kernel.
The multimedia/Internet keys now generate input events. (Yay!)
However, they generate the wrong ones. (Boo!)
The following table lists the name of each key (as printed on the
keyboard), the keycode recorded by xev when using PS/2, and the keycode
recorded by xev when using USB:
Label PS/2 USB
----- ---- ---
Back 234 234
Forward 233 233
Stop 232 232
Refresh 231 121 (mismatch)
Search 229 229
Favorites 230 230
Web/Home 178 130 (mismatch)
Mail 236 236
Mute 160 166 (mismatch)
- (volume) 174
165
(mismatch)
+ (volume) 176
158
(mismatch)
Play/Pause 162 159 (mismatch)
Stop 164 151 (mismatch)
Prev Track 144 164 (mismatch)
Next Track 153 162 (mismatch)
Media 237 129 (mismatch)
My Computer 235 111
(mismatch)
Calculator 161 161
Sleep 223 227 (mismatch)
Observe that for more than half of these keys, USB yields a different
keycode than PS/2. Is that a bug? I don't know. I would have expected
the keycodes to remain the same, but perhaps that is a mistaken
assumption on my part.
If at all possible, though, I do believe it would be a good idea for the
keycodes to remain consistent. If one configures XFree86 to use the
"microsoftpro" XkbModel, then the PS/2 keycodes are all correctly mapped
to appropriate keysyms. For example, the Media key produces keycode 237
which is mapped to the XF86AudioMedia keysym. The USB keycode, 129, has
no keysym mapping.
I understand how to use xmodmap to establish new keycode-to-keysym
mappings. But when using a standard keyboard, it's really nice when the
existing XkbModel just works out of the box. One shouldn't different
XkbModels for different jacks.
Comments?
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-users
