Hi,
On 30-01-15 23:35, Dmitry Torokhov wrote:
Hi Hans,
On Fri, Jan 23, 2015 at 11:24:50AM +0100, Hans de Goede wrote:
Some mice report buttons with a MSC_SCAN of 589833 this maps to a button code
of 0x118, add a BTN_SIDE2 define for this to input.h
BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=1184352
Signed-off-by: Hans de Goede <hdegoede@xxxxxxxxxx>
Looking at Logitech website it seems that one of teh buttons is marked
"zoom", so why don't we use KEY_ZOOM for it?
Because we don't do any smart mapping on HID mouse buttons, see:
drivers/hid/hid-input.c: hidinput_configure_usage() where we do:
switch (usage->hid & HID_USAGE_PAGE) {
...
case HID_UP_BUTTON:
code = ((usage->hid - 1) & HID_USAGE);
switch (field->application) {
case HID_GD_MOUSE:
case HID_GD_POINTER: code += BTN_MOUSE; break;
...
}
map_key(code);
break;
...
}
So we basically assign any mouse buttons to evdev code 0x110 - 0x11f,
which by itself is the best the kernel can do really as the HUT is
not helpful here, simply saying the following on these buttons:
Button 1 (primary/trigger)
Button 2 (secondary)
Button 3 (tertiary)
Button 4
...
Button 65535
IOW it gives no guidance which button is what on mouse other then for
left / right / middle. This e.g. can clearly be seen on a cheap mouse
with prev / next buttons I have where the buttons send hut button code
4 & 5 which gets translated to BTN_SIDE / BTN_EXTRA
Given how these buttons are specified in the HUT, the kernel should
really just be sending BTN_0 (for hut button 4) - BTN_9 for them, as they
have no clearly assigned meaning.
It is too late to change this now though, as that would be a clear ABI
break.
Also note that given that the kernel just sends code + BTN_MOUSE, it is
already sending 0x118 (and potentially higher codes), and this is already
picked up by e.g. Xorg, all I'm trying to do here is assign a label to the
codes already being send.
Maybe we should do something like this:
--- a/include/uapi/linux/input.h
+++ b/include/uapi/linux/input.h
@@ -494,10 +494,23 @@ struct input_keymap_entry {
#define BTN_RIGHT 0x111
#define BTN_MIDDLE 0x112
#define BTN_SIDE 0x113
+#define BTN_EXTRA0 0x113 /* Exact usage unknown on HID devices */
#define BTN_EXTRA 0x114
+#define BTN_EXTRA1 0x114 /* Exact usage unknown on HID devices */
#define BTN_FORWARD 0x115
+#define BTN_EXTRA2 0x115 /* Exact usage unknown on HID devices */
#define BTN_BACK 0x116
+#define BTN_EXTRA3 0x116 /* Exact usage unknown on HID devices */
#define BTN_TASK 0x117
+#define BTN_EXTRA4 0x117 /* Exact usage unknown on HID devices */
+#define BTN_EXTRA5 0x118
+#define BTN_EXTRA6 0x119
+#define BTN_EXTRA7 0x11a
+#define BTN_EXTRA8 0x11b
+#define BTN_EXTRA9 0x11c
+#define BTN_EXTRA10 0x11d
+#define BTN_EXTRA11 0x11e
+#define BTN_EXTRA12 0x11f
#define BTN_JOYSTICK 0x120
#define BTN_TRIGGER 0x120
Note that we could use udev hwdb entries to fix up things, I could e.g.
fix the prev / next buttons on my mouse to send the right codes by remapping
them through hwdb. I'm a bit divided on this, as this is going to be a bit of
whack a mole. This is somewhat orthogonal though, because we should at least
give some symbolic name to codes which we are actually emitting now already,
and will keep emitting for new mice without a hwdb entry.
Regards,
Hans
---
include/uapi/linux/input.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/include/uapi/linux/input.h b/include/uapi/linux/input.h
index a1d7e93..27b0f6a 100644
--- a/include/uapi/linux/input.h
+++ b/include/uapi/linux/input.h
@@ -498,6 +498,7 @@ struct input_keymap_entry {
#define BTN_FORWARD 0x115
#define BTN_BACK 0x116
#define BTN_TASK 0x117
+#define BTN_SIDE2 0x118
#define BTN_JOYSTICK 0x120
#define BTN_TRIGGER 0x120
--
2.1.0
Thanks.
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
