X.org: two mice + xmodmap. How to?

Aysen · Apprentice Joined: 18 May 2005 Posts: 187 Location: Poland

Hi everyone!

I have just discovered that I still own a nice wireless mouse with a ball (supposed to work like two scrollwheels) that I completely forgot about and decided to try it out. The mouse works perfectly except one thing. My primary mouse is Logitech MX900 and it's set as the core pointer. Its thumb buttons generate codes 8 and 9, so I use xmodmap to make them 6 and 7 to work as back and forward. Now the weird thing - the new mouse's (MT1007) thumb buttons also generate codes 8 and 9 (checked with xmodmap pointer = default), but xmodmap doesn't change them. Without xmodmap it's like this (left column = generated code numbers, right column = buttons that generate them):
MX900:

placeholder · Advocate Joined: 07 Feb 2004 Posts: 2500

Perhaps you should try out evdev. Emerge it(not sure what it is called in Gentoo so run a search), and then set them up like so:

Aysen · Apprentice Joined: 18 May 2005 Posts: 187 Location: Poland

Thank you very much emma, with evdev my MX900 works perfectly (each button has its own code)

However, my main problem is still unsolved. What's more, I've discovered a few other issues.

Using evdev as the MX900 driver (like this:
Code:

Section "InputDevice"
Identifier "Logitech MX900"
Driver "evdev"
Option "Device" "/dev/input/event2"
EndSection
) breaks my keyboard's multimedia keys. They no longer generate any events in xev. I'll investigate this later (OK, I've found a few other threads about it... looks like it's unsolved yet).
evdev doesn't recognize MT1007's ball properly. The horizontal and vertical directions generate the same codes (4 and 5). The "mouse" driver recognizes them properly.

There seems to be a nasty bug in X.org - when both mice were using evdev, I unplugged one of the receivers and restarted X. It failed miserably and totally froze - no ALT-CTRL-BS, no ALT-CTRL-Fn, no nothing. The last thing in the log was:

Code:

(**) Option "CorePointer"
(**) Logitech MX900: Core Pointer
(**) Logitech MX900: Device: "/dev/input/event2"
(II) Logitech MX900: Found x and y relative axes
(II) Logitech MX900: Found mouse buttons
(II) Logitech MX900: Found keys
(II) Logitech MX900: Configuring as mouse
(II) Logitech MX900: Configuring as keyboard
(**) Option "SendCoreEvents"
(**) Mediatech MT1007: always reports core events
(**) Mediatech MT1007: Device: "/dev/input/event3"
(EE) Unable to open evdev device "/dev/input/event3".
(EE) PreInit failed for input device "Mediatech MT1007"
(II) UnloadModule: "evdev"

*** If unresolved symbols were reported above, they might not
*** be the reason for the server aborting.

Backtrace:
0: /usr/bin/X(xf86SigHandler+0x9e) [0x80baa3e]

Fatal server error:
Caught signal 11. Server aborting

Looks like it unloaded the evdev driver when it found out that the second mouse was not present, not caring that the first mouse was actually using the driver (those 2 EEs are not critical - when MX900 was using the evdev and MT1007 the mouse driver, those EEs were also present, but the evdev driver wasn't unloaded and X was running fine). I guess I'll file a bug (if my interpretation of the error is correct).

xmodmap definitely works only for the core pointer - when I made MT1007 the core pointer, its thumb buttons got mapped as 6 and 7 and MX900's stayed as 8 and 9, but when MX900 was core then it was exactly the opposite. Does anyone know a way to suprass this limitation?

Hardware details if anyone's interested:

The mouse (MX900) and keyboard (Cordless Elite (uh, what a name... model number Y-RP20, anyway)) are from the Logitech Cordless Desktop MX for Bluetooth.
The Mediatech MT1007 mouse identifies itself as Acrox RF Ball Mouse in sysfs. Media-Tech's details.

I'll be thankful if someone helps me with the xmodmap problem.
In the meantime, I'll look into the multimedia keys problem.

Regards!