Odd mouse issue with new xorg-server

[Mel Flynn]

On 26-4-2012 4:30, Kevin Oberman wrote:
> I hit a weird one with the new xorg-server. I have previously been
> running the xorg-devel from about 4or 5 months ago, so it's something
> that has happened since then.
> 
> After updating to the new xorg, my mouse was dead. I looked at the log
> and found:
> (II) config/hal: Adding input device PS/2 Mouse
> (II) LoadModule: "mouse"
> (II) Loading /usr/local/lib/xorg/modules/input/mouse_drv.so
> (II) Module mouse: vendor="X.Org Foundation"
>         compiled for 1.7.7, module version = 1.7.1
>         Module class: X.Org XInput Driver
>         ABI class: X.Org XInput driver, version 7.0
> (WW) PS/2 Mouse: No Device specified, looking for one...
> (II) PS/2 Mouse: Setting Device option to "/dev/psm0"
> (--) PS/2 Mouse: Device: "/dev/psm0"
> (==) PS/2 Mouse: Protocol: "Auto"
> (**) PS/2 Mouse: always reports core events
> (**) Option "Device" "/dev/psm0"
> (EE) xf86OpenSerial: Cannot open device /dev/psm0
>         Device busy.
> (EE) PS/2 Mouse: cannot open input device
> (II) UnloadModule: "mouse"
> (EE) PreInit returned NULL for "PS/2 Mouse"
> (EE) config/hal: NewInputDeviceRequest failed (8)
> 
> But I have the following in xorg.conf:
> Section "InputDevice"
>         Identifier  "Mouse0"
>         Driver      "mouse"
>         Option      "Protocol" "auto"
>         Option      "Device" "/dev/sysmouse"
>         Option      "ZAxisMapping" "4 5 6 7"
> EndSection
> 
> If I simply remove my xorg.conf, it works fine. I tried 'Xorg
> -configure', but that failed to write an output file.
> 
> Any idea what is going on? It is OK on this system as it is fine with
> no config file, but my two-headed work system is likely to require a
> config, so I am worried about hitting it there.

+Me too and analysis:
hald provides the mouse, however the mouse_drv.so randomly probes
protocols. There's no way to hint at the protocol in the xorg.conf
(bug!) because the entire input sections are skipped when AutoAddDevices
(hald) is in effect.
I'm saying random, cause there's really no set order in which protocols
are tried and I may have to startx several times for the driver to
settle on sysmouse.
Disabling moused mitigates the issue, since the device node won't be busy.
Real solution is to alter mouse_drv's scanning order on FreeBSD to be
sysmouse first, ps/2 second, then the rest and serial last. This should
do the right thing for most of the installations. Secondly, there needs
to be a way to give protocol hints in xorg.conf in the cases that
aforementioned logic does not work.
Thirdly mousedrv should take a look at hald's advice. Opening a serial
mouse driver on something classified as 'info.product = PS/2 Mouse' does
not strike me as logical. Also, the input.device = /dev/sysmouse, should
trigger SysMouse protocol.

-- 
Mel