Bug 12013

Summary:	VMWare mouse integration does not work in Cauldron
Product:	Mageia	Reporter:	David Walser <luigiwalser>
Component:	RPM Packages	Assignee:	Colin Guthrie <mageia>
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	Normal	CC:	mageia, thierry.vignaud
Version:	Cauldron
Target Milestone:	---
Hardware:	All
OS:	Linux
Whiteboard:
Source RPM:	udev, x11-server	CVE:
Status comment:
Bug Depends on:	12022
Bug Blocks:
Attachments:	Xorg.0.log

Description David Walser 2013-12-16 15:00:49 CET

With Mageia 3 and earlier, when X was running in a VMWare VM, you could freely and seamlessly move the mouse between the guest and host.  In Cauldron, you have to Ctrl-Alt for the mouse to escape the VM, just as if you didn't have any VMWare-specific drivers installed.

Reproducible: 

Steps to Reproduce:

Comment 1 Thierry Vignaud 2013-12-17 13:30:56 CET

Works for me (using Mageia-4-beta2-x86_64-DVD.iso & VMware Player 5.0.1 build-894247)

Keywords: (none) => NEEDINFO

Comment 2 Thierry Vignaud 2013-12-17 13:55:54 CET

I meant it works for the installer.

Comment 3 Thierry Vignaud 2013-12-17 14:53:04 CET

But it doesn't after indeed...

Keywords: NEEDINFO => (none)
CC: (none) => mageia
Source RPM: x11-driver-input-vmmouse-13.0.0-3.mga4.src.rpm => udev, x11-server

Thierry Vignaud 2013-12-17 14:53:13 CET

Hardware: i586 => All

Comment 4 Thierry Vignaud 2013-12-17 14:55:36 CET

Created attachment 4633 [details]
Xorg.0.log

Colin, note the following:

(II) The server relies on udev to provide the list of input devices.
  If no devices become available, reconfigure udev or disable AutoAddDevices.
(WW) Hotplugging is on, devices using drivers 'kbd', 'mouse' or 'vmmouse' will be disabled.

vmmouse is configured in /etc/X11/xorg.conf but is ignored due to x11 server using udev. However the later doesn't make the server to load vmmouse...

Comment 5 Thierry Vignaud 2013-12-17 14:56:55 CET

We will need an udev rule like Debian:
http://www.slax.org/en/blog/21063-Mouse-grab-for-vmware.html

"So how to make Xorg use the vmmouse driver? The following udev rule makes the trick, and will be included in Slax:

 
ACTION=="add|change", ENV{ID_INPUT_MOUSE}=="?*", ATTRS{description}=="i8042 AUX port", KERNEL=="event[0-9]*", PROGRAM="/usr/bin/vmmouse_detect", ENV{ID_INPUT.tags}="vmmouse"

 This is in debian for very long time, but didn't make its way to the official xorg input driver"

Status: NEW => ASSIGNED
Assignee: thierry.vignaud => mageia

Comment 6 Colin Guthrie 2013-12-17 15:05:22 CET

Looks like it yeah. Or we could just add a dropin conf xorg snippet in /etc/X11/xorg.conf.d/ perhaps?

How does fedora do it? I don't really like the vmmouse_detect program bit. We may need to add that to the initrd too if an add/change event is not triggered later when userspace takes over...

Comment 7 Colin Guthrie 2013-12-17 15:10:11 CET

So it seems, that udev support just isn't compiled in our package.

A simple buildrequires should fix it.

Comment 8 Colin Guthrie 2013-12-17 15:12:38 CET

I've added the BR and submitted to cauldron. Lets see if it works OK without any initrd tweaks...

Thierry Vignaud 2013-12-17 15:21:18 CET

CC: (none) => thierry.vignaud
Depends on: (none) => 12022

Comment 9 Thierry Vignaud 2013-12-17 16:08:08 CET

It's not enough.
The udev rule is needed else it fails the same way

Comment 10 Colin Guthrie 2013-12-17 16:36:05 CET

But enabling udev support includes the udev rule (I think the comment about it not making it into the upstream package is outdated - it's clearly there!)... or am I missing something else?

http://svnweb.mageia.org/packages/cauldron/x11-driver-input-vmmouse/current/SPECS/x11-driver-input-vmmouse.spec?r1=558291&r2=558290&pathrev=558291

Comment 11 David Walser 2013-12-17 17:00:32 CET

I've installed x11-driver-input-vmmouse-13.0.0-4.mga4 and it works for me.

Comment 12 Thierry Vignaud 2013-12-17 18:20:31 CET

Strange. I'll try again tomorrow. Closing for now.

Status: ASSIGNED => RESOLVED
Resolution: (none) => FIXED

Comment 13 David Walser 2013-12-17 20:33:29 CET

You need to reboot after installing it before it'll actually work.

Thanks Thierry and Colin!

Comment 14 Thierry Vignaud 2013-12-17 21:25:15 CET

I'd reboot of course