Switching between different systems with a KVM switch makes Xinput lose connection.

[Tobin Davis]

Binary package hint: xorg

I am using a Trendnet TM408 VGA & USB 4-port switch to control 4 systems from one keyboard, mouse, and VGA monitor. After switching multiple times between systems, X will fail to respond on main desktop system. Main desktop runs 24/7 with minimal reboots (updates and occasional hangs like this). One machine is running a console only, the other two are used for testing and are reimaged frequently.

System is fully responsive via ssh, but stopping kdm and restarting fails to properly reload nVidia driver, thus reboot is required. Error messages in Xorg.0.log indicate loss of communication between Xinput drivers and kernel usb drivers. dmesg log doesn't show any abnormal issue at kernel level.

ProblemType: Bug
Architecture: i386
Date: Thu Nov 5 10:50:50 2009
DistroRelease: Ubuntu 9.10
NonfreeKernelModules: nvidia
Package: xserver-xorg-input-all 1:7.4+3ubuntu7
ProcCmdLine: root=UUID=b14ec690-ed3a-4ed7-a694-c11d02196e01 ro quiet splash
ProcEnviron:
 LANGUAGE=
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-14.48-generic
RelatedPackageVersions:
 xserver-xorg 1:7.4+3ubuntu7
 libgl1-mesa-glx 7.6.0-1ubuntu4
 libdrm2 2.4.14-1ubuntu1
 xserver-xorg-video-intel 2:2.9.0-1ubuntu2
 xserver-xorg-video-ati 1:6.12.99+git20090929.7968e1fb-0ubuntu1
SourcePackage: xorg
Uname: Linux 2.6.31-14-generic i686
XsessionErrors: (polkit-gnome-authentication-agent-1:3141): GLib-CRITICAL **: g_once_init_leave: assertion `initialization_value != 0' failed
dmi.bios.date: 10/02/2007
dmi.bios.vendor: Intel Corp.
dmi.bios.version: DPP3510J.86A.0293.2007.1002.1519
dmi.board.asset.tag: Base Board Asset Tag
dmi.board.name: DP35DP
dmi.board.vendor: Intel Corporation
dmi.board.version: AAD81073-207
dmi.chassis.type: 2
dmi.modalias: dmi:bvnIntelCorp.:bvrDPP3510J.86A.0293.2007.1002.1519:bd10/02/2007:svn:pn:pvr:rvnIntelCorporation:rnDP35DP:rvrAAD81073-207:cvn:ct2:cvr:
fglrx: Not loaded
system:
 distro: Ubuntu
 architecture: i686kernel: 2.6.31-14-generic

[Bryce Harrington (bryce)]

(II) config/hal: Adding input device No brand 4 Port KVMSwicther
(**) No brand 4 Port KVMSwicther: always reports core events
(**) No brand 4 Port KVMSwicther: Device: "/dev/input/event6"
(II) No brand 4 Port KVMSwicther: Found keys
(II) No brand 4 Port KVMSwicther: Configuring as keyboard
(II) XINPUT: Adding extended input device "No brand 4 Port KVMSwicther" (type: KEYBOARD)
(**) Option "xkb_rules" "evdev"
(**) Option "xkb_model" "pc105"
(**) Option "xkb_layout" "us"
(EE) BTC USB Multimedia Keyboard: Failed to reopen device after 10 attempts.
(EE) BTC USB Multimedia Keyboard: Failed to reopen device after 10 attempts.
(EE) Logitech USB Optical Mouse: Failed to reopen device after 10 attempts.
(EE) No brand 4 Port KVMSwicther: Failed to reopen device after 10 attempts.

Looks like a bug in -evdev that's been fixed upstream.

[Bryce Harrington (bryce)]

PPA with the patch I think will fix it.

https://edge.launchpad.net/~bryceharrington/+archive/blue

This is a pretty significant patch that chunks out a whole feature (well, "feature") so it may be too much to digest for an SRU. But maybe if other bugs come to light that this fixes we could evaluate it.

[Bryce Harrington (bryce)]

Assigning myself to make sure this gets into lucid one way or another

[Bryce Harrington (bryce)]

sorry, patch didn't build due to it being against a newer -evdev codebase than is in karmic. the patch will have to be rebased against karmic's -evdev.

[Tobin Davis (gruemaster)]

Here's the patch rebased against xserver-xorg-input-evdev_2.2.5-1ubuntu6. Will install and start testing to see if it fixes this issue.

[Bryce Harrington (bryce)]

Hi Tobin,

Thanks for posting that. Have you found that it solved the issue for you?

[Tobin Davis (gruemaster)]

Need to test more. Will know by Monday if it was effective.

[Tobin Davis (gruemaster)]

As of Monday, no issues. Will spend Tuesday changing systems more often so if the issue still persists, it will soon be noticeable.

[Timo Aaltonen (tjaalton)]

Lucid has 2.3.2 which includes this fix.

[Timo Aaltonen (tjaalton)]

sigh, no it doesn't, reopening.

[Timo Aaltonen (tjaalton)]

Closing again, this is in -evdev since maverick.