Computers 'n Stuff

Cleaning up Windows XP

2008-05-20T03:22:00.000-07:00

Windows XP needs some maintenance from time to time when performance drops too far. Here are the major things I check and some utilites I use to fix those issues.

Check for Malware and remove if neccesary (Freeware: Avira Private edition and SpyBotS&D, HighJackThis)
Remove all not requred Software froma the autostart folder (Windows tools: msconfig)
Remove unnessacary data form your harddrive. Deinstall Sofware you dont require, Remove restore points, old windows update files and clean up temorary files (Freeware: CClean, EmpTemp; Windows tools: Cleanmgr)
Defrag harddrive, (Windows tools: defrag). Consider doing this in safe mode so more files can be moved. Use PageDefrag (Sysinternals) for offline Defragmentation of Systemfiles (Registry hives, Pagefiles and so on). Chek master file table (MFT) fragmentation. Few fragments (e.g. 3) dont matter. Many fragments ( e.g 100+) require MFT defragmentation for no feeware tools are avialable (AFAIK) DiskKeeper Professional can do the trick.
Check for disk errors. Usually an NTFS drive that is not flagged dirty by Windows will not require checking. I use chkdsk /f to invoke autochk. exe at boot time to fix file system corruption on the system, however, since I feel that many file system errors will compromise hard disk performance.
Fix registry. (consider creating a restore point before) (Freeware: CClean, Wiser registry cleaner, RegClean) I feel that this makes the system usually more resposnive if you remove serveral hundrets of wrong entries.
Defragment internal structure of the Registry. (Ntregopt, Auslogistic registry defrag). This may decrease the size of the registry which decrease boot time.
Use HDtune to check disk performance. If disk performance is poor, check whether drivers are installed correctly in the device manager. On Intel chipset based mainboards the inel application accelerater may improve Disk performance.
Rebuild icon database. This will improve poor loading time for icons when this is caused by a corrupted icon database. (TweakUI[windowsXP powertoys])
Remove "Ghost drivers": SET DEVMGR_SHOW_NONPRESENT_DEVICES=1 (Enable view hidden devices environment varibale) fire up device manager and check view>show hidden devices. Deinstall hidden devices that are not required any more.
Bootvis. Downlaod and run Bootvis (Microsoft) to optimize boot time.
Check for error messages in the management console (Events>applications) and debug if applicable.
remove all unused protocoll from your network connections. In particular, deinstall microsoft network client and server or QoS if you are not using it. In the canse for micorsoft network, this probably increases security since your computer can probably not be accesed via SMB protocoll and in addition it may increase network throughput. In particular QoS has been suggested to reserve up to 20% of bandwidth for priorititized applications. However, you'll probably notice a difference only if you transfer large files over high speed LAN (100Mbit+) a lot.

Computer won't Boot after Reinstalling Windows XP on a Linux/Windows Dual Boot System: Another Windows XP Bug

2008-04-07T05:27:00.000-07:00

Using a dual boot system is sometimes a little tedious. In particular since the windows installer is keen to screw up any operating systems that are installed on your hard drive besides windows. Here is what happened to me after I had to re install a broken windows instatllation on a dual boot system.

If you mess around too much with windows it will sooner or later cease to boot. Usually it is sufficient to boot with the windows installation disk and to choose the repair installation.
However, in my case repair installation did not help, and I had to do a clean install. Yet, after finishing the Windows XP SP 2 installation my computer was completely screwed up. After the Windows XP isntall it bacame impossible to boot any operating system, neither the old windows, the old linux nor the new windows. Usually in this cases it helps to boot the computer from a rescue boot disk and then boot the required partition. But in this case it didnt help at all. Also the repair utility featured on the Linux installation disk did not help at all. So I thought my harddrive would be lost.

However, after taking a look at the master boot record (MBR) of the hard drive I figured out what the problem was:

The windows xp installer flagged both partitions of my harddrive as active after the clean install. Because after the linux installation I used GRUB as a boot loader, the linux partition was set as active in the MBR.
So the Windows XP installer set the windwos partition as the active partition and leaving the active flag for the linux partition in place. Thus, after finishing the Windows XP reinstallation my MBR ended up with two partitions flaged as active. This is a condition which should never be present in a master boot record.
To fix the MPR you have to edit it to flag the only that partition active you want to boot from, e.g. the partition containig the GRUB boot-loader. Caution: Never use the fixmbr command from the recovery console to fix the MBR. This command will overwrite your partition table and screw up your entire hard drive.
I consider this as another nasty Windows XP bug. Considering the Microsoft marketing strategy it is conseivable that Microsoft intentionally put this bug into their installation routine to deliberately screw up systems that hold other than microsoft operating systems on their harddrive.

Using wget to convert a server side CMS into Offline/client site/desktop CMS

2007-06-20T13:44:00.000-07:00

I have a little bit of webspace for hosting a private homepage, but the provider supports pretty much nothing, no root acces not, CMS not php or mysql. Here are some thoughts how to use a client side cms to build static html.

The Idea is as follows. I want to use my webspace to host a little private host site. But the privider supports no php cgi or mysql. He doesent have content management or wikis or the like. So the only option is static html. But for linux there are not so many tools that help you in generating static html websites. Exept may be html preprocessors. However, today the world is using php so why not using php as the html preprocessor. So I ended up hosting MySQL Apache and PHP locally on the client side and then ripped my own site using wget. I use OpenSuse 10.2 and all this runs out of the box.


wget --mirror -r –p -k --html-extension –-convert-links –P /home/ich/Downloads/mirror/mdiawiki http://localhost:8080/mylinux

This way i can use server side includes in convert them to static html. It works also for zope or plone generated sites but setting up a zope site is not as easy as I thought. So I think its easier to use includes to seperate the index and template from the content so that I dont need to edit everythin by hand.

Microsoft "theBad" windows vs. OpenSource Linux: Not eveything is lost

2007-06-15T12:10:00.000-07:00

The Bad or like thy like to call themselves, microsoft, is tyining to bully Linux. Howerver, not everything is lost. Research reveals that all companies supporting Linux have a greater financial power than Microsoft.
The new goes that a couple of companies supporting Linux are depating wheather they should form a alliance against Microsoft which is threatening the open SourcCommunity with copyright infringement law sues. Should the participants of meeting choose to form a alliance to oppose "the Bad" which continues to gain momentum in the open source market, they will have almost 9 times tha market power than microsoft (397 billion USD compared to 44.3 billion). But what could the strategy against thue new bottem up attack of microft be?

Installation of Moneyplex 2007 with Rainer SCT cyberjack in Linux in OpenSuse 10.2 64Bit

2007-06-15T08:10:00.001-07:00

Moneyplex is a monymanagement software supporting Smart Card identification for online Banking. Here I write what you have to keep in mind if you want to install it.

The problem with Moneyplex 2007 is that is still 32Bit. But compared to earlyer versions of OpenSuse and Moneyplex its now pretty easy to get Moneyplex running in a 64bit environment. No need to use chroot dirs or virual mashines. Just install the Moneyples software using the installer and download the updates.

Now get the 32 bit driver for the smart card reader Reiner SCT Cyperjack for Suse 10.2
install it with

# rpm -i [drivername.rpm]

now connect the Reader and do

# modprobe cyberjack

and check with

# lsmod | grep cyberjack

wheather driver has loaded.

Now add your user account to the group cyberjack e.g. using yast.

In moneyplex you have to use the CTAPI driver you'll find in
/usr/lib/readers/

and everything should work fine now.

Is Microsoft racketeering the Open Source community?

2007-06-12T10:11:00.000-07:00

Microsoft is asking people to pay them for patents, but they won't say which ones. If a guy walks into a shop and says: 'It's an unsafe neighborhood, why don't you pay me 20 bucks and I'll make sure you're okay,' that's illegal. It's racketeering. What Microsoft is doing with intellectual property is exactly the same. It's a great company and I have great admiration for it, but this was not a well considered position," said Shuttleworth.

From What the Microsoft/Xandros deal means for Linux

This is a pretty funny comparison but I can't belive that the plans of Microsoft concerning the opens source community are that simple. I can't tell exactly what it is but I believe it 's a part of a long term evil strategy of Microsoft to destoy the open source community. There is enough evidence that Open Source has appeard already on the radar of Microsoft constanly detecting potential competitors. I has always been and it will allways be the goal of microsoft to defeat all competitiors that appear on their radar. Why should it be any different with open source? There is no obvious reason why Microsoft shoulde start to cooperate with some kind of sofware project that competes with their market. I believe its not the plan of microsoft just to earn some money I blelieve this is a strategie to attack a distributed target like the open source community. I belive its the long term goal of microsoft to infect the open source with closed source payware. These companes like Xandros, Novell and LG will serve as Trojan horses and virus droppers for Microsoft. I will be the goal of microsoft to sneak licensed technologies into the code of linux so that they can take leagl actions against any user of Linux and open source. The more companies Microsoft buys the more companies will be allowed to use this code and the more companies will distribute the infected open source Linux which will be something like Windux or Lindows that will compete with the pure opensource linux therby reducing the market share of which.

Building a Xen aware Nvidia grafic card driver: 6800GS SLI in Xen under OpenSuse10.2 64Bit

2007-06-11T03:39:00.000-07:00

Xen is a highly professional Open Source virtualization technology which promises the least performance loss. Heres how I got the Closed Source Nvidia driver going unde OpenSuse-Xen 64bit AMD X2.

3D acceleration and SLI mode in particular are only supported by the closed source nvidia drivers. However, Xen requires a modified "Xen-Aware" kernel to run in a pravirtualized environment. Only the most recent CPUs support (SVM for AMD)fully virtulized virtual mashines that run without kernel modifications. On many websites you'll find the infromation that blah blah blah X2 CPUs from AMD would support full virtualization (SVM) blah blah. But this is wrong. Don't be mistaken only the AM2 version of X2 processors support SVM. If its not AM2 its not SVM. So don't make the same mistake like me and get a socket 939 CPU.

The Nvidia driver does not support Xen Kernels and its closed source, so you can't modifie it yourself. But not everything is lost. The nvidia driver uses a kernel module as a compatibility layer to load the binary driver. This kernel module is open source, so in theory it should be possible to modify the kernel module so that it works with Xen. And belive it or not some brave guys really hacked the kernel module to support xen-kernels.

So I had to get the hypervisor and the default-xen kernel of opoen suse 10.2 running. This is pretty easy you can use Yast to do the installation stuff (isntall hypervisor, install default-xen and have it change the bootloader settings) And of course you'll need all the devel packeges to compile the kernels.

Now boot with the default kernel and install install the binary nvidia drivers you are going to patch later.

sh NVIDIA-Linux-x86-1.0-9746-pkg2.run --x-module-path=/usr/lib64/xorg/modules/ -q

The version of the compatibility layer and the binaries have to match. the nvidia install script will do the job for you. However, it will refuse to run in a paravirualized environment. So for the Xen-kernel, be prepared to do everything manually.

Boot into the dom0 domain. you'll only have the console available because the xserver won't work. without the propreitary nvidiea driver with the GPUs in SLI mode.


init 3 # stops Xserver and swiches to runlevel 3

make cloneconfig
make prepare
wget NVIDIA_xen-9746.patch.gz
wget http://us.download.nvidia.com/XFree8...-9746-pkg1.run
sh NVIDIA-Linux-x86-1.0-9746-pkg2.run -x # extract sources
cd NVIDIA-Linux-x86-1.0-9746-pkg2/usr/src/nv
zcat ../../../../NVIDIA_xen-9746.patch.gz | patch -p4 #patch sources

Sombody stated you have to modify usr/src/nv/os-interface.c line 1288 modified to:

#if 0
to avoid compiling erroers. but this was not true for me. The driver compiled fine without any modifications. So first try without modification.


make SYSSRC=/usr/src/linux module # make nvidia kernel module
cp nvidia.ko /lib/modules/`uname -r`/updates/nvidia.ko
depmod -a
modprobe nvidia
init 5  # when successful enter runlevel 5 to start x server

Now the driver should load.
I also read this command to install the modules:
install -D -o root -g root -m 0644 nvidia.ko /lib/modules/`uname -r`/updates/nvidia.ko
But for me copying to the /udate folder of the kernel module folder was sufficient.
However my 3D performace was poor (1fps with glxgears)
So i'll try some older patches and see what happens.
The other patch I tried did not any better. It turned out they dont support SLI. If you turn of SLI by
nvidia-xconfig --no-sli
3D work descent (12800 fps for glxgears) which is about the same than without Xen.
Links to Nvidia kernel module Xen patches
another Nvidia xen patch howto for open suse

Java directory structure in OpenSuse 10.2 Linux

2007-06-09T07:46:00.001-07:00

Here is the directory structure of Java in OpenSuse 10.2 as a refereence.

# which java
/usr/bin/java
/usr/bin/java -> /etc/alternatives/java*
/etc/alternatives/java -> /usr/lib64/jvm/jre1.6.0_01/bin/java

this means the 64 bit java executable lives in /usr/lib64/jvm/[version]/bin/java. To enable the java version as a defualt adjust the link /etc/alternatives/java to point to the desired java version.

/etc/alternatives/jre -> /usr/lib64/jvm/[version]/

this is the link to your default jre. adust as desired.

/etc/alternatives/[version] -> /usr/lib64/jvm/[version]/
link pointing from alternatives to the real dir of the version. The real version can be a link too. if the versions are compatible e.g.
/usr/lib64/jvm/[version1] -> java-1.5.0-sun-1.5.0_update10/jre/

thats where the jar live

/usr/share/java

needed directorys but no content ...
/usr/shar/[version]
/usr/lib64/[version]

No firefox java plugin 64bit available exept Blackdown but couldn't get it working exept with <=32megs heap.. strange. nspluginwrapper doesent work with libjavaplugin_oj.so since it use open java OJAPI and not NAPI. Interestingly, Konquerer has its own java plugin that works put is incopatible with firefox though...

browser-blugins live in /usr/lib64/browser-plugins
or /usr/lib/browser-plugins
respectively. Thats the default plugin dir for firefox if you use the suse startupscript for firefox.

Enable SLI in Linux: My Xorg.conf as an example

2007-06-09T07:16:00.000-07:00

The SLI mode was a little harder to enable. Here is how i got it working

Install the driver according to the nvidia howto.


sax2 -r -m 0=nvidia doesent work with SLI setup

use


sax2 -r -c0 -m0=nvidia

instead. assuming that your disply is hooked up to card 0. Else you can use Card 1 but card 1 will not work for SLI

Now here was that catch: when I wanted to enable SLI GPU 0 as detectd by

nvidia-settings

needs to be the same as chip0 as reported by


sax2 -p

the can be identified by ther pci adress. And the display has to be hooked up to card 0 too. else the display will be scrambled when you starup the display.

I had to change the Init Disply first to PEG(slot2) to get this accomplished.

then you type


nvidiea-xconfig --sli=on

and your set. It si imortant that the frst available Xscreen is the screen with you monitor hooked up to it if have more that one configured in you xorg.config.

Heres the sections relevant to xscreen config of xorg.conf as a reference with sli enabeled:


Section "Module"
Load         "dbe"
Load         "type1"
Load         "freetype"
Load         "extmod"
Load         "glx"
Load         "v4l"
EndSection


Section "Monitor"
Option       "CalcAlgorithm" "XServerPool"
DisplaySize  338 270
HorizSync    30-82
Identifier   "Monitor[0]"
ModelName    "ACER AL1714"
Option       "DPMS"
VendorName   "ACR"
VertRefresh  43-75
UseModes     "Modes[0]"
EndSection

Section "Device"
BoardName    "GeForce 6800 GS"
BusID        "4:0:0"
Driver       "nvidia"
Identifier   "Device[0]"
VendorName   "NVIDIA"
EndSection


Section "Screen"
DefaultDepth 24
SubSection "Display"
  Depth      15
  Modes      "1280x1024" "1152x864" "1024x768" "800x600"
EndSubSection
SubSection "Display"
  Depth      16
  Modes      "1280x1024" "1152x864" "1024x768" "800x600"
EndSubSection
SubSection "Display"
  Depth      24
  Modes      "1280x1024" "1152x864" "1024x768" "800x600"
EndSubSection
SubSection "Display"
  Depth      8
  Modes      "1280x1024" "1152x864" "1024x768" "800x600"
EndSubSection
Device       "Device[0]"
Identifier   "Screen[0]"
Monitor      "Monitor[0]"
Option         "SLI" "on"
EndSection


Section "ServerLayout"
Identifier   "Layout[all]"
InputDevice  "Keyboard[0]" "CoreKeyboard"
InputDevice  "Mouse[1]" "CorePointer"
Screen       "Screen[0]"
EndSection


Section "DRI"
  Group      "video"
  Mode       0660
EndSection

Nvidie Linux Driver README

Quake 4: enable SLI in linux

2007-06-09T07:05:00.000-07:00

Quake 4 also runs faster in SLI mode in Linux. However, it needs to configured properly to increase performance. Here is my Quake starup script with SMP enableled.


#!/bin/sh
# Needed to make symlinks/shortcuts work.
# the binaries must run with correct working directory
cd "/usr/local/games/quake4"
export __GL_DOOM3=1
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:."
exec ./quake4smp.x86 "$@"
export __GL_DOOM3=1

important is to add export __GL_DOOM3=1 to improve SLI performance

Enable SMP for Quake 4 Linux Versin

2007-06-05T15:06:00.001-07:00

Finaly I got myself an X2 core. Now I urgently need to test ist performance using a slick sesiion of Quake for. However SMP support needs to be enabeled in Quake 4 firset. And now lets rock

clipped from www.phoronix.com

To enable Symmetric
Multi-Processing support, simply run quake4-smp rather than quake4.
The option remains dimmed for the multi CPU/core area within the advanced settings
of the game.

My .xbindkeysrc.scm configruation file I use with the MX1000 lasermouse (for now)

2007-06-03T13:25:00.000-07:00

This is my .xbindkeysrc I'm using with my mx1000 lasermouse. It's quite useful. with firefox. The thumbkeys can scroll in history and if you press the appbutton and use the scrollwheel at the same time you scroll through tabs.. If the mouse is setup properly it should be possible to do horizontal scrolling by the tilting th scroll wheel. However Firefox needs some adjustments for theis to function since he will scroll through history when the horizontal scroll buttons are pressed. The appbutton pushed and released alone will toggle the last active app.

The the Guru repository xbindkey suse rpm are compiled without guile support.To use xbindkeys it has to be compiled with guile support. The Guru repository does provide a xbindkeys-1.8.2-1-src.rpm that can be compiled with :

# rpmbuild --with guile --rebuild sourcepackage

to provide guile support. where sourcepackgage has to be replaced by the name of the sourcpackage, in this matter by xbindkeys-1.8.2-1-src.rpm. You will need to install the guile-devel package in order to compile xbindkeys in this way however. The compild RPM will show up in /usr/src/packages/RPMS/`architecture` where architecture has to corrospond your hardware architechture, in my case x86_64-


;;   This configuration is guile based.
;;   http://www.gnu.org/software/guile/guile.html
;;   This config script us supposed to live in the homedirectory.
;;   This was a not so quick but probably yet dirty hack by Vee Lee.(I know nothing about coding but at least it works)
;;   You'll need xmacroplay and xbindkeys with -guile support compiled for this to work..
;;   This is heavily based on the domoscripts (http://hocwp.free.fr/xbindkeys/xbindkeysrc-combo.scm.html)
;;   of the author of xbindkeys (hocwp at free dot fr) Many thanks and keep up the good work.
;;   It assigns keybindings to the scroll wheel  on the fly after the appwindow button is pressed on the mx1000 lasermous!

define (display-n str)
"Display a string then newline"
(display str)
(newline))

(define (first-binding)
"First binding"
(xbindkey-function '("b:10") second-binding)
(xbindkey-function '("m:0" "b:9") (lambda () (run-command "echo -e 'KeyStrPress Alt_L\nKeyStrPress Right\nKeyStrRelease Right\nKeyStrRelease Alt_L\n' | xmacroplay :0 &")))
(xbindkey-function '("m:0" "b:8") (lambda () (run-command "echo -e 'KeyStrPress Alt_L\nKeyStrPress Left\nKeyStrRelease Left\nKeyStrRelease Alt_L\n' | xmacroplay :0 &")))
)

(define (reset-first-binding)
"reset first binding"
(display-n "resetting first binding")
(ungrab-all-keys)
(remove-all-keys)
(first-binding)
(grab-all-keys))

;;This allows to scroll through tabs in firefox
(define (second-binding)
"Second binding"
(display-n "setting second binding")
(ungrab-all-keys)
(remove-all-keys)
(xbindkey-function '("b:4")
                (lambda ()
                   (display-n "Scroll up")
                   (run-command "echo -e 'KeyStrPress Control_L\nKeyStrPress Tab\nKeyStrRelease Tab\nKeyStrRelease Control_L' | xmacroplay :0 &")
))

(xbindkey-function '("b:5")
                (lambda ()
                   (display-n "Scroll down")
                  (run-command "echo -e 'KeyStrPress Control_L\nKeyStrPress Shift_L\nKeyStrPress Tab\nKeyStrRelease Tab\nKeyStrRelease Shift_L\nKeyStrRelease Control_L' | xmacroplay :0 &")
))

(xbindkey-function '(release "b:10") (lambda ()
 (run-command "echo -e 'KeyStrPress Alt_L\nKeyStrPress Tab\nKeyStrRelease Tab\nKeyStrRelease Alt_L' | xmacroplay :0 &")
(reset-first-binding)))
(grab-all-keys))

(define (scroll-apps-binding)
"scroll-apps-binding"
(display-n "scroll apps binding success")
(ungrab-all-keys)
(remove-all-keys)
(xbindkey-function '("b:4")(run-command "echo -e 'KeyStrPress Tab\nKeyStrRelease Tab' | xmacroplay :0 &"))
(xbindkey-function '("b:5")(run-command "echo -e 'KeyStrPress Shift_L\nKeyStrPress Tab\nKeyStrRelease Tab\nKeyStrRelease Shift_L' | xmacroplay :0 &"))
(xbindkey-function '("b:1") reset-first-binding)
(grab-all-keys))
(debug)
(grab-all-keys)
(first-binding)

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; End of xbindkeys configuration ;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

I'm looking for a way to alt + tab through the application windows using the scroll wheel but I can't get this working. Suggestions are welcome

Related Posts:
Mouse setup in linux links

setup of the MX1000: .xbindkeysrc with xmacroplay

2007-06-02T13:21:00.000-07:00

Binding mouse events of the thumbbuttens to XVkdb by xbindkeys kind of sucks. I makes Firefox freeze. Hers my new .xbindkeyrc using xmacroplay.

Ok, this is my .xbindkeysrc


# Backward and Forward buttons
# "xvkbd -text "\A\[Left]"" # freezes firefox and makes mouse cursor jump
"echo -e 'KeyStrPress Alt_L\nKeyStrPress Left\nKeyStrRelease Left\nKeyStrRelease Alt_L' | xmacroplay :0 &" #better
 m:0x10 + b:8"
#"xvkbd -text "\A\[Right]""
"echo -e 'KeyStrPress Alt_L\nKeyStrPress Right\nKeyStrRelease Right\nKeyStrRelease Alt_L' | xmacroplay :0 &"
 m:0x10 + b:9

the idea was stolen from Stephan Reichelt Nice Work!
Get xmacro sources and make. copy xmacroplay to /usr/bin. I couldn't find a Suse rpm for this.

Related Posts
MX1000 Setup
MX1000 setup part 2

How the Logitech MX1000 setup works in OpenSuse 10.2

2007-06-02T03:40:00.001-07:00

Some more information about the setup of the MX1000 lasermouse in OpenSuse. Its a guide to use lomoco and keybindings with xbindkeys in conjunction with xvkdb.

In Suse 10.2 udev detects the mx1000 with this ruleset found in:
/etc/udev/rules.d/40-lomco.rules

# "C-BN34",  "Receiver for MX1000 Laser"
BUS=="usb", SUBSYSTEM=="usb_device", SYSFS{idVendor}=="046d", SYSFS{idProduct}=="c50e", RUN="/lib/udev/lomoco.sh"

Then it runs the script:

/lib/udev/lomoco.sh

This script takes options for smartscroll enable/disable and resolution from this cofnig file:

/etc/sysconfig/logitech_mouse

So this is probalby the place to modify the lomoco settings at boot time systems wide.
For keybinding you can use xbindkeys it is a package of OpenSuse and you don't need to compile and install lmwheels. Its not maintained any more from what I understand. But check yourself if in doubt. Create .xbindkeysrc in you homedir with

# 'xbindkeys --defaults > /home/yourhomedir/.xbindkeysrc'.

Now edit .xbindkeysrc and add the following lines:

# Application-toggle button
"xvkbd -text "\[Alt_L]\[Tab]""
m:0x10 + b:10
# Backward and Forward buttons
"xvkbd -text "\[Alt_L]\[Left]""
m:0x10 + b:8
"xvkbd -text "\[Alt_L]\[Right]""
m:0x10 + b:9

The .xbindkeyserc was stolen with modification from here So how it works is that xbindkeys traces the events you trigger by pressing keys. When it catches a defind event then it launches an application which is here above the respective event in the .xbindkeysrc. In our case of the mouse events it starts xvkbd which is a virtual keybord that sends keyboard events to the xserver you define behind the -text optons. you kould define any keystroke you like instead the events here. For lmwheels you could probably define different keystrokes depending on the application you use the mouse in if you need it The b:10 and b:8 and so on are the mouse events where the number is the event the mouse sends. Here I created a listo of events triggerd by the MX1000 buttons if you need to know which number you need to use. Now execute xbindkeys and test the key bindings.

When satisfied you can make an autostart entry in your KDE Atotstart folder to start xbindkeys each time you boot. To do so execute e.g.:

# ln -s `which xbindkeys` `kde-config --userpath autostart`

links to lots of mouse setup guides

Technorati Tags: Linux, MX1000, Guide

Logitech MX1000 USB mouse configuration / setup for X11 in OpenSuse 10.2 linux

2007-06-01T14:59:00.000-07:00

The MX1000 Lasermouse is supported pretty well by OpenSuse. It was just a little hard to find out what you need. Here I write how I did the setup of the MX1000

To support all mouse buttons of a modern USB mouse like the Logitech MX1000 lasermouse in X11 I use the evdev driver. This driver is compiled into X11 by Suse linux so you just need to activate that driver. To do this I first modified the /etc/X11/xorg.conf now it looks like:


Section "InputDevice" 
 Identifier  "Mouse0"
 Driver      "evdev"
 Option      "evBits" "+1-2"
 Option      "keyBits" "~272-287"
 Option      "relBits" "~0-2 ~6 ~8"
 Option      "Pass" "3"
EndSection

The options are stolen from the man page of evdev and are said to work for virtually any mouse. forget all about these mapping and ZAxixMapping Options, they are not supported by evdev, Check the manpage. In some forums you can read to use Options protocoll evdev. But this is obsolete and only used for older veresions of evdev. However, the identifier Mouse0 is crucial I guess, since with another identifier like mouse[0] the xserver refused to start. My explaination is that is has to match the entry in /dev/input/mouse0. This entry (node) done by udev. The configuratin is probalbly done at installation of Suse. Sombody wrote that Suse 10.2 does not support evdev any more. I think this is not true. You don't need to load it as a kernel module. evdev is compiled into the Xorg server.

I use lomoco, to control the feaures of the MX1000 lasermouse. lomoco is installed per default, it can be used to incerase the resolution from 400dpi (default) to 800 dpi:

# lomoco -8

and to turn on/off smart scrolling


# lomoco --sms
# lomoco --no-sms

in Suse linux /etc/sysconfig/logitech_mouse holds the configruation file of the lomoco config scrip that is executed on detection of a logitech mouse to configure smart scrolling and resolution of Logitech mice at startup.

What happens when you hit a mouse key depends on the xevent that is triggered and translated by the application into action. The xev command can be used to decipher the events triggerd by the mouse buttons. these xevents are generated by the mouse depending on sms or no-sms (use

# xev

in a console to test):


button               event with smart scroll     event without smart scroll
left                            1                          1
wheel pressed                   2                          2
right                           3                          3
wheel up                        4                          4
wheel down                      5                          5
wheel tilt right                6                         14
wheel tilt right release       14                         n/a
wheel tilt left                 7                         13
wheel tilt right release       13                         n/a
thujb back                      8                          8
thumb front                     9                          9
app change                     10                         10
smart scroll up                 4(as long as hold down)   11
smart scroll up release        11                         n/a
smart scroll down               5(as long as hold down)   12
smart scroll down release      12                         n/a

this is a list of x events triggerd by the mousbuttens in different modes of the logitech mouse.
Intersitng to note that although the MX1000 is a 12 button mouse (10 button + two for the wheel) the mouse can generate 14 events. So actually these Options Buttons 12 you can find in many forums are probably wrong anyway.

Xmodmap is used to remap the events of the mouse to other events. This technique comes in handy when an application has a hard wired reaction to a mouse event, but you want to trigger that event by another moue button. So what xomodmap does is to translate one event to another and sends that event to the application instead of the original. For example you could use xmodmap to map the thumbbuttons 8 and 9 on xevents 7 and 8 by

# xmodmap -e "pointer = 1 2 3 4 5 8 9 6 7 10 11 12 13 14 15 16 17 18 19 20"

In this row the position of the digit is the value of original x event and the value is the new xevent that will be sent in case that event is trigged, in this matter (digit at position 6 has the value 8) event 6 (tilt mouse wheel with smart scroll enableled) will be mapped to event 8 (thumb button) . Event 8 (thumb button) is now tranlated ot event 6 (tilt mouse wheel)
This above mapping is interesting since firefox can be configured to do horizontal scrolling or to scroll forward and backward history on events 6 and 7 depending on the about:config settings when you don't need horizontal scrolling a lot. . This is the default setting (type about:config in the adress field of firefox to check, I used Firefox 2.0 here) which does history scrolling on events 6 and 7 , that are usually the events triggered by the horizontal mouse wheel in X11.

mousewheel.horizscroll.withnokey.action Standart Integer 2
mousewheel.horizscroll.withnokey.sysnumlines Standard boolean false

of course you can also do horziontal scrolling now with the thumb buttons if you like but you have to change Integer to 1 and boolean to true.

Other customizations of mouse xevents that x applications do not recoginze by default can be done by imwheel or xbindkeys in combinatin with xvkdb or xbindkeys in combination with xmacroplay

imwheel : imwheel will directly translate mouse events to (virual) keystrokes (or xevents) that are sent to and recognized by the application. e.g. alt + left will scroll one item back in history of firefox and you can configure imwheel to sent this keystrokes to the application when the thumb buttons are pressed which are not recognized by default. The advantage of imwheels is that it can be configured to send different keystrokes when a different application is used and that it doesen't need helper applications.

xbindkeys: xbindkeys is usually used to execute applications when a key or a combination of keys is pressed. This is usually referred to as keyboard shortcuts. The advantage over imwheels is that it is highly configurable. You can use chord keys with the mouse and you can change the keybindings on the fly. With the mouse it can be used to execute an application that itself sends (virtual) keystrokes (or xevents) to an application similar to imwheel. Applications that are capable of sending virual keystrokes would be xvkdb and xmacroplay. I prefer to use xmacroplay since it is lightwight, quicker and causes less overhead when mouseevents are triggered in rapid fashion.

Related Posts:
MX1000 Lasermouse

Usefule mouse setup links:
linklinst of Lots of linux mouse setup guides:

Customizing Blogger "Post a comment" link and open it in a Popup Window

2007-05-29T13:06:00.000-07:00

The "post a comment line" can not be customized using the classical layout in blogger unless you hack your template as described here. So if you like the "post a comment" form open in a new Window than this solution is ok. The option to open it in a popup window is not documented. Here I discribe how the customized link can be hacked to open the blogger comment form in a a popup window on click.
First you have to follow the instructions of bloger help and find out your Blog ID. If you want to open the post comment field in a new window you could use the code as suggested by blogger. It should look similar to this:

<a href="http://www.blogger.com/comment.g?blogID=1234567&postID=<$BlogItemNumber$>">You want to talk to me?</a>

where 1234567 should be replaced by you blog id and "You want to talk to me?" by the line you want to appear as the link to you comments form of you blog.
However, there is no way documented to open that form in a pop up window with a customized link. To do that you have to replace <$BlogItemCreate$> with the following code in your template:

<a class="comment-link" href="http://www.blogger.com/comment.g?blogID=1234567&postID=<$BlogItemNumber$>&amp;isPopup=true" onclick="window.open(this.href, 'bloggerPopup', 'toolbar=0,scrollbars=1,location=0,statusbar=1,menubar=0,resizable=1,width=400,height=450');return false;">You want to talk to me?</a>

Again, you have to edit 1234567 to match your blog ID and replace "You want to talk to me?" by the text you want to appear as the link to your comments input form.
Try yourself >> You want to talk to me?
Et voilá, if you now click on the link your comments form will open in a popup
window.

The Novell Microsoft deal

2007-05-29T07:13:00.000-07:00

I have read about that windows Novell deal. Since I use OpenSuse I am a little concerned about that thing. What is all this Linux Novell Microsoft thing about. I understand that Microsoft and Novell siged an agreement, but I'm not positive about the reasons why. What will be the long term goal of Microsoft and Novell concerning open source, what would a scenario look like where Microsoft could achiever ther goals concerning the open source community.

To start this analysis we first have to collect a few thing we know about Microsoft:

We now that Novell is sells server applications and systems that integrate into Windows architecture.
We know that Microsoft is a hostile Company strieving to dominate the market. Microsoft has used its markt power to desroy competitors
Microsoft trys to replace exisiting standards by their own standards or to establish their own proprietary standards on the market.
Microsoft has never been innvoative, they have always been reactive in developping standards and software and they only distribute closed source code. Patents are usually based on innovations and coping copyrighted code usually needs knowledge of the source code its unlikely that open source is using windows copyrighted code.
Microsoft is leader as a provider for desktop systems but is completely out of the loop when it comes to server systems.
Xen Virtualisation is more suitable for server systems
If you want to migrate from one opteration system to the other it might be helpful to be able to operate both systems on the same mashine the same time.

From the above mentioned facts I can not imagine that Microsoft is likely to cooperate with to open source communitiy. They are trying to gain momentum on the server market wich is at the moment dominated the open source community. Though it is understandable from the Novell history that they want to support micorsoft operating systemes and that it might be useful for them for marketing purposes right know. However, it is highly likely that Novell will become unwillingly the Trojan horse that Microsoft will unleash to destroy the Open Sorce community. Without beeing supported by the open source community I am pretty sure that microsoft would have disapeard from the surface of our earth but now things are different.
Microsofe states that it will cooperate with tha open source community on the development of cross platform solutions but in reality they are trying to sneak in proprietary code maintained by Novell Suse and my be OpenSuse which will be indespensible for the function. In parallel they will and promote it and market it untill it they have a market share their system becomes the standard and nobody can use untainted openssource any more.
In conclusion I think it might be the strategy of Microsoft to take over the open source and taint it with closed source and then continue to maintain the code by themselves. As a result Microsoft has taken over the Linux project ....

Have fear ....

Linux toolbox: useful basic console commands to start using Linux

2007-05-27T01:52:00.000-07:00

LLinux features an extremely powerful command line interface with an amount of commands available that windows users can only dream of. Most of the time I think the command line is faster, more flexible and powerful than any grafical user interface (GUI), especially when it comes to system administration. However, for me as a beginner it was a struggle to find and understand the basic commands you need to maintain and configure a linux desktop. Thus, I decided to write down the commands I personally found to be most frequently used. Maybe this is a kick start for some users who start to use Linux.

General syntax remarks
First of all its important to know that linux commands usually accept options which come after the command itself. The short form of the option is usually preceded by a - (one hyphen) and the long form by -- (two hyphens). If you are not sure what a command does or what its syntax is type the command followed by -h ore --help. eg:


# man -h
# man --help

this will remind you of the syntax and options the command man accepts
man
man is a lifeline for linux. man brings up the manual or man page of any command available on your system instructing you how to use that command and what it is doing. So if you stumble upon a command and you don't now how its used or what its doing then type man in combination with the command you need the information about. For example if you want to know what - man - does or which options for - man - are available type:

# man man

cd
cd is use to change the current directory. If you want to change to the root directory type

# cd /

The prompt will show you the directory you're currently in. Typing cd followed by two dots (..) will bring you one level up in the directory tree.

# cd ..

Typing cd with out any options will bring you to your home directory.
the tab key
hitting the tab key in the console will expand your input depending on the context youre using it in. e.g. linux is trying to guess what you want to type. So if you type m and the tab key linux will list all available commands starting with m. The tab key is particularly useful if you're typing long directory path-names. just type the initial letters and then press the tab key to see if linux can already expand your input to full length.

tab tab
hitting the tab key twice will list you all posible input options in the context your just in.

CTRL + C
pressing ctrl and c at the same time breaks a programm you starte from the command line.

up-arrow and down-arrow key
the up-arrow will bring up your recent iput commands. the down-arrow does the oposite. this way you can browse your input history.

Shift page-up and shift page-down
use this to scroll up and down the console text.

CTRL Alt F1, CTRL Alt F2, CTRL Alt F3 ....CTRL Alt F7
pressing CTRL Alt and one of the function keys F1 - F7 lets you swich to other consoles. The GUI (X window) lives in consel F7

su
If you're working in a console you started from a Xwindow console and you need root access to you system you can use the su command: su will prompt you for the root password and than log you in to a root console.

# su -c "somecommand"

will ask you for your your root password to execute somecommand with root privileges and will exit right after the command has finished. This is useful if you need the root privileges for only a few commands.

ls
this will give you a short list of the directory your currently in. it will return only the directory names.

ll
this will give you a long list of the directory your currently in. it will retrurn the file and directory names including the respective owner and group.

l
will give a similar output like ll put will also list hidden files and directories which are usually preceded by a . (dot).

tar
this command comes in handy if you need to extract tar archives. A lot of software for linux is dirtributed as tar archives.

# tar xvzf somefile.tar.gz

This is the command you will usually have to use to extract an tar.gz archive of the name e.g. somefile. the option x means extract, v triggers the the output of all extracted filenames, z tells tar that is has to extract a zipped archive and the flag f inicates that the filname of the archive to be processed follows.
chown
this will set the owner of a file or directory e.g.

# chown me somefile

This will set the owner of somefile to be me. Test if successful with:

# l somefile

chgrp
this will set which group a file or directory belongs to e.g.

# chgrp users somefile

will set the group of somefile to be users.

chmod
chmod is used to set if a file is writeable (w) readable (r) or execuatable (x) by owner, member of the group and the world. look for example at the following line of the ll output:

-rw-r--r--   6 root root  4096 26. Mai 14:56 somefile

The first is a hyphen since this is a file. d would indicate a directory and l a link. the first triplet rw- sets the rights for the owner of this file which is root as indicted by the first entry in the list which is the owner. Hes allowed to read and write this file but not to execute it. The second triplet r-- indicates that members of the group root can read and execute this file but not write to it. The same access ist granted to this file to all others which is set by the last triplet of letters r--. now we assume you want to make this file executable. to do this you could use e.g.

chmod a+x somefile

In addition you want to allow members of the group root to write to this file. this could be done by

chmod g+w somefile

In depth information about Linux permissions

md
This command is used to create a directory with the name somedir

# md somedir

rd
This command is used to remove somedir. this works only if the directory is empty

# rd somedir

If you want to remove a directory with all its contained files and subfolders you could use

# rm -r somedir

rm
This command is used to remove a file with the name somefile

# rm somefile

It can also be used to remove folders with all its contents:

# rm -r somedir

find
Find is useful if you know the name of a file or directory but can't remeber where that file lives. to search your complete directory tree for somefile type

find / -name somefile

To find all files of a certen kind e.g. jpg files which end in .jpg use the asterisk * e.g.

find / -name *.jpg

lsmod
Sometimes to install hardware and stuff you need to know which modules are loaded. typing lsmod will return you a list of loaded modules. this list can be rather long so if you're looking for a module you know the name of grep in combination with a pipe comes in handy
| (the pipe)
The commands you used so far read from the standard input and print the output to the standard output which is usually the console so you can read it. But sometimes you prefer to redirect the output form the standardout to the standard input of a second command to process it further. this can be done by a pipe. To commands are linked together by a pipe in this way, e.g.:

# onecommand | nextcommand

This will send the output of onecommand to the nextcommand wich will do some further processing and sends the output to the standard out which usually happens to be your console.
grep
grep will help you to find lines containig some letters or words your interested in. So if you want to know if somemodule is loaded in your mashine you could type:

# lsmod | grep somemodule

grep is also useful if you want to search your logfiles for messages your interested in if you use it in combination with cat.

cat
Cat will print the contents of a file to the standard output which is usually your console.e.g.

# cat /var/log/messages

Will print the content of your messages logfile.this can be quite long. So if you you are only interested in lines of your log file that contain somephrase you could use :

# cat /var/log/messages | grep somephrase

to filter the output of cat
tail
tail somefile outputs the last ten lines of somefile. The -f option will print updates to the file directly to the console so you can trace whats going on. e.g.

# tail -f /var/log/messages

head
head does pretty much the same as tail but outputs the first ten lines
dmesg
This prints out all your kernel messages. Checking kernel messages is good for debugging, e.g. if some modules failed to load. this is how the linux kernel talks to you and tells you what he has been seeing.

write standard output to file >
sometimes you may need to post parts of your logs to a newsgroup for debugging purposes. To do this you can write the output of cat piped through grep to a file e.g.

# cat /var/log/messages | grep somephrase > news.txt

and the post the output which you can find in news.txt

CTRL + Z
pressing ctrl and z at the same time will interupt a programm and fork it to background you started form the command line and that is still running. Use the command fg to bring it back to the foreground or bg to continue the program in the background.

shutdown
shutdown without options does pretty much what you expect. interesting options are:
# shutdown -h now does preety much the same as:
# halt
this asks all applications to gracefully exit and stops all processes in an ordered fashion.
# shutdown -r now
this will reboot the system. Does pretty much the same as
# init 6

init

this is useful for switching runlevels
# init 5
Initiates the runvlevel with everything, network firewoall GUI login and the like.

# init 3 swichtes to a lowere runlevel and stops Xserver and GUI. this is good for debugging of Xserver.
kill
Use kill to terminate a process.

kill -term pid

where pid has to be replaced by the process id. you can find out about the process id by typing:

ps aux

use in combination with grep to reduce output

Some Links to Guides to start with:
Linux Quickstart More absolute Basic you'll need for Linux
Network commands: list of commmands and short explaination
Console commands list of console commands and brief explaination.
start with linux If you are interested to operate Linux than this would be a good point to start.
system administrator guide This is a good guide to learn how to administrate your linux
the file system learn where Linux stores stuff

Installation and setup of the em28xx linux driver for analogue devices like the PCTV analog USB 2 (150e/55e) in opensuse 10.2

2007-05-26T14:25:00.000-07:00

Ihave an old Pinnacle analog USB video grabber device which I bought some time ago for no particular reason. I tested It in windows but picture quality and software sucked. Channel switching took forever and nothing worked the way I wanted it to, so I stopped using it. But now I tried to install it in OpenSuse Linux just for the hack of it. Here I write my experiences with this device.

First I have to say that Linux actually supports this device and the picture quality of the linux driver and tv sofware is much better than that provided by pinnacle for windows. So if you want to use this PCTV analog usb its well worth switching to linux.

The PCTV analog USB is em2820 videodecoder chip based vido grabber with an LG TAPE series tuner. I once cracked it open to see the chips inside the PCTV analog USB 105e 55e

If you want to use the 150e/55e with OpenSuse 10.2 the em28xx driver which is a successor of the em2820 is the right choice. As far as I understand it, this driver is part of the v4l project and is maintained by Markus Rechberger, Mauro Carvalho Chehab, Aapo Tahkola, Ludovico Cavedon, Sascha Sommer, Aidan Thornton. Thanks to them we can now use this budged video grabber with a great quality driver.

The em28xx is loaded as a kernel module in OpenSuse 10.2 as soon as a PCTV analog USB is connected to a USB port. It then automatcally creates device nodes /dev/dsp (digital audion/pcm) and /dev/video (raw video).

You first have to update the em28xx driver that comes out of the box of OpenSuse 10.2, this means you have to compile them from source and install them. This sounds harder than it is, so don't get scared. The result is well worth the effort.

The updated drivers from may 2007 will allow you to use the following software in addition to mplayer which is also supported by the old Suse driver:

MythTV, TVtime, Kdetv, freevo. In addition the picture quality is markedly improved, setting the resolution of the grabbed video image and setting of the correct videonorm is now possible. Last but not least no VBI device is initialized any more which might also cause some strange errors.

To update the driver you will have to install gcc, kernel-sources, kernel-headers and mercurial packages using your favorite packet manager e.g. yast2 for OpenSuse. In addition you need SOX for the operation of the digital audio. Now you are prepped to download and build the drivers. This is quite easy since the PCTV analog USB 150e 55e doesen't need firmware or any kind of DVB support like DVB cards would require which are also supported by the em28xx. To dowload the sources and to build the driver type:

# hg clone http://mcentral.de/hg/~mrec/v4l-dvb-kernel
# cd v4l-dvb-kernel
# make
# make ./install

If this driver update doesen't solve your problem then you might consider to try the experimental tree of the driver:

# hg clone http://mcentral.de/hg/~mrec/v4l-dvb-experimental
# cd v4l-dvb-kernel
# make
# make ./install

Markus Rechberger also suggested somwhere (cant remember the link, though) to reinstall the media/lib to troubleshoot problems with using tvtime. This helped me too to get TVtime working. This could be done by:

# rm -rf  /lib/modules/`uname -r`/kernel/drivers/media

and then reinstalling the drivers. May be that will help if you are experiencing problems which can't be solved otherwise.

In OpenSuse the module (modprobe or insmod) options live in /etc/modprobe.d/tv. The driver has to be loaded to with the correct options so that It will support the PCTV correctly. My /etc/modprobe.d/tv looks like this:

alias char-major-81 videodev
alias char-major-81-0 off
alias char-major-81-1 em28xx
alias char-major-81-2 off
alias char-major-81-3 off
options em28xx card=8 tuner=47 disable_ir=0

All available valid insmod options for card= and tuner= for em28xx are listed here

Without the tuner=47 option I could not tune UHF channels with my device.

Testing video playback

If you have mplayer installed then you can use it to test for correct setup. An mplayer command which is successfully used to start mplayer for tv playback was:

# mplayer -tv device=/dev/video0:amode=0:input=0:norm=PAL:driver=v4l2:outfmt=yuy2 tv://E6

this should tune channel E6 and open a tv window. Use h and k to switch channels. Note that /dev/video0 has to be adjusted to match your PCTV videograbber. It might also be /dev/video1 or the like if you have multiple video input devices.

This one starts mplayer with audioplayback enabeled without the need of SOX

# mplayer -tv driver=v4l2:device=/dev/video0:amode=1:input=0:norm=PAL:adevice=/dev/dsp1:immediatemode=0:audiorate=48000 tv://

Note that /dev/dsp1 needs to match the Audioinput of your PCTV Videograber. It might be /dev/dsp0 or the like.

Audio setup
The PCTV analog USB features no analog output which you could plug into the line in input of your audio card. The PCTV features digital audio (PCM) which has to be piped to the digital sound input of you sound card. Of the software I have tested only mplayer and MythTV can read directly from the digital sound device of the PCTV analog USB and output it over your soundcard. Kdetv, xawtv, motv and freevo remain silent because they expect an analgo output to be connected to the line in of your sound card. For these tv viewers you'll need a workaround. You have to install and use sox to pipe the digtal audio from the tv card to your sound card.
these commands worked for me:

# sox -t ossdsp -r 48000 -b -c 2 /dev/dsp1 -t ossdsp /dev/dsp
# v4lctl -c /dev/video0 volume mute off

(/dev/dsp1 is the pcm device of the analog USB and /dev/dsp is the PCM in of you audiocard. /dev/video0 should be the PCTV analog USB) Change this if the names of the devices on your box differ) now you should hear the sound of the channel you just tuned to with your PCTV analog USB. If you hear only static then try to tune to another channel. Now you can fire up your TV-viewer software e.g kdetv to swich the channels.

Links:
em28xx driver FAQ A lot of problems get addressed here
em2820 driver FAQ this is the old version of the em28xx driver. All development is done on the em28xx now I guess, but you can find some troubleshooting hints for the PCTV analog USB 10e 55e since it is a em2820 based device.
Official em28xx Project Wiki

How to install AppArmor and Dazuko in OpenSuse 10.2

2007-05-23T06:10:00.001-07:00

So you want to run Dazuko on a OpenSuse 10.2 System. However, This will not work out of the box, since Dazuko and Apparmor are using the same Linux Security Module (LSM) API. If you try to load dazuko while apparmor is running you'll get:

dazuko: failed to register
modprobe: FATAL: Error inserting dazuko (/lib/modules/xxxxx/weak-updates/dazuko.ko): Invalid argument

But there are ways to get Dazuko up and running and Apparmor even at the same time.

To run dazuko in OpenSuse 10.2 you have basically three options:

Unload the apparmor module before you load the dazuko module
recompile the dazuko module with syscalls enabeled (this doesen't work for chroot directories
recompile the kernel to export __d_path

Here I discribe all three of these options: How to do it and how to choose the right option.

Option 1:
This is the right option if you don't need AppArmor. You need to install Antivir using the ./install script that comes with the installation file which can be downloaded from avira. In addition you have to install the dazuko module using e.g. the Yast utility from the OpenSuse 10.2 distribution.

# modprobe -r apparmor
# modprobe dazuko
# avguard start

Use the Yast runtim configuration to make the changes permanent.

Since the default SuseLinux 10.2 default kernel is an SMP kernel for both single and multi CPU systems which does not export __d_path you there are only the following ways to get dazuko running together with AppArmor:

Option 2:
download the sources of the dazuko module and compile the module with syscalls enableled to circumvent the LSM API apparmor uses , local dpath disableled to avoid problems with the SMP kernel and disabbeled chroot support to get it running without __d_path. This is the right choice if you don't need chroot'd directories.

# ./configure --disable-local-dpath --disable-chroot-support --enable-syscalls --mapfile=/boot/System.map-`uname -r`
# make

now install and load the module

Option 3:
This assumes your in the folder wher the patch lives. Its in the extracted source folder of the dazuko moudIf you need support for chroot directories you have to patch kernel source e.g.

# patch /lib/modules/`uname -r`/build/fs/dcache.c patch_dpath.diff

and recompile the kernel in order to make him export __d_path. However this command seems not to work when you never build a kernel yourself. In this case use

# patch /usr/src/linux/fs/dcache.c patch_dpath.diff
# make clean
# make oldconfig
# make
# make modules_install
# make install

The dazuko module can only be compiled after compiling and loading the kernel so that the dazuko module will be compiled with the correct System.map. Else you get

dazuko: loaded, version=2.3.2
dazuko: no version for "__d_path" found: kernel tainted.

when loading the module

The Dazuko module can be configured and installed in this way to support chroot'd directories:

# cd /usr/lib/Antivir
# ./configure --disable-local-dpath --enable-syscalls --mapfile=/boot/System.map-`uname -r`
# mv /lib/modules/2.6.18.2-34-default/updates/dazuko.ko /lib/modules/2.6.18.2-34-default/updates/dazuko.ko.orig
# cp dazuko.ko /lib/modules/2.6.18.2-34-default/updates

everything is detailed in the Suse README of the AntiVir Distribution:

AntiVir README for users of the SUSE distribution
=================================================

There are issues with the on access scan feature on recent SUSE systems. More
thorough and up to date information is available at the www.dazuko.org site.
See the FAQ at http://www.dazuko.org/faq.shtml and the installation
instructions at http://www.dazuko.org/howto-install.shtml for more details.
The RELEASE_NOTES file that came with this installation package discusses the
issues with the latest Linux 2.6 kernels in more depth.

The on access scan feature of the AntiVir Server and AntiVir Workstation
products (commonly referred to as "Guard") runs on top of the Dazuko kernel
module. That's why it is essential that this module is available and fully
functional on a system where on access scans will be used.

By default Dazuko prefers the LSM API to interface with Linux 2.6 kernels to
capture file events. Unfortunately the AppArmor and SE Linux projects block
this API and prevent other modules from accessing it, which is why the
AppArmor and SE Linux features cannot be employed together with other LSM
using modules such as Dazuko.

In these setups it's necessary to switch to the so called syscall hooking
method to interface with a Linux 2.6 kernel. It's suggested to use at least
Dazuko version 2.3.1 and run the configure script with the --enable-syscalls
option. It's essential to specify a System.map file with the --mapfile=
option which exactly fits the kernel which the module gets built for.

Unfortunately there are System.map files which declare kernel data pages as
read only while they actually are not. Taking action at runtime based on this
information when it's incorrect leads to a kernel BUG() while not taking
action at runtime while the information is correct results in Oopses and
crashes. This situation cannot be detected by software without actually
running into the problem. Neither at build time nor at run time is it
possible to determine the correctness of this information.

At this point it's important that the administrator does know whether the
kernel data pages on this system actually are read only or whether they are
not and were wrongly declared so. With this information the Dazuko configure
script can be invoked with the appropriate options. Should this information
not be available, a test system should be used to try which approach works for
such a configuration (this is not suggested to be done on a production
system).

SUSE distributions usually do not have read only kernel data pages. Which is
why the Dazuko configure script assumes that the System.map information is not
correct and issues a warning message to this effect. Should loading a module
which was built this way result in a kernel BUG(), the module unfortunately
cannot be unloaded and the machine needs to be rebooted instead. The module
then needs to be built with the additional --sct-readonly flag passed to the
configure script.

and here:

==============================
README for Linux 2.6 kernels
==============================

In order to lookup full file paths (even in chroot environments), Dazuko
uses the kernel function __d_path(). Unfortunately this function is no
longer exported to kernel modules in the Linux 2.6 kernel.

Dazuko has its own copy of this function but unfortunately it is only safe
for UP (uni-processor or non-SMP) kernels. This is because the real __d_path()
function uses a kernel lock, which is not available to kernel modules.

If your kernel is not configured as SMP, then it is safe to use Dazuko's
copy of the __d_path() function. This is default.

If your kernel is SMP then you should _not_ use Dazuko's copy of __d_path().
You can disable Dazuko's local copy of __d_path() by configuring Dazuko with:

$ ./configure --disable-local-dpath

Since Dazuko requires __d_path(), your kernel will have to export it if you
disable the local copy. Some GNU/Linux distributions have a modified kernel
that does export this function. The "vanilla" Linux kernel does _not_ export
this function. You can check if your kernel exports this function with:

$ grep __d_path /proc/kallsyms

The results can be interpreted as:

xxxxx T __d_path <= exported xxxxx t __d_path <= not exported If you are running an SMP kernel and __d_path() is not exported, then you will have to modify your kernel source code to export the symbol and rebuild your kernel. The modification is very simply and can be made by using the included patch to modify the fs/dcache.c file. Here is an example: $ patch /lib/modules/`uname -r`/build/fs/dcache.c patch_dpath.diff The location of your Linux kernel source code may be different. Once you have patched the fs/dcache.c file, you can rebuild and install your kernel and kernel modules. If your GNU/Linux distribution doesn't export the __d_path() function, you may want to write them email suggesting that this function be exported in the next release. Note: As a workaround, Dazuko now offers the configuration option --disable-chroot-support. This will disable Dazuko from trying to resolve chroot'd filenames, thus making Dazuko safe for SMP kernels that do not export __d_path(). However, if this option is used, registered applications will only be able to work with chroot'd filenames if they resolve the filenames from within userland. If the application you plan to use with Dazuko requires handling chroot'd filenames (and you have an SMP kernel that cannot export __d_path()), then you should contact the application developer to make sure it supports this option.

Azureus I2p Network plugin configuration Howto for Linux

2007-05-22T16:04:00.001-07:00

So Azureus comes with a lot of nice plugins for download. Let's see which of them I'll install .... wait, what is this I2P network plug in... allows use of the I2P network ... Hm, no idea what I2p network really is, sounds useful though ok let's install the plugin and see what happens.

These were my first thoughts when I stumbled upon the I2P network and the corresponding Azureus plugin. Had I only known how much work i took to get it up and running I probably wouldn't have installed it. Here I write what It takes to set up the Azureus I2P Network plugin.

Outline:

Install the i2prouter
configure the server tunnel
configure Azureus to send the i2pnetwork destination hash of the server tunnel to the tracker instead of the local ip adress.
configure Azureus to listen on the target port of the server tunnel of the 12prouter
configure Azureus and the i2p network plugin to communicate with each other via an proxy network relay port

Introduction:

Ok, what is the I2P network

The I2p network is a anonymous Network inside the Internet. It is a distributed network technology to make it harder to the the server to find out your IP address. I like to think of it as a Network layer on top of the Internet protocols. The Internet is usually not as anonymous as you might think since if you connect to e.g. a server over the Internet a connection is established such that the server knows your IP address. This IP address and the data transmitted from and to this address can be logged. Since ISPs (Internet service provider) log which IP address has been assigned to which user its no problem to trace back the person who established this connection. To cut long things short, your IP address usually reveals you real identity

Things get a little different if you connect to the internet through a proxy. In this case only the proxy knows your IP adress but the servers in the internet only know the IP address of the proxy. Every computer behind such a proxy appears to the internet as only one IP address. And heres what the I2P Netwokr does: It turns your computer into a proxy and all your internet communication has to go through that proxy and the peers are identified by hash keys to each other and not by their real IP address and treaffic is dynamically tunneled through multiple proxys. Since the traffic is encrypted and no single proxy knows all IP addresses of the proxys participating in the tunnel and from the outside its not clear if your only routing the tunnel or are the end of the tunnel it becoms extremely hard to trace which hash key belongs to which IP address.

Ok, what does the Azureus I2P Network Plugin do?

Well, I like to think the I2P Network plugin serves as some kind of proxy to which Azureus connects.In this way the plugin can trace all Azureus traffic and if it detects that Azureus is trying to connect to the I2P Network it redirects the traffic to the port of the I2P router/proxy which knows to handle the torrent protocol in the I2P network. (usually 7654 on your localhost). In other words the I2P network plugin redirects the outgoing I2P traffic of azureus to your I2P router that is if Azureus is trying to connect to some peer outside.

The Incomming traffic is not handeled by the plug in. To allow for incomming traffic you have to configure a server tunnel in your I2P-router and forward this tunnel to some port on the machine where azureus lives and listens. For this to work you have to instruct azureus to overwrite the address sent to the tracker with the hash key for that server tunnel

So, what's next?

Configuration and installation of the i2prouter:

Download the I2P proxy and install it to /usr/local/i2p and fire it up

# java -jar i2pinstall.exe
# cd /usr/local/i2p
# ./postinstall.sh
# ./i2prouter start

now fire up your browser and connect to the i2pconsole at:

http://localhost:7657/index.jsp

If you can see the the I2p proxy console now your i2p router should be up and running. Make sure you do the general I2P router configuration and itis working properly (configuring NAT and the like where applicable) Inbound TCP connection need to be enabeled if you want remote source to connct to Azureus (NAT OK)

Now do the Azureus specific I2P router configuration:

Click on "I2P tunnel" Or you can click here

Select add new server tunnel standard and click create.
Enter a Name e.g. Azureus and a Description e.g. Azureus server tunnel.
Check Auto Start (yes) so that the Server tunnel gets opened automatically when you fire up your i2p router.
enter the target host which is the IP address or name of the machine where Azureus lives. If your I2P router and Azureus running on the same box its usually ok to enter localhost here.
As Target host Port enter the port number your Azureus is listening on incoming TCP connections e.g. 57431
Enter a name for the Private key file e.g. AzureusPriv.dat.

The rest can remain unchanged.

Now click save and the I2P router will generate a local destination hash.
Copy the local destination hash to the clipboard that will appear in the status message window. The hash will be the address or identifier which will be announced to the tracker and wich peers from the i2p network will use to connect to you.
If you lost this hash you can also copy it from the Local Server Tunnel entry. You will find a new entry und local server tunnels with the name you have chosen above e.g. Azureus. Click on the Name and a Tunnel configuration window will pop open. Now you can copy the hash from the local destination field . Make sure you get the whole thing its very long.

Now your finished with the I2P router configuration

Configuration of Azureus

go to Azureus/Tools/Ḱonfiguration/Tracker/Client
Click on overwrite announce ip address options and paste the local destination hash in here which you copied from the server tunnel you just created.
Move the curser to the end of the hash and append

.i2p

This is important. If you don't do this correctly it will render Azureus unresponsive when you open a torrent.

Now go to Connection and enter port 57431 in the incoming TCP field. The port should be the same as your entered as the target port field when you configured the server tunnel in your I2P router.
Now go to proxy settings and
activate use proxy for comunication with Tracker
activate have SOCKS proxy
enter localhost as host address
enter an unused port which azureus should use to communicate with the I2P network plugin (which serves as proxy) e.g. 15223
leave user and pass blank
(check) activate Proxying for source communication
uncheck (deactivate) imform tracker ..
select SOCKS Version V4a (do not use V4, this will not work)
check (activate) use the same Proxy setting for Tracker and Source communication.
Now go to Plugins/I2P Network Plugin and

Configuration of the I2P Network plugin

Check "enable" to activate the plugin
enter the path to the directory where you have installed the i2p router to e.g. /usr/local/i2prouter
if you don't do this correctly then you get
I2P install location not defind, plugin initialisation failed
Enter the Port you which the I2p Network plugin to listen on In Network relay port. This port must mach the port you entered as the port for tracker communication.

The rest can remain unchanged.

Configure the UPnP functionality if you which so that the i2p network plugin opens the i2p router ports for you.

Now restart Azureus

When lucky then you get: "i2p router connection succeeded" messge
Attantion: This only means that your i2p plugin works. Make sure that Azreus does not accept connections from outsinde i2p and that azureus does not attemt to connect to networks outside the 12p. this might compromise you ananymity since in this case you're in danger transmit your ip and your hasch to the same server.

Now go to tracker.postman.i2p

Troubleshooting:
If Azureus freezes after you opened a terrent then you dident copy the destination hash of your server tunnel to Azureus correctly. Check that .i2p is appended to the end of the hash.
If your NAT status light remains yellow this most likely means that you haven't configured the incomming connections properly. Then you have to check if you configured the server tunnel correctly. Check furthermore if the target port of the Azureus server tunnel of the i2prouter and the TCP port Azureus listens on do match each other. More over the i2prouter needs to be configured to accept incoming TCP connections.
Links:
Azureus I2p Plugin

MythTV: StartRecorder() -- timed out waiting for recorder to start

2007-05-21T08:20:00.000-07:00

This problem occured after installation of MythTV under OpenSuse 10.2 using the PCTV analog USB video grabber. Here I write how I solved this issue.

This problem turned out to result from a driver problem for the analogeus USB TV grabber which comes with OpenSuse 10.2. This dirver is outdated. I guess the MythTV can't communicate with the driver. The v4l2 driver for the PCTV USB 150e/55e PAL/SECAM for Linux is the em28xx. The current version fixes this problem.

To install the current driver you need to install the kernel source and kernel headings and the mercurial package which will download the drvier source for you.
In addition its a good idea to a clean reinstall of the media library. I used the following commands which probably can do the trick for you too however don't do the "rm" stuff unless you know that the vf4l driver source compiles without problems:

# rm -rf  /lib/modules/`uname -r`/kernel/drivers/media
# hg clone http://mcentral.de/hg/~mrec/v4l-dvb-kernel
# cd v4l-dvb-kernel
# make
# make ./install

In fact I use the experimentel tree of the driver which also woks fine for me.(http://mcentral.de/hg/~mrec/v4l-dvb-experimental)

The experimental driver works for me with MythTV, Freevo, TVtime, mplayer, and kdetv so it can't get much better, I guess.
releated posts

PCTV analog USB2: "open /dev/vbi: Device or resource busy issue" solved

2007-05-20T06:58:00.001-07:00

VBI means Vertical Blank interval of a TV signal in the US terminology. This is the interval between frames where no picture is displayed but information like teletext can be broadcasted. I guess the VBI device in linux is the interface to read out such data from the TV signal provided that you TV grabber can extract you the VBI data from the video signal. Here I write why many people report that the VBI device of the PCTV analog USB is always busy though no application is using it.

the 150e/55e has an em2820 on board which does not support teletext ( raw VBI capture, sliced VBI capture)!!!

Although the em28xx driver supports em2820 based tv grabbers quite well it initializes an vbi device although it probably shouldn't. This is likely to the reason for the erreor you get when starting scnatv:

#scantv

vbi: open failed [/dev/vbi]

open /dev/vbi: Device or resource busy

#dmesg

em28xx #0: this driver can be opened only once

This issue has been fixed in the new version of the em28xx driver. If you update the em28xx driver it wont initialize a vbi device any more. However, after update you are likely to get anoher error message after starting scantv:

open /dev/vbi: no such device or directory

obviously scantv refuses to work without a vbi device which is probably a bug?

But as a workaround you could now start kdetv or xaw tv and scan the channels. This will work and you can enter the addtitional imformation for the channels manually.

Technorati Tags: PCTV analog USB2, em28xx

Howto solve PCTV analog USB 2 (150e/55e) tuner issue: can't scan UHF

2007-05-19T10:45:00.001-07:00

For me the em28xx detected tuner #37 which is probably wrong cause tuning UHF channels din't work. Here I explain how to solve this problem.

My tuner reads LG tape something so I use #47 for the tuner= insmod option instead. Now I can tune in UHF channels.

For Suse you have to edit /etc/modprobe.d/tv and enter the module options for the em28xx. Look for options and change tuner=37 to tuner=47. Now UHF works fine for me. Heres how my /etc/modprobe.d/tv looks like.

alias char-major-81 videodev
options em28xx card=8 disable_ir=0 tuner=47
alias char-major-81-0 off
alias char-major-81-1 em28xx
alias char-major-81-2 off
alias char-major-81-3 off

wiki discussing the em28xx UHF tuning issue for other cards

In addition I use card=8. Here is a list of valid choices for the card= insmod option:

em28xx #0:     card=0 -&gt; Generic EM2800 video grabber
em28xx #0:     card=1 Generic EM2820 video grabber
em28xx #0:     card=2 Generic EM2821 video grabber
em28xx #0:     card=3 Generic EM2870 video grabber
em28xx #0:     card=4 Generic EM2881 video grabber
em28xx #0:     card=5 Generic EM2860 video grabber
em28xx #0:     card=6 Generic EM2861 video grabber
em28xx #0:     card=7 Terratec Cinergy 250 USB
em28xx #0:     card=8 Pinnacle PCTV USB 2 (LG PAL TAPC)
em28xx #0:     card=9 Hauppauge WinTV USB 2
em28xx #0:     card=10 MSI VOX USB 2.0
em28xx #0:     card=11 Terratec Cinergy 200 USB
em28xx #0:     card=12 Leadtek Winfast USB II
em28xx #0:     card=13 Kworld USB2800
em28xx #0:     card=14 Pinnacle Dazzle DVC 90
em28xx #0:     card=15 Hauppauge WinTV HVR 900
em28xx #0:     card=16 Terratec Hybrid XS
em28xx #0:     card=17 Terratec Hybrid XS Secam
em28xx #0:     card=18 Kworld PVR TV 2800 RF
em28xx #0:     card=19 Terratec Prodigy XS
em28xx #0:     card=20 Videology 20K14XUSB USB2.0
em28xx #0:     card=21 Usbgear VD204v9
em28xx #0:     card=22 Terratec Cinergy T XS
em28xx #0:     card=23 Pinnacle PCTV DVB-T
em28xx #0:     card=24 DNT DA2 Hybrid
em28xx #0:     card=25 Pinnacle Hybrid Pro
em28xx #0:     card=26 Hercules Smart TV USB 2.0
em28xx #0:     card=27 Compro, VideoMate U3
em28xx #0:     card=28 KWorld DVB-T 310U
em28xx #0:     card=29 SIIG AVTuner-PVR/Prolink PlayTV USB 2.0
em28xx #0:     card=30 Terratec Cinergy T XS (MT2060)
em28xx #0:     card=31 MSI DigiVox A/D
em28xx #0:     card=32 D-Link DUB-T210 TV Tuner
em28xx #0:     card=33 Gadmei UTV310
em28xx #0:     card=34 Kworld 355 U DVB-T
em28xx #0:     card=35 Supercomp USB 2.0 TV
em28xx #0:     card=36 Hauppauge WinTV HVR Rev. 1.2
em28xx #0:     card=37 Gadmei UTV330
em28xx #0:     card=38 V-Gear PocketTV
em28xx #0:     card=39 Kworld 350 U DVB-T
em28xx #0:     card=40 Terratec Hybrid XS (em2882)
em28xx #0:     card=41 Pinnacle Dazzle DVC 100
em28xx #0:     card=42 Generic EM2750 video grabber
em28xx #0:     card=43 Yakumo MovieMixer
em28xx #0:     card=44 Huaqi DLCW-130
em28xx #0:     card=45 Generic EM2883 video grabber
em28xx #0:     card=46 Hauppauge WinTV HVR 950
em28xx #0:     card=47 Pinnacle PCTV HD Pro
em28xx #0:     card=48 Pinnacle Hybrid Pro (2)
em28xx #0:     card=49 Hauppauge WinTV USB 2 (R2)
em28xx #0:     card=50 NetGMBH Cam
em28xx #0:     card=51 Leadtek Winfast USB II Deluxe
em28xx #0:     card=52 MSI DigiVox A/D II
em28xx #0:     card=53 Typhoon DVD Maker
em28xx #0:     card=54 Pinnacle PCTV USB 2 (Philips FM1216ME)

Here is a list for valid options for the insmod tuner= option. I use tuner=47

tuner=0 - Temic PAL (4002 FH5)
tuner=1 - Philips PAL_I (FI1246 and compatibles)
tuner=2 - Philips NTSC (FI1236,FM1236 and compatibles)
tuner=3 - Philips (SECAM+PAL_BG) (FI1216MF, FM1216MF, FR1216MF)
tuner=4 - NoTuner
tuner=5 - Philips PAL_BG (FI1216 and compatibles)
tuner=6 - Temic NTSC (4032 FY5)
tuner=7 - Temic PAL_I (4062 FY5)
tuner=8 - Temic NTSC (4036 FY5)
tuner=9 - Alps HSBH1
tuner=10 - Alps TSBE1
tuner=11 - Alps TSBB5
tuner=12 - Alps TSBE5
tuner=13 - Alps TSBC5
tuner=14 - Temic PAL_BG (4006FH5)
tuner=15 - Alps TSCH6
tuner=16 - Temic PAL_DK (4016 FY5)
tuner=17 - Philips NTSC_M (MK2)
tuner=18 - Temic PAL_I (4066 FY5)
tuner=19 - Temic PAL* auto (4006 FN5)
tuner=20 - Temic PAL_BG (4009 FR5) or PAL_I (4069 FR5)
tuner=21 - Temic NTSC (4039 FR5)
tuner=22 - Temic PAL/SECAM multi (4046 FM5)
tuner=23 - Philips PAL_DK (FI1256 and compatibles)
tuner=24 - Philips PAL/SECAM multi (FQ1216ME)
tuner=25 - LG PAL_I+FM (TAPC-I001D)
tuner=26 - LG PAL_I (TAPC-I701D)
tuner=27 - LG NTSC+FM (TPI8NSR01F)
tuner=28 - LG PAL_BG+FM (TPI8PSB01D)
tuner=29 - LG PAL_BG (TPI8PSB11D)
tuner=30 - Temic PAL* auto + FM (4009 FN5)
tuner=31 - SHARP NTSC_JP (2U5JF5540)
tuner=32 - Samsung PAL TCPM9091PD27
tuner=33 - MT20xx universal
tuner=34 - Temic PAL_BG (4106 FH5)
tuner=35 - Temic PAL_DK/SECAM_L (4012 FY5)
tuner=36 - Temic NTSC (4136 FY5)
tuner=37 - LG PAL (newer TAPC series)
tuner=38 - Philips PAL/SECAM multi (FM1216ME MK3)
tuner=39 - LG NTSC (newer TAPC series)
tuner=40 - HITACHI V7-J180AT
tuner=41 - Philips PAL_MK (FI1216 MK)
tuner=42 - Philips 1236D ATSC/NTSC dual in
tuner=43 - Philips NTSC MK3 (FM1236MK3 or FM1236/F)
tuner=44 - Philips 4 in 1 (ATI TV Wonder Pro/Conexant)
tuner=45 - Microtune 4049 FM5
tuner=46 - Panasonic VP27s/ENGE4324D
tuner=47 - LG NTSC (TAPE series)
tuner=48 - Tenna TNF 8831 BGFF)
tuner=49 - Microtune 4042 FI5 ATSC/NTSC dual in
tuner=50 - TCL 2002N
tuner=51 - Philips PAL/SECAM_D (FM 1256 I-H3)
tuner=52 - Thomson DTT 7610 (ATSC/NTSC)
tuner=53 - Philips FQ1286
tuner=54 - tda8290+75
tuner=55 - TCL 2002MB
tuner=56 - Philips PAL/SECAM multi (FQ1216AME MK4)
tuner=57 - Philips FQ1236A MK4
tuner=58 - Ymec TVision TVF-8531MF/8831MF/8731MF
tuner=59 - Ymec TVision TVF-5533MF
tuner=60 - Thomson DTT 761X (ATSC/NTSC)
tuner=61 - Tena TNF9533-D/IF/TNF9533-B/DF
tuner=62 - Philips TEA5767HN FM Radio
tuner=63 - Philips FMD1216ME MK3 Hybrid Tuner
tuner=64 - LG TDVS-H06xF
tuner=65 - Ymec TVF66T5-B/DFF
tuner=66 - LG TALN series
tuner=67 - Philips TD1316 Hybrid Tuner
tuner=68 - Philips TUV1236D ATSC/NTSC dual in
tuner=69 - Tena TNF 5335 and similar models
tuner=70 - Samsung TCPN 2121P30A
tuner=71 - Xceive xc3028
tuner=72 - Thomson FE6600
tuner=73 - Samsung TCPG 6121P30A
tuner=75 - Philips TEA5761 FM Radio

the original list can be found here:

em28xx tuner insmod option list: I use no 47 for LG TAPE series
Related Posts can be found here

Solved: nForce4 CK804 (Intel8x0) ALSA audiodriver uses only on two speakers for stereo playback when 5.1 6 channel speakers are connected

2007-05-17T20:38:00.001-07:00

I have a 5.1 surround speaker set connected to my K8N ultra SLI mainboard with a CK804 sound on board. When playing MP3 I realized that Linux used only the front speaker for stereo playback, which is pretty annoying since the front speakers are not full range an desperately need the woofer to run for a nice sound. Here I write how I solved this issue.

It turned out that this issue is not a driver issue, however. This is due to a, in my opinion, odd behavior of the ALSA drivers. ALSA feeds the channel from the source to the corresponding speaker. Since you have only to two channels in stereo sound alsa sends only sound to the front (left and right) speaker. Although this behavior is very logical its verry confusing for a easyly confused noob like me and it took me quite a while to fix this problem.

First I had to make sure that the hardware is set up properly. Therefore I used a little helper applet that comes with the ALSA package:

# speaker-test -c 6 -r 48000

Now I heard sound (noise) on all speakers. If you hear something then the driver and ALSA should work.

Getting a driver update from nvidia is not necessary since they only distribute only sata and lan drivers for their nForce chip set. They included their driver sources for their on board sound to the ALSA project so it should be fine to use their driver. The CK804 is supported as as intel8x0. Anyway, here are the sites for the nvidia downloads:

http://www.nvidia.com/object/linux_nforce_1.21.html
http://http.download.nvidia.com/XFree86/nforce/1.0-0310/ReleaseNotes.html

But I did an ALSA driver update. However, this update is optional I guess.

---snip---

To do this you need to install autoconf, kernel-source, gnu make, kernel-headers with yast if not done yet.

To compile and install the latest drivers from source, first download the CVS
(Remark: devfs is depreciated in Suse 10.2 since it uses udev so we don't need to execute ./snddevices script here)

# rsync -avz --delete rsync://alsa.alsa-project.org/hg your_directory
# ./cvscompile --with-cards=intel8x0 --with-sequencer=yes
# make install
# modprobe snd-intel8x0;modprobe snd-pcm-oss;modprobe snd-mixer-oss;modprobe snd-seq-oss

---snip----

The most important step is to edit the .asoundrc. The .asoundrc is a configuration file which is read in by default and lives in your home directory.

# vi /home/xxx/.asoundrc

vi is an editor but you can use whichever you you prfer to perorm this task. xxx denotes the name of your home directory. Then add the the following code to your .asoundrc :

pcm.ch51dup {
type route
slave.pcm surround51
slave.channels 6
ttable.0.0 1
ttable.1.1 1
ttable.0.2 1
ttable.1.3 1
ttable.0.4 0.5
ttable.1.4 0.5
ttable.0.5 0.5
ttable.1.5 0.5
}

This will duplicate stereo channels to the sourround speakers but you have to use ch51dup as device for output of stereo files. In Amarok I use the configuration shown in the screenshot:

I used the following links as source:

alsa-project home site
Information about the alsa Module intel8x0 which is the relevant driver for the nForce CK804
Additional Wiki page of the ALSA project about the Intel8xo module for the nForce CK80x
Playing stereo on surround sound setup with alsa
general infromation about the alsa .asoundrc configuration file

Technorati Tags: linux, nforce3, sound, stereo, surround, intel8x0, NVidia CK804

Installing MythTV on OpenSuse Linux 10.2 64 Bit w/ USB 2.0

2007-05-15T21:43:00.001-07:00

Myth TV is a Media Center for Linux software. It fetches an electronic programm Guide for you, schedules recordings, seraches your favorite soaps in the electronic programm guide (EPG) and records them for you so you nerver miss a thing. In addition you can load a lot of modules for DVD CD Phosto Mp3 playback and stuff. Pretty neat.

I use MythTV 0.2 and kernel 2.6.18.8-0.3-default SMP with openSUSE patches. The em28xx driver needs to be updated though. Don't try MythTV 0.19 that won't work with the PCTV usb2. probably none of the em28xx devices will work with versions prior to 0.2.

Mythbackend is the server application of MythTV. it controles the TV grabber starts stop the recording and stores media data and stores and the Electronic Program Guide (EPG). All this data is stored in the MySQL database. The forntend is the client application. Its a remotecontrol for the backend. The backend will start stop recordings, swich channels on request of the frontend and stream the movie data to the frontend which will display it to you. The frontend also queries the MySQL datbase for EPG data, I guess. Here I write how I installed MythTV for OpenSuse 10.2

A precompiled *.rpm is provided by Packman

The Packman MythTV rpm v0.2 won't work with SuseLinux 10.2 with PCTV analog USB2 out of the box since the em28xx (driver for the 150e/55e) of OpenSuse 10.2 is outdated. You have to compile you own em28xx V4l2 from sources

follow the MythTV installation instructions for OpenSuse outlined here:

OpenSuse MythTV wiki

or here:

Installing MythTV

Some additional Notes:

Enter as root

# rcmysql start

(issue # chkconfig --level 35 mysql on
when mysql wount start at boot though enabled in the runleveleditor of Yast2 )
then initialize the database:
(Without mysql running mythtv-setup says it cant connect to the mysql server)

# mysql -u root < /usr/share/doc/packages/mythtv/database/mc.sql

List of packqges and dependencys is not complete in the Howto. You have to install xml-grabbers package!!!! this package contains the XML grabbers you need for the EPG channel list!!!! without you get tv_grab_de_tvtoday no such file or directory (or whichever the name of you grabber is)
you'll need to Alt + tab to the console and answer the questions of the grabber. It will ask you which Channels it should fetch for you. only check those channels you'll really need e.g which are available in your Cable netl.
You have to create directory where you want to save your digitally recorded tv programmes and make it world readable/writable before you start the mythtv-setup. you need to specifiy that directory during mythtv-setup. e.g.

# mkdir /home/video
# chmod a+rwx /home/video

start the setup from a console so you can see error messages on the console output!!!!

# mythtv-setup

The grabber is a pice of software which rips the Programm Guide from some website and outputs it in XML (extendable markup language, a flexible file format that mythtv can read)
during setup Mythtv will invoke the grabber

tv_grab_de_tvtoday --configure

This will only configure the grabber. Here you select wich Channels you want to grab from the electronic programm guide from the internet. Configuration will be saved to file

~/.xmltv/tv_grab_de_tvtoday.conf

here you can edit the file in case you made mistakes,
use as root

# rcmythtvbackend start

to start the mythtv backend
After the end of mythtv-setup you can do a

# mythfilldatabase --manual

to enter the frequency channel information for the respective station.
The

--manual

directive triggers an interactive mode where you can provide some information. MythTV needs to know to which channel it has to set the tuner of you tv Grabber (ur PCTV analog USB) if you want to watch a certen Broadcasting station. I will ask you on which channel it can find a Broadcasting station. The unique identifieer of the EPG data is the XMLTV.ID Thus MythTV needs to which XMLTV.ID is found on which channel. In additiont you can enter a lot of other informative stuff like long name and short name of the broadcasting station.
do the mythfilldatabase only for one day with the

--manual

directive else you have to go through all channel by for each day and grabbing takes forever (may be hours +).

use

# mythfilldatabase --update

to grab the following days instead. Once the channels are defined thy will be imported by mythfilldatabase automatically.

Mythfilldatabase takes forever. For me if was faster to run the grabber first for 1 dayThen improt EPG from *.xml file using:

# tv_grab_de_tvtoday --list-channels --output channels.xml
# mythfilldatabase --manual --preset --file 1 -1 channels.xml

have a channels list ready to assaign each station a channel from the channel freequncy list.
you can find this out by using mplayer
start mplayer from the console

mplayer -tv driver=v4l2:device=/dev/video0:amode=1:input=0:norm=PAL:adevice=/dev/dsp1:immediatemode=0:audiorate=48000:fps=30   tv://

and swich channels with h and k. this will require theat the PCTV USB is /dev/video0 and the audio device of the PCTV usb is /dev/dsp1. In the console (STD OUT) mplayer will print the current channel and you can see at the same time what si goin on.
Kdetv can also do a channal scan and you have to add the names manually. the em2820 of the PCTV USB2 150e/55e does not support Teletext/vbi. So dont bother.
then grab the rest

# tv_grab_de_tvtoday  --slow --output EPG.xml

and import

# mythfilldatabase -file 1 -1 EPG.xml --update

the 1 denotes the input source. If its not the first input source you want to grab then adjust this number accordingly -1 is ths offset which imports everything up to 7 day in advance from now on.
When lucky you can now start frontend to watch tv and do recordings and stuff.

to debug start frontend and backend from a console.

the em28xx and V4l2 drivers which come with OpenSUSE 10.2 are outdated. They don't work for PCTV USB 2 PCTV analog USB2 150e/55e and MythTV
if you start the backend and start live TV you'll get a black screen and get a :

StartRecorder() -- timed out waiting for recorder to start

on the console you started the backend from.

Compile the recent stable version of em28xx

see here: em2880 wiki about mythtv

and here

repository ot the em28xx

you dont need to add any fimware to the source for analog devices like the 150e/55e. Firmware is only needed for digital devices.

To download the sources of the em28xx you will need a special programm which mnages sources. This is called mercurial (check if package is installed). It manages so called mercurial sources.

This is a package wich manages sources of large decentralized projects. It will download the sources you'll need and und update your sources automatically
enter:
You need to install kernel sources and kernel headlins before you can compile the em28xx driver

# hg clone http://mcentral.de/hg/~mrec/v4l-dvb-kernel
# cd v4l-dvb-kernel
# make
# make ./install

now reboot

after reboot MythTV should work.

Note. Kdetv also works only after driver update;)
link that download from v4linux domain is outdated ... can show up on old wiki or forum entries

some useful links:

forum
troubleshooting

Technorati Tags: PCTV 150e/55e analog USB, OpenSuse, 64Bit, Linux, MythTv, em28xx

Azureus lives!

2007-05-13T09:34:00.000-07:00

Azureus is a great torrent tool with a little blue frog as a mascot in the logo. I diden't now that those blue frogs really existed in the wild until recently. The frog is named dendrobates azureus and carries the color azureus in its scientific name.

Testing the SLI mode in Linux

2007-05-13T08:18:00.000-07:00

If you have installed the Nvidia Drivers for Linux you want to know if the SLI mode is running. To do this you can enable the SLI Heads-Up-Display options Nvidia X Sever Settings GUI which can be installed by Yast. To do this start the utility select "OpenGL settings and check "enable sli heads up dispaly"

Now Select "X Server Display Configuration" and click on "Save to X Configuration File".

Now start an Application that should use SLI and on the left an HUD marked SLI should appear.

this should show you the overhead of you GPU but here you see no GPU load becaus the applicatiion does not put much strain on such a setup, however if you run DOOM3 or Quak4 which can both be installed on Linux you will the green bars filling up the white lines.

Install Nvidia 6800GS Grafics cards in SLI mode for OpenSuse

2007-05-13T08:06:00.001-07:00

When I first installed OpenSuse on my System it didn't set up my graphics card properly. I use a Gigabyte K8N Ultra SLI with a Nforce4 Chipset with two 6800GS in SLI mode.
Here I wrie what I've done to get my Grafic cards running. This might be also helpful for others who want to use a NVIDIA driver.
To activate the SLI mode you first have to enable the SLI mode on your mainboard by putting the SLI selector card firmly into place. Some people complain that their SLI mode doesent work and the secoind GPU is not recognised although both are installed into the Mainboard. This happens because without the SLI selector module properly in place the mainboard falls back into normal mode and thus the second PCIe_16 slot with the second SLI device is not available. In addition the SLI cards should be connected using a SLI adaptor for optimal performance.

However the setup of the GPU by Yast/Sax was not succesful so a grafical login was not possible after installation. However, It was possible to fix the Xorg.conf written by Yast using the nvidia Confugration utitlity.

First you'll have to leave the XServer using CTRL + Alt + F2 since display is messy and you can't use it. The reason is probably that Yast/Sax is configuring an Xserver on each GPU with one display for each Xserver which doesent work for me. Switch to the text console. If this fails you have to boot in debug mode with GRUB or LILO in the boot menu. For the next steps you can use the text based yast utility to do the installations.

To enable SLI mode you have to install the Closed Source Nvidia driver for OpenSuse.

Download the 64Bit Nvidia Linux Driver for the 6800GS and follow the installation instructions for Suse.

In the README they tell you to use the Sax utility to configure the Xorg.conf but this doesen't work for SLI devices I guess.

Install the nvidia-xconfig utility using Yast. This can do the modification to the /etc/X11/xorg.conf for you. Type in a root console

# nvidia-xconfig --only-one-x-screen --no-twinview --logo --sli=auto

This configures the xorg.conf to use only one display and to turn on SLI whenever posiible.

If you followed the readme of Nvidia (shutting down Xserver with Init 3 and unloading grafic driver moduls with modprobe -r then adding the repository and instlling the rpm for Suse with yast) you should now be able to start the X server with

# X

or

# Init 5

ScribeFire and : A FireFox plugin for Bloggers. Probably the best bloggin solution for Linux Users

2007-05-12T14:49:00.001-07:00

I was interested in finding a blogging tool for Linux.

I checked the tools which came with the OpenSuse distro. Namely:

Gnome blog poster
kBlog

Whereas both are extremely basic Kblog has the advantage of integrating in the KDE (kicker) task bar which enables your to blog right from your desktop I missed a lot of features I like.

I heard of Flok which is a supposed to be a nice blogging tool but actually is a webbrowser on its own based on the Firefox engine.

Firefox? Hm, maybe there is a plug in which can do the same, so I don't need to install another browser. So I started to look for a blogging plug in for firefox. And then I found one: ScribeFire.

This is a pretty good bloggin tool. It supports technorati tags, del.icio.us bookmarking. track back URL and blogger categories. Scribfire Plugin can bet toggeld using the F8 key. So you can copy and past your Blogpost while surfing.

However, the support of categories is incomplete yet, since it doesn't lists the categories you have used up to now. this would be a nice feature though.

It also posts the pictures from you desktop to blogger for you. However, it tells you that blogger does not support upload of pictures via API. . For now you have to fall back to the URL method as a blogger user i guess. (publish as dreaft first and upload the picture using the HTML interface upload of blogger.

Since it is a Firfox plugin you can use the Firefox build in spell check which is good !

Another nice tool I found is the clipmarks plugin. With this tool you can use the clipmark webservice to clip a pice of a website and insert it directly to your blog as a citation and then add a comment. This is handy when you want to blog any online material.

Yet another quick search hack for Firefox

2007-05-12T12:41:00.001-07:00

Finally a good way to search databases on Firefox. Although this firefox quick search plug in is really nice, however it gets a bit bulky from time to time. You need to nave a search plugin for each datablase and you need to activate it from the dropdown menu. That tends to be a little time consuming. This hack is great since you just need to enter an Keyword for you database and your search phrase. This is much quicker and more flexible when you want to search different databases than choosing it from the dropdown bar and you don't need a plugin

clipped from www.mozilla.org

Visit the page that has the search field that you search with. Right click on
the search field. Choose Add a Keyword for this Search...
The Add Bookmark dialog will appear. Give the bookmark a name, e.g.
"IMDB Search" and create a keyword e.g. "imdb" and file the Bookmark somewhere.
Firefox comes with a Quick Searches folder so you could use that. Click OK

Now enter imdb Matchstick Men into the Location bar and press
Enter.

Linux compared to Windoze, a personal experience

2007-05-09T17:26:00.000-07:00

Now That I use OpenSuse for a while I want to summarize what the advantages and disadvantages between OpenSuse and Windows are.

For the moment it boils down to these major points:

Installation of OS: Linux might sometimes a bit more difficult
Availability of drivers and setup of Hardware: Many manufacturers support only Windows (So choose your Hardware carefully :D)
Availability- and Setup of Software: That depends on your needs.
For server and network applications: Linux rules
Productivity: Media player and Open Office should be fine for most users
Security: Linux rules

Well, Installation: Basic insstallation is easy in the first place. Yast does pretty good job in detecting the hardware, repartitioning the hard drive to set up a dual boot system and configuring the hardware. However, for me installation went not completely smooth. I took a little hacking before eveything worked properly. Here write why I think Linux can be used by anyone and why I think its better than Windows as far as the above points are concerned.

A major issue was the set up of the graphics adapter. I'm using an odd SLI setup with two Nvidia 6800GS cards. Sax/Yast2 did detected my setup correctly but apperently were not prepared to write a working Xorg.conf. Since the grafical log in was configured automatically I could see nothing on my display at all after starting my mashine the first time. It took me quite a while to fix that problem.

Another major problem for Linux users arises from the fact that the manufacturers usualy do not provide drivers for Linux. Development of the drivers relays on the Open Source community in large parts. This usually causes a delay in availability of drivers and as a result can cause troubles during setup. (See e.g. bcm43xx, PCTV USB 2). Setup of my hardware for Windoze was easier in some instances, since all drivers were easy to set up by a GUI.

Some linux users say, well Linux is not Winodoze and Linux is not supposed to do everything for the user because its Open Source. Yet, in my opinion, Open Source and usability do not exclude each other. Automatic installation routines are helpful for the unexperienced and easily confused user and therefore should be further developed to include those users into the open source community. For the experienced user the existence of automatic installation tools (Yast and the like) usually doesn't mean a loss of freedom, flexibility or control since he still free to configure everything by hand, if he wishes.

However, you get a lot of help from the Linux community with this hacking stuff and if you are willing to Google around a little I'm confident that you can find a fix for almost any problem you have. I myself like this hacking a lot, so I have no problems with it.

Next thing is the installation and set up of the Software. As long as there is a binary available for your distro, the installation is no problem and I think it can't get any easier than this. The KDE desktop looks great and comes with a lot of goodies. To be honest with you, I like the KDE GUI much more the the Windows GUI. I would say its superior. If I think of the Beryl window manager for Linux (its beta but works great for me) I would say it far far far ahead of windows. If you havent seen it, that's well worth a try.

Thunderbird, Firefox, DigiCam and OpenOffice work out of the box. And I think the average Joe won't need much more than this in 99% of the cases. So I would say a preconfigured Linux system which circumvents the installation hassle might be a good choice for some users since it saves you the buck for your copy of Windoze.

However, things might look different in rare cases where you desire a software for which no binaries for your distro available and you have to do the installation by hand or compile the sources yourself. This is usually no problem with the right HowTo, though you have to spent more time on it. Only on rare occasions these installations things will cause you some headache and in the case of (installation of AntiVir) i head to do alot of reading.

I think some of these problems arise because there actually is no such thing as Linux. Instead there are so and so many Kernel versions, so and so many library versions and distributions that make it rather impossible for an manufacturer to provide the user with all rpm or binaries. In this respect the high flexibility of Linux might be a disadvantage.

The Availability of software is a difficult question. I guess If you insist on using Microsoft Office, Linx apparently is not the right choice for you. Also a lot of comercial software is not ported to Linux. However, depending on your demands there is an Open Source alternative for virtually any application you might think of. In addition, you get a lot of hight quality software you don't need to pay for which would cost you a fortune if you had to buy it for windows. I'm thinking of the great video processing software, the top notch server applications and the like. Thus, as far as applications are concerned, with Linux you might get not exactly what you are used to but may be you get much more than you are used to in the end.

A major issue of Windoze is security. The malware constantly targeting your Windoze machine is a nightmare. Linux comes with a firewall which is much mor powerful than the Windoze firewall, and what is more there is almost no malware targeting Linux. Moreover, theres is usually no chance for malware to execute commands with root privileges. This is in contrast to Windoze, where pretty much everything is executed as root. Windoze is a constant threat to your privacy cause you don't know which kind of data it sends back to Redmond during online updates (and who knows, may be Mr Gates opend some backdors in his Windoze on request of the NSA so they can spy on you .... Warning: joke :D)

In many respects Linux is far ahead of Windoze. The Linux 64Bit version is around for years now, whereas Windoze only begins to support 64Bit architecture. Windoze only recently implemented a Firewall whereas Firewalls are second nature to Linux for ages. The Beryl Winwos manager is for Linux is just great and there is nothing comparable for windows I would know of (correct me if you wihch). There are probalbly a lot of more examples to continue.

In conclusion I definitely think that Linux is the better operating system which gives more freedom and flexibility by far. In addition you get extremely powerful applications for free. However, Linux is not supported by a lot of software- and hardware manufacturers. Of course this is not a problem of Linux itself but might make life a little more difficult every now and then. Depending on your Hardware and Sofware requirements Linux might not be the right choice. Though, I believe that for the average user who is doing some Internet surfing, graphics processing. letter writing, multi media playing .... Linux can satisfy all his needs for a much better price and with more security.

So why not give it a try and get a copy of OpenSuse or any other distro you like and put it in your DVD drive. Linux will install a dual boot system for you on its own most likely without any trouble, so you can play around a little. Its no big deal any more.

Configuration Belkin 802.11g 54g Wireless Desktop Network Card (BCM4306) for linux using native drivers

2007-05-09T14:27:00.000-07:00

So I got myself one of these WLAN desktop cards from Belkin a long time ago when I was just using Windoze. No I learned how set it up in linux.

When I installed Linux this card identified itself to the system as:

#lspci
01:08.0 Network controller: Broadcom Corporation BCM4306 802.11b/g Wireless LAN Controller (rev 03)

and (suprise suprise) it turned out that Broadcom refuses to support the Linux Open Sorce community with drivers (neither closed nor open source) or the specifications about the chip.

Luckily with Suse 10.0 I managed to use this card by the help of the ndiswrapper which is the remedy for most Linuxusers forgotten by the Hardware manufacturers.

But now Suse 10.2 comes with an open source driver for BCM4306 devices bcm43xx. Thanks to those nice guys who reverse engeneered the specs so the driver could be developed. --> THANKS GUYS :D. The developent of the driver is of particular importance, sicne my SE515 WlanRouter also runs with Broadcom chips and probably all efforts to develop an firmware for this device by openWRT will depend on these drivers.

Theres onle one catch. The bcm43xx driver needs firmware in /lib/firmware before it can be activated. Luckyly there is a tool to download and install the firmware autmomatically.

Just install the packet "bcm43xx-fwcutter and execute as root:
# /usr/sbin/install_bcm43xx_firmware

thats all. N0w you can configure and activate the device using Yast or which ever technique you prefer.

Links:
OpenSuse Forum Thread
bcm43xx Linux driver home site

Howto install the Pinnacle PCTV analog USB (PCTV USB 2, 150e / 55e) OpenSuse 10.2 64Bit

2007-05-07T14:42:00.000-07:00

PCTV analog USB (Pinnacle Systems GmbH 150e/55e PAL/SEC) Works! It is detected as Pinnacle PCTV USB2 and supported by the em28xx driver.
The TV decoder chip on this device is

em2820.

The label on the tuner reads:

LG TAPE-S701D 601F0514.

Other chips:

EMP202.
Phillips
SAA7113H,
TIPS6:120,
PIC16C54C

# dmesg
em28xx v4l2 driver version 0.0.1 loaded
em28xx new video device (2304:0208): interface 0, class 255
em28xx #0: Alternate settings: 8
em28xx #0: Alternate setting 0, max size= 0
em28xx #0: Alternate setting 1, max size= 1024
em28xx #0: Alternate setting 2, max size= 1448
em28xx #0: Alternate setting 3, max size= 2048
em28xx #0: Alternate setting 4, max size= 2304
em28xx #0: Alternate setting 5, max size= 2580
em28xx #0: Alternate setting 6, max size= 2892
em28xx #0: Alternate setting 7, max size= 3072
saa7115 8-0025: saa7113 found @ 0x4a (em28xx #0)
tuner 8-0043: chip found @ 0x86 (em28xx #0)
tda9887 8-0043: tda988[5/6/7] found @ 0x43 (tuner)
tuner 8-0063: chip found @ 0xc6 (em28xx #0)
tuner 8-0063: type set to 37 (LG PAL (newer TAPC series))
em28xx #0: i2c eeprom 00: 1a eb 67 95 04 23 08 02 10 00 1e 03 98 1e 6a 2e
em28xx #0: i2c eeprom 10: 00 00 06 57 6e 00 00 00 8e 00 00 00 07 00 00 00
em28xx #0: i2c eeprom 20: 16 00 01 00 00 00 00 00 00 00 00 00 00 00 00 00
em28xx #0: i2c eeprom 30: 00 00 20 40 20 80 02 20 10 01 00 00 00 00 00 00
em28xx #0: i2c eeprom 40: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
em28xx #0: i2c eeprom 50: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
em28xx #0: i2c eeprom 60: 00 00 00 00 00 00 00 00 00 00 2e 03 50 00 69 00
em28xx #0: i2c eeprom 70: 6e 00 6e 00 61 00 63 00 6c 00 65 00 20 00 53 00
em28xx #0: i2c eeprom 80: 79 00 73 00 74 00 65 00 6d 00 73 00 20 00 47 00
em28xx #0: i2c eeprom 90: 6d 00 62 00 48 00 00 00 1e 03 50 00 43 00 54 00
em28xx #0: i2c eeprom a0: 56 00 20 00 55 00 53 00 42 00 32 00 20 00 50 00
em28xx #0: i2c eeprom b0: 41 00 4c 00 00 00 06 03 31 00 00 00 00 00 00 00
em28xx #0: i2c eeprom c0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
em28xx #0: i2c eeprom d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
em28xx #0: i2c eeprom e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
em28xx #0: i2c eeprom f0: 00 00 00 00 00 00 00 00 0e 5d 62 39 03 ad 97 e4
EEPROM ID= 0x9567eb1a
Vendor/Product ID= 2304:0208
AC97 audio (5 sample rates)
500mA max power
Table at 0x06, strings=0x1e98, 0x2e6a, 0x0000
registered VBI
em28xx #0: V4L2 device registered as /dev/video0 and /dev/vbi0
em28xx #0: Found Pinnacle PCTV USB 2
usbcore: registered new driver em28xx

# lsusb
Bus 001 Device 005: ID 2304:0208 Pinnacle Systems, Inc. [hex] Pinnacle Studio PCTV USB2

Suse 10.2 Recognizes the divice and loads the em28xx driver automatically when device is plugged into USB 2.0 port.

Install: mplayer, video4linux2 (v4l2), sox

Successfull commands to start tv-playback:


# sox -t ossdsp -r 48000 -b -c 2 /dev/dsp1 -t ossdsp /dev/dsp

REM: this will pipe PCTV USB2 audio stream (dsp1) to installed Soundcard device (dsp2)


# v4lctl -c /dev/video0 volume mute off

REM: turns muting of video0 device (PCTV USB2 v4l2 device) off = sound on.


# mplayer -tv device=/dev/video0:amode=0:input=0:norm=PAL:driver=v4l2:outfmt=yuy2 tv://

REM: starts the mplayer with tv playback from video0 with PAL norm. This way i can get audio and video. channels can be swiched with H and K/L on the keyboard when the focus is on mplayer window.

Another possibility to watch TV where audio playback needs no sox, yet channelswitching is buggy:


# mplayer -tv driver=v4l2:device=/dev/video0:amode=1:input=0:norm=PAL:adevice=/dev/dsp1:immediatemode=0:audiorate=48000 tv://

command options will be explained in the man pages of mplayer under -tv option
scantv mythtv and kdetv tvtime don't work with the opensuse 10.2 version of the em28xx driver.

Installing AVira Antivir for SuseLinux 10.2 64 Bit

2007-05-06T05:34:00.000-07:00

If you need to install an antivirus scanner for linux to protect windows mashines via a linux box Antivir might be the right choice for you. It provides very good virus signatures, on-access scanning and is absolutly free. Keep in mind that the free version can't do on-access scanning of archieves. Archieves will only be scanned by the command line scanner. Here I describe how I installed the AVira Scanner.

Prerquisites:

You need a working dazuko module if you like to use AVguard.

If you need to install the Dazuko module then you might find my Dazuko instllation HOWTO helpful
You will need a working java funtime (JRE) environment installed on your machine to run the GUI included in the Avira workstation packages. (avguard-gui)

Downlaod the free Antivir workstation for linux file from Avira at free-av.de. You can not use the packet of the Suse Distro since this is an evaluation version for enterprises. It will not allow you to update the virus signature database and will expire.
Extratct the archieve. You have to apply a workaroud of a bug in the installation routine of the 64 Bit GUI of AVira as outlined in the Avira Linux Forum.
# cd

# cp gui/gui_workstation_linux_glibc22.tgz
gui/gui_workstation_linux_glibc22_x86_64.tgz
this is probably because the forgot to include the 64Bit version (gui_workstation_linux_glibc22_x86_64.tgz) of the GUI in the installation archieve which is appearently the same as the 32Bit version at the moment.
Now you are ready to install. In the Folder of the extracted archieve type
#./install
and follow the instructions on the screen.
If you are asked how the avguard should be installed choose option [m] meaning that dazuko is loaded as a kernel.
find the location of your dazuko.ko module in another shell with:
# find /lib/modules -mane dazuko.ko
and copy paste the path to your dazuko.ko module in the installation consle.
follow the on-screen istruktions.
to enable the GUI support for AV guard youll have to make sure that in the /etc/avguard.conf the following lines are uncommented:
# Enable and configure GUI support
GuiSupport yes
GuiCAFile /usr/lib/AntiVir/gui/cert/cacert.pem
GuiCertFile /usr/lib/AntiVir/gui/cert/server.pem
GuiCertPass antivir_default
Otherwise AVGuard can not connect to the avguard-gui since both communicate using a SSL connection and the *.perm files contain the keys.
If you don't do this you will find the antivir-gui to report the avguard to be stopped although it is running and it can't be started/stopped from the gui

Using the Yast utility make sure that all users who should be able to start avguard are added to the group antivir
Configure /etc/avguard.conf to suit your needs. The config file is well commented.
Now you can start the antivir-gui by

hit Alt+F2 and enter
antivir-gui
now press return and the GUI should open.
open a command shell and enter
> antivir-gui

To add an Avira Icon to your KDE menu first download the Avira umbrella icon.
Then open it with Konquerer and save as as PNG file format (Document/Save File as/ chose png from the format dropdown menu and chosse save). Then move *.png to /usr/share/pixmaps
# chown root:root avira_logo.png
# mv avira_logo.png /usr/share/pixmaps
Then open the KDE Menu-Editor and go to System/Security. Chose File/NewItem. Chose Name eg. AntiVir. Klick on the Icon button select other icons and find the umbrella and select it and klick OK. Enter antivir-gui in the command field and klick the disk icon to save.

Commads and default Files:

antivir-gui will start the antivir gui
antivir -h will list all command line options of antivr command line scanner
antivir -z -s --scan-mode=smart --heur-macro --heur-level=2 --moveto=/home/unwanted [path]
this command will envoce the antivir virus scanning utility such that it will scan all files whis extentios likely to contain viruses of [path] and its subfolders for viruses and malicious macros with heurisitcs level 2. Suspisous files will be renamed and moved to the default quarantine directory of antivir.
antivir --update will update the virus signatures of antivir.
avguard start will start the on-access scanning tool of antivir and will load the dazuko module. Note: the dazuko module will need to be properly installed in order to allow avguard to be operational. if dazuko and avguard were poperly loaded

#dmesg
should read something like:
[...]
dazuko: info: using chroot events for chroot'd processes
dazuko: loaded, version=2.3.3
(-note:-
"dazuko: info: using chroot events for chroot'd processes" means that dazuko modulae was configured using the "--disable-chroot-support" directive at compile time. Needed for SMP kernels that do not export d_path. see: Dazuko in OpenSuse 10.2
-note-)
# tail /var/log/messages should read something like:
[...]
May 6 19:14:21 localhost antivir[20149]: excluding "/home/unwanted/" from scan (quarantine directory)
May 6 19:14:21 localhost antivir[20149]: excluding "/sys/" from scan (special file system)
May 6 19:14:21 localhost antivir[20149]: excluding "/proc" from scan (special file system)
May 6 19:14:21 localhost antivir[20149]: Information: Dazuko kernel module version 2.3.3 detected

avguard status will report the status of avguard
avguard stop will stop the on-acces scanner

avupdater start will start the av updatedemon wich will update the virus signatures
configantivir script to configure the avupdater deamon. follow the on screen instructions.

/etc/avupdater.conf holds the configuration file for the avupdater daemon.
/etc/avguard.conf holds the configuration of avguard

/usr/lib/Antivir default installation path of antivir avguard and antivir-gui
/home/unwanted The default quarantine directory of avguard

/usr/lib/Antivir/MANUAL.avguard manual on istallation and configuring avguard.

Note:

The On-Access Scanner AVguard will not scan compressed files for viruses. I think this is a major drawback since most users want to protect their windwos installation from viruses that come in files they downladed with Linux. Use the scanner as a workaroud to scan compressed files.
The GUI will only allow for scanning of files in the home directory. For other files you have to use the command line scanner. However, users who want to protect their windwos installation from viruses the download via Linux may want to scan /media and /windows folder as well if the devices are mounted with write permission. Possible workarounds include adding a link form the home directory to other directories you want to inlcude in your scan. The command line utility can be configured to follow links.

Links:
Avira Forum
Free AntiVire download site

Howto Install Dazuko (with AppArmor and AntiVir or ClamAV) with the default kernel in OpenSuse 10.2 64Bit version

2007-05-05T05:31:00.000-07:00

Issues:
On-Access scanning (AntiVir: AVGuard, ClamAV:Clamd with Clamuko enabeled) does not work using OpenSuse 10.2 as AppArmor is enabeled by default

Why?:

Clamd and AVGuard require Dazuko which is a kernel module (dazuko.ko). (Note: dazuko.ko is the version for new kernel. dazuko.o the old kernel version) Dazuko and AppArmor are using the same API. Since AppArmor is enabeled per default in OpenSuse 10.2 the Dazuko module can not be loaded and as a result AVGuard and On-Access scanning by Clamd/Clamuko will fail.
As Dazuko support of ClamAV is experimentel (Clamuko) the feature is disabeled in the /etc/clamd.conf and needs to be enabled first.
For security reasons Dazuko only talks to processes of root. The OpenSuse default /etc/clamd.conf starts clamd as user vscan. This has to be disabeled in the clamd.conf.

suggested reading:
known issues in OpenSuse:
HowTo install dazuko
Dazuko doesent run on my 64 bit mashine

Workaround:
Install the Dazuko module of OpenSuse 10.2 using the Yast utility.

Unload AppArmor e.g.
# rmmod apparmor
Load Dazuko module e.g.
# modprobe dazuko

(--Note:

To permanently avoid loading of AppArmor at boot time you can use the Yast/Novell AppArmor/AppArmor Control Panel Utility.

The dazuko module can be installed as a binary package using the Yast Utilitiy

After installation the binary can be found in standard installation at:
/lib/modules/2.6.18.2-34-default/updates/dazuko.ko

if you are using different kernels
# find /lib -name dazuko.ko
will show you all occurences of the dazuko.ko module

Note--)

Resolution:
If you need/want to use AppArmor and AVGuard/Clamuko in parrallel you need to downlaoad the Dazuko sources and to compile the dazuko module (I used version 2.3.3) using the following directives:

# ./configure --disable-local-dpath --disable-chroot-support --enable-syscalls --mapfile=/boot/System.map-xxxx
# make

Now you should be abel to load the module
# insmod ./dazuko.ko

To check wheather its loaded successful
# lsmod | grep dazuko
dazuko
# dmesg
[...]
dazuko: info: using chroot events for chroot'd processes
dazuko: loaded, version=2.3.3
#

Finally replace old dazuko.ko with new dazuko.ko version.

(---I've done this after installation of AntiVir and after installation of the dazuko module of the OpenSuse distribution using the Yast utility. This might avoid some ajustments you might need to do otherwise like installing a node in /dev/modules and adjusting the module dependecies and the like. to make the device node first get the major number of the dazuko module which needs to loaded already:
# cat /proc/devices | grep mdp
254 mdp
where 254 is usually the major number. Now make a device node
# mknod -m 600 /dev/dazuko c 254 0
# chown root:root /dev/dazuko
----)

(--Note:

System.map-xxxx has to be replaced by the actual fileneme of your systme map. You can find it using e.g.
# find /boot -name System.map*
Chose the System.map of the kernel you which to compile the module for. This is probably the kernel you boot and for standard installations this will be the default kernel. The Mapfile will look something like System.map-xxxxx-default where xxxxx is the version number of your kernel

to find out which version of the kernel you are using right now you can type e.g.
# cat /proc/version

Note--)

Why?
Dazoko has to be compiled with syscalls enabeled (--enable-syscalls directive) to bypath the API occupied by AppArmor.( no idea what syscalls are ...)
Therfore dazuko needs to now the Systemmap of your kernal (--mapfile=/boot/System.map.xxxx) (Don't ask met a System map is, though)
To use syscalls dazuko needs the "dpath" (don't ask me what that is. either :). Yet with the default OpenSuse 10.2 kernels dpath can not be mad available to dazuko:
Dazuko can use the d_path exported by the kernel. However, OpenSuse 10.2 kernel does not export the dpath.

(--Note:

check dpath export with:
# grep __d_path /proc/kallsyms

The results can be interpreted as:

xxxxx T __d_path <= capital T means dpath is exported xxxxx t __d_path <= lower t means dpath is not exported --Note) Alternatively, Dazuko can use a local copy of dpath. Yet this does not work for SMP (Multi-CPU) kernels. Unfortunately the OpenSuse 10.2 default kernel is SMP for single and multiprocessor systmes. So for the default Kernel we have to disable the use of the local dpath (--disable-local-dpath directive). Luckyly dazuko offers an wokraroud to work without the dpath which only works if you don't need a chroot environment so you can disable dazukos chroot support (--disable-chroot-support direktive). As far as I know chroot is some kind of virtual file system inside a running filesystem. I think that most users will not nedd chroot support. In fact the OpenSuse dazuko.ko module was compiled with chroot-support disabeled. If you need chroot directory support in dazuko than you probalbly can't use default kernels in OpenSuse and you have to patch the kernel sources in order to compile a kernel which exports dpath.

Releated posts:
apparmor and dazuko
Links:
Dazuko home
Free AntiVir Download
ClamAV home

Read se515 firmware root_fs using cramfs

2006-03-26T12:33:00.000-08:00

The firmware of the SE515 can be decomposed with the SE515 tools. resulting in kernel, hader and root_fs files of which root_fs is the root file system of the embedded linux running on the SE515. A brief howto is found at the SE515 forum.

root_fs is big endian (big value on low adress) cramfs compressed. opening this file on a little endian host (little value on low adress) like intel ix86 PC results in following error message:

# cramfsck: superblock magic not found

So cramfsswap can be used to convert from big endian to little endian.

cramfsswap found at http://labs.fqdn.org/page/cramfsswap swaps file from big endian (mips, se515) to little endian (iX86, Linux)

After swapping endianess cramfsck generates new error message
# cramfsck: memory access error

Yet, after
#modprobe cramfs

(works for suse 9.1 where cramfs support seems to be supported as a module) the SE515 root_fs can be mounted:

#mount -o loop -t cramfs image_file /mount/point

Though, the file structure can be accessed the files can not be opened. They seem to be empty. This is problably due to LZMA compression of the root_fs.

I turned out that we have to use patched cramfs.tools patched by the broadcom sourceforge project to extract the root_fs of the firmware. The patched version of cramfsck can detect and swap endianess on the fly. Moreover it supports LZMA compression wich is used by the SE515 but not implemented in the official cramfs.tools.

Related Post:
TOS bit manipulation on the Siemens Gigaset SE515

Links:
Broadcom 6345 Linux distribution - Open source firmware project for routers with Broadcom 6345 chipset like the SE515
broadcom router kernel cross-compilation - introduction to cross compilation of kernel and firmware for the broadcom based routers
MIPS toolchain RPMs - precompiled cross compile tools for mips architecture
SE515 forum lots of useful information and downloads concerning the SE515 router
SE515 Wake on Lan - thread on wake on lan for SE515
Download SE515 firmware source code source code download for the Siemens SE515
Darkman firmware release Darkmans inofficial firmware release for the SE515
HPH Firmware releaseHeins Peter Hippenstiels inofficial firmware release. Contains also a firmware archieve with all old and up to date official SE515 firmwares, source codes.

Techorati Tags:
Siemens | SE515 |DSL router |cramfs |Firmware |

Resizing NTFS partition on a HP-Omnibook XE3-GF to install SUSE Linux 9.1

2006-03-25T06:17:00.000-08:00

With only one Windows partition on my laptop I wonted to set up a dual boot system without installing windows new. SUSE v9.1 pro claimed to support dual boot installations over windows installations by default, so it should be a pice of cake. Though, it turned out that this was a harder task than i thought.
During installation, YAST failed to resize the NTFS partition C:\ wich Windows XP was installed to, though I followed the documentation:
I defragmented the hard drive to move empty space to the end of the partition.
I swiched of virual memory because windows defrag cant defragment (pagefile.sys).
Moreover I Figured out that you also have ot disable hibernating
- (Start/settings/Control Pannel-Power Options
Then on page Hibernate uncheck the checkbox "Enable hibernate support") -
since the hibernating file on the harddrive (hyperfil.sys) can not be moved by defrag either.
Though Yast still failed to resize the NTFS partition.

Consulting the 90 days installation support was worthless. They answered my support question by copy pasting the answeres from their faq database. This is pretty much useless since i already read all these faqs before wich didn't help and then decided to contact the support. And they only came up with the same crap. (omg X-() Finaly they said this would be a windows problem. IMHO this could not be true since the resizing algorithm of Yast failed and not the defragmentation. So I had to solve this problem on my own althoug I payed for this worhtless 90day installation support. crazy.

Finally I found out what the problem was. HP Omnibooks have a hidden recovery partition that you cant detect with winows explorer so I wasent aware of it. This recovery partition happens to be the first partition on the hardrive /dev/hda1 Win95 FAT32 LBA. reported to be free and Yast needs the windows partition c:\ to be the first partition.

You can change the size of the windows partition only in expert mode but then it works.

Links:
SUSE LINUX on IBM Laptops from pro-com
How to use ntfsresize from the command line
Dual Booting Linux and Windows NT/2000/XP
NTFSresize manpage
Techorati Tags:
NTFSresize | Omnibook |

Advanced Siemens Gigaset SE515 configuration

2006-03-23T12:04:00.000-08:00

The Siemens Gigaset SE515 runs some kind of Linux as the operatign system. This is interesting since it fetures a tty terminal to which you could log on using telnet.

using a dos- or unix console type:

#telnet se515

and the SE515 will talk to you.

However, its a little bit difficult to access the console on the SE515. First you have to downgrade to some firmware version (dont remember which one though) and log on as the user admin. The password will be the same as the pass you use to log on using the HTML interface. Then you 'll need to create a new user. If your happy with that firmware, you don't need to bother.

Now you can upgrade and log on again. If you log on as admin again you wont have root rights but if you log on as the new user you created you'll have root privileges.

With root provileges you have some more power over the firewall than you would have using the HTML interface alone if know how to use

#iptables

Now you can activate MSS clamping on the router which comes in handy if you don't want to adjust MTU of your network device but still like to avoid TCP packet fragmentation when using the DSL uplink provided by the SE515 Wlan Router.

# iptables -A FORWARD -p tcp --tcp-flags SYN,RST SYN -j TCPMSS  --clamp-mss-to-pmtu

Zone Alarm causes considerable CPU load on mashines with heavy traffic

2006-02-27T07:33:00.000-08:00

Reacently I was updating my local Gene Ontology database for MySQL. Using a little java applet wich was updating the local references wich accesses the MySQL server via networkport on the localhost I realized that vsmon.exe an integral part of zone alarm was constantly draing 33% of CPU power as assesd by task manager.
This was probably a result of zone alarm monitoring the heavy network traffic on the local mashine caused by the requests of the softwareagent to the MySQL server on the local mashine.
Shutting down the zone alarm agent dereased the processing time for the databaseupdate considerably and the computer became usable again. Before it was frozen due to the heavy CPU load.

New attack using faked Telekom-Rechnung (telco-bill) to distribute trojan dropper as e-mail attachment rechnung.pdf.exe

2005-12-06T11:04:00.000-08:00

The E-mail attack contaied the TR/Dldr.Agent.zm trojan dropper.
The Trojan dropper was sent as an attachment to the following E-Mail:

Guten Tag,

die Gesamtsumme für Ihre Rechnung im Monat Oktober 2005 beträgt: 710.89 Euro.
Mit dieser E-Mail erhalten Sie Ihre aktuelle Rechnung und - soweit von Ihnen beauftragt - die Einzelverbindungsübersicht.
Sind Sie Unternehmer und benötigen unsere Rechnung zur Geltendmachung von Vorsteuerabzug? Bitte beachten Sie dann, dass Sie seit 29.12.2004 die Möglichkeit haben, Ihre Rechnung per E-Mail mit einer qualifizierten elektronischen Signatur zu erhalten. Sie können diese im Bereich "persönliche Einstellungen" aktivieren.
Sollten Sie dem Finanzamt bisher eine von Ihnen zusätzlich beauftragte Rechnung in Papierform zum Vorsteuerabzug vorgelegt haben, bitten wir außerdem zu beachten, dass wir Ihnen diese nur noch in Form eines "Rechungsdoppels" bieten können, da nur so vermieden werden kann, dass T-Com mehrere Rechnungsoriginale ausstellt.

Antworten auf Ihre weiteren Fragen zur digitalen Signatur finden Sie auch in unseren FAQs unter dem Stichwort "Digitale Signatur".
=================================
RECHNUNG ONLINE - TIPP DES MONATS
Die neuen WünschDirWas Tarife sind jetzt da! Jetzt online anmelden unter www.t-com.de/reo/WuenschDirWas und bis zu 10,- Euro sparen.
Die aktuellen Top-Angebote der Deutschen Telekom finden Sie unter:
www.t-com.de/aktuell.
=================================

Bei Fragen zu Rechnung Online oder zum Rechnungsinhalt klicken Sie bitte unter www.t-com.de/rechnung (oben links) auf "Kontakt".

Mit freundlichen Grüßen

Ihre T-Com

------------------------------------------------------
Aktuelle Informationen zu den Allgemeinen Geschäftsbedingungen finden Sie unter www.t-com.de/aktuell-agb.

The header of the E-Mail had the following content

From - Sun Dec 04 19:34:07 2005
X-Account-Key: account3
X-UIDL: e3e839a5f59cd48a0eca02f3d4522229
X-Mozilla-Status: 0001
X-Mozilla-Status2: 10000000
Return-Path:
X-Flags: 0000
Delivered-To: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Received: (qmail invoked by alias); 04 Dec 2005 17:28:28 -0000
Received: from p2173-ipbf303sapodori.hokkaido.ocn.ne.jp (HELO p2173-ipbf303sapodori.hokkaido.ocn.ne.jp) [58.91.49.173]
by mx0.gmx.net (mx034) with SMTP; 04 Dec 2005 18:28:28 +0100
Received: from t-com.net (tcmail23.telekom.de [217.6.95.237])
by p2173-ipbf303sapodori.hokkaido.ocn.ne.jp (Postfix) with ESMTP id 189E40F124
for xxxxxxxxxxxxxxxxxxxxxxxx; Sun, 04 Dec 2005 12:42:38 -0500
From: Telekom
To: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Subject: Telekom Rechnung Online Monat Oktober 2005
Date: Sun, 04 Dec 2005 12:42:38 -0500
Message-ID: <011101c5f8fa$2412c41b$1ba88a30@t-com.net>
MIME-Version: 1.0
Content-Type: multipart/related;
boundary="----=_NextPart_000_0014_4075B805.18F48070"
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook, Build 10.0.4510
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1123

The filename of the attachment was Rechnung.pdf.exe. The attackers use an extremly high telekom bill to prompt the anxious user to double click the attachment wich seeks to conceal by the file ending. *.pdf.exe that it is an executable file since Windows don't show file endings to the user as a default. this means the visible name of the file to some users will be Rechnung.pdf. Executing the file will install a malicious code wich is identified by AV-Antivier (VDF version = 6.32.18.79, 03)as the trojan dropper TR/Dldr.Agent.zm. In The Windows properties menu the file was assigned the name "Zulu"

Validate the source code of your blog so it displays correctly in more browsers

2005-11-29T06:31:00.000-08:00

Finally my blog inSUBordination is now XHTML1.0 Transitional compliant and approved by the W3C validator. If you like go and validate your blog too!

It's easy because you just have to enter your URL and the W3C validator will automatically load the source code of your blog. Don't paste your template directly because the blogger tags won't validate. Blogger tags have to be translated into HTML by the blogger engine first.

The most frequent errors I encountered where:

XHTML Tags and attributes have to be lowercase. Old code generators produced uppercase HTML Tags. e.g. <META CONTENT="foo"> (wrong) instead of <meta content="foo" /> (correct)

empty tags have to be closed using a slash (/), Empty tags do not enclose other elements, e.g. <br /> not <br>. Applies also to the <img> and <meta> tag.

Don't use the sign & because the browser thinks it follows a special sign which is delimited by a semicolon (;).

...foo.com&action=foo... instead of ...foo.com&action=foo...

the <img> tag needs an mandatory alt="" element, a text describing the image that will be displayed if the image fails to load.

All values for elements within a tag need to be literal, i.e. they must be written between parentheses. This means use <img src="http://www.foo.com/foo.gif" border="0" /> instead of <img src=http://www.foo.com/foo.gif border=0 />.

Isn't it ironic ... SONY illegally uses code published under GPL for their DRM software

2005-11-27T05:18:00.000-08:00

Check out this post at the fsfe.org website stating that sony BMG ignores copyrights of software published under the GPL to produce their spyware rootkit.

The board of IFPI consists of representatives of EMI, Universal, Warner and drumroll Sony BMG. Yes, the very same company that has infected the computers of its customers with a self-installing, hidden rootkit that threatened the security and data-integrity of the machines it was installed on. Fortunately the EFF -- along with several others -- is now taking them to court.

Did I mention that this rootkit contained source code that was taken illegally from Free Software programs under the GNU General Public License (GPL)? So in order to preserve their distribution monopoly on music they've show themselves quite willing to ignore the copyright of software authors

Rootkits are cloaking technologies that hide files, Registry keys, and other system objects from diagnostic and security software, and they are usually employed by malware attempting to keep their implementation hidden. The Trojan of Sony was first discovered at Synsinternals by Mark Russinovich. I think this is not merly a copyprotection since it hooks up into your system calls, scans your executable files each second and moreover uses a filterdriver on top of your CD-ROM driver to guard what you do with your CD-Rom drive.
The sony software phones home and with this type of connection the sony servers could record each time a copy-protected CD is played and the IP address of the computer.
playing it. Thats what a trojan or spyware does. The Sony Spyware installes without propper information in the EULA or registration in the systems deinstallations routines. Thats self evident since it makes no sense to inform the user about cloaking copyprotection and spyware files on his system if you want to cloak them.
The cloaking functionallity of the rootkit has been discovered to be exploited by three viruses so far.
The pathetic thing about the unistaller provided by sony is that it more or less only updates the copyprotection files.

Wake up Neo, the matrix has you.

Links:
Discovery of the root kit by Marc Russinovic
EFF Files Class Action Lawsuit Against Sony BMG
Illegal use of source code by Sony/BMG rootkit published under the GNU General Public License (GPL)
Virus exploits Sony/BMG rootkit
legal issues of Sony's DRM rootkit software

Techorati Tags:
rootkit | sony | GPL

Useful CSS design links

2005-11-25T04:41:00.000-08:00

positioniseverything.net/
http://www.alistapart.com/
http://glish.com/css/
http://css.fractatulum.net/sample/layout_zum_mitnehmen.htm
http://www.andreas-kalt.de/webdesign/tutorials/css-layout-erstellen
http://www.bluerobot.com/web/layouts/
http://www.csscreator.com/version2/pagelayout.php
http://www.webreference.com/authoring/style/sheets/layout/advanced/
http://www.inknoise.com/experimental/layoutomatic.php
http://www.thenoodleincident.com/tutorials/box_lesson/boxes.html
http://www.benmeadowcroft.com/webdev/
http://www.webreference.com/authoring/style/sheets/layout/
http://web-graphics.com/mtarchive/001668.php
http://www.browsercam.com/default.aspx
http://uwmike.com/articles/2005/05/04/jello-liquid-layout

max-width in internet-explorer
design of css menus

The Firefox BioBar. An Indespensible tool for the molecular biologist

2005-11-23T17:22:00.000-08:00

The firefox BioBar is an extention which integrates seamlessly into Firefox. It allows you to querry virtually any database on the internet. It works like the google toolbar and submits your querry directly form the browser to the search engine or database. Moreover you save the time to bookmark each database by hand and thus the Biobar serves a great link collection as well. I myself didn't know how many databases are out there up to now.

Pimp up your blog! Dancing letters!

2005-11-22T05:03:00.000-08:00

This is a perfect idea to pimp up any blog. Just a great idea. Code can be generated here at fancy letters. I fond this site via Männer Blog

Why we need Firefox to take back the Internet

2005-11-22T02:30:00.000-08:00

While working on the new template for my blog I realized why browsers like Firefox are indespensible to maintain a free and open Internet. Learn why.

Setting up my new layout for my blog I realized that the internet explorer 6.0 (IE) - which should actually be termed microsoft's Internet Restrictor - does not interpret the tags the same way like Firefox 1.0.7 does. At first I thought I messed something up, but then understood that its microsoft's "buggy" implementation of the W3C recommandations. Yet, as we learned from microsoft's history of ignoring non-microsoft standards (be it commercial or open-source)I'm actually convinced that microsoft intentionally ignors the W3C recommandations to establish their own standard.
Since so many users to today are surfing the internet with IE, the designers have to support the Internet explorers implementation of the HTML. In pre-firefox times the IE had a market share of 95%+. Ohter browsers had to implement microsoft's "interpretation" of the W3C recommandations for compatibility reasosns. In other terms: the world is using de-facto the Microsoft standard for HTML but not the one of the W3C any more.
With Mozilla Firefox now we have a browser wich strictly implements the W3C recommandations which means the results of your layout are predictable since the tags do what they are supposed to do. Thats why I chosse to check the layout only on Firefox and to use no hacks to make it compatible wiht IE. If webmasters continue to support IE, microsoft will go on to change standards at will.
Only increasing marketshare of a good free browser like Firefox can rescue the web foram beeing take over by microsoft. On many weblogs the statistics form the counters indicate that usage numbers of Firefox surpassed those of IE. This makes me hope for the future.
In conclusion we the users will decide which will be the destiny of the internet by choosing the browser. It's up to you. If you want microsoft to take over the web, then use IE. If you want the intenet to be open-source and free for anyone then use a free browser like Firefox.

MyBlinkies

2005-11-06T15:10:00.000-08:00

My Hi-Fi System

2005-11-06T12:28:00.000-08:00

Here I post the components of my Hi-Fi systems for those who are interested. I shopped around a while and it took me a lot of listening and testing to find this set-up. May be its a help for someone who wants to set up its own system. Also I would appreciate all comments and ideas to improve mine.

Power Supply: AudioPlan PowerStar
Rack: Spider by finite elements
Speakers: B&W Nautilus 804 ->
Driven by a horizontal bi-amped set-up using:

(i) Integrated Amplifier: Electrocompaniet ECL-3 (drives woofers, serves as pre-amp)
(ii) Power Amplifier: Electrocompaniet AW60 FFT (drives tweeters and mid-ranges)

Both ampliefiers use the same internal set-up for the power amplifier. This makes them very suitable for be amping. Though, the needed some special adjustment/upgreat to adjust the output power (amplification factor) of both apmplifiers to the same level.

CD-Player: Audionet ART-V2

Characteristics:
Very focused, clear warm and transparent sound with a convincing stage. Most bad recordings sound too sharp. Tough, this is a general draw back af adiophile hi-fi systems.

Step by Step Howto on using BitTorrent to download files

2005-04-30T13:25:00.000-07:00

BitTorrent is a new peer to peer file sharing network protocol developed by Bram Cohen. If you lik it you should donate to him.

1. You will need a BitTorrent client to use BitTorrent.
Here are the most commonly used and probably the best:

Azureus A Java Based Bittorent
BitComet a powerful C++ client which is working automatically from behind NAT (Network Adress Translation, IP-masqherading) routers

Further not so commonly used yet very good stabele and fast BitTorrent clients which do not use much CPU power, though these are no eye catcher except may be ABC:

BitTorrent The origianl BitTorrent Client from the developper of the BitTorrent protocol Bram Cohen Thanks a lot for the protocol.
BitTornado A client extending the BitTorrent core thus supporting some experimental code for new sharing protocols.
ABC [yet another Bittorent Client] A client based on the BitTornado client with an imporved front end supporting imporved queing capabilities.

More information on not so common clients here.

2. What you need next is to find the torrent.

Search for Torrent Files

A torrent is a small file containing the meta data of the file you actually want to download. It is used to feed the BitTorrent client with the information it needs to download the file you want. You can find the most torrents here:

Each WebSite offers a search option. When you find more thant one torrent for the file you need then choose the one with the most seeders in relation to the leechers. Seeds have the complete file and only upload it whereas leachers have only partially downloaded the file and will usually upload to those peers where they get the most downloadbandwitdth from. Then choose the torrent with most participants. Note, Though it is possible that the swarm of leechers holds the complete file it is also possible that you wont be able to complete the file if you download a torrent without as seeder. A seeder in other words proves that all files of the torrent are present among the peers perticipating in the torrent.

3. Download the *.torrent file

Next you use the File download capability of your Browser to download the "*.torrent" file where * denotes the name of the torrent file which is usually a discription of the file you want to download. The download of the *.torrent file is usually initiated when you klick on the download or start download button next to the discription of the torrent. Most BitTorrent Clients integrate seemless into you browser since the browser downloads the torrent file and opens it with its default application which happens to be your Torrent client and so the download should start right away. If not, open the torrent-file saved to your hard drive from within the torrent client (e.g. Azureus or BitComet).

4. What happens now, and what does the tracker.

With the information the Torrent Client finds in the *.torrent file it will handle the download of the file you actually want. It will connect to the tracker, a server knowing all the IP-adresses of the peers sharing the file of a torrent, that is handling your torrent. Your client finds the IP-adress of the handling tracker in the *.torrent file you loaded into the client.
The Client will transmit (announce) its own IP adress to the tracker and recieve ip adresses of other clients who have the same file to share. Your client will then attempt to establish connections to peers (other computers running a torrent client) he got to know from the tracker to download the files you want.
Furthermore since your IP-Adress is announced by the tracker to other peers they will try to establish a remote connection to your peer.

5. Open ports 6881-6889

Since it is important that remote connections to your computer can be established to optimize download speed you should make sure that your computer can be reached from the internet. Using a Firewall ports 6881 to 6889 have to be opened. Otherwise inbound traffic comming from the internet to your computer will be blocked.

If you are using NAT these ports have to be forwarded to the IP-Adress of the mashine your torrent Client is running on. If these ports are not forwarded peers from the internet will only reach you router. The NAT router decides depending upon the port the packets are adressed to what to do whith the IP-packets. If they are adressed to the port so and so (eg 6881) the router will send those IP-packets to the mashine you are running your Client on if you activated port forwarding. A port will always be forwarded to only one ip-Adress behind the router.

Today many Bittorrent client support Universal Plug & Play (UnPnP) protocol. If you also have a router with an UnPnP option , you can enable the option on both client and router and the ports specified in your client will be forwarded to your mashine automatically.

With most bittorent clients it is possible to configure the port the client should connect to, so you don't have to use 6881 all the time. Changing ports might be useful for two reasons:
1. Internet providers (ISP) use traffic fiters to block bittorent traffic by blocking the standard ports of this protocoll.
2. I you want to use more that 1 client behind a NAT router assign an individual port range to each client and forward it to them by the router accordingly.

6. Tweak your upload speed

If you want to downlaod using a TCP connection then for each packet you recieve will be acknowledged by your computer sending a ACK packet to the sender of the packet your computer just recieved. My experiance is that 10 kB/s downloads use aproximately 1 kB/s upstream bandwidth. Then you have to conisder the system overhead. The system uses also some upstream bandwidth to keep up the connections and so on. So ist usually not a good idea to use all your upoadbandwith to upload to the torrent swarm because this makes downloads almost impossible. But remember that using BitTorrent you trade in your uploads for downloads from other peers so if your decrease ur upload bandwith too much you will loose download bandwith as well. So you have to find the maximium upload bandwith, which is not decreasing download bandwith.

A good idea is to use traffic shaping. This will give your ACK packages a higher priority and make sure they get through even tough you have uploads. This way you can set upload speed close to maximum and the traffic shaper will automatically reduce upload bandwidth for clients dynamically if more bandwith is used to maintain download speed. The following tools can accoplish this task:

1. cFos: traffic shaping for windows
2. wondershaper: traffic shaping for linux
3. prioritizing interactive traffic for linux

An aditional benefit of traffic shaping is that applications using HTML and FTP get a higher priority. This means that Web pages load almost as quick as they would load without bittorrent. Since bittorent dowloads take forever this keeps your connection usable at all times. Yet, for onlinegames like Doom you can still feel the slower connection although it becmes playable again by the use of traffic shapers.

The drawback of traffic shapers certainly is that they limit the maximal bandwidth. Therfore, traffic shaping makes no sense if you have only one application which needs the full bandwidth in one direction only. But the improvement on responsiveness of interactive application outweigth the perforamnce loss by far if

1. multiple applications with different sevice types/prioritis acces your internet connection at the same time or
2. if you are downloading and uploading at the same time.

7. Share

"give and you shalt recieve". Remember to share your downloaded files as long as you can. Otherwise the a peer to peer network doesen't work. Share files at least until you reach a share ratio of 1. The share ratio represents the upload to download ratio of your BitTorrent client:

filesize uploaded divided by filesize downloaded eg. :
1MB ul /2 MB dl = share ratio of 0.5 (50%)

Fixing hard disk errors

2005-04-26T03:53:00.000-07:00

I experienced a real bad hard disk crash recently and i was afraid to loose all my data. Fortunately I managed to fix the hard drive and to recover my data.

First the computer wouldnt boot to windows any more. It said shell32.dll was damaged. At the same time the hard disk made strange noises like taktaktaktak... and eventally the computer crashed. The reason for this behaviour was probably an overheated drive. I was running the UD client on my laptop and because it was standing on the couch the heat couldn't dissapate.

Even after letting the computer cool down for a day I was unable to boot: shell32.dll damage. So what did I do to fix the problem?

First I downloaded the emergency boot cd and burned the image to CD-Rom and booted the computer from CD-drive. The Disk contains a tool called Drive fitness test DFT which is a freeware utility form Toshiba. You can download it from there as well. I found plenty of corrupted sectors on my hard drive (10+) which where probalbly the cause of the problem. The DFT utility comes with a tool to repair the bad sectors. I run it two times on the hard drive before it corrected all sectors since the programm stops whith an error message if it find too many errors. After all fixable sectors were fixed i went on.

I booted into the recovery console of windows. But somehow I needed a administrator password although i was not aware of ever setting one. Anyhow I used a linux based tool on the emergency boot cd to remove the password an thus I managed to log into the recovery console to run "chkdisk /r". Again I repeated the chkdisk command until it reporded to have found no errors on the harddrive any more. Then I issued the "fixboot" command to fix the unmountable_boot_volume message which meanwhile apeared when I tried to boot my computer.

What I did next was to do a windows installation with the repair option to replace the broken shell32.dll. May be just too replace the broken file selectively would have worked too, but I figured a repair install would be more simple and replace all broken system files. After finishing installation and completing the crappy windows activation procedure I still had too fix some driver issues but the system was up and running ;).

The strange thing was that after a run of the diskeeper set-it and forget-it tool i had plenty of corrupted clusters on the hard drive again. Which I had to fix following the above mentioned procedure again. Since then im not running diskeeper only with smart I/O option disabled since I figured this too be the main difference between manual defragmentation which caused no problems and the set-it and forget-it defragmentation which messed up my harddrive.

Up to now my hard disk runs smoothly and no tool chkdsk, drive fitness test or the seatools
could find an error on my hard drive at all.

Using Sony Clie PEG NR70V as a Memory-Stick Reader under Linux

2005-02-17T11:08:00.000-08:00

Make sure appropriate Kernel and modules are installed: You will need usbserial, usb_storage, usbcore, sg and scsi_mod( #lsmod). Start the MS-import utility on the Clié. Check If Clié was detected as scsi device ( #cat /proc/scsi/scsi). If not, leave MS-Import running and turn off the Clié and swich on again. This is because the MS-Imprort starts too slow to be recoginiszed by Linux as scsi device sometimes. If the Sony was detected as a scsi device you can just mount it.

I reacently tried to install the card reading option of my sony clie nr70V so that it workes for linux. I was amazed as I realized, that linux has native support for the card reading option of the sony clie.
What I did was the following:
Just plug the sony into the craddle which is connected to some USB port of the linux box. Then turn on the clie and run MS-import on the sony. Than linxux shoud automatically detect the PDA as a USB mass storage device and mount it if you have hotplugging functionallity installed.
But on my device it didnt work. Then I Googled a bit and found out that a lot of users had the same problem, althogh the have the appropriate Kernel and modules installed ( You will need usbserial, usb_storage, usbcore, sg and scsi_mod)
but I couldnt find the solution.
Then I checked /var/log/messages. It reads:
-snip-
Feb 17 21:35:59 linux kernel: usb 3-1: new full speed USB device using address 4
Feb 17 21:35:59 linux kernel: usb 3-1: Product: Sony PEG Mass Storage
Feb 17 21:35:59 linux kernel: usb 3-1: Manufacturer: Sony
Feb 17 21:35:59 linux /etc/hotplug/usb.agent[6143]: need a device for this command
Feb 17 21:35:59 linux kernel: scsi2 : SCSI emulation for USB Mass Storage devices
Feb 17 21:35:59 linux kernel: USB Mass Storage device found at 4
-snip-

So the device was recognized but it was not installed as a scsi device (Generic scsi - sg module)
# cat /proc/scsi/scsi revealed no entries. This probalbly means the USB device is recognized but scsi support is not installed. Actually I never figured out why, yet I found a work around.

What I did was the following:

Start and Leave MS-Import running on the palm while connected to the PC. Turn off the palm.

kernel messages:
-snip-
Feb 17 21:51:16 linux kernel: usb 3-1: USB disconnect, address 4
Feb 17 21:51:16 linux /etc/hotplug/usb.agent[6345]: need a device for this command
Feb 17 21:51:17 linux kpilotDaemon: resmgr: server response code 501
-snip-

Then turn on the palm again, e voila:

-snip-
Feb 17 21:52:04 linux kernel: usb 3-1: new full speed USB device using address 5
Feb 17 21:52:04 linux kernel: usb 3-1: Product: Sony PEG Mass Storage
Feb 17 21:52:04 linux kernel: usb 3-1: Manufacturer: Sony
Feb 17 21:52:04 linux /etc/hotplug/usb.agent[6398]: need a device for this command
Feb 17 21:52:04 linux kernel: scsi3 : SCSI emulation for USB Mass Storage devices
Feb 17 21:52:05 linux kernel: Vendor: Sony Model: CLIE MGMS Rev: 1.00
Feb 17 21:52:05 linux kernel: Type: Direct-Access ANSI SCSI revision: 02
Feb 17 21:52:05 linux kernel: SCSI device sde: 253696 512-byte hdwr sectors (130 MB)
Feb 17 21:52:05 linux kernel: sde: assuming Write Enabled
Feb 17 21:52:05 linux kernel: sde: assuming drive cache: write through
Feb 17 21:52:05 linux /etc/hotplug/block.agent[6458]: try 1 while waiting for /block/sde's bus_id 3:0:0:0
Feb 17 21:52:05 linux kernel: sde: sde1
Feb 17 21:52:05 linux kernel: Attached scsi removable disk sde at scsi3, channel 0, id 0, lun 0
Feb 17 21:52:05 linux kernel: Attached scsi generic sg4 at scsi3, channel 0, id 0, lun 0, type 0
Feb 17 21:52:05 linux kernel: USB Mass Storage device found at 5
Feb 17 21:52:05 linux /etc/hotplug/block.agent[6480]: new block device /block/sde/sde1
Feb 17 21:52:06 linux /etc/hotplug/block.agent[6458]: waiting for /var/lock/block.agent.lock, process 6480 holds it
Feb 17 21:52:13 linux /etc/hotplug/block.agent[6480]: mount by-path/usb-storage-odd-Sony-SonyPEGMassStorage:0:0:0p1
Feb 17 21:52:14 linux /etc/hotplug/block.agent[6458]: new block device /block/sde
-snip-
and
# cat /proc/scsi/scsi
Attached devices:
ANSI SCSI revision: 02
Host: scsi3 Channel: 00 Id: 00 Lun: 00
Vendor: Sony Model: CLIE MGMS Rev: 1.00
Type: Direct-Access ANSI SCSI revision: 02

And a window opend on by KDE by whichcraft which revealed all the contents of the memory stick of my Sony Clie. Good!

PC security: Firewalling, virus protection and adware protection for free.

2005-01-12T11:22:00.000-08:00

Today you need to secure your PC when you're connected to the internet. Here I'll post a collection of software you can use for free which will provide the nessecary security. I think you'll need (i) a firewall (ii) a virus scanner (iii) a trojan scanner (iv) and last not least a spyware scanner.

First you should start by setting up a firewall. This will secure your PC from direct attacks from the internet like portscan. You could use the windows firewall, but I prefer using Zone Alarm. This is a free personal Firewall wich listens on all ports of your PC and asks you to permit outgoing or incoming traffic based on the programs which asks for access. This makes this firewall pretty easy to configure since you don't have to mess with ports. Additionally bacause of the logging capability you get an overview of the traffic on your PC. By the way, make sure that windows file and print server are not associated with the tcp/ip protocoll of the network device you established the connection to the internet. Zone Alarm will provide you with some firewall capabilities. If you want to go 1 step further you should use a router as firewall which supports NAT (network adress translation) or ip-masqureading. This technology is supposed to be very secure since you're PC is invisible from the internet. The firewall should moreover not reply to ICMP service requests like finger or ping because that way the hacker probably won't even notice that your online.

Next I would suggest you install a virus scanner. The best scanner I know which is free for personal use it AntiVir. This software will includes free updates of the virus definitions. In addition to scanning capabilities the software comes with an virusguard running in deamon mode controlling all files accessed (read or write) during a session for putative virus infections. In addition he can detect some trojans.

next step would be to install a trojan scanner. The only software for free I found so far is a-squared. After installation you have to register before you can use it. After this you can scan your complete harddrive and updates of trojan signatures are free. Trojan background guard is disabeled in freeware mode though.

Last I would install an anti spyware tool. I prefer Spybot S&D. For my feeling it has the most features and offers most comprehensive protection. Best of all, its absolutely free.

If you use Windows-XP you could furthermor disable administrator mode for the account you surf the internet with. This limits access to crucial files like registry for malicious software and provide additional security.

Additinally I would stop to use outlook and Internet Explorer since these pieces of software are vulnerable and are often attacked by hackers. If you use for example firefox, thunderbird or the mozilla suite you have less risk beeing attacked, since this software is more secure for the time beeing.

So this is the secury software I would suggest to make you surfing safer. I you have a question just post a comment and I'll update this post if I can help.

Microsoft came from the stone age and trys to bring us back there !!

2005-01-09T13:22:00.000-08:00

Does anyone know if Microsoft has ever invented a piece of software technology? I can't think of one.
MS- Dos was bought from some other company and was originally named QD-Dos. QD-DOS which means quick and dirty disk operating system was a quick and dirty remake of CPM. Mulitmedia Application for PCs where first developped by Apple. Graphical user interfaces mulit-tasking and windows technology where first invented by Apple.
Not only that Microsoft doesen't invent anything it takes them forever to implement innovations into their software. They only start to improve their software if they spot a competitor. In this case their improvement is ment to destroy this competitor. A good example for this behavior is the browser war wich is now going into the second round.
The business practice of Microsoft is meant to destroy every competitor. This developement would mean that their would't be any improvement in software at all in the future. We would have to get used to buggy unstable und incomfortable software.
Good luck we have the open source community. This is our last chance to have better software in the future. As a result of their non-profit nature, Microsoft will never be able to destroy the open source project. I would advice everybody to support the open source project if he is interested in progress of software technology and comfortabel software.

If you are a internet explorer user a first step would be to switch to firefox. This is the most modern, secure, comfortable and configurabel browser you can get at the moment.

Violoence on TV and in computer Games causes death!

2005-01-08T13:27:00.000-08:00

Why would somebody kill a man? Usually not for fun, I guess. Yet, in Egoshooters like Doom, Unreal Tournament, Counterstrike and others people kill for fun.

We all know this is not real, only computer games. Though, players with vulnerable personality might loose the contact to reality. By playing egoshooters the player learns that killing is fun and that killing very efficient makes him a winner. In conclusion killing makes him feel good and gives him the feeling of superiority.

Most people can distinguish between game and reality, but how can you rule out that that some can't? Those people who can't, might be out of the loop. They don't have the oportunity to experience different ways to assert themselves. They don't make many experiences to improve their social skills. They might be underdogs and don't know how to cope with their agressions. The only teacher for them who teaches them social skills and coping strategies is TV and computer games. What do they learn there? Violence! They get used to violence and killing. For them murder looses its terror.

So what will such a person do if he is frustrated, agressive or has problems and is seeing no future for himself??? He will do the only thing he knows! He will react agressively. Eventually he will kill, because he learned from the egoshooters killing makes him feel good and he thinks the same behoviour might have the same effect in real life. From TV he knows that the heros in the movies kill to achieve their goals. Then why not just try to immitate what he saw?

For a mature personality with an intact social background TV and violent computer games are no danger. Yet, for a vulnerable personality they sure are and it might trigger tragedies like in columbine.