Monday, June 21, 2010

Fixing Wireless Updated

Reader llenchikk left a comment stating that installing linux-backports-modules-wireless-lucid-generic would fix the wireless problem but for me it didn't. First I uninstalled the self compiled modules via 'make uninstall' and then I installed the backports package and rebooted but the wireless behaved as after the clean install. Even scanning is not possible.

Thanks to reader bash0 the current compat-wireless does compile again. He/She figured out a bug in one of the scripts and provided a fix in the comments of article Fixing wireless:

To solve that, you'll have to change one line in scripts/gen-compat-autoconf.sh. It is written:
if [ $SUBLEVEL -lt 32 ]; then

I changed it to
if [ $SUBLEVEL -le 32 ]; then

It works for me and now I enjoy the latest bleeding edge driver. Thank you bash0.

Thursday, May 6, 2010

Internal microphone workaround

I found a workaround for the internal microphone. The problem is that the current driver sees the microphone as a stereo device, so you have to make it mono. After the workaround it's usable but it has a lot of background noise.

1. Install pulseaudio volume control
sudo apt-get install pavucontrol
2.  Open it
Applications -> Sound & Video -> PulseAudio volume control 
3. Goto tab  'Input Devices' and select 'Hardware Input Devices' from the Show control at the bottom.

4. Unlock the stereo channels by clicking on on lock icon

5. Adjust one channel to zero and the other to 90% to 100%


Skype


1. Install Skype

2. Go to Skype's Options -> Sound Devices and uncheck 'Allow Skype to automatically adjust mixer levels'

3. Hit apply and make a test call


Kernel update 2.6.32-22

Today I've checked the ubuntu updates again and... voila there was a new kernel. So I downloaded the current compat-wireless to prepare for my wireless fix and I updated.

Wireless was still not working , so I compiled compat-wireless bleeding edge: compat-wireless-2010-05-05. Works. Interesting is that two finger scrolling works now on the mousepad. But honestly I fiddled so much around to have this working that I don't know now if it's because of my workarounds or the new kernel.

Is this working for you guys now? Please tell me.

Chromium doesn't drain battery

Well, as I said I filed a bug but soon I learned that I didn't think of the possibility that one of my extensions could be the reason. And so it was. My beloved google mail checker 'Google Mail Checker Plus" was the trouble. Since I uninstalled it chrome is running fine. I posted a report at the developer's forum.

Monday, April 26, 2010

Chromium drains battery

I love google's chrome browser and since I found a corresponding extension to all my firefox addons I really switched to chrome. So I'm using it on my netbook, too. But recently I found out that chrome causes way too much CPU wake ups. I used powertop for the analysis rather than the battery applet's Processor information.

Idling, with chromium running, one page open (my blog):

And now compare with firefox. Also idling with only my blog opened:

As you can see firefox doesn't even appear in the list of processes waking up the CPU when idling. The CPU stays in the C4 state for 97.2% of the time, which is quite decent. But chromium wakes up the CPU 217 times per second and makes more CPU load balancing necessary, too. All in all it makes the CPU staying in C4 state 77.3 % of the time, only, which is very bad.

Well, that means that I'll be using firefox again. I guess I'll file a bug about that.

Update
Well, as I said I filed a bug but soon I learned that I didn't think of the possibility that one of my extensions could be the reason. And so it was. My beloved google mail checker 'Google Mail Checker Plus" was the trouble. Since I uninstalled it chrome is running fine.

Latest compat-wireless works again

As reader Cliff Wells mentioned in his comment to the article Fixing wireless my solution works again with the current compat-wireless bleeding edge sources compat-wireless-2010-04-12 and the lucid kernel 2.6.32-21.

Friday, April 2, 2010

Panning screen at 1024x768

Sometimes I encountered the problem that I had a dialog on the screen which was too large for the display's native 1024x600 resolution. In windows the EeePC tools offer a panning mode that switches to a virtual 1024x768 resolution where the visible section of the screen follows the mouse pointer. So I thought it must be possible to do the same with xrandr. Some playing around with xrandr was a success. Searching for solutions of others led me to a convenience script:

#!/bin/bash

function pan-mode {
echo `xrandr --prop | grep "current"`
}

function pan-enable {
echo Enabling VGA output
xrandr --output LVDS1 --panning 1024x768
}
function pan-disable {
echo Disabling VGA output
xrandr --output LVDS1 --panning 1024x600
}

### MAIN ###
case $1 in 
on)  pan-enable ;;
off) pan-disable ;;
status)  pan-mode ;;
*)   echo "*usage: $0 on|off|status"    ;;
esac
Create an executable shell script in the bin dir:
  1. gedit ~/bin/panning (then paste the script into the file and save)
  2. chmod a+x ~/bin/panning
Now you can switch with panning on or panning off in the console.


If you want to make the panning default create or edit the file ~/.xprofile and add the following line to it:
xrandr --output LVDS1 --panning 1024x768
Next time you reboot or restart X the screen will be in panning mode.