Klaus Knopper is the creator of Knoppix and co-founder of the LinuxTag expo. He currently works as a teacher, programmer, and consultant. If you have a configuration problem, or if you just want to learn more about how Linux works, send your questions to: email@example.com
I want to test the latest Knoppix Live CD, but I'm having problems with the screen resolution.
I have an nVidia graphics card and an Acer screen (resolution 1440x900). When I boot the Live CD, the best resolution I get is 800x600. I have booted with knoppix screen=1440x900 and knoppix fb1440x900, but so far I am without success. Also, I have tried xrandr --output VGA --mode 1440x900 without any luck. Do you have any suggestions?
Newer X.org versions use the randr extension, which lets you change screen resolution without restarting the X server, having multiple screens on different monitors and setting up two or more desktops beside each other.
This extension is automatically active with no additional configuration. In fact, the previous Modes configurations in /etc/X11/xorg.conf is completely ignored. The X server automatically detects monitor resolutions and configures what it thinks are the best resolution settings.
Utilities such as xrandr let you change display settings on the fly.
That's the good part.
The bad part is that if anything goes wrong in this automatic setup (for example, incorrect detection of supported screen resolutions), the problem might be more difficult to fix unless you just disable randr, set up the old-style modelines, and restart the X server.
I'll start with (x)randr:
It could be that the X server detects a maximum resolution that is too small for your graphics card/display.
To find out, enter
for all detected resolutions. The output will be similar to Listing 1.
xrandr -q Output
01 Screen 0: minimum 320 x 200, current 1400 x 1050, maximum 1400 x 1200 02 VGA-0 disconnected (normal left inverted right x axis y axis) 03 LVDS connected 1400x1050+0+0 (normal left inverted right x axis y axis) 0mm x 0mm 04 1400x1050 60.8*+ 05 1280x1024 59.9 06 1280x800 60.0 07 1280x768 60.0 08 1024x768 60.0 59.9 09 800x600 60.3 59.9 10 640x480 59.9 59.4 11 S-video disconnected (normal left inverted right x axis y axis)
In this example, 1400x1050 is detected as the highest possible resolution and has been chosen as the current value.
If your desired resolution of 1440x900 is not listed, xrandr will refuse to set it. If 800x600 is listed as the highest possible mode, X might not reserve enough memory to support higher resolutions. To fix this problem, add the line
Virtual 2048 2048
to the Display subsection of /etc/X11/xorg.conf. This line will let the system switch to a higher resolution if a new monitor is connected. For some Intel graphics chipsets, virtual screen sizes of modes higher than 2048x2048 will disable DRI, and Compiz won't run.
Because you have an nVidia card, the problem might just be that the current xorg nv driver cannot handle the 1440x900 resolution, compared with the better supported ATI and Intel card drivers. nVidia still has a restrictive policy about specifications and API descriptions for their cards, which makes it hard to develop an open source driver that supports all features. Therefore, you might be better off using a proprietary nVidia Linux driver that you can download and install from the nVidia web page.
For Knoppix, if installing the nVidia driver, start in text mode with
and run the installation from there. If you have booted from USB flash memory with the persistent Knoppix image feature, your changes will be written to flash memory and will not be lost when restarting the computer. After completing the installation, type
to switch to the graphical desktop. The xrandr command should now report all resolutions supported by the new driver, and you can switch to native resolution with
xrandr --output VGA --mode 1440x900
as you tried before. Instead of --mode, you might have to use the -s flag if the nv driver only supports randr 1.1 instead of 1.2.
An alternative, if you can get ahold of a modeline for your desired resolution, is to add a new mode via xrandr:
xrandr --newmode "1440x900" 106.5 1440 1520 1672 1904 900 901 904 932 -HSync +VSync xrandr --addmode VGA 1440x900 xrandr --output VGA --mode 1440x900
Now compare the modeline given in the first xrandr command with the specification for your graphics card and monitor and change it accordingly.
The standard nv X.org module might or might not accept the new modeline.
Eee PC Keyboard Combinations
How can I turn on/off the card reader, WiFi, webcam on my Eee PC? The keyboard combinations don't seem to work.
Kernels 2.6.28 and up feature special hardware support for the Asus Eee PC and other systems through the virtual /sys directory. By writing to the files shown in Table 1, you can control parts of the computer. To get the current setting, use the cat command:
cat /sys/devices/platform/eeepc/camera 1
In this example, the value returned as 1 indicates that the webcam is powered up and therefore available to programs (provided the uvcvideo driver is installed), so you could use
to display a webcam video.
To change the state of a device, use echo (as root, because of write access):
echo 1 > /sys/class/rfkill/rfkill0/state
This command will activate the integrated WiFi card and also turn on the blue LED. The udev daemon should now automatically load the WiFi driver (ath5k), and Network Manager will start scanning for networks. This technique is not Eee PC--specific; it might also work for other notebooks with newer Atheros WiFi chipsets.
The reason the keyboard combinations don't work with some distributions is their default /etc/acpi/ configuration, which is sometimes set very conservatively, so you might want to add these combinations by monitoring ACPI events during a key press and putting matching scripts for hardware activation in /etc/acpi.
USB Flash Drive
I have been experimenting with installing and booting Linux from flash drives. For this purpose, I bought myself a cheap 8GB USB stick on which is written ZIXV. The device reports to /var/log/ as:
scsi 3:0:0:0: Direct-Access CION AP192 1.01 PQ: 0 ANSI: 4 sd 3:0:0:0: [sdb] 15925248 512-byte hardware sectors (8154 MB)
I was able to repartition this drive (ext2 fs) for Linux using gparted, and I successfully installed (and booted) Ubuntu (Intrepid) from it. But then I had some problems (not related to the drive). I attempted to install an older version of Mandrake Linux. During installation, the partitioner program failed, and thereafter, I have not been able to access the flash drive at all. I include the full log below (Listing 2); the kernel drivers cannot read the partition block on the drive, so it must have been corrupted.
If the kernel drivers are unable to pick up /dev/sdb, no other program (such as gparted) is able to access the drive either. So now my question is: Is there a way to restore a "raw" partition table such that the kernel drivers can pick up /dev/sdb and enable the use of gparted? I have tried using dd without success.
USB Stick Problems
01 Mar 13 16:09:13 M2-laptop kernel: [ 488.512962] usb 2-3: not running at top speed; connect to a high speed hub 02 Mar 13 16:09:13 M2-laptop kernel: [ 488.531979] usb 2-3: configuration #1 chosen from 1 choice 03 Mar 13 16:09:13 M2-laptop kernel: [ 488.535387] scsi4 : SCSI emulation for USB Mass Storage devices 04 Mar 13 16:09:18 M2-laptop kernel: [ 493.544174] scsi 4:0:0:0: Direct-Access CION AP192 1.01 PQ: 0 ANSI: 4 05 Mar 13 16:09:18 M2-laptop kernel: [ 493.555151] sd 4:0:0:0: [sdb] 15925248 512-byte hardware sectors (8154 MB) 06 Mar 13 16:09:18 M2-laptop kernel: [ 493.562137] sd 4:0:0:0: [sdb] Write Protect is off 07 Mar 13 16:09:18 M2-laptop kernel: [ 493.612181] sd 4:0:0:0: [sdb] 15925248 512-byte hardware sectors (8154 MB) 08 Mar 13 16:09:18 M2-laptop kernel: [ 493.619180] sd 4:0:0:0: [sdb] Write Protect is off 09 Mar 13 16:09:49 M2-laptop kernel: [ 493.619765] sdb:<6>usb 2-3: reset full speed USB device using ohci_hcd and address 10 10 Mar 13 16:10:19 M2-laptop kernel: [ 554.544029] usb 2-3: reset full speed USB device using ohci_hcd and address 10 11 Mar 13 16:10:50 M2-laptop kernel: [ 585.280030] usb 2-3: reset full speed USB device using ohci_hcd and address 10 12 Mar 13 16:11:00 M2-laptop kernel: [ 595.700029] usb 2-3: reset full speed USB device using ohci_hcd and address 10 13 Mar 13 16:11:11 M2-laptop kernel: [ 605.944055] usb 2-3: USB disconnect, address 10 14 Mar 13 16:11:11 M2-laptop kernel: [ 605.946247] sd 4:0:0:0: Device offlined - not ready after error recovery 15 Mar 13 16:11:11 M2-laptop kernel: [ 605.946268] sd 4:0:0:0: [sdb] Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK,SUGGEST_OK 16 Mar 13 16:11:11 M2-laptop kernel: [ 605.946339] Dev sdb: unable to read RDB block 8 17 Mar 13 16:11:11 M2-laptop kernel: [ 605.946398] unable to read partition table 18 Mar 13 16:11:11 M2-laptop kernel: [ 605.946530] sd 4:0:0:0: [sdb] Attached SCSI removable disk 19 Mar 13 16:11:11 M2-laptop kernel: [ 605.946643] sd 4:0:0:0: Attached scsi generic sg2 type 0
The syslog messages in Listing 2 do not look good. The USB drive might not be able to handle repartitioning (i.e., changes of the first sector). This might not be a problem with the storage itself. Possibly, the integrated controller is not able to handle different partition tables other than the default. Another (equally unlikely) problem might be the connection between the flash stick and the computer that causes transmission errors, or it could be a USB speed set too high. But at this capacity, the flash drive should be USB 2.0 capable. Usually, flash drives of 16GB and higher seem to have performance problems with write operations.
In any case, if the flash drive locks up during a write procedure, resist the urge to just pull it from the computer. Many flash drives enter a yet unrecoverable mode if they are disconnected during a write. Some card readers have a file under /sys that allows you to shut down the reader without physically disconnecting the card:
echo 0 > /sys/bus/platform/devices/eeepc/cardr
(Also, see the question below in "Eee PC Keyboard Combinations.")
For USB flash, you can try scsi-spin or eject to force an already killed write attempt to really terminate,
scsi-spin -d /dev/sda
eject -f /dev/sda
provided the flash drive is connected as /dev/sda. Both commands might wait until the currently writing process is either complete, or at least until it ceases to lock the device, which takes a long time. Just watch the output of dmesg for timeout or device reset errors.
Now the problem might be how to get the partition table back into its initial state after a failed partitioning attempt. If the flash drive is still shown with its correct size with fdisk -l /dev/sda, it could help to repartition it with fdisk as root (for the final ioctl system call that tells the kernel to physically reread the partition table) and then use one or both of the above commands to tell the kernel to complete all operations on this device and remove it from the system. Even then, flash drives might need some time to reorganize segments internally, so take a flashing/blinking light on the drive as a hint of activity, even if Linux tells you the device is completely synchronized and can be removed.
If you are not afraid of the cost, you can buy the exact same flash drive again and copy its partition table
dd if=/dev/sda of=sda.mbr bs=512 count=1
and, after switching out the new drive for the one that has the defective partition table, use the same command with reversed in/out options:
dd of=/dev/sda if=sda.mbr bs=512 count=1
Again, make sure you are reading/writing the right device files. Check dmesg after you insert the drive for the detected device file name.
As with other USB stick operations, when you use the eject and scsi-spin commands, make sure all writing activity is finished before removing the stick.
In general, changing the original partition table might just be impossible for some flash media (because of bad hardware design). In my opinion, this is a warranty case, and you should be able to get a free replacement.
The default partitioning of flash drives, which is usually a single primary FAT16 or FAT32 partition, is suitable for booting Linux. For example, you can use
to create the boot loader on the first partition. A bootable master boot record is created with
ms-sys -s /dev/sda
but the same caveat about changing the partition table applies.
Because some USB controllers will still boot from the first partition when they find an empty MBR, you might want to try going with the default empty MBR first.
Buy this article as PDF
Linux Foundation's big event celebrates the 25th anniversary of Linux
Linux has evolved from a “won’t be a professional” project to one of the most professional software projects in the history of computers.
Competitors get in the game with RHEL without Red Hat
Security researchers have already notified Microsoft; some fixes are available
The company is collaborating with Google and Intel to use Kubernetes as an engine for Fuel
Customers can take a free test drive of SLES for HPC on the Azure Cloud
San Francisco-based chip company announces their first fully open source chip platform.
The whole distro gets rebuilt on glibc 2.3
Ubuntu Vendor tries to solve app packaging and distribution problem across distributions.