Loading Linux Images to a Compact Flash Disk

When changing firmware images or creating custom firmware images for your Compact Flash-based machine, it is necessary to perform a few steps to load the firmware image onto the CF card.

Necessary Tools

To perform this procedure, the following items are required:

• The Compact Flash card to hold the firmware.
• A Compact Flash card reader attached to a computer.
• The filesystem tar.
• A fully functional Linux desktop.
• Linux tools installed: tar, and either gzip or bzip2 (depending on the image).
• The EMAC put-image package which provides the put-image script. The README.TXT file in the package details the steps needed to install the script properly.

Unpack the put-image tar after downloading:

developer@ldc:~$ tar zxvf ~/Downloads/put-image.tar.gz

Procedure

Perform the following steps to load the firmware onto the Compact Flash card:

• Insert the Compact Flash card into the card reader.
• Navigate (from within the shell) to the directory which contains the firmware image to be used.
• Determine where the Compact Flash card was mounted.

developer@ldc:~$ dmesg | tail -n 10
[23236.042944] sdc: detected capacity change from 4110188544 to 0
[23243.783467] sd 12:0:0:0: [sdc] 8027712 512-byte logical blocks: (4.11 GB/3.82 GiB)
[23243.785199] sd 12:0:0:0: [sdc] No Caching mode page present
[23243.785204] sd 12:0:0:0: [sdc] Assuming drive cache: write through
[23243.787314] sd 12:0:0:0: [sdc] No Caching mode page present
[23243.787326] sd 12:0:0:0: [sdc] Assuming drive cache: write through
[23243.790625]  sdc: sdc1
[23244.290093] kjournald starting.  Commit interval 5 seconds
[23244.293646] EXT3-fs (sdc1): using internal journal
[23244.293651] EXT3-fs (sdc1): mounted filesystem with ordered data mode
developer@ldc:~$

The output of the dmesg command shows that the root device node for the Compact Flash card in this case is sdc, and that the only partition found is sdc1. Inspect the output shown above to see the particular device node on the development PC where the Compact Flash card is mounted. This will be different depending on the configuration of the development PC. The /dev/ prefix will go before the device name, but is not shown in the output of dmesg. The device node will start with either sd or hd. The third letter will specify which of these devices is assigned to the CF card.

• At this point, it is wise to double check that device node is actually the CF card and not a hard drive in the system. Specifying the wrong device node could cause a complete loss of data on a hard drive.
  
  Continuing with /dev/sdc1 as the example device node, type the following command:

developer@ldc:~$ mount | grep sdc1
/dev/sdc1 on /media/EMAC-OE type ext3 (rw,nosuid,nodev,uhelper=udisks)

This shows that the device node, /dev/sdc1, is in fact the Compact Flash card. This is because:

• It is not mounted on one of the standard Linux filesystem mountpoints, such as /, /boot, /usr, or /home.
• It is mounted in the /media directory, where it is expected. It may alternatively get mounted in /mnt, depending upon the configuration of the Linux distribution in use.

• Use fdisk -l to inspect the device node. Use sudo with fdisk to gain the required root privileges to run fdisk

developer@ldc:~$ sudo fdisk -l /dev/sdc
Disk /dev/sdc: 4009 MB, 4009549824 bytes
77 heads, 56 sectors/track, 1816 cylinders, total 7831152 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00090707 
  
 Device Boot      Start         End      Blocks   Id  System
/dev/sdc1              62     7831151     3915545   83  Linux
developer@ldc:~$

• As can be seen from the output of the fdisk -l command above, the disk is 4009 MB in size, which corresponds with the size of the 4 GB Compact Flash being used in this example.

• The device to specify for the target boot argument must also be determined prior to installing the image onto the Compact Flash card. Images obtained from EMAC will contain a text file that specifies what boot device node to use. For a custom image, use the fdisk -l command on the target device to determine the boot device. In this example, the target hardware specifies /dev/hdc for the boot device.

• Now, run the put-image script as below.
  • The --target option specifies where the card is mounted.
  • The --boot option specifies the target boot argument.
  • The last option is the file name of the compressed target root filesystem.

After completion, the put-image script will output a message indicating success or failure.

developer@ldc:~$ /path/to/put-image --target=/dev/sdc --boot=/dev/sda emac-firmware-image.tar.gz

• The put-image script should automatically unmount the Compact Flash card after the process completes. It is best to check that the Compact Flash card is actually unmounted before unplugging it. Using the /dev/sdc1 example from above, type the following command.

developer@ldc:~$ mount | grep sdc1

• If the command does not return anything, the Compact Flash card is not mounted and can be removed and inserted into the target device. However, if the command returns something similar to the following listing, the card will need to be unmounted manually.

/dev/sdc1 on /media/EMAC-OE type ext3 (rw,nosuid,nodev,uhelper=udisks)

• To unmount the Compact Flash card manually, run the following command:

developer@ldc:~$ umount /dev/sdc1

See Also

• Linux Boot Process Customization