U-Boot Reference Manual

90000852_K
Digi International Inc. 2011. All Rights Reserved.
The Digi logo is a registered trademark of Digi International, Inc.
All other trademarks mentioned in this document are the property of their respective owners.
Information in this document is subject to change without notice and does not represent a commitment on the part of Digi International.
Digi provides this document as is, without warranty of any kind, either expressed or implied, including, but not limited to, the implied
warranties of fitness or merchantability for a particular purpose. Digi may make improvements and/or changes in this manual or in the
product(s) and/or the program(s) described in this manual at any time.

This product could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein;
these changes may be incorporated in new editions of the publication.

Digi International Inc.

11001 Bren Road East
Minnetonka, MN 55343 (USA)
+1 877 912-3444 or +1 952 912-3444
http://www.digi.com

2
Contents

1. Conventions used in this manual ................................................. 5

2. Acronyms and Abbreviations ....................................................... 6
3. Introduction .................................................................................... 7
3.1 What is a boot loader? ........................................................................................................... 7
3.2 What is U-Boot?..................................................................................................................... 7
3.3 Features of U-Boot ................................................................................................................ 7
3.3.1 Customizable footprint ......................................................................................................................... 7
3.3.2 Monitor ................................................................................................................................................. 7
3.3.3 Variables .............................................................................................................................................. 8
3.3.4 Ethernet and USB ................................................................................................................................ 8
3.3.5 Numbers............................................................................................................................................... 8
3.4 The boot process ................................................................................................................... 8

4. U-Boot commands ......................................................................... 9

4.1 Overview................................................................................................................................ 9
4.2 Built-in commands ................................................................................................................. 9
4.2.1 Information commands ...................................................................................................................... 10
4.2.2 MII commands ................................................................................................................................... 11
4.2.3 Network commands ........................................................................................................................... 11
4.2.4 USB commands ................................................................................................................................. 12
4.2.5 Memory commands ........................................................................................................................... 12
4.2.6 Serial port commands ........................................................................................................................ 14
4.2.7 I2C commands ................................................................................................................................... 14
4.2.8 Environment variables commands .................................................................................................... 14

5. Environment variables ................................................................ 15

5.1 Overview.............................................................................................................................. 15
5.2 Simple and recursive variables ............................................................................................ 15
5.3 Scripts.................................................................................................................................. 15
5.4 System variables ................................................................................................................. 16
5.4.1 Common system variables ................................................................................................................ 16
5.4.2 Dynamic variables..............................................................................................................................17
5.4.3 User keys ........................................................................................................................................... 17
5.4.4 Protected variables ............................................................................................................................ 18

6. Bootscript ..................................................................................... 19
6.1 Bootscript process ............................................................................................................... 19
6.2 Creating a bootscript ............................................................................................................ 19
6.2.1 Creating a bootscript in Windows ...................................................................................................... 20
6.3 Configuration for launching the bootscript ............................................................................ 20
6.4 Bootscript restrictions .......................................................................................................... 20

7. Boot commands ........................................................................... 21

7.1 Overview.............................................................................................................................. 21
7.2 Reading images into RAM ................................................................................................... 21
7.2.1 From Ethernet .................................................................................................................................... 21
7.2.2 From USB .......................................................................................................................................... 21
7.2.3 From MMC ......................................................................................................................................... 21
7.2.4 From flash .......................................................................................................................................... 22
7.3 Booting images in RAM ....................................................................................................... 22
7.4 Direct booting ...................................................................................................................... 23
7.4.1 Boot examples ................................................................................................................................... 23
7.5 Automatic booting ................................................................................................................ 23

8. Using NVRAM ............................................................................... 25

3
 8.1 The 'flpart' command ........................................................................................................... 25
8.1.1 A partition table entry ......................................................................................................................... 25
8.1.2 Changing the partition table ............................................................................................................... 26
8.2 The intnvram command...................................................................................................... 26
8.2.1 Mappings of variables ........................................................................................................................ 28

9. Firmware update commands ...................................................... 29

9.1 Overview.............................................................................................................................. 29
9.2 Updating flash with images in RAM ..................................................................................... 29
9.3 Direct updating..................................................................................................................... 30
9.3.1 Update limits ...................................................................................................................................... 31

10.Customize U-Boot ........................................................................ 32

10.1Overview.............................................................................................................................. 32
10.2JTAG Console ..................................................................................................................... 32
10.3Silent Console...................................................................................................................... 33
10.4Splash screen support ......................................................................................................... 33
10.4.1 Enable splash screen support in U-Boot ........................................................................................... 34
10.4.2 Creating a splash partition ................................................................................................................. 34
10.4.3 Uploading a splash image ................................................................................................................. 35
10.4.4 Initialize video interface ..................................................................................................................... 35

11.U-Boot development .................................................................... 37

12.Troubleshooting ........................................................................... 38
12.1Work with FIMs on NS921x based modules ........................................................................ 38
12.2Problems booting from NFS ................................................................................................. 38
12.3USB memory sticks ............................................................................................................. 38

4
1. Conventions used in this manual
This list shows the typographical conventions used in this guide:
Style Used for file and directory names, variables in commands, URLs
and new terms.

Style In examples, to show the contents of files, the output from

commands, the C code.
Variables to be replaced with actual values are shown in italics.

Style Variable's names and commands.

In examples, to show the text that should be typed literally by the
user.
# A prompt that indicates the action is performed in the target device.
$ A prompt that indicates the action is performed in the host
computer.
<field> A mandatory field that must be replaced with a value
[field] An optional field
[a|b|c] A field that can take one of several values

This manual also uses these frames and symbols:

This is a warning. It helps solve or to avoid common mistakes

or problems

This is a hint. It contains useful information about a topic

$ This is a host computer session

$ Bold text indicates what must be input

# This is a target session

# Bold text indicates what must be input

This is an excerpt from a file

Bold text indicates what must be input

5
2. Acronyms and Abbreviations
BIOS Basic Input Output System
CPU Central Processing Unit
FAT File Allocation Table
I2C Inter-Integrated Circuit
MBR Master Boot Record
MII Media Independent Interface
NVRAM Non Volatile RAM
OS Operating System
PC Personal Computer
RAM Random Access Memory
TFTP Trivial File Transfer Protocol
USB Universal Serial Bus

6
3. Introduction

3.1 What is a boot loader?

Microprocessors can execute only code that exists in memory (either ROM or RAM), while
operating systems normally reside in large-capacity devices such as hard disks, CD-ROMs, USB
disks, network servers, and other permanent storage media.
When the processor is powered on, the memory doesn't hold an operating system, so special
software is needed to bring the OS into memory from the media on which it resides. This software
is normally a small piece of code called the boot loader. On a desktop PC, the boot loader resides
on the master boot record (MBR) of the hard drive and is executed after the PC's basic input output
system (BIOS) performs system initialization tasks.
In an embedded system, the boot loaders role is more complicated because these systems rarely
have a BIOS to perform initial system configuration. Although the low-level initialization of the
microprocessor, memory controllers, and other board-specific hardware varies from board to board
and CPU to CPU, it must be performed before an OS can execute.
At a minimum, a boot loader for an embedded system performs these functions:
Initializing the hardware, especially the memory controller
Providing boot parameters for the OS
Starting the OS
Most boot loaders provide features that simplify developing and updating firmware; for example:
Reading and writing arbitrary memory locations
Uploading new binary images to the board's RAM from mass storage devices
Copying binary images from RAM into flash

3.2 What is U-Boot?

U-Boot is an open-source, cross-platform boot loader that provides out-of-box support for hundreds
of embedded boards and many CPUs, including PowerPC, ARM, XScale, MIPS, Coldfire, NIOS,
Microblaze, and x86.
For more information about the U-Boot project see http://sourceforge.net/projects/u-boot/ and
http://www.denx.de/wiki/DULG/Manual.

3.3 Features of U-Boot

3.3.1 Customizable footprint

U-Boot is highly customizable to provide both a rich feature set and a small binary footprint.

3.3.2 Monitor
U-Boot has a command shell (also called a monitor) in which you work with U-Boot commands to
create a customized boot process.

7
3.3.3 Variables
U-Boot uses environment variables that can be read or written to and from non-volatile media. Use
these variables to create scripts of commands (executed one after the other) and to configure the
boot process.

3.3.4 Ethernet and USB

Because U-Boot can download a kernel image using either Ethernet or USB, no flash programming
is needed to test a new kernel. This prevents the deterioration of flash caused by repeated flash
erases and writes.

3.3.5 Numbers
Numbers used by U-Boot are always considered to be in hexadecimal format. For example, U-Boot
understands number 30100000 as 0x30100000.

3.4 The boot process

After power-up or reset, the processor loads the U-Boot boot loader in several steps.
The processor does these steps:
Executes a primary bootstrap that configures the interrupt and exception vectors, clocks,
and SDRAM
Decompresses the U-Boot code from flash to RAM
Passes execution control to the U-Boot
U-Boot does these steps:
Configures the Ethernet MAC address, flash, and, serial console
Loads the settings stored as environment variables in non-volatile memory
After a few seconds (a length of time you can program), automatically boots the pre-
installed kernel
To stop the automatic booting (autoboot) of the pre-installed kernel, send a character to the serial port
by pressing a key from the serial console connected to the target. If U-Boot is stopped, it displays a
command line console (also called monitor).
UU-Boot 1.1.6 (Aug 28 2009 - 14:03:27 - GCC 4.3.2) DUB-RevF3
for ConnectCore 9M 2443 on Development Board

DRAM: 64 MB
NAND: 128 MB
CPU: S3C2443@534MHz
Fclk = 534MHz, Hclk = 133MHz, Pclk = 66MHz
Autoscript from TFTP... [not available]
Hit any key to stop autoboot: 0

8
4. U-Boot commands

4.1 Overview
U-Boot has a set of built-in commands for booting the system, managing memory, and updating an
embedded systems firmware. By modifying U-Boot source code, you can create your own built-in
commands.

4.2 Built-in commands

For a complete list and brief descriptions of the built-in commands, at the U-Boot monitor prompt,
enter either of these commands:
help
?
You see a list like this one
# help
? - alias for 'help'
autoscr - run script from memory
base - print or set address offset
bdinfo - print Board Info structure
boot - boot default, i.e., run 'bootcmd'
bootd - boot default, i.e., run 'bootcmd'
bootelf - Boot from an ELF image in memory
bootm - boot application image from memory
bootp - boot image via network using BootP/TFTP protocol
bootvx - Boot vxWorks from an ELF image
clock - Set Processor Clock
cmp - memory compare
coninfo - print console devices and information
cp - memory copy
crc32 - checksum calculation
date - get/set/reset date & time
dboot - Digi ConnectCore modules boot commands
dcache - enable or disable data cache
dhcp - invoke DHCP client to obtain IP/boot params
echo - echo args to console
envreset- Sets environment variables to default setting
fatinfo - print information about filesystem
fatload - load binary file from a dos filesystem
fatls - list files in a directory (default /)
flpart - displays or modifies the partition table.
fsinfo - print information about filesystems
fsload - load binary file from a filesystem image
go - start application at address 'addr'
help - print online help
icache - enable or disable instruction cache
icrc32 - checksum calculation
iloop - infinite loop on address range
imd - i2c memory display
iminfo - print header information for application image
imm - i2c memory modify (auto-incrementing)
imw - memory write (fill)
inm - memory modify (constant address)
intnvram- displays or modifies NVRAM contents like IP or partition table
iprobe - probe to discover valid I2C chip addresses
itest - return true/false on integer compare
loadb - load binary file over serial line (kermit mode)
loads - load S-Record file over serial line
loady - load binary file over serial line (ymodem mode)
loop - infinite loop on address range

9
 ls - list files in a directory (default /)
md - memory display
mm - memory modify (auto-incrementing)
mtest - simple RAM test
mw - memory write (fill)
nand - NAND sub-system
nboot - boot from NAND device
nfs - boot image via network using NFS protocol
nm - memory modify (constant address)
ping - send ICMP ECHO_REQUEST to network host
printenv- print environment variables
printenv_dynamic- Prints all dynamic variables
rarpboot- boot image via network using RARP/TFTP protocol
reset - Perform RESET of the CPU
run - run commands in an environment variable
saveenv - save environment variables to persistent storage
setenv - set environment variables
sleep - delay execution for some time
sntp - synchronize RTC via network
tftpboot- boot image via network using TFTP protocol
update - Digi ConnectCore modules update commands
usb - USB sub-system
usbboot - boot from USB device
version - print monitor version
The available commands can vary according to the capabilities of your hardware platform.
For more information about a command, enter:
help command name
# help run
run var [...]
- run the commands in the environment variable(s) 'var'

As you enter the first letters of a command, U-Boot searches its list of built-in
commands until it finds a match. For example, if you enter save or sav or even
sa, U-Boot executes the saveenv command.
You need to enter enough letters for U-Boot to determine the command to
execute. For example, if you enter loa U-Boot cannot tell whether to execute
loadb, loads or loady, and you get an 'Unknown command' message.

4.2.1 Information commands

To get information, use these commands:
Command Description
bdinfo Prints board info structure.
coninfo Prints console devices and information.
date [MMDDhhmm[[CC]YY][.ss]] Gets / sets / resets system date/time.
fatinfo <interface> <dev[:part]> Prints information about the file system from 'dev' on 'interface.'
flinfo [bank] Prints information about the flash memory banks.
fsinfo Prints information about file systems.
iminfo [addr ...] Prints header information for the application image starting at the
'addr' address in memory, including verification of the image
contents (magic number, header, and payload checksums).
This command works only for Linux kernel images.

10
 nand bad Shows NAND bad blocks.
nand info Shows available NAND devices.
mii info <addr> Prints MII PHY info
version Displays U-Boot version and timestamp.

4.2.2 MII commands

To access the Ethernet PHY use this commands:
Command Description
mii device Lists available devices.
mii device <device name> Set current device.
mii read <addr > <reg> Reads register 'reg' from MII PHY 'addr'.
mii write <addr > <reg> <data> Writes 'data' to register 'reg' at MII PHY 'addr'.
mii dump <addr > <reg> Displays data of register 'reg' from MII PHY 'addr'.

The parameter addr and reg can be range e.g. 2-7.

The command mii dump is only usable for register 0-5.

4.2.3 Network commands

This table shows the network-related commands:
Command Description
bootp [loadAddress] [bootFilename] Boots the image over the network using the BootP/TFTP
protocol. If no argument is given, bootp takes the values
from the 'loadaddr' and 'bootfile' environment variables.
dhcp Requests an IP address from a DHCP server.
If the autoload variable is set to yes, also transfers the file
to which the bootfile environment variable points to the
loadaddr RAM memory address by TFTP.
ping <pingAddress> Pings the IP address passed as parameter. If the other end
responds, you see this message:
"host <pingAddress> is alive".
tftpboot [loadAddress] [bootfilename] Using FTP, transfers image 'bootfilename' into the RAM
address 'loadAddress'.
nfs [loadAddress] [host ip addr:bootfilename] Using NFS, transfers image 'bootfilename' into the RAM
address 'loadAddress'.
rarpboot [loadAddress] [bootfilename] Using RARP/TFTP, transfers image into the RAM address
'loadAddress'.
sntp Gets the date and time from the NTP server to which the
'ntpserverip' environment variable' points..

11
 Netconsole is not supported, because of poor performance.

If the autostart variable is set to 'yes', all these commands

(except ping and sntp) boot the transferred image by calling
the bootm command.
bootm does not work for WinCE images. If you are working
with a WinCE image file, either set the autostart variable to
'no' or delete it before executing these network commands.

4.2.4 USB commands

To access the USB subsystem, use the usb command, followed by its operations:
Command Description
usb reset Resets (rescans) USB controller
usb stop [f] Stops USB [f]=force stop
usb tree Shows USB device tree
usb info [dev] Shows available USB devices
usb storage Shows details of USB storage devices
usb dev [dev] Shows or set current USB storage device
usb part [dev] Prints the partition table of one or all USB storage devices
usb read addr blk# cnt Reads 'cnt' blocks starting at block 'blk#' to RAM address
'addr'
fatload usb <dev[:part]> <addr> <filename> Reads 'filename' image from FAT partition 'part' of USB
device 'dev' into the RAM memory address 'addr'.
If part is not specified, partition 1 is assumed.
ext2load usb <dev[:part]> <addr> Reads 'filename' image from EXT2/3 partition 'part' of USB
<filename> device 'dev' into the RAM memory address 'addr'.
If part is not specified, partition 1 is assumed.

4.2.5 Memory commands

These commands manage RAM memory:
Command Description
cmp[.b, .w, .l] addr1 addr2 count Compares memory contents from address 'addr1' to 'addr2'
for as many 'count' bytes, words, or long words.
cp[.b, .w, .l] source target count Copies memory contents from address 'source' to 'target'
for as many 'count' bytes, words, or long words.
dcache [on|off] Turns data cache on or off.
eeprom read <addr> <off> <cnt> Copies 'cnt' bytes memory contents from eeprom offset 'off'
to RAM address 'addr'.
eeprom write <addr> <off> <cnt> Copies 'cnt' bytes memory contents from RAM address
'addr' to eeprom offset 'off'.
erase_pt <name> Erases the partition 'name'. With flpart the 'name' can be
found.
go addr [arg ...] Starts the application at address 'addr' passing 'arg' as
arguments.

12
 imls Prints information about all images found at sector
boundaries in flash.
icache [on|off] Turns instruction cache on or off.
md[.b, .w, .l] <address> [# of Displays the contents of the memory at address 'addr' for as
objects] many '[#of objects]' bytes, words, or long words.
mm[.b, .w, .l] <address> Lets you modify locations of memory, beginning at
address, which gets auto-incremented.
mw[.b, .w, .l] <address> <value > Writes 'value' into 'address' for as many 'count' bytes,
[count] words, or long words.
nm[.b, .w, .l] address Lets you modify a fixed location of memory.
nand[.jffs2] read <addr> <off> Copies the memory contents from flash address 'off' to
<size> RAM address 'addr' for as many 'size' bytes (only for NAND
flash memories). Bad block management is used, when
using .jffs2. The bad block management detects bad blocks
and skips them.
nand[.jffs2] write <addr> <off> Copies the memory contents from RAM address 'addr' to
<size> flash address 'off' for as many 'size' bytes (NAND flash
memories only). Bad block management is used, when
using .jffs2. The bad block management detects bad blocks
and skips them.
nand erase [off size] Erases 'size' bytes from address 'off'. Erases the entire
device if no parameters are specified (NAND flash
memories only).
U-Boot skips bad blocks and shows their addresses.
nand dump[.oob] off Dumps NAND page at address 'off' with optional out-of-
band data (only for NAND flash memories).
nand markbad <off> Marks block at 'off' as bad.
nand unmarkbad <off> Erases bad block at 'off'.
nboot address dev [off] Boots image from NAND device dev at offset off
(transferring it first to RAM address).
protect [on|off] ... Protects/unprotects NOR sector(s).

The eeprom command uses default eeprom configured by

CFG_I2C_EEPROM_ADDR.

When writing your own boot macro, make sure data cache and
instruction cache are turned off right before booting the OS.

13
4.2.6 Serial port commands
Use these commands to work with the serial line:
Command Description
loadb [off] [baud] Loads binary file over serial line with offset 'off' and baud
rate 'baud' (Kermit mode)
loads [off] Loads S-Record file over the serial line with offset 'off'
loady [off] [baud] Loads binary file over the serial line with offset 'off' and
baud rate 'baud' (Ymodem mode)

4.2.7 I2C commands

These commands interface with the I2C interface:
Command Description
iloop chip address[.0, .1, .2] [# of objects] Loops, reading a set of I2C addresses
imd chip address[.0, .1, .2] [# of objects] Displays I2C memory
imm chip address[.0, .1, .2] Lets you modify I2C memory, with auto-incremented
address
imw address[.0, .1, .2] value [count] Fills with 'value' an I2C memory range
inm chip address[.0, .1, .2] Memory modify, read, and keep address
iprobe Discovers valid I2C chip addresses
itest [.b, .w, .l, .s] [*]value1 <op> [*]value2 Returns TRUE/FALSE on integer compare

4.2.8 Environment variables commands

To read, write, and save environment variables, use these commands:
Command Description
printenv [name ...] If no variable is given as argument, prints all U-Boot
environment variables.
If a list of variable names is passed, prints only those variables.
printenv_dynamic Prints all dynamic variables
envreset Overwrites all current variables values to factory default values.
Does not reset the 'wlanaddr' or 'ethaddr' variables or any other
persistent settings stored in NVRAM (see topic 8.2).
saveenv Writes the current variable values to non-volatile memory
(NVRAM).
setenv name [value] If no value is given, the variable is deleted. If the variable is
dynamic, it is reset to the default value.
If a value is given, sets variable 'name to value 'value'.

14
5. Environment variables

5.1 Overview
U-Boot uses environment variables to tailor its operation. The environment variables configure
settings such as the baud rate of the serial connection, the seconds to wait before auto boot, the
default boot command, and so on.
These variables must be stored in either non-volatile memory (NVRAM) such as an EEPROM or a
protected flash partition.
The factory default variables and their values also are stored in the U-Boot binary image itself. In
this way, you can recover the variables and their values at any time with the envreset command.
Environment variables are stored as strings (case sensitive). Custom variables can be created as
long as there is enough space in the NVRAM.

5.2 Simple and recursive variables

Simple variables have a name and a value (given as a string):
# setenv myNumber 123456
# printenv myNumber
myNumber=123456

To expand simple variables, enclose them in braces and prefix a dollar sign:
# setenv myNumber 123456
# setenv var This is my number: ${myNumber}
# printenv var
var=This is my number: 123456
Recursive variables (or scripts) contain one or more variables within their own value. The inner
variables are not expanded in the new variable. Instead, they are expanded when the recursive
variable is run as a command, as shown here:
# setenv dumpaddr md.b \${addr} \${bytes}
# printenv dumpaddr
dumpaddr=md.b ${addr} ${bytes}
# setenv addr 2C000
# setenv bytes 5
# run dumpaddr
0002c000: 00 00 00 00 00 .....
You must use the back slash '\' before '$' to prevent variables from being expanded into other
variables values.

5.3 Scripts
In U-Boot, a script is made up of variables that contain a set of commands; the commands are
executed one after another.
Consider this variable:
# printenv cmd1
setenv var val;printenv var;saveenv

If you were to run this script, with run cmd1 the var variable would be created with val value, the
value would be printed to the console, and the variables would be saved to either the EEPROM or
flash partition dedicated to variables.
# run cmd1
var=val
Saving Environment to Flash...

15
 Un-Protected 1 sectors
Erasing Flash...
. done
Erased 1 sectors
Writing to Flash... done
Protected 1 sectors
Separate the commands in a script with semi-colons. (;). As with recursive variables, this sign must
be preceded by a back-slash sign or it is considered the termination of the first command itself.
This is how you would save cmd1:

# setenv cmd1 setenv var val\;printenv var\;saveenv

For running commands stored in variables, use the run command and its variables separated by
spaces:
# setenv cmd1 setenv var val
# setenv cmd2 printenv var
# setenv cmd3 saveenv
# run cmd1 cmd2 cmd3

See how to create a bootscript that automatically executes at start at topic 6.

5.4 System variables

U-Boot uses several built-in variables:

5.4.1 Common system variables

Variable Description
autoload If set to "no" (or any string beginning with 'n'), the rarpboot, bootp, or
dhcp command performs a configuration lookup from the BOOTP /
DHCP server but does not try to load any image using TFTP.
autostart If set to "yes", an image loaded using the rarpboot, bootp, dhcp or
tftpboot commands is automatically started (by internally calling the
bootm command).
baudrate The baud rate of the serial connection.
bootcmd Defines a command string that is automatically executed when the
initial countdown is not interrupted.
Executed only when the bootdelay variable is also defined.
bootdelay Seconds to wait before running the automatic boot process in
bootcmd.
bootfile Name of the default image to load with TFTP.
filesize Contains the size of the last file transferred by TFTP or USB.
fileaddr The RAM address where the last file transferred by TFTP was placed.
ntpserverip NTP server IP address (for getting the date/time).
stdin Standard input system.
stdout Standard output system.
stderr Standard error output system.

16
 verify If set to 'n' or 'no,' disables the checksum calculation over the complete
image in the bootm command to trade speed for safety in the boot
process. Note that the header checksum is still verified.
ipaddr IP address of the target's Ethernet interface.
ipaddr_wlan IP address of the target's WLAN interface (for modules that have it).
netmask Subnet mask of Ethernet interface.
netmask_wlan Subnet mask of WLAN interface (for modules that have it).
gatewayip IP address used as network gateway.
serverip IP address of the host PC (for remote connections like TFTP transfers).

5.4.2 Dynamic variables

Depending on the module, the partitioning information, and so on, U-Boot generates some
variables "on the fly" if they do not already exist in U-Boot.
These variables can be overwritten with setenv thus becoming standard U-Boot variables.
Dynamic variables which are not set with setenv also exist (they are automatically created), but
they cannot be printed with printenv.
Some of these variables are OS-specific for different OS implementations (Linux, Windows CE,
NET+OS). They provide special functionality for the OS running in the platform.

For more information, see the boot loader development chapter of your
development kit's documentation.

5.4.3 User keys

The development board in the kit may have two user buttons. If it does, U-Boot can detect which
one is pressed when it starts.
If you press either key when the boot loader is starting, the key1 or key2 variable is executed
before the bootcmd. This lets you have different boot scripts, depending on the key pressed during
boot, so you can boot two different kernels, such as a dual Linux/Windows CE or two versions of
the same OS.
If the key1 and key2 variables do not exist, the normal bootcmd is executed.
When the two keys are pressed during boot, both are detected as pressed, and both scripts are
launched. The script in variable key1 is always executed before the one in variable key2.

You can disable detection of user keys for customized hardware where these
keys don't exist. To do so, you need to reconfigure and recompile U-Boot. See
chapter 11 for information about U-Boot development.

For the Digi Connect ME 9210 only user key 2 is enabled by

default.

17
5.4.4 Protected variables
Several variables are of great relevance for the system and are stored in a protected section of
NVRAM.
Some of these protected variables are, for example, the serial number of the module and the MAC
addresses of the network interfaces, which are programmed during production and normally should
not be changed.

18
6. Bootscript
The bootscript is an script that is automatically executed when the boot loader starts, and before
the OS auto boot process.
The bootscript allows the user to execute a set of predefined U-Boot commands automatically
before proceeding with normal OS boot. This is especially useful for production environments and
targets which dont have an available serial port for showing the U-Boot monitor.

6.1 Bootscript process

The bootscript works in the following way:
1. U-Boot checks the variable loadbootsc. If set to no, it continues normal execution.
2. If the variable loadbootsc is set to yes (factory default value) U-Boot tries to download
the bootscript file with the filename stored in variable bootscript from the TFTP server IP
address defined at variable serverip (by default 192.168.42.1).
The default value of the bootscript variable is <platformname>-bootscript.
3. If the bootscript file is successfully downloaded, it is executed.
4. If any of the commands in the bootscript fails, the rest of script is cancelled.
5. When the bootscript has been fully executed (or cancelled) U-Boot continues normal
execution.

6.2 Creating a bootscript

To create a bootscript file:
1. Create a plain text file with the sequence of U-Boot commands. Usually, it is recommended
that the last command sets the variable loadbootsc to no, to avoid the bootscript from
executing a second time.
For example, create a file called myscript.txt with the following contents:
setenv company digi
setenv bootdelay 1
printenv company
setenv loadbootsc no
saveenv
This script creates a new variable called company with value digi and sets the bootdelay
to one second. Finally it sets the variable loadbootsc to no so that U-Boot doesnt try to
execute the bootscript in the future, and saves the changes.
2. Execute the mkimage tool (provided with U-Boot) with the file above as input file. Syntax is:
mkimage -T script -n "Bootscript" -C none -d <input_file> <output_file>

The name of the output file must be in the form <platformname>-bootscript, where
<platformname> must be replaced with your targets platform name.
For example, to create the bootscript from the text file above and for a Connect ME 9210
platform, go to the U-Boot directory and execute:
$ tools/mkimage -T script -n Bootscript -C none -d myscript.txt
cme9210js-bootscript

19
6.2.1 Creating a bootscript in Windows
There are some important caveats when producing a bootscript in a Windows host PC.

6.2.1.1 Final carriage return

When creating the plain text file with the sequence of commands, make sure that the last command
in the sequence contains a final carriage return (in other words, a blank line at the end of the file).
Otherwise, the last command in the sequence may not execute.

6.2.1.2 Windows line-end characters

If developing U-Boot in a Windows host (using Cygwin), the plaintext file with the list of commands
might contain incorrect line-end characters. If this is the case, the plaintext file must be converted
first to UNIX line-end format before generating the bootscript. This can be done using the dos2unix
application over the file, before calling mkimage:

$ dos2unix.exe myscript.txt

The prebuilt mkimage tool is not included with the Cygwin

development environment. This tool is built the first time you
compile U-Boot under Cygwin.

6.3 Configuration for launching the bootscript

Once the bootscript file has been created two more steps are needed to let the target run the
bootscript at start.
1. Place the bootscript file into the TFTP exposed folder, so that the target is able to find it
when it boots.
2. The U-Boot variable serverip of the target must point to the host PC with the TFTP server.
You have two options:
a. Connect to the targets U-Boot monitor and set the serverip variable to the IP of
your host PC.
b. If you dont have access to the U-Boot monitor or simply dont want to have any
user interaction with the target (for example in a production environment), configure
the host PC Ethernet cards IP to the factory default IP address stored in variable
serverip, which is 192.168.42.1.
Once done with all the steps, power up the target and it will connect to the host PC, will download
the bootscript to RAM, execute it, and continue booting as usual.

6.4 Bootscript restrictions

The Digi U-Boot command flpart (for partitioning the Flash) is a menu-driven program, which
expects key presses for different user selections. This command may not work in a bootscript. For
repartitioning the Flash, use the command intnvram instead (refer to chapter 8.2 for more
information).

20
7. Boot commands

7.1 Overview
U-Boot runs code placed in RAM, although it can also read data from other media. The boot
process normally takes place in two steps:
Reading the OS image from media (Ethernet, flash, USB, MMC) into RAM
Jumping to the first instruction of the image in RAM

7.2 Reading images into RAM

7.2.1 From Ethernet

The most common way to boot an image during development is by transferring it using TFTP over
the Ethernet interface. You do this with the tftpboot command, passing:
The address of RAM in which to place the image
The image file name

# tftpboot <loadAddress> <bootfilename>

The TFTP transfer takes place between the serverip address (host) and the ipaddr address
(target). The host must be running a TFTP server and have bootfilename archive placed in the
TFTP-exposed directory.
For Linux kernel images, if the autostart variable is set to yes, this command directly boots the
kernel after downloading it.

7.2.2 From USB

Another way to boot an image is by reading it from a USB flash storage device. The USB disk must
be formatted either in FAT, ext2, or ext3 file system.
To read an image from a USB flash disk formatted in FAT, enter:
# usb reset
# fatload usb <dev>[:partition] <loadAddress> <bootfilename>

If the flash disk is formatted in ext2/ext3:

# usb reset
# ext2load usb <dev>[:partition] <loadAddress> <bootfilename>

This command reads file bootfilename from device dev, partition partition of the USB flash disk into
the RAM address loadAddress. Device and partition are given as a number (0, 1, 2...).
If no partition is specified, partition 1 is assumed.

7.2.3 From MMC

If the target has an MMC or HSMMC (High Speed MMC) interface U-Boot can also read from it.
The MMC card must be formatted either in FAT, ext2, or ext3 file system.

21
 To read an image from an MMC card formatted in FAT, enter:
# mmcinit
# fatload mmc <dev>[:partition] <loadAddress> <bootfilename>

If the card is formatted in ext2/ext3:

# mmcinit
# ext2load mmc <dev>[:partition] <loadAddress> <bootfilename>

This command reads file bootfilename from device dev, partition partition of the MMC card into the
RAM address loadAddress. Device and partition are given as a number (0, 1, 2...).
If no partition is specified, partition 1 is assumed.

7.2.4 From flash

For standalone booting, the device can read the image from flash, avoiding dependency on any
external hardware.
In targets with NOR flash memories, do this with memory commands:

# cp.[b/w/l] <sourceAddress> <loadAddress> <count>

This command copies count bytes, words, or long words (depending on the suffix used -: b, w, l -
from sourceAddress into loadAddress.
In targets with NAND flash memories, the special NAND commands must be used:

# nand read <loadAddress> <sourceAddress> <count>

This command copies count bytes from sourceAddress into loadAddress.

7.3 Booting images in RAM

After the image is transferred to RAM, you can boot it in either of two ways, depending on the OS:
For Windows CE images:

# go <loadAddress>

For Linux images:

# bootm <loadAddress>

where loadAddress (in both cases) is the address in RAM at which the image resides.
Windows CE images must be compiled with the information
about the address in RAM from which they will be booted. For
example, if a WinCE kernel is compiled with a boot address of
0x2C0000, it can be transferred to a different address, but the
system can boot only from the compiled-in address.

22
7.4 Direct booting
To simplify the boot process, Digi's U-Boot version includes the dboot built-in command, which
reads the OS image from the media and runs it from RAM in a single step.
The syntax for the dboot command is:

# dboot <os> <media> <dev>[:partition] <filesystem> <bootfilename>

where
os is either linux, wce or netos.
media is either flash, tftp, nfs, usb, mmc, or hsmmc.
dev[:partition] is the device index (only for USB, MMC, HSMMC media) starting at 0 and the
partition number (starting at 1) where the image to boot resides. If not provided, device 0 and
partition 1 are assumed
filesystem is either fat, ext2, or ext3 (only for USB, MMC, HSMMC media) and must match the
file system of the partition that holds the image to boot. If not provided, FAT is assumed.
bootfilename is the name of the kernel image file to download and boot. If not provided, the
filename is taken from kimg variable for Linux, or wimg for Windows CE

If booting from a network media (tftp, nfs) and the dhcp variable is set to yes or
on, the command first gets an IP address from a DHCP.

7.4.1 Boot examples

Boot default Windows CE image stored in a USB disk with partition 1 formatted in FAT, plugged in
the USB host device 0:

# dboot wce usb

Boot default Linux image stored in an SD card with partition 3 formatted in ext3, plugged in the
USB host device 0:

# dboot linux mmc 0:3 ext3

Refer to your OS user manual for further instructions on booting your kernel
image.

7.5 Automatic booting

If U-Boot is not interrupted after the delay established in bootdelay, the automatic boot process
takes place. Automatic booting consists of running what is specified in the bootcmd environment
variable.

23
In other words, automatic booting has the same effect as doing either of the next two examples:

# run bootcmd

# boot

If, for example, if you want to automatically boot a WinCE image from TFTP server, set bootcmd
like this:

# setenv bootcmd dboot wce tftp

Or, if you want to automatically boot a Linux image from flash, set bootcmd like this:

# setenv bootcmd dboot linux flash

If bootdelay is set to 0, the autoboot happens immediately

after U-Boot starts. To stop the process and enter the monitor,
press a key as soon as the first U-Boot output lines appear.

24
8. Using NVRAM
An embedded OS requires some persistent settings; for example, MAC address, IP address, Internet
gateway, flash partition table, and U-Boot environment variables. You change some of these only in
production and others only during custom setup.
These settings must be stored in non-volatile memory (NVRAM) so they are not lost when you power the
target off.
A partition called NVRAM on the flash memory is used to store these settings. The contents are protected
by a CRC32 checksum and they are also mirrored to a different location in the partition. This way, if
anything goes wrong reading these data or data becomes corrupted, the information can be restored from
the mirrored data.

8.1 The 'flpart' command

To print, modify, or restore the partitions table, use the flpart command. This U-Boot command
requires no arguments; you create the partitions table using a menu of options.

8.1.1 A partition table entry

A partition table entry contains these fields:
Field Description
Number Index of partition in the table
Name Name of the partition
Chip Index of flash chip (normally, only one)
Start Physical start address of the partition (in hex)
Size Size of the partition (in hex)
Type Partition type (what it will contain)
U-Boot
NVRAM
FPGA
Linux-Kernel
WinCE-EBoot
WinCE-Kernel
Net+OS-Kernel
Net+OS-Loader
Net+OS-NVRAM
File system
WinCE-Registry
Splash-Screen
Unknown
FS File system that the partition contains:
YAFFS
JFFS2
CRAMFS
SQUASHFS
INITRD

25
 FlashFX
Unknown
Flags Flags (non-exclusive):
read-only
mount read-only
rootfs

8.1.2 Changing the partition table

To modify the partition table, use the flpart command in U-Boot:
# flpart
Commands:
a) Append partition
d) Delete partition
m) Modify partition
p) Print partition table
r) Reset partition table
q) Quit
Cmd (? for help)> p
Nr | Name | Start | Size | Type | FS | Flags
----------------------------------------------------------------------
0 | U-Boot | 0 | 768 KiB | U-Boot | None | fixed
1 | NVRAM | 768 KiB | 256 KiB | NVRAM | None | fixed
2 | FPGA | 1 MiB | 1 MiB | FPGA | None | fixed
3 | EBoot | 2 MiB | 1 MiB | WinCE-EBoot | None |
4 | Registry | 3 MiB | 1 MiB | WinCE-Registry | None |
5 | Kernel | 4 MiB | 20 MiB | WinCE-Kernel | None |
6 | FFX | 24 MiB | 2 MiB | Filesystem | FlashFX |

You add, modify, or delete partitions step-by-step; the command prompts you for the necessary
information.

Start and Size values can be given as hexadecimal numbers (prefixed with 0x)
or as decimal numbers followed with k (for KiB) or m (for MiB).

The partition table also can be reset to the default values. In this case, because the partition table
differs according to the targets OS, you select the OS you want.
Changes take effect only after quitting 'flpart' and saving the
changes.
When the size or start address of a partition has been changed,
it is always necessary to erase it and write a new image to it.

8.2 The intnvram command

Most of the variables stored in NVRAM can be read with the printenv command, modified or
erased with the setenv command, stored with the saveenv command, and reset with the envreset
command. There are, however, protected variables in the NVRAM which are read-only. These are,
for example, the MAC address of the module, the serial number, the boot and NVRAM partitions,
the wireless calibration data, etc.
Protected variables stored in NVRAM can be read, modified, erased, stored, or reset with the
intnvram command.

26
 DO NOT USE THE intnvram COMMAND UNLESS YOU ARE
COMPLETELY SURE OF WHAT YOU ARE DOING. INCORRECT USE
CAN DESTROY SENSITIVE NVRAM DATA FOREVER AND MAKE
THE MODULE UNUSABLE

Changes made to NVRAM with the intnvram command are kept in RAM. U-Boot writes the
changes to NVRAM only when you execute the saveenv command or intnvram save command.
Here is the syntax of the intnvram command:
Usage: intnvram help|print <params>|printall|repair|reset|save|set <params>

help : prints this

print : prints selected parameters.
E.g.: print module mac serialnr
printall : prints complete contents and metainfo
repair : Repairs the contents. If one image is
bad, the good one is copied onto it.
If both are good or bad, nothing happens.
reset : resets everything to factory default values.
save : saves the parameters
set : sets parameters.

For help with this command, enter intnvram help.

To print the complete contents of the NVRAM settings, enter intnvram printall.
You can set or print either one parameter or a set of parameters. Parameters are grouped in
blocks. This is the complete parameters list with the possible values some of them can take:
params for "set" or "print" can be
module [producttype=] [serialnr=] [revision=] [patchlevel=]
[ethaddr1=] [ethaddr2=]
network [gateway=] [dns1=] [dns2=] [server=] [netconsole=] [ip1=]
[netmask1=] [dhcp1=] [ip2=] [netmask2=] [dhcp2=]
partition [add] [del] [select=] [name=] [chip=] [start=] [size=]
[type=] [flag_fixed=] [flag_readonly=]
[flag_fs_mount_readonly=] [flag_fs_root=] [flag_fs_type=]
[flag_fs_version=]
os [add] [del] [select=] [type=] [start=] [size=]

Params trailed with '=' require a value in the set command. In the print
command, '=' mustn't be used.

Possible Values are

os type: None,Critical,OS-Meta,U-Boot,Linux,EBoot,WinCE,Net+OS,
Unknown,Application
partition type: U-Boot,NVRAM,FPGA,Linux-Kernel,WinCE-EBoot,WinCE-Kernel,
Net+OS-Kernel,Filesystem,WinCE-Registry,Unknown,
Splash-Screen
flag_fs_type: None,JFFS2,CRAMFS,INITRD,FlashFX,Unknown

Specify the group of the parameter before the parameter itself. For example, to print the module IP
for the wired Ethernet interface, execute:
# intnvram print network ip1
ip1=192.168.42.30

27
 For printing different parameters of a block, the block must be used only once. For example, to print
the module's MAC address and serial number, execute:
# intnvram print module ethaddr1 serialnr
ethaddr1=00:40:9D:2E:92:D4
serialnr=0700-94000329A

To set a parameter a valid value must be provided, as shown here:

# intnvram set module serialnr=REVA-6_001

To access a partition parameter, address the specific partition with the parameter select=n, where
n is the index to the partition. This example prints the names of partitions 1 and 2:
# intnvram print partition select=0 name select=1 name
name=U-Boot
name=NVRAM

The reset command will completely erase the MAC addresses

and the wireless calibration data of your module. Do not use
the reset command unless you have a backup file with the
wireless calibration data for future restoration.

8.2.1 Mappings of variables

Some of the protected variables in NVRAM are mapped to U-Boot environment variables.
Therefore, modifying them with intnvram command is the same as doing so with setenv
command. For security reasons, however, some variables cannot be modified with the
setenv command.
This table lists the mapped variables:
U-Boot variable NVRAM parameter Blocked for 'setenv'
ethaddr ethaddr1 X
wlanaddr ethaddr2 X
netmask netmask1
netmask_wlan netmask2
ipaddr ip1
ipaddr_wlan ip2
dnsip dns1
dnsip2 dns2
dhcp dhcp1
dhcp2 dhcp2
serverip server
gatewayip gateway

28
9. Firmware update commands

9.1 Overview
The boot loader, kernel, and other data stored in flash form the firmware of the device. Because
U-Boot can write any part of flash, its flash commands can be used to reprogram (update) any part
of the firmware. This includes the boot loader itself.
The update process normally takes place in three steps:
Reading image from media (Ethernet, USB, MMC) into RAM memory
Erasing the flash that is to be updated
Copying the image from RAM into flash

9.2 Updating flash with images in RAM

Flash memory must be updated with images located in RAM memory. You can move images to
RAM using either Ethernet, USB, or MMC (see section 7.2 for more information).
To erase flash and copy the images from RAM to flash, use these commands:
For NOR flash memory:
# erase address +size
# cp.[b|w|l] sourceAddress targetAddress count

The first command erases size bytes beginning at address. The second command copies count
bytes, words or long words (depending on the suffix used: b, w, l) from sourceAddress into
targetAddress.
For NAND flash memory:
# nand erase address size
# nand write sourceAddress targetAddress count

The first command erases size bytes beginning at address. The second command copies count
bytes from sourceAddress into targetAddress.

The erasure of the flash comprises whole erase-blocks.

The address and size parameters must be multiples of the
erase-blocks of the flash memory. See your module's flash
datasheet for the erase-block size.

29
9.3 Direct updating
Digi's U-Boot version includes the built-in update command. This command copies the image from
the media to RAM, erases the flash size needed for the image, and moves the image from RAM
into flash in a single step, simplifying the update process.
Here is the syntax for update:
# help update
update partition [source [device:part filesystem] [file]]
- updates 'partition' via 'source'
values for 'partition': uboot, linux, rootfs, userfs, eboot, wce, wcez,
netos, netos_loader, splash, or any partition name
values for 'source': tftp, nfs, usb, mmc, hsmmc
'device:part': number of device and partition, for 'usb', 'mmc', 'hsmmc'
sources
values for 'filesystem': fat|vfat, ext2|ext3
values for 'file' : the file to be used for updating
source is the place to take the image from. If not provided, tftp is assumed.
dev[:partition] is the device index (only for USB, MMC, HSMMC media) starting at 0, and the
partition number (starting at 1) where the image to update resides. If not provided, device 0 and
partition 1 are assumed
filesystem is either fat, ext2, or ext3 (only for USB, MMC, HSMMC media). If not provided, FAT
is assumed.
file is the name of the image to download and update. If not provided, the filename is taken
from one of the following U-Boot environment variables (depending on the partition to be
updated):
o kimg: for the Linux kernel image
o wimg: for the Windows CE kernel image
o nimg: for the NET+OS kernel image
o uimg: for the boot loader image
o usrimg: for the user image
o rimg: for the Linux root file system image
o simg: for the Splash screen image
o fimg: for the FPGA image
o nloader: for the Windows CE kernel image

If updating from a network media (tftp, nfs) and the dhcp variable is set to yes
or on, the command first gets an IP address from a DHCP server.

For example, to update the Splash screen partition using a file called mylogo.bmp that resides on
the second partition (formatted in ext3) of an SD card which is plugged in the first MMC device
(index 0), the update command would be:

# update splash mmc 0:2 ext3 mylogo.bmp

30
 To update the boot loader from the TFTP exposed folder with the default name image stored in
variable uimg, the update command would be:

# update uboot

9.3.1 Update limits

The update command in U-Boot transfers files to RAM, erases the flash partition, and writes the
files from RAM into flash memory.
The file that is transferred is copied to a specific physical address in RAM; therefore, the maximum
length of the file to update is:
Update file size limit = Total RAM memory RAM offset where the file was loaded
As a general rule, U-Boot does not let you update a flash partition with a file whose size exceeds
the available RAM memory. This means that, for example, if you have a module with 32MB RAM
and 64MB flash and you want to update a partition with a file that is 35MB, U-Boot will not do it.
Note that this limitation is due to the RAM memory size, as U-Boot first needs to transfer the file to
RAM before copying it to flash.

For updating partitions with files larger than the available RAM memory, see
your OS-specific update flash tool.

31
10. Customize U-Boot

10.1 Overview
U-Boot has a lot of functionalities, which can only be enabled before compiling U-Boot. To
configure U-Boot options and customize the boot loader, refer to your OS specific user manual.
Description of some available configurations follow:

10.2 JTAG Console

The JTAG console uses the JTAG interface instead of the serial line, which is used by the default
console. The Requirements for the JTAG console are a debugger and a host application to
communicate with the Direct Communication Channel. The BDI2000 supports input/output and the
Segger jlink supports output only.

This configuration does not work together with silent mode.

Before you can start using JTAG-Console you have to configure your debugger for using the Direct
Communication Channel. Using a BDI200 you have to add the following line in section [TARGET]
to your BDI configuration file:
DCC 7
After that type:
$ telnet <bdi2000 ip-address> 7
Connected to bdi2000.
Escape character is '^]'.

Using a Segger jlink the version 3.87i or higher of jlink-software is required. Start the
jlinkcommander and type:

$ term

Only stdout can be used with the Segger jlink.

After starting your target you can switch output and input independently to JTAG console by typing
# setenv stdout jtag
# setenv stdin jtag
# setenv stderr jtag
# saveenv

After reboot the JTAG console is used by default, if you have stored the
environment.

32
10.3 Silent Console
The target does not display any output when the console is set to silent mode. Before using a boot
loader with silent console, you should first define a way to recover from it. Otherwise you will not be
able to disable it in the future.
The first possibility is using the key environments variables (see topic 5.4.3):
# setenv key1 setenv silent no\;saveenv
# saveenv

The sequence to recover from silent mode is the following:

1. Keep Key1 pressed while the target is booting
2. After a short time (about 4 seconds) press the reset button
3. The target now boots with output on the console

The second possibility to recover from silent mode is using a gpio. To use this functionality you
have to define the gpio number and the level of the gpio, which signals the console to leave the
silent mode. The number of the gpios can be found in the hardware reference.

You will need the JTAG interface to flash the firmware, when
the function to recover from silent mode is undefined.

The console is switch to silent mode after reboot by typing:

# setenv silent yes
# saveenv

After reboot you can recover from silent mode, by setting the gpio to the defined level short after
the target starts to boot or when the target end to boot. The second point is only reached, when
autoboot is not used or the execution fails.

10.4 Splash screen support

The U-Boot splash screen support allows the user to display a picture at boot time on an LCD or
monitor that is connected to the target. The picture is read from a partition in the flash.
Some platforms have splash support already enabled in U-Boot and some have not. To determine
if your platform has splash support enabled, check if the variable video exists:

# printenv_dynamic

If it exists and has some value, there should also be a partition for the splash screen image. Check
it with command flpart and option p to print the partition table. If there is no splash screen partition,
refer to chapter 8.1 to create one. If there is, you can go to chapter 10.4.3 to see how to upload a
splash image to the flash partition.

33
10.4.1 Enable splash screen support in U-Boot
Go to your OS configuration tool for U-Boot and enable the Display support and splash screen
support. You will need to select the displays you want to support (more than one can be
supported):
Display resolution
Sharp LQ057Q3DC12I 320x240
Sharp LQ064V3DG01 640x480
VGA monitor 640x480
EDT 8-inch QVGA 320x240
Sharp LQ070Y3DG3B1 800x480
Note that, depending on your platform, some displays might not be available.
Save the changes and recompile U-Boot. Then update the U-Boot image in your target.

10.4.2 Creating a splash partition

A partition is needed to hold the splash screen image. If your partition table doesnt have one, you
need to create it with the flpart command:
# flpart
Nr | Name | Start | Size | Type | FS |
Flags
---------------------------------------------------------------------
0 | U-Boot | 0 |768 KiB |U-Boot | | fixed
1 | NVRAM | 768 KiB |512 KiB |NVRAM | | fixed
2 | Kernel | 1280 KiB | 3 MiB |Linux-Kernel | |
3 | RootFS-JFFS2 | 4352 KiB | 16 MiB |Filesystem | JFFS2 | rootfs
4 | User-JFFS2 |20736 KiB | 11 MiB | Filesystem | JFFS2 |
Commands:
a) Append partition
d) Delete partition
m) Modify partition
p) Print partition table
r) Reset partition table
q) Quit
Cmd (? for help)> a
Last partition 4 had already maximum size.
Size (in MiB, 0 for auto, 11 MiB max) (11 MiB): 10 MiB
Adding partition # 5
Name (): Splash
Chip (0): 0
Start (in MiB, 0 for auto) (0): 0
--> Set to 31 MiB
Size (in MiB, 0 for auto, 1 MiB max) (0): 1MiB
Partition Types
Partition Type (U-Boot, ? for help)> s
Fixed (n): n
Readonly (n): n
Partition 5 added
Cmd (? for help)> p
Nr | Name | Start | Size | Type | FS | Flags
---------------------------------------------------------------------
0 | U-Boot | 0 |768 KiB |U-Boot | | fixed
1 | NVRAM | 768 KiB |512 KiB |NVRAM | | fixed
2 | Kernel | 1280 KiB | 3 MiB |Linux-Kernel | |
3 | RootFS-JFFS2 | 4352 KiB | 16 MiB |Filesystem | JFFS2 | rootfs

34
 4 | User-JFFS2 |20736 KiB | 10 MiB |Filesystem | JFFS2 |
5 | Splash | 31 MiB | 1 MiB |Splash-Screen | |
Cmd (? for help)>q
Partition table has been modified. Save? (y): y
Writing Parameters to NVRAM

10.4.3 Uploading a splash image

After the partition is created the splash image has to be flashed. Only 8-bit BMP images can be
used. Images must match the display resolution.
Uploading is done by following command (in the example, the bitmap file digi.bmp is used):
# update Splash tftp digi.bmp
TFTP from server 192.168.42.1; our IP address is 192.168.42.30
Filename 'digi.bmp'.
Load address: 0x200000
Loading: ######
done
Bytes transferred = 77878 (13036 hex)
Calculated checksum = 0x2a13f7f
Erasing: complete
Writing: complete
Verifying: complete
Update successful

Or, if the variable simg contains the BMP filename, you can simply execute:

# update splash

A splash image takes about 75 KiB for QVGA resolution and 300 KiB for VGA
resolution.

If the splash image does not match the resolution of the

display or it is not 8-bit pallet bitmap, U-Boot will print an
error message and will not show the splash screen.

10.4.4 Initialize video interface

To initialize the video interface in U-Boot, the following variables must be set:
# setenv videoinit yes
# setenv video displayfb:DISPLAYNAME
# saveenv

where DISPLAYNAME is the name of the connected display:

LQ057Q3DC12I

LQ064V3DG01

VGA

EDT28

35
 LQ070Y3DG3B1

In order to change the selected display without recompiling U-Boot, each display
must have support enabled. See in topic 10.4.1.

The default location of the frame buffer is at the end of U-Boot in RAM. The location can be
changed by setting:

# setenv fb_base <address>

The default is also used if the passed address for the frame buffer is inside of a protected area (for
example if the address points to the U-Boot code).

Platforms cc9m2443 and ccw9m2443 always initialize the

video interface, regardless of the value of videoinit variable or
the existence of a splash partition.

36
11. U-Boot development
U-Boot is an open source project. Sources are freely distributed, and you can modify them to meet
your requirements for a boot loader.
The project sources are ready to be installed and compiled in a Linux environment. If you do not
have a Linux machine for development, you can install the Cygwin X-Tools software
(http://www.cygwin.com). The X-Tools provide a Unix-like development environment for Windows,
based on Cygwin and the GNU toolchain, to cross-compile the boot loader.
For information about installing the U-Boot sources, modifying platform-specific sources, and
recompiling the boot loader, see your development kit documentation. Procedures may vary
according to hardware platform and OS.

37
12. Troubleshooting

12.1 Work with FIMs on NS921x based modules

Modules using NS921x processors can use the FIMs in the processor as UART and/or SD
controller.
When using the FIMs, instruction cache must be enabled, otherwise they may malfunction.
Instruction cache is enabled by default in U-Boot, so do not disable it when using the FIMs.

12.2 Problems booting from NFS

When booting or updating directly from NFS, sometimes U-Boot can experience problems
mounting or unmounting the NFS server. To work around this problem, edit the file /etc/hosts of
the host PC serving the NFS path and add a line with the IP address of the target.

192.168.42.40 mytarget

12.3 USB memory sticks

U-Boot might experience problems detecting certain USB memory sticks due to different response
timings.

38