Getting Started with StarEast Board

 

Robert Heath and Adam Moerschell

Department of Electrical & Computer Engineering

University of California, Davis, CA 95616

 

Hardware Setup

 

You will need at least:

            - 2 computers (One must have a serial port, both must have Ethernet ports)

            - 2 Ethernet cables

            - 1 StarEast board with Intel IXP425 533Mhz Processor

            - 1 RS232 Serial-to-Ethernet cable (included with StarEast board)

            - 1 StarEast CD (for Linux build)

 

The StarEast board has three Ethernet interfaces on it.  Two have LEDs on their left hand side.  These two ports are for actual traffic and are called ixp0 and ixp1.  Ixp1 is in the middle and ixp0 is on the left hand side.  The remaining port is used with the RS232 cable to configure the board via a terminal connection.  This is called the UART port.

 

Your setup should connect one computer directly via Ethernet cable to each ixp0 and ixp1.  The computer with the serial port should connect to the UART port via the RS232 cable.


 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1 -   The Ports of Interest on the StarEast Board
 
 

 

 

 


Connecting to Terminal Interface

 

You can use any terminal connection program you want to use to connect to the UART port.  Windows users commonly use Hyperterminal, and in linux you can use minicom.  The following connection settings are required:

                 

                  115200 baud

                  8 bits

                  no parity

                  1 stop bit

                  no hardware control

                  no software control

 

 

If using minicom make sure to execute the following command to give access to the serial port:

chmod 666 /dev/ttyS0        (requires permissions)

 

Once you have set up your connection, connect then plug the StarEast board in.  You should see some boot up messages then the Redboot prompt:

 

                  Redboot>

 

Building snapgear Images

 

The ÒsnapgearÓ directory on the Stareast cd contains the compressed files that you will need for your snapgear installation.  This process was done on Fedora Core 3, and should work for all versions of fedora core and RedHat up to FC3.

 

1.)   Copy the snapgear directory from the CD to the computer you will be building the images on.  If the copied files are read only, modify the permissions to allow modification of the files and directories.

 

2.)   In the snapgear directory, decompress the snapgear source code:

 

tar vxzf snapgear-3.1.1.tar.gz

 

This will create another snapgear directory that contains the source code.

 

3.)   Change to the snapgear directory that was just created:

 

cd snapgear

 

4.)   The IXP400 Access Library integration support files need to be installed.

 

tar vxzf ../snapgear-3.1.x-ixp400-v1.4.2-support.tar.gz

 

cp ../ixp400AccessLibrary-1_4.zip

     vendors/Intel/IXDP425/ixp400-1.4-uclinux/insall

 

Note: Destination starts on next line in the command above

 

vendors/Intel/IXDP425/ixp400-1.4-uclinux/install

 

5.)   The tools for compiling the images need to be installed.  These tools are in the arm-linux-tools-20030927.tar.gz file.  Decompress this file and it will create a new directory called ÒusrÓ.  Move this directory to your home directory and add it to your path:

 

export PATH=/home/<usrname>/usr/local/bin:${PATH}

 

6.)   From the decompressed snapgear directory, the kernel configuration needs to be run:

 

 

make makeconfig

 

Make the following changes in the menu that is displayed:

 

Vendor/Product Selection --->

---Select the Vendor you wish to target

Intel

Kernel/Library/Defaults Selections --->

---Select the Product you wish to target

IXDP425

Kernel/Library/Defaults Selection --->

(linux-2.4.x) Kernel Version

(glibc) Libc Version

[*]Customize Kernel Settings

[*]Customize Vendor/Usr Settings

 

After you exit and save, a new window will appear, make the following changes in that window.

 

General setup --->

[*]Support for hot-pluggable devices

General setup --->

Default kernel command string:

"console=ttyS1,115200 root=/dev/ram0 initrd=0x00800000,16M

mem=64M@0x00000000"

 

(The commands below are what is changed from the original string in the string above.)

"ttyS0 -> ttyS1"

"8M -> 16M"

 

Save and exit this window, and the last window will appear, make the following change:

 

Network device support ---> Wireless LAN (non-hamradio)

[*] Wireless LAN (non-hamradio)

 

Exit and save, the menu configuration portion is now complete.

 

7.)    Before building the images, a few files in the snapgear source need to be replaced and modified. From the snapgear directory:

 

cd linux-2.4.x/arch/arm/mach-ixp425

 

vi ixdp425-pci.c

 

Make the following changes:

 

Line 63:  Ò#define IXP425_PCI_MAX_DEV_4Ó, change Ò4Ó to Ò5Ó

Line 70:  Add line Ò{INTD, INTA, INTB, INTC},Ó

 

8.)    Now change back to the snapgear directory and build the image.

 

make dep

 

make

 

The images are in /root/stareast/working4/snapgear/images.

 

NOTE: If during the ÒmakeÓ process you receive an error about string.h, copy the string.h file from /usr/include to /usr/include/linux:

 

cp /usr/include/string.h /usr/include/linux/string.h

 

This is necessary for Fedora Core, and possibly later versions of RedHat.

 

Loading Linux Images

 

            Images:         

                                          zImage – Linux kernel.

                                                      ramdisk.gz – compressed ramdisk that contains all of the binaries and other files that are used by linux when it is running.

 

1.)    Open a terminal program to communicate with the Stareast board. (setup of the terminal program is discussed in the Hardware Setup Section).

2.)    Turn on the power to the Stareast board and wait until the RedBoot prompt appears.

3.)    Load the kernel into RAM using the following command:

 

load –r –v –b –m ymodem –b 0x01600000

 

a.)   Using your terminal program send a file using Ymodem, and select the zImage file from your images directory.

 

4.)    Load the ramdisk into RAM using the following command:

 

load –r –v –b –m ymodem –b 0x00800000

 

a.)  Using your terminal program send a file using Ymodem, and select the              ramdisk.gz file from your images directory.

                                   

5.)    After the images are loaded into RAM they needed to be written to the flash memory so that they do not have to be uploaded again after a reboot.

 

a.)   At the RedBoot prompt type:

 

fis create –b 0x01600000 –l 0xb0000 zImage

 

This stores the data at 0x01600000 of size 0xb0000 into the flash memory and is titled zImage.

 

 

 

b.)   At the RedBoot prompt type:

 

fis create –b 0x00800000 –l 0x300000 ramdisk

 

This stores the data at 0x00800000 of size 0x300000 into the flash memory and is titled zImage.  For the ramdisk that we have used 0x300000 is an overestimate of the size of the image, but will not be a problem.  If you notice that after loading an image from flash memory that it has errors when running, you should check that the size specified when creating the flash file is large enough, otherwise only part of the image is stored.

 

Running Linux

 

Now that you have a kernel and a ramdisk loaded on the processor, connect via the terminal program.  To load up linux there are three quick commands:

           

      fis load kernel

      fis load ramdisk

      go –n 0x01600000

 

This will now load linux and you will be left with a prompt that is a hash(#).  There is a directory structure and a small amount of standard linux commands available to you.  If you have added your own executables to the ramdisk (see the Adding a Custom Program section) you will have access to run them from the prompt here.

 

Configuring Network Interfaces

 

The StarEast board has two network interfaces, ixp0 and ixp1.  To configure these with IP addresses simply use the ifconfig command as follows:

 

            ifconfig ixp0 10.0.0.1

      ifconfig ixp1 192.168.0.1

 

We ran into some trouble with the interfaces dropping their configurations.  By just typing ifconfig at any time you can see which interfaces are configured.  If either ixp0 or ixp1 does not show up it has lost its configuration and you will need to type in one of the above commands again.  You should set up the computers connected to these ports to have IP addresses in the same subnet.  The two ports themselves do not have to be on the same subnet though.  For example

           

      IXP0:  192.168.0.1 <--------> Computer 1: 192.168.0.5

      IXP1:  10.0.0.1    <--------> Computer 2: 10.0.0.7

 

Adding a custom program

 

In order to add your own custom program, you will have to modify several files and reconfigure and rebuild linux as outlined in the Building Linux section. 

 

1) Modify user/Makefile

 

Add the following line to the makefile:

 

            dir_$(CONFIG_USER_DIRNAME_PNAME)  += dirname

 

This will be the name of the directory that will hold your source code to be compiled.  The directory will be located at user/dirname

 

2) Modify config/Configure.help

 

This will modify the text that is used during the configuration process.  Add lines as follows:

 

            CONFIG_USER_DIRNAME_PNAME

  This program does stuff.

     

                  Notice that the second line is indented two spaces.

 

3) Modify config/config.in

Add a line in whichever menu section you want your program to show up in.  Towards the bottom is the Miscellaneous section, this can be a safe bet:

 

            bool ÔpnameÕ            CONFIG_USER_DIRNAME_PNAME

 

You can have more than one program in a directory, so you can add another line as follows:

 

      bool ÔotherpnameÕ CONFIG_USER_DIRNAME_OTHERPNAME

 

4) Create a program directory user/dirname, and place your source code in it.

 

5) Create a Makefile for your code.  Follow this template:

 

  

   EXEC = pname

   OBJS = pname.o

 

   all: $(EXEC)

 

   $(EXEC): $(OBJS)

        $(CC) $(LDFLAGS) -o $@ $(OBJS) $(LDLIBS)

 

   romfs:

        $(ROMFSINST)    /bin/$(EXEC)

 

   clean:

        -rm -f $(EXEC) *.elf *.gdb *.o

 

 

 

If your code has multiple executables follow this Makefile:

 

 

 

 

 

EXECS = pname otherpname

   OBJS = pname.o otherpname.o

 

   all: $(EXECS)

 

   $(EXECS): $(OBJS)

        $(CC) $(LDFLAGS) -o $@ $@.o $(LDLIBS)

 

   romfs:

        $(ROMFSINST) -e CONFIG_USER_DIRNAME_PNAME       /bin/pname

        $(ROMFSINST) -e CONFIG_USER_DIRNAME_OTHERPNAME  /bin/otherpname

 

 

 

6) All that is left to do is to rerun the configuration.  When doing so look in the menu you added your program to and make sure to check the checkbox.  Next run make dep and make.  This will give you a proper ramdisk.gz to upload to the IXP425.  You will be able to type pname at the prompt to execute the code.  The executable will be located in /bin .

 

 

 

References

 

For building Images:

 

[1] ÒStarEast Multi-Radio Wireless Platform Reference Manual.Ó  Included on the StarEast cd with the processor.

 

For adding custom Programs:

 

[2] D. P. Siddons, ÒAdding User Applications to the uClinux Distribution.Ó

http://cvs.uclinux.org/cgi-bin/cvsweb/~checkout~/uClinux-dist/Documentation/Adding-User-Apps-HOWTO?content-type=text/plain

 

For hardware setup and installing/running linux:

[3] ÒIXP4XX Open Source Developers GuideÓ http://ixp4xx-osdg.sourceforge.net/

 

Questions/Comments

 

This document was prepared by Robert Heath (rdheath@ucdavis-alumni.com)  and Adam Moerschell (atmoerschell@ucdavis.edu) under the supervision of Prof. Venkatesh Akella as part of the EEC-282 Class on embedded systems in Spring Quarter 2005.