| Line 8... |
Line 8... |
vmlinux.dis.bz2 Kernel disassembly file, bzip2 compressed
|
vmlinux.dis.bz2 Kernel disassembly file, bzip2 compressed
|
vmlinux.mem.bz2 Kernel .mem file for Verilog simulations, bzip2 compressed
|
vmlinux.mem.bz2 Kernel .mem file for Verilog simulations, bzip2 compressed
|
If you build the kernal from source these 2 files
|
If you build the kernal from source these 2 files
|
get replaced.
|
get replaced.
|
|
|
|
|
# +++++++++++++++++++++++++++++++++++++++++++
|
# +++++++++++++++++++++++++++++++++++++++++++
|
# How to run Amber Linux kernel on a development board
|
# How to run Amber Linux kernel on a development board
|
# +++++++++++++++++++++++++++++++++++++++++++
|
# +++++++++++++++++++++++++++++++++++++++++++
|
1. Download the bitfile to configure the FPGA using Impact ir Chipscope
|
1. Download the bitfile to configure the FPGA using Impact or Chipscope
|
2. Connect HyperTerminal to the serial port on the FPGA to connect to the boot loader
|
2. Connect HyperTerminal to the serial port on the FPGA to connect to the boot loader
|
3. Download the disk image
|
3. Download the disk image
|
> b 800000
|
> b 800000
|
Then select the file $AMBER_BASE/sw/vmlinux/initrd to transfer
|
Then select one of the provided disk image files to transfer, e.g.
|
|
$AMBER_BASE/sw/vmlinux/initrd-200k-hello-world
|
|
|
4. Download the kernel image
|
4. Download the kernel image
|
> l
|
> l
|
Then select the file $AMBER_BASE/sw/vmlinux/vmlinux to transfer
|
Then select the file $AMBER_BASE/sw/vmlinux/vmlinux to transfer
|
|
|
5. Execute th ekernel
|
5. Execute th ekernel
|
> j
|
> j 80000
|
|
|
|
|
# +++++++++++++++++++++++++++++++++++++++++++
|
# +++++++++++++++++++++++++++++++++++++++++++
|
# How to build Amber Linux kernel from source
|
# How to build Amber Linux kernel from source
|
# +++++++++++++++++++++++++++++++++++++++++++
|
# +++++++++++++++++++++++++++++++++++++++++++
|
| Line 64... |
Line 68... |
|
|
cp vmlinux vmlinux_unstripped
|
cp vmlinux vmlinux_unstripped
|
${AMBER_CROSSTOOL}-objcopy -R .comment -R .note vmlinux
|
${AMBER_CROSSTOOL}-objcopy -R .comment -R .note vmlinux
|
${AMBER_CROSSTOOL}-objcopy --change-addresses -0x02000000 vmlinux
|
${AMBER_CROSSTOOL}-objcopy --change-addresses -0x02000000 vmlinux
|
$AMBER_BASE/sw/tools/amber-elfsplitter vmlinux > vmlinux.mem
|
$AMBER_BASE/sw/tools/amber-elfsplitter vmlinux > vmlinux.mem
|
|
|
# Add the ram disk image to the .mem file
|
# Add the ram disk image to the .mem file
|
$AMBER_BASE/sw/tools/amber-bin2mem ${AMBER_BASE}/sw/vmlinux/initrd 800000 >> vmlinux.mem
|
# You can use one of the provided disk images or generate your own (see below)
|
|
$AMBER_BASE/sw/tools/amber-bin2mem ${AMBER_BASE}/sw/vmlinux/initrd-200k-hello-world 800000 >> vmlinux.mem
|
${AMBER_CROSSTOOL}-objdump -C -S -EL vmlinux_unstripped > vmlinux.dis
|
${AMBER_CROSSTOOL}-objdump -C -S -EL vmlinux_unstripped > vmlinux.dis
|
cp vmlinux.mem $AMBER_BASE/sw/vmlinux/vmlinux.mem
|
cp vmlinux.mem $AMBER_BASE/sw/vmlinux/vmlinux.mem
|
cp vmlinux.dis $AMBER_BASE/sw/vmlinux/vmlinux.dis
|
cp vmlinux.dis $AMBER_BASE/sw/vmlinux/vmlinux.dis
|
|
|
# Run the Linux simulation to verify that you have a good kernel image
|
# Run the Linux simulation to verify that you have a good kernel image
|
| Line 89... |
Line 95... |
|
|
# Pick a directory on your system where you want to build Linux
|
# Pick a directory on your system where you want to build Linux
|
export LINUX_WORK_DIR=/proj/amber2-linux
|
export LINUX_WORK_DIR=/proj/amber2-linux
|
|
|
|
|
|
# Create the Linux build directory
|
|
test -e ${LINUX_WORK_DIR} || mkdir ${LINUX_WORK_DIR}
|
cd ${LINUX_WORK_DIR}
|
cd ${LINUX_WORK_DIR}
|
|
|
# Need root permissions to mount disks
|
# Need root permissions to mount disks
|
su root
|
su root
|
dd if=/dev/zero of=initrd bs=200k count=1
|
dd if=/dev/zero of=initrd bs=200k count=1
|
mke2fs -F -m0 -b 1024 initrd
|
mke2fs -F -m0 -b 1024 initrd
|
|
|
| Line 113... |
Line 122... |
mknod ${LINUX_WORK_DIR}/mnt/dev/null c 1 3
|
mknod ${LINUX_WORK_DIR}/mnt/dev/null c 1 3
|
mknod ${LINUX_WORK_DIR}/mnt/dev/loop0 b 7 0
|
mknod ${LINUX_WORK_DIR}/mnt/dev/loop0 b 7 0
|
chmod 600 ${LINUX_WORK_DIR}/mnt/dev/*
|
chmod 600 ${LINUX_WORK_DIR}/mnt/dev/*
|
|
|
cp $AMBER_BASE/sw/hello-world/hello-world.flt ${LINUX_WORK_DIR}/mnt/sbin/init
|
cp $AMBER_BASE/sw/hello-world/hello-world.flt ${LINUX_WORK_DIR}/mnt/sbin/init
|
|
#cp $AMBER_BASE/sw/dhry/dhry.flt ${LINUX_WORK_DIR}/mnt/sbin/init
|
chmod +x ${LINUX_WORK_DIR}/mnt/sbin/init
|
chmod +x ${LINUX_WORK_DIR}/mnt/sbin/init
|
|
|
# Check
|
# Check
|
df ${LINUX_WORK_DIR}/mnt
|
df ${LINUX_WORK_DIR}/mnt
|
|
|
# Unmount
|
# Unmount
|
umount ${LINUX_WORK_DIR}/mnt
|
umount ${LINUX_WORK_DIR}/mnt
|
rm -rf ${LINUX_WORK_DIR}/mnt
|
rm -rf ${LINUX_WORK_DIR}/mnt
|
exit
|
exit # from being root
|
|
|
|
cp initrd $AMBER_BASE/sw/vmlinux/initrd-
|
|
|
|
---
|
|
|
|
If 200k is not large enough, you can change the size as follows.
|
|
You'll need to change a couple of values in the ATAG data structure defined in the
|
|
boot loader. Specifically the ATAG_RAMDISK_SIZE parameter and the ATAG_INITRD_SIZE
|
|
parameter in file $AMBER_BASE/sw/boot-loader/start.S. Then create an initrd image
|
|
with a different bs number, for example;
|
|
dd if=/dev/zero of=initrd bs=400k count=1
|
|
|
|
The initrd image size gets picked up automatically by the kernel, as long as the
|
|
ram disk defined in the ATAG data is large enough to contain it.
|
|
|
cp initrd $AMBER_BASE/sw/vmlinux
|
|
