Compiling Haiku for Arm
ARM Compiler Toolset
Building the ARM compiler toolset is quite easy and involves generating gcc binaries for your platform.
For a complete list of flags for the configure script, see Haiku's Configure Options
From the haiku source directory, run the following. (be sure to adjust the options to match your build environment.)
./configure --build-cross-tools arm ../buildtools --target-board beagle
If you want to run configure again to tweak some more options, you need to tell it to configure for ARM. This is done with the --cross-tools-prefix option:
../configure --use-gcc-pipe --cross-tools-prefix cross-tools/bin/arm-unknown-haiku- --target-board beagle
The target-board parameter will set the configuration for both the compiler and the Haiku image built from the generated Jamfile.
Available target boards are:
- beagle: build Haiku for the BeagleBoard
- raspberry_pi: build Haiku for the Raspberry Pi
- verdex: build Haiku for the Gumstix Verdex
ARM Haiku Builds
These builds require a valid ARM compiler toolset. The ARM ports are in an extremely early state, don't expect these to work.
Compiling Haiku for the Verdex (mostly qemu)
The Verdex is an extremely small ARM machine marketed by Gumstix. It utilizes a PXA270 ARM processor made by Marvell. Testing with QEMU requires a firmware image in addition to the SD card image. This is done by jamming haiku-flash-uimage.
jam -q haiku-mmc-image haiku-flash-uimage
Compiling Haiku for the Beagle Board
The Beagle Board is a 75x75 mm ARM computer with all the features you would find on a normal x86 desktop machine. The OMAP3530 includes an Cortex-A8 ARM processor and 256 MB of flash memory paired with 256 MB of RAM.
jam -q haiku-mmc-image
Compiling Haiku for the Raspberry Pi
The Raspberry Pi is a popular small ARM computer geared towards the educational market. Given the extremely low price (US$ 35) of the Raspberry Pi, demand has been quite high. The Pi includes a BCM2835 system on a chip (SoC), containing an ARM1176JZF-S processor bolted to the side of a VideoCore IV GPU. The more popular B model contains 256MB RAM with a configurable split between the ARM processor and the VideoCore GPU.
While the Raspberry Pi is attractive due to it's extremely low cost compared to it's moderate performance, it does have the drawback of the firmware being very closed source. The Raspberry Pi foundation does however work with developers to make improvements to the closed source firmware.
There is an extremely early port of Haiku to the Raspberry Pi.
jam -q -sHAIKU_BOOT_PLATFORM=raspberrypi_arm haiku_loader
jam -q -sHAIKU_BOOT_PLATFORM=raspberrypi_arm haiku-floppyboot.tgz