Once your code is built and you have verified that all necessary directories exist, power on and test your device with basic bring up, as described below. Bring up tests are typically designed to stress certain aspects of your system and allow you to characterize the device's behavior.
1. Confirm a Clean Installation of a Basic Linux Kernel
Before considering Android-specific modifications to the Linux kernel, verify that you can build, deploy, and boot a core Linux kernel on your target hardware.
Your kernel configuration file should include the following:
# CONFIG_ANDROID_GADGET is not set
# CONFIG_ANDROID_RAM_CONSOLE is not set
# CONFIG_ANDROID_TIMED_GPIO is not set
3. Write Drivers
Android ships with default drivers for all basic functionality but you'll likely want to write your own drivers (or at least customize the default drivers) for your own device depending on your hardware configuration. See the following topics for examples of how to write your own drivers.
An image represents the state of a system or part of a system stored in non-volatile memory. The build process should produce the following system images:
bootloader: The bootloader is a small program responsible for initiating loading of the operating system.
system: The system image stores a snapshot of the Android operating system.
data: The data image stores user data. Anything not saved to the device/data directory will be lost on reboot.
kernel: The kernel represents the most basic element of an operating system. Android's Linux kernel is responsible for managing the system's resources and acts as an abstraction layer between hardware and a system's applications.
ramdisk: RAMdisk defines a portion of Random Access Memory (RAM) that gets used as if it were a hard drive.
Configure the bootloader to load the kernel and RAMdisk into RAM and pass the RAMdisk address to the kernel on startup.
5. Boot the kernel and mount the RAMdisk.
6. Debug Android-specific init programs on RAMdisk
Android-specific init programs are found in device/system/init. Add LOG messages to help you debug potential problems with the LOG macro defined in device/system/init/init.c.
The init program directly mounts all filesystems and devices using either hard-coded file names or device names generated by probing the sysfs filesystem (thereby eliminating the need for a /etc/fstab file in Android). After device/system files are mounted, init reads /etc/init.rc and invokes the programs listed there (one of the first of which is the console shell).
Each of these applications is embedded Linux C/C++ and you can use any standard Linux debugging tool to troubleshoot applications that aren't running. Execute % make showcommands to determine precise build commands. gdbserver (the GNU debugger) is available in the bin directory of the system partition (please see http://sourceware.org/gdb/ for more information).
8. Pulling it all together
If bring up was successful, you should see the following Java applications (with icons) visible on the LCD panel:
com.google.android.phone: The Android contact application.
If they are not visible or unresponsive to keypad control, run the framebuffer/keypad tests.
Android Init Language
The Android Init Language consists of four broad classes of statements:
The language syntax includes the following conventions:
All classes are line-oriented and consist of tokens separated by whitespace. c-style backslash escapes may be used to insert whitespace into a token. Double quotes may also be used to prevent whitespace from breaking text into multiple tokens. A backslash
appearing as the last character on a line is used for line-folding.
Lines that start with a # (leading whitespace allowed) are comments.
Actions and Services implicitly declare new sections. All commands or options belong to the section most recently declared. Commands or options before the first section are ignored.
Actions and Services have unique names. If a second Action or Service is declared with the same name as an existing one, it is ignored as an error.
Actions are named sequences of commands. Actions have a trigger used to determine when the action should occur. When an event occurs which matches an action's trigger, that action is added to the tail of a to-be-executed queue (unless it is already on the queue).
Each action in the queue is dequeued in sequence. Each command in an action is executed in sequence. Init handles other activities (such as, device creation/destruction, property setting, process restarting) "between" the execution of the commands in activities.
Actions take the form of:
Services are programs that init launches and (optionally) restarts when they exit.
Services take the form of:
Options are modifiers to services that affect how and when init runs a service. Options are described in the table below:
This service will not automatically start with its class. It must be explicitly started by name.
socket [ [ ] ]
Create a unix domain socket named /dev/socket/ and pass its fd to the launched process. Valid values include dgram and stream. user and group default to 0.
Change to username before exec'ing this service. Currently defaults to root.
group [ ]*
Change to groupname before exec'ing this service. Additional groupnames beyond the first, which is required, are used to set additional groups of the process (with setgroups()). Currently defaults to root.
capability [ ]+
Set linux capability before exec'ing this service
Do not restart the service when it exits.
Specify a class name for the service. All services in a named class must start and stop together. A service is considered of class "default" if one is not specified via the class option.
Triggers are strings used to match certain kinds of events that cause an action to occur.
This is the first trigger that occurs when init starts (after /init.conf is loaded).