Programming Android in Ada

Slide: 1

Copyright
©
2013 AdaCore

José F. Ruiz
FOSDEM 2013, Brussels
Senior Software
Engineer

AdaCore

Programming Android in Ada
3 February 2013
Slide: 2

Copyright
©
2013 AdaCore

Outline
•
 
The
execution

platform


•
 
How to
build
applications
–
 
Using
the
virtual
machine
–
 
Interfacing
the
virtual
machine
–
 
Ada
compilers
and
related
technologies

•
 
The show
–
 
Or how to use
it

Slide: 3

Copyright
©
2013 AdaCore

Android
•
 
Linux-based operating system
–
 
Open source
•
 
Targets smartphones and tablets
–
 
But not only
•
 
Applications implemented mostly in Java
–
 
Available SDK, APIs, IDE
•
 
Main hardware platform is ARM
–
 
But also x86, MIPS
•
 
There are nice emulators to facilitate development and testing
Slide: 4

Copyright
©
2013 AdaCore

•
 
Linux kernel
•
 
Middleware
•
 
Libraries
•
 
APIs
•
 
Application framework
–
 
Includes Java-compatible libraries
–
 
Uses the
Dalvik
virtual machine with just-in-time compilation
–
 
Run
Dalvik
'
dex
-code' (
Dalvik
Executable)
–
 
Usually translated from Java
bytecode

Android software architecture
Slide: 5

Copyright
©
2013 AdaCore

Android SDK
•
 
Android Software Development Kit (SDK)
–
 
API libraries and developer tools necessary to build, test, and debug apps for Android
–
 
Includes simulator
–
 
http://developer.android.com/sdk/index.html

•
 
Android Native Development Kit (NDK)
–
 
Implement parts of your app using native-code languages (C, C++,
Ada
,…)
–
 
http://developer.android.com/tools/sdk/ndk/index.html

Slide: 6

Copyright
©
2013 AdaCore

The strategy for Ada
•
 
Use the Android SDK for what it handles well
–
 
Structure of the application, graphical interface, …
–
 
Java at this point
–
 
On the
Dalvik
VM
•
 
Use
GNATbench
to create a mixed Java/Ada application
–
 
Write the Ada subprogram as usual
–
 
Automatic creation of the glue code for the interface to Java
–
 
Use the generated Java code from the Android application
package
Hello
is
function

Msg
return
String;
end
Hello
;

import

com.example.Hello.Hello_Package
;

public class
MainActivity

extends
Activity {
…

view.SetText
(
Hello_Package.Msg
().
toString
());
…
}
Slide: 7

Copyright
©
2013 AdaCore

Interfacing Ada and Java
•
 
Goal: applications not written entirely in Java
–
 
Standard Java class library missing platform-dependent features
–
 
Reuse already existing code
–
 
Implement time-critical part of the application
•
 
A communication layer between the two languages
–
 
Java Native Interface (JNI)
–
 
Low level interface between the JVM and the native environment
–
 
Expensive and error-prone to use manually
•
 
An interface in the target language
–
 
Binding generation tool (Ada2Java)
•
 
A builder for the whole application
–
 
GPRbuild

•
 
A debugger for the whole application
–
 
Eclipse (with GDB & Java debugger underneath)
Slide: 8

Copyright
©
2013 AdaCore

How it works
I have code written in Ada

I want to use this code in a
Java project

JNI

procedure
Proc (P : Rec);

void
Proc (Rec_Bound P);

On a single machine…

Slide: 9

Copyright
©
2013 AdaCore

Ada2Java
•
 
Takes a set of Ada packages as input
–
 
The interface that we want to make available to Java
•
 
Process the public specifications, using ASIS
•
 
Generates Ada and Java connection to JNI
•
 
Generates a Java high level and type safe view of the Ada package
•
 
The user does not deal with JNI directly
Slide: 10

Copyright
©
2013 AdaCore

Example of binding generation
package
Test
is



function
Addition (A, B :
Integer
)
return

Integer
;

end
Test;
package
Test;

public

class
Test_Package {


static

public

int
Addition (
int
A,
int
B){
// Code …
}
}
public

static

void
main (String [] argv) {
int
r = Test_Package.Addition(10, 20);

System.out.println ("10 + 20 = " + r);
}
Slide: 11

Copyright
©
2013 AdaCore

Ada part of the toolset
•
 
GNATbench

–
 
Official Android IDE is Eclipse
–
 
Eclipse plug-in to build Ada libraries on Android
•
 
Ada-Java Interfacing Suite (AJIS)
–
 
Java applications on Android can invoke Ada code using interfaces automatically
generated by AJIS

•
 
GNAT for Android
–
 
For the part of the application running on Linux/Android (outside the VM)
–
 
Configured for ARM/Linux
Slide: 12

Copyright
©
2013 AdaCore

Ada tasks on Android
•
 
Ada task runs on top of operating system thread
–
 
One-to-one correspondence
•
 
Task dispatching policy
–
 
Can be selected with
pragma
Dispatching_Policy

–
 
SCHED_OTHER by default
–
 
Preemptive priority scheduling
–
 
pragma
Dispatching_Policy
(
FIFO_Within_Priorities
)
•
 
pthread_setschedparam
(Thread, SCHED_FIFO,
Param’Access
)
–
 
Means run until blocked (or preempted), no time slicing
–
 
Reduces non-determinism
•
 
Multiprocessing
–
 
Unfortunately, the thread affinity interface does not seem to be there
Slide: 13

Copyright
©
2013 AdaCore

Synchronization
•
 
A protected type is a data object with locks
–
 
Data encapsulation, accessible through interface (locked access routines)
–
 
functions (read the data with read lock)
–
 
procedures (read/write the data with write lock)
–
 
entries (wait until some condition is met, then read/write the data with write lock)
•
 
Priority inheritance
–
 
Guard against priority inversion
–
 
low priority task grabs resource X
–
 
high priority task needs resource X, waits
–
 
medium priority task preempts low priority task, and runs for a long time, holding up high priority task
–
 
Solution, while high priority task is waiting, lend high priority to low priority task
•
 
Implementation
–
 
Mutual exclusion
–
 
Mutex

–
 
Waiting / Signaling operations
–
 
Conditional variable
Slide: 14

Copyright
©
2013 AdaCore

System programming
•
 
Clock and delay
–
 
Clock uses
gettimeofday

–
 
Delay operations use timed conditional variables
–
 
We can wakeup the task before expiration if needed
•
 
Interrupt handling
–
 
Underlying signal mechanism
–
 
A server task per interrupt served
–
 
With specific mask to serve the required signal
–
 
Call to
sigwait

Slide: 15

Copyright
©
2013 AdaCore

Demo
•
 
Lunar Lander
–
 
Sample provided by the Android SDK
•
 
Reimplement
the computation of physics behavior in Ada
–
 
Position and speed of the spacecraft
Slide: 16

Copyright
©
2013 AdaCore

Conclusion
•
 
Simple way to benefit from the Android SDK and Ada
•
 
Do as much as is convenient with the Android SDK
•
 
Then use Ada for
–
 
CPU intensive, critical, portable, … parts
•
 
The trick is
–
 
To realize that the Android platform is a Linux platform on with there is a JVM
–
 
To use the automatic generation of Ada/Java bindings
•
 
Nice integration with Eclipse to facilitate usage