Software for Embedded Systems

Designed & Presented by,

Anand K

Embedded Software

Software interacting with the physical worldo it takes time.

o it consumes power

o it does not terminate.

Embedded softwareo hidden in watches, mobile phones, iPods

o guided missiles, controls satellites

o used in automotive instruments.

Varying requirements and hardware platform.

Requirements of Software for

Embedded Systems Reliability

Human intervention not possible for error handling

Cost- effectiveness

Low power consumption

Efficient use of memory.

Performance requirement Timeliness

Features of Embedded Software

Timeliness Computation takes time Even Infinitely fast computer embedded software needs to deal with time

because the physical processes with which it interacts evolves over time.

Concurrency Engage with the physical world where multiple things happen at once.

Liveliness Program must not terminate or block waiting for events that will never

occur.

Heterogeneity Different computational styles and implementation technologies

Interact with events occurring

Reactivity React continuously to the environment at the speed of the environment.

Real Time

The time between presentation of a set of inputs to a system and the realization of the required behavior including availability of the associated outputs, is called the response time of the system.

A real-time system is a system that must satisfy explicit bounded response time constraints or risk severe consequences including failure Soft Real-time systemsHard real-time systems

Software Quality Measurement

Dynamic efficiency Number of CPU cycles required.

Static efficiency RAM size Global variables and stack.

ROM size Fixed constants and program code.

Power consumption. Correctness. Easy to understand Easy to change Flexibility and maintainability

Developing Software

Determine requirements Product functionality

Design the System Architecture Select the Operating System(if any) Choose the development platform Code the application Optimize the code according to the requirement. Verify the software on the host system

Execution within a simulator environment.

Execution within an emulator environment.

Verify the software on the target system.

Make Process

Compilingo Parsing, Semantic Analysis and Object Code Generationo Processor specific characteristics to be exploited for

efficient code generationo Cross-compiler or cross-assembler

Run on host to produce code for target.

Linkingo Merge sections from multiple object fileso Unresolved reference to symbols replaced by reference to actual

variables or function callso A special object file that contains compiled start-up code also

includedo Start-up code:

Small block assembly language code that prepares the way for the execution of the code

Startup code

o Start-up code usually consists of following actions in sequence. Disable all interrupts.

Copy any predefined data from ROM to RAM.

Zero the uninitialized data area.

Allocate space and initialize the stack.

Initialize processor’s stack pointer

Create and initialize the heap

Enable Interrupts.

Call main

oProgrammer may need to write his own start-up code and link(if necessary)

Img.jpg

Memory Allocation

o RAM Memory Global variables.

Heap

Stack• Local variables

• Return addresses

• Temporary data.

o Content changes at run-time

Memory Allocation

o ROM memory Machine Codes

Fixed constants

o If both ROM and EPROM partitioning depending upon the requirements of the

design. Configuration data

Flash Data store(RAM buffer)

o Slower compared to RAM

Desirable HLL features

o Versatile parameter passing Call by value

Copied into stack at considerable execution time

Large data structures never passed by value.

Call by reference

Indirect instructions used; time for execution of procedure more.

Desirable features

Global variables Vs parameter lists

References to global variables faster- direct addressing

Parameters make interfaces clearly defined

Parameters can increase interrupt latency.• interrupts disabled during parameter passing

C for Embedded Systems

o The practices that characterize Embedded C development:

▫ In-line Assembly language▫ Device knowledge

▫ #pragma directive▫ Endianness

▫ Mechanical knowledge▫ knowledge of specific device/peripheral operation

o Processor dependent objects like characters, bytes, bits and addresses can be handled directly

to be manipulated to control interrupt controllers, CPU registers etc.

o The @ operator associates port register with identifier in #pragma port statements

o C provides special variable types like register, volatile, static, and constant

C for Embedded Systems

o Storage modifiers register – variables to be frequently used, compiler

places in a register. static – private communication in modules. Static

variables initialized only once, visible within a single compilation unit.

extern – extern variables are meant to be allocated memory only once across several modules.

auto – by default all local variables are declared as ‘auto’

o Data type Identifiers const

• Allocated in ROM or any other NV memory

volatile – variables whose values may change outside of the executing software.

Memory management in C

o Static Allocation Global variables, variables with static qualifier

Memory allocated during compile or link time

Advantage of modularising the data within a program design.

o Automatic Allocation At run time in stack for variables defined in every

function

Compiler provides support for allocation and access to these variables

Dynamic Allocation

Dynamic allocation Allocation of memory at run-time on demand

Allocation and de-allocation under program control.

Allocated from a large pool of unused memory area called Heap

Memory accessed indirectly, via pointer reference

Dynamic Allocation

Memory space in the heap is not always allocated and deallocated from one end (like the stack).

If a block of sufficient size is found then it is allocated and the pointer returned.

If no block is found to satisfy the request then a heap overflow error is returned.

Allocation in heap area Memory in heap as chunk of unit sizes 16,32, etc.

Memory is actually acquired or released in the units in which the memory is divided.o wastage of memory

o Internal fragmentation

Allocation of Stack and Heap Space for the Declaration of Array A

Heap Management

A map of space allocated is kept so that, when a block is deleted, the correct amount of space is returned to the free list.

Smaller unit size minimizes wastage but increases memory requirement for maintaining data about the allocated blocks

free() returns unit to heapo using free() more than once on the same pointer can

result in inconsistency

Difficulties

Memory leak memory allocated, not freed

not available subsequently

“Run forever” can result in system failure due to cumulative effect of small memory leak

Dangling pointers

Thank U