MPLAB® XC8 C Compiler

User’s Guide

 2012-2013 Microchip Technology Inc. DS50002053D

Note the following details of the code protection feature on Microchip devices:
• Microchip products meet the specification contained in their particular Microchip Data Sheet.

• Microchip believes that its family of products is one of the most secure families of its kind on the market today, when used in the
intended manner and under normal conditions.

• There are dishonest and possibly illegal methods used to breach the code protection feature. All of these methods, to our
knowledge, require using the Microchip products in a manner outside the operating specifications contained in Microchip’s Data
Sheets. Most likely, the person doing so is engaged in theft of intellectual property.

• Microchip is willing to work with the customer who is concerned about the integrity of their code.

• Neither Microchip nor any other semiconductor manufacturer can guarantee the security of their code. Code protection does not
mean that we are guaranteeing the product as “unbreakable.”

Code protection is constantly evolving. We at Microchip are committed to continuously improving the code protection features of our
products. Attempts to break Microchip’s code protection feature can be a violation of the Digital Millennium Copyright Act. If such acts
allow unauthorized access to your software or other copyrighted work, you can have a right to sue for relief under that Act.

Information contained in this publication regarding device Trademarks

applications and the like is provided only for your convenience The Microchip name and logo, the Microchip logo, dsPIC,
and can be superseded by updates. It is your responsibility to FlashFlex, KEELOQ, KEELOQ logo, MPLAB, PIC, PICmicro,
ensure that your application meets with your specifications. PICSTART, PIC32 logo, rfPIC, SST, SST Logo, SuperFlash
MICROCHIP MAKES NO REPRESENTATIONS OR and UNI/O are registered trademarks of Microchip Technology
WARRANTIES OF ANY KIND WHETHER EXPRESS OR Incorporated in the U.S.A. and other countries.
IMPLIED, WRITTEN OR ORAL, STATUTORY OR
OTHERWISE, RELATED TO THE INFORMATION, FilterLab, Hampshire, HI-TECH C, Linear Active Thermistor,
INCLUDING BUT NOT LIMITED TO ITS CONDITION, MTP, SEEVAL and The Embedded Control Solutions
QUALITY, PERFORMANCE, MERCHANTABILITY OR Company are registered trademarks of Microchip Technology
FITNESS FOR PURPOSE. Microchip disclaims all liability Incorporated in the U.S.A.
arising from this information and its use. Use of Microchip Silicon Storage Technology is a registered trademark of
devices in life support and/or safety applications is entirely at Microchip Technology Inc. in other countries.
the buyer’s risk, and the buyer agrees to defend, indemnify and
Analog-for-the-Digital Age, Application Maestro, BodyCom,
hold harmless Microchip from any and all damages, claims,
chipKIT, chipKIT logo, CodeGuard, dsPICDEM,
suits, or expenses resulting from such use. No licenses are
[Link], dsPICworks, dsSPEAK, ECAN,
conveyed, implicitly or otherwise, under any Microchip
ECONOMONITOR, FanSense, HI-TIDE, In-Circuit Serial
intellectual property rights.
Programming, ICSP, Mindi, MiWi, MPASM, MPF, MPLAB
Certified logo, MPLIB, MPLINK, mTouch, Omniscient Code
Generation, PICC, PICC-18, PICDEM, [Link], PICkit,
PICtail, REAL ICE, rfLAB, Select Mode, SQI, Serial Quad I/O,
Total Endurance, TSHARC, UniWinDriver, WiperLock, ZENA
and Z-Scale are trademarks of Microchip Technology
Incorporated in the U.S.A. and other countries.
SQTP is a service mark of Microchip Technology Incorporated
in the U.S.A.
GestIC and ULPP are registered trademarks of Microchip
Technology Germany II GmbH & Co. KG, a subsidiary of
Microchip Technology Inc., in other countries.
All other trademarks mentioned herein are property of their
respective companies.
© 2012-2013, Microchip Technology Incorporated, Printed in
the U.S.A., All Rights Reserved.
Printed on recycled paper.
ISBN: 978-1-62077-723-7

QUALITY MANAGEMENT SYSTEM Microchip received ISO/TS-16949:2009 certification for its worldwide
CERTIFIED BY DNV headquarters, design and wafer fabrication facilities in Chandler and
Tempe, Arizona; Gresham, Oregon and design centers in California
and India. The Company’s quality system processes and procedures
== ISO/TS 16949 == are for its PIC® MCUs and dsPIC® DSCs, KEELOQ® code hopping
devices, Serial EEPROMs, microperipherals, nonvolatile memory and
analog products. In addition, Microchip’s quality system for the design
and manufacture of development systems is ISO 9001:2000 certified.

DS50002053D-page 2  2012-2013 Microchip Technology Inc.

 MPLAB® XC8 C COMPILER
USER’S GUIDE

Table of Contents
Preface ........................................................................................................................... 7
Chapter 1. Compiler Overview
1.1 Introduction ................................................................................................... 13
1.2 Compiler Description and Documentation .................................................... 13
1.3 Device Description ....................................................................................... 14
Chapter 2. Common C Interface
2.1 Introduction ................................................................................................... 15
2.2 Background – The Desire for Portable Code ............................................... 15
2.3 Using the CCI ............................................................................................... 18
2.4 ANSI Standard Refinement .......................................................................... 19
2.5 ANSI Standard Extensions ........................................................................... 27
2.6 Compiler Features ........................................................................................ 41
Chapter 3. How To’s
3.1 Introduction ................................................................................................... 43
3.2 Installing and Activating the Compiler .......................................................... 43
3.3 Invoking the Compiler ................................................................................... 45
3.4 Writing Source Code .................................................................................... 48
3.5 Getting My Application to Do What I Want ................................................... 59
3.6 Understanding the Compilation Process ...................................................... 64
3.7 Fixing Code That Does Not Work ................................................................. 72
Chapter 4. XC8 Command-line Driver
4.1 Introduction ................................................................................................... 77
4.2 Invoking the Compiler ................................................................................... 78
4.3 The Compilation Sequence .......................................................................... 81
4.4 Runtime Files ............................................................................................... 87
4.5 Compiler Output ........................................................................................... 89
4.6 Compiler Messages ...................................................................................... 91
4.7 MPLAB XC8 Driver Options ......................................................................... 96
4.8 Option Descriptions ...................................................................................... 97
4.9 MPLAB X Option Equivalents ..................................................................... 127
Chapter 5. C Language Features
5.1 Introduction ................................................................................................. 135
5.2 ANSI C Standard Issues ............................................................................ 135
5.3 Device-Related Features ............................................................................ 137
5.4 Supported Data Types and Variables ........................................................ 147
5.5 Memory Allocation and Access .................................................................. 168
5.6 Operators and Statements ......................................................................... 186

 2012-2013 Microchip Technology Inc. DS50002053D-page 3

MPLAB® XC8 C Compiler User’s Guide

5.7 Register Usage ........................................................................................... 188

5.8 Functions .................................................................................................... 189
5.9 Interrupts .................................................................................................... 198
5.10 Main, Runtime Startup and Reset ............................................................ 204
5.11 Library Routines ....................................................................................... 208
5.12 Mixing C and Assembly Code .................................................................. 210
5.13 Optimizations ............................................................................................ 219
5.14 Preprocessing .......................................................................................... 222
5.15 Linking Programs ..................................................................................... 234
Chapter 6. Macro Assembler
6.1 Introduction ................................................................................................. 255
6.2 Assembler Usage ....................................................................................... 255
6.3 Options ....................................................................................................... 256
6.4 MPLAB XC8 Assembly Language .............................................................. 260
6.5 Assembly-Level Optimizations ................................................................... 286
6.6 Assembly List Files ..................................................................................... 287
Chapter 7. Linker
7.1 Introduction ................................................................................................. 297
7.2 Operation .................................................................................................... 297
7.3 Relocation and Psects ................................................................................ 306
7.4 Map Files .................................................................................................... 307
Chapter 8. Utilities
8.1 Introduction ................................................................................................. 313
8.2 Librarian ..................................................................................................... 314
8.3 HEXMATE .................................................................................................. 317
Appendix A. Library Functions
A.1 Introduction ................................................................................................ 327
Appendix B. Embedded Compiler Compatibility Mode
B.1 Introduction ................................................................................................ 421
B.2 Compiling in Compatibility Mode ................................................................ 421
B.3 Syntax Compatibility .................................................................................. 422
B.4 Data Type .................................................................................................. 423
B.5 Operator ..................................................................................................... 423
B.6 Extended Keywords ................................................................................... 424
B.7 Intrinsic Functions ...................................................................................... 425
B.8 Pragmas ..................................................................................................... 426
Appendix C. Error and Warning Messages
C.1 Introduction ................................................................................................ 427
Appendix D. Implementation-Defined Behavior
D.1 Introduction ................................................................................................ 541
D.2 Translation (G.3.1) ..................................................................................... 541
D.3 Environment (G.3.2) .................................................................................. 541
D.4 Identifiers (G.3.3) ....................................................................................... 542

DS50002053D-page 4  2012-2013 Microchip Technology Inc.

 D.5 Characters (G.3.4) ..................................................................................... 542
D.6 Integers (G.3.5) .......................................................................................... 543
D.7 Floating-Point (G.3.6) ................................................................................ 544
D.8 Arrays and Pointers (G.3.7) ....................................................................... 544
D.9 Registers (G.3.8) ....................................................................................... 544
D.10 Structures, Unions, Enumerations, and Bit-Fields (G.3.9) ....................... 545
D.11 Qualifiers (G.3.10) ................................................................................... 545
D.12 Declarators (G.3.11) ................................................................................ 545
D.13 Statements (G.3.12) ................................................................................ 545
D.14 Preprocessing Directives (G.3.13) ........................................................... 546
D.15 Library Functions (G.3.14) ....................................................................... 547
Glossary ..................................................................................................................... 549
Index ........................................................................................................................... 569
Worldwide Sales and Service .................................................................................. 580

 2012-2013 Microchip Technology Inc. DS50002053D-page 5

MPLAB® XC8 C Compiler User’s Guide
NOTES:

DS50002053D-page 6  2012-2013 Microchip Technology Inc.

 MPLAB® XC8 C COMPILER
USER’S GUIDE

Preface

NOTICE TO CUSTOMERS
All documentation becomes dated, and this manual is no exception. Microchip tools and
documentation are constantly evolving to meet customer needs, so some actual dialogs
and/or tool descriptions can differ from those in this document. Please refer to our web site
([Link]) to obtain the latest documentation available.
Documents are identified with a “DS” number. This number is located on the bottom of each
page, in front of the page number. The numbering convention for the DS number is
“DSXXXXXA”, where “XXXXX” is the document number and “A” is the revision level of the
document.
For the most up-to-date information on development tools, see the MPLAB® IDE online help.
Select the Help menu, and then Topics to open a list of available online help files.

INTRODUCTION
This chapter contains general information that will be useful to know before using the
MPLAB® XC8 C Compiler User’s Guide. Items discussed in this chapter include:
• Document Layout
• Conventions Used in this Guide
• Recommended Reading
• Recommended Reading
• The Microchip Web Site
• Development Systems Customer Change Notification Service
• Customer Support
• Document Revision History

DOCUMENT LAYOUT
The MPLAB XC8 C Compiler User’s Guide is organized as follows:
• Chapter 1. Compiler Overview
• Chapter 2. Common C Interface
• Chapter 3. How To’s
• Chapter 4. XC8 Command-line Driver
• Chapter 5. C Language Features
• Chapter 6. Macro Assembler
• Chapter 7. Linker
• Chapter 8. Utilities
• Appendix A. Library Functions
• Appendix B. Embedded Compiler Compatibility Mode
• Appendix C. Error and Warning Messages
• Appendix D. Implementation-Defined Behavior
• Glossary
• Index

 2012-2013 Microchip Technology Inc. DS50002053D-page 7

MPLAB® XC8 C Compiler User’s Guide

CONVENTIONS USED IN THIS GUIDE

This manual uses the following documentation conventions:

DOCUMENTATION CONVENTIONS
Description Represents Examples
Arial font:
Italic characters Referenced books MPLAB® IDE User’s Guide
Emphasized text ...is the only compiler...
Initial caps A window the Output window
A dialog the Settings dialog
A menu selection select Enable Programmer
Quotes A field name in a window or “Save project before build”
dialog
Underlined, italic text with A menu path File>Save
right angle bracket
Bold characters A dialog button Click OK
A tab Click the Power tab
N‘Rnnnn A number in verilog format, 4‘b0010, 2‘hF1
where N is the total number of
digits, R is the radix and n is a
digit.
Text in angle brackets < > A key on the keyboard Press <Enter>, <F1>
Courier New font:
Plain Courier New Sample source code #define START
Filenames [Link]
File paths c:\mcc18\h
Keywords _asm, _endasm, static
Command-line options -Opa+, -Opa-
Bit values 0, 1
Constants 0xFF, ‘A’
Italic Courier New A variable argument file.o, where file can be
any valid filename
Square brackets [ ] Optional arguments mcc18 [options] file
[options]
Curly brackets and pipe Choice of mutually exclusive errorlevel {0|1}
character: { | } arguments; an OR selection
Ellipses... Replaces repeated text var_name [,
var_name...]
Represents code supplied by void main (void)
user { ...
}

DS50002053D-page 8  2012-2013 Microchip Technology Inc.

 Preface

RECOMMENDED READING
This user’s guide describes how to use MPLAB XC8 C Compiler. Other useful docu-
ments are listed below. The following Microchip documents are available and
recommended as supplemental reference resources.
Readme for MPLAB XC8 C Compiler
For the latest information on using MPLAB XC8 C Compiler, read MPLAB® XC8 C
Compiler Release Notes (a PDF file) in the Docs subdirectory of the compiler’s instal-
lation directory. The release notes contain update information and known issues that
cannot be included in this user’s guide.
Readme Files
For the latest information on using other tools, read the tool-specific Readme files in
the Readmes subdirectory of the MPLAB IDE installation directory. The Readme files
contain update information and known issues that cannot be included in this user’s
guide.

THE MICROCHIP WEB SITE

Microchip provides online support via our web site at [Link]. This web
site is used as a means to make files and information easily available to customers.
Accessible by using your favorite Internet browser, the web site contains the following
information:
• Product Support – Data sheets and errata, application notes and sample
programs, design resources, user’s guides and hardware support documents,
latest software releases and archived software
• General Technical Support – Frequently Asked Questions (FAQs), technical
support requests, online discussion groups, Microchip consultant program
member listing
• Business of Microchip – Product selector and ordering guides, latest Microchip
press releases, listing of seminars and events, listings of Microchip sales offices,
distributors and factory representatives

 2012-2013 Microchip Technology Inc. DS50002053D-page 9

MPLAB® XC8 C Compiler User’s Guide

DEVELOPMENT SYSTEMS CUSTOMER CHANGE NOTIFICATION SERVICE

Microchip’s customer notification service helps keep customers current on Microchip
products. Subscribers will receive e-mail notification whenever there are changes,
updates, revisions or errata that are related to a specified product family or
development tool of interest.
To register, access the Microchip web site at [Link], click on Customer
Change Notification and follow the registration instructions.
The Development Systems product group categories are:
• Compilers – The latest information on Microchip C compilers, assemblers, linkers
and other language tools. These include all MPLAB C compilers; all MPLAB
assemblers (including MPASM™ assembler); all MPLAB linkers (including
MPLINK™ object linker); and all MPLAB librarians (including MPLIB™ object
librarian).
• Emulators – The latest information on Microchip in-circuit [Link]
includes the MPLAB REAL ICE™ and MPLAB ICE 2000 in-circuit emulators.
• In-Circuit Debuggers – The latest information on the Microchip in-circuit
debuggers. This includes MPLAB ICD 3 in-circuit debuggers and PICkit™ 3
debug express.
• MPLAB® IDE – The latest information on Microchip MPLAB IDE, the Windows®
Integrated Development Environment for development systems tools. This list is
focused on the MPLAB IDE, MPLAB IDE Project Manager, MPLAB Editor and
MPLAB SIM simulator, as well as general editing and debugging features.
• Programmers – The latest information on Microchip programmers. These include
production programmers such as MPLAB REAL ICE in-circuit emulator, MPLAB
ICD 3 in-circuit debugger and MPLAB PM3 device programmers. Also included
are nonproduction development programmers such as PICSTART® Plus and
PICkit 2 and 3.

CUSTOMER SUPPORT
Users of Microchip products can receive assistance through several channels:
• Distributor or Representative
• Local Sales Office
• Field Application Engineer (FAE)
• Technical Support
Customers should contact their distributor, representative or field application engineer
(FAE) for support. Local sales offices are also available to help customers. A listing of
sales offices and locations is included in the back of this document.
Technical support is available through the web site at:
[Link]

DS50002053D-page 10  2012-2013 Microchip Technology Inc.

 Preface

DOCUMENT REVISION HISTORY

Revision D (Dec 2013)
• Added new information relating to the software stack and function reentrancy.
• Added information relating to code profiling features offered by the compiler.
• Removed information pertaining to MPLAB IDE v8.
• Added new “How To’s”
• Removed sections on OBJTOHEX and CROMWELL.
• Added additional information relating to assembly code formats and operators.
• Corrected Fletcher algorithms used by HEXMATE.
• Added new driver options and updated existing option descriptions.
• Added and updated macros, built-ins and functions in Library Function chapter.
• Updated compiler warning and error messages.

Revision C (May 2013)

• Added Embedded Compiler Compatibility Mode chapter.
• Added information relating to new ELF/DWARF debugging files.
• Added new driver options and updated existing option descriptions.
• Updated MPLAB X IDE option dialog descriptions relating to compiler options.
• Expanded information relating to the available optimizations.
• Added code to illustrate checksum algorithms used by HEXMATE.
• Updated compiler warning and error messages.
• Updated information relating to list and map file contents.
• Added information about multiplication routines.
• Expanded information about eeprom variables and bit objects.
• Expanded information relating to the configuration pragma.
• Added information and examples using the __section() specifier.
• Expanded and extended information relating to assembly code deviations and
assembler directives.

Revision B (July 2012)

• Added How To’s chapter.
• Expanded section relating to PIC18 erratas.
• Updated the section relating to compiler optimization settings.
• Updated MPLAB v8 and MPLAB X IDE project option dialogs.
• Added sections describing PIC18 far qualifier and in-line function qualifier.
• Expanded section describing the operation of the main() function
• Expanded information about equivalent assembly symbols for Baseline parts.
• Updated the table of predefined macro symbols.
• Added section on #pragma addrqual
• Added sections to do with in-lining functions
• Updated diagrams and text associated with call graphs in the list file
• Updated library function section to be consistent with packaged libraries
• Added new compiler warnings and errors.
• Added new chapter describing the Common C Interface Standard (CCI)

Revision A (February 2012)

Initial release of this document.

 2012-2013 Microchip Technology Inc. DS50002053D-page 11

MPLAB® XC8 C Compiler User’s Guide
NOTES:

DS50002053D-page 12  2012-2013 Microchip Technology Inc.

 MPLAB® XC8 C COMPILER
USER’S GUIDE

Chapter 1. Compiler Overview

1.1 INTRODUCTION
This chapter is an overview of the MPLAB® XC8 C Compiler, including these topics.
• Compiler Description and Documentation
• Device Description

1.2 COMPILER DESCRIPTION AND DOCUMENTATION

The MPLAB XC8 C Compiler is a free-standing, optimizing ISO C90 (popularly known
as ANSI C) compiler. It supports all 8-bit PIC® microcontrollers: PIC10, PIC12, PIC16
and PIC18 series devices, as well as the PIC14000 device.
The compiler is available for several popular operating systems, including 32- and
64-bit Windows® (excluding Windows Server), Linux® and Mac OS® X.
The compiler is available in three operating modes: Free, Standard or PRO. The Stan-
dard and PRO operating modes are licensed modes and require a serial number to
enable them. Free mode is available for unlicensed customers. The basic compiler
operation, supported devices and available memory are identical across all modes.
The modes only differ in the level of optimization employed by the compiler.

1.2.1 Conventions
Throughout this manual, the term “compiler” is used. It can refer to all, or a subset of,
the collection of applications that comprise the MPLAB XC8 C Compiler. When it is not
important to identify which application performed an action, it will be attributed to “the
compiler”.
Likewise, “compiler” is often used to refer to the command-line driver. Although specif-
ically, the driver for the MPLAB XC8 C Compiler package is called xc8. The driver and
its options are discussed in Section 4.7 “MPLAB XC8 Driver Options”. Accordingly,
“compiler options” commonly relates to command-line driver options.
In a similar fashion, “compilation” refers to all or a selection of steps involved in
generating source code into an executable binary image.

 2012-2013 Microchip Technology Inc. DS50002053D-page 13

MPLAB® XC8 C Compiler User’s Guide

1.3 DEVICE DESCRIPTION

This compiler supports 8-bit Microchip PIC devices with baseline, mid-range,
Enhanced mid-range, and PIC18 cores. The following descriptions indicate the
distinctions within those device cores:
The baseline core uses a 12-bit-wide instruction set and is available in PIC10, PIC12
and PIC16 part numbers.
The enhanced baseline core also use a 12-bit instruction set, but this set includes addi-
tional instructions. Some of the enhanced baseline chips support interrupts and the
additional instructions used by interrupts. These devices are available in PIC12 and
PIC16 part numbers.
The mid-range core uses a 14-bit-wide instruction set that includes more instructions
than the baseline core. It has larger data memory banks and program memory pages,
as well. It is available in PIC12, PIC14 and PIC16 part numbers.
The Enhanced mid-range core also uses a 14-bit-wide instruction set but incorporates
additional instructions and features. There are both PIC12 and PIC16 part numbers
that are based on the Enhanced mid-range core.
The PIC18 core instruction set is 16 bits wide and features additional instructions and
an expanded register set. PIC18 core devices have part numbers that begin with
PIC18.
The compiler takes advantage of the target device’s instruction set, addressing modes,
memory, and registers whenever possible.
See Section 4.8.19 “--CHIPINFO: Display List of Supported Devices” for
information on finding the full list of devices that are supported by the compiler.

DS50002053D-page 14  2012-2013 Microchip Technology Inc.

 MPLAB® XC8 C COMPILER
USER’S GUIDE

Chapter 2. Common C Interface

2.1 INTRODUCTION
The Common C Interface (CCI) is available with all MPLAB XC C compilers and is
designed to enhance code portability between these compilers. For example,
CCI-conforming code would make it easier to port from a PIC18 MCU using the MPLAB
XC8 C compiler to a PIC24 MCU using the MPLAB XC16 C compiler.
The CCI assumes that your source code already conforms to the ANSI Standard. If you
intend to use the CCI, it is your responsibility to write code that conforms. Legacy proj-
ects will need to be migrated to achieve conformance. A compiler option must also be
set to ensure that the operation of the compiler is consistent with the interface when the
project is built.
The following topics are examined in this chapter of the MPLAB XC8 C Compiler User’s
Guide:
• Background – The Desire for Portable Code
• Using the CCI
• ANSI Standard Refinement
• ANSI Standard Extensions
• Compiler Features

2.2 BACKGROUND – THE DESIRE FOR PORTABLE CODE

All programmers want to write portable source code.
Portability means that the same source code can be compiled and run in a different
execution environment than that for which it was written. Rarely can code be one hun-
dred percent portable, but the more tolerant it is to change, the less time and effort it
takes to have it running in a new environment.
Embedded engineers typically think of code portability as being across target devices,
but this is only part of the situation. The same code could be compiled for the same
target but with a different compiler. Differences between those compilers might lead to
the code failing at compile time or runtime, so this must be considered as well.
You can only write code for one target device and only use one brand of compiler, but
if there is no regulation of the compiler’s operation, simply updating your compiler
version can change your code’s behavior.
Code must be portable across targets, tools, and time to be truly flexible.
Clearly, this portability cannot be achieved by the programmer alone, since the com-
piler vendors can base their products on different technologies, implement different fea-
tures and code syntax, or improve the way their product works. Many a great compiler
optimization has broken many an unsuspecting project.
Standards for the C language have been developed to ensure that change is managed
and code is more portable. The American National Standards Institute (ANSI) pub-
lishes standards for many disciplines, including programming languages. The ANSI C
Standard is a universally adopted standard for the C programming language.

 2012-2013 Microchip Technology Inc. DS50002053D-page 15

MPLAB® XC8 C Compiler User’s Guide

2.2.1 The ANSI Standard

The ANSI C Standard has to reconcile two opposing goals: freedom for compilers ven-
dors to target new devices and improve code generation, with the known functional
operation of source code for programmers. If both goals can be met, source code can
be made portable.
The standard is implemented as a set of rules which detail not only the syntax that a
conforming C program must follow, but the semantic rules by which that program will
be interpreted. Thus, for a compiler to conform to the standard, it must ensure that a
conforming C program functions as described by the standard.
The standard describes implementation, the set of tools, and the runtime environment
on which the code will run. If any of these change, e.g., you build for, and run on, a dif-
ferent target device, or if you update the version of the compiler you use to build, then
you are using a different implementation.
The standard uses the term behavior to mean the external appearance or action of the
program. It has nothing to do with how a program is encoded.
Since the standard is trying to achieve goals that could be construed as conflicting,
some specifications appear somewhat vague. For example, the standard states that an
int type must be able to hold at least a 16-bit value, but it does not go as far as saying
what the size of an int actually is; and the action of right-shifting a signed integer can
produce different results on different implementations; yet, these different results are
still ANSI C compliant.
If the standard is too strict, device architectures cannot allow the compiler to conform.1
But, if it is too weak, programmers would see wildly differing results within different
compilers and architectures, and the standard would lose its effectiveness.
The standard organizes source code whose behavior is not fully defined into groups
that include the following behaviors:
Implementation-defined behavior
This is unspecified behavior in which each implementation documents how the choice
is made.
Unspecified behavior
The standard provides two or more possibilities and imposes no further requirements
on which possibility is chosen in any particular instance.
Undefined behavior
This is behavior for which the standard imposes no requirements.
Code that strictly conforms to the standard does not produce output that is dependent
on any unspecified, undefined, or implementation-defined behavior. The size of an
int, which was used as an example earlier, falls into the category of behavior that is
defined by implementation. That is to say, the size of an int is defined by which com-
piler is being used, how that compiler is being used, and the device that is being
targeted.
All the MPLAB XC compilers conform to the ANS X3.159-1989 Standard for program-
ming languages (with the exception of the MPLAB XC8 compiler’s inability to allow
recursion, as mentioned in the footnote). This is commonly called the C89 Standard.
Some features from the later standard, C99, are also supported.

1. For example, the mid-range PIC® microcontrollers do not have a data stack. Because a compiler
targeting this device cannot implement recursion, it (strictly speaking) cannot conform to the ANSI
C Standard. This example illustrates a situation in which the standard is too strict for mid-range
devices and tools.

DS50002053D-page 16  2012-2013 Microchip Technology Inc.

 Common C Interface
For freestanding implementations (or for what we typically call embedded applications),
the standard allows non-standard extensions to the language, but obviously does not
enforce how they are specified or how they work. When working so closely to the
device hardware, a programmer needs a means of specifying device setup and inter-
rupts, as well as utilizing the often complex world of small-device memory
architectures. This cannot be offered by the standard in a consistent way.
While the ANSI C Standard provides a mutual understanding for programmers and
compiler vendors, programmers need to consider the implementation-defined behavior
of their tools and the probability that they may need to use extensions to the C language
that are non-standard. Both of these circumstances can have an impact on code
portability.

2.2.2 The Common C Interface

The Common C Interface (CCI) supplements the ANSI C Standard and makes it easier
for programmers to achieve consistent outcomes on all Microchip devices when using
any of the MPLAB XC C compilers.
It delivers the following improvements, all designed with portability in mind.
Refinement of the ANSI C Standard
The CCI documents specific behavior for some code in which actions are implemen-
tation-defined behavior under the ANSI C Standard. For example, the result of
right-shifting a signed integer is fully defined by the CCI. Note that many
implementation-defined items that closely couple with device characteristics, such as
the size of an int, are not defined by the CCI.
Consistent syntax for non-standard extensions
The CCI non-standard extensions are mostly implemented using keywords with a uni-
form syntax. They replace keywords, macros and attributes that are the native com-
piler implementation. The interpretation of the keyword can differ across each com-
piler, and any arguments to the keywords can be device specific.
Coding guidelines
The CCI can indicate advice on how code should be written so that it can be ported
to other devices or compilers. While you may choose not to follow the advice, it will
not conform to the CCI.

 2012-2013 Microchip Technology Inc. DS50002053D-page 17

MPLAB® XC8 C Compiler User’s Guide

2.3 USING THE CCI

The CCI allows enhanced portability by refining implementation-defined behavior and
standardizing the syntax for extensions to the language.
The CCI is something you choose to follow and put into effect, thus it is relevant for new
projects, although you can choose to modify existing projects so they conform.
For your project to conform to the CCI, you must do the following things.
Enable the CCI
Select the MPLAB IDE widget Use CCI Syntax in your project, or use the
command-line option that is equivalent.
Include <xc.h> in every module
Some CCI features are only enabled if this header is seen by the compiler.
Ensure ANSI compliance
Code that does not conform to the ANSI C Standard does not confirm to the CCI.
Observe refinements to ANSI by the CCI
Some ANSI implementation-defined behavior is defined explicitly by the CCI.
Use the CCI extensions to the language
Use the CCI extensions rather than the native language extensions
The next sections detail specific items associated with the CCI. These items are seg-
regated into those that refine the standard, those that deal with the ANSI C Standard
extensions, and other miscellaneous compiler options and usage. Guidelines are
indicated with these items.
If any implementation-defined behavior or any non-standard extension is not discussed
in this document, then it is not part of the CCI. For example, GCC case ranges, label
addresses, and 24-bit short long types are not part of the CCI. Programs which use
these features do not conform to the CCI. The compiler may issue a warning or error
to indicate a non-CCI feature has been used and the CCI is enabled.

DS50002053D-page 18  2012-2013 Microchip Technology Inc.

 Common C Interface

2.4 ANSI STANDARD REFINEMENT

The following topics describe how the CCI refines the implementation-defined
behaviors outlined in the ANSI C Standard.

2.4.1 Source File Encoding

Under the CCI, a source file must be written using characters from the 7-bit ASCII set.
Lines can be terminated using a line feed ('\n') or carriage return ('\r') that is immediately
followed by a line feed. Escaped characters can be used in character constants or
string literals to represent extended characters that are not in the basic character set.

[Link] EXAMPLE
The following shows a string constant being defined that uses escaped characters.
const char myName[] = "Bj\370rk\n";

[Link] DIFFERENCES
All compilers have used this character set.

[Link] MIGRATION TO THE CCI

No action required.

2.4.2 The Prototype for main

The prototype for the main() function is:
int main(void);

[Link] EXAMPLE
The following shows an example of how main() might be defined:
int main(void)
{
while(1)
process();
}

[Link] DIFFERENCES
The 8-bit compilers used a void return type for this function.

[Link] MIGRATION TO THE CCI

Each program has one definition for the main() function. Confirm the return type for
main() in all projects previously compiled for 8-bit targets.

2.4.3 Header File Specification

Header file specifications that use directory separators do not conform to the CCI.

[Link] EXAMPLE
The following example shows two conforming include directives.
#include <usb_main.h>
#include "global.h"

 2012-2013 Microchip Technology Inc. DS50002053D-page 19

MPLAB® XC8 C Compiler User’s Guide
[Link] DIFFERENCES
Header file specifications that use directory separators have been allowed in previous
versions of all compilers. Compatibility problems arose when Windows-style separa-
tors “\” were used and the code was compiled under other host operating systems.
Under the CCI, no directory separators should be used.

[Link] MIGRATION TO THE CCI

Any #include directives that use directory separators in the header file specifications
should be changed. Remove all but the header file name in the directive. Add the direc-
tory path to the compiler’s include search path or MPLAB IDE equivalent. This will force
the compiler to search the directories specified with this option.
For example, the following code:
#include <inc/lcd.h>
should be changed to:
#include <lcd.h>
and the path to the inc directory added to the compiler’s header search path in your
MPLAB IDE project properties, or on the command-line as follows:
-Ilcd

2.4.4 Include Search Paths

When you include a header file under the CCI, the file should be discoverable in the
paths searched by the compiler that are detailed below.
Header files specified in angle bracket delimiters < > should be discoverable in the
search paths that are specified by -I options (or the equivalent MPLAB IDE option), or
in the standard compiler include directories. The -I options are searched in the
order in which they are specified.
Header files specified in quote characters " " should be discoverable in the current
working directory or in the same directories that are searched when the header files are
specified in angle bracket delimiters (as above). In the case of an MPLAB X project, the
current working directory is the directory in which the C source file is located. If unsuc-
cessful, the search paths should be to the same directories searched when the header
file is specified in angle bracket delimiters.
Any other options to specify search paths for header files do not conform to the CCI.

[Link] EXAMPLE
If including a header file, as in the following directive:
#include "myGlobals.h"
The header file should be locatable in the current working directory, or the paths spec-
ified by any -I options, or the standard compiler directories. A header file being located
elsewhere does not conform to the CCI.

[Link] DIFFERENCES
The compiler operation under the CCI is not changed. This is purely a coding guideline.

[Link] MIGRATION TO THE CCI

Remove any option that specifies header file search paths other than the -I option (or
the equivalent MPLAB IDE option), and use the -I option in place of this. Ensure the
header file can be found in the directories specified in this section.

DS50002053D-page 20  2012-2013 Microchip Technology Inc.

 Common C Interface

2.4.5 The Number of Significant Initial Characters in an Identifier

At least the first 255 characters in an identifier (internal and external) are significant.
This extends upon the requirement of the ANSI C Standard that states a lower number
of significant characters are used to identify an object.

[Link] EXAMPLE
The following example shows two poorly named variables, but names which are
considered unique under the CCI.
int stateOfPortBWhenTheOperatorHasSelectedAutomaticModeAndMotorIsRunningFast;
int stateOfPortBWhenTheOperatorHasSelectedAutomaticModeAndMotorIsRunningSlow;

[Link] DIFFERENCES
Former 8-bit compilers used 31 significant characters by default, but an option allowed
this to be extended.
The 16- and 32-bit compilers did not impose a limit on the number of significant
characters.

[Link] MIGRATION TO THE CCI

No action required. You can take advantage of the less restrictive naming scheme.

2.4.6 Sizes of Types

The sizes of the basic C types, e.g., char, int and long, are not fully defined by the
CCI. These types, by design, reflect the size of registers and other architectural fea-
tures in the target device. They allow the device to efficiently access objects of this type.
The ANSI C Standard does, however, indicate minimum requirements for these types,
as specified in <limits.h>.
If you need fixed-size types in your project, use the types defined in <stdint.h>, e.g.,
uint8_t or int16_t. These types are consistently defined across all XC compilers,
even outside of the CCI.
Essentially, the C language offers a choice of two groups of types: those that offer sizes
and formats that are tailored to the device you are using, or those that have a fixed size,
regardless of the target.

[Link] EXAMPLE
The following example shows the definition of a variable, native, whose size will allow
efficient access on the target device; and a variable, fixed, whose size is clearly indi-
cated and remains fixed, even though it may not allow efficient access on every device.
int native;
int16_t fixed;

[Link] DIFFERENCES
This is consistent with previous types implemented by the compiler.

[Link] MIGRATION TO THE CCI

If you require a C type that has a fixed size, regardless of the target device, use one of
the types defined by <stdint.h>.

 2012-2013 Microchip Technology Inc. DS50002053D-page 21

MPLAB® XC8 C Compiler User’s Guide

2.4.7 Plain char Types

The type of a plain char is unsigned char. It is generally recommended that all
definitions for the char type explicitly state the signedness of the object.

[Link] EXAMPLE
The following example
char foobar;
defines an unsigned char object called foobar.

[Link] DIFFERENCES
The 8-bit compilers have always treated plain char as an unsigned type.
The 16- and 32-bit compilers used signed char as the default plain char type. The
-funsigned-char option on those compilers changed the default type to be
unsigned char.

[Link] MIGRATION TO THE CCI

Any definition of an object defined as a plain char and using the 16- or 32-bit compilers
needs review. Any plain char that was intended to be a signed quantity should be
replaced with an explicit definition, for example.
signed char foobar;
You can use the -funsigned-char option on MPLAB XC16 and XC32 to change the
type of plain char, but since this option is not supported on MPLAB XC8, the code is
not strictly conforming.

2.4.8 Signed Integer Representation

The value of a signed integer is determined by taking the two’s complement of the
integer.

[Link] EXAMPLE
The following shows a variable, test, that is assigned the value -28 decimal.
signed char test = 0xE4;

[Link] DIFFERENCES
All compilers have represented signed integers in the way described in this section.

[Link] MIGRATION TO THE CCI

No action required.

2.4.9 Integer Conversion

When converting an integer type to a signed integer of insufficient size, the original
value is truncated from the most-significant bit to accommodate the target size.

[Link] EXAMPLE
The following shows an assignment of a value that is truncated.
signed char destination;
unsigned int source = 0x12FE;
destination = source;
Under the CCI, the value of destination after the alignment is -2 (i.e., the bit pattern
0xFE).

DS50002053D-page 22  2012-2013 Microchip Technology Inc.

 Common C Interface
[Link] DIFFERENCES
All compilers have performed integer conversion in an identical fashion to that
described in this section.

[Link] MIGRATION TO THE CCI

No action required.

2.4.10 Bitwise Operations on Signed Values

Bitwise operations on signed values act on the two’s complement representation,
including the sign bit. See also Section 2.4.11 “Right-shifting Signed Values”.

[Link] EXAMPLE
The following shows an example of a negative quantity involved in a bitwise AND
operation.
signed char output, input = -13;
output = input & 0x7E;
Under the CCI, the value of output after the assignment is 0x72.

[Link] DIFFERENCES
All compilers have performed bitwise operations in an identical fashion to that
described in this section.

[Link] MIGRATION TO THE CCI

No action required.

2.4.11 Right-shifting Signed Values

Right-shifting a signed value will involve sign extension. This will preserve the sign of
the original value.

[Link] EXAMPLE
The following shows an example of a negative quantity involved in a bitwise AND
operation.
signed char input, output = -13;
output = input >> 3;
Under the CCI, the value of output after the assignment is -2 (i.e., the bit pattern
0xFE).

[Link] DIFFERENCES
All compilers have performed right-shifting as described in this section.

[Link] MIGRATION TO THE CCI

No action required.

 2012-2013 Microchip Technology Inc. DS50002053D-page 23

MPLAB® XC8 C Compiler User’s Guide

2.4.12 Conversion of Union Member Accessed Using Member With

Different Type
If a union defines several members of different types and you use one member identi-
fier to try to access the contents of another (whether any conversion is applied to the
result) is implementation-defined behavior in the standard. In the CCI, no conversion is
applied and the bytes of the union object are interpreted as an object of the type of the
member being accessed, without regard for alignment or other possible invalid
conditions.

[Link] EXAMPLE
The following shows an example of a union defining several members.
union {
signed char code;
unsigned int data;
float offset;
} foobar;
Code that attempts to extract offset by reading data is not guaranteed to read the
correct value.
float result;
result = [Link];

[Link] DIFFERENCES
All compilers have not converted union members accessed via other members.

[Link] MIGRATION TO THE CCI

No action required.

2.4.13 Default Bit-field int Type

The type of a bit-field specified as a plain int is identical to that of one defined using
unsigned int. This is quite different from other objects where the types int, signed
and signed int are synonymous. It is recommended that the signedness of the
bit-field be explicitly stated in all bit-field definitions.

[Link] EXAMPLE
The following shows an example of a structure tag containing bit-fields that are
unsigned integers and with the size specified.
struct OUTPUTS {
int direction :1;
int parity :3;
int value :4;
};

[Link] DIFFERENCES
The 8-bit compilers have previously issued a warning if type int was used for bit-fields,
but would implement the bit-field with an unsigned int type.
The 16- and 32-bit compilers have implemented bit-fields defined using int as having
a signed int type, unless the option -funsigned-bitfields was specified.

DS50002053D-page 24  2012-2013 Microchip Technology Inc.

 Common C Interface
[Link] MIGRATION TO THE CCI
Any code that defines a bit-field with the plain int type should be reviewed. If the inten-
tion was for these to be signed quantities, then the type of these should be changed to
signed int. For example, in the following example:
struct WAYPT {
int log :3;
int direction :4;
};
the bit-field type should be changed to signed int, as in:
struct WAYPT {
signed int log :3;
signed int direction :4;
};

2.4.14 Bit-fields Straddling a Storage Unit Boundary

The standard indicates that implementations can determine whether bit-fields cross a
storage unit boundary. In the CCI, bit-fields do not straddle a storage unit boundary; a
new storage unit is allocated to the structure, and padding bits fill the gap.
Note that the size of a storage unit differs with each compiler, as this is based on the
size of the base data type (e.g., int) from which the bit-field type is derived. On 8-bit
compilers this unit is 8-bits in size; for 16-bit compilers, it is 16 bits; and for 32-bit
compilers, it is 32 bits in size.

[Link] EXAMPLE
The following shows a structure containing bit-fields being defined.
struct {
unsigned first : 6;
unsigned second :6;
} order;
Under the CCI and using MPLAB XC8, the storage allocation unit is byte sized. The
bit-field, second, is allocated a new storage unit since there are only 2 bits remaining
in the first storage unit in which first is allocated. The size of this structure, order,
is 2 bytes.

[Link] DIFFERENCES
This allocation is identical with that used by all previous compilers.

[Link] MIGRATION TO THE CCI

No action required.

2.4.15 The Allocation Order of Bit-fields

The memory ordering of bit-fields into their storage unit is not specified by the ANSI C
Standard. In the CCI, the first bit defined is the least significant bit of the storage unit in
which it is allocated.

 2012-2013 Microchip Technology Inc. DS50002053D-page 25

MPLAB® XC8 C Compiler User’s Guide
[Link] EXAMPLE
The following shows a structure containing bit-fields being defined.
struct {
unsigned lo : 1;
unsigned mid :6;
unsigned hi : 1;
} foo;
The bit-field lo is assigned the least significant bit of the storage unit assigned to the
structure foo. The bit-field mid is assigned the next 6 least significant bits; and hi, the
most significant bit of that same storage unit byte.

[Link] DIFFERENCES
This is identical with the previous operation of all compilers.

[Link] MIGRATION TO THE CCI

No action required.

2.4.16 The NULL Macro

The NULL macro is defined in <stddef.h>; however, its definition is
implementation-defined behavior. Under the CCI, the definition of NULL is the expres-
sion (0).

[Link] EXAMPLE
The following shows a pointer being assigned a null pointer constant via the NULL
macro.
int * ip = NULL;
The value of NULL, (0), is implicitly converted to the destination type.

[Link] DIFFERENCES
The 32-bit compilers previously assigned NULL the expression ((void *)0).

[Link] MIGRATION TO THE CCI

No action required.

2.4.17 Floating-point Sizes

Under the CCI, floating-point types must not be smaller than 32 bits in size.

[Link] EXAMPLE
The following shows the definition for outY, which is at least 32-bit in size.
float outY;

[Link] DIFFERENCES
The 8-bit compilers have allowed the use of 24-bit float and double types.

[Link] MIGRATION TO THE CCI

When using 8-bit compilers, the float and double type will automatically be made
32 bits in size once the CCI mode is enabled. Review any source code that may have
assumed a float or double type and may have been 24 bits in size.
No migration is required for other compilers.

DS50002053D-page 26  2012-2013 Microchip Technology Inc.

 Common C Interface

2.5 ANSI STANDARD EXTENSIONS

The following topics describe how the CCI provides device-specific extensions to the
standard.

2.5.1 Generic Header File

A single header file <xc.h> must be used to declare all compiler- and device-specific
types and SFRs. You must include this file into every module to conform with the CCI.
Some CCI definitions depend on this header being seen.

[Link] EXAMPLE
The following shows this header file being included, thus allowing conformance with the
CCI, as well as allowing access to SFRs.
#include <xc.h>

[Link] DIFFERENCES
Some 8-bit compilers used <htc.h> as the equivalent header. Previous versions of
the 16- and 32-bit compilers used a variety of headers to do the same job.

[Link] MIGRATION TO THE CCI

Change:
#include <htc.h>
previously used in 8-bit compiler code, or family-specific header files, e.g., from:
#include <p32xxxx.h>
#include <p30fxxxx.h>
#include <p33Fxxxx.h>
#include <p24Fxxxx.h>
#include "p30f6014.h"
to:
#include <xc.h>

2.5.2 Absolute Addressing

Variables and functions can be placed at an absolute address by using the __at()
construct. Stack-based (auto and parameter) variables cannot use the __at()
specifier.

[Link] EXAMPLE
The following shows two variables and a function being made absolute.
int scanMode __at(0x200);
const char keys[] __at(123) = { ’r’, ’s’, ’u’, ’d’};

int modify(int x) __at(0x1000) {

return x * 2 + 3;
}

[Link] DIFFERENCES
The 8-bit compilers have used an @ symbol to specify an absolute address.
The 16- and 32-bit compilers have used the address attribute to specify an object’s
address.

[Link] MIGRATION TO THE CCI

Avoid making objects and functions absolute if possible.

 2012-2013 Microchip Technology Inc. DS50002053D-page 27

MPLAB® XC8 C Compiler User’s Guide
In MPLAB XC8, change absolute object definitions, e.g., from:
int scanMode @ 0x200;
to:
int scanMode __at(0x200);
In MPLAB XC16 and XC32, change code, e.g., from:
int scanMode __attribute__(address(0x200)));
to:
int scanMode __at(0x200);

[Link] CAVEATS
If the __at() and __section() specifiers are both applied to an object when using
MPLAB XC8, the __section() specifier is currently ignored.

2.5.3 Far Objects and Functions

The __far qualifier can be used to indicate that variables or functions are located in
‘far memory’. Exactly what constitutes far memory is dependent on the target device,
but it is typically memory that requires more complex code to access. Expressions
involving far-qualified objects usually generate slower and larger code.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this qualifier.
Some devices may not have such memory implemented; in which case, use of this
qualifier is ignored. Stack-based (auto and parameter) variables cannot use the
__far specifier.

[Link] EXAMPLE
The following shows a variable and function qualified using __far.
__far int serialNo;
__far int ext_getCond(int selector);

[Link] DIFFERENCES
The 8-bit compilers have used the qualifier far to indicate this meaning. Functions
could not be qualified as far.
The 16-bit compilers have used the far attribute with both variables and functions.
The 32-bit compilers have used the far attribute with functions, only.

DS50002053D-page 28  2012-2013 Microchip Technology Inc.

 Common C Interface
[Link] MIGRATION TO THE CCI
For 8-bit compilers, change any occurrence of the far qualifier, e.g., from:
far char template[20];
to:
__far, i.e., __far char template[20];
In the 16- and 32-bit compilers, change any occurrence of the far attribute, e.g., from:
void bar(void) __attribute__ ((far));
int tblIdx __attribute__ ((far));
to:
void __far bar(void);
int __far tblIdx;

[Link] CAVEATS
None.

2.5.4 Near Objects

The __near qualifier can be used to indicate that variables or functions are located in
‘near memory’. Exactly what constitutes near memory is dependent on the target
device, but it is typically memory that can be accessed with less complex code. Expres-
sions involving near-qualified objects generally are faster and result in smaller code.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this qualifier.
Some devices may not have such memory implemented; in which case, use of this
qualifier is ignored. Stack-based (auto and parameter) variables cannot use the
__near specifier.

[Link] EXAMPLE
The following shows a variable and function qualified using __near.
__near int serialNo;
__near int ext_getCond(int selector);

[Link] DIFFERENCES
The 8-bit compilers have used the qualifier near to indicate this meaning. Functions
could not be qualified as near.
The 16-bit compilers have used the near attribute with both variables and functions.
The 32-bit compilers have used the near attribute for functions, only.

 2012-2013 Microchip Technology Inc. DS50002053D-page 29

MPLAB® XC8 C Compiler User’s Guide
[Link] MIGRATION TO THE CCI
For 8-bit compilers, change any occurrence of the near qualifier, e.g., from:
near char template[20];
to:
__near, i.e., __near char template[20];
In 16- and 32-bit compilers, change any occurrence of the near attribute, e.g., from:
void bar(void) __attribute__ ((near));
int tblIdx __attribute__ ((near));
to:
void __near bar(void);
int __near tblIdx;

[Link] CAVEATS
None.

2.5.5 Persistent Objects

The __persistent qualifier can be used to indicate that variables should not be
cleared by the runtime startup code.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this qualifier.

[Link] EXAMPLE
The following shows a variable qualified using __persistent.
__persistent int serialNo;

[Link] DIFFERENCES
The 8-bit compilers have used the qualifier, persistent, to indicate this meaning.
The 16- and 32-bit compilers have used the persistent attribute with variables to
indicate they were not to be cleared.

[Link] MIGRATION TO THE CCI

With 8-bit compilers, change any occurrence of the persistent qualifier, e.g., from:
persistent char template[20];
to:
__persistent, i.e., __persistent char template[20];
For the 16- and 32-bit compilers, change any occurrence of the persistent attribute,
e.g., from:
int tblIdx __attribute__ ((persistent));
to:
int __persistent tblIdx;

[Link] CAVEATS
None.

DS50002053D-page 30  2012-2013 Microchip Technology Inc.

 Common C Interface

2.5.6 X and Y Data Objects

The __xdata and __ydata qualifiers can be used to indicate that variables are
located in special memory regions. Exactly what constitutes X and Y memory is depen-
dent on the target device, but it is typically memory that can be accessed independently
on separate buses. Such memory is often required for some DSP instructions.
Use the native keywords discussed in the Differences section to look up information on
the semantics of these qualifiers.
Some devices may not have such memory implemented; in which case, use of these
qualifiers is ignored.

[Link] EXAMPLE
The following shows a variable qualified using __xdata, as well as another variable
qualified with __ydata.
__xdata char data[16];
__ydata char coeffs[4];

[Link] DIFFERENCES
The 16-bit compilers have used the xmemory and ymemory space attribute with
variables.
Equivalent specifiers have never been defined for any other compiler.

[Link] MIGRATION TO THE CCI

For 16-bit compilers, change any occurrence of the space attributes xmemory or
ymemory, e.g., from:
char __attribute__((space(xmemory)))template[20];
to:
__xdata, or __ydata, i.e., __xdata char template[20];

[Link] CAVEATS
None.

2.5.7 Banked Data Objects

The __bank(num) qualifier can be used to indicate that variables are located in a par-
ticular data memory bank. The number, num, represents the bank number. Exactly what
constitutes banked memory is dependent on the target device, but it is typically a sub-
division of data memory to allow for assembly instructions with a limited address width
field.
Use the native keywords discussed in the Differences section to look up information on
the semantics of these qualifiers.
Some devices may not have banked data memory implemented; in which case, use of
this qualifier is ignored. The number of data banks implemented will vary from one
device to another.

[Link] EXAMPLE
The following shows a variable qualified using __bank().
__bank(0) char start;
__bank(5) char stop;

 2012-2013 Microchip Technology Inc. DS50002053D-page 31

MPLAB® XC8 C Compiler User’s Guide
[Link] DIFFERENCES
The 8-bit compilers have used the four qualifiers bank0, bank1, bank2 and bank3 to
indicate the same, albeit more limited, memory placement.
Equivalent specifiers have never been defined for any other compiler.

[Link] MIGRATION TO THE CCI

For 8-bit compilers, change any occurrence of the bankx qualifiers, e.g., from:
bank2 int logEntry;
to:
__bank(), i.e., __bank(2) int logEntry;

[Link] CAVEATS
This feature is not yet implemented in MPLAB XC8.

2.5.8 Alignment of Objects

The __align(alignment) specifier can be used to indicate that variables must be
aligned on a memory address that is a multiple of the alignment specified. The align-
ment term must be a power of 2. Positive values request that the object’s start address
be aligned; negative values imply the object’s end address be aligned.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this specifier.

[Link] EXAMPLE
The following shows variables qualified using __align() to ensure they end on an
address that is a multiple of 8, and start on an address that is a multiple of 2,
respectively.
__align(-8) int spacer;
__align(2) char coeffs[6];

[Link] DIFFERENCES
An alignment feature has never been implemented on 8-bit compilers.
The 16- and 32-bit compilers used the aligned attribute with variables.

[Link] MIGRATION TO THE CCI

For 16- and 32-bit compilers, change any occurrence of the aligned attribute, e.g.,
from:
char __attribute__((aligned(4)))mode;
to:
__align, i.e., __align(4) char mode;

[Link] CAVEATS
This feature is not yet implemented on MPLAB XC8.

DS50002053D-page 32  2012-2013 Microchip Technology Inc.

 Common C Interface

2.5.9 EEPROM Objects

The __eeprom qualifier can be used to indicate that variables should be positioned in
EEPROM.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this qualifier.
Some devices may not implement EEPROM. Use of this qualifier for such devices
generates a warning. Stack-based (auto and parameter) variables cannot use the
__eeprom specifier.

[Link] EXAMPLE
The following shows a variable qualified using __eeprom.
__eeprom int serialNos[4];

[Link] DIFFERENCES
The 8-bit compilers have used the qualifier, eeprom, to indicate this meaning for some
devices.
The 16-bit compilers have used the space attribute to allocate variables to the memory
space used for EEPROM.

[Link] MIGRATION TO THE CCI

For 8-bit compilers, change any occurrence of the eeprom qualifier, e.g., from:
eeprom char title[20];
to:
__eeprom, i.e., __eeprom char title[20];
For 16-bit compilers, change any occurrence of the eedata space attribute, e.g.,
from:
int mainSw __attribute__ ((space(eedata)));
to:
int __eeprom mainSw;

[Link] CAVEATS
MPLAB XC8 does not implement the __eeprom qualifiers for any PIC18 devices; this
qualifier works as expected for other 8-bit devices.

 2012-2013 Microchip Technology Inc. DS50002053D-page 33

MPLAB® XC8 C Compiler User’s Guide

2.5.10 Interrupt Functions

The __interrupt(type) specifier can be used to indicate that a function is to act as
an interrupt service routine. The type is a comma-separated list of keywords that indi-
cate information about the interrupt function.
The current interrupt types are:
<empty>
Implement the default interrupt function
low_priority
The interrupt function corresponds to the low priority interrupt source
(MPLAB XC8 – PIC18 only)
high_priority
The interrupt function corresponds to the high priority interrupt source (MPLAB XC8)
save(symbol-list)
Save on entry and restore on exit the listed symbols (XC16)
irq(irqid)
Specify the interrupt vector associated with this interrupt (XC16)
altirq(altirqid)
Specify the alternate interrupt vector associated with this interrupt (XC16)
preprologue(asm)
Specify assembly code to be executed before any compiler-generated interrupt code
(XC16)
shadow
Allow the ISR to utilize the shadow registers for context switching (XC16)
auto_psv
The ISR will set the PSVPAG register and restore it on exit (XC16)
no_auto_psv
The ISR will not set the PSVPAG register (XC16)
Use the native keywords discussed in the Differences section to look up information on
the semantics of this specifier.
Some devices may not implement interrupts. Use of this qualifier for such devices
generates a warning. If the argument to the __interrupt specifier does not make
sense for the target device, a warning or error is issued by the compiler.

[Link] EXAMPLE
The following shows a function qualified using __interrupt.
__interrupt(low_priority) void getData(void) {
if (TMR0IE && TMR0IF) {
TMR0IF=0;
++tick_count;
}
}

DS50002053D-page 34  2012-2013 Microchip Technology Inc.

 Common C Interface
[Link] DIFFERENCES
The 8-bit compilers have used the interrupt and low_priority qualifiers to
indicate this meaning for some devices. Interrupt routines were, by default, high priority.
The 16- and 32-bit compilers have used the interrupt attribute to define interrupt
functions.

[Link] MIGRATION TO THE CCI

For 8-bit compilers, change any occurrence of the interrupt qualifier, e.g., from:
void interrupt myIsr(void)
void interrupt low_priority myLoIsr(void)
to the following, respectively:
void __interrupt(high_priority) myIsr(void)
void __interrupt(low_priority) myLoIsr(void)
For 16-bit compilers, change any occurrence of the interrupt attribute, e.g., from:
void __attribute__((interrupt,auto_psv,(irq(52)))) myIsr(void);
to:
void __interrupt(auto_psv,(irq(52)))) myIsr(void);
For 32-bit compilers, the __interrupt() keyword takes two parameters, the vector
number and the (optional) IPL value. Change code that uses the interrupt attribute,
similar to these examples:
void __attribute__((vector(0), interrupt(IPL7AUTO), nomips16))
myisr0_7A(void) {}

void __attribute__((vector(1), interrupt(IPL6SRS), nomips16))

myisr1_6SRS(void) {}

/* Determine IPL and context-saving mode at runtime */

void __attribute__((vector(2), interrupt(), nomips16))
myisr2_RUNTIME(void) {}
to:
void __interrupt(0,IPL7AUTO) myisr0_7A(void) {}

void __interrupt(1,IPL6SRS) myisr1_6SRS(void) {}

/* Determine IPL and context-saving mode at runtime */

void __interrupt(2) myisr2_RUNTIME(void) {}

[Link] CAVEATS
None.

 2012-2013 Microchip Technology Inc. DS50002053D-page 35

MPLAB® XC8 C Compiler User’s Guide

2.5.11 Packing Objects

The __pack specifier can be used to indicate that structures should not use memory
gaps to align structure members, or that individual structure members should not be
aligned.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this specifier.
Some compilers cannot pad structures with alignment gaps for some devices, and use
of this specifier for such devices is ignored.

[Link] EXAMPLE
The following shows a structure qualified using __pack, as well as a structure where
one member has been explicitly packed.
__pack struct DATAPOINT {
unsigned char type;
int value;
} x-point;
struct LINETYPE {
unsigned char type;
__pack int start;
long total;
} line;

[Link] DIFFERENCES
The __pack specifier is a new CCI specifier that is available with MPLAB XC8. This
specifier has no apparent effect since the device memory is byte addressable for all
data objects.
The 16- and 32-bit compilers have used the packed attribute to indicate that a
structure member was not aligned with a memory gap.

[Link] MIGRATION TO THE CCI

No migration is required for MPLAB XC8.
For 16- and 32-bit compilers, change any occurrence of the packed attribute, e.g.,
from:
struct DOT
{
char a;
int x[2] __attribute__ ((packed));
};
to:
struct DOT
{
char a;
__pack int x[2];
};
Alternatively, you can pack the entire structure, if required.

[Link] CAVEATS
None.

DS50002053D-page 36  2012-2013 Microchip Technology Inc.

 Common C Interface

2.5.12 Indicating Antiquated Objects

The __deprecate specifier can be used to indicate that an object has limited longev-
ity and should not be used in new designs. It is commonly used by the compiler vendor
to indicate that compiler extensions or features can become obsolete, or that better fea-
tures have been developed and should be used in preference.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this specifier.

[Link] EXAMPLE
The following shows a function that uses the __deprecate keyword.
void __deprecate getValue(int mode)
{
//...
}

[Link] DIFFERENCES
No deprecate feature was implemented on 8-bit compilers.
The 16- and 32-bit compilers have used the deprecated attribute (note the different
spelling) to indicate that objects should be avoided, if possible.

[Link] MIGRATION TO THE CCI

For 16- and 32-bit compilers, change any occurrence of the deprecated attribute,
e.g., from:
int __attribute__(deprecated) intMask;
to:
int __deprecate intMask;

[Link] CAVEATS
None.

2.5.13 Assigning Objects to Sections

The __section() specifier can be used to indicate that an object should be located
in the named section (or psect, using the MPLAB XC8 terminology). This is typically
used when the object has special and unique linking requirements that cannot be
addressed by existing compiler features.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this specifier.

[Link] EXAMPLE
The following shows a variable which uses the __section keyword.
int __section("comSec") commonFlag;

 2012-2013 Microchip Technology Inc. DS50002053D-page 37

MPLAB® XC8 C Compiler User’s Guide
[Link] DIFFERENCES
The 8-bit compilers have used the #pragma psect directive to redirect objects to a
new section, or psect. The operation of the __section() specifier differs from this
pragma in several ways, as described below.
Unlike with the pragma, the new psect created with __section() does not inherit the
flags of the psect in which the object would normally have been allocated. This means
that the new psect can be linked in any memory area, including any data bank. The
compiler also makes no assumptions about the location of the object in the new sec-
tion. Objects redirected to new psects using the pragma must always be linked in the
same memory area, albeit at any address in that area.
The __section() specifier allows objects that are initialized to be placed in a different
psect. Initialization of the object is still performed, even in the new psect. This requires
the automatic allocation of an additional psect (its name is the same as the new psect,
prefixed with the letter i), that will contain the initial values. The pragma cannot be used
with objects that are initialized.
Objects allocated a different psect with __section()are cleared by the runtime
startup code, unlike objects that use the pragma.
You must reserve memory, and locate via a linker option, for any new psect created with
a __section() specifier in the current MPLAB XC8 compiler implementation.
The 16- and 32-bit compilers have used the section attribute to indicate a different
destination section name. The __section() specifier works in a similar way to the
attribute.

[Link] MIGRATION TO THE CCI

For MPLAB XC8, change any occurrence of the #pragma psect directive, such as:
#pragma psect text%%u=myText
int getMode(int target) {
//...
}
to the __section() specifier, as in:
int __section ("myText") getMode(int target) {
//...
}
For 16- and 32-bit compilers, change any occurrence of the section attribute, e.g.,
from:
int __attribute__((section("myVars"))) intMask;
to:
int __section("myVars") intMask;

[Link] CAVEATS
None.

2.5.14 Specifying Configuration Bits

The #pragma config directive can be used to program the Configuration bits for a
device. The pragma has the form:
#pragma config setting = state|value
where setting is a configuration setting descriptor (e.g., WDT), state is a descriptive
value (e.g., ON) and value is a numerical value.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this directive.

DS50002053D-page 38  2012-2013 Microchip Technology Inc.

 Common C Interface
[Link] EXAMPLE
The following shows Configuration bits being specified using this pragma.
#pragma config WDT=ON, WDTPS = 0x1A

[Link] DIFFERENCES
The 8-bit compilers have used the __CONFIG() macro for some targets that did not
already have support for the #pragma config.
The 16-bit compilers have used a number of macros to specify the configuration set-
tings.
The 32-bit compilers supported the use of #pragma config.

[Link] MIGRATION TO THE CCI

For the 8-bit compilers, change any occurrence of the __CONFIG() macro, e.g.,
__CONFIG(WDTEN & XT & DPROT)
to the #pragma config directive, e.g.,
#pragma config WDTE=ON, FOSC=XT, CPD=ON
No migration is required if the #pragma config was already used.
For the 16-bit compilers, change any occurrence of the _FOSC() or _FBORPOR()
macros attribute, e.g., from:
_FOSC(CSW_FSCM_ON & EC_PLL16);
to:
#pragma config FCKSMEM = CSW_ON_FSCM_ON, FPR = ECIO_PLL16
No migration is required for 32-bit code.

[Link] CAVEATS
None.

2.5.15 Manifest Macros

The CCI defines the general form for macros that manifest the compiler and target
device characteristics. These macros can be used to conditionally compile alternate
source code based on the compiler or the target device.
The macros and macro families are details in Table 2-1.

TABLE 2-1: MANIFEST MACROS DEFINED BY THE CCI

Name Meaning if defined Example
__XC__ Compiled with an MPLAB XC compiler __XC__
__CCI__ Compiler is CCI compliant and CCI enforce- __CCI__
ment is enabled
__XC##__ The specific XC compiler used (## can be 8, __XC8__
16 or 32)
__DEVICEFAMILY__ The family of the selected target device __dsPIC30F__
__DEVICENAME__ The selected target device name __18F452__

 2012-2013 Microchip Technology Inc. DS50002053D-page 39

MPLAB® XC8 C Compiler User’s Guide
[Link] EXAMPLE
The following shows code that is conditionally compiled dependent on the device hav-
ing EEPROM memory.
#ifdef __XC16__
void __interrupt(__auto_psv__) myIsr(void)
#else
void __interrupt(low_priority) myIsr(void)
#endif

[Link] DIFFERENCES
Some of these CCI macros are new (for example __CCI__), and others have different
names to previous symbols with identical meaning (e.g., __18F452 is now
__18F452__).

[Link] MIGRATION TO THE CCI

Any code that uses compiler-defined macros needs review. Old macros will continue to
work as expected, but they are not compliant with the CCI.

[Link] CAVEATS
None.

2.5.16 In-line Assembly

The asm() statement can be used to insert assembly code in-line with C code. The
argument is a C string literal that represents a single assembly instruction. Obviously,
the instructions contained in the argument are device specific.
Use the native keywords discussed in the Differences section to look up information on
the semantics of this statement.

[Link] EXAMPLE
The following shows a MOVLW instruction being inserted in-line.
asm("MOVLW _foobar");

[Link] DIFFERENCES
The 8-bit compilers have used either the asm() or #asm ... #endasm constructs to
insert in-line assembly code.
This is the same syntax used by the 16- and 32-bit compilers.

[Link] MIGRATION TO THE CCI

For 8-bit compilers, change any instance of #asm ... #endasm, so that each instruction
in the #asm block is placed in its own asm() statement, e.g., from:
#asm
MOVLW 20
MOVWF _i
CLRF Ii+1
#endasm
to:
asm("MOVLW20");
asm("MOVWF _i");
asm("CLRFIi+1");
No migration is required for the 16- or 32-bit compilers.

DS50002053D-page 40  2012-2013 Microchip Technology Inc.

 Common C Interface
[Link] CAVEATS
None.

2.6 COMPILER FEATURES

The following item details the compiler options used to control the CCI.

2.6.1 Enabling the CCI

It is assumed that you are using the MPLAB X IDE to build projects that use the CCI.
The widget in the MPLAB X IDE Project Properties to enable CCI conformance is Use
CCI Syntax in the Compiler category.
If you are not using this IDE, then the command-line options are --EXT=cci for
MPLAB XC8 or -mcci for MPLAB XC16 and XC32.

[Link] DIFFERENCES
This option has never been implemented previously.

[Link] MIGRATION TO THE CCI

Enable the option.

[Link] CAVEATS
None.

 2012-2013 Microchip Technology Inc. DS50002053D-page 41

MPLAB® XC8 C Compiler User’s Guide
NOTES:

DS50002053D-page 42  2012-2013 Microchip Technology Inc.

 MPLAB® XC8 C COMPILER
USER’S GUIDE

Chapter 3. How To’s

3.1 INTRODUCTION
This section contains help and references for situations that are frequently encountered
when building projects for Microchip 8-bit devices. Click the links at the beginning of
each section to assist in finding the topic relevant to your question. Some topics are
indexed in multiple sections.
Start here:
• Installing and Activating the Compiler
• Invoking the Compiler
• Writing Source Code
• Getting My Application to Do What I Want
• Understanding the Compilation Process
• Fixing Code That Does Not Work

3.2 INSTALLING AND ACTIVATING THE COMPILER

This section details questions that might arise when installing or activating the compiler.
• How Do I Install and Activate My Compiler?
• How Can I Tell if the Compiler has Activated Successfully?
• Can I Install More Than One Version of the Same Compiler?

3.2.1 How Do I Install and Activate My Compiler?

Installation of the compiler and activation of the license are performed simultaneously
by the XC compiler installer. The guide Installing and Licensing MPLAB XC C Compil-
ers (DS52059) is available on [Link]. It provides details on single-user
and network licenses, as well as how to activate a compiler for evaluation purposes.

3.2.2 How Can I Tell if the Compiler has Activated Successfully?

If you think the compiler cannot have installed correctly or is not working, it is best to
verify its operation outside of MPLAB IDE to isolate possible problems. Try running the
compiler from the command line to check for correct operation. You do not have to
actually compile code.
From your terminal or DOS-prompt, run the compiler driver xc8 (see
Section 4.2 “Invoking the Compiler”) with the option --VER. This option instructs the
compiler to print version information and exit. Under Windows, for example, type the
following line (replacing the path information with a path that is relevant to your instal-
lation).
"C:\Program Files\Microchip\xc8\v1.00\bin\xc8" --ver
The compiler should run, print an informative banner and quit.
The operating mode is printed by the compiler each time you build. Note that if it is not
activated properly, the compiler will continue to operate, but only in the Free mode. If
an error is displayed, or the compiler indicates Free mode, your activation was not
successful.

 2012-2013 Microchip Technology Inc. DS50002053D-page 43

MPLAB® XC8 C Compiler User’s Guide

3.2.3 Can I Install More Than One Version of the Same Compiler?
Yes, the compilers and installation process has been designed to allow you to have
more than one version of the same compiler installed, and you can easily move
between the versions by changing options in MPLAB IDE; see Section 3.3.4 “How
Can I Select Which Compiler I Want to Build With?”.
Compilers should be installed into a directory whose name is related to the compiler
version. This is reflected in the default directory specified by the installer. For example,
the 1.00 and 1.10 MPLAB XC8 compilers would typically be placed in separate
directories.
C:\Program Files\Microchip\xc8\v1.00\
C:\Program Files\Microchip\xc8\v1.10\

DS50002053D-page 44  2012-2013 Microchip Technology Inc.

 How To’s

3.3 INVOKING THE COMPILER

This section discusses how the compiler is run, on the command-line or from the
MPLAB IDE. It includes information about how to get the compiler to do what you want
it to do, in terms of options and the build process itself.
• How Do I Compile From Within MPLAB X IDE?
• How Do I Compile on the Command-line?
• How Do I Compile Using a Make Utility?
• How Can I Select Which Compiler I Want to Build With?
• How Can I Change the Compiler's Operating Mode?
• How Do I Build Libraries?
• How Do I Know What Compiler Options Are Available and What They Do?
• How Do I Know What the Build Options in MPLAB X IDE Do?
• What is Different About an MPLAB X IDE Debug Build?
See, also, the following linked information in other sections.
• What Do I Need to Do When Compiling to Use a Debugger?
• How Do I Use Library Files in My Project?
• How Do I Stop the Compiler From Using Certain Memory Locations?
• What Optimizations Are Employed by the Compiler?

3.3.1 How Do I Compile From Within MPLAB X IDE?

MPLAB X IDE user’s guide and online help provide directions for setting up a project
in the MPLAB X integrated development environment.
If you have one or more MPLAB XC8 compilers installed, you select the compiler you
wish to use in the Configuration category in the Project Properties dialog. The options
for that compiler are then shown in the XC8 Compiler and XC8 Linker categories. Note
that each of these compiler categories have several Option categories.

3.3.2 How Do I Compile on the Command-line?

The compiler driver is called xc8 for all 8-bit PIC devices; e.g., in Windows, it is named
[Link]. This application should be invoked for all aspects of compilation. It is located
in the bin directory of the compiler distribution. Avoid running the individual compiler
applications (such as the assembler or linker) explicitly. You can compile and link in the
one command, even if your project is spread among multiple source files.
The driver is introduced in Section 4.2 “Invoking the Compiler”. See
Section 3.3.4 “How Can I Select Which Compiler I Want to Build With?”, to ensure
you are running the correct driver if you have more than one installed. The com-
mand-line options to the driver are detailed in Section 4.7 “MPLAB XC8 Driver
Options”. The files that can be passed to the driver are listed and described in
Section 4.2.3 “Input File Types”.

3.3.3 How Do I Compile Using a Make Utility?

When compiling using a make utility (such as make), the compilation is usually per-
formed as a two-step process: first generating the intermediate files, then the final com-
pilation and link step to produce one binary output. This is described in
Section 4.3.3 “Multi-Step Compilation”.
The MPLAB XC8 compiler uses a unique technology called OCG that uses an interme-
diate file format that is different than traditional compilers (including XC16 and XC32).
The intermediate file format used by XC8 is a p-code file (.p1 extension), not an object
file. Generating object files as an intermediate file for multi-step compilation defeats
many of the advantages of this technology.

 2012-2013 Microchip Technology Inc. DS50002053D-page 45

MPLAB® XC8 C Compiler User’s Guide

3.3.4 How Can I Select Which Compiler I Want to Build With?

The compilation and installation process has been designed to allow you to have more
than one compiler installed at the same time. You can create a project in MPLAB X IDE
and then build this project with different compilers by simply changing a setting in the
project properties.
To select which compiler is actually used when building a project under MPLAB X IDE,
go to the Project Properties dialog. Select the Configuration category in the Project
Properties dialog (Conf: [default]). A list of MPLAB XC8 compilers is shown in the
Compiler Toolchain, on the far right. Select the compiler that you require.
Once selected, the controls for that compiler are then shown by selecting the MPLAB
XC8 global options, MPLAB XC8 Compiler and MPLAB XC8 Linker categories. These
reveal a pane of options on the right. Note that each category has several panes which
can be selected from a pull-down menu that is near the top of the pane.

3.3.5 How Can I Change the Compiler's Operating Mode?

The compiler’s operating mode (Free, Standard or PRO, see Section 1.2 “Compiler
Description and Documentation”) can be specified as a command line option when
building on the command line; see Section 4.8.37 “--MODE: Choose Compiler Oper-
ating Mode”. If you are building under MPLAB X IDE, there is a Project Properties
selector in the XC8 Compiler category, under the Optimizations option selector; see
Section 4.9.1 “Global Category”.
You can only select modes that your license entitles you to use. The Free mode is
always available; Standard or PRO can be selected if you have purchased a license for
those modes.

3.3.6 How Do I Build Libraries?

Note that XC8 uses a different code generation framework (OCG) that uses additional
library files to those used by traditional compilers (including XC16 and XC32). See
Section 4.3.1 “The Compiler Applications”, for general information on the library
types available and how they fit into the compilation process.
When you have functions and data that are commonly used in applications, you can
either make all the C source and header files available so that other developers can
copy these into their projects. Alternatively you can bundle these source files up into a
library which, along with the accompanying header files, can be linked into a project.
Libraries are more convenient because there are fewer files to deal with. Compiling
code from a library can also be fractionally faster. However, libraries do need to be
maintained. XC8 must use LPP libraries for library routines written in C; the old-style
LIB libraries are used for library routines written in assembly source. It is recommended
that even these libraries be rebuilt if your project is moving to a new compiler version.
Using the compiler driver, libraries can be built by listing all of the files that are to be
included into the library on the command line. None of these files should contain a
main() function, nor settings for Configuration bits or any other such data. Use the
--OUTPUT=lpp option; see Section 4.8.44 “--OUTPUT= type: Specify Output File
Type”, to indicate that a library file is required. For example:
XC8 --chip=16f877a --output=lpp lcd.c utils.c io.c
creates a library file called [Link]. You can specify another name using the -O
option; see Section 4.8.9 “-O: Specify Output File”, or just rename the file.

DS50002053D-page 46  2012-2013 Microchip Technology Inc.

 How To’s
To build a library in MPLAB X IDE, create a regular project.1 Add your source files in
the usual way. Add in the option --OUTPUT=lpp to the Additional Options field in the
MPLAB XC8 Linker category. Click Build. The IDE will issue a warning about the HEX
file being missing, but this can be ignored. The library output can be found in the
dist/default/production folder of the project directory.

3.3.7 How Do I Know What Compiler Options Are Available and What
They Do?
A list of all compiler options can be obtained by using the --HELP option on the com-
mand line; see Section 4.8.33 “--HELP: Display Help”. If you give the --HELP option
an argument, being an option name, it will give specific information on that option, for
example --HELP=runtime.
Alternatively, all options are all listed in Section 4.8 “Option Descriptions” in this
user’s guide. If you are compiling in MPLAB X IDE, see Section 4.9 “MPLAB X
Option Equivalents”.

3.3.8 How Do I Know What the Build Options in MPLAB X IDE Do?
Each of the widgets and controls, in the MPLAB X IDE Project Properties, map directly
to one command-line driver option or suboption, in most instances.
Section 4.8 “Option Descriptions” in this user’s guide lists all command-line driver
options and includes cross references, where appropriate, to corresponding sections
that relate to accessing those options from the IDE. (see Section 4.9 “MPLAB X
Option Equivalents”).

3.3.9 What is Different About an MPLAB X IDE Debug Build?

In MPLAB X, there are distinct build buttons and menu items to build (production) a
project and to debug a project.
While there are many differences between the builds in the IDE – in the XC8 compiler,
there is very little that is different between the two types of build. The main difference
is the setting of a preprocessor macro called __DEBUG, which is assigned 1 when a
performing a debug build. This macro is not defined for production builds.
You can make code in your source conditional on this macro using #ifdef directives,
etc., (see Section 5.14.2 “Preprocessor Directives”); so that you can have your pro-
gram behave differently when you are still in a development cycle. Some compiler
errors are easier to track down after performing a debug build.
In MPLAB X IDE, memory is reserved for your debugger (if selected) only when you
perform a debug build. See Section 3.5.3 “What Do I Need to Do When Compiling
to Use a Debugger?” for more information.

1. At present, the IDE library projects are incompatible with MPLAB XC8.

 2012-2013 Microchip Technology Inc. DS50002053D-page 47

MPLAB® XC8 C Compiler User’s Guide

3.4 WRITING SOURCE CODE

This section presents issues that pertain to the source code you write. It has been
subdivided into the sections listed below.
• C Language Specifics
• Device-Specific Features
• Memory Allocation
• Variables
• Functions
• Interrupts
• Assembly Code

3.4.1 C Language Specifics

This section discusses source code issues that directly relate to the C language itself,
but are commonly asked.
• When Should I Cast Expressions?
• Can Implicit Type Conversions Change the Expected Results of My Expressions?
• How Do I Enter Non-English Characters Into My Program?
• How Can I Use a Variable Defined in Another Source File?

[Link] WHEN SHOULD I CAST EXPRESSIONS?

Expressions can be explicitly case using the cast operator -- a type in round brackets,
e.g., (int). In all cases, conversion of one type to another must be done with caution
and only when absolutely necessary.
Consider the example:
unsigned long l;
unsigned int i;

i = l;
Here, a long type is being assigned to an int type, and the assignment will truncate
the value in l. The compiler will automatically perform a type conversion from the type
of the expression on the right of the assignment operator (long) to the type of the
lvalue on the left of the operator (int).This is called an implicit type conversion. The
compiler typically produces a warning concerning the potential loss of data by the trun-
cation.
A cast to type int is not required and should not be used in the above example if a
long to int conversion was intended. The compiler knows the types of both operands
and performs the conversion accordingly. If you did use a cast, there is the potential for
mistakes if the code is later changed. For example, if you had:
i = (int)l;
the code works the same way; but if, in future, the type of i is changed to a long, for
example, then you must remember to adjust the cast, or remove it, otherwise the con-
tents of l will continue to be truncated by the assignment, which cannot be correct.
Most importantly, the warning issued by the compiler will not be produced if the cast is
in place.

DS50002053D-page 48  2012-2013 Microchip Technology Inc.

 How To’s
Only use a cast in situations where the types used by the compiler are not the types
that you require. For example, consider the result of a division assigned to a floating
point variable:
int i, j;
float fl;

fl = i/j;
In this case, integer division is performed, then the rounded integer result is converted
to a float format. So, if i contained 7 and j contained 2, the division yields 3 and this
is implicitly converted to a float type (3.0) and then assigned to fl. If you wanted the
division to be performed in a float format, then a cast is necessary:
fl = (float)i/j;
(Casting either i or j forces the compiler to encode a floating-point division.) The
result assigned to fl now is 3.5.
An explicit cast can suppress warnings that might otherwise have been produced. This
can also be the source of many problems. The more warnings the compiler produces,
the better chance you have of finding potential bugs in your code.

[Link] CAN IMPLICIT TYPE CONVERSIONS CHANGE THE EXPECTED

RESULTS OF MY EXPRESSIONS?
Yes! The compiler will always use integral promotion and there is no way to disable this;
see Section 5.6.1 “Integral Promotion”. In addition, the types of operands to binary
operators are usually changed so that they have a common type, as specified by the C
Standard. Changing the type of an operand can change the value of the final expres-
sion, so it is very important that you understand the type C Standard conversion rules
that apply when dealing with binary operators. You can manually change the type of an
operand by casting; see Section [Link] “When Should I Cast Expressions?”.

[Link] HOW DO I ENTER NON-ENGLISH CHARACTERS INTO MY PROGRAM?

The ANSI standard and HI-TECH C do not support extended characters set in charac-
ter and string literals in the source character set. See Section 5.4.6 “Constant Types
and Formats”, to see how these characters can be entered using escape sequences.

[Link] HOW CAN I USE A VARIABLE DEFINED IN ANOTHER SOURCE FILE?

Provided the variable defined in the other source file is not static (see
Section [Link].1 “Static Variables”) or auto (see Section [Link] “Auto Variable
Allocation and access”), then adding a declaration for that variable into the current
file will allow you to access it. A declaration consists of the keyword extern in addition
to the type and the name of the variable, as specified in its definition, e.g.,
extern int systemStatus;
This is part of the C language. Your favorite C textbook will give you more information.
The position of the declaration in the current file determines the scope of the variable.
That is, if you place the declaration inside a function, it will limit the scope of the variable
to that function. If you place it outside of a function, it allows access to the variable in
all functions for the remainder of the current file.
Often, declarations are placed in header files and then they are #included into the C
source code; see Section 5.14.2 “Preprocessor Directives”.

 2012-2013 Microchip Technology Inc. DS50002053D-page 49

MPLAB® XC8 C Compiler User’s Guide

3.4.2 Device-Specific Features

This section discusses the code that needs to be written to set up or control a feature
that is specific to Microchip PIC devices.
• How Do I Set the Configuration Bits?
• How Do I Use the PIC Device’s ID Locations?
• How Do I Determine the Cause of Reset on Mid-range Parts?
• How Do I Access SFRs?
• How Do I Stop the Compiler From Using Certain Memory Locations?
See, also, the following linked information in other sections.
• What Do I Need to Do When Compiling to Use a Debugger?

[Link] HOW DO I SET THE CONFIGURATION BITS?

These should be set in your code using either a macro or a pragma. Earlier versions of
MPLAB IDE allowed you to set these bits in a dialog, but MPLAB X IDE requires that
they be specified in your source code. See Section 5.3.5 “Configuration Bit
Access”, for details about how these are set.

[Link] HOW DO I USE THE PIC DEVICE’S ID LOCATIONS?

There is a supplied macro or pragma that allows these values to be programmed; see
Section 5.3.7 “ID Locations”.

[Link] HOW DO I DETERMINE THE CAUSE OF RESET ON MID-RANGE

PARTS?
The TO and PD bits in the STATUS register allow you to determine the cause of a
Reset. However, these bits are quickly overwritten by the runtime startup code that is
executed before main is executed; see Section 5.10.1 “Runtime Startup Code”. You
can have the STATUS register saved into a location that is later accessible from C
code, so that the cause of Reset can be determined by the application after it is running
again; see Section “If the software stack is being used by the program, the stack
pointer (FSR1) is also initialized by the runtime startup code. See
Section [Link].2 “Software Stack Operation”.”.

[Link] HOW DO I ACCESS SFRS?

The compiler ships with header files; see Section 5.3.3 “Device Header Files”, that
define the variables that are mapped over the top of memory-mapped SFRs. Since
these are C variables, they can be used like any other C variables and no new syntax
is required to access these registers.
Bits within SFRs can also be accessed. Individual bit-wide variables are defined that
are mapped over the bits in the SFR. Bit-fields are also available in structures that map
over the SFR as a whole. You can use either in your code; see Section 5.3.6 “Using
SFRs From C Code”.
The name assigned to the variable is usually the same as the name specified in the
device data sheet. See Section [Link] “How Do I Find The Names Used to
Represent SFRs and Bits?”, if these names are not recognized.

DS50002053D-page 50  2012-2013 Microchip Technology Inc.

 How To’s
[Link] HOW DO I FIND THE NAMES USED TO REPRESENT SFRS AND BITS?
Special function registers and the bits within them are accessed via special variables
that map over the register; see Section [Link] “How Do I Access SFRs?”. However,
the names of these variables sometimes differ from those indicated in the data sheet
for the device you are using.
If required, you can examine the <xc.h> header file to find the device-specific header
file that is relevant for your device. This file will define the variables that allow access
to these special variables. However, an easier way to find these variable names is to
look in any of the preprocessed files left behind from a previous compilation. Provided
the corresponding source file included <xc.h>, the preprocessed file will show the
definitions for all the SFR variables and bits for your target device.
If you are compiling under MPLAB X IDE, the preprocessed file(s) are left under the
build/default/production directory of your project for regular builds, or under
build/default/debug for debug builds. They are typically left in the source file
directory if you are compiling on the command line. These files have a .pre extension.

3.4.3 Memory Allocation

Here are questions relating to how your source code affects memory allocation.
• How Do I Position Variables at an Address I Nominate?
• How Do I Position Functions at an Address I Nominate?
• How Do I Place Variables in Program Memory?
• How Do I Stop the Compiler From Using Certain Memory Locations?
See, also, the following linked information in other sections.
• Why Are Some Objects Positioned Into Memory That I Reserved?

[Link] HOW DO I POSITION VARIABLES AT AN ADDRESS I NOMINATE?

The easiest way to do this is to make the variable absolute by using the @ address
construct, see Section 5.5.4 “Absolute Variables”. This means that the address you
specify is used in preference to the variable’s symbol in generated code. Since you
nominate the address, you have full control over where objects are positioned. But, you
must also ensure that absolute variables do not overlap. Variables placed in the middle
of banks can cause havoc with the allocation of other variables and lead to “Can’t find
space” errors; see Section 3.7.6 “How Do I Fix a “Can’t find space...” Error?”. See
also, Section [Link].4 “Changing the Default Auto Variable Allocation”, for infor-
mation on moving auto variables; Section [Link].3 “Changing the Default
Non-Auto Variable Allocation”, for moving non-auto variables; and
Section [Link] “Changing the Default Allocation”, for moving program-space
variables.

[Link] HOW DO I POSITION FUNCTIONS AT AN ADDRESS I NOMINATE?

The easiest way to do this is to make the functions absolute by using the @ address
construct, see Section 5.8.4 “Changing the Default Function Allocation”. This
means that the address you specify is used in preference to the function’s symbol in
generated code. Since you nominate the address, you have full control over where
functions are positioned, but must also ensure that absolute functions do not overlap.
Functions placed in the middle of pages can cause havoc with the allocation of other
functions and lead to "Can’t find space" errors, see Section 3.7.6 “How Do I Fix a
“Can’t find space...” Error?”.

 2012-2013 Microchip Technology Inc. DS50002053D-page 51

MPLAB® XC8 C Compiler User’s Guide
[Link] HOW DO I PLACE VARIABLES IN PROGRAM MEMORY?
The const qualifier implies that the qualified variable is read-only. As a consequence
of this, any variables (except for auto variables or function parameters) that are
qualified const are placed in program memory, thus freeing valuable data RAM. See
Section 5.5.3 “Variables in Program Space”, for more information. Variables that are
qualified const can also be made absolute, so that they can be positioned at an
address you nominate; see Section [Link] “Absolute Objects in Program
Memory”.

[Link] HOW DO I STOP THE COMPILER FROM USING CERTAIN MEMORY

LOCATIONS?
Memory can be reserved when you build. The --RAM and --ROM options allow you to
adjust the ranges of data and program memory, respectively, when you build; see
Section 4.8.49 “--RAM: Adjust RAM Ranges”, and Section 4.8.50 “--ROM: Adjust
ROM Ranges”. By default, all the available on-chip memory is available for use.
However, these options allow you to reserve parts of this memory.

3.4.4 Variables
This sections examines questions that relate to the definition and usage of variables
and types within a program.
• Why Are My Floating-point Results Not Quite What I Am Expecting?
• How Can I Access Individual Bits of a Variable?
• How Long Can I Make My Variable and Macro Names?
See, also, the following linked information in other sections.
• How Do I Share Data Between Interrupt and Main-line Code?
• How Do I Position Variables at an Address I Nominate?
• How Do I Place Variables in Program Memory?
• How Do I Place Variables in the PIC18 Device’s External Program Memory?
• How Can I Rotate a Variable?
• How Do I Utilize All the RAM Banks on My Device?
• How Do I Utilize the Linear Memory on Enhanced Mid-range PIC Devices?
• How Do I Find Out Where Variables and Functions Have Been Positioned?

[Link] WHY ARE MY FLOATING-POINT RESULTS NOT QUITE WHAT I AM

EXPECTING?
First, if you are watching floating-point variables in MPLAB X IDE, make sure that their
type and size agree with the way in which they are defined. For 24-bit floating point vari-
ables (whether they have type float or double), ensure that in MPLAB X IDE the
Display Column Value As popup menu to IEEE Float (24 bit). If the variable is a 32-bit
floating point object, set the types to IEEE Float.
The size of the floating point type can be adjusted for both float and double types;
see Section 4.8.31 “--FLOAT: Select Kind of Float Types”, and
Section 4.8.23 “--DOUBLE: Select Kind of Double Types”.
Since floating-point variables only have a finite number of bits to represent the values
they are assigned, they will hold an approximation of their assigned value; see
Section 5.4.3 “Floating-Point Data Types”. A floating-point variable can only hold
one of a set of discrete real number values. If you attempt to assign a value that is not
in this set, it is rounded to the nearest value. The more bits used by the mantissa in the
floating-point variable, the more values can be exactly represented in the set, and the
average error due to the rounding is reduced.

DS50002053D-page 52  2012-2013 Microchip Technology Inc.

 How To’s
Whenever floating-point arithmetic is performed, rounding also occurs. This can also
lead to results that do not appear to be correct.

[Link] HOW CAN I ACCESS INDIVIDUAL BITS OF A VARIABLE?

There are several ways of doing this. The simplest and most portable way is to define
an integer variable and use macros to read, set, or clear the bits within the variable
using a mask value and logical operations, such as the following.
#define testbit(var, bit) ((var) & (1 <<(bit)))
#define setbit(var, bit) ((var) |= (1 << (bit)))
#define clrbit(var, bit) ((var) &= ~(1 << (bit)))
These, respectively, test to see if bit number, bit, in the integer, var, is set; set the
corresponding bit in var; and clear the corresponding bit in var. Alternatively, a
union of an integer variable and a structure with bit-fields (see
Section [Link] “Bit-Fields in Structures”) can be defined, e.g.,
union both {
unsigned char byte;
struct {
unsigned b0:1, b1:1, b2:1, b3:1, b4:1, b5:1, b6:1, b7:1;
} bitv;
} var;
This allows you to access byte as a whole (using [Link]), or any bit within that
variable independently (using [Link].b0 through [Link].b7).
Note that the compiler does support bit variables (see Section [Link] “Bit Data Types
and Variables”), as well as bit-fields in structures.

[Link] HOW LONG CAN I MAKE MY VARIABLE AND MACRO NAMES?

The C Standard indicates that a only a specific number of initial characters in an iden-
tifier are significant, but it does not actually state what this number is and it varies from
compiler to compiler. For XC8, the first 255 characters are significant, but this can be
reduced using the -N option; see Section 4.8.8 “-N: Identifier Length”. The fewer
characters there are in your variable names, the more portable your code. Using the
-N option allows the compiler to check that your identifiers conform to a specific length.
This option affects variable and function names, as well as preprocessor macro names.
If two identifiers only differ in the non-significant part of the name, they are considered
to represent the same object, which will almost certainly lead to code failure.

3.4.5 Functions
This section examines questions that relate to functions.
• What is the Optimum Size For Functions?
• How Do I Stop An Unused Function Being Removed?
• How Do I Make a Function Inline?
See, also, the following linked information in other sections.
• How Can I Tell How Big a Function Is?
• How Do I Position Functions at an Address I Nominate?
• How Do I Know What Resources Are Being Used by Each Function?
• How Do I Find Out Where Variables and Functions Have Been Positioned?
• How Do I Use Interrupts in C?

 2012-2013 Microchip Technology Inc. DS50002053D-page 53

MPLAB® XC8 C Compiler User’s Guide
[Link] WHAT IS THE OPTIMUM SIZE FOR FUNCTIONS?
Generally speaking, the source code for functions should be kept small, as this aids in
readability and debugging. It is much easier to describe and debug the operation of a
function that performs a small number of tasks. And, they typically have fewer side
effects, which can be the source of coding errors.
In the embedded programming world, a large number of small functions, and the calls
necessary to execute them, can result in excessive memory and stack usage, so a
compromise is often necessary.
The PIC10/12/16 devices use pages in the program memory as the location where the
function code is stored and executed. Although the compiler will allow, and can encode,
functions whose size (the size of the assembly code they generate) exceeds that of a
program memory page, functions of such a size should be avoided and split into
smaller routines where possible. The assembly call and jump sequences to locations
in other pages are much longer than those made to destinations in the same page. If a
function is so large as to cross a page boundary, then loops (or other code constructs
that require jumps within that function) can use the longer form of jump on each
iteration; see Section 5.8.3 “Allocation of Executable Code”.
PIC18 devices are less affected by internal memory paging and the instruction set
allows for calls and jumps to any destination with no penalty. But you should still
endeavor to keep functions as small as possible.
With all devices, the smaller the function, the easier it is for the linker to allocate them
to memory without errors.

[Link] HOW DO I STOP AN UNUSED FUNCTION BEING REMOVED?

If a C function’s symbol is referenced in hand-written assembly code, the function will
never be removed, even if it is not called or never had its address taken in C code.
Create an assembly source file and add this file to your project. You only have to
reference the symbol in this file; so, the file can contain the following
GLOBAL _myFunc
where myFunc is the C name of the function in question (note the leading underscore
in the assembly name, see Section [Link] “Equivalent Assembly Symbols”). This
is sufficient to prevent the function removal optimization from being performed.

[Link] HOW DO I MAKE A FUNCTION INLINE?

You can ask the compiler to inline a function by using the inline specifier (see
Section [Link] “Inline Specifier”) or #pragma inline. This is only a suggestion to
the compiler and cannot always be obeyed. Do not confuse this specifier/pragma with
the intrinsic pragma1 (see Section [Link] “The #pragma Intrinsic Directive”),
which is for functions that have no corresponding source code and which will be
specifically expanded by the code generator during compilation.

1. This specifier was originally named in-line but was changed to avoid confusion.

DS50002053D-page 54  2012-2013 Microchip Technology Inc.

 How To’s

3.4.6 Interrupts
Interrupt and interrupt service routine questions are discussed in this section.
• How Do I Use Interrupts in C?
See, also, the following linked information in other sections.
• How Can I Make My Interrupt Routine Faster?
• How Do I Share Data Between Interrupt and Main-line Code?

[Link] HOW DO I USE INTERRUPTS IN C?

First, be aware of what sort of interrupt hardware is available on your target device.
Baseline PIC devices do not implement interrupts at all; mid-range devices utilize a sin-
gle interrupt vector, and PIC18 devices implement two separate interrupt vector
locations and use a simple priority scheme.
In C source code, a function can be written to act as the interrupt service routine by
using the interrupt qualifier; see Section 5.9.1 “Writing an Interrupt Service
Routine”. Such functions save/restore program context before/after executing the
function body code and a different return instruction is used; see
Section 5.9.4 “Context Switching”. There must be no more than one interrupt func-
tion for each interrupt vector implemented on the target device.
Aside from the interrupt qualifier, the function prototype must specify no parameters
and a void return type. If you wish to implement the low priority interrupt function on
PIC18 devices, use the low_priority keyword as well as the interrupt qualifier.
Code inside the interrupt function can do anything you like, but see
Section 3.6.7 “How Can I Make My Interrupt Routine Faster?” for suggestions to
enhance real-time performance.
Prior to any interrupt occurring, your program must ensure that peripherals are
correctly configured and that interrupts are enabled; see Section 5.9.5 “Enabling
Interrupts”. On PIC18 devices, you must specify the priority of interrupt sources by
writing the appropriate SFRs.

 2012-2013 Microchip Technology Inc. DS50002053D-page 55

MPLAB® XC8 C Compiler User’s Guide

3.4.7 Assembly Code

This section examines questions that arise when writing assembly code as part of a C
project.
• How Should I Combine Assembly and C Code?
• What Do I Need Other than Instructions in an Assembly Source File?
• How Do I Access C Objects from Assembly Code?
• How Can I Access SFRs from Within Assembly Code?
• What Things Must I Manage When Writing Assembly Code?

[Link] HOW SHOULD I COMBINE ASSEMBLY AND C CODE?

Ideally, any hand-written assembly should be written as separate routines that can be
called. This offers some degree of protection from interaction between compiler-gener-
ated and hand-written assembly code. Such code can be placed into a separate
assembly module that can be added to your project; see Section 5.12.1 “Integrating
Assembly Language Modules”.
If necessary, assembly code can be added in-line with C code using either of two meth-
ods; see Section 5.12.2 “#asm, #endasm and asm()”. The code added in-line should
ideally be limited to instructions such as NOP, SLEEP or CLRWDT. Macros are already
provided which in-line all these instructions; see Appendix A. “Library Functions”.
More complex in-line assembly that changes register contents and the device state can
cause code failure if precautions are not taken and should be used with caution. See
Section 5.7 “Register Usage” for those registers used by the compiler.

[Link] WHAT DO I NEED OTHER THAN INSTRUCTIONS IN AN ASSEMBLY

SOURCE FILE?
Assembly code typically needs assembler directives as well as the instructions them-
selves. The operation of all the directives are described in the subsections of
Section 6.4.9 “Assembler Directives”. Common directives required are mentioned
below.
All assembly code must be placed in a psect so it can be manipulated as a whole by
the linker and placed in memory. See Section 5.15.1 “Program Sections” for general
information on psects; see Section [Link] “PSECT” for information on the directive
used to create and specify psects.
The other commonly used directive is GLOBAL, defined in Section [Link] “GLOBAL”
which is used to make symbols accessible across multiple source files.

[Link] HOW DO I ACCESS C OBJECTS FROM ASSEMBLY CODE?

Most C objects are accessible from assembly code. There is a mapping between the
symbols used in the C source and those used in the assembly code generated from
this source. Your assembly should access the assembly-equivalent symbols which are
detailed in Section 5.12.3 “Interaction between Assembly and C Code”.
Instruct the assembler that the symbol is defined elsewhere by using the GLOBAL
assembler directive; see Section [Link] “GLOBAL”. This is the assembly equivalent
of a C declaration, although no type information is present. This directive is not needed
and should not be used if the symbol is defined in the same module as your assembly
code.
Any C variable accessed from assembly code will be treated as if it were qualified
volatile; see Section [Link] “Volatile Type Qualifier”. Specifically specifying the
volatile qualifier in C code is preferred as it makes it clear that external code can
access the object.

DS50002053D-page 56  2012-2013 Microchip Technology Inc.

 How To’s
[Link] HOW CAN I ACCESS SFRS FROM WITHIN ASSEMBLY CODE?
The safest way to gain access to SFRs in assembly code is to have symbols defined
in your assembly code that equate to the corresponding SFR address. Header files are
provided with the compiler so that you do not need to define these yourselves, and they
are detailed in Section [Link] “Accessing Registers from Assembly Code”.
There is no guarantee that you will be able to access symbols generated by the com-
pilation of C code, even code that accesses the SFR you require.

[Link] WHAT THINGS MUST I MANAGE WHEN WRITING ASSEMBLY CODE?

If you are hand-writing assembly code there are several things that you must take con-
trol of.
• Whenever you access a RAM variable, you must ensure that the bank of the vari-
able is selected before you read or write the location. This is done by one or more
assembly instructions. The exact code is based on the device you are using and
the location of the variable. Bank selection is not be required if the object is in
common memory, (which is called the access bank on PIC18 devices) or if you
are using an instruction that takes a full address (such as the MOVFF instruction on
PIC18 devices). Check your device data sheet to see the memory architecture of
your device, and the instructions and registers which control bank selection. Fail-
ure to select the correct bank will lead to code failure.
The BANKSEL pseudo instruction can be used to simplify this process; see
Section [Link] “Bank and Page Selection”.
• You must ensure that the address of the RAM variable you are accessing has
been masked so that only the bank offset is being used as the instruction’s file
register operand. This should not be done if you are using an instruction that takes
a full address (such as the MOVFF instruction on PIC18 devices). Check your
device data sheet to see what address operand instructions requires. Failure to
mask an address can lead to a fixup error (see Section 3.7.8 “How Do I Fix a
Fixup Overflow Error?”) or code failure.
The BANKMASK macro can truncate the address for you; see
Section [Link] “Accessing Registers from Assembly Code”.
• Before you call or jump to any routine, you must ensure that you have selected
the program memory page of this routine using the appropriate instructions. You
can either use the PAGESEL pseudo instruction; see Section [Link] “Bank and
Page Selection”, or the FCALL or LJMP pseudo instructions (not required on
PIC18 devices); see Section [Link] “Long Jumps and Calls” which will
automatically add page selection instructions, if required.
• You must ensure that any RAM used for storage has memory reserved. If you are
only accessing variables defined in C code, then reservation is already done by
the compiler. You must reserve memory for any variables you only use in the
assembly code using an appropriate directive such as DS or DABS; see
Section [Link] “DS” or Section [Link] “DABS”. It is often easier to define
objects in C code rather than in assembly.

 2012-2013 Microchip Technology Inc. DS50002053D-page 57

MPLAB® XC8 C Compiler User’s Guide
• You must place any assembly code you write in a psect (see
Section [Link] “PSECT”, for the directive to do this and
Section 5.15.1 “Program Sections”, for general information about psects). A
psect you define can need flags (options) to be specified. Pay particular note to
the delta, space and class flags (see Section [Link].4 “Delta”, and
Section [Link].15 “Space” and Section [Link].3 “Class”). If these are not set
correctly, compile errors or code failure will almost certainly result. If the psect
specifies a class and you are happy with it being placed anywhere in the memory
range defined by that class (see Section 7.2.1 “-Aclass =low-high,...”), it does
not need any additional options to be linked; otherwise, you will need to link the
psect using a linker option (see Section 7.2.19 “-Pspec” for the usual way to link
psects and Section 4.8.6 “-L-: Adjust Linker Options Directly” which indicates
how you can specify this option without running the linker directly).
Assembly code that is placed in-line with C code will be placed in the same psect
as the compiler-generated assembly and you should not place this into a separate
psect.
• You must ensure that any registers you write to in assembly code are not already
in used by compiler-generated code. If you write assembly in a separate module,
then this is less of an issue as the compiler will, by default, assume that all regis-
ters are used by these routines (see Section 5.7 “Register Usage”). No
assumptions are made for in-line assembly (see Section 5.12.2 “#asm,
#endasm and asm()”) and you must be careful to save and restore any
resources that you use (write) and which are already in use by the surrounding
compiler-generated code.

DS50002053D-page 58  2012-2013 Microchip Technology Inc.

 How To’s

3.5 GETTING MY APPLICATION TO DO WHAT I WANT

This section provides programming techniques, applications and examples. It also
examines questions that relate to making an application perform a specific task.
• What Can Cause Glitches on Output Ports?
• How Do I Link Bootloaders and Downloadable Applications?
• What Do I Need to Do When Compiling to Use a Debugger?
• How Can I Have Code Executed Straight After Reset?
• How Do I Share Data Between Interrupt and Main-line Code?
• How Can I Prevent Misuse of My Code?
• How Do I Use Printf to Send Text to a Peripheral?
• How Do I Calibrate the Oscillator on My Device?
• How Do I Place Variables in the PIC18 Device’s External Program Memory?
• How Can I Implement a Delay in My Code?
• How Can I Rotate a Variable?
• How Can I Stop Variables Being Cleared at Startup?

3.5.1 What Can Cause Glitches on Output Ports?

In most cases, this is caused by using ordinary variables to access port bits or the entire
port itself. These variables should be qualified volatile.
The value stored in a variable mapped over a port (hence the actual value written to
the port) directly translates to an electrical signal. It is vital that the values held by these
variables only change when the code intends them to, and that they change from their
current state to their new value in a single transition. See Section [Link] “Volatile
Type Qualifier”. The compiler attempts to write to volatile variables in one operation.

3.5.2 How Do I Link Bootloaders and Downloadable Applications?

Exactly how this is done depends on the device you are using and your project require-
ments, but the general approach when compiling applications that use a bootloader is
to allocate discrete program memory space to the bootloader and application so they
have their own dedicated memory. In this way the operation of one cannot affect the
other. This will require that either the bootloader or the application is offset in memory.
That is, the Reset and interrupt location are offset from address 0 and all program code
is offset by the same amount.
On PIC18 devices, typically the application code is offset, and the bootloader is linked
with no offset so that it populates the Reset and interrupt code locations. The boot-
loader Reset and interrupt code merely contains code which redirects control to the real
Reset and interrupt code defined by the application and which is offset.
On mid-range devices, this is not normally possible to perform when interrupts are
being used. Consider offsetting all of the bootloader with the exception of the code
associated with Reset, which must always be defined by the bootloader. The applica-
tion code can define the code linked at the interrupt location. The bootloader will need
to remap any application code that attempts to overwrite the Reset code defined by the
bootloader.
The option --CODEOFFSET, (see Section 4.8.21 “--CODEOFFSET: Offset Program
Code to Address”), allows the program code (Reset and vectors included) to be
moved by a specified amount. The option also restricts the program from using any pro-
gram memory from address 0 (Reset vector) to the offset address. Always check the
map file; see Section 7.4.2 “Contents”, to ensure that nothing remains in reserved
areas.

 2012-2013 Microchip Technology Inc. DS50002053D-page 59

MPLAB® XC8 C Compiler User’s Guide
The contents of the HEX file for the bootloader can be merged with the code of the
application by adding the HEX file as a project file, either on the command line, or in
MPLAB IDE. This results in a single HEX file that contains the bootloader and applica-
tion code in the one image. HEX files are merged by the HEXMATE application; see
Section 8.3 “HEXMATE”. Check for warnings from this application about overlap,
which can indicate that memory is in use by both bootloader and the downloadable
application.

3.5.3 What Do I Need to Do When Compiling to Use a Debugger?

You can use debuggers, such as ICD3 or REALICE, to debug code built with the
MPLAB XC8 compiler. These debuggers use some of the data and program memory
of the device for its own use, so it is important that your code does not also use these
resources.
There is a command-line option; see Section 4.8.22 “--DEBUGGER: Select
Debugger Type”, that can be used to tell the compiler which debugger is to be used.
The compiler can then reserve the memory used by the debugger so that your code will
not be located in these locations.
In the MPLAB X IDE, the appropriate debugger option is specified if you perform a
debug build. It will not be specified if you perform a regular Build Project or Clean and
Build.
Since some device memory is being used up by the debugger, there is less available
for your program and it is possible that your code or data might not fit in the device when
a debugger is selected.
Note that which specific memory locations used by the debuggers is an attribute of
MPLAB IDE, not the device. If you move a project to a new version of the IDE, the
resources required can change. For this reason, you should not manually reserve
memory for the debugger, or make any assumptions in your code as to what memory
is used. A summary of the debugger requirements is available in the MPLAB IDE help
files.
To verify that the resources reserved by the compiler match those required by the
debugger, do the following. Compile your code with and without the debugger selected
and keep a copy of the map file produced for both builds. Compare the linker options
in the map files and look for changes in the -A options; see Section 7.2.1 “-Aclass
=low-high,...”. For example, the memory defined for the CODE class with no debugger
might be specified by this option:
-ACODE=00h-0FFh,0100h-07FFh,0800h-0FFFhx3
and with the ICD3 selected as the debugger, it becomes:
-ACODE=00h-0FFh,0100h-07FFh,0800h-0FFFhx2,01800h-01EFFh
This shows that a memory range from 1F00 to 1FFF has been removed by the compiler
and cannot be used by your program. See also Section 3.6.16 “Why Are Some
Objects Positioned Into Memory That I Reserved?”.

3.5.4 How Can I Have Code Executed Straight After Reset?

A special hook has been provided so you can easily add special “powerup” assembly
code which will be linked to the Reset vector; see Section 5.10.2 “The Powerup Rou-
tine”. This code will be executed before the runtime startup code is executed, which in
turn is executed before the main function; see Section 5.10 “Main, Runtime Startup
and Reset”.

DS50002053D-page 60  2012-2013 Microchip Technology Inc.

 How To’s

3.5.5 How Do I Share Data Between Interrupt and Main-line Code?

Variables accessed from both interrupt and main-line code can easily become cor-
rupted or mis-read by the program. The volatile qualifier (see
Section [Link] “Volatile Type Qualifier”) tells the compiler to avoid performing opti-
mizations on such variables. This will fix some of the issues associated with this prob-
lem.
The other issues relates to whether the compiler/device can access the data atomically.
With 8-bit PIC devices, this is rarely the case. An atomic access is one where the entire
variable is accessed in only one instruction. Such access is uninterruptable. You can
determine if a variable is being accessed atomically by looking at the assembly code
the compiler produces in the assembly list file; see Section 6.6 “Assembly List
Files”. If the variable is accessed in one instruction, it is atomic. Since the way vari-
ables are accessed can vary from statement to statement it is usually best to avoid
these issues entirely by disabling interrupts prior to the variable being accessed in
main-line code, then re-enable the interrupts afterwards; see Section 5.9.5 “Enabling
Interrupts”.

3.5.6 How Can I Prevent Misuse of My Code?

First, many devices with flash program memory allow all or part of this memory to be
write protected. The device Configuration bits need to be set correctly for this to take
place; see Section 5.3.5 “Configuration Bit Access” and your device data sheet.
Second, you can prevent third-party code being programmed at unused locations in the
program memory by filling these locations with a value rather than leaving them in their
default unprogrammed state. You can chose a fill value that corresponds to an instruc-
tion or set all the bits so as the values cannot be further modified. (Consider what will
happen if you program somehow reaches and starts executing from these filled values.
What instruction will be executed?)
The compiler’s HEXMATE utility (see Section 8.3 “HEXMATE”) has the capability to
fill unused locations and this operation can be requested using a command-line driver
option; see Section 4.8.30 “--FILL: Fill Unused Program Memory”. As HEXMATE
only works with HEX files, this feature is only available when producing HEX/COF file
outputs (as opposed to binary, for example), which is the default operation.
And last, if you wish to make your library files or intermediate p-code files available to
others but do not want the original source code to be viewable, then you can obfuscate
the files using the --SHROUD option; see Section 4.8.55 “--SHROUD: Obfuscate
P-code Files”

 2012-2013 Microchip Technology Inc. DS50002053D-page 61

MPLAB® XC8 C Compiler User’s Guide

3.5.7 How Do I Use Printf to Send Text to a Peripheral?

The printf function does two things: it formats text based on the format string and
placeholders you specify, and sends (prints) this formatted text to a destination (or
stream); see Appendix A. “Library Functions”. The printf function performs all the
formatting; then it calls a helper function, called putch, to send each byte of the for-
matted text. By customizing the putch function you can have printf send data to any
peripheral or location; see Section 5.11.1 “The printf Routine”. You can choose the
printf output go to an LCD, SPI module or USART, for example.
A stub for the putch function can be found in the compiler’s sources directory. Copy
it into your project then modify it to send the single byte parameter passed to it to the
required destination. Before you can use printf, peripherals that you use will need to
be initialized in the usual way. Here is an example of putch for a USART on a mid-range
device.
void putch(char data) {
while( ! TXIF) // check buffer
continue; // wait till ready
TXREG = data; // send data
}
You can get printf to send to one of several destinations by using a global variable
to indicate your choice. Have the putch function send the byte to one of several des-
tinations based on the contents of this variable.

3.5.8 How Do I Calibrate the Oscillator on My Device?

Some devices allow for calibration of their internal oscillators; see your device data
sheet. The runtime startup code generated by the compiler, (see
Section 5.10.1 “Runtime Startup Code”), will by default provide code that performs
oscillator calibration. This can be disabled, if required, using an option; see
Section 4.8.51 “--RUNTIME: Specify Runtime Environment”.

3.5.9 How Do I Place Variables in the PIC18 Device’s External

Program Memory?
If all you mean to do is place read-only variables in program memory, qualify them as
const; see Section 5.5.3 “Variables in Program Space”. If you intend the variables
to be located in the external program memory then use the far qualifier and specify
the memory using the --RAM option; see Section 4.8.49 “--RAM: Adjust RAM
Ranges”. The compiler will allow far-qualified variables to be modified. Note that the
time to access these variables will be longer than for variables in the internal data mem-
ory. The access mode to external memory can be specified with an option; see
Section 4.8.25 “--EMI: Select External Memory Interface Operating Mode”.

DS50002053D-page 62  2012-2013 Microchip Technology Inc.

 How To’s

3.5.10 How Can I Implement a Delay in My Code?

If an accurate delay is required, or if there are other tasks that can be performed during
the delay, then using a timer to generate an interrupt is the best way to proceed.
If these are not issues in your code, then you can use the compiler’s in-built delay
pseudo-functions: _delay, __delay_ms or __delay_us; see Appendix A. “Library
Functions”. These all expand into in-line assembly instructions or a (nested) loop of
instructions which will consume the specified number of cycles or time. The delay argu-
ment must be a constant and less than approximately 179,200 for PIC18 devices and
approximately 50,659,000 for other devices.
Note that these code sequences will only use the NOP instruction and/or instructions
which form a loop. The alternate PIC18-only versions of these pseudo-functions, e.g.,
_delaywdt, can use the CLRWDT instruction as well. See also, Appendix A. “Library
Functions”.

3.5.11 How Can I Rotate a Variable?

The C language does not have a rotate operator, but rotations can be performed using
the shift and bitwise OR operators. Since the PIC devices have a rotate instruction, the
compiler will look for code expressions that implement rotates (using shifts and ORs)
and use the rotate instruction in the generated output wherever possible; see
Section 5.6.2 “Rotation”.

3.5.12 How Can I Stop Variables Being Cleared at Startup?

Use the persistent qualifier (see Section [Link] “Persistent Type Qualifier”),
which will place the variables in a different psect that is not cleared by the runtime
startup code.

 2012-2013 Microchip Technology Inc. DS50002053D-page 63

MPLAB® XC8 C Compiler User’s Guide

3.6 UNDERSTANDING THE COMPILATION PROCESS

This section tells you how to find out what the compiler did during the build process,
how it encoded output code, where it placed objects, etc. It also discusses the features
that are supported by the compiler.
• What’s the Difference Between the Free, Standard and PRO Modes?
• How Can I Make My Code Smaller?
• How Can I Reduce RAM Usage?
• How Can I Make My Code Faster?
• How Can I Speed Up Programming Times
• How Does the Compiler Place Everything in Memory?
• How Can I Make My Interrupt Routine Faster?
• How Big Can C Variables Be?
• How Do I Utilize All the RAM Banks on My Device?
• How Do I Utilize the Linear Memory on Enhanced Mid-range PIC Devices?
• What Devices are Supported by the Compiler?
• How Do I Know What Code the Compiler Is Producing?
• How Can I Tell How Big a Function Is?
• How Do I Know What Resources Are Being Used by Each Function?
• How Do I Find Out Where Variables and Functions Have Been Positioned?
• Why Are Some Objects Positioned Into Memory That I Reserved?
• How Do I Know How Much Memory Is Still Available?
• How Do I Use Library Files in My Project?
• What Optimizations Are Employed by the Compiler?
• Why Do I Get Out-of-memory Errors When I Select a Debugger?
• How Do I Know Which Stack Model the Compiler Has Assigned to a Function?
• How Do I Know What Value Has Been Programmed in the Configuration Bits or ID
Location?
See, also, the following linked information in other sections.
• How Do I Find Out What an Warning/error Message Means?
• What is Different About an MPLAB X IDE Debug Build?
• How Do I Stop An Unused Function Being Removed?
• How Do I Build Libraries?

3.6.1 What’s the Difference Between the Free, Standard and PRO
Modes?
These modes (see Section 1.2 “Compiler Description and Documentation”)
mainly differ in the optimizations that are performed when compiling. Compilers oper-
ating in Free (formerly called Lite) and Standard mode can compile for all the same
devices as supported by the Pro mode. The code compiled in Free and Standard mode
can use all the available memory for the selected device. What will be different is the
size and speed of the generated compiler output. Free mode output will be much less
efficient when compared to that produced in Standard mode, which in turn will be less
efficient than that produce when in Pro mode.
All these modes use the OCG compiler framework, so the entire C program is compiled
in one step and the source code does not need many non-standard extensions.

DS50002053D-page 64  2012-2013 Microchip Technology Inc.

 How To’s
There are a small number of command-line options disabled in Free mode, but these
do not relate to code features; merely how the compiler can be executed. Most
customers never need to use these options. The options are --GETOPTION (see
Section 4.8.32 “--GETOPTION: Get Command-line Options”) and --SETOPTION
(see Section 4.8.54 “--SETOPTION: Set the Command-line Options for
Application”).

3.6.2 How Can I Make My Code Smaller?

There are a number of ways that this can be done, but results vary from one project to
the next. Use the assembly list file, (see Section 6.6 “Assembly List Files”), to
observe the assembly code produced by the compiler to verify that the following tips
are relevant to your code.
Use the smallest data types possible as less code is needed to access these. (This also
reduces RAM usage.) Note that a bit type and non-standard 24-bit integer type
(short long) exists for this compiler. Avoid multi-bit bit-fields whenever possible. The
code used to access these can be very large. See Section 5.4 “Supported Data
Types and Variables”, for all data types and sizes.
There are two sizes of floating-point type, as well, and these are discussed in the same
section. Avoid floating-point if at all possible. Consider writing fixed-point arithmetic
code.
Use unsigned types, if possible, instead of signed types; particularly if they are used in
expressions with a mix of types and sizes. Try to avoid an operator acting on operands
with mixed sizes whenever possible.
Whenever you have a loop or condition code, use a “strong” stop condition, i.e., the fol-
lowing:
for(i=0; i!=10; i++)
is preferable to:
for(i=0; i<10; i++)
A check for equality (== or !=) is usually more efficient to implement than the weaker
< comparison.
In some situations, using a loop counter that decrements to zero is more efficient than
one that starts at zero and counts up by the same number of iterations. This is more
likely to be the case if the loop index is a byte-wide type. So you might be able to rewrite
the above as:
for(i=10; i!=0; i--)
There might be a small advantage in changing the order of function parameters so that
the first parameter is byte sized. A register is used if the first parameter is byte-sized.
For example consider:
char calc(char mode, int value);
over
char calc(int value, char mode);
Ensure that all optimizations are enabled; see Section 4.8.42 “--OPT: Invoke Com-
piler Optimizations”. Be aware of what optimizations the compiler performs (see
Section 5.13 “Optimizations”, and Section 6.5 “Assembly-Level Optimizations”)
so you can take advantage of them and don’t waste your time manually performing opti-
mizations in C code that the compiler already handles, e.g., don’t turn a multiply-by-4
operation into a shift-by-2 operation as this sort of optimization is already detected.

 2012-2013 Microchip Technology Inc. DS50002053D-page 65

MPLAB® XC8 C Compiler User’s Guide

3.6.3 How Can I Reduce RAM Usage?

Use the smallest data types possible. (This also reduces code size as less code is
needed to access these.) Note that a bit type and non-standard 24-bit integer type
(short long) exists for this compiler. See Section 5.4 “Supported Data Types and
Variables” for all data types and sizes. There are two sizes of floating-point type, as
well, and these are discussed in the same section.
Consider using auto variables over global or static variables as there is the poten-
tial that these can share memory allocated to other auto variables that are not active
at the same time. Memory allocation of auto variables is made on a compiled stack,
described in Section [Link] “Auto Variable Allocation and access”.
Rather than pass large objects to, or from, functions, pass pointers which reference
these objects. This is particularly true when larger structures are being passed, but
there might be RAM savings to be made even when passing long variables.
Objects that do not need to change throughout the program can be located in program
memory using the const qualifier; see Section [Link] “Const Type Qualifier”, and
Section 5.5.3 “Variables in Program Space”. This frees up precious RAM, but slows
execution.
Ensure that all optimizations are enabled; see Section 4.8.42 “--OPT: Invoke Com-
piler Optimizations”. Be aware of which optimizations the compiler performs (see
Section 5.13 “Optimizations”), so that you can take advantage of them and don’t
waste your time manually performing optimizations in C code that the compiler already
handles.

3.6.4 How Can I Make My Code Faster?

To a large degree, smaller code is faster code, so efforts to reduce code size often
decrease execution time; see Section 3.6.2 “How Can I Make My Code Smaller?”.
See also, Section 3.6.7 “How Can I Make My Interrupt Routine Faster?”. However,
there are ways some sequences can be sped up at the expense of increased code size.
One of the compiler optimization settings is for speed (the alternate setting is for
space), so ensure this is selected; see Section 4.8.42 “--OPT: Invoke Compiler
Optimizations”. This will use alternate output in some instances that is faster, but
larger.
Some library multiplication routines operate faster when one of their operands is a
smaller value. See Section 5.3.9 “Hardware Multiply Instructions” for more
information on how to take advantage of this.
Generally, the biggest gains to be made in terms of speed of execution come from the
algorithm used in a project. Identify which sections of your program need to be fast.
Look for loops that might be linearly searching arrays and choose an alternate search
method such as a hash table and function. Where results are being recalculated, con-
sider if they can be cached.

3.6.5 How Can I Speed Up Programming Times

The linker may build up some program output at low addresses in the HEX file, working
their way to higher address; and some output at high address, working their way down
in memory. This does not affect code operation and makes linking easier, but it can pro-
duce a HEX file covering the entire device memory space. Programming this HEX file
into the device may take a long time.
To reduce programming times in this situation, instruct the linker to not use all the
device’s program memory. Use the --ROM option to reserve the upper part of program
memory, see Section 4.8.50 “--ROM: Adjust ROM Ranges”.

DS50002053D-page 66  2012-2013 Microchip Technology Inc.

 How To’s

3.6.6 How Does the Compiler Place Everything in Memory?

In most situations, assembly instructions and directives associated with both code and
data are grouped into sections, called psects, and these are then positioned into con-
tainers which represent the device memory. An introductory explanation into this pro-
cess is given in Section 5.15.1 “Program Sections”. The exception is for absolute
variables (see Section 5.5.4 “Absolute Variables”), which are placed at a specific
address when they are defined and which are not placed in a psect.

3.6.7 How Can I Make My Interrupt Routine Faster?

Consider suggestions made in Section 3.6.2 “How Can I Make My Code Smaller?”
(code size) for any interrupt code. Smaller code is often faster code.
In addition to the code you write in the ISR there is the code the compiler produces to
switch context. This is executed immediately after an interrupt occurs and immediately
before the interrupt returns, so must be included in the time taken to process an inter-
rupt; see Section 5.9.4 “Context Switching”. This code is optimal in that only regis-
ters used in the ISR will be saved by this code. Thus, the less registers used in your
ISR will mean potentially less context switch code to be executed.
Mid-range devices have only a few registers that are used by the compiler, and there
is little context switch code. Even fewer registers are considered for saving when com-
piling for enhanced mid-range device. PIC18 devices will benefit most from the above
suggestion as they use a larger set of registers in generated code; see
Section 5.7 “Register Usage”.
Generally simpler code will require less resources than more complicated expressions.
Use the assembly list file to see which registers are being used by the compiler in the
interrupt code; see Section 6.6 “Assembly List Files”.
Consider having the ISR simply set a flag and return. The flag can then be checked in
main-line code to handle the interrupt. This has the advantage of moving the compli-
cated interrupt-processing code out of the ISR so that it no longer contributes to its reg-
ister usage. Always use the volatile qualifier (see Section [Link] “Volatile Type
Qualifier”for variables shared by the interrupt and main-line code; see
Section 3.5.5 “How Do I Share Data Between Interrupt and Main-line Code?”.

3.6.8 How Big Can C Variables Be?

This question specifically relates to the size of individual C objects, such as arrays or
structures. The total size of all variables is another matter.
To answer this question you need to know in which memory space the variable will be
located. Objects qualified const will be located in program memory; other objects will
be placed in data memory. Program memory object sizes are discussed in
Section [Link] “Size Limitations of Const Variables”. Objects in data memory are
broadly grouped into autos and non-autos and the size limitations of these objects,
respectively, are discussed in Section [Link].2 “Software Stack Operation” and
Section [Link].2 “Non-Auto Variable Size Limits”.

3.6.9 How Do I Utilize All the RAM Banks on My Device?

The compiler will automatically use all the available RAM banks on the device you are
programming. It is only if you wish to alter the default memory allocation that you need
take any action. Special bank qualifiers; see
Section “--RAM=default,+20000-2FFFF.”, and an option (see
Section 4.8.15 “--ADDRQUAL: Set Compiler Response to Memory Qualifiers”) to
indicate how these qualifiers are interpreted are used to manually allocate variables.
Note that there is no guarantee that all the memory on a device can be utilized as data
and code is packed in sections, or psects.

 2012-2013 Microchip Technology Inc. DS50002053D-page 67

MPLAB® XC8 C Compiler User’s Guide

3.6.10 How Do I Utilize the Linear Memory on Enhanced Mid-range PIC

Devices?
The linear addressing mode is a means of accessing the banked data memory as one
contiguous and linear block; see Section 5.5.1 “Address Spaces”. Use of the linear
memory is fully automatic. Objects that are larger than a data bank can be defined in
the usual way and will be accessed using the linear addressing mode; see
Section [Link].2 “Software Stack Operation”, and Section [Link].2 “Non-Auto
Variable Size Limits”. If you define absolute objects at a particular location in memory,
you can use a linear address, if you prefer, or the regular banked address; see
Section [Link] “Absolute Variables in Data Memory”.

3.6.11 What Devices are Supported by the Compiler?

Support for new devices usually takes place with each compiler release. To find
whether a device is supported by your compiler, you can do several things; see also,
Section 5.3.1 “Device Support”.
• HTML listings are provided in the compiler’s docs directory. Open these in your
favorite web browser. They are called pic_chipinfo.html and
pic18_chipinfo.html.
• Run the compiler driver on the command line (see Section 4.2 “Invoking the
Compiler”) with the --CHIPINFO option; see Section 4.8.19 “--CHIPINFO: Dis-
play List of Supported Devices”. A full list of all devices is printed to the screen.

3.6.12 How Do I Know What Code the Compiler Is Producing?

The assembly list file (see Section 6.6 “Assembly List Files”) shows the assembly
output for almost the entire program, including library routines linked in to your pro-
gram, as well a large amount of the runtime startup code; see
Section 5.10.1 “Runtime Startup Code”. The list file is produced by default if you are
using MPLAB IDE. If you are using the command-line, the option --ASMLIST will pro-
duce this file for you; see Section 4.8.16 “--ASMLIST: Generate Assembler List
Files”. The assembly list file will have a .lst extension.
The list file shows assembly instructions, some assembly directives and information
about the program, such as the call graph (see Section 6.6.6 “Call Graph”), pointer
reference graph (see Section 6.6.5 “Pointer Reference Graph”), and information for
every function. Not all assembly directives are shown in the list file if the assembly
optimizers are enabled (they are produced in the intermediate assembly file).
Temporarily disable the assembly optimizers (see Section 4.8.42 “--OPT: Invoke
Compiler Optimizations”), if you wish to see all the assembly directives produced by
the compiler.

3.6.13 How Can I Tell How Big a Function Is?

This size of a function (the amount of assembly code generated for that function) can
be determined from the assembly list file; see Section 6.6 “Assembly List Files”, or
a ‘funclist’ file generated by the compiler. Recent compilers define a symbol whose
assigned value is equal to the size of the function. The symbol has the form
__size_of_func, where func is the name of the function. The units of this symbol
will be the same as the addressability of the program memory for the particular device:
words for PIC10/12/16 and bytes for PIC18. You can also search for the labels that
mark the beginning and end of the function. The function starts at the label _func:,
where func is the name of the function, and ends just prior to the label
__end_of_func. For example, the function main can have associated symbols
__size_of_main, _main and __end_of_main. These will be found in the symbol
table at the end of the assembly list file.

DS50002053D-page 68  2012-2013 Microchip Technology Inc.

 How To’s
The list of functions, memory location, and size is available in a file called funclist.
Each function will have a line similar to the following.
_main: CODE, 2012 0 30
This indicates that generated assembly code associated with the function, main, was
placed in the CODE linker class (see Section [Link].3 “Class”), was located at
address 2012 (decimal), in address space number 0 (see
Section [Link].15 “Space”), and was 30 (again decimal) words/bytes long. An
introduction to psects is given in Section 5.15.1 “Program Sections”.

3.6.14 How Do I Know What Resources Are Being Used by Each

Function?
In the assembly list file there is information printed for every C function, including library
functions; see Section 6.6 “Assembly List Files”. This information indicates what
registers the function used, what functions it calls (this is also found in the call graph;
see Section 6.6.6 “Call Graph”), and how many bytes of data memory it requires.
Note that auto, parameter and temporary variables used by a function can overlap with
those from other functions as these are placed in a compiled stack by the compiler; see
Section [Link].1 “Compiled Stack Operation”.

3.6.15 How Do I Find Out Where Variables and Functions Have Been
Positioned?
You can determine where variables and functions have been positioned from either the
assembly list file, see Section 6.6 “Assembly List Files”; or the map file, see
Section 7.4 “Map Files”. Only global symbols are shown in the map file; all symbols
(including locals) are listed in the assembly list file, but only for the code represented
by that list file. (Each assembly module has its own list file.)
There is a mapping between C identifiers and the symbols used in assembly code,
which are the symbols shown in both of these files; see Section [Link] “Equivalent
Assembly Symbols”. The symbol associated with a variable is assigned the address
of the lowest byte of the variable; for functions it is the address of the first instruction
generated for that function.

3.6.16 Why Are Some Objects Positioned Into Memory That I

Reserved?
The memory reservation options (see Section [Link] “How Do I Stop the Compiler
From Using Certain Memory Locations?”) will adjust the range of addresses asso-
ciated with classes used by the linker. Most variables and function are placed into
psects (see Section 5.15.1 “Program Sections”) that are linked anywhere inside
these class ranges and so are affected by these reservation options.
Some psects are explicitly placed at an address rather than being linked anywhere in
an address range, e.g., the psect that holds the code to be executed at Reset is always
linked to address 0 because that is where the Reset location is defined to be for 8-bit
devices. Such a psect will not be affected by the --ROM option, even if you use it to
reserve memory address 0. Psects that hold code associated with Reset and interrupts
can be shifted using the --CODEOFFSET option; see Section 4.8.20 “--CLIST:
Generate C Listing File”.
Check the assembly list file (see Section 6.6 “Assembly List Files”) to determine the
names of psects that hold objects and code. Check the linker options in the map file;
see Section 7.4 “Map Files”, to see if psects have been linked explicitly or if they are
linked anywhere in a class. See also, the linker options -p (Section 7.2.19 “-Pspec”)
and -A (Section 7.2.1 “-Aclass =low-high,...”).

 2012-2013 Microchip Technology Inc. DS50002053D-page 69

MPLAB® XC8 C Compiler User’s Guide

3.6.17 How Do I Know How Much Memory Is Still Available?

Although the memory summary printed by the compiler after compilation, (see
Section 4.8.58 “--SUMMARY: Select Memory Summary Output Type” options), or
the memory gauge available in MPLAB IDE both indicate the amount of memory used
and the amount still available, neither of these features indicate whether this memory
is one contiguous block or broken into many small chunks. Small blocks of free memory
cannot be used for larger objects and so out-of-memory errors can be produced even
though the total amount of memory free is apparently sufficient for the objects to be
positioned. (See Section 3.7.6 “How Do I Fix a “Can’t find space...” Error?”.)
The “UNUSED ADDRESS RANGES” section (see Section [Link] “Unused Address
Ranges”) in the map file indicates exactly what memory is still available in each linker
class. It also indicated the largest contiguous block in that class if there are memory
bank or page divisions.

3.6.18 How Do I Use Library Files in My Project?

See Section 3.3.6 “How Do I Build Libraries?” for information on how you build your
own library files. The compiler will automatically include any applicable standard library
into the build process when you compile, so you never need to control these files.
To use one or more library files that were built by yourself or a colleague, include them
in the list of files being compiled on the command line. The library files can be specified
in any position in the file list relative to the source files, but if there is more than one
library file, they will be searched in the order specified in the command line. The LPP
libraries do not need to be specified if you are compiling to an intermediate file, i.e.,
using the --PASS1 option (see Section 4.8.45 “--PARSER: Specify Parser Mode”).
For example:
xc8 --chip=16f1937 main.c int.c [Link]
If you are using MPLAB X IDE to build a project, add the library file(s) to the Libraries
folder that will shown in your project, in the order in which they should be searched. The
IDE will ensure that they are passed to the compiler at the appropriate point in the build
sequence.

3.6.19 What Optimizations Are Employed by the Compiler?

Optimizations are employed at both the C and assembly level of compilation. This is
described in Section 5.13 “Optimizations” and Section 6.5 “Assembly-Level Opti-
mizations”, respectively. The options that control optimization are described in
Section 4.8.42 “--OPT: Invoke Compiler Optimizations”.

3.6.20 Why Do I Get Out-of-memory Errors When I Select a Debugger?

If you use a hardware tool debugger, such as the REAL ICE or ICD3, these require
memory for the on-board debug executive. When you select a debugger using the com-
piler’s --DEBUGGER option (Section 4.8.22 “--DEBUGGER: Select Debugger
Type”), or the IDE equivalent, the memory required for debugging is removed from that
available to your project. See Section 3.5.3 “What Do I Need to Do When Compiling
to Use a Debugger?”

DS50002053D-page 70  2012-2013 Microchip Technology Inc.

 How To’s

3.6.21 How Do I Know Which Stack Model the Compiler Has Assigned
to a Function?
Look in the function information section in the assembly list file, see
Section 6.6.3 “Function Information”. The last line of this block will indicate whether
the function uses a reentrant or non-reentrant model.

3.6.22 How Do I Know What Value Has Been Programmed in the

Configuration Bits or ID Location?
Check the file [Link]. This contains the output of the #pragma config direc-
tive. You will see the numerical value programmed to the appropriate locations. In the
following example, the configuration value programmed is 0xFFBF. A breakdown of
what this value means is also printed.
; Config register CONFIG @ 0x2007
; BOREN = OFF, BOR disabled
; ...
; PWRTE = 0x1, unprogrammed default

psect config
org 0x0
dw 0xFFBF

 2012-2013 Microchip Technology Inc. DS50002053D-page 71

MPLAB® XC8 C Compiler User’s Guide

3.7 FIXING CODE THAT DOES NOT WORK

This section examines issues relating to projects that do not build due to compiler
errors, or those that build, but do not work as expected.
• How Do I Find Out What an Warning/error Message Means?
• How Do I Find the Code that Caused Compiler Errors or Warnings in My
Program?
• How Can I Stop Spurious Warnings From Being Produced?
• Why Can’t I Even Blink an LED?
• How Do I Know If the Hardware Stack Has Overflowed?
• How Do I Fix a “Can’t find space...” Error?
• How Do I Fix a “Can’t generate code...” Error?
• How Do I Fix a Fixup Overflow Error?
• What Can Cause Corrupted Variables and Code Failure When Using Interrupts?

3.7.1 How Do I Find Out What an Warning/error Message Means?

Each warning or error message has a description, and possibly sample code that might
trigger such an error, listed in the messages chapter, see Appendix C. “Error and
Warning Messages”. The compiler prints with each message a unique ID number in
brackets. Use this number to look up the message in the manual. This number also
allows you to control message behavior using options and pragmas, see
Section 4.6.5 “Changing Message Behavior”.

3.7.2 How Do I Find the Code that Caused Compiler Errors or

Warnings in My Program?
In most instances, where the error is a syntax error relating to the source code, the
message produced by the compiler indicates the offending line of code, see
Section 4.6 “Compiler Messages”. If you are compiling in MPLAB IDE, then you can
double-click the message and have the editor take you to the offending line. But
identifying the offending code is not always so easy.
In some instances, the error is reported on the line of code following the line that needs
attention. This is because a C statement is allowed to extend over multiple lines of the
source file. It is possible that the compiler cannot be able to determine that there is an
error until it has started to scan to statement following. So in the following code
input = PORTB // oops - forgot the semicolon
if(input>6)
// ...
The missing semicolon on the assignment statement will be flagged on the following
line that contains the if() statement.
In other cases, the error might come from the assembler, not the code generator. If the
assembly code was derived from a C source file then the compiler will try to indicate
the line in the C source file that corresponds to the assembly that is at fault. If the
source being compiled is an assembly module, the error directly indicates the line of
assembly that triggered the error. In either case, remember that the information in the
error relates to some problem is the assembly code, not the C code.
Finally, there are errors that do not relate to any particular line of code at all. An error
in a compiler option or a linker error are examples of these. If the program defines too
many variables, there is no one particular line of code that is at fault; the program as a
whole uses too much data. Note that the name and line number of the last processed
file and source can be printed in some situations even though that code is not the direct
source of the error.

DS50002053D-page 72  2012-2013 Microchip Technology Inc.

 How To’s
To determine the application that generated the error or warning, take a note of its
unique number printed in the message, see Section 4.6.1 “Messaging Overview”,
and check the message section of the manual, see Appendix C. “Error and Warning
Messages”. At the top of each message description, on the right in brackets, is the
name of the application that produced this message. Knowing the application that pro-
duced the error makes it easier to track down the problem. The compiler application
names are indicated in Section 4.3 “The Compilation Sequence”. If you need to see
the assembly code generated by the compiler, look in the assembly list file, see
Section 6.6 “Assembly List Files”. For information on where the linker attempted to
position objects, see the map file discussed in Section 7.4 “Map Files”.

3.7.3 How Can I Stop Spurious Warnings From Being Produced?

Warnings indicate situations that could possibly lead to code failure. In many situations
the code is valid and the warning is superfluous. Always check your code to confirm
that it is not a possible source of error and in cases where this is so, there are several
ways that warnings can be hidden.
• The warning level threshold can be adjusted so that only warnings of a certain
importance are printed, see Section [Link] “Disabling Messages”
• All warnings with a specified ID can be inhibited
• In some situations, a pragma can be used to inhibit a warning with a specified ID
for certain lines of source code, see Section [Link] “The #pragma warning
Directive”.

3.7.4 Why Can’t I Even Blink an LED?

Even if you have set up the TRIS register and written a value to the port, there are
several things that can prevent such a seemingly simple program from working.
• Make sure that the device’s Configuration registers are set up correctly, see
Section 5.3.5 “Configuration Bit Access”. Make sure that you explicitly specify
every bit in these registers and don’t just leave them in their default state. All the
configuration features are described in your device data sheet. If the Configura-
tion bits that specify the oscillator source are wrong, for example, the device clock
cannot even be running.
• If the internal oscillator is being used, in addition to Configuration bits there can be
SFRs you need to initialize to set the oscillator frequency and modes, see
Section 5.3.6 “Using SFRs From C Code” and your device data sheet.
• Either turn off the Watch Dog Timer in the Configuration bits or clear the Watch
Dog Timer in your code (see Section Appendix A. “Library Functions”) so that
the device does not reset. If the device is resetting, it can never reach the lines of
code in your program that blink the LED. Turn off any other features that can
cause device Reset until your test program is working.
• The device pins used by the port bits are often multiplexed with other peripherals.
A pin might be connected to a bit in a port, or it might be an analog input, or it
might the output of a comparator, for example. If the pin connected to your LED is
not internally connected to the port you are using, then your LED will never oper-
ate as expected. The port function tables shown in your device data sheets will
show other uses for each pin that will help you identify peripherals to investigate.

 2012-2013 Microchip Technology Inc. DS50002053D-page 73

MPLAB® XC8 C Compiler User’s Guide
• Make sure you do not have a “read-modify-write” problem. If the device you are
using does not have a separate “latch” register (as is the case with mid-range PIC
devices) this problem can occur, particularly if the port outputs are driving large
loads, such as an LED. You can see that setting one bit turns off another or other
unusual events. Create your own latch by using a temporary variable. Rather than
read and write the port directly, make modifications to the latch variable. After
modifications are complete, copy the latch as a whole to the port. This means you
are never reading the port to modify it. Check the device literature for more
detailed information.

3.7.5 How Do I Know If the Hardware Stack Has Overflowed?

The 8-bit PIC devices have a limited hardware stack that is only used for function (and
interrupt function) return addresses, see Section 5.3.4 “Stacks”. If the nesting of func-
tion calls and interrupts is too deep, the stack will overflow (wraps around and over-
writes previous entries). Code will then fail at a later point — sometimes much later in
the call sequence — when it accesses the corrupted return address.
The compiler attempts to track stack depth and, when required, swap to a method of
calling that does not need the hardware stack (PIC10/12/16 devices only). You have
some degree of control over what happens when the stack depth has apparently over-
flowed, see Section 4.8.51 “--RUNTIME: Specify Runtime Environment” and the
stackcall suboption.
A call graph shows the call hierarchy and depth that the compiler has determined. This
graph is shown in the assembly list file. To understand the information in this graph, see
Section 6.6.6 “Call Graph”.
Since the runtime behavior of the program cannot be determined by the compiler, it can
only assume the worst case and can report that overflow is possible even though it is
not. However, no overflow should go undetected if the program is written entirely in C.
Assembly code that uses the stack is not considered by the compiler and this must be
taken into account.

3.7.6 How Do I Fix a “Can’t find space...” Error?

There are a number of different variants of this message, but all essentially imply a sim-
ilar situation. They all relate to there being no free space large enough to place a block
of data or instructions. Due to memory paging, banking or other fragmentation, this
message can be issued when seemingly there is enough memory remaining. See
Appendix C. “Error and Warning Messages” for more information on your particular
error number.

3.7.7 How Do I Fix a “Can’t generate code...” Error?

This is a catch-all message which is generated if the compiler has exhausted all possi-
ble means of compiling a C expression, see Appendix C. “Error and Warning Mes-
sages”. It does not usually indicate a fault in your code. The inability to compile the
code can be a deficiency in the compiler, or an expression that requires more registers
or resources than are available at that point in the code. This is more likely to occur on
baseline devices. In any case, simplifying the offending expression, or splitting a state-
ment into several smaller statements, usually allows the compilation to continue. You
can need to use another variable to hold the intermediate results of complicated
expressions.

DS50002053D-page 74  2012-2013 Microchip Technology Inc.

 How To’s

3.7.8 How Do I Fix a Fixup Overflow Error?

Fixup – the process of replacing a symbolic reference with an actual address – can
overflow if the address assigned to the symbol is too large to fit in the address field of
the assembly instruction. Most 8-bit PIC assembly instructions specify a file address
that is an offset into the currently selected memory bank. If a full unmasked address is
specified with these instructions, the linker will be unable to encode the large address
value into the instruction and this error will be generated. For example, a mid-range
device instruction only allows for file addresses in the range of 0 to 0x7F. However, if
such a device has 4 data banks of RAM, the address of variables can range from 0 to
0x1FF.
If the symbol of a variable that will be located at address 1D0, for example, is specified
with one of these instructions, when the symbol is replaced with its final value, this
value will not fit in the address field of the instruction.
In most cases, these errors are caused by hand-written assembly code. When writing
assembly, it is the programmer’s responsibility to add instructions to select the destina-
tion bank and then mask the address being used in the instruction, see
Section [Link] “What Things Must I Manage When Writing Assembly Code?”. It
is important to remember that this is an issue with an assembly instruction and you
need to find the instruction at fault before you can proceed. See the relevant error num-
ber in Appendix C. “Error and Warning Messages” for specific details on to how to
track down the offending instruction.

3.7.9 What Can Cause Corrupted Variables and Code Failure When
Using Interrupts?
This is usually caused by having variables used by both interrupt and main-line code.
If the compiler optimizes access to a variable or access is interrupted by an interrupt
routine, then corruption can occur. See Section 3.5.5 “How Do I Share Data Between
Interrupt and Main-line Code?” for more information.

 2012-2013 Microchip Technology Inc. DS50002053D-page 75

MPLAB® XC8 C Compiler User’s Guide
NOTES:

DS50002053D-page 76  2012-2013 Microchip Technology Inc.

 MPLAB® XC8 C COMPILER
USER’S GUIDE

Chapter 4. XC8 Command-line Driver

4.1 INTRODUCTION
The name of the command-line driver is xc8. MPLAB XC8 can be invoked to perform
all aspects of compilation, including C code generation, assembly, and link steps. Even
if an IDE is used to assist with compilation, the IDE will ultimately call xc8.
Although the internal compiler applications can be called explicitly from the command
line, the xc8 driver is the recommended way to use the compiler as it hides the com-
plexity of all the internal applications used and provides a consistent interface for all
compilation steps.
This chapter describes the steps that the driver takes during compilation, the files that
the driver can accept and produce, as well as the command-line options that control
the compiler’s operation. The relationship between these command-line options and
the controls in the MPLAB IDE Build Options dialog is also described.
The following topics are examined in this chapter of the MPLAB XC8 C Compiler User’s
Guide:
• Invoking the Compiler
• The Compilation Sequence
• Runtime Files
• Compiler Output
• Compiler Messages
• MPLAB XC8 Driver Options
• MPLAB X Option Equivalents

 2012-2013 Microchip Technology Inc. DS50002053D-page 77

MPLAB® XC8 C Compiler User’s Guide

4.2 INVOKING THE COMPILER

This section explains how to invoke xc8 on the command line, as well as the files that
it can read.

4.2.1 Driver Command-line Format

The xc8 driver has the following basic command format:
xc8 [options] files [libraries]
Throughout this manual, it is assumed that the compiler applications are in the con-
sole’s search path or that the full path is specified when executing an application. The
compiler’s location can be added to the search path when installing the compiler by
selecting the Add to environment checkbox at the appropriate time during the
installation.
It is customary to declare options (identified by a leading dash “-” or double dash “–”)
before the files’ names. However, this is not mandatory.
The formats of the options are supplied in Section 4.7 “MPLAB XC8 Driver Options”,
along with corresponding descriptions of the options.
The files can be an assortment of C and assembler source files, and precompiled
intermediate files, such as relocatable object (.obj) files or p-code (.p1) files. While
the order in which the files are listed is not important, it can affect the order in which
code or data appears in memory, and can affect the name of some of the output files.
Libraries is a list of user-defined object code or p-code library files that will be
searched by the code generator (in the case of p-code libraries) or the linker (for object
code libraries), in addition to the standard C libraries. The order of these files will deter-
mine the order in which they are searched. It is customary to insert the Libraries list
after the list of source file names. However, this is not mandatory.
If you are building code using a make system, familiarity with the unique intermediate
p-code file format, as described in Section 4.3.3 “Multi-Step Compilation”, is
recommended. Object files are seldom used with the MPLAB XC8 C Compiler, unless
assembly source modules are in the project.

[Link] LONG COMMAND LINES

The xc8 driver is capable of processing command lines exceeding any operating sys-
tem limitation if the driver is passed options via a command file. The command file is
specified by the @ symbol, which should be immediately followed (i.e., no intermediate
space character) by the name of the file containing the command-line arguments that
are intended for the driver.
Each command-line argument must be separated by one or more spaces and can
extended to several lines by using a space and backslash character to separate lines.
The file can contain blank lines, which are simply skipped by the driver.
The use of a command file means that compiler options and source code filenames can
be permanently stored for future reference without the complexity of creating a make
utility.
In the following example, a command file xyz.xc8 was constructed in a text editor to
contain both the options and the file names that are required to compile a project.
--chip=16F877A -m \
--opt=all -g \
main.c isr.c
After it is saved, the compiler can be invoked with the following command:
xc8 @xyz.xc8

DS50002053D-page 78  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.2.2 Environment Variables

When hosted on a Windows environment, the compiler uses the registry to store infor-
mation relating to the compiler installation directory and activation details, along with
other configuration settings. That information is required whether the compiler is run on
the command line or from within an IDE.
Under Linux® and Mac OS® X environments, the registry is replaced by an XML file that
stores the same information.
On non-Windows hosts, the compiler searches for the XML file in the following ways:
1. The compiler looks for the presence of an environment variable called XC_XML.
If present, this variable should contain the full path to the XML file (including the
file’s name).
2. If this variable is not defined, the compiler then searches for an environment vari-
able called HOME. This variable typically contains the path to the user’s home
directory. The compiler looks for the XML with a name .[Link] in the directory
indicated by the HOME variable.
3. If the HOME environment variable is not defined, the compiler tries to open the file
/etc/[Link].
4. If none of these methods finds the XML file, an error is generated.
When running the compiler on the command line, you can wish to set the PATH envi-
ronment variable. This allows you to run the compiler driver without specifying the full
compiler path with the driver name. Note that the directories specified by the PATH vari-
able are only used to locate the compiler driver. Once the driver is running, it uses the
registry or XML file, described above, to locate the internal compiler applications, such
as the parser, assembler and linker, etc. The directories specified in the PATH variable
do not override the information contained in the registry or XML file. The MPLAB IDE
allows the compiler to be selected via a dialog and execution of the compiler does not
depend on the PATH variable.

 2012-2013 Microchip Technology Inc. DS50002053D-page 79

MPLAB® XC8 C Compiler User’s Guide

4.2.3 Input File Types

xc8 distinguishes source files, intermediate files, and library files solely by the file type,
or extension. Recognized file types are listed in Table 4-1. Alphabetic case of the
extension is not important from the compiler’s point of view, but most operating system
shells are case sensitive.

TABLE 4-1: xc8 INPUT FILE TYPES

File Type Meaning
.c C source file
.p1 p-code file
.lpp p-code library file
.as or .asm Assembler source file
.obj Relocatable object code file
.lib Relocatable object library file
.hex Intel HEX file

This means, for example, that a C source file must have a .c extension. Assembler
files can use either .as or .asm extensions.
There are no compiler restrictions imposed on the names of source files, but be aware
of case, name-length, and other restrictions that are imposed by your operating sys-
tem. If you are using an IDE, avoid assembly source files whose base name is the
same as the base name of any project in which the file is used. This can result in the
source file being overwritten by a temporary file during the build process.
The terms “source file” and “module” are often used when talking about computer
programs. They are often used interchangeably, but they refer to the source code at
different points in the compilation sequence.
A source file is a file that contains all or part of a program. They can contain C code, as
well as preprocessor directives and commands. Source files are initially passed to the
preprocessor by the driver.
A module is the output of the preprocessor, for a given source file, after inclusion of any
header files (or other source files) which are specified by #include preprocessor
directives. All preprocessor directives and commands (with the exception of some com-
mands for debugging) have been removed from these files. These modules are then
passed to the remainder of the compiler applications. Thus, a module can be the amal-
gamation of several source and header files. A module is also often referred to as a
translation unit. These terms can also be applied to assembly files, as they can include
other header and source files.

DS50002053D-page 80  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.3 THE COMPILATION SEQUENCE

When you compile a project, many internal applications are called to do the work. This
section looks at when these internal applications are executed, and how this relates to
the build process of multiple source files. This section should be of particular interest if
you are using a make system to build projects.

4.3.1 The Compiler Applications

The main internal compiler applications and files are illustrated in Figure 4-1.
You can consider the large underlying box to represent the whole compiler, which is
controlled by the command line driver, xc8. You can be satisfied just knowing that C
source files (shown on the far left) are passed to the compiler and the resulting output
files (shown here as a HEX and COFF debug file on the far right) are produced; how-
ever, internally there are many applications and temporary files being produced. An
understanding of the internal operation of the compiler, while not necessary, does
assist with using the tool.
To simplify the compiler design, some of the internal applications come in a PIC18 and
PIC10/12/16 variant. The appropriate application is executed based on the target
device. In fact, the xc8 driver delegates the build commands to one of two com-
mand-line drivers: PICC or PICC18. This operation is transparent and xc8 can be
considered as “the driver” which does all the work.
The driver will call the required compiler applications. These applications are shown as
the smaller boxed inside the large driver box. The temporary file produced by each
application can also be seen in this diagram.

FIGURE 4-1: COMPILER APPLICATIONS AND FILES

assembly
p-code .lpp .p1 p-code .as source
libraries files files
Command-line driver
processed p-code relocatable
files (module) files assembly file object file

C source .pre .p1 .as .obj

files

.c code
preprocessor
p or parser assembler
generator

debug file

cromwell .cof
.c

linker objtohex

hexmate .hhex
.obj .hex
hex file
absolute hex file
object file

relocatable .obj .lib object hex .hex

object files libraries files

 2012-2013 Microchip Technology Inc. DS50002053D-page 81

MPLAB® XC8 C Compiler User’s Guide
Table 4-2 lists the compiler applications. The names shown are the names of the exe-
cutables, which can be found in the bin directory under the compiler’s installation
directory.

TABLE 4-2: COMPILER APPLICATION NAMES

Name Description
xc8 (calls PICC or PICC18) Command line driver; the interface to the compiler
CLIST Text file formatter
CPP The C preprocessor
P1 C code parser
CGPIC or CGPIC18 Code generator (based on the target device)
ASPIC or ASPIC18 Assembler (based on the target device)
HLINK Linker
OBJTOHEX Conversion utility to create HEX files
CROMWELL Debug file converter
HEXMATE HEX file utility
LIBR Librarian
DUMP Object file viewer

For example, C source files (.c files) are first passed to the C preprocessor, CPP. The
output of this application is .pre files. These files are then passed to the parser appli-
cation, P1, which produces a p-code file output with extension .p1. The applications
are executed in the order specified and temporary files are used to pass the output of
one application to the next.
The compiler can accept more than just C source files. Table 4-1 lists all the possible
input file types, and these files can be seen in this diagram, on the top and bottom,
being passed to different compilation applications. They are processed by these
applications and then the application output joins the normal flow indicated in the
diagram.
For example, assembly source files are passed straight to the assembler application1
and are not processed at all by the code generator. The output of the assembler (an
object file with .obj extension) is passed to the linker in the usual way. You can see
that any p-code files (.p1 extension) or p-code libraries (.lpp extension) that are
supplied on the command line are initially passed to the code generator.
Other examples of input files include object files (.obj extension) and object libraries
(.lib extension), both of which are passed initially to the linker, and even HEX files
(.hex extension), which are passed to one of the utility applications, called HEXMATE,
which is run right at the end of the compilation sequence.
Some of the temporary files shown in this diagram are actually preserved and can be
inspected after compilation has concluded. There are also driver options to request that
the compilation sequence stop after a particular application and the output of that
application becomes the final output.

1. Assembly file will be preprocessed before being passed to the assembler if the -P option is
selected.

DS50002053D-page 82  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver
FIGURE 4-2: MULTI-FILE COMPILATION
Intermediate files

preprocess
p-
C file & code
parse

preprocess
p- code
C file & code assemble link
generation
parse

library
files

First stage of compilation Second stage of compilation

4.3.2 Single-Step Compilation

Figure 4-1 showed us the files that are generated by each application and the order in
which these applications are executed. However this does not indicate how these
applications are executed when there is more than one source file being compiled.
Consider the case when there are two C source files that form a complete project and
that are to be compiled, as is the case shown in Figure 4-2. If these files are called
main.c and io.c, these could be compiled with a single command, such as:
xc8 --chip=16F877A main.c io.c
This command will compile the two source files all the way to the final output, but
internally we can consider this compilation as consisting of two stages.
The first stage involves processing of each source file separately, and generating some
sort of intermediate file for each source file. The second stage involves combining all
these intermediate files and further processing to form the final output. An intermediate
file is a particular temporary file that is produced and marks the midpoint between the
first and second stage of compilation.
The intermediate file used by xc8 is the p-code (.p1 extension) file output by the
parser, so there will be one p-code file produced for each C source file. As indicated in
the diagram, CPP and then P1 are executed to form this intermediate file. (For clarity,
the CPP and P1 applications have been represented by the same block in the diagram.)
In the second stage, the code generator reads in all the intermediate p-code files and
produces a single assembly file output, which is then passed to the subsequent
applications that produce the final output.
The desirable attribute of this method of compilation is that the code generator, which
is the main application that transforms from the C to the assembly domain, sees the
entire project source code via the intermediate files.
Traditional compilers have always used intermediate files that are object files output by
the assembler. These intermediate object files are then combined by the linker and fur-
ther processed to form the final output. This method of compilation is shown in
Figure 4-3. It shows that the code generator is executed once for each source file. So,
the code generator can only analyze that part of the project that is contained in the
source file that is currently being compiled. The MPLAB XC16 and XC32 compilers
work in this fashion.
Using object files as the intermediate file format with MPLAB XC8 C Compiler will
defeat many features the compiler uses to optimize code. Always use p-code files as
the intermediate file format if you are using a make system to build projects.

 2012-2013 Microchip Technology Inc. DS50002053D-page 83

MPLAB® XC8 C Compiler User’s Guide
FIGURE 4-3: THE TRADITIONAL COMPILATION SEQUENCE
Intermediate files

preprocess
code .obj
C file & assemble files
generation
parse

preprocess
code .obj
C file & assemble files link
generation
parse

library Second stage

First stage of compilation files
of compilation

When compiling files of mixed types, this can still be achieved with just one invocation
of the compiler driver. As discussed in Section 4.3 “The Compilation Sequence”, the
driver will pass each input file to the appropriate compiler application.
For example, the files, main.c, io.c, [Link] and c_sb.lpp are to be compiled.
To perform this in a single step, the following command line could be used.
xc8 --chip=16F877A main.c io.c [Link] c_sb.lpp
As shown in Figure 4-1 and Figure 4-2, the two C files (main.c and io.c) will be com-
piled to intermediate p-code files; these, along with the p-code library file (c_sb.lpp)
will be passed to the code generator. The output of the code generator, as well as the
assembly source file ([Link]), will be passed to the assembler.
The driver will recompile all source files, regardless of whether they have changed
since the last build. IDEs (such as MPLAB® IDE) and make utilities must be employed
to achieve incremental builds. See also, Section 4.3.3 “Multi-Step Compilation”.
Unless otherwise specified, a HEX file and Microchip COFF file are produced as the
final output. All intermediate files remain after compilation has completed, but most
other temporary files are deleted, unless you use the --NODEL option (see
Section 4.8.40 “--NODEL: Do Not Remove Temporary Files”) which preserves all
generated files except the run-time start-up file. Note that some generated files can be
in a different directory than your project source files. See Section 4.8.43 “--OUTDIR:
Specify a Directory for Output Files”, and Section 4.8.41 “--OBJDIR: Specify a
Directory for Intermediate Files”, which can both control the destination for some
output files.

DS50002053D-page 84  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.3.3 Multi-Step Compilation

Make utilities and IDEs, such as MPLAB IDE, allow for an incremental build of projects
that contain multiple source files. When building a project, they take note of which
source files have changed since the last build and use this information to speed up
compilation.
For example, if compiling two source files, but only one has changed since the last
build, the intermediate file corresponding to the unchanged source file need not be
regenerated.
MPLAB IDE is aware of the different compilation sequence employed by xc8 and takes
care of this for you. From MPLAB IDE you can select an incremental build (Build Project
icon), or fully rebuild a project (Clean and Build Project icon).
If the compiler is being invoked using a make utility, the make file will need to be con-
figured to recognized the different intermediate file format and the options used to gen-
erate the intermediate files. Make utilities typically call the compiler multiple times: once
for each source file to generate an intermediate file, and once to perform the second
stage compilation.
You can also wish to generate intermediate files to construct your own library files.
However, xc8 is capable of constructing libraries in a single step, so this is typically not
necessary. See Section 4.8.44 “--OUTPUT= type: Specify Output File Type”, for
more information on library creation.
The option --PASS1 (see Section 4.8.45 “--PARSER: Specify Parser Mode”) is
used to tell the compiler that compilation should stop after the parser has executed.
This will leave the p-code intermediate file behind on successful completion.
For example, the files main.c and io.c are to be compiled using a make utility. The
command lines that the make utility should use to compile these files might be
something like:
xc8 --chip=16F877A --pass1 main.c
xc8 --chip=16F877A --pass1 io.c
xc8 --chip=16F877A main.p1 io.p1
If is important to note that the code generator needs to compile all p-code or p-code
library files associated with the project in the one step. When using the --PASS1 option
the code generator is not being invoked, so the above command lines do not violate
this requirement.
Using object files as the intermediate file format with MPLAB XC8 C Compiler will
defeat many features the compiler uses to optimize code. Always use p-code files as
the intermediate file format if you are using a make system to build projects.

 2012-2013 Microchip Technology Inc. DS50002053D-page 85

MPLAB® XC8 C Compiler User’s Guide

4.3.4 Compilation of Assembly Source

Since the code generator performs many tasks that were traditionally performed by the
linker, there could be complications when assembly source is present in a project.
Assembly files are traditionally processed after C code, but it is necessary to have this
performed first so that specific information contained in the assembly code can be
conveyed to the code generator.
The specific information passed to the code generator is discussed in more detail in
Section 5.12.3 “Interaction between Assembly and C Code”.
When assembly source is present, the order of compilation is as shown in Figure 4-4.

FIGURE 4-4: COMPILATION SEQUENCE WITH ASSEMBLY FILES

preprocess
p-
C file & code
parse
preprocess
p-
C file & code
parse
code
library assemble link
generation
files

ASM
assemble
file

driver

OBJ
file

First, any assembly source files are assembled to form object files. These files, along
with any other objects files that are part of the project, are scanned by the
command-line driver and the information is passed to the code generator; where it
subsequently builds the C files, as has been described earlier.

[Link] INTERMEDIATE FILES AND ASSEMBLY SOURCE

The intermediate file format associated with assembly source files is the same as that
used in traditional compilers; i.e., an object file (.obj extension). Assembly files are
never passed to the code generator and so the code generator technology does not
alter the way these files are compiled.
The -C option (see Section 4.8.1 “-C: Compile to Object File”) is used to generate
object files and to halt compilation after the assembly step.

4.3.5 Printf Check

An extra execution of the code generator is performed prior to the actual code genera-
tion phase. This pass is part of the process by which the printf library function is
customized; see Section 5.11.1 “The printf Routine”, for more details.
This pass is only associated with scanning the C source code for printf placeholder
usage and you will see the code generator being executed if you select the verbose
option when you build; see Section 4.8.14 “-V: Verbose Compile”.

DS50002053D-page 86  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.4 RUNTIME FILES

In addition to the C and assembly source files specified on the command line, there are
also compiler-generated source files and pre-compiled library files which might be
compiled into the project by the driver. These files contain:
• C Standard library routines
• Implicitly called arithmetic routines
• User-defined library routines
• The runtime startup code
• The powerup routine
• The printf routine.
Strictly speaking, the powerup routine is neither a compiler-generated source, nor a
library routine. It is fully defined by the user. Because it is very closely associated with
the runtime startup module, it is discussed with the other runtime files in the following
sections.

4.4.1 Library Files

The names of the C standard library files appropriate for the selected target device, and
other driver options, are determined by the driver and passed to the code generator and
linker. You do not need to include library files into your project manually. P-code librar-
ies (.lpp libraries) are used by the code generator, and object code libraries (.lib
files) are used by the linker. Most library routines are derived from p-code libraries.
By default, xc8 will search the lib directory under the compiler installation directory
for library files that are required during compilation.

[Link] STANDARD LIBRARIES

The C standard libraries contain a standardized collection of functions, such as string,
math and input/output routines. The range of these functions is described in
Appendix A. “Library Functions”. Although it is considered a library function, the
printf function’s code is not found in these library files. C source code for this func-
tion is generated from a special C template file that is customized after analysis of the
user’s C code. See “PRINTF, VPRINTF”, for more information on using the printf
library function, and Section 5.11.1 “The printf Routine”, for information on how the
printf function is customized when you build a project.
The libraries also contain C routines that are implicitly called by the output code of the
code generator. These are routines that perform tasks such as floating-point opera-
tions, integer division and type conversions, and that cannot directly correspond to a C
function call in the source code.
The library name format is [Link], where the following apply.
• family can be pic18 for PIC18 devices, or pic for all other 8-bit PIC devices
• type indicates the sort of library functionality provided and can be stdlib for the
standard library functions, or trace, etc.
• options indicates hyphen-separated names to indicate variants of the library to
accommodate different compiler options or modes, e.g., htc for HI-TECH C
compatibility, d32 for 32-bit doubles, etc.
For example, the standard library for baseline and midrange devices using 24-bit
double types is [Link].
All the libraries are present in the lib directory of the compiler installation. Search this
directory for the full list of all libraries supplied.

 2012-2013 Microchip Technology Inc. DS50002053D-page 87

MPLAB® XC8 C Compiler User’s Guide
[Link] USER-DEFINED LIBRARIES
User-defined libraries can be created and linked in with programs as required. Library
files are easier to manage and can result in faster compilation times, but must be com-
patible with the target device and options for a particular project. Several versions of a
library might need to be created to allow it to be used for different projects.
Libraries can be created manually using the compiler and the librarian, LIBR. See
Section 8.2 “Librarian” for more information on the librarian and creating library files
using this application. Alternatively, library files can be created directly from the
compiler by specifying a library output using the --OUTPUT option; see
Section 4.8.44 “--OUTPUT= type: Specify Output File Type”.
User-created libraries that should be searched when building a project can be listed on
the command line along with the source files.
As with Standard C library functions, any functions contained in user-defined libraries
should have a declaration added to a header file. It is common practice to create one
or more header files that are packaged with the library file. These header files are then
included into source code when required.
Library files specified on the command line are initially scanned for unresolved sym-
bols; so, these files can redefine anything that is defined in the C standard libraries. See
also, Section 5.15.4 “Changing and Linking the Allocated Section”.

4.4.2 Startup and Initialization

A C program requires certain objects to be initialized and the device to be in a particular
state before it can begin execution of its function main. It is the job of the runtime
startup code to perform these tasks. Section 5.10.1 “Runtime Startup Code” details
the specific actions taken by this code and how it interacts with programs you write.
Rather than the traditional method of linking in a generic, precompiled routine, the
MPLAB XC8 C Compiler determines what runtime startup code is required from the
user’s program and then generates this code each time you build.
Both the driver and code generator are involved in generating the runtime startup code.
The driver creates the code that handles device setup. This code is placed into a sep-
arate assembly startup module. The code generator produces code that initializes the
C environment, such as clearing uninitialized C variables and copying initialized C
variables. This code is output along with the rest of the C program.
The runtime startup code is regenerated every time you build a project. The file created
by the driver can be deleted after compilation, and this operation can be controlled with
the keep suboption to the --RUNTIME option. The default operation of the driver is to
keep the startup module; however, if using MPLAB IDE to build, the file will be deleted
unless you indicate otherwise in the Project Properties dialog.
If the startup module is kept, it will be called [Link] and will be located in the
current working directory. If you are using an IDE to perform the compilation, the
destination directory can be dictated by the IDE itself. MPLAB X IDE stores this file in
the dist/default/production directory in your project directory.
Generation of the runtime startup code is an automatic process that does not require
any user interaction; however, some aspects of the runtime code can be controlled, if
required, using the --RUNTIME option. Section 4.8.51 “--RUNTIME: Specify Run-
time Environment” describes the use of this option. See Section 5.10.1 “Runtime
Startup Code”, which describes the functional aspects of the code contained in this
module and its effect on program operation.
The runtime startup code is executed before main. However, if you require any special
initialization to be performed immediately after Reset, you should use the powerup
feature described later in Section 5.10.2 “The Powerup Routine”.

DS50002053D-page 88  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.5 COMPILER OUTPUT

There are many files created by the compiler during the compilation. A large number of
these are intermediate files. Some are deleted after compilation is complete, but many
remain and are used for programming the device, or for debugging purposes.

4.5.1 Output Files

The names of many output files use the same base name as the source file from which
they were derived. For example, the source file input.c will create a p-code file called
input.p1.
Some of the output files contain project-wide information and are not directly associ-
ated with any one particular input file, e.g., the map file. If the names of these output
files are not specified by a compiler option, their base name is derived from the first C
source file listed on the command line. If there are no files of this type specified, the
name is based on the first input file (regardless of type) on the command line.
If you are using an IDE, such as MPLAB X IDE, to specify options to the compiler, there
is typically a project file that is created for each application. The name of this project is
used as the base name for project-wide output files, unless otherwise specified by the
user. However, check the manual for the IDE you are using, for more details.

Note: Throughout this manual, the term project name will refer to either the name
of the project created in the IDE, or the base name (file name without
extension) of the first C source file specified on the command line.

The compiler is directly able to produce a number of the output file formats that are
used by the 8-bit PIC development tools.
The default behavior of xc8 is to produce a Microchip format COFF and Intel HEX out-
put. Unless changed by a driver option, the base names of these files will be the project
name. The default output file types can be controlled by compiler options, e.g., the
--OUTPUT option. The extensions used by these files are fixed and are listed together
with this option’s description in Section 4.8.44 “--OUTPUT= type: Specify Output
File Type”.
The COFF file is used by debuggers to obtain debugging information about the project.
The compiler can produce ELF/DWARF debugger files, although these are not com-
patible with MPLAB IDE v8 and early versions of MPLAB X IDE. You must specifically
select ELF output for these files to be produced. ELF/DWARF files allow for more accu-
rate debugging. Use of these files correct several COFF-related issues that prevent
you from correctly viewing objects, in particular pointer variables, in the IDE. Ensure the
IDE version you are using supports ELF before selecting this option.
Table 4-16 shows all output format options available with xc8 using the --OUTPUT
option. The File Type column lists the filename extension that is used for the output file.

[Link] SYMBOL FILES

By default, xc8 creates symbol files that are used to generate the debug output files,
such as COFF and ELF files. These files include a SYM file (.sym extension) and a
CMF file (.cmf extension), and both are produced by the linker. In addition, there is a
SDB file (.sdb extension) produced by the code generator.
The SDB file contains type information, and the SYM and CMF files contain address
information. The SDB and SYM/CMF files, in addition to the HEX file, are combined by
the CROMWELL application to produce the output debug files, such as the COFF file.
The CMF file largely replaces the older SYM file format. They contain similar informa-
tion, but CMF files are more detailed and enable more accurate debug files to be
generated.

 2012-2013 Microchip Technology Inc. DS50002053D-page 89

MPLAB® XC8 C Compiler User’s Guide

4.5.2 Diagnostic Files

Two valuable files produced by the compiler are the assembly list file, produced by the
assembler, and the map file, produced by the linker.
The compiler options --ASMLIST (see Section 4.8.15 “--ADDRQUAL: Set Compiler
Response to Memory Qualifiers”) generates a list file, and the -M option (see
Section 4.8.7 “-M: Generate Map File”) specifies generation of a map file.
The assembly list file contains the mapping between the original source code and the
generated assembly code. It is useful for information such as how C source was
encoded, or how assembly source can have been optimized. It is essential when con-
firming if compiler-produced code that accesses objects is atomic, and shows the
psects in which all objects and code are placed. For an introductory guide to psects,
see Section 5.15.1 “Program Sections”. Also, see Section 6.5 “Assembly-Level
Optimizations”, for more information on the contents of this file.
There is one list file produced for the entire C program, including C library files. It is
assigned the project name and the extension .lst. One additional list file is produced
for each assembly source file compiled in the project.
The map file shows information relating to where objects were positioned in memory. It
is useful for confirming if user-defined linker options were correctly processed, and for
determining the exact placement of objects and functions. It also shows all the unused
memory areas in a device and memory fragmentation. See Section 7.4 “Map Files”,
for complete information on the contents of this file.
There is one map file produced when you build a project, assuming the linker was
executed and ran to completion. The file is assigned the project name and a .map
extension.

DS50002053D-page 90  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.6 COMPILER MESSAGES

All compiler applications, including the command-line driver, xc8, use textual mes-
sages to report feedback during the compilation process. A centralized messaging sys-
tem is used to produce the messages, which allows consistency during all stages of the
compilation process. The messaging system is described in this section and a com-
plete list of all warning and error messages can be found in Appendix C. “Error and
Warning Messages”.

4.6.1 Messaging Overview

A message is referenced by a unique number that is passed to the messaging system
by the compiler application that needs to convey the information. The message string
corresponding to this number is obtained from Message Description Files (MDF), which
are stored in the dat directory in the compiler’s installation directory.
When a message is requested by a compiler application, its number is looked up in the
MDF that corresponds to the currently selected language. The language of messages
can be altered as discussed in Section 4.6.2 “Message Language”.
Once found, the alert system can read the message type and the string to be displayed
from the MDF. Several different message types are described in
Section 4.6.3 “Message Type”; and the type can be overridden by the user, as
described in that same section.
The user is also able to set a threshold for warning message importance, so that only
those that the user considers significant will be displayed. In addition, messages with
a particular number can be disabled. A pragma can also be used to disable a particular
message number within specific lines of code. These methods are explained in
Section [Link] “Disabling Messages”.
Provided the message is enabled and it is not a warning message whose level is below
the current warning threshold, the message string will be displayed.
In addition to the actual message string, there are several other pieces of information
that can be displayed, such as the message number, the name of the file for which the
message is applicable, the file’s line number and the application that issued the
message, etc.
If a message is an error, a counter is incremented. After a specific amount of errors has
been reached, compilation of the current module will cease. The default number of
errors that will cause this termination can be adjusted by using the --ERRORS option;
see Section 4.8.28 “--ERRORS: Maximum Number of Errors”. This counter is reset
for each internal compiler application, thus specifying a maximum of five errors will
allow up to five errors from the parser, five from the code generator, five from the linker,
five from the driver, etc.
Although the information in the MDF can be modified with any text editor, this is not rec-
ommended. Message behavior should only be altered using the options and pragmas
described in the following sections.

 2012-2013 Microchip Technology Inc. DS50002053D-page 91

MPLAB® XC8 C Compiler User’s Guide

4.6.2 Message Language

The xc8 driver supports more than one language for displayed messages. There is one
MDF for each language supported.
Under Windows®, the default language can be specified when installing the compiler.
The default language can be changed on the command line using the --LANG option;
see Section 4.8.35 “--LANG: Specify the Language for Messages”. Alternatively, it
can be changed permanently by using the --LANG option together with the --SETUP
option which will store the default language in either the registry, under Windows, or in
the XML configuration file on other systems. On subsequent builds, the default
language used will be that specified.
Table 4-3 shows the MDF applicable for the currently supported languages.

TABLE 4-3: SUPPORTED LANGUAGES

Language MDF name
English en_msgs.txt
German de_msgs.txt
French fr_msgs.txt

If a language other than English is selected, and the message cannot be found in the
appropriate non-English MDF, the alert system tries to find the message in the English
MDF. If an English message string is not present, a message is displayed that is similar
to this one:
error/warning (*) generated, but no description available
where * indicates the message number that was generated that will be printed;
otherwise, the message in the requested language will be displayed.

4.6.3 Message Type

There are four types of messages. These are described below along with the com-
piler’s behavior when encountering a message of each type.
Advisory Messages convey information regarding a situation the compiler has en-
countered or some action the compiler is about to take. The information is
being displayed “for your interest”, and typically requires no action to be
taken. Compilation will continue as normal after such a message is issued.
Warning Messages indicate source code or some other situation that can be com-
piled, but is unusual and can lead to a runtime failure of the code. The code
or situation that triggered the warning should be investigated; however, com-
pilation of the current module will continue, as will compilation of any
remaining modules.
Error Messages indicate source code that is illegal or that compilation of this code
cannot take place. Compilation will be attempted for the remaining source
code in the current module, but no additional modules will be compiled and
the compilation process will then conclude.
Fatal Error Messages indicate a situation in which the compilation cannot proceed
and requires that the compilation process to stop immediately.

DS50002053D-page 92  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.6.4 Message Format

By default, messages are printed in a human-readable format. This format can vary
from one compiler application to another, since each application reports information
about different file formats.
Some applications (for example, the parser) are typically able to pinpoint the area of
interest down to a position on a particular line of C source code, whereas other appli-
cations, such as the linker, can at best only indicate a module name and record number,
which is less directly associated with any particular line of code. Some messages relate
to issues in driver options that are in no way associated with any source code.
There are several ways of changing the format in which message are displayed, which
are discussed below.
The driver option -E (with or without a filename) alters the format of all displayed mes-
sages. See Section 4.8.3 “-E: Redirect Compiler Errors to a File”, for details. Using
this option produces messages that are better suited to machine parsing, and are less
user-friendly. Typically, each message is displayed on a single line. The general form
of messages produced when using the -E option is:
filename line: (message number) message string (type)
The -E option also has another effect. When used, the driver first checks to see if spe-
cial environment variables have been set. If so, the format dictated by these variables
is used as a template for all messages that will be produced by all compiler
applications. The names of these environment variables are given in Table 4-4.

TABLE 4-4: MESSAGING ENVIRONMENT VARIABLES

Variable Effect
HTC_MSG_FORMAT All advisory messages
HTC_WARN_FORMAT All warning messages
HTC_ERR_FORMAT All error and fatal error messages

The value of these environment variables are strings that are used as templates for the
message format. Printf-like placeholders can be placed within the string to allow the
message format to be customized. The placeholders, and what they represent, are
presented in Table 4-5.

TABLE 4-5: MESSAGING PLACEHOLDERS

Placeholder Replacement
%a Application name
%c Column number
%f Filename
%l Line number
%n Message number
%s Message string (from MDF)

If these options are used in a DOS batch file, two percent characters will need to be
used to specify the placeholders, as DOS interprets a single percent character as an
argument and will not pass this on to the compiler. For example:
SET HTC_ERR_FORMAT="file %%f: line %%l"

 2012-2013 Microchip Technology Inc. DS50002053D-page 93

MPLAB® XC8 C Compiler User’s Guide
Environment variables, in turn, can be overridden by the driver options: --MSGFORMAT,
--WARNFORMAT and --ERRFORMAT; see Section 4.8.27 “--ERRFORMAT: Define
Format for Compiler Messages”. These options take a string as their argument. The
option strings are formatted, and can use the same placeholders as their variable
counterparts.
For example, a project is compiled, but, as shown, produces a warning from the parser
and an error from the linker (numbered 362 and 492, respectively).
main.c: main()
17: ip = &b;
^ (362) redundant "&" applied to array (warning)
(492) attempt to position absolute psect "text" is illegal
Notice that the parser message format identifies the particular line and position of the
offending source code.
If the -E option is now used and the compiler issues the same messages, the compiler
will output:
main.c: 12: (362) redundant "&" applied to array (warning)
(492) attempt to position absolute psect "text" is illegal (error)
The user now uses the --WARNFORMAT in the following fashion:
--WARNFORMAT="%a %n %l %f %s"
When recompiled, the following output will be displayed:
parser 362 12 main.c redundant "&" applied to array
(492) attempt to position absolute psect "text" is illegal (error)
Notice that the format of the warning was changed, but that of the error message was
not. The warning format now follows the specification of the environment variable. The
application name (parser) was substituted for the %a placeholder, the message
number (362) substituted the %n placeholder, etc.

4.6.5 Changing Message Behavior

Both the attributes of individual messages and general settings for the messaging sys-
tem can be modified during compilation. There are both driver options and C pragmas
that can be used to achieve this.

[Link] DISABLING MESSAGES

Each warning message has a default number indicating a level of importance. This
number is specified in the MDF and ranges from -9 to 9. The higher the number, the
more important the warning.
Warning messages can be disabled by adjusting the warning level threshold using the
--WARN driver option; see Section 4.8.61 “--WARN: Set Warning Level”. Any
warnings whose level is below that of the current threshold are not displayed.
The default threshold is 0 which implies that only warnings with a warning level of 0 or
higher will be displayed by default. The information in this option is propagated to all
compiler applications, so its effect will be observed during all stages of the compilation
process.
Warnings can also be disabled by using the --MSGDISABLE option; see
Section 4.8.38 “--MSGDISABLE: Disable Warning Messages”. This option takes a
comma-separated list of warning numbers. The warnings corresponding to the num-
bers listed are disabled and will never be issued, regardless of the current warning level
threshold. If the special message number 0 is specified, then all warning messages are
disabled.

DS50002053D-page 94  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver
Some warning messages can also be disabled by using the warning pragma. This
pragma will only affect warnings that are produced by either the parser or the code gen-
erator; i.e., errors directly associated with C code. See Section [Link] “The
#pragma warning Directive” for more information on this pragma.
Error messages can also be disabled; however, a more verbose form of the above
command is required to confirm the action. To specify an error message number in the
--MSGDISABLE command, each error number must be followed by :off to ensure
that it is disabled. For example:
--MSGDISABLE=1257,195:off,194:off
will disable warning 1257, and errors 195 and 194.

Note: Disabling error or warning messages in no way fixes the condition that
triggered the message. Always use extreme caution when exercising these
options.

[Link] CHANGING MESSAGE TYPES

It is also possible to change the type of some messages. This can only be done for
messages generated by the parser or code generator. See Section [Link] “The
#pragma warning Directive”, for more information on this pragma.

 2012-2013 Microchip Technology Inc. DS50002053D-page 95

MPLAB® XC8 C Compiler User’s Guide

4.7 MPLAB XC8 DRIVER OPTIONS

This section looks at the general form of xc8 command-line options and what action
the compiler will perform if no option is specified for a certain feature.

4.7.1 General Option Formats

All single letter options are identified by a leading dash character, “-”, for example: -C.
Some single letter options specify an additional data field that follows the option name
immediately and without any whitespace, for example: -Ddebug. In this manual,
options are written in upper case and suboptions are written in lower case.
Multi-letter, or word, options have two leading dash characters, for example:
--ASMLIST. (Because of the double dash, the driver can determine that the option
--DOUBLE, for example, is not a -D option followed by the argument OUBLE.)
Some of these word options use suboptions which typically appear as a comma-sepa-
rated list following an equal character, =, for example: --OUTPUT=hex,cof. The exact
formats of the options vary. The options and formats are described in detail in the
following sections.
Some commonly used suboptions include default, which represent the default spec-
ification that would be used if this option was absent altogether; all, which indicates
that all the available suboptions should be enabled as if they had each been listed; and
none, which indicates that all suboptions should be disabled. For example:
--OPT=none
will turn off all optimizers.
Some suboptions can be prefixed with a plus character, +, to indicate that they are in
addition to the other suboptions present; or a minus character “-”, to indicate that they
should be excluded. For example:
--OPT=default,-asm
indicates that the default optimization be used, but that the assembler optimizer should
be disabled. If the first character after the equal sign is + or -, then the default keyword
is implied. For example:
--OPT=-asm
is the same as the previous example.
See the –-HELP option, Section 4.8.33 “--HELP: Display Help”, for more information
about options and suboptions.

4.7.2 Default Options

If you run the compiler driver from the command line and do not specify the option for
a feature, it will default to a certain state. You can also specify the default suboption
(to double-dash options) that will also invoke the default behavior. You can check what
the default behavior is by using the --HELP=option on the command line; see
Section 4.8.33 “--HELP: Display Help”.
If you are compiling from within the MPLAB X IDE, it will, by default, issue explicit
options to the compiler (unless changed in the Project Properties dialog), and these
options can be different to those that are the default on the command line. For example,
unless you specify the --ASMLIST option on the command line, the default operation
of the compiler is not to produce an assembly list file. But, if you are compiling from
within the MPLAB X IDE, the default operation (this, in fact, cannot be disabled) is to
always produce an assembly list file.
If you are compiling the same project from the command line and from the MPLAB X
IDE, always check that all options are explicitly specified.

DS50002053D-page 96  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8 OPTION DESCRIPTIONS

Most aspects of the compilation can be controlled using the command-line driver, xc8.
The driver will configure and execute all required applications, such as the code
generator, assembler and linker.
xc8 recognizes the compiler options which are tabled below and are explained in detail
in the sections following. The case of the options is not important; however, command
shells in most operating systems are case sensitive when it comes to the names of
files.

TABLE 4-6: DRIVER OPTIONS

Option Meaning
-C Compile to object file and stop
-Dmacro Define preprocessor macro symbol
-Efilename Redirect compile errors
-G[filename] Generate symbolic debug information
-Ipath Specify include path
-Largument Set linker option
-M[filename] Generate map file
-Nnumber Specify identifier length
-Ofile Specify output filename and type
-P Preprocess assembly source
-Q Quiet mode
-S Compile to assembly file and stop
-Umacro Undefine preprocessor macro symbol
-V Verbose mode
--ADDRQUAL=qualifier Specify address space qualifier handling
--ASMLIST Generate assembly list file
--CHAR=type Default character type (defunct)
--CHECKSUM=specification Calculate a checksum and store the result in program
memory
--CHIP=device Select target device
--CHIPINFO Print device information
--CODEOFFSET=value Specify ROM offset address
--DEBUGGER=type Set debugger environment
--DOUBLE=size Size of double type
--ECHO Echo command line
--EMI=mode Select external memory interface operating mode
--ERRATA=type Specify errata workarounds
--ERRFORMAT=format Set error format
--ERRORS=number Set maximum number of errors
--EXT=extensions Specify C language extensions
--FILL=specification Specify a ROM-fill value for unused memory
--FLOAT=size Size of float type
--GETOPTION=argument Get advanced options
--HELP=option Help
--HTML=file Generate HTML debug files
--LANG=language Specify language
--MEMMAP=mapfile Display memory map

 2012-2013 Microchip Technology Inc. DS50002053D-page 97

MPLAB® XC8 C Compiler User’s Guide
TABLE 4-6: DRIVER OPTIONS (CONTINUED)
Option Meaning
--MODE=mode Choose operating mode
--MSGDISABLE=list Disable warning messages
--MSGFORMAT=specification Set advisory message format
--NODEL Do not remove temporary files
--NOEXEC Do not execute compiler applications
--OBJDIR=path Set object files directory
--OPT=optimizations Control optimization
--OUTDIR=path Set output directory
--OUTPUT=path Set output formats
--PARSER=mode Specify parser mode
--PASS1 Produce intermediate p-code file and stop
--PRE Produce preprocessed source files and stop
--PROTO Generate function prototypes
--RAM=ranges Adjust RAM ranges
--ROM=ranges Adjust ROM ranges
--RUNTIME=options Specify runtime options
--SCANDEP Scan for dependencies
--SERIAL=specification Insert a hexadecimal code or serial number
--SETOPTION=argument Set advanced options
--SETUP=specification Setup the compiler
--SHROUD Shroud (obfuscate) generated p-code files
--STACK=type[:sizes] Specify data stack type and sizes
--STRICT Use strict ANSI keywords
--SUMMARY=type Summary options
--TIME Report compilation times
--VER Show version information
--WARN=number Set warning threshold level
--WARNFORMAT=specification Set warning format

4.8.1 -C: Compile to Object File

The -C option is used to halt compilation after executing the assembler, leaving a relo-
catable object file as the output. It is frequently used when compiling assembly source
files using a make utility. It cannot be used unless all C source files are present on the
command line. Use --PASS1 to generate intermediate files from C source, see
Section 4.8.45 “--PARSER: Specify Parser Mode”.
See Section 4.3.3 “Multi-Step Compilation”, for more information on generating and
using intermediate files.

DS50002053D-page 98  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.2 -D: Define Macro

The -D option is used to define a preprocessor macro on the command line, exactly as
if it had been defined using a #define directive in the source code. This option can
take one of two forms, -Dmacro which is equivalent to:
#define macro 1
placed at the top of each module compiled using this option, or -Dmacro= text which
is equivalent to:
#define macro text
where text is the textual substitution required. Thus, the command:
xc8 --CHIP=16F877AA -Ddebug -Dbuffers=10 test.c
will compile test.c with macros defined exactly as if the C source code had included
the directives:
#define debug 1
#define buffers 10
Defining macros as C string literals requires bypassing any interpretation issues in the
operating system that is being used. To pass the C string, "hello world", (including
the quote characters) in the Windows environment, use: "-DMY_STRING=\\\"hello
world\\\"" (you must include the quote characters around the entire option, as there
is a space character in the macro definition). Under Linux or Mac OS X, use:
-DMY_STRING=\"hello\ world\".
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.3 -E: Redirect Compiler Errors to a File

This option has two purposes. The first is to change the format of displayed messages.
The second is to optionally allow messages to be directed to a file, as some editors do
not allow the standard command line redirection facilities to be used when invoking the
compiler.
The general form of messages produced with the -E option in force is:
filename line_number: (message number) message string (type)
If a filename is specified immediately after -E, it is treated as the name of a file to which
all messages (errors, warnings, etc.) will be printed. For example, to compile x.c and
redirect all errors to [Link], use the command:
xc8 --CHIP=16F877AA -[Link] x.c
The -E option also allows errors to be appended to an existing file by specifying an
addition character, +, at the start of the error filename, for example:
xc8 --CHIP=16F877AA -E+[Link] y.c
If you wish to compile several files and combine all of the errors generated into a single
text file, use the -E option to create the file then use -E+ when compiling all the other
source files. For example, to compile a number of files with all errors combined into a
file called [Link], you could use the - E option as follows:
xc8 --CHIP=16F877AA -[Link] -O --PASS1 main.c
xc8 --CHIP=16F877AA -E+[Link] -O --PASS1 part1.c
xc8 --CHIP=16F877AA -E+[Link] -C [Link]
Section 4.6 “Compiler Messages” has more information regarding this option as well
as an overview of the messaging system and other related driver options.

 2012-2013 Microchip Technology Inc. DS50002053D-page 99

MPLAB® XC8 C Compiler User’s Guide

4.8.4 -I: Include Search Path

Use -I to specify an additional directory to search for header files which have been
included using the #include directive. The directory can either be an absolute or rel-
ative path. The -I option can be used more than once if multiple directories are to be
searched.
The compiler’s include directory containing all standard header files is always
searched, even if no -I option is present. If header filenames are specified using quote
characters rather than angle brackets, as in #include "lcd.h", then the current
working directory is searched in addition to the compiler’s include directory. Note that
if compiling within MPLAB IDE, the search path is relative to the output directory, not
the project directory.
These default search paths are searched after any user-specified directories have
been searched. For example, the following code:
xc8 --CHIP=16F877AA -C -Ic:\include -Id:\myapp\include test.c
will search the directories c:\include and d:\myapp\include for any header files
included into the source code, then search the default include directory.
Under Windows OS, be aware that use of the directory backslash character may unin-
tentionally form an escape sequence. For example, to specify an include file path that
ends with a directory separator character and which is quoted, use -I"E:\\" instead
of -I"E:\", to avoid the escape sequence \". Note that MPLAB X IDE will quote any
include file path you specify in the project properties.
This option has no effect for files that are included into assembly source using the
assembly INCLUDE directive. See Section [Link] “INCLUDE”, for details.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.5 -L: Scan Library

The -L option is used to specify additional libraries that are to be scanned by the linker.
Libraries specified using the -L option are scanned before the standard C library, allow-
ing additional versions of standard library functions to be accessed.
The argument to -L is a library keyword to which the prefix pic; numbers representing
the device range, number of ROM pages and the number of RAM banks; and the suffix
.lib are added.
In this way, the option -Ll, when compiling for a 16F877A, will, for example, scan the
library [Link] and the option -Lxx will scan a library called
[Link].
All libraries must be located in the lib directory of the compiler installation directory.
As indicated, the argument to the -L option is not a complete library filename. If you
wish the linker to scan libraries whose names do not follow the naming convention pre-
viously mentioned or whose locations are not in the lib subdirectory, simply include
the libraries’ names on the command line along with your source files, or add these to
your project.

DS50002053D-page 100  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.6 -L-: Adjust Linker Options Directly

The -L driver option can be used to specify an option that will be passed directly to the
linker. If -L is followed immediately by text starting with a dash character “-”, the text
will be passed directly to the linker without being interpreted by the xc8 command-line
driver. In the event that the -L option is not followed immediately by a dash character,
it is assumed the option is the library scan option; see Section 4.8.5 “-L: Scan
Library” for more information.
For example, if the option -L-N is specified, the -N option will be passed on to the linker
without any subsequent interpretation by the driver. The linker will then process this
option, when, and if, it is invoked, and perform the appropriate operation.
Take care with command-line options. The linker cannot interpret command-line driver
options; similarly, the driver cannot interpret linker options. In most situations, it is
always the command-line driver, xc8, that is being executed. If you need to add alter-
nate linker settings in the Linker category of the Project Properties dialog, you must add
driver options (not linker options). These driver options will be used by the driver to gen-
erate the appropriate linker options during the linking process. The -L option is a
means of allowing a linker option to be specified via a driver option.
The -L option is especially useful when linking code which contains non-standard pro-
gram sections (or psects), as can be the case if the program contains hand-written
assembly code which contains user-defined psects (see [Link] “PSECT”, and
Section 5.15.1 “Program Sections”), or C code which uses the #pragma psect
directive (see [Link] “The #pragma psect Directive”). Without this -L option, it
would be necessary to invoke the linker manually to allow the linker options to be
adjusted.
This option can also be used to replace default linker options. If the string starting from
the first character after the -L option, up to the first equal character, “=”, matches a
psect or class name in the default options, then (the reference to the psect or class
name in the default option, and the remainder of that option, are deleted) that default
linker option is replaced by the option specified by the -L. For example, if a default
linker option was:
-preset_vec=00h,intentry,init,end_init
the driver option -L-pinit=100h would result in the following options being passed
to the linker: -pinit=100h -preset_vec=00h. Note that the end_init linker
option has been removed entirely. If there are no characters following the first equal
character in the -L option, then no replacement will be made for the default linker
options that will be deleted. For example, the driver option -L-pinit= will adjust the
default options passed to the linker, as above; but, the -pinit linker option would be
removed entirely.
No warning is generated if such a default linker option cannot be found. The default
option that you are deleting or replacing must contain an equal character.

 2012-2013 Microchip Technology Inc. DS50002053D-page 101

MPLAB® XC8 C Compiler User’s Guide

4.8.7 -M: Generate Map File

The -M option is used to request the generation of a map file. The map file is generated
by the linker and includes detailed information about where objects are located in
memory. See Section 7.4 “Map Files” for information regarding the content of these
files.
If no filename is specified with the option, then the name of the map file will have the
project name (see Section 4.3 “The Compilation Sequence”), with the extension
.map.
This option is on by default when compiling from within MPLAB X IDE and using the
Universal Toolsuite.

4.8.8 -N: Identifier Length

This option allows the significant C identifier length (used by functions and variables)
to be decreased from the default value of 255. Valid sizes for this option are from 32 to
255. The option has no effect for all other values.
This option also controls the significant length of identifiers used by the preprocessor,
such as macro names. The default length is also 255, and can be adjusted to a
minimum of 31.
If the --STRICT option is used, the default significant identifier length is reduced to 31.
Code that uses a longer identifier length will be less portable.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.9 -O: Specify Output File

This option allows the base name of the output file(s) to be specified. If no -O option is
given, the base name of output file(s) will be the same as the project name, see
Section 4.3 “The Compilation Sequence”. The files whose names are affected by
this option are those files that are not directly associated with any particular source file,
such as the HEX file, MAP file and SYM file.
The -O option can also change the directory in which the output file is located by includ-
ing the required path before the filename. This will then also specify the output directory
for any files produced by the linker or subsequently run applications. Any relative paths
specified are with respect to the current working directory.
For example, if the option -Oc:\project\output\first is used, the MAP and
HEX file, etc., will use the base name first, and will be placed in the directory
c:\project\output.
Any extension supplied with the filename will be ignored.
If a path is specified with the option that enables MAP file creation, -M, (see
Section 4.8.7 “-M: Generate Map File”), this overrides any name or path information
provided by -O.
To change the directory in which all output and intermediate files are written, use the
--OUTDIR option; see Section Section 4.8.43 “--OUTDIR: Specify a Directory for
Output Files”. Note that if -O specifies a path that is inconsistent with the path
specified in the --OUTDIR option, it will result in an error.

DS50002053D-page 102  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.10 -P: Preprocess Assembly Files

The -P option causes assembler source files to be preprocessed before they are
assembled, thus allowing the use of preprocessor directives, such as #include, and
C-style comments with assembler code.
By default, assembler files are not preprocessed.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.11 -Q: Quiet Mode

This option places the compiler in a quiet mode that suppresses the Microchip
Technology Incorporated copyright notice from being displayed.

4.8.12 -S: Compile to Assembler Code

The -S option stops compilation after generating an assembly output file. One
assembly file will be generated for all the C source code, including p-code library code.
The command:
xc8 --CHIP=16F877A -S test.c
will produce an assembly file called [Link], which contains the assembly code gen-
erated from test.c. The generated file is valid assembly code that could be passed
to xc8 as a source file, however this should only be done for exploratory reasons. To
take advantage of the benefits of the compilation technology in the compiler, it must
compile and link all the C source code in a single step. See the --PASS1 option
(Section 4.8.45 “--PARSER: Specify Parser Mode”) to generate intermediate files if
you wish to compile code using a two-step process or use intermediate files.
This option is useful for checking assembly code output by the compiler. The file pro-
duced by this option differs to that produced by the --ASMLIST option (see
Section 4.8.15 “--ADDRQUAL: Set Compiler Response to Memory Qualifiers”) in
that it does not contain op-codes or addresses and it can be used as a source file in
subsequent compilations. The assembly list file is more human readable, but is not a
valid assembly source file.

4.8.13 -U: Undefine a Macro

The -U option, the inverse of the -D option, is used to undefine predefined macros.
This option takes the form -Umacro, where macro is the name of the macro to be
undefined
The option, -Udraft, for example, is equivalent to:
#undef draft
placed at the top of each module compiled using this option.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

 2012-2013 Microchip Technology Inc. DS50002053D-page 103

MPLAB® XC8 C Compiler User’s Guide

4.8.14 -V: Verbose Compile

The -V option specifies verbose compilation. When used, the compiler will display the
command lines used to invoke each of the compiler applications described in
Section 4.3 “The Compilation Sequence”.
The name of the compiler application being executed will be displayed, plus all the
command-line arguments to this application. This option is useful for confirming options
and files names passed to the compiler applications.
If this option is used twice (-V -V), it will display the full path to each compiler applica-
tion as well as the full command-line arguments. This would be useful to ensure that
the correct compiler installation is being executed, if there is more than one compiler
installed.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.15 --ADDRQUAL: Set Compiler Response to Memory Qualifiers

The --ADDRQUAL option indicates the compiler’s response to non-standard memory
qualifiers in C source code.
By default, these qualifiers are ignored; i.e., they are accepted without error, but have
no effect. Using this option allows these qualifiers to be interpreted differently by the
compiler.
The near qualifier is affected by this option. On PIC18 devices, this option also affects
the far qualifier; and for other 8-bit devices, the bankx qualifiers (bank0, bank1,
bank2, etc.) are affected.
The suboptions are detailed in Table 4-7.

TABLE 4-7: COMPILER RESPONSES TO MEMORY QUALIFIERS

Selection Response
require The qualifiers will be honored. If they cannot be met, an error will be
issued.
request The qualifiers will be honored, if possible. No error will be generated if they
cannot be followed.
ignore The qualifiers will be ignored and code compiled as if they were not used.
reject If the qualifiers are encountered, an error will be immediately generated.

For example, when using the option --ADDRQUAL=request the compiler will try to
honor any non-standard qualifiers, but silently ignore them if they cannot be met.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

DS50002053D-page 104  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.16 --ASMLIST: Generate Assembler List Files

The --ASMLIST option tells xc8 to generate assembler listing files for the C and
assembly source modules being compiled. One assembly list file is produced for the
entire C program, including code from the C library functions.
Additionally, one assembly list file is produced for each assembly source file in the
project, including the runtime startup code. For more information, see
Section 4.4.2 “Startup and Initialization”.
Assembly list files use a .lst extension and, due to the additional information placed
in these files, cannot be used as assembly source files.
In the case of listings for C source code, the list file shows both the original C code and
the corresponding assembly code generated by the compiler. See
Section 6.5 “Assembly-Level Optimizations”, for full information regarding the
content of these files.
The same information is shown in the list files for assembly source code.
This option is on by default when compiling under MPLAB IDE.

 2012-2013 Microchip Technology Inc. DS50002053D-page 105

MPLAB® XC8 C Compiler User’s Guide

4.8.17 --CHECKSUM: Calculate a Checksum

This option will perform a checksum over the address range specified and store the
result at the destination address specified. The general form of this option is as follows.
-CHECKSUM=start-end@destination[,offset=][,width=w][,code=c][,algorithm=a]
Additional specifications are appended as a comma-separated list to this option. Such
specifications are:
width=n selects the width of the checksum result in bytes for non-Fletcher algo-
rithms. A negative width will store the result in little-endian byte order. Result
widths from one to four bytes are permitted.
offset=nnnn specifies an initial value or offset to be added to this checksum.
algorithm=n selects one of the checksum algorithms implemented in HEXMATE. The
selectable algorithms are described in Table 8-4.
code=nn is a hexadecimal code that will trail each byte in the checksum result. This
can allow each byte of the checksum result to be embedded within an
instruction.
The start, end and destination attributes are, by default, hexadecimal constants.
If an accompanying --FILL option has not been specified, unused locations within the
specified address range will be filled with FFFh for baseline devices, 3FFFh for
mid-range devices, or FFFF for PIC18 devices. This is to remove any unknown values
from the equation and ensure the accuracy of the checksum result.
For example:
--checksum=800-fff@20,width=1,algorithm=2
will calculate a 1 byte checksum from address 0x800 to 0xfff and store this at address
0x20. A 16-bit addition algorithm will be used. See Table 4-8, for the available
algorithms.

TABLE 4-8: CHECKSUM ALGORITHM SELECTION

Selector Algorithm description
-4 Subtraction of 32 bit values from initial value
-3 Subtraction of 24 bit values from initial value
-2 Subtraction of 16 bit values from initial value
-1 Subtraction of 8 bit values from initial value
1 Addition of 8 bit values from initial value
2 Addition of 16 bit values from initial value
3 Addition of 24 bit values from initial value
4 Addition of 32 bit values from initial value
7 Fletcher’s checksum (8 bit calculation, 2-byte result width)
8 Fletcher’s checksum (16 bit calculation, 4-byte result width)

The checksum calculations are performed by the HEXMATE application. The informa-
tion in this driver option is passed to the HEXMATE application when it is executed.

4.8.18 --CHIP: Define Device

This option must be used to specify the target device, or device, for the compilation.
This is the only compiler option that is mandatory when compiling code.
To see a list of supported devices that can be used with this option, use the
--CHIPINFO option described in Section 4.8.19 “--CHIPINFO: Display List of Sup-
ported Devices”.

DS50002053D-page 106  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.19 --CHIPINFO: Display List of Supported Devices

The --CHIPINFO option displays a list of devices the compiler supports. The names
listed are those chips that are defined in the chipinfo file and which can be used with
the --CHIP option.
Compiler execution will terminate after this list has been printed.

4.8.20 --CLIST: Generate C Listing File

Use of this option will generate a C listing file for each C source file specified on the
command line.
The listing files produced consist of the original C source code prepended with a line
number. Do not confuse these files with assembly list files; see
Section 4.8.16 “--ASMLIST: Generate Assembler List Files”.

4.8.21 --CODEOFFSET: Offset Program Code to Address

In some circumstances, such as bootloaders, it is necessary to shift the program image
to an alternative address. This option is used to specify a base address for the program
code image and to reserve memory from address 0 to that specified in the option.
When using this option, all code psects (including Reset and interrupt vectors and con-
stant data) will be adjusted to the address specified. The address is assumed to be a
hexadecimal constant. A leading 0x, or a trailing h hexadecimal specifier can be used,
but is not necessary.
This option differs from the --ROM option in that it will move the code associated with
the Reset and interrupt vectors. That cannot be done using the --ROM option; see
Section 4.8.50 “--ROM: Adjust ROM Ranges”.
For example, if the option --CODEOFFSET=600 is specified, the Reset vector will be
moved from address 0 to address 0x600; the interrupt vector will be moved from
address 4 to 0x604, in the case of mid-range PIC devices, or to the addresses 0x608
and 0x618 for PIC18 devices. No code will be placed between address 0 and 0x600.
As the Reset and interrupt vector locations are fixed by the PIC device, it is the pro-
grammer’s responsibility to ensure code that can redirect control to the offset Reset
and interrupt routines is written and located at the original locations.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

 2012-2013 Microchip Technology Inc. DS50002053D-page 107

MPLAB® XC8 C Compiler User’s Guide

4.8.22 --DEBUGGER: Select Debugger Type

This option is intended for use for compatibility with development tools that can act as
a debugger. xc8 supports several debuggers and using this option will configure the
compiler to conform to the requirements of that selected. The possible selections for
this option are defined in Table 4-9.

TABLE 4-9: SELECTABLE DEBUGGERS

Suboption Debugger selected
none No debugger (default)
icd2 MPLAB® ICD 2
icd3 MPLAB ICD 3
pickit2 PICkit™ 2
pickit3 PICkit 3
realice MPLAB REAL ICE™ in-circuit emulator

For example:
xc8 --CHIP=16F877AA --DEBUGGER=icd2 main.c
Choosing the correct debugger is important as they can use memory resources that
could otherwise be used by the project if this option is omitted. Such a conflict would
lead to runtime failure.
If the debugging features of the development tool are not to be used (for example, if the
MPLAB ICD 3 is only being used as a programmer), then the debugger option can be
set to none, because memory resources are not being used by the tool.
MPLAB X IDE will automatically apply this option for debug builds once you have
indicated the hardware tool in the Project Preferences.

4.8.23 --DOUBLE: Select Kind of Double Types

This option allows the kind of double-precision floating-point types to be selected. By
default, the compiler will choose the truncated IEEE754 24-bit implementation for
double types. With this option, it can be changed to the full 32-bit IEEE754
implementation.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.24 --ECHO: Echo Command Line Before Processing

Use of this option will result in the driver command line being echoed to the stdout
stream before compilation commences. Each token of the command line will be printed
on a separate line and they will appear in the order in which they are placed on the
command line.

DS50002053D-page 108  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.25 --EMI: Select External Memory Interface Operating Mode

The external memory interface available on some PIC18 devices can be operated in
several modes. The interface can operate in 16-bit modes; Word-write and Byte-select
modes or in an 8-bit mode: Byte-write mode. These modes are represented by those
specified in Table 4-10.

TABLE 4-10: EXTERNAL MEMORY INTERFACE MODES

Mode Operation
wordwrite 16-bit Word-write mode (default)
byteselect 16-bit Byte-select mode
bytewrite 8-bit Byte-write mode

The selected mode will affect the code generated when writing to the external data
interface. In word write mode, dummy reads and writes can be added to ensure that an
even number of bytes are always written. In Byte-select or Byte-write modes, dummy
reads and writes are not generated and can result in more efficient code.
Note that this option does not pre-configure the device for operation in the selected
mode. See your device data sheet for the registers and settings that are used to
configure the device’s external interface mode.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

 2012-2013 Microchip Technology Inc. DS50002053D-page 109

MPLAB® XC8 C Compiler User’s Guide

4.8.26 --ERRATA: Specify Errata Workarounds

This option allows specification of software workarounds to documented silicon errata
issues. A default set of errata issues apply to each device, but this set can be adjusted
by using this option and the arguments presented in Table 4-11.

TABLE 4-11: ERRATA WORKAROUNDS

Bit
Symbol Workaround
pos.
4000 0 Program memory accesses/jumps across 4000h address boundary
fastints 1 Fast interrupt shadow registers corruption
lfsr 2 Broken LFSR instruction
minus40 3 Program memory reads at -40 degrees
reset 4 GOTO instruction cannot exist at Reset vector
bsr15 5 Flag problems when BSR holds value 15
daw 6 Broken DAW instruction
eedatard 7 Read EEDAT in immediate instruction after RD set
eeadr 8 Don't set RD bit immediately after loading EEADR
ee_lvd 9 LVD must stabilize before writing EEPROM
fl_lvd 10 LVD must stabilize before writing Flash
tblwtint 11 Clear interrupt registers before tblwt instruction
fw4000 12 Flash write exe must act on opposite side of 4000h boundary
resetram 13 RAM contents can corrupt if async. Reset occurs during write access
fetch 14 Corruptible instruction fetch. – apply FFFFh (NOP) at required locations
clocksw 15 Code corruption if switching to external oscillator clock source – apply
switch to HFINTOSC high-power mode first

At present, workarounds are mainly employed for PIC18 devices, but the clocksw
errata is only applicable for some enhanced mid-range devices.
To disable all software workarounds, use the following.
--ERRATA=none
To apply the default set of workarounds, but to specifically disable the jump across
4000 errata, for example, use the following.
--ERRATA=default,-4000
A preprocessor macro ERRATA_TYPES (see Section 5.14.3 “Predefined Macros”) is
set to a value to indicate the errata applied. Each errata listed in Table 4-11 represents
a bit position in the macro’s value, with the topmost errata in the table being the least
significant. The bit position is indicated in the table and is set if the corresponding errata
is applied. That is, if the 4000, reset and bsr15 errata were applied, the value
assigned to the ERRATA_TYPES macro would be 0x31.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

DS50002053D-page 110  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.27 --ERRFORMAT: Define Format for Compiler Messages

If the --ERRFORMAT option is not used, the default behavior of the compiler is to dis-
play any errors in a “human readable” form. This standard format is perfectly accept-
able to a person reading the error output, but is not generally usable with environments
that support compiler error handling.
This option allows the exact format of printed error messages to be specified using spe-
cial placeholders embedded within a message template. See Section 4.6 “Compiler
Messages” for full details of the messaging system employed by xc8, and the
placeholders which can be used with this option.
This section is also applicable to the --WARNFORMAT and --MSGFORMAT options,
which adjust the format of warning and advisory messages, respectively.
If you are compiling using MPLAB IDE, the format of the compiler messages is auto-
matically configured to what the IDE expects. It recommended that you do not adjust
the message formats if compiling using this IDE.

4.8.28 --ERRORS: Maximum Number of Errors

This option sets the maximum number of errors each compiler application, as well as
the driver, will display before execution is terminated. By default, up to 20 error
messages will be displayed by each application.
See Section 4.6 “Compiler Messages” for full details of the messaging system
employed by xc8.

4.8.29 --EXT: Specify C Language Extensions

The compiler can accept several different sets of non-standard C language extensions.
The suboption to --EXT indicates the set and these are shown in Table 4-12.

TABLE 4-12: ACCEPTABLE C LANGUAGE EXTENSIONS

Suboption Meaning
xc8 The native XC8 extensions (default)
cci A common C interface acceptable by all MPLAB XC compilers
iar Extensions defined by the IAR C/C++ Compiler for ARM

Enabling the cci suboption requests the compiler to check all source code and com-
piler options for compliance with the Common C Interface (CCI). Code that complies
with this interface is portable across all MPLAB XC compilers. Code or options that do
not conform to the CCI will be flagged by compiler warnings. See Chapter
2. “Common C Interface”, for information on the features that are covered by this
interface.
The iar suboption enables conformance with the non-standard extensions supported
by the IAR C/C++ Compiler™ for ARM. This is discussed further in
Appendix B. “Embedded Compiler Compatibility Mode”.

 2012-2013 Microchip Technology Inc. DS50002053D-page 111

MPLAB® XC8 C Compiler User’s Guide

4.8.30 --FILL: Fill Unused Program Memory

This option allows specification of a hexadecimal opcode that can be used to fill all
unused program memory locations. This option utilizes the features of the HEXMATE
application, so it is only available when producing a HEX output file, which is the default
operation.
This driver feature allows you to compile and fill unused locations in one step. If you
prefer not to use the driver option and you prefer to fill unused locations after compila-
tion, then you need to use the HEXMATE application. Note that the corresponding option
in HEXMATE is -FILL (one leading dash) as opposed to the drivers --FILL option.
Note, also, that the unused tag that can be specified with the driver option does not
exist in the HEXMATE options.
The usage of the driver option is:
--FILL=[const_width:]fill_expr[@address[:end_address]]
where:
• const_width has the form wn and signifies the width (n bytes) of each constant
in fill_expr. If const_width is not specified, the default value is the native
width of the architecture. That is, --FILL=w1:1 with fill every byte with the value
0x01.
• fill_expr can use the syntax (where const and increment are n-byte
constants):
- const fill memory with a repeating constant; i.e., --FILL=0xBEEF becomes
0xBEEF, 0xBEEF, 0xBEEF, 0xBEEF
- const+=increment fill memory with an incrementing constant; i.e.,
--fill=0xBEEF+=1 becomes 0xBEEF, 0xBEF0, 0xBEF1, 0xBEF2
- const-=increment fill memory with a decrementing constant; i.e.,
--fill=0xBEEF-=0x10 becomes 0xBEEF, 0xBEDF, 0xBECF, 0xBEBF
- const,const,...,const fill memory with a list of repeating constants; i.e.,
--FILL=0xDEAD,0xBEEF becomes 0xDEAD,0xBEEF,0xDEAD,0xBEEF
• The options following fill_expr result in the following behavior:
- @unused (or nothing) fill all unused memory with fill_expr; i.e.,
--FILL=0xBEEF@unused fills all unused memory with 0xBEEF. The driver
will expand this to the appropriate ranges and pass these to HEXMATE.
- @address fill a specific address with fill_expr; i.e.,
--FILL=0xBEEF@0x1000 puts 0xBEEF at address 1000h
- @address:end_address fill a range of memory with fill_expr; i.e.,
--FILL=0xBEEF@0:0xFF puts 0xBEEF in unused addresses between 0 and
255
All constants can be expressed in (unsigned) binary, octal, decimal or hexadecimal, as
per normal C syntax. For example, 1234 is a decimal value, 0xFF00 is hexadecimal,
and FF00 is illegal.
See Section 4.9 “MPLAB X Option Equivalents”, or Section 4.9 “MPLAB X Option
Equivalents”, for information on using this option in MPLAB IDE.

DS50002053D-page 112  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.31 --FLOAT: Select Kind of Float Types

This option allows the size of float types to be selected. The types available to be
selected are given in Table 4-13.
See also, the --DOUBLE option in Section 4.8.23 “--DOUBLE: Select Kind of Dou-
ble Types”.

TABLE 4-13: FLOATING-POINT SELECTIONS

Suboption Effect
double Size of float matches size of double type
24 24-bit float (default)
32 32-bit float (IEEE754 format)

See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.32 --GETOPTION: Get Command-line Options

This option is used to retrieve the command line options that are used for named com-
piler application. The options are then saved into the given file. This option is not
required for most projects, and is disabled when the compiler is operating in Free mode
(see Section 4.8.37 “--MODE: Choose Compiler Operating Mode”).
The options take an application name and a filename to store the options, for example:
--GETOPTION=hlink,[Link]

4.8.33 --HELP: Display Help

This option displays information on the xc8 compiler options. The option --HELP will
display all options available. To find out more about a particular option, use the option’s
name as a parameter. For example:
xc8 --help=warn
will display more detailed information about the --WARN option, the available
suboptions, and which suboptions are enabled by default.

4.8.34 --HTML: Generate HTML Diagnostic Files

This option will generate a series of HTML files that can be used to explore the compi-
lation results of the latest build. The files are stored in a directory called html, located
in the output directory. The top-level file (which can be opened with your favorite web
browser) is called [Link].
Use this option at all stages of compilation to ensure files associated with all
compilation stages are generated.
The index page is a graphical representation of the compilation process. Each file icon
is clickable and will open to show the contents of that file. This is available for all inter-
mediate files, and even binary files will open in a human-readable form. Each applica-
tion icon can also be clicked to show a page that contains information about the
application’s options and build results.
The list of all preprocessor macros (preprocessor icon) and a graphical memory usage
map (Linker icon) provide information that is not otherwise readily accessible.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

 2012-2013 Microchip Technology Inc. DS50002053D-page 113

MPLAB® XC8 C Compiler User’s Guide

4.8.35 --LANG: Specify the Language for Messages

This option allows the compiler to be configured to produce error, warning and some
advisory messages in languages other than English.
English is the default language unless this has been changed at installation, or by the
use of the --SETUP option. Some messages are only ever printed in English regard-
less of the language specified with this option. For more information, see
Section 4.6.2 “Message Language”.
Table 4-14 shows those languages currently supported.

TABLE 4-14: SUPPORTED LANGUAGES

Suboption Language
en, english English (default)
fr, french, francais French
de, german, deutsch German

4.8.36 --MEMMAP: Display Memory Map

This option will display a memory map for the map file specified with this option. The
information printed is controlled by the --SUMMARY option, see
Section 4.8.58 “--SUMMARY: Select Memory Summary Output Type”, for exam-
ple:
xc8 --memmap=[Link] --summary=psect,class,file
This option is seldom required, but would be useful if the linker is being driven explicitly;
i.e., instead of in the normal way through the command-line driver. This command
would display the memory summary that is normally produced at the end of compilation
by the driver.

4.8.37 --MODE: Choose Compiler Operating Mode

This option selects the basic operating mode of the compiler. The available types are
pro, std, and free. For legacy projects, the mode lite is accepted to mean the Free
operating mode.
A compiler operating in PRO mode uses full optimization and produces the smallest
code size. Standard mode uses limited optimizations, and Free mode only uses a min-
imum optimization level and will produce relatively large code. See
Section 5.13 “Optimizations” and Section 6.5 “Assembly-Level Optimizations”,
for more information on which optimizations are available in each mode.
Only those modes permitted by the compiler license status will be accepted. For exam-
ple if you have purchased a Standard compiler license, that compiler can be run in
Standard or Free mode, but not the PRO mode.
See Section 4.9 “MPLAB X Option Equivalents” for use of this option in MPLAB
IDE.

DS50002053D-page 114  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.38 --MSGDISABLE: Disable Warning Messages

This option allows error, warning or advisory messages to be disabled during compila-
tion of a project.
The option is passed a comma-separated list of message numbers that are to be dis-
abled. If the number of an error message is specified in this list, it must be followed by
:off; otherwise, it will be ignored. If the message list is specified as 0, then all warn-
ings are disabled.
For full information on the compiler’s messaging system and use of this option; see
Section 4.6 “Compiler Messages”, and; see Section 4.6.5 “Changing Message
Behavior”, for other ways to disable messages.

4.8.39 --MSGFORMAT: Set Advisory Message Format

This option sets the format of advisory messages produced by the compiler. Warning
and error messages are controlled separately by other options. See
Section 4.8.27 “--ERRFORMAT: Define Format for Compiler Messages” and
Section 4.8.62 “--WARNFORMAT: Set Warning Message Format” for information
on changing the format of these sorts of messages.
See Section 4.6 “Compiler Messages” for full information on the compiler’s messag-
ing system.
If you are compiling using MPLAB IDE, the format of the compiler messages is auto-
matically configured to what the IDE expects. It recommended that you do not adjust
the message formats if compiling using this IDE.

4.8.40 --NODEL: Do Not Remove Temporary Files

Specifying --NODEL when building will instruct xc8 not to remove the intermediate and
temporary files that were created during the build process.

4.8.41 --OBJDIR: Specify a Directory for Intermediate Files

This option allows a directory to be nominated in xc8 to locate its intermediate files. If
this option is omitted, intermediate files will be created in the current working directory.
This option will not set the location of output files, instead use --OUTDIR. See
Section 4.8.43 “--OUTDIR: Specify a Directory for Output Files” and
Section 4.8.9 “-O: Specify Output File” for more information.

 2012-2013 Microchip Technology Inc. DS50002053D-page 115

MPLAB® XC8 C Compiler User’s Guide

4.8.42 --OPT: Invoke Compiler Optimizations

The --OPT option allows control of all the compiler optimizers. If this option is not spec-
ified, or it is specified as --OPT=all, the space and asm optimizations are enabled
(see below). Optimizations can be disabled by using --OPT=none, or individual opti-
mizers can be controlled, for example: --OPT=asm will only enable some assembler
optimizations.
Table 4-15 lists the available optimization types.

TABLE 4-15: OPTIMIZATION OPTIONS

Option name Function
asm Select optimizations of assembly code derived from C source (default)
asmfile Select optimizations of assembly source files
debug Favor accurate debugging over optimization
speed Favor optimizations that result in faster code
space Favor optimizations that result in smaller code (default)
all Enable all compiler optimizations
none Do not use any compiler optimizations

Note that different suboptions control assembler optimizations of assembly source files
and intermediate assembly files produced from C source code.
The speed and space suboptions are contradictory. Space optimizations are the
default. If speed and space suboptions are both specified, then speed optimizations
takes precedence. If all optimizations are requested, the space optimization is
enabled. These optimizations affect procedural abstraction, which is performed by the
assembler, and other optimizations at the code generation stage.
The asmfile selection optimizes assembly source files, which are otherwise not opti-
mized by the compiler. By contrast, the asm control allows for optimization of assembly
code that was derived from C code, an optimization that is enabled by default.
Some compiler optimizations can affect the ability to debug code. Enabling the debug
suboption can restrict some optimizations that would otherwise take place, and which
would affect debugging.
See Section 4.9 “MPLAB X Option Equivalents” for use of this option in MPLAB
IDE.

DS50002053D-page 116  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.43 --OUTDIR: Specify a Directory for Output Files

This option allows a directory to be nominated for xc8 to locate its output files. If this
option is omitted, output files will be created in the current working directory. See also
Section 4.8.41 “--OBJDIR: Specify a Directory for Intermediate Files” and
Section 4.8.9 “-O: Specify Output File” for more information.

4.8.44 --OUTPUT= type: Specify Output File Type

This option allows the type of the output file(s) to be specified. If no --OUTPUT option
is specified, the output file’s name will be the same as the project name (see
Section 4.3 “The Compilation Sequence”).
The available output file formats are shown in Table 4-16. More than one output format
can be specified by supplying a comma-separated list of tags. Not all formats are
supported by Microchip development tools.
For debugging, the ELF/DWARF format is preferred, but this format is not supported by
MPLAB IDE v8 or early versions of MPLAB X IDE. Before selecting the ELF file output,
ensure your IDE version has support for this format. Microchip COFF is the default
debugging file output.
Output file types that specify library formats stop the compilation process before the
final stages of compilation are executed. So, specifying an output file format list that
contains, for example: lib or all, will prevent the other formats from being created.

TABLE 4-16: OUTPUT FILE FORMATS

Type tag File format
lib Object library file (for assembly source)
lpp P-code library file (for C source)
intel, inhx32 Intel HEX (default)
inhx032 Intel Hex with initialization of upper extended linear address to
zero
tek Tektronix Hex
aahex American Automation symbolic HEX file
mot, motorola, s19 Motorola S19 HEX file
bin, binary Binary file
mcof, mcoff, mpcoff Microchip COFF (default)
elf ELF/DWARF file

So, for example:

xc8 --CHIP=16F877AA --OUTPUT=lpp lcd_init.c lcd_data.c lcd_msgs.c
will compile the three names files into an LPP (p-code) library.

 2012-2013 Microchip Technology Inc. DS50002053D-page 117

MPLAB® XC8 C Compiler User’s Guide

4.8.45 --PARSER: Specify Parser Mode

This option controls which symbols are stripped from intermediate (p-code) files pro-
duced by the parser. The default is to remove unused symbols, which can also be
specified using the --PARSER=lean mode. The rich mode will not remove any
symbols.
Note that the rich mode will generate larger intermediate files and will considerably
slow down the compilation, particularly if there are many SFRs defined for a device.
PIC18 devices typically have a large number of SFRs. Use the rich setting if you have
in-line assembly code that accesses symbols that are not referenced by C code other-
wise undefined symbol errors can be produced by the assembler application.

4.8.46 --PASS1: Compile to P-code

The --PASS1 option is used to generate p-code intermediate files (.p1 files) from the
parser, and then stop compilation. Such files need to be generated if creating p-code
library files, however the compiler is able to generate library files in one step, if required.
See Section 4.8.44 “--OUTPUT= type: Specify Output File Type” for specifying a
library output file type.)

4.8.47 --PRE: Produce Preprocessed Source Code

The --PRE option is used to generate preprocessed C source files (also called mod-
ules or translation units) with an extension .pre. This can be useful to ensure that pre-
processor macros have expanded to what you think they should. Use of this option can
also create C source files that do not require any separate header files. If the .pre files
are renamed to .c files, they can be passed to the compiler for subsequent processing.
This is useful when sending files to a colleague or to obtain technical support without
sending all the header files, which can reside in many directories.
If you wish to see the preprocessed source for the printf() family of functions, do
not use this option. The source for this function is customized by the compiler, but only
after the code generator has scanned the project for printf() usage. Thus, as the
–-PRE option stops compilation after the preprocessor stage, the code generator will
not execute and no printf() code will be processed. If this option is omitted, the
preprocessed source for printf() will be automatically retained in the file
[Link].

DS50002053D-page 118  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.48 --PROTO: Generate Prototypes

The --PROTO option is used to generate .pro files containing both ANSI C and K&R
style function declarations for all functions within the specified source files. Each .pro
file produced will have the same base name as the corresponding source file. Proto-
type files contain both ANSI C-style prototypes and old-style C function declarations
within conditional compilation blocks.
The extern declarations from each .pro file should be edited into a global header file,
which can then be included into all the C source files in the project. The .pro files can
also contain static declarations for functions that are local to a source file. These
static declarations should be edited into the start of the source file.
To demonstrate the operation of the --PROTO option, enter the following source code
as file test.c:
#include <stdio.h>
add(arg1, arg2)
int * arg1;
int * arg2;
{
return *arg1 + *arg2;
}

void printlist(int * list, int count)

{
while (count--)
printf("d " *list++);
putchar(’\n’);
}
If compiled with the command:
xc8 --CHIP=16F877AA --PROTO test.c
xc8 will produce [Link] containing the following declarations, which can then be
edited as necessary:
/* Prototypes from test.c */
/* extern functions - include these in a header file */
#if PROTOTYPES
extern int add(int *, int *);
extern void printlist(int *, int);
#else /* PROTOTYPES */
extern int add();
extern void printlist();
#endif /* PROTOTYPES */

 2012-2013 Microchip Technology Inc. DS50002053D-page 119

MPLAB® XC8 C Compiler User’s Guide

4.8.49 --RAM: Adjust RAM Ranges

This option is used to adjust the default RAM, which is specified for the target device.
The default memory will include all the on-chip RAM specified for the target
PIC10/12/16 device, thus this option only needs be used if there are special memory
requirements. Typically this option is used to reserve memory (reduce the amount of
memory available). Specifying additional memory that is not in the target device will
typically result in a successful compilation, but can lead to code failures at runtime.
The default RAM memory for each target device is specified in the chipinfo file,
[Link].
Strictly speaking, this option specifies the areas of memory that can be used by writable
(RAM-based) objects; but, not necessarily those areas of memory that contain physical
RAM. The output that will be placed in the ranges specified by this option are typically
variables that a program defines.
For example, to specify an additional range of memory to that already present on-chip,
use:
--RAM=default,+100-1ff
This will add the range from 100h to 1ffh to the on-chip memory. To only use an external
range and ignore any on-chip memory, use:
--RAM=0-ff
This option can also be used to reserve memory ranges already defined as on-chip
memory in the chipinfo file. To do this, supply a range prefixed with a minus character,
-, for example:
--RAM=default,-100-103
will use all the defined on-chip memory, but not use the addresses in the range from
100h to 103h for allocation of RAM objects.
This option is also used to specify RAM for far objects on PIC18 devices. These
objects are stored in the PIC18 extended memory. Any additional memory specified
with this option whose address is above the on-chip program memory is assumed to
be extended memory implemented as RAM.
For example, to indicate that RAM has been implemented in the extended memory
space at addresses 0x20000 to 0x20fff, use the following option.
--RAM=default,+20000-20fff
This option will adjust the memory ranges used by linker classes; see
Section 7.2.1 “-Aclass =low-high,...”; so, any object that is in a psect is placed in this
class. Any objects contained in a psect that are explicitly placed at a memory address
by the linker (see Section 7.2.19 “-Pspec”) i.e., are not placed into a memory class,
are not affected by the option. For an introductory guide to psects, see
Section 5.15.1 “Program Sections”.
See Section 4.9 “MPLAB X Option Equivalents” for use of this option in MPLAB
IDE.

DS50002053D-page 120  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.50 --ROM: Adjust ROM Ranges

This option is used to change the default ROM that is specified for the target device.
The default memory will include all the on-chip ROM specified for the target
PIC10/12/16 device, thus this option only needs to be used if there are special memory
requirements. Typically this option is used to reserve memory (reduce the amount of
memory available). Specifying additional memory that is not in the target device will
typically result in a successful compilation, but can lead to code failures at runtime.
The default ROM memory for each target device is specified in the chipinfo file,
[Link].
Strictly speaking, this option specifies the areas of memory that can be used by
read-only (ROM-based) objects; but, not necessarily those areas of memory that con-
tain physical ROM. When producing code that can be downloaded into a system via a
bootloader, the destination memory can be some sort of (volatile) RAM. The output that
will be placed in the ranges specified by this option are typically executable code and
any data variables that are qualified as const.
For example, to specify an additional range of memory to that on-chip, use:
--ROM=default,+100-2ff
This will add the range from 100h to 2ffh to the on-chip memory. To only use an external
range and ignore any on-chip memory, use:
--ROM=100-2ff
This option can also be used to reserve memory ranges already defined as on-chip
memory in the chip configuration file. To do this supply a range prefixed with a minus
character, -, for example:
--ROM=default,-100-1ff
will use all the defined on-chip memory, but not use the addresses in the range from
100h to 1ffh for allocation of ROM objects.
This option will adjust the memory ranges used by linker classes, see
Section 7.2.1 “-Aclass =low-high,...”; so, any object that is in a psect is placed in this
class. Any objects which are contained in a psect that are explicitly placed at a memory
address by the linker (see Section 7.2.19 “-Pspec”), i.e., are not placed into a mem-
ory class, are not affected by the option. For an introductory guide to psects, see
Section 5.15.1 “Program Sections”.
Note that some psects must be linked above a threshold address, most notably some
psects that hold const data. Using this option to remove the upper memory ranges can
make it impossible to place these psects.
See Section 4.9 “MPLAB X Option Equivalents” for use of this option in MPLAB
IDE.

 2012-2013 Microchip Technology Inc. DS50002053D-page 121

MPLAB® XC8 C Compiler User’s Guide

4.8.51 --RUNTIME: Specify Runtime Environment

The --RUNTIME option is used to control what is included as part of the runtime envi-
ronment. The runtime environment encapsulates any code that is present at runtime
that has not been defined by the user, instead supplied by the compiler, typically as
library code or compiler-generated source files.
All required runtime features are enabled by default and this option is not required for
normal compilation.
Note that the code that clears or initializes variables, which is included by default, will
clobber the contents of the STATUS register. For mid-range and baseline devices, if
you need to check the cause of Reset using the TO or PD bits in this register, then you
must enable the resetbits suboption as well. See Section [Link] “STATUS Reg-
ister Preservation”, for how this feature is used. The usable suboptions include those
shown in Table 4-17.

TABLE 4-17: RUNTIME ENVIRONMENT SUBOPTIONS

Default
Suboption Controls On (+) Implies
State

init the code present in the main program The ROM image is copied into RAM and
module that copies the ROM-image of initialized variables will contain their initial On
initial values to RAM variables value at main().
clib the inclusion of library files into the output Library files are linked into the output.
On
code by the linker.
clear the code present in the main program Uninitialized variables are cleared and will
On
module that clears uninitialized variables contain 0 at main().
config programming the device with default Configuration bits not specified will be
Off
config bytes assigned a default value. (PIC18 only)
download conditioning of the Intel HEX file for use Data records in the Intel HEX file are padded
with bootloaders out to 16-byte lengths and will align on 16-byte
Off
boundaries. Startup code will not assume
Reset values in certain registers.
flp additional code to provide function pro Diagnostic code will be embedded into the
Off
filing output to allow function profiling.
no_startup whether the startup module is linked in Startup module will not be linked in.
Off
with user-defined code
osccal initialize the oscillator with the oscillator Oscillator will be calibrated (PIC10/12/16
On
constant only).
oscval: set the internal clock oscillator calibration Oscillator will be calibrated with value sup-
n/a
value value plied (PIC10/12/16 only).
keep whether the startup module source file The startup module is not deleted.
On
([Link]) is deleted after compilation
plib whether the peripheral library is linked in. The peripheral library will be linked in to the
On
build (PIC18 only).
resetbits Preserve Power-down and Time-out STATUS bits are preserved (PIC10/12/16
Off
STATUS bits at start up only).
stackcall Allow function calls to use a table look-up Functions called via CALL instruction while
method after the hardware stack has the stack was not exhausted, then called via a
filled. (Ignored if reentrant or hybrid look-up table (PIC10/12/16 devices only). Off
function model is used, or if function uses
the reentrant specifier.)
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB IDE.

DS50002053D-page 122  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.8.52 --SCANDEP: Scan for Dependencies

When this option is used, .dep and .d dependency files are generated. The depen-
dency file lists those files on which the source file is Dependant. Dependencies result
when one file is #included into another. The .d file format is used by GCC-based
compilers and it contains the same information as the .dep file.
Compilation will stop after the preprocessing stage if this option is used.

4.8.53 --SERIAL: Store a Value at this Program Memory Address

This option allows a hexadecimal code to be stored at a particular address in program
memory. A typical task for this option might be to position a serial number in program
memory.
The byte-width of data to store is determined by the byte-width of the hexcode param-
eter in the option. For example, to store a one-byte value, 0, at program memory
address 1000h, use --SERIAL=00@1000. To store the same value as a four byte
quantity use --SERIAL=00000000@1000.
This option is functionally identical to the corresponding HEXMATE option. For more
detailed information and advanced controls that can be used with this option, refer to
Section [Link] “-SERIAL”.
The driver will also define a label at the location where the value was stored, and which
can be referenced from C code as _serial0. To enable access to this symbol,
remember to declare it, for example:
extern const int _serial0;
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.54 --SETOPTION: Set the Command-line Options for Application

This option is used to supply alternative command line options for the named
application when compiling. The general form of this option is shown below.
--SETOPTION=app,file
where the app component specifies the application that will receive the new options,
and the file component specifies the name of the file that contains the additional
options that will be passed to the application. This option is not required for most
projects.
If specifying more than one option to a component, each option must be entered on a
new line in the option file. This option can also be used to remove an application from
the build sequence. If the file parameter is specified as off, execution of the named
application will be skipped. In most cases, this is not desirable as almost all applications
are critical to the success of the build process. Disabling a critical application will result
in catastrophic failure. However, it is permissible to skip a non-critical application such
as CLIST or HEXMATE, if the final results are not relying on their function.

 2012-2013 Microchip Technology Inc. DS50002053D-page 123

MPLAB® XC8 C Compiler User’s Guide

4.8.55 --SHROUD: Obfuscate P-code Files

This option should be used in situations where either p-code files or p-code libraries
are to be distributed and are built from confidential source code.
C comments, which are normally included into these files, as well as line numbers and
variable names will be removed, or obfuscated, so that the original source code cannot
be reconstructed from the distributed files.

4.8.56 --STACK: Specify Data Stack Type For Entire Program

This option allows selection of the stack type to be used by a program’s stack-based
(auto and parameter) variables. The data stacks available are called a compiled stack
and a software stack, and they are described in Section [Link] “Data Stacks”. The
stack types that can be used with this option are described in Table 4-18.

TABLE 4-18: --STACK SUBOPTIONS

Stack types Default Allocation for Stack-based Variables
compiled or nonreentrant Use the compiled stack for all functions; functions are
non-reentrant (default).
software or reentrant Use the software stack for eligible functions and devices;
such functions are reentrant.
hybrid Use the compiled stack for functions not called reentrantly;
use the software stack for all other eligible functions and
devices; functions are only reentrant if required.

Suboptions that specify reentrancy only affect target devices that support a software
stack. Functions encoded for baseline and mid-range devices always use the compiled
stack. In addition, not all functions can use a software stack. Interrupt functions must
use the compiled stack, but functions they call may use the software stack.
The hybrid setting forces the compiler to consider both a compiled and software stack
for the program’s stack-based variables. The software stack will only be used if the
functions and device supports reentrancy. This mode allows for reentrancy, when
required, but takes advantage of the efficiency of the compiled stack for the majority of
the program’s functions. A function is compiled to use the software stack if it is called
reentrantly in the program; otherwise, it will use a compiled stack.
Any of these option settings can be overridden for individual functions by using function
specifiers, described in Section [Link] “Reentrant and nonreentrant Specifiers”.

Note: Use the software (reentrant) setting with caution. The maximum run-
time size of the software stack is not accurately known at compile time, so
the compiler cannot warn of memory overwrites. The stack can overflow
and corrupt objects or data memory used by something outside the pro-
gram (such as hardware or another independently-compiled applications).
When all functions are forced to use the software stack, the stack size will
increase substantially.
In addition to the stack type, this option can be used to specify the maximum size of
memory reserved by the compiler for the software stack. This option configuration only
affects the software stack; there are no controls for the size of the compiled stack.
Distinct memory areas are allocated for the software stack used by main-line code and
each interrupt function. In essence there are separate stacks for each interrupt and
main-line code, but this is transparent at the program level. The compiler automatically
manages the allocation of memory to each stack.

DS50002053D-page 124  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver
You can manually specify the maximum space allocated for each stack by following the
stack type with a colon-separated list of decimal values, each value being the maximum
size, in bytes, of the memory to be reserved. The sizes specified correspond to the
main-line code, the lowest priority interrupt through the highest priority interrupt. (PIC18
devices have two separate interrupts; other devices have only one.) Alternatively, you
can explicitly state you have no size preference by using a size of auto or leave the
numeric field blank. For PIC18 devices, the following example:
--STACK=reentrant:auto:30:50
will arrange the stack starting locations so that the low-priority interrupt stack can grow
to, at most, 30 bytes (before overflow); the high-priority interrupt stack can grow to, at
most, 50 bytes (before overflow); and the main-line code stack can consume the
remainder of the free memory that can be allocated to the stack (before overflow).
If you do specify the stack sizes using this option, each size must be specified numer-
ically or you can use the auto token. Do not leave a size field empty. If you try to use
this option to allocate more stack memory than is available, a warning is issued and
only the available memory will be utilized.

4.8.57 --STRICT: Strict ANSI Conformance

The --STRICT option is used to enable strict ANSI C conformance of all special,
non-standard keywords.
The MPLAB XC8 C compiler supports various special keywords (for example the
persistent type qualifier). If the --STRICT option is used, these keywords are
changed to include two underscore characters at the beginning of the keyword (for
example, __persistent) so as to strictly conform to the ANSI standard. Thus if you
use this option, you will need to use the qualifier __persistent in your code, not
persistent.
Be warned that use of this option can cause problems with some standard header files
(e.g., <xc.h>) as they contain special keywords.
See Section 4.9 “MPLAB X Option Equivalents”, for use of this option in MPLAB
IDE.

4.8.58 --SUMMARY: Select Memory Summary Output Type

Use this option to select the type of memory summary that is displayed after compila-
tion. By default, or if the mem suboption is selected, a memory summary is shown. This
shows the total memory usage for all memory spaces.
A psect summary can be shown by enabling the psect suboption. This shows individ-
ual psects, after they have been grouped by the linker, and the memory ranges they
cover. Table 4-19 shows what summary types are available. The output printed when
compiling normally corresponds to the mem setting.

TABLE 4-19: MEMORY SUMMARY SUBOPTIONS

Suboption Controls
psect A summary of psect names and the addresses where they were
linked will be shown.
mem A concise summary of memory used will be shown. (default)
class A summary of all classes in each memory space will be shown.
hex A summary of addresses and HEX files that make up the final out-
put file will be shown.
file Summary information will be shown on screen and saved to a file.

See Section 4.9 “MPLAB X Option Equivalents” for use of this option in MPLAB
IDE.

 2012-2013 Microchip Technology Inc. DS50002053D-page 125

MPLAB® XC8 C Compiler User’s Guide

4.8.59 --TIME: Report Time Taken For Each Phase of Build Process
Adding the --TIME option when building generates a summary that shows how much
time it took for each stage of the build process to complete.

4.8.60 --VER: Display the Compiler’s Version Information

The --VER option will display what version of the compiler is running and then exit the
compiler.

4.8.61 --WARN: Set Warning Level

The --WARN option is used to set the compiler warning level threshold. Allowable warn-
ing levels range from -9 to 9. The warning level determines how pedantic the compiler
is about dubious type conversions and constructs. Each compiler warning has a des-
ignated warning level; the higher the warning level, the more important the warning
message. If the warning message’s warning level exceeds the set threshold, the warn-
ing is printed by the compiler. The default warning level threshold is 0 and will allow all
normal warning messages.
Use this option with care as some warning messages indicate code that is likely to fail
during execution, or compromise portability.
Warning message can be individually disabled with the --MSGDISABLE option, see
Section 4.8.38 “--MSGDISABLE: Disable Warning Messages”. See also
Section 4.6 “Compiler Messages” for full information on the compiler’s messaging
system.
See Section 4.9 “MPLAB X Option Equivalents” for use of this option in MPLAB
IDE.

4.8.62 --WARNFORMAT: Set Warning Message Format

This option sets the format of warning messages produced by the compiler. See
Section 4.8.27 “--ERRFORMAT: Define Format for Compiler Messages” for more
information on this option. For full information on the compiler’s messaging system, see
Section 4.6 “Compiler Messages”.
If you are compiling using MPLAB IDE, the format of the compiler messages is auto-
matically configured to what the IDE expects. It recommended that you do not adjust
the message formats if compiling using this IDE.

DS50002053D-page 126  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver

4.9 MPLAB X OPTION EQUIVALENTS

When compiling under the MPLAB X IDE, it is still the compiler’s command-line driver
that is being executed and compiling the program. The MPLAB XC8 compiler plugins
control the MPLAB X IDE Properties dialog that is used to access the compiler options,
however these graphical controls ultimately adjust the command-line options passed
to the command-line driver when compiling. You can see the command-line options
being used when building in MPLAB X IDE in the Output window.
The following dialogs and descriptions identify the mapping between the MPLAB X IDE
dialog controls and command-line options. Click any option to see online help and
examples shown in the Option Description field in the lower part of the Project
Properties dialog.

4.9.1 Global Category

The options in the panel in this category control the final output of the compiler.
See Figure 4-5 in conjunction with the following command-line option equivalent.

FIGURE 4-5: GLOBAL OPTIONS

Output file format

This selector specifies the output source-level debug format that will be used by
debuggers; see Section 4.8.44 “--OUTPUT= type: Specify Output File Type”.

4.9.2 Compiler Category

The panels in this category control aspects of compilation of C source.

[Link] PREPROCESSING AND MESSAGES

These options relate to the C preprocessor and messages produced by the compiler;
see Section 4.6 “Compiler Messages” for more information.
See Figure 4-6 in conjunction with the following command-line option equivalents.

 2012-2013 Microchip Technology Inc. DS50002053D-page 127

MPLAB® XC8 C Compiler User’s Guide
FIGURE 4-6: PREPROCESSING AND MESSAGES OPTIONS

1
2
3
4
5
6
7
8
9
10
11

1. Define macros
The button and field on this line can be used to define preprocessor macros; see
Section 4.8.2 “-D: Define Macro”.
2. Undefine macros
The button and field on this line can be used to undefine preprocessor macros;
see Section 4.8.13 “-U: Undefine a Macro”.
3. Preprocess assembly files
This checkbox controls whether assembly source files are scanned by the pre-
processor; see Section 4.8.10 “-P: Preprocess Assembly Files”.
4. Identifier length
Not implemented; see Section 4.8.8 “-N: Identifier Length”.
5. Include directories
This selection uses the buttons and fields grouped in the bracket to specify
include (header) file search directories; see Section 4.8.4 “-I: Include Search
Path”.
6. Strict ANSI Conformance
This forces the compiler to reject any non-standard keywords; see
Section 4.8.56 “--STACK: Specify Data Stack Type For Entire Program”.
7. Verbose
This checkbox controls whether the full command lines for the compiler applica-
tions are displayed when building; see Section 4.8.14 “-V: Verbose Compile”.
8. Warning level
This selector allows the warning level print threshold to be set; see
Section 4.8.61 “--WARN: Set Warning Level”.
9. Use CCI Syntax
This option indicates that the compiler should use the Common C Interface
compiler extensions; see Section 4.8.29 “--EXT: Specify C Language
Extensions”, and Chapter 2. “Common C Interface”, for more information.

DS50002053D-page 128  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver
10. Use IAR Syntax
This option indicates that the compiler should use the IAR compiler extensions;
see Section 4.8.29 “--EXT: Specify C Language Extensions” for more
information.
11. Generate the ASM listing file
This option indicates that the compiler should generate an assembly listing file.
This file should be used to examine the assembly code produced by the
compiler; see Section 4.8.16 “--ASMLIST: Generate Assembler List Files”.

[Link] OPTIMIZATIONS
These options, shown in Figure 4-7, relate to optimizations performed by the compiler,

1. Optimization controls
These controls adjust the optimizations employed by the compiler (see
Section 4.8.42 “--OPT: Invoke Compiler Optimizations”). The Optimization
set widget, if present, is not used; and, you can choose any setting for this. Select
the custom options below (--OPT suboptions: asm, asmfile, speed/space
and debug). The Speed checkbox indicates your preference for any applicable
optimizations to be focused on speed or space.
2. Address qualifiers
This selector allows the user to select the behavior of the address qualifiers; see
Section 4.8.15 “--ADDRQUAL: Set Compiler Response to Memory
Qualifiers”.
3. Operation mode
This selector allows the user to force another available operating mode (free,
std, or pro) other than the default; see Section 4.8.37 “--MODE: Choose
Compiler Operating Mode”. The operating mode will affect OCG-optimizations
FIGURE 4-7: OPTIMIZATIONS OPTIONS

1
{ 2
3

 2012-2013 Microchip Technology Inc. DS50002053D-page 129

MPLAB® XC8 C Compiler User’s Guide

4.9.3 Linker Category

The options in this dialog control the aspects of the second stage of compilation
including code generation and linking.

[Link] RUNTIME
All the widgets in Figure 4-8 correspond to suboptions of the --RUNTIME option, see
Section 4.8.51 “--RUNTIME: Specify Runtime Environment”. Respectively, these
map to the clear, init, keep, no_startup, osccal, oscval, resetbits,
download, stackcall, config, clib and plib suboptions of the --RUNTIME
option.

FIGURE 4-8: RUNTIME OPTIONS

DS50002053D-page 130  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver
[Link] MEMORY MODEL
The panel in this category, shown in Figure 4-9, controls settings that apply to the entire
project.

FIGURE 4-9: MEMORY MODEL OPTIONS

1
2
3
4
5

1. Size of Double
This selector allows the size of the double type to be selected; see
Section 4.8.23 “--DOUBLE: Select Kind of Double Types”.
2. Size of Float
This selector allows the size of the float type to be selected; see
Section 4.8.31 “--FLOAT: Select Kind of Float Types”.
3. External memory
This option allows specification of how external memory access is performed.
This only affects those devices that can access external memory; see
Section 4.8.25 “--EMI: Select External Memory Interface Operating Mode”.
4. RAM ranges
This field allows the default RAM (data space) memory used to be adjusted; see
Section 4.8.49 “--RAM: Adjust RAM Ranges”.
5. ROM ranges
This field allows the default ROM (program memory space) memory used to be
adjusted; see Section 4.8.50 “--ROM: Adjust ROM Ranges”.

 2012-2013 Microchip Technology Inc. DS50002053D-page 131

MPLAB® XC8 C Compiler User’s Guide
[Link] FILL FLASH MEMORY
All the controls shown in Figure 4-10 relate to options associated with filling unused
program memory. See Section 4.8.30 “--FILL: Fill Unused Program Memory”, for
more information on the different fields.

FIGURE 4-10: FILL FLASH MEMORY OPTIONS

DS50002053D-page 132  2012-2013 Microchip Technology Inc.

 XC8 Command-line Driver
[Link] ADDITIONAL
The options shown in Figure 4-11 relate to miscellaneous options.
1. Extra linker options
This field allows you to enter additional options to the linker. Enter the entire -L-
option in this field, for example -L-pmytest=300h. See Section 4.8.6 “-L-:
Adjust Linker Options Directly”.
2. Serial
This option allows you to specify a string that can be inserted into your output
HEX file. See Section 4.8.53 “--SERIAL: Store a Value at this Program Mem-
ory Address”, for details.
3. Codeoffset
This field allows an offset for the program to be specified; see
Section 4.8.20 “--CLIST: Generate C Listing File”.
4. Checksum
This field allows the checksum specification to be specified; see
Section 4.8.17 “--CHECKSUM: Calculate a Checksum”.
5. Errata
This allows customization of the errata workarounds applied by the compiler; see
Section 4.8.26 “--ERRATA: Specify Errata Workarounds”.
6. Trace type
Not implemented. Native trace supported.
7. Extend address 0 in HEX file
This option specifies that the Intel HEX file should have initialization to zero of
the upper address; see Section 4.8.44 “--OUTPUT= type: Specify Output File
Type”.
8. Use response file to link
This option allows a file name to be specified. The file must contain
command-line options which are then used by MPLAB XC8 during the link step
and in preference to the other link-step settings in the project properties; see
Section [Link] “Long Command Lines”. This option is only relevant when
running MPLAB X IDE under Windows.

FIGURE 4-11: ADDITIONAL OPTIONS

1
2
3
4
5
6
7
8

 2012-2013 Microchip Technology Inc. DS50002053D-page 133

MPLAB® XC8 C Compiler User’s Guide
[Link] REPORTING
These options, shown in Figure 4-12 relate to information produced during and after
compilation.

1. Display memory usage after compilation

These checkboxes allow you to specify what information is displayed after
compilation. The correspond to the psect, class, mem and hex suboptions to
the --SUMMARY option; see Section 4.8.58 “--SUMMARY: Select Memory
Summary Output Type”.
2. Create summary file
Selecting this checkbox will send the information you have selected above to a
file, as well as to the standard output. This corresponds to the file suboption to
the --SUMMARY option; see Section 4.8.58 “--SUMMARY: Select Memory
Summary Output Type”.
3. Create html files
This will create HTML files summarizing the previous build; see
Section 4.8.34 “--HTML: Generate HTML Diagnostic Files”.

FIGURE 4-12: REPORTING OPTIONS

1
{ 2
3

DS50002053D-page 134  2012-2013 Microchip Technology Inc.

 MPLAB® XC8 C COMPILER
USER’S GUIDE

Chapter 5. C Language Features

5.1 INTRODUCTION
MPLAB XC8 C Compiler supports a number of special features and extensions to the
C language which are designed to ease the task of producing ROM-based applications
for 8-bit PIC devices. This chapter documents the special language features which are
specific to these devices.
• ANSI C Standard Issues
• Device-Related Features
• Supported Data Types and Variables
• Memory Allocation and Access
• Operators and Statements
• Register Usage
• Functions
• Interrupts
• Main, Runtime Startup and Reset
• Library Routines
• Mixing C and Assembly Code
• Optimizations
• Preprocessing
• Linking Programs

5.2 ANSI C STANDARD ISSUES

This compiler conforms to the ISO/IEC 9899:1990 Standard for programming lan-
guages. This is commonly called the C90 Standard. It is referred to as the ANSI C
Standard in this manual.
Some violations to the ANSI C Standard are indicated below in
Section 5.2.1 “Divergence from the ANSI C Standard”. Some features from the
later standard C99 are also supported.

5.2.1 Divergence from the ANSI C Standard

The C language implemented on MPLAB XC8 C Compiler can diverge from the ANSI
C Standard in several areas.
Due to limited memory and no hardware implementation of a data stack, recursion is
not supported and functions are not reentrant on baseline and some mid-range
devices. Functions can be encoded reentrantly for enhanced mid-range and PIC18
devices. See Section 5.3.4 “Stacks” for more information on the stack models used
by the compiler for each device family.
For those devices that do not support reentrancy, the compiler can make functions
called from main-line and interrupt code appear to be reentrant via a duplication
feature. See Section 5.9.6 “Function Duplication”.

 2012-2013 Microchip Technology Inc. DS50002053D-page 135

MPLAB® XC8 C Compiler User’s Guide
You cannot reliably use the C sizeof operator with pointer types; however, this oper-
ator may be used with pointer variable identifiers. This is a result of the dynamic size
of pointers assigned by the compiler. So for the following code:
char * cp;
size_t size;
size = sizeof(char *);
size = sizeof(cp);
size in the first example will be assigned the maximum size a pointer can be for the
particular target device you have chosen. In the second example, size will be
assigned the actual size of the pointer variable, cp.

5.2.2 Implementation-Defined behavior

Certain features of the ANSI C standard have implementation-defined behavior. This
means that the exact behavior of some C code can vary from compiler to compiler. The
exact behavior of the compiler is detailed throughout this manual, and is fully
summarized in Appendix D. “Implementation-Defined Behavior”.

5.2.3 Common C Interface Standard

This compiler conforms to the Microchip XC compiler Common C Interface standard
(CCI). This is a further refinement of the ANSI standard that attempts to standardize
implementation-defined behavior and non-standard extensions across the entire
MPLAB XC compiler family. It is described in Chapter 2. “Common C Interface”.
If you choose to write code which conforms to this standard, a compiler option (see
Section 4.8.29 “--EXT: Specify C Language Extensions”) should be enabled. This
will to indicate that the compiler should enforce conformance. Alternatively, you can
continue to write code using the non-standard ANSI extensions provided by the com-
piler.

DS50002053D-page 136  2012-2013 Microchip Technology Inc.

 C Language Features

5.3 DEVICE-RELATED FEATURES

MPLAB XC8 has several features which relate directly to the 8-bit PIC architectures
and instruction sets. These are detailed in the following sections.

5.3.1 Device Support

MPLAB XC8 C Compiler aims to support all 8-bit PIC devices. However, new devices
in these families are frequently released. There are several ways you can check
whether the compiler you are using supports a particular device.
From the command line, run the compiler you wish to use and pass it the option
--CHIPINFO (See Section 4.8.19 “--CHIPINFO: Display List of Supported
Devices”). A list of all devices will be printed.
If you use the -V option in addition to the --CHIPINFO option, more detailed
information my be shown about each device.
You can also see the supported devices in your favorite web browser. Open the files
pic_chipinfo.html for a list of all supported baseline or mid-range device, or
pic18_chipinfo.html for all PIC18 devices. Both these files are located in the
DOCS directory under your compiler’s installation directory.

5.3.2 Instruction Set Support

The compiler support all instruction sets for PIC10/12/16 devices as well as the stan-
dard (legacy) PIC18 instruction set. The extended instruction mode available on some
PIC18 devices is not currently supported. Ensure you set the Configuration bits to use
the PIC18 legacy instruction mode when appropriate.

5.3.3 Device Header Files

There is one header file that is typically included into each C source file you write. The
file is <xc.h> and is a generic header file that will include other device- and architec-
ture-specific header files when you build your project.
Inclusion of this file will allow access to SFRs via special variables, as well as macros
which allow special memory access or inclusion of special instructions, like CLRWDT.
Legacy projects can continue to use the <htc.h> header file.
Avoid including chip-specific header files into your code as this will reduce portability.
The header files shipped with the compiler are specific to that compiler version. Future
compiler versions can ship with modified header files. If you copy compiler header files
into your project, particularly if you modify these files, be aware that they cannot be
compatible with future versions of the compiler.
For information about assembly include files (.inc), see
Section [Link] “Accessing Registers from Assembly Code”.

 2012-2013 Microchip Technology Inc. DS50002053D-page 137

MPLAB® XC8 C Compiler User’s Guide

5.3.4 Stacks
Stacks are used for two different purposes by programs running on 8-bit devices: one
stack is for storing function return addresses, and one or two other stacks are used for
data allocation.

[Link] FUNCTION RETURN ADDRESS STACK

The 8-bit PIC devices use what is referred to in this user’s guide as a hardware stack.
This stack is limited in depth and cannot be manipulated directly. It is only used for
function return addresses and cannot be used for program data.
You must ensure that the maximum hardware stack depth is not exceeded; otherwise,
code can fail. Nesting function calls too deeply will overflow the stack. It is important to
take into account implicitly called library functions and interrupts, which also use levels
of the stack. The compiler can be made to manage stack usage for some devices using
the stackcall suboption to the --RUNTIME compiler option, see
Section 4.8.51 “--RUNTIME: Specify Runtime Environment”. This enables an alter-
nate means of calling functions to prevent stack overflow.
A call graph is provided by the code generator in the assembler list file, see
Section 6.6.6 “Call Graph”. This will indicate the stack levels at each function call and
can be used as a guide to stack depth. The code generator can also produce warnings
if the maximum stack depth is exceeded.
The warnings and call graphs are guides to stack usage. Optimizations and the use of
interrupts can decrease or increase the program’s stack depth over that determined by
the compiler.

[Link] DATA STACKS

The compiler can implement two types of data stack: a compiled stack and a software
stack. Both these stacks are for storing stack-based variables, such as a function’s
auto, parameter, and temporary variables.
Either one or both of these types of stacks may be used by a program. Compiler
options, specifiers, and how the functions are called will dictate which stacks are used.
See Section [Link] “Auto Variable Allocation and access”, for more information on
how the compiler allocates a function’s stack-based objects.
A compiled stack is a static allocation of memory for stack-based objects that can be
built up in multiple data banks. See Section [Link].1 “Compiled Stack Operation”,
for information about how objects are allocated to this stack. Objects in the stack are in
fixed locations and can be accessed using an identifier (hence it is a static allocation).
Thus, there is no stack pointer. The size of the compiled stack is known at compile time,
and so available space can be confirmed by the compiler. The compiled stack is allo-
cated to psects that use the basename cstack; for example, cstackCOMMON,
cstackBANK0. See Section 5.15.2 “Compiler-Generated Psects”, for more
information on the naming convention for compiler-generated psects.
By contrast, the software stack has a size that is dynamic and varies as the program is
executed. The maximum size of the stack is not exactly known at compile time and the
compiler typically reserves as much space as possible for the stack to grow during pro-
gram execution. The stack is always allocated a single memory range, which may cross
bank boundaries, but within this range it may be segregated into one area for main-line
code and an area for each interrupt routine, if required. A stack pointer is used to
indicate the current position in the stack. This pointer is permanently allocated to FSR1.
A psect is used as a placeholder to reserve the memory used by the stack. This psect
is called stack.

DS50002053D-page 138  2012-2013 Microchip Technology Inc.

 C Language Features

5.3.5 Configuration Bit Access

The PIC devices have several locations which contain the Configuration bits or fuses.
These bits specify fundamental device operation, such as the oscillator mode, watch-
dog timer, programming mode and code protection. Failure to correctly set these bits
can result in code failure, or a non-running device.
These bits can be set using a configuration pragma. The pragma has the following
forms.
#pragma config setting = state|value
#pragma config register = value
where setting is a configuration setting descriptor, e.g., WDT, and state is a textual
description of the desired state, e.g., OFF. The value field is a numerical value that can
be used in preference to a descriptor.
Consider the following examples.
#pragma config WDT = ON // turn on watchdog timer
#pragma config WDTPS = 0x1A // specify the timer postscale value
One pragma can be used to program several settings by separating each setting-value
pair with a comma. For example, the above could be specified with one pragma, as in
the following.
#pragma config WDT=ON, WDTPS = 0x1A
The setting-value pairs can also be quoted to ensure that the preprocessor does not
perform substitution of these tokens, for example:
#pragma config "BOREN=OFF"
Without the quotes and with the preprocessor macro OFF defined, for example, substi-
tution within the pragma would take place. You should never assume that the OFF and
ON tokens used in configuration macros equate to 0 and 1, respectively, as that is often
not the case.
Rather than specify individual settings, the entire register can be programmed with one
numerical value, for example:
#pragma config CONFIG1L = 0x8F
The upper and lower half of each register must be programmed separately.
The settings and values associated with each device can be determined from an HTML
guide. Open the file pic_chipinfo.html or pic18_chipinfo.html, which are
located in the DOCS directory of your compiler installation. Click the link to your target
device, and the page will show you the settings and values that are appropriate with
this pragma. Check your device data sheet for more information.

 2012-2013 Microchip Technology Inc. DS50002053D-page 139

MPLAB® XC8 C Compiler User’s Guide
[Link] CONFIGURATION BIT LEGACY SUPPORT
You can continue to use the configuration macros for legacy projects, but use the
pragma for new projects.
The compiler supports the __CONFIG and __PROG_CONFIG macros, which allow con-
figuration bit symbols or a Configuration Word value, respectively, to be specified, for
example:
#include <xc.h>
__CONFIG(WDTDIS & HS & UNPROTECT);
For PIC10/12/16 devices that have more than one Configuration Word, each subse-
quent invocation of __CONFIG() will modify the next Configuration Word in sequence.
When using the legacy macros for these devices, the order of the macros must match
the order of the Configuration Words. Typically this might look like:
#include <xc.h>
__CONFIG(WDTDIS & XT & UNPROTECT); // Program config. word 1
__CONFIG(FCMEN);
The __CONFIG macro used for PIC18 devices takes an additional argument being the
number of the Configuration Word location. For example:
__CONFIG(2, BW8 & PWRTDIS & WDTPS1 & WDTEN); // specify symbols
If you want to use a literal value to program the entire Configuration Word, you must
use the __PROG_CONFIG macro. For PIC10/12/16 devices, that might appear as fol-
lows:
__PROG_CONFIG(0xFFFA);
and with PIC18 devices, you must again specify the word being programmed, as in the
following:
__PROG_CONFIG(1, 0xFE57); // specify a literal constant value
You cannot use the setting symbols in the __PROG_CONFIG macro, nor can you use a
literal value in the __CONFIG macro.
The configuration locations do not need to be programmed in order, except as noted
above for multi-word PIC10/12/16 devices using the legacy macros.
To use the legacy macros, ensure you include <xc.h> in your source file. Symbols for
the macros can be found in the .cfgmap files contained in the dat/cfgmap directory
of your compiler installation.

[Link] CONFIGURATION CONSIDERATIONS

Neither the config pragma nor the __CONFIG macro produce executable code. They
should both be placed outside function definitions so as not to interfere with the
operation of the function’s code.
MPLAB X IDE does not allow the Configuration bits to be adjusted. They must be
specified in your source code using the pragma (or legacy macro).
All the bits in the Configuration Words should be programmed to prevent erratic pro-
gram behavior. Do not leave them in their default/unprogrammed state. Not all Config-
uration bits have a default state of logic high; some have a logic low default state.
Consult your device data sheet for more information.

DS50002053D-page 140  2012-2013 Microchip Technology Inc.

 C Language Features

5.3.6 Using SFRs From C Code

The Special Function Registers (SFRs) are registers which control aspects of the MCU
operation or that of peripheral modules on the device. Most of these registers are mem-
ory mapped, which means that they appear at, and can be accessed using, specific
addresses in the device’s data memory space. Individual bits within some registers
control independent features. Some registers are read-only; some are write-only. See
your device data sheet for more information.
Memory-mapped SFRs are accessed by special C variables that are placed at the
address of the register. (Variables that are placed at specific addresses are called
absolute variables and are described in Section 5.5.4 “Absolute Variables”.) These
variables can be accessed like any ordinary C variable so that no special syntax is
required to access SFRs.
The SFR variables are predefined in header files and are accessible once you have
included the <xc.h> header file (see Section 5.3.3 “Device Header Files”) into your
source code. Both bit variables and structures with bit-fields are defined, so you can
use either of them in your source code to access bits within a register.
The names given to the C variables that map over registers and bits within those reg-
isters are based on the names specified in the device data sheet. However, as there
can be duplication of some bit names within registers, there can be differences in the
nomenclature.
The names of the structures that hold the bit-fields will typically be those of the corre-
sponding register followed by bits. For example, the following shows code that
includes the generic header file, clears PORTA as a whole, sets bit 0 of PORTA using
a bit variable and sets bit 2 of PORTA using the structure/bit-field definitions.
#include <xc.h>
void main(void)
{
PORTA = 0x00;
RA0 = 1;
PORTAbits.RA2 = 1;
}
To confirm the names that are relevant for the device you are using, check the
device-specific header file that <xc.h> will include for the definitions of each variable.
These files will be located in the include directory of the compiler and will have a
name that represents the device. There is a one-to-one correlation between device and
header file name that will be included by <xc.h>, e.g., when compiling for a
PIC16LF1826 device, <xc.h> will include the header file <pic16lf1826.h>.
Remember that you do not need to include this chip-specific file into your source code;
it is automatically included by <xc.h>.
Care should be taken when accessing some SFRs from C code or from assembly
in-line with C code. Some registers are used by the compiler to hold intermediate val-
ues of calculations, and writing to these registers directly can result in code failure. The
compiler does not detect when SFRs have changed as a result of C or assembly code
that writes to them directly. The list of registers used by the compiler and further
information can be found in Section 5.7 “Register Usage”.
SFRs associated with peripherals are not used by the compiler to hold intermediate
results and can be changed as you require. Always ensure that you confirm the
operation of peripheral modules from the device data sheet.

 2012-2013 Microchip Technology Inc. DS50002053D-page 141

MPLAB® XC8 C Compiler User’s Guide
[Link] SPECIAL BASELINE/MID-RANGE REGISTER ISSUES
Some SFRs are not memory mapped, do not have a corresponding variable defined in
the device specific header file, and cannot be directly accessed from C code.
For example, the W register is not memory mapped on baseline devices. Some devices
use OPTION and TRIS registers, that are only accessible via special instructions and
that are also not memory mapped. See Section 5.3.9 “Hardware Multiply Instruc-
tions” on how these registers are accessed by the compiler.

[Link] SPECIAL PIC18 REGISTER ISSUES

Some of the SFRs associated with the PIC18 can be grouped to form multi-byte values,
e.g., the TMRxH and TMRxL register combined form a 16-bit timer count value.
Depending on the device and mode of operation, there can be hardware requirements
to read these registers in certain ways, e.g., often the TMRxL register must be read
before trying to read the TMRxH register to obtain a valid 16-bit result.
Although it is possible to define a word-sized C variable to map over such registers, i.e.,
an int variable TMRx that maps over both TMRxL and TMRxH, the order in which the
compiler would read the bytes of such an object will vary from expression to expres-
sion. Some expressions require that the Most Significant Byte (MSB) is read first;
others start with the Least Significant Byte (LSB) first.
It is recommended that the existing SFR definitions in the chip header files be used.
Each byte of the SFR should be accessed directly, and in the required order, as dictated
by the device data sheet. This results in a much higher degree of portability.
The following code copies the two timer registers into a C unsigned variable count
for subsequent use.
count = TMR0L;
count += TMR0H << 8;
Macros are also provided to perform reading and writing of the more common timer reg-
isters. See the macros READTIMERx and WRITETIMERx in Appendix A. “Library
Functions”. These guarantee the correct byte order is used.

5.3.7 ID Locations
The 8-bit PIC devices have locations outside the addressable memory area that can be
used for storing program information, such as an ID number. The config pragma is
also used to place data into these locations by using a special register name. The
pragma is used as follows.
#pragma config IDLOCx = value
where x is the number (position) of the ID location, and value is the nibble or byte
which is to be positioned into that ID location. If value is larger than the maximum
value allowable for each location on the target device, the value will be truncated and
a warning message issued. The size of each ID location varies from device to device.
See your device data sheet for more information.
For example:
#pragma config IDLOC0 = 1
#pragma config IDLOC1 = 4
will attempt fill the first two ID locations with 1 and 4. One pragma can be used to pro-
gram several locations by separating each register-value pair with a comma. For
example, the above could also be specified as shown below.
#pragma config IDLOC0 = 1, IDLOC1 = 4
The config pragma does not produce executable code and so should ideally be placed
outside function definitions.

DS50002053D-page 142  2012-2013 Microchip Technology Inc.

 C Language Features
[Link] ID LOCATION LEGACY SUPPORT
The compiler also has legacy support for the __IDLOC macro. The macro is used in a
manner similar to:
#include <xc.h>
__IDLOC(x);
where x is a list of hexadecimal digits, which are positioned into the ID locations. Do
not use the usual 0x hexadecimal radix specifier with these values. If an invalid char-
acter is encountered, the value 0 will be programmed into the corresponding location.
Only the lower four bits of each ID location are programmed, so the following:
__IDLOC(15F0);
will attempt to fill ID locations with the hexadecimal values: 1, 5, F and 0.
To use this macro, ensure you include <xc.h> in your source file.
The __IDLOC macro does not produce executable code and so should ideally be
placed outside function definitions.
Some devices permit programming up to seven bits within each ID location. The
__IDLOC() macro is not suitable for such devices and the __IDLOC7(a,b,c,d)
macro should be used instead. The parameters a to d must be constants which repre-
sent the values to be programmed. The values can be entered in either decimal or
hexadecimal format, such as:
__IDLOC7(0x7f,1,70,0x5a);
It is not appropriate to use the __IDLOC7() macro on a device that does not permit
seven-bit programming of ID locations. The __IDLOC7 macro does not produce exe-
cutable code and so should ideally be placed outside function definitions.

5.3.8 Bit Instructions

Wherever possible, the MPLAB XC8 C Compiler will attempt to use bit instructions,
even on non-bit integer values. For example, when using a bitwise operator and a mask
to alter a bit within an integral type, the compiler will check the mask value to determine
if a bit instruction can achieve the same functionality.
unsigned int foo;
foo |= 0x40;
will produce the instruction:
BSF _foo,6
To set or clear individual bits within integral type, the following macros could be used:
#define bitset(var, bitno) ((var) |= 1UL << (bitno))
#define bitclr(var, bitno) ((var) &= ~(1UL << (bitno)))
To perform the same operation on foo as above, the bitset macro could be
employed as follows:
bitset(foo,6);

 2012-2013 Microchip Technology Inc. DS50002053D-page 143

MPLAB® XC8 C Compiler User’s Guide

5.3.9 Hardware Multiply Instructions

The PIC18 instruction set includes several 8-by-8-bit hardware multiple instructions,
and these are used by the compiler in some situations. Non-PIC18 targets always use
an iterative library routine for multiplication operations.
The hardware multiply instructions are used for all 8-bit (char) multiplications by the
compiler.
A special form of the int (16-bit) multiplication library routine (__wmul) has been
employed to decompose the 16-bit operation into the addition of several 8-bit multipli-
cation results. Thus, the benefits of hardware multiplication are also utilized for 16-bit
multiplication expressions.
A similar algorithm has not been implemented for long, 32-bit, multiplies (__lmul) as
the decomposition overhead outweighs the advantage of the hardware multiply
instruction.
For all devices, when the hardware multiply instructions are not available, a library rou-
tine consisting of an iterative algorithm is called to perform the multiply. The smaller the
first argument to these routines (i.e., the smaller the left operand to the * operator), the
faster the routine will run. For example, for the code:
x = 10;
y = 200;
result = x * y; // first multiply
result = y * x; // second multiply
the variable result will be assigned the same value in both statements, but the first
multiplication expression will be performed faster than the second.

5.3.10 Baseline PIC MCU Special Instructions

The Baseline devices have some registers which are not in the normal SFR space and
cannot be accessed using an ordinary file instruction. These are the OPTION and TRIS
registers.
Both registers are write-only and cannot be used in expression that read their value.
They can only be accessed using special instructions which the compiler will use
automatically.
The definition of the variables that map to these registers make use of the control
qualifier. This qualifier informs the compiler that the registers are outside of the normal
address space and that a different access method is required. You should not use this
qualifiers for any other registers.
When you write to either of these SFR variables, the compiler will use the appropriate
instruction to load the value. So, for example, to load the TRIS register, the following
code:
TRIS = 0xFF;
can be encoded by the compiler as:
MOVLW 0ffh
TRIS
Those PIC devices which have more than one output port can have definitions for
objects: TRISA, TRISB and TRISC, depending on the exact number of ports available.
This objects are used in the same manner as described above.

DS50002053D-page 144  2012-2013 Microchip Technology Inc.

 C Language Features

5.3.11 Oscillator Calibration Constants

Some Baseline and Mid-range devices come with an oscillator calibration constant
which is pre-programmed into the device’s program memory. This constant can be read
from program memory and written to the OSCCAL register to calibrate the internal RC
oscillator.
On some Baseline PIC devices, the calibration constant is stored as a MOVLW instruc-
tion at the top of program memory, e.g., the PIC10F509 device. On Reset, the program
counter is made to point to this instruction and it is executed first before the program
counter wraps around to 0x0000, which is the effective Reset vector for the device. The
default runtime startup routine (see Section 5.10.1 “Runtime Startup Code”) will
automatically include code to load the OSCCAL register with the value contained in the
W register after Reset on such devices. No other code is required.
For other chips, such as PIC12F629 device, the oscillator constant is also stored at the
top of program memory, but as a RETLW instruction. The compiler’s startup code will
automatically generate code to retrieve this value and perform the configuration.
At runtime, the calibration value stored as a RETLW instruction can be read using the
‘function’ __osccal_val(),as a label is assigned the RETLW instruction address. A
prototype for the function is provided in <xc.h>. For example:
calVal = __osccal_val();
Loading of the calibration value can be turned off via the osccal suboption to the
--RUNTIME option (see Section 4.8.51 “--RUNTIME: Specify Runtime Environ-
ment”).
At runtime, this calibration value can be read using the macro
_READ_OSCCAL_DATA(). To be able to use this macro, make sure that <xc.h> is
included into the relevant modules of your program. This macro returns the calibration
constant which can then be stored into the OSCCAL register, as follows:
OSCCAL = _READ_OSCCAL_DATA();
Note: The location which stores the calibration constant is never code protected
and will be lost if you reprogram the device. Thus, if you are using a win-
dowed or Flash device, the calibration constant must be saved from the last
ROM location before it is erased. The constant must then be reprogrammed
at the same location along with the new program and data.
If you are using an in-circuit emulator (ICE), the location used by the cali-
bration RETLW instruction cannot be programmed. Calling the
_READ_OSCCAL_DATA() macro will not work and will almost certainly not
return correctly. If you wish to test code that includes this macro on an ICE,
you will have to program a RETLW instruction at the appropriate location in
program memory. Remember to remove this instruction when programming
the actual part so you do not destroy the calibration value.

 2012-2013 Microchip Technology Inc. DS50002053D-page 145

MPLAB® XC8 C Compiler User’s Guide

5.3.12 REAL ICE Support

The compiler supports log and trace functions (instrumented trace) when using the
Microchip REAL ICE debugger. See the REAL ICE documentation for more information
on the instrumented trace features.
Only native trace is currently supported by the compiler. Not all devices support instru-
mented trace, and the IDE you are using also needs to have instrumented trace
support for your target device, as well.
The log and trace macro calls need to be added by hand to your source code in MPLAB
IDE. They have the following form.
__TRACE(id);
__LOG(id, expression);
MPLAB X IDE will automatically substitute an appropriate value for id when you com-
pile; however, you can specify these by hand if required. The trace id should be a con-
stant in the range of 0x40 to 0x7F, and the log id is a constant in the range of 0x0 to
0x7F. Each macro should be given a unique number so that it can be properly identified.
The same valid number can be used for both trace and log macros.
The expression can be any integer or 32-bit floating point expression. Typically, this
expression is simply a variable name so the variable’s contents are logged.
Macros should be placed in the C source code at the desired locations. They will trigger
information to be sent to the debugger and IDE when they are executed. Adding trace
and log macros will increase the size of your code as they contribute to the program
image that is downloaded to the device.
Here is an example of these macros that you might add.
inpStatus = readUser();
if(inpStatus == 0) {
__TRACE(id);
recovery();
}
__LOG(id, inpStatus);

5.3.13 Function profiling

The compiler can generate function registration code for the MPLAB REAL ICE debug-
ger to provide function profiling. The flp suboption to the --RUNTIME option (see
Section 4.8.51 “--RUNTIME: Specify Runtime Environment”) enables this feature.
To obtain profiling results, you must also use a Power Monitor Board and MPLAB X IDE
and power monitor plugin that support code profiling for the MPLAB XC8 C Compiler.
When enabled, the compiler inserts assembly code into the prologue and epilogue of
each function. This code communicates runtime information to the debugger to signal
when a function is being entered and when it exits. This information, along with further
measurements made by a Microchip Power Monitor Board, can determine how much
energy each function is using. This feature is transparent, but note the following points
when profiling is enabled:
• The program will increase in size and run slower due to the profiling code
• One extra level of hardware stack is used
• Some additional RAM memory is consumed
• Inlining of functions will not take place for any profiled function
If a function cannot be profiled (due to hardware stack constraints) but is qualified
inline, the compiler might inline the function. See Section [Link] “Inline Specifier”
for more information on inlining functions.

DS50002053D-page 146  2012-2013 Microchip Technology Inc.

 C Language Features

5.4 SUPPORTED DATA TYPES AND VARIABLES

5.4.1 Identifiers
A C variable identifier (the following is also true for function identifiers) is a sequence
of letters and digits, where the underscore character “_” counts as a letter. Identifiers
cannot start with a digit. Although they can start with an underscore, such identifiers are
reserved for the compiler’s use and should not be defined by your programs. Such is
not the case for assembly domain identifiers, which often begin with an underscore,
see Section [Link] “Equivalent Assembly Symbols”.
Identifiers are case sensitive, so main is different to Main.
Not every character is significant in an identifier. The maximum number of significant
characters can be set using an option, see Section 4.8.8 “-N: Identifier Length”. If
two identifiers differ only after the maximum number of significant characters, then the
compiler will consider them to be the same symbol.

5.4.2 Integer Data Types

The MPLAB XC8 compiler supports integer data types with 1, 2, 3 and 4 byte sizes as
well as a single bit type. Table 5-1 shows the data types and their corresponding size
and arithmetic type. The default type for each type is underlined.

TABLE 5-1: INTEGER DATA TYPES

Type Size (bits) Arithmetic Type
bit 1 Unsigned integer
signed char 8 Signed integer
unsigned char 8 Unsigned integer
signed short 16 Signed integer
unsigned short 16 Unsigned integer
signed int 16 Signed integer
unsigned int 16 Unsigned integer
signed short long 24 Signed integer
unsigned short long 24 Unsigned integer
signed long 32 Signed integer
unsigned long 32 Unsigned integer
signed long long 32 Signed integer
unsigned long long 32 Unsigned integer

The bit and short long types are non-standard types available in this implementa-
tion. The long long types are C99 Standard types, but this implementation limits their
size to only 32 bits.
All integer values are represented in little endian format with the Least Significant bit
(LSb) at the lower address.
If no signedness is specified in the type, then the type will be signed except for the
char types which are always unsigned. The bit type is always unsigned and the
concept of a signed bit is meaningless.
Signed values are stored as a two’s complement integer value.
The range of values capable of being held by these types is summarized in Table 5-2
The symbols in this table are preprocessor macros which are available after including
<limits.h> in your source code.

 2012-2013 Microchip Technology Inc. DS50002053D-page 147

MPLAB® XC8 C Compiler User’s Guide
As the size of data types are not fully specified by the ANSI Standard, these macros
allow for more portable code which can check the limits of the range of values held by
the type on this implementation.
The macros associated with the short long type are non-standard macros available
in this implementation; those associated with the long long types are defined by the
C99 Standard.

TABLE 5-2: RANGES OF INTEGER TYPE VALUES

Symbol Meaning Value
CHAR_BIT bits per char 8
CHAR_MAX max. value of a char 127
CHAR_MIN min. value of a char -128
SCHAR_MAX max. value of a signed char 127
SCHAR_MIN min. value of a signed char -128
UCHAR_MAX max. value of an unsigned char 255
SHRT_MAX max. value of a short 32767
SHRT_MIN min. value of a short -32768
USHRT_MAX max. value of an unsigned short 65535
INT_MAX max. value of an int 32767
INT_MIN min. value of a int -32768
UINT_MAX max. value of an unsigned int 65535
SHRTLONG_MAX max. value of a short long 8388607
SHRTLONG_MIN min. value of a short long -8388608
USHRTLONG_MAX max. value of an unsigned short 16777215
long
LONG_MAX max. value of a long 2147483647
LONG_MIN min. value of a long -2147483648
ULONG_MAX max. value of an unsigned long 4294967295
LLONG_MAX max. value of a long long 2147483647
LLONG_MIN min. value of a long long -2147483648
ULLONG_MAX max. value of an unsigned long 4294967295
long
Macros are also available in <stdint.h> which define values associated with
fixed-width types.
When specifying a signed or unsigned short int, short long int, long int
or long long int type, the keyword int can be omitted. Thus a variable declared
as short will contain a signed short int and a variable declared as unsigned
short will contain an unsigned short int.
It is a common misconception that the C char types are intended purely for ASCII char-
acter manipulation. However, the C language makes no guarantee that the default
character representation is even ASCII. (This implementation does use ASCII as the
character representation.)
The char types are the smallest of the multi-bit integer sizes, and behave in all
respects like integers. The reason for the name “char” is historical and does not mean
that char can only be used to represent characters. It is possible to freely mix char
values with values of other types in C expressions. With the MPLAB XC8 C Compiler,
the char types are used for a number of purposes – as 8-bit integers, as storage for
ASCII characters, and for access to I/O locations.

DS50002053D-page 148  2012-2013 Microchip Technology Inc.

 C Language Features
[Link] BIT DATA TYPES AND VARIABLES
The MPLAB XC8 C Compiler supports bit integral types which can hold the values 0
or 1. Single bit variables can be declared using the keyword bit (or __bit), for
example:
bit init_flag;
You can also use the These variables cannot be auto or parameters to a function, but
can be qualified static, allowing them to be defined locally within a function. For
example:
int func(void) {
static bit flame_on;
// ...
}
A function can return a bit object by using the bit keyword in the function’s prototype
in the usual way. The 1 or 0 value will be returned in the carry flag in the STATUS reg-
ister.
The bit variables behave in most respects like normal unsigned char variables, but
they can only contain the values 0 and 1, and therefore provide a convenient and effi-
cient method of storing flags. Eight bit objects are packed into each byte of memory
storage, so they don’t consume large amounts of internal RAM.
Operations on bit objects are performed using the single bit instructions (bsf and
bcf) wherever possible, thus the generated code to access bit objects is very
efficient.
It is not possible to declare a pointer to bit types or assign the address of a bit object
to any pointer. Nor is it possible to statically initialize bit variables so they must be
assigned any non-zero starting value (i.e., 1) in the code itself. Bit objects will be
cleared on startup, unless the bit is qualified persistent.
When assigning a larger integral type to a bit variable, only the LSb is used. For
example, if the bit variable bitvar was assigned as in the following:
int data = 0x54;
bit bitvar;
bitvar = data;
it will be cleared by the assignment since the LSb of data is zero. This sets the bit
type apart from the C99 Standard __Bool, which is a boolean type, not a 1-bit wide
integer. The __Bool type is not supported on the MPLAB XC8 compiler. If you want to
set a bit variable to be 0 or 1 depending on whether the larger integral type is zero
(false) or non-zero (true), use the form:
bitvar = (data != 0);
The psects in which bit objects are allocated storage are declared using the bit
PSECT directive flag, see Section [Link] “PSECT”. All addresses assigned to bit
objects and psects will be bit addresses. For absolute bit variables (see
Section 5.5.4 “Absolute Variables”), the address specified in code must be a bit
address. Take care when comparing these addresses to byte addresses used by all
other variables.
If the xc8 flag --STRICT is used, the bit keyword becomes unavailable, but you can
use the __bit keyword.

 2012-2013 Microchip Technology Inc. DS50002053D-page 149

MPLAB® XC8 C Compiler User’s Guide

5.4.3 Floating-Point Data Types

The MPLAB XC8 compiler supports 24- and 32-bit floating-point types. Floating point
is implemented using either a IEEE 754 32-bit format, or a modified (truncated) 24-bit
form of this. Table 5-3 shows the data types and their corresponding size and arithmetic
type.

TABLE 5-3: FLOATING-POINT DATA TYPES

Type Size (bits) Arithmetic Type
float 24 or 32 Real
double 24 or 32 Real
long double same as double Real

For both float and double values, the 24-bit format is the default. The options
--FLOAT=24 and --DOUBLE=24 can also be used to specify this explicitly. The 32-bit
format is used for double values if the --DOUBLE=32 option is used and for float
values if --FLOAT=32 is used.
Variables can be declared using the float and double keywords, respectively, to
hold values of these types. Floating-point types are always signed and the unsigned
keyword is illegal when specifying a floating-point type. Types declared as long
double will use the same format as types declared as double. All floating-point
values are represented in little endian format with the LSb at the lower address.
This format is described in Table 5-4, where:
• Sign is the sign bit which indicates if the number is positive or negative
• The exponent is 8 bits which is stored as excess 127 (i.e., an exponent of 0 is
stored as 127).
• Mantissa is the mantissa, which is to the right of the radix point. There is an
implied bit to the left of the radix point which is always 1 except for a zero value,
where the implied bit is zero. A zero value is indicated by a zero exponent.
The value of this number is (-1)sign x 2(exponent-127) x 1. mantissa.

TABLE 5-4: FLOATING-POINT FORMATS

Format Sign Biased exponent Mantissa
IEEE 754 32-bit x xxxx xxxx xxx xxxx xxxx xxxx xxxx xxxx
modified IEEE 754 x xxxx xxxx xxx xxxx xxxx xxxx
24-bit

Here are some examples of the IEEE 754 32-bit formats shown in Table 5-5. Note that
the Most Significant Bit (MSb) of the mantissa column (i.e., the bit to the left of the radix
point) is the implied bit, which is assumed to be 1 unless the exponent is zero (in which
case the float is zero).

TABLE 5-5: FLOATING-POINT FORMAT EXAMPLE IEEE 754

Format Number Biased exponent [Link] Decimal
32-bit 7DA6B69Bh 11111011b 1.0100110101101101 2.77000e+37
0011011b
(251) (1.302447676659) —
24-bit 42123Ah 10000100b 1.001001000111010b 36.557
(132) (1.142395019531) —

Use the following process to manually calculate the 32-bit example in Table 5-5.

DS50002053D-page 150  2012-2013 Microchip Technology Inc.

 C Language Features
The sign bit is zero; the biased exponent is 251, so the exponent is 251-127=124. Take
the binary number to the right of the decimal point in the mantissa. Convert this to dec-
imal and divide it by 223 where 23 is the number of bits taken up by the mantissa, to
give 0.302447676659. Add 1 to this fraction. The floating-point number is then given
by:
-1021241.302447676659
which becomes:
12.126764793256e+371.302447676659
which is approximately equal to:
2.77000e+37
Binary floating-point values are sometimes misunderstood. It is important to remember
that not every floating-point value can be represented by a finite sized floating-point
number. The size of the exponent in the number dictates the range of values that the
number can hold, and the size of the mantissa relates to the spacing of each value that
can be represented exactly. Thus the 24-bit format allows for values with approximately
the same range of values representable by the 32-bit format, but the values that can be
exactly represented by this format are more widely spaced.
So, for example, if you are using a 24-bit wide floating-point type, it can exactly store
the value 95000.0. However, the next highest number it can represent is 95002.0 and
it is impossible to represent any value in between these two in such a type as it will be
rounded. This implies that C code which compares floating-point type cannot behave
as expected. For example:
volatile float myFloat;
myFloat = 95002.0;
if(myFloat == 95001.0) // value will be rounded
PORTA++; // this line will be executed!
in which the result of the if() expression will be true, even though it appears the two
values being compared are different.
Compare this to a 32-bit floating-point type, which has a higher precision. It also can
exactly store 95000.0 as a value. The next highest value which can be represented is
(approximately) 95000.00781.
The characteristics of the floating-point formats are summarized in Table 5-6. The sym-
bols in this table are preprocessor macros which are available after including
<float.h> in your source code.
Two sets of macros are available for float and double types, where XXX represents
FLT and DBL, respectively. So, for example, FLT_MAX represents the maximum float-
ing-point value of the float type. It can have two values depending on whether float
is a 24 or 32 bit wide format. DBL_MAX represents the same values for the double
type.

 2012-2013 Microchip Technology Inc. DS50002053D-page 151

MPLAB® XC8 C Compiler User’s Guide
As the size and format of floating-point data types are not fully specified by the ANSI
Standard, these macros allow for more portable code which can check the limits of the
range of values held by the type on this implementation.

TABLE 5-6: RANGES OF FLOATING-POINT TYPE VALUES

Symbol Meaning 24-bit Value 32-bit Value
XXX_RADIX Radix of exponent representation 2 2
XXX_ROUNDS Rounding mode for addition 0 0
XXX_MIN_EXP Min. n such that FLT_RADIXn-1 is -125 -125
a normalized float value
XXX_MIN_10_EXP Min. n such that 10n is a -37 -37
normalized float value
XXX_MAX_EXP Max. n such that FLT_RADIXn-1 128 128
is a normalized float value
XXX_MAX_10_EXP Max. n such that 10n is a 38 38
normalized float value
XXX_MANT_DIG Number of FLT_RADIX mantissa 16 24
digits
XXX_EPSILON The smallest number which 3.05176e-05 1.19209e-07
added to 1.0 does not yield 1.0

5.4.4 Structures and Unions

MPLAB XC8 C Compiler supports struct and union types. Structures and unions
only differ in the memory offset applied to each member.
These types will be at least 1 byte wide. The members of structures and unions cannot
be objects of type bit, but bit-fields are fully supported.
Structures and unions can be passed freely as function arguments and function return
values. Pointers to structures and unions are fully supported.

[Link] STRUCTURE AND UNION QUALIFIERS

The compiler supports the use of type qualifiers on structures. When a qualifier is
applied to a structure, all of its members will inherit this qualification. In the following
example the structure is qualified const.
const struct {
int number;
int *ptr;
} record = { 0x55, &i };
In this case, the entire structure will be placed into the program space and each mem-
ber will be read-only. Remember that all members are usually initialized if a structure
is const as they cannot be initialized at runtime.
If the members of the structure were individually qualified const, but the structure was
not, then the structure would be positioned into RAM, but each member would be
read-only. Compare the following structure with the above.
struct {
const int number;
int * const ptr;
} record = { 0x55, &i };

DS50002053D-page 152  2012-2013 Microchip Technology Inc.

 C Language Features
[Link] BIT-FIELDS IN STRUCTURES
MPLAB XC8 C Compiler fully supports bit-fields in structures.
Bit-fields are always allocated within 8-bit words, even though it is usual to use the type
unsigned int in the definition.
The first bit defined will be the LSb of the word in which it will be stored. When a bit-field
is declared, it is allocated within the current 8-bit unit if it will fit; otherwise, a new byte
is allocated within the structure. Bit-fields can never cross the boundary between 8-bit
allocation units. For example, the declaration:
struct {
unsigned lo : 1;
unsigned dummy : 6;
unsigned hi : 1;
} foo;
will produce a structure occupying 1 byte. If foo was ultimately linked at address 10H,
the field lo will be bit 0 of address 10H; hi will be bit 7 of address 10H. The LSb of
dummy will be bit 1 of address 10H and the MSb of dummy will be bit 6 of address 10h.

Note: Accessing bit-fields larger than a single bit can be very inefficient. If code
size and execution speed are critical, consider using a char type or a char
structure member, instead. Be aware that some SFRs are defined as
bit-fields. Most are single bits, but some can be multi-bit objects.

Unnamed bit-fields can be declared to pad out unused space between active bits in
control registers. For example, if dummy is never referenced, the structure above could
have been declared as:
struct {
unsigned lo : 1;
unsigned : 6;
unsigned hi : 1;
} foo;
A structure with bit-fields can be initialized by supplying a comma-separated list of initial
values for each field. For example:
struct {
unsigned lo : 1;
unsigned mid : 6;
unsigned hi : 1;
} foo = {1, 8, 0};
Structures with unnamed bit-fields can be initialized. No initial value should be supplied
for the unnamed members, for example:
struct {
unsigned lo : 1;
unsigned : 6;
unsigned hi : 1;
} foo = {1, 0};
will initialize the members lo and hi correctly.
A bit-field that has a size of 0 is a special case. The Standard indicates that no further
bit-field is to be packed into the allocation unit in which the previous bit-field, if any, was
placed.

 2012-2013 Microchip Technology Inc. DS50002053D-page 153

MPLAB® XC8 C Compiler User’s Guide
[Link] ANONYMOUS UNIONS
The MPLAB XC8 compiler supports anonymous unions. These are unions with no
identifier and whose members can be accessed without referencing the enclosing
union. These unions can be used when placing inside structures. For example:
struct {
union {
int x;
double y;
};
} aaa;

void main(void)
{
aaa.x = 99;
// ...}
Here, the union is not named and its members accessed as if they are part of the struc-
ture.
Anonymous unions are not part of the ISO C90 C Standard. Their use limits the
portability of any code, and they are not recommended

5.4.5 Pointer Types

There are two basic pointer types supported by the MPLAB XC8 C Compiler: data
pointers and function pointers. Data pointers hold the addresses of variables which can
be indirectly read, and possible indirectly written, by the program. Function pointers
hold the address of an executable function which can be called indirectly via the pointer.
To conserve memory requirements and reduce execution time, pointers are made dif-
ferent sizes and formats. The MPLAB XC8 C Compiler uses sophisticated algorithms
to track the assignment of addresses to all pointers, and, as a result, non-standard
qualifiers are not required when defining pointer variables. The standard qualifiers
const and volatile can still be used and have their usual meaning. Despite this, the
size of each pointer is optimal for its intended usage in the program.

[Link] COMBINING TYPE QUALIFIERS AND POINTERS

It is helpful to first review the ANSI C standard conventions for definitions of pointer
types.
Pointers can be qualified like any other C object, but care must be taken when doing
so as there are two quantities associated with pointers. The first is the actual pointer
itself, which is treated like any ordinary C variable and has memory reserved for it. The
second is the target, or targets, that the pointer references, or to which the pointer
points. The general form of a pointer definition looks like the following:
target_type_&_qualifiers * pointer’s_qualifiers pointer’s_name;
Any qualifiers to the right of the * (i.e., next to the pointer’s name) relate to the pointer
variable itself. The type and any qualifiers to the left of the * relate to the pointer’s tar-
gets. This makes sense since it is also the * operator that dereferences a pointer, which
allows you to get from the pointer variable to its current target.
Here are three examples of pointer definitions using the volatile qualifier. The fields
in the definitions have been highlighted with spacing:
volatile int * vip ;
int * volatile ivp ;
volatile int * volatile vivp ;

DS50002053D-page 154  2012-2013 Microchip Technology Inc.

 C Language Features
The first example is a pointer called vip. It contains the address of int objects that
are qualified volatile. The pointer itself — the variable that holds the address — is
not volatile; however, the objects that are accessed when the pointer is derefer-
enced are treated as being volatile. In other words, the target objects accessible via
the pointer can be externally modified.
The second example is a pointer called ivp which also contains the address of int
objects. In this example, the pointer itself is volatile, that is, the address the pointer
contains can be externally modified; however, the objects that can be accessed when
dereferencing the pointer are not volatile.
The last example is of a pointer called vivp which is itself qualified volatile, and
which also holds the address of volatile objects.
Bear in mind that one pointer can be assigned the addresses of many objects; for
example, a pointer that is a parameter to a function is assigned a new object address
every time the function is called. The definition of the pointer must be valid for every
target address assigned.

Note: Care must be taken when describing pointers. Is a “const pointer” a pointer
that points to const objects, or a pointer that is const itself? You can talk
about “pointers to const” and “const pointers” to help clarify the definition,
but such terms cannot be universally understood.

[Link] DATA POINTERS

The MPLAB XC8 compiler monitors and records all assignments of addresses to each
data pointer the program contains. This includes assignment of the addresses of
objects to pointers; assignment of one pointer to another; initialization of pointers when
they are defined; and takes into account when pointers are ordinary variables and func-
tion parameters, and when pointers are used to access basic objects, or structures or
arrays.
The size and format of the address held by each pointer is based on this information.
When more than one address is assigned to a pointer at different places in the code, a
set of all possible targets the pointer can address is maintained. This information is spe-
cific to each pointer defined in the program, thus two pointers with the same C type can
hold addresses of different sizes and formats due to the way the pointers were used in
the program.
The compiler tracks the memory location of all targets, as well as the size of all targets
to determine the size and scope of a pointer. The size of a target is important as well,
particularly with arrays or structures. It must be possible to increment a pointer so it can
access all the elements of an array, for example.
There are several pointer classifications used with the MPLAB XC8 C Compiler, such
as those indicated below.
For baseline and mid-range devices:
• An 8-bit pointer capable of accessing common memory and two consecutive
banks, e.g., banks 0 and 1, or banks 7 and 8, etc.
• A 16-bit pointer capable of accessing the entire data memory space
• An 8-bit pointer capable of accessing up to 256 bytes of program space data
• A 16-bit pointer capable of accessing up to 64 Kbytes of program space data
• A 16-bit mixed target space pointer capable of accessing the entire data space
memory and up to 64 Kbytes of program space data
For PIC18 devices:
• An 8-bit pointer capable of accessing the access bank
• A 16-bit pointer capable of accessing the entire data memory space

 2012-2013 Microchip Technology Inc. DS50002053D-page 155

MPLAB® XC8 C Compiler User’s Guide
• An 8-bit pointer capable of accessing up to 256 bytes of program space data
• A 16-bit pointer capable of accessing up to 64 Kbytes of program space data
• A 24-bit pointer capable of accessing the entire program space
• A 16-bit mixed target space pointer capable of accessing the entire data space
memory and up to 64 Kbytes of program space data
• A 24-bit mixed target space pointer capable of accessing the entire data space
memory and the entire program space
Each data pointer will be allocated one of the available classifications after preliminary
scans of the source code. There is no mechanism by which the programmer can spec-
ify the style of pointer required (other than by the assignments to the pointer). The C
code must convey the required information to the compiler.
Information about the pointers and their targets are shown in the pointer reference
graph, which is described in Section 6.6.5 “Pointer Reference Graph”. This graph is
printed in the assembly list file, which is controlled by the option described in
Section 4.8.16 “--ASMLIST: Generate Assembler List Files”.
Consider the following mid-range device program in the early stages of development.
It consists of the following code:
int i, j;

int getValue(const int * ip) {

return *ip;
}

void main(void) {
j = getValue(&i);
// ... code that uses j
}
A pointer, ip, is a parameter to the function getValue(). The pointer target type uses
the qualifier const because we do not want the pointer to be used to write to any
objects whose addresses are passed to the function. The const qualification serves
no other purpose and does not alter the format of the pointer variable.
If the compiler allocates the variable i (defined in main()) to bank 0 data memory, it
will also be noted that the pointer ip (parameter to getValue()) only points to one
object that resides in bank 0 of the data memory. In this case, the pointer, ip, is made
an 8-bit wide data pointer. The generated code that dereferences ip in getValue()
will be generated assuming that the address can only be to an object in bank 0.
As the program is developed, another variable, x, is defined and (unknown to the pro-
grammer) is allocated space in bank 2 data memory. The main() function now looks
like:
int i, j; // allocated to bank 0 in this example
int x; // allocated to bank 2 in this example

int getValue(const int * ip) {

return *ip;
}

void main(void) {
j = getValue(&i);
// ... code that uses j
j = getValue(&x);
// ... code that uses j
}

DS50002053D-page 156  2012-2013 Microchip Technology Inc.

 C Language Features
The pointer, ip, now has targets that are in bank 0 and in bank [Link] be able to accom-
modate this situation, the pointer is made 16 bits wide, and the code used to derefer-
ence the pointer will change accordingly. This takes place without any modification to
the source code.
One positive aspect of tracking pointer targets is less of a dependence on pointer qual-
ifiers. The standard qualifiers const and volatile must still be used in pointer defi-
nitions to indicate a read-only or externally-modifiable target object, respectively.
However, this is in strict accordance with the ANSI C standard. Non-standard qualifiers,
like near and bank2, are not required to indicate pointer targets, have no effect, and
should be avoided. Omitting these qualifiers will result in more portable and readable
code, and reduce the chance of extraneous warnings being issued by the compiler.
[Link].1 Pointers to Both Memory Spaces
When a pointer is assigned the address of one or more objects that have been allo-
cated memory in the data space, and also assigned the address of one or more const
objects, the pointer will fall into one of the mixed target space pointers listed in
Section [Link] “Data Pointers”, and the address will be encoded so that the target
memory space can be determined at runtime. The encoding of these pointer types are
as follows.
For the Baseline/Mid-range 16-bit mixed target space pointer, the MSb of the address
(i.e., bit number 15) indicates the memory space that the address references. If this bit
is set, it indicates that the address is of something in program memory; clear indicates
an object in the data memory. The remainder of this address represents the full address
in the indicated memory space.
For the PIC18 16-bit mixed target space pointer, any address above the highest data
space address is that of an object in the program space memory; otherwise, the
address is of a data space memory object.
For the PIC18 24-bit mixed target space pointer, bit number 21 indicates the memory
space that the address references. If this bit is set, it indicates that the address is of an
object residing in data memory; if it is clear, it indicates an object in the program mem-
ory. The remainder of this address represents the full address in the indicated memory
space. Note that for efficiency reasons, the meaning of the memory space bit is the
opposite to that for baseline and mid-range devices.
To extend the mid-range device example given in Section [Link] “Data Pointers”,
the code is now developed further. The function getValue() is now called with the
address of an object that resides in the program memory, as shown.
int i, j; // allocated to bank 0 in this example
int x; // allocated to bank 2 in this example
const int type = 0x3456;

int getValue(const int * ip) {

return *ip;
}

void main(void) {
j = getValue(&i);
// ... code that uses j
j = getValue(&x);
// ... code that uses j
j = getValue(&type);
// ... code that uses j
}

 2012-2013 Microchip Technology Inc. DS50002053D-page 157

MPLAB® XC8 C Compiler User’s Guide
Again, the targets to the pointer, ip, are determined, and now the pointer is made of
the class that can access both data and program memory. The generated code to
dereference the pointer will be such that it can determine the required memory space
from the address, and access either space accordingly. Again, this takes place without
any change in the definition of the pointer.
If assembly code references a C pointer, the compiler will force that pointer to become
a 16-bit mixed target space pointer, in the case of baseline or mid-range programs, or
a 24-bit mixed target space pointer, for PIC18 programs. These pointer types have
unrestricted access to all memory areas and will operate correctly, even if assignments
(of a correctly formatted address) are made to the pointer in the assembly code.

[Link] FUNCTION POINTERS

The MPLAB XC8 compiler fully supports pointers to functions, which allows functions
to be called indirectly. These are often used to call one of several function addresses
stored in a user-defined C array, which acts like a lookup table.
For baseline and mid-range devices, function pointers are always one byte in size and
hold an offset into a jump table that is output by the compiler. This jump table contains
jumps to the destination functions.
For PIC18 devices, function pointers are either 16 or 24 bits wide. The pointer size is
purely based on the amount of program memory available on the target device.
As with data pointers, the target assigned to function pointers is tracked. This is an eas-
ier process to undertake compared to that associated with data pointers as all function
instructions must reside in program memory. The pointer reference graph (described in
Section 6.6.5 “Pointer Reference Graph”) will show function pointers, in addition to
data pointers, as well as all their targets. The targets will be names of functions that
could possibly be called via the pointer.
One notable runtime feature for baseline and mid-range devices is that a function
pointer which contains null (the value 0) and is used to call a function indirectly will
cause the code to become stuck in a loop which branches to itself. This endless loop
can be used to detect this erroneous situation. Typically calling a function via a null
function would result in the code crashing or some other unexpected behavior. The
label to which the endless loop will jump is called fpbase.

[Link] SPECIAL POINTER TARGETS

Pointers and integers are not interchangeable. Assigning an integer constant to a
pointer will generate a warning to this effect. For example:
const char * cp = 0x123; // the compiler will flag this as bad code
There is no information in the integer constant, 0x123, relating to the type, size or mem-
ory location of the destination. There is a very good chance of code failure if pointers
are assigned integer addresses and dereferenced, particularly for PIC devices that
have more than one memory space. Is 0x123 an address in data memory or program
memory? How big is the object found at address 0x123?
Always take the address of a C object when assigning an address to a pointer. If there
is no C object defined at the destination address, then define or declare an object at
this address which can be used for this purpose. Make sure the size of the object
matches the range of the memory locations that are to be accessed by the pointer.

DS50002053D-page 158  2012-2013 Microchip Technology Inc.

 C Language Features
For example, a checksum for 1000 memory locations starting at address 0x900 in pro-
gram memory is to be generated. A pointer is used to read this data. You can be
tempted to write code such as:
const char * cp;
cp = 0x900; // what resides at 0x900???
and increment the pointer over the data.
However, a much better solution is this:
const char * cp;
const char inputData[1000] @ 0x900;
cp = &inputData;
// cp is incremented over inputData and used to read values there
In this case, the compiler can determine the size of the target and the memory space.
The array size and type indicates the size of the pointer target, the const qualifier on
the object (not the pointer) indicates the target is located in program memory space.
Note that the const array does not need initial values to be specified in this instance,
see Section [Link] “Const Type Qualifier” and can reside over the top of other
objects at these addresses.
If the pointer has to access objects in data memory, you need to define a different object
to act as a dummy target. For example, if the checksum was to be calculated over 10
bytes starting at address 0x90 in data memory, the following code could be used.
const char * cp;
extern char inputData[10] @ 0x90;
cp = &inputData;
// cp is incremented over inputData and used to read values there
No memory is consumed by the extern declaration, and this can be mapped over the
top of existing objects.
User-defined absolute objects will not be cleared by the runtime startup code and can
be placed over the top of other absolute variables.
Take care when comparing (subtracting) pointers. For example:
if(cp1 == cp2)
; // take appropriate action
The ANSI C standard only allows pointer comparisons when the two pointer targets are
the same object. One exception is that the address can extend to one element past the
end of an array.
Comparisons of pointers to integer constants are even more risky, for example:
if(cp1 == 0x246)
; // take appropriate action
Never compare pointers with integer constants.
A null pointer is the one instance where a constant value can be assigned to a pointer
and this is handled correctly by the compiler. A null pointer is numerically equal to 0
(zero), but this is a special case imposed by the ANSI C standard. Comparisons with
the macro NULL are also allowed.
If null is the only value assigned to a pointer, the pointer will be made as small as
possible.

 2012-2013 Microchip Technology Inc. DS50002053D-page 159

MPLAB® XC8 C Compiler User’s Guide

5.4.6 Constant Types and Formats

A constant is used to represent an immediate value in the source code, as opposed to
a variable that could hold the same value. For example 123 is a constant.
Like any value, a constant must have a C type. In addition to a constant’s type, the
actual value can be specified in one of several formats.

[Link] INTEGRAL CONSTANTS

The format of integral constants specifies their radix. MPLAB XC8 supports the ANSI
standard radix specifiers, as well as ones which enables binary constants to be
specified in C code.
The formats used to specify the radices are given in Table 5-7. The letters used to spec-
ify binary or hexadecimal radices are case insensitive, as are the letters used to specify
the hexadecimal digits.

TABLE 5-7: RADIX FORMATS

Radix Format Example
binary 0b number or 0B number 0b10011010
octal 0 number 0763
decimal number 129
hexadecimal 0x number or 0X number 0x2F

Any integral constant will have a type of int, long int or long long int, so that
the type can hold the value without overflow. Constants specified in octal or hexadeci-
mal can also be assigned a type of unsigned int, unsigned long int or
unsigned long long int if the signed counterparts are too small to hold the value.
The default types of constants can be changed by the addition of a suffix after the digits;
e.g., 23U, where U is the suffix. Table 5-8 shows the possible combination of suffixes
and the types that are considered when assigning a type. So, for example, if the suffix
l is specified and the value is a decimal constant, the compiler will assign the type
long int, if that type will hold the constant; otherwise, it will assigned long long
int. If the constant was specified as an octal or hexadecimal constant, then unsigned
types are also considered.

TABLE 5-8: SUFFIXES AND ASSIGNED TYPES

Suffix Decimal Octal or Hexadecimal
u or U unsigned int unsigned int
unsigned long int unsigned long int
unsigned long long int unsigned long long int
l or L long int long int
long long int unsigned long int
long long int
unsigned long long int
u or U, and l or L unsigned long int unsigned long int
unsigned long long int unsigned long long int
ll or LL long long int long long int
unsigned long long int
u or U, and ll or LL unsigned long long int unsigned long long int

DS50002053D-page 160  2012-2013 Microchip Technology Inc.

 C Language Features
Here is an example of code that can fail because the default type assigned to a
constant is not appropriate:
unsigned long int result;
unsigned char shifter;

void main(void)
{
shifter = 20;
result = 1 << shifter;
// code that uses result
}
The constant 1 (one) will be assigned an int type, hence the result of the shift opera-
tion will be an int. Even though this result is assigned to the long variable, result,
it can never become larger than the size of an int, regardless of how much the con-
stant is shifted. In this case, the value 1 shifted left 20 bits will yield the result 0, not
0x100000.
The following uses a suffix to change the type of the constant, hence ensure the shift
result has an unsigned long type.
result = 1UL << shifter;

[Link] FLOATING-POINT CONSTANT

Floating-point constants have double type unless suffixed by f or F, in which case it
is a float constant. The suffixes l or L specify a long double type which is
considered an identical type to double by MPLAB XC8.

[Link] CHARACTER AND STRING CONSTANTS

Character constants are enclosed by single quote characters, ’, for example ’a’. A
character constant has int type, although this can be later optimized to a char type
by the compiler.
To comply with the ANSI C standard, the compiler does not support the extended char-
acter set in characters or character arrays. Instead, they need to be escaped using the
backslash character, as in the following example.
const char name[] = "Bj\370rk";
printf("%s's Resum\351", name); \\ prints "Bjørk's Resumé"
Multi-byte character constants are not supported by this implementation.
String constants, or string literals, are enclosed by double quote characters “, for exam-
ple “hello world”. The type of string constants is const char * and the character
that make up the string are stored in the program memory, as are all objects qualified
const.
A common warning relates to assigning a string literal to a pointer that does not specify
a const target, for example:
char * cp = "hello world\n";
The string characters cannot be modified, but this type of pointer allows writes to take
place, hence the warning. To prevent yourself from trying to overwrite the string,
qualifier the pointer target as follows. See also Section [Link] “Combining Type
Qualifiers and Pointers”.
const char * cp = "hello world\n";

 2012-2013 Microchip Technology Inc. DS50002053D-page 161

MPLAB® XC8 C Compiler User’s Guide
Defining and initializing an array (i.e., not a pointer) with a string is an exception. For
example:
char ca[]= "hello world\n";
will actually copy the string characters into the RAM array, rather than assign the
address of the characters to a pointer, as in the previous examples. The string literal
remains read-only, but the array is both readable and writable.
The MPLAB XC8 compiler will use the same storage location and label for strings that
have identical character sequences, except where the strings are used to initialize an
array residing in the data space. For example, in the code snippet
if(strncmp(scp, "hello", 6) == 0)
fred = 0;
if(strcmp(scp, "world") == 0)
fred--;
if(strcmp(scp, "hello world") == 0)
fred++;
the characters in the string “world” and the last 6 characters of the string “hello
world” (the last character is the null terminator character) would be represented by
the same characters in memory. The string “hello” would not overlap with the same
characters in the string “hello world” as they differ in terms of the placement of the
null character.
Two adjacent string constants (i.e., two strings separated only by white space) are
concatenated by the compiler. Thus:
const char * cp = "hello" "world";
will assign the pointer with the address of the string “hello world “.

DS50002053D-page 162  2012-2013 Microchip Technology Inc.

 C Language Features

5.4.7 Standard Type Qualifiers

Type qualifiers provide additional information regarding how an object can be used.
The MPLAB XC8 compiler supports both ANSI C qualifiers and additional special qual-
ifiers which are useful for embedded applications and which take advantage of the 8-bit
PIC MCU architecture.

[Link] CONST TYPE QUALIFIER

MPLAB XC8 supports the use of the ANSI type qualifiers const and volatile.
The const type qualifier is used to tell the compiler that an object is read only and will
not be modified. If any attempt is made to modify an object declared const, the com-
piler will issue a warning or error.
User-defined objects declared const are placed in a special psect linked into the pro-
gram space. Objects qualified const can be absolute. The @ address construct is
used to place the object at the specified address in program memory, as in the following
example which places the object tableDef at address 0x100.
const int tableDef[] @ 0x100 = { 0, 1, 2, 3, 4};
Usually a const object must be initialized when it is declared, as it cannot be assigned
a value at any point at runtime. For example:
const int version = 3;
will define version as being an int variable that will be placed in the program mem-
ory, will always contain the value 3, and which can never be modified by the program.
However, uninitialized const objects can be defined and are useful if you need to place
an object in program memory over the top of other objects at a particular location. Usu-
ally uninitialized const objects will be defined as absolute, as in the following example.
const char checksumRange[0x100] @ 0x800;
will define the object checksumRange as a 0x100 byte array of characters located at
address 0x800 in program memory. This definition will not place any data in the HEX
file.

 2012-2013 Microchip Technology Inc. DS50002053D-page 163

MPLAB® XC8 C Compiler User’s Guide
[Link] VOLATILE TYPE QUALIFIER
The volatile type qualifier is used to tell the compiler that an object cannot be guar-
anteed to retain its value between successive accesses. This prevents the optimizer
from eliminating apparently redundant references to objects declared volatile
because it can alter the behavior of the program to do so.
Any SFR which can be modified by hardware or which drives hardware is qualified as
volatile, and any variables which can be modified by interrupt routines should use
this qualifier as well. For example:
volatile static unsigned int TACTL @ 0x160;
The volatile qualifier does not guarantee that any access will be atomic, which is
often not the case with the 8-bit PIC MCU architecture. All these devices can only
access a maximum of 1 byte of data per instruction.
The code produced by the compiler to access volatile objects can be different to
that to access ordinary variables, and typically the code will be longer and slower for
volatile objects, so only use this qualifier if it is necessary. However, failure to use
this qualifier when it is required can lead to code failure.
Another use of the volatile keyword is to prevent variables being removed if they
are not used in the C source. If a non-volatile variable is never used, or used in a
way that has no effect on the program’s function, then it can be removed before code
is generated by the compiler.
A C statement that consists only of a volatile variable’s name will produce code that
reads the variable’s memory location and discards the result. For example the entire
statement:
PORTB;
will produce assembly code the reads PORTB, but does nothing with this value. This is
useful for some peripheral registers that require reading to reset the state of interrupt
flags. Normally such a statement is not encoded as it has no effect.
Some variables are treated as being volatile even though they cannot be qualified
in the source code. See Section [Link] “Undefined Symbols” if you have assem-
bly code in your project.

DS50002053D-page 164  2012-2013 Microchip Technology Inc.

 C Language Features

5.4.8 Special Type Qualifiers

The MPLAB XC8 C Compiler supports special type qualifiers to allow the user to control
placement of static and extern class variables into particular address spaces.

[Link] PERSISTENT TYPE QUALIFIER

By default, any C variables that are not explicitly initialized are cleared on startup. This
is consistent with the definition of the C language. However, there are occasions where
it is desired for some data to be preserved across a Reset.
The persistent type qualifier (or __persistent) is used to qualify variables that
should not be cleared by the runtime startup code.
In addition, any persistent variables will be stored in a different area of memory to
other variables. Different psects are used to hold these objects. See
5.15.2 “Compiler-Generated Psects” for more information.
This type qualifier cannot be used on variables of class auto; however, statically
defined local variables can be qualified persistent. For example, you should write:
void test(void)
{
static persistent int intvar; /* must be static */
// ...
}
If the xc8 option, --STRICT is used, you cannot use the persistent qualifier, but
you can continue to use __persistent.

[Link] NEAR TYPE QUALIFIER

Some of the 8-bit PIC architectures implement data memory which can be always
accessed regardless of the currently selected bank. This common memory can be
used to reduce code size and execution times as the bank selection instructions that
are normally required to access data in banked memory are not required when access-
ing the common memory. PIC18 devices refer to this memory as the access bank mem-
ory. Mid-range and baseline devices have very small amounts of this memory, if it is
present at all. PIC18 devices have substantially more common memory, but the amount
differs between devices. See your device data sheet for more information.
The near type qualifier (or __near) can be used to place global variables in common
memory. This qualifier cannot be used with auto or static local objects.
The compiler automatically uses the common memory for frequently accessed
user-defined variables so this qualifier would only be needed for special memory place-
ment of objects, for example if C variables are accessed in hand-written assembly code
that assumes that they are located in this memory.
This qualifier is controlled by the compiler option --ADDRQUAL, which determines its
effect, see Section 4.8.15 “--ADDRQUAL: Set Compiler Response to Memory
Qualifiers”. Based on this option’s settings, this qualifier can be binding or ignored
(which is the default operation). Qualifiers which are ignored will not produce an error
or warning, but will have no effect.
Here is an example of an unsigned char object qualified as near:
near unsigned char fred;
Note that the compiler can store some temporary objects in the common memory, so
not all of this space can be available for user-defined variables.
If the xc8 option, --STRICT is used, the near qualifier is no longer available, but you
can continue to use __near.

 2012-2013 Microchip Technology Inc. DS50002053D-page 165

MPLAB® XC8 C Compiler User’s Guide
[Link] FAR TYPE QUALIFIER
The far type qualifier (or __far) is used to place global variables into the program
memory space for those PIC18 devices that can support external memory. It will be
ignored when compiling for PIC10/12/16 targets. This qualifier cannot be used with
auto or static local objects.
The compiler assumes that far variables will be located in RAM that is implemented
in the external memory space.
Access of far variables are less efficient than that of internal variables and will result
in larger, slower code.
This qualifier is controlled by the compiler option --ADDRQUAL, which determines its
effect on PIC18 devices, see Section 4.8.15 “--ADDRQUAL: Set Compiler
Response to Memory Qualifiers”. Based on this option’s settings, this qualifier can
be binding or ignored (which is the default operation). Qualifiers which are ignored will
not produce an error or warning, but will have no effect.
Here is an example of an unsigned int object placed into the device’s external
program memory space:
far unsigned int farvar;
If the --STRICT is used, you can only use the __far form of the qualifier.
Note that not all PIC18 devices support external memory in their program memory
space and, thus, the far qualifier is not applicable to all PIC18 devices. On supported
devices, the address range where the additional memory will be mapped must first be
specified with the --RAM option, Section 4.8.49 “--RAM: Adjust RAM Ranges”. For
example, to map additional data memory from 20000h to 2FFFFh use
--RAM=default,+20000-2FFFF.

[Link] BANK0, BANK1, BANK2 AND BANK3 TYPE QUALIFIERS

The bank0, bank1, bank2 and bank3 type qualifiers are recognized by the compiler
and allow some degree of control of the placement of objects in the device’s data mem-
ory banks. When compiling for PIC18 targets, these qualifiers are only accepted for
portability and have no effect on variable placement; on other devices they can be used
to define C objects that are assumed to be located in certain memory banks by
hand-written assembly code. The compiler automatically allocates variables to all data
banks, so these qualifiers are not normally needed.
Although a few devices implement more than 4 banks of data RAM, bank qualifiers to
allow placement into these upper banks are not currently available.
These qualifiers are controlled by the compiler option --ADDRQUAL, which determines
their effect, see Section 4.8.15 “--ADDRQUAL: Set Compiler Response to Memory
Qualifiers”. Based on this option’s settings, these qualifiers can be binding or ignored
(which is the default operation). Qualifiers which are ignored will not produce an error
or warning, but will have no effect.
Objects qualified with any of these qualifiers cannot be auto or parameters to a func-
tion, but can be qualified static, allowing them to be defined locally within a function,
as in:
void myFunc(void) {
static bank1 unsigned char play_mode;
If the xc8 option, --STRICT is used, these qualifiers are changed to 0, 1, 2 and 3.

DS50002053D-page 166  2012-2013 Microchip Technology Inc.

 C Language Features
[Link] EEPROM TYPE QUALIFIER
The eeprom type qualifier (or __eeprom) is recognized by the compiler for baseline
and mid-range devices only and indicates that objects should be placed in the
EEPROM memory. Not all devices implement EEPROM memory. Check your device
data sheet for more information.
Objects qualified with this qualifier cannot be auto or parameters to a function, but can
be qualified static, allowing them to be defined locally within a function, as in:
void myFunc(void) {
static eeprom unsigned char inputData[3];
See Section 5.5.5 “Variables in EEPROM” for more information on these variables
and other ways of accessing the EEPROM.
If the --STRICT option is used, only the __eeprom form of this qualifier is available.

 2012-2013 Microchip Technology Inc. DS50002053D-page 167

MPLAB® XC8 C Compiler User’s Guide

5.5 MEMORY ALLOCATION AND ACCESS

There are two broad groups of RAM-based variables: auto/parameter variables, which
are allocated to some form of stack, and global/static variables, which are positioned
freely throughout the data memory space at static locations. The memory allocation of
these two groups is discussed separately in the following sections.

5.5.1 Address Spaces

All 8-bit PIC devices have a Harvard architecture, which has a separate data memory
(RAM) and program memory space (often flash). Some devices also implement
EEPROM.
The data memory uses banking to increase the amount of available memory (referred
to in the data sheets as the general purpose register file) without having to increase the
assembly instruction width. One bank is “selected” by setting one or more bits in an
SFR. (Consult your device data sheet for the exact operation of the device you are
using.) Most instructions which access a data address use only the offset into the cur-
rently selected bank to access data. The exception is the PIC18 instruction MOVFF,
which takes a full banked address and operates independently of the selected bank.
Some devices only have one bank but many have more than one.
Both the general purpose RAM and SFRs both share the same data space and can
appear in all available memory banks. PIC18 devices have all SFRs in the one data
bank, but mid-range and baseline devices have SFRs at the lower addresses of each
bank. Due to the location of SFRs in these devices, the general purpose memory
becomes fragmented and this limits the size of most C objects.
The Enhanced mid-range devices overcome this fragmentation by allowing a linear
addressing mode, which allows the general purpose memory to be accessed as one
contiguous chunk. Thus, when compiling for these devices, the maximum allowable
size of objects typically increases. Objects defined when using PIC18 devices can also
typically use the entire data memory. See Section [Link].2 “Software Stack
Operation” and Section [Link].2 “Non-Auto Variable Size Limits”.
Many devices have several bytes which can be accessed regardless of which bank is
currently selected. This memory is called common memory. The PIC18 data sheets
refer to the bank in which this memory is stored as the access bank, and hence it is
often referred to as the access bank memory. Since no code is required to select a bank
before accessing these locations, access to objects in this memory is typically faster
and produces smaller code. The compiler always tries to use this memory if possible.
The program memory space is primarily for executable code, but data can also be
located here. There are several ways the different device families locate and read data
from this memory, but all objects located here will be read-only.

DS50002053D-page 168  2012-2013 Microchip Technology Inc.

 C Language Features

5.5.2 Variables in Data Space Memory

Most variables are ultimately positioned into the data space memory. The exceptions
are non-auto variables which are qualified as const, which are placed in the program
memory space, or eeprom qualified variables.
Due to the fundamentally different way in which auto variables and non-auto vari-
ables are allocated memory, they are discussed separately. To use the C language ter-
minology, these two groups of variables are those with automatic storage duration and
those with permanent storage duration, respectively.

Note: The terms “local” and “global” are commonly used to describe variables, but
are not ones defined by the language Standard. The term “local variable” is
often taken to mean a variable which has scope inside a function, and
“global variable” is one which has scope throughout the entire program.
However, the C language has three common scopes: block, file (i.e., inter-
nal linkage) and program (i.e., external linkage), so using only two terms to
describe these can be confusing. For example, a static variable defined
outside a function has scope only in that file, so it is not globally accessible,
but it can be accessed by more than one function inside that file, so it is not
local to any one function, either.

[Link] NON-AUTO VARIABLE ALLOCATION

Non-auto variables (those with permanent storage duration) are located by the com-
piler into any of the available data banks. This is done in a two-stage process: placing
each variable into an appropriate psect and later linking that psect into a predetermined
bank. See Section 5.15.1 “Program Sections” for an introductory guide to psects.
Thus, during compilation, the code generator can determine which bank will hold each
variable and encode the output accordingly, but it will not know the exact location within
that bank.
The compiler will attempt to locate all variables in one bank (i.e., place all variables in
the psect destined for this bank), but if this fills (i.e., if the compiler detects that the psect
has become too large for the free space in a bank), variables will be located in other
banks via different psects. Qualifiers are not required to have these variables placed in
banks other than bank 0 but can be used if you want to force a variable to a particular
bank. See “--RAM=default,+20000-2FFFF.” and Section 4.8.15 “--ADDRQUAL: Set
Compiler Response to Memory Qualifiers” for more information on how to do this.
If common memory is available on the target device, this will also be considered for
variables. This memory can be limited in size and can be reserved for special use, so
only a few variables can be allocated to it.

 2012-2013 Microchip Technology Inc. DS50002053D-page 169

MPLAB® XC8 C Compiler User’s Guide
The compiler considers three categories of non-auto variables, which all relate to the
value the variable should contain by the time the program begins. Each variable cate-
gory has a corresponding psect which is used to hold the output code which reserves
memory for each variable. The base name of each psect category is tabulated below.
A full list of all psect names are in Section 5.15.2 “Compiler-Generated Psects”.
nv These psects are used to store variables qualified persistent, whose val-
ues should not be altered by the runtime startup code. They are not cleared
or otherwise modified at startup.
bss These psects contain any uninitialized variables, which are not assigned a
value when they are defined, or variables which should be cleared by the
runtime startup code.
data These psects contain the RAM image of any initialized variables, which are
assigned a non-zero initial value when they are defined and which must have
a value copied to them by the runtime startup code.
As described in Section 5.10 “Main, Runtime Startup and Reset”, the base name of
data space psects is always used in conjunction with a linker class name to indicate the
RAM bank in which the psect will be positioned. This section also lists other variants of
these psects and indicates where these psect must be linked. See also
Section 5.15.2 “Compiler-Generated Psects” for more information on how initial
values are assigned to the variables.
Note that the data psect used to hold initialized variables is the psect that holds the
RAM variables themselves. There is a corresponding psect (called idata) that is
placed into program memory (so it is non-volatile) and which is used to hold the initial
values that are copied to the RAM variables by the runtime startup code.
All non-auto variables, except for static variables, discussed in
Section [Link].1 “Static Variables”, always use their lexical name with a leading
underscore character as the assembly identifier used for this object. See
Section [Link] “Equivalent Assembly Symbols” for more information on the
mapping between C- and assembly-domain symbols.

[Link].1 Static Variables

All static variables have permanent storage duration, even those defined inside a
function which are “local static” variables. Local static variables only have scope in
the function or block in which they are defined, but unlike auto variables, their memory
is reserved for the entire duration of the program. Thus they are allocated memory like
other non-auto variables.
Static variables can be accessed by other functions via pointers since they have
permanent duration.
Variables which are static are guaranteed to retain their value between calls to a
function, unless explicitly modified via a pointer.
Variables which are static and which are initialized only have their initial value
assigned once during the program’s execution. Thus, they can be preferable over ini-
tialized auto objects which are assigned a value every time the block in they are
defined begins execution. Any initialized static variables are initialized in the same
way as other non-auto initialized objects by the runtime startup code, see
Section 4.4.2 “Startup and Initialization”.
The assembly symbols used to access static objects in assembly code are discussed
in Section [Link] “Equivalent Assembly Symbols”.

DS50002053D-page 170  2012-2013 Microchip Technology Inc.

 C Language Features
[Link].2 Non-Auto Variable Size Limits
Arrays of any type (including arrays of aggregate types) are fully supported by the com-
piler. So too are the structure and union aggregate types, see 5.4.4 “Structures and
Unions”. These objects can often become large in size and can affect memory
allocation.
When compiling for enhanced mid-range PIC devices, the size of an object (array or
aggregate object) is typically limited only by the total available data memory. Single
objects that will not fit into any of the available general purpose RAM ranges will be
allocated memory in several RAM banks and accessed using the device’s linear GPR
(general purpose RAM).
Note that the special function registers (which reside in the data memory space) or
memory reservations in general purpose RAM can prevent objects from being allo-
cated contiguous memory in the one bank. In this case objects that are smaller than
the size of a RAM bank can also be allocated across multi-banks. The generated code
to access multi-bank objects will always be slower and the associated code size will be
larger than for objects fully contained within a single RAM bank.
When compiling for PIC18 devices, the size of an object is also typically limited only by
the data memory available. Objects can span several data banks.
On baseline and other mid-range devices, arrays and structures are limited to the max-
imum size of the available GPR memory in each RAM bank, not the total amount of
memory remaining. An error will result if an array is defined which is larger than this
size.
With any device, reserving memory in general purpose RAM (see
Section 4.8.49 “--RAM: Adjust RAM Ranges”), or defining absolute variables in the
middle of data banks (see Section 5.5.4 “Absolute Variables”), further restricts the
contiguous memory in the data banks and can reduce the maximum size of objects you
can define.

[Link].3 Changing the Default Non-Auto Variable Allocation

There are several ways in which non-auto variables can be located in locations other
than those chosen by the compiler.
Variables can be placed in other memory spaces by the use of qualifiers. For example
if you wish to place variables in the program memory space, then the const specifier
should be used (see Section [Link] “Const Type Qualifier”). The eeprom qualifier
(see [Link] “Eeprom Type Qualifier”) can be used to allocate variables to the
EEPROM, if such memory exists on your target device.
If you wish to prevent variables from using one or more data memory locations so that
these locations can be used for some other purpose, you are best reserving the mem-
ory using the memory adjust options. See Section 4.8.49 “--RAM: Adjust RAM
Ranges” for information on how to do this.
If only a few non-auto variables are to be located at specific addresses in data space
memory, then the variables can be made absolute. This allows individual variables to
be explicitly positioned in memory at an absolute address. Absolute variables are
described in Section 5.5.4 “Absolute Variables”. Once variables are made absolute,
their address is hard coded in generated output code, they are no longer placed in a
psect and do not follow the normal memory allocation procedure.

 2012-2013 Microchip Technology Inc. DS50002053D-page 171

MPLAB® XC8 C Compiler User’s Guide
The psects in which the different categories of non-auto variables (the nv, bss and
data psects described in Section [Link] “Non-Auto Variable Allocation”) can be
shifted as a whole by changing the default linker options. So, for example, you could
move all the persistent variables. However, typically these psects can only be moved
within the data bank in which they were allocated by default. See Section 5.10 “Main,
Runtime Startup and Reset”, for more information on changing the default linker
options for psects. The code generator makes assumptions as to the location of these
psects and if you move them to a location that breaks these assumptions, code can fail.
Non-auto can also be placed at specific positions by using the psect pragma, see
Section [Link] “The #pragma psect Directive”. The decision whether variables
should be positioned this way or using absolute variables should be based on the
location requirements.

[Link] AUTO VARIABLE ALLOCATION AND ACCESS

This section discusses allocation of auto variables (those with automatic storage dura-
tion) to a data stack. This also includes function parameter variables, which behave like
auto variables, in terms of their storage duration and scope. Temporary variables
defined by the compiler also fall into this group. They are identical to auto variables,
except they are defined by the compiler and, hence, have no C name. Together, these
objects are often called stack-based objects.
The auto (short for automatic) variables are the default type of local variable. Unless
explicitly declared to be static, a local variable will be made auto. The auto key-
word can be used if desired.
The auto variables, as their name suggests, automatically come into existence when
a function is executed, then disappear once the function returns. Since they are not in
existence for the entire duration of the program, there is the possibility to reclaim mem-
ory they use when the variables are not in existence and allocate it to other variables
in the program.
Typically such variables are stored on some sort of a dynamic data stack where mem-
ory can be easily allocated and deallocated by each function. This is not possible on all
8-bit devices supported by MPLAB XC8. Nor is it the most efficient means of storing
objects.
MPLAB XC8 has two methods of implementing data stacks for stack-based variables:
a compiled stack and a software stack1. Section 5.3.4 “Stacks” describes all the
stacks used by MPLAB XC8 and the 8-bit PIC devices.
Each C function is compiled to use exactly one of these stacks. The stack used affects
whether a function allows reentrancy. If a function is encoded to place its stack-based
objects on the software stack, it is said to be using a reentrant function model. A
function uses a non-reentrant function model if it places its stack-based objects on the
compiled stack. This information is summarized in Table 5-9 along with the devices that
support each model. The function model directly implies the stack used by a function.
See subsections below for specific details on how the compiled stack and software
stack operate.

TABLE 5-9: FUNCTION MODELS IMPLEMENTATION

Function Model Data stack used Supported device families
Non-reentrant Compiled stack All devices
Reentrant Software stack Enhanced mid-range and
PIC18 devices

1. What is referred to as a software stack in this user’s guide is the typical dynamic stack arrangement
employed by most computers and is ordinary data memory accessed by some sort of push and pop
instructions, and a stack pointer register.

DS50002053D-page 172  2012-2013 Microchip Technology Inc.

 C Language Features
When compiling for those devices that do not support the reentrant function model, all
functions are encoded to use the compiler stack, and these functions are
non-reentrant.
For the enhanced mid-range and PIC18 devices, by default the compiler will use the
non-reentrant model for all functions. Alternatively the user can dictate which functions
are to be compiled reentrantly (and those which are not) by using compiler options or
function specifiers. There is also a hybrid stack mode which allows the compiler to
choose which functions need to be compiled using a reentrant model and which can
use the non-reentrant model. The hybrid mode allows the program to use recursion but
still take advantage of the more efficient compiled stack.
The --STACK option (see Section 4.8.56 “--STACK: Specify Data Stack Type For
Entire Program”) can be used to change the compiler’s default behavior when assign-
ing function models. Set the --STACK option to software so the compiler will always
choose the reentrant model (software stack) for each function. If the --STACK option
is set to compiled or this option is omitted, all functions are encoded to use the
non-reentrant (compiled stack) function model. Set this option to hybrid for hybrid
stack mode and to allow the compiler to decide how each function should be encoded.
In hybrid mode the compiler will choose a function model based on how the function is
called in the program. If the function is not reentrantly called, then it will be encoded to
use the non-reentrant model and the compiled stack. If the function appears in more
than one call graph (i.e., it is called from main-line and interrupt code), or it appears in
a loop in a call graph (i.e., it is called recursively), then the compiler will use the
reentrant model.
The --STACK option’s software and compiled settings changes the function model
for all functions. You can change the function model for individual functions by using
function specifiers when you define the function.
Use either the compiled or nonreentrant specifier (identical meanings) to indicate
that the specified function must use the compiled stack, without affecting any other
function. Alternatively, use either the software or reentrant specifier to indicate a
function must be encoded to use the software stack.
The function specifiers have precedence over the --STACK option setting. If, for exam-
ple, the option --STACK=compiled has been used, but one function uses the
software (or reentrant) specifier, then the specified function will use the software
stack and all the remaining functions will use the compiled stack. These functions
specifiers also override any choice made by the compiler in hybrid mode.
If a function has been specified as compiled (or nonreentrant), or the
--STACK=compiled option has been issued, and that function appears in more than
one call graph in the program, then the usual function duplication feature automatically
comes into effect. See Section 5.9.6 “Function Duplication”, for more information on
how this is performed. Duplicating a non-reentrant function allows it to be called from
multiple call graphs, but cannot be used if the function is called recursively.
The auto variables defined in a function will not necessarily be allocated memory in
the order declared, in contrast to parameters which are always allocated memory
based on their lexical order. In fact, auto variables for one function can be allocated in
many RAM banks.
The standard qualifiers: const and volatile can both be used with auto variables
and these do not affect how they are positioned in memory. This implies that a local
const-qualified object is still an auto object and, as such, will be allocated memory in
the compiled stack in the data space memory, not in the program memory like with
non-auto const objects.

 2012-2013 Microchip Technology Inc. DS50002053D-page 173

MPLAB® XC8 C Compiler User’s Guide
The compiler will try to locate the stack in one data bank, but if this fills (i.e., if the com-
piler detects that the stack psect has become too large), it can build up the stack into
several components (each with their own psect) and link each in a different bank.
Each auto object is referenced in assembly code using a special symbol defined by
the code generator. If accessing auto variables defined in C source code, you must use
these symbols, which are discussed in Section 5.12.3 “Interaction between
Assembly and C Code”.
[Link].1 Compiled Stack Operation
A compiled stack consists of fixed memory areas that are usable by each function’s
stack-based variables. When a compiled stack is used, functions are not re-entrant
since stack-based variables in each function will use the same fixed area of memory
every time the function is invoked.
Fundamental to the generation of the compiled stack is the call graph, which defines a
tree-like hierarchy of function calls, i.e it shows what functions can be called by each
function.
There will be one graph produced for each root function. A root function is typically not
called, but which is executed via other means and contains a program entry point. The
function main() is an example of a root function that will be in every project. Interrupt
functions which are executed when a hardware interrupt occurs, are another example.

FIGURE 5-1: FORMATION OF CALL GRAPH

main {
F1(…);
F2(…); Call graph
F3(…);
} main
F1
F4
code F2
F1 { generator F3
F4(…); isr
} F5
F6

isr {
F5(…); Analysis of program
F6(…);
}

Figure 5-1 shows sections of a program being analyzed by the code generator to form
a call graph. In the original source code, the function main() calls F1(), F2() and
F3(). F1() calls F4(), but the other two functions make no calls. The call graph for
main() indicates these calls. The symbols F1, F2 and F3 are all indented one level
under main. F4 is indented one level under F1.

DS50002053D-page 174  2012-2013 Microchip Technology Inc.

 C Language Features
This is a static call graph which shows all possible calls. If the exact code for function
F1() looked like:
int F1(void) {
if(PORTA == 44)
return F4();
return 55;
}
the function F4() will always appear in the call graph, even though it is conditionally
executed in the actual source code. Thus, the call graph indicates all functions that
might be called.
In the diagram, there is also an interrupt function, isr(), and it too has a separate
graph generated.
The term main-line code is often used, and refers to any code that is executed as a
result of the main() function being executed. In the above figure, F1(), F2(), F3()
and F4() are only ever called by main-line code.
The term interrupt code refers to any code that is executed as a result of an interrupt
being generated, in the above figure, F5() and F6() are called by interrupt code.
Figure 5-2 graphically shows an example of how the compiled stack is formed.

FIGURE 5-2: FORMATION OF THE COMPILED STACK

Formation of auto-parameter block (APB)

1
for function F2

a
int aa , int
F2(int int bb ) { F2
b
char c ; c
} 2 Analysis of call graph

main
F1
main F4
F3
F 3 F2
1
F1
compiled

F3
stack

F2
F 2
F4
4 isr
F5
isr
F6
F5
5 F6
F 6

Overlap of non-concurrently active APBs

3
to form compiled stack

Each function in the program is allocated a block of memory for its parameter, auto
and temporary variables. Each block is referred to as an auto-parameter block (APB).
The figure shows the APB being formed for function F2(), which has two parameters,
a and b, and one auto variable, c.
The parameters to the function are first grouped in an order strictly determined by the
lexical order in which they appear in the source code. These are then followed by any
auto objects; however, the auto objects can be placed in any order. So we see
memory for a is followed by that for b and lastly c.

 2012-2013 Microchip Technology Inc. DS50002053D-page 175

MPLAB® XC8 C Compiler User’s Guide
Once these variables have been grouped, the exact location of each object is not
important at this point and we can represent this memory by one block — the APB for
this function.
The APBs are formed for all functions in the program. Then, by analyzing the call graph,
these blocks are assigned positions, or base values, in the compiled stack.
Memory can be saved if the following point is observed: If two functions are never
active at the same time, then their APBs can be overlapped.
In the example shown in the figure, F4() and F1() are active at the same time, in fact
F1() calls F4(). However, F2(), F3() and F1() are never active at the same time;
F1() must return before F2() or F3() can be called by main(). The function main()
will always be active and so its APB can never overlap with that of another function.
In the compiled stack, you can see that the APB for main() is allocated unique mem-
ory. The blocks for F1(), F2() and F3() are all placed on top of each other and the
same base value in the compiled stack; however, the memory taken up by the APBs
for F1() and F4() are unique and do not overlap.
Our example also has an interrupt function, isr(), and its call graph is used to assem-
ble the APBs for any interrupt code in the same way. Being the root of a graph, isr()
will always be allocated unique memory, and the APBs for interrupt functions will be
allocated memory following.
The end result is a block of memory which forms the compiled stack. This block can
then be placed into the device’s memory by the linker.
Once auto variables have been allocated a relative position in the compiled stack, the
stack itself is then allocated memory in the data space. This is done is a similar fashion
to the way other variables are assigned memory: a psect is used to hold the stack and
this psect is placed into the available data memory by the linker. The psect base name
used to hold the compiled stack is called cstack, and, like with other psects, the base
name is always used in conjunction with a linker class name to indicate the RAM bank
in which the psect will be positioned. See Section 5.15.2 “Compiler-Generated
Psects” for the limitations associated with where this psect can be linked.
For devices with more than one bank of data memory, the compiled stack can be built
up into components, each located in a different memory bank. The compiler will try to
allocate the compiled stack in one bank, but if this fills, it will consider other banks. The
process of building these components of the stack is the same, but each function can
have more than one APB and these will be allocated to one of the stack components
based on the remaining memory in the component’s destination bank.
Human readable symbols are defined by the code generator which can be used to
access auto and parameter variables in the compiled stack from assembly code, if
required. See Section 5.12.3 “Interaction between Assembly and C Code” for full
information between C domain and assembly domain symbols.
[Link].2 Software Stack Operation
Functions using a software stack (reentrant model) dynamically allocate memory for
their stack-based variables in a region of memory specifically reserved for this software
stack.
Allocation starts at one end of this reserved area, and the stack memory grows as new
function instances come into existence. When a function using the reentrant model
exits, any stack memory it used is freed and made available for other functions. The
stack grows up in memory, toward larger addresses.
Main-line code and each interrupt routine use unique areas in the stack space. The
maximum size allocated to each area can be specified using the --STACK option, see
Section 4.8.56 “--STACK: Specify Data Stack Type For Entire Program”.

DS50002053D-page 176  2012-2013 Microchip Technology Inc.

 C Language Features
MPLAB XC8 designates a register, known as the stack pointer, which always holds the
address of the next free location in the software stack. The register used by the stack
pointer is FSR1 for both enhanced mid-range and PIC18 devices. The address held by
the stack pointer is increased when variables are allocated (pushed) to the stack; it is
decreased when a function returns and variables are removed (popped) from the stack.
Note that if there are any functions in the program that are reentrantly encoded, the
FSR1 register is reserved for the stack pointer for the entire program’s duration, even
when executing code associated with non-reentrant functions. With this register
unavailable for use with general statements, the code generated may be less efficient
or "Can’t generate code" errors may result.
The stack pointer is reloaded when an interrupt occurs so it accesses the interrupt func-
tion’s unique stack area. It is restored by the interrupt context switch code when the
interrupt routine is complete.
There is no register assigned to hold a frame pointer. All access of stack-based objects
must use an address that is an offset from the stack pointer.
When a function is called, any arguments to that function are pushed onto the stack by
the calling function, in a reverse order to that in which the corresponding parameters
appear in the function’s prototype. If required, the called function will increase the value
stored in the stack pointer to allocate storage for any auto or temporary variables it
needs to allocate.
If the reentrant function returns a value on the stack (this might happen for return values
larger than 4 bytes in size), the calling function will adjust the stack to remove the return
value.
Recall that a function’s return address is not stored on this stack. It is automatically
stored on the hardware stack by the device, see Section [Link] “Function Return
Address Stack”.
The compiler can detect if the software stack memory requirements for each function
will exceed set limits. These limits are 256 bytes for PIC18 devices and typically 31
bytes for enhanced mid-range devices. Note that the compiler cannot detect for over-
flow of the memory reserved for the stack as a whole. There is no runtime check made
for software stack overflows. If the software stack overflows, data corruption and code
failure might result.

[Link].3 Size Limits of Auto Variables

The compiled stack is built up as one contiguous block which can be placed into one
of the available data banks. However, if the stack becomes too large for this space, it
can be assembled into several blocks, with each block being positioned in a different
bank of memory. Thus the total size of the stack is roughly limited only by the available
memory on the device.
Unlike with non-auto variables, it is not efficient to access auto variables within the
compiled stack using the linear memory of enhanced mid-range devices. For all
devices, including PIC18 and Enhanced mid-range PIC MCUs, each component of the
compiled stack must fit entirely within one bank of data memory on the target device
(however, you can have more than one component, each allocated to a different bank).
This limits the size of objects within the stack to the maximum free space of the bank
in which it is allocated. The more auto variables in the stack; the more restrictive the
space is to large objects. Recall that SFRs on mid-range devices are usually present
in each data bank, so the maximum amount of GPR available in each bank is typically
less than the bank size for these devices.

 2012-2013 Microchip Technology Inc. DS50002053D-page 177

MPLAB® XC8 C Compiler User’s Guide
The software stack is always allocated one block of memory. This memory may cross
bank boundaries. The size is typically limited by the amount of free data space
remaining. An auto object placed on the software stack may be any size, providing it
fits in the allocated stack space. It may be allocated memory that crosses a bank
boundary and will always be accessed via the stack pointer, FSR1.
If a program requires large objects that should not be accessible to the entire program,
consider leaving them as local objects, but using the static specifier. Such variables
are still local to a function, but are no longer auto and have fewer size limitations. They
are allocated memory as described in Section [Link] “Non-Auto Variable
Allocation”.
[Link].4 Changing the Default Auto Variable Allocation
As auto variables are placed in a stack, there is no means to move individual vari-
ables. They cannot be made absolute, nor can they be moved using the #pragma
psect directive.
The psects in which the auto variables reside can be shifted as a whole by changing
the default linker options. However, these psects can only be moved within the data
bank in which they were allocated by default. See Section 5.10 “Main, Runtime
Startup and Reset” for more information on changing the default linker options for
psects. The code generate makes assumptions as to the location of these psects and
if you move them to a location that breaks these assumptions, code can fail.

5.5.3 Variables in Program Space

The only variables that are placed into program memory are those that are not auto
and which have been qualified const. Any auto variables qualified const are placed
in the compiled stack along with other auto variables, and all components of the com-
piled stack will only ever be located in the data space memory.
Any const-qualified (auto or non-auto) variable will always be read-only and any
attempt to write to these in your source code will result in an error being issued by the
compiler.
On some 8-bit PIC devices, the program space is not directly readable by the device.
For these devices, the compiler stores data in the program memory by means of RETLW
instructions which can be called, and which will return a byte of data in the W register.
The compiler will generate the code necessary to make it appear that program memory
is being read directly.
Enhanced mid-range PIC devices can directly read their program memory, although
the compiler will still usually store data as RETLW instructions. This way the compiler
can either produce code that can call these instructions to obtain the program memory
data as with the ordinary mid-range devices, or directly read the operand to the instruc-
tion (the LSB of the RETLW instruction). The most efficient access method can be
selected by the compiler when the data needs to be read.

DS50002053D-page 178  2012-2013 Microchip Technology Inc.

 C Language Features
Data can be stored as individual bytes in the program memory of PIC18 devices. This
can be read using table read instructions.
On all devices, accessing data located in program memory is much slower than
accessing objects in the data memory. The code associated with the access is also
larger.
A const object is usually defined with initial values, as the program cannot write to
these objects at runtime. However, this is not a requirement. An uninitialized const
object can be defined to define a symbol, or label, but not make a contribution to the
output file. Uninitialized const objects are often made absolute, see
Section 5.5.4 “Absolute Variables”. Here are examples of const object definitions.
const char IOtype = ’A’; // initialized const object
const char buffer[10]; // I just define a label
The data held by non-auto const variables is placed in one of several psects, based
on the target device. See Section 5.15.2 “Compiler-Generated Psects” for the
limitations associated with where these psects can be linked.
See Section 5.12.3 “Interaction between Assembly and C Code” for the equivalent
assembly symbols that are used to represent const-qualified variables in program
memory.

[Link] SIZE LIMITATIONS OF CONST VARIABLES

Arrays of any type (including arrays of aggregate types) can be qualified const and
placed in the program memory. So too can structure and union aggregate types, see
5.4.4 “Structures and Unions”. These objects can often become large in size and can
affect memory allocation.
For baseline PIC devices, the maximum size of a single const object is 255 bytes.
However, you can define as many const objects as required provided the total size
does not exceed the available program memory size of the device. Note that as well as
other program code, there is also code required to be able to access const-qualified
data in the program memory space. Thus, you can need additional program memory
space over the size of the object itself. This additional code to access the const data
is only included once, regardless of the amount or number of const-qualified objects.
For all other 8-bit devices, the maximum size of a const-qualified object is limited only
by the available program memory. These devices also use additional code that
accesses the const data. PIC18 devices need additional code each time an object is
accessed, but this is typically small. The mid-range devices include a larger routine, but
this code is also only included once, regardless of the amount or number of
const-qualified objects.

 2012-2013 Microchip Technology Inc. DS50002053D-page 179

MPLAB® XC8 C Compiler User’s Guide
[Link] CHANGING THE DEFAULT ALLOCATION
If you only intend to prevent all variables from using one or more program memory loca-
tions so that you can use those locations for some other purpose, you are best reserv-
ing the memory using the memory adjust options. See Section 4.8.50 “--ROM: Adjust
ROM Ranges” for information on how to do this.
If only a few non-auto const variables are to be located at specific addresses in pro-
gram space memory, then the variables can be made absolute. This allows individual
variables to be explicitly positioned in memory at an absolute address. Absolute vari-
ables are described in Section 5.5.4 “Absolute Variables”. Once variables are made
absolute, their address is hard coded in generated output code, they are no longer
placed in a psect and do not follow the normal memory allocation procedure.
The psects in which the different categories of non-auto const variables can be
shifted as a whole by changing the default linker options. However, there are limitations
in where these psects can be moved to. See Section 5.10 “Main, Runtime Startup
and Reset” for more information on changing the default linker options for these
psects.
Variables in program memory can also be placed at specific positions by using the
psect pragma, see Section [Link] “The #pragma psect Directive”. The decision
whether variables should be positioned this way or using absolute variables should be
based on the location requirements.

DS50002053D-page 180  2012-2013 Microchip Technology Inc.

 C Language Features

5.5.4 Absolute Variables

Most variables can be located at an absolute address by following its declaration with
the construct @ address, where address is the location in memory where the
variable is to be positioned. Such a variable is known as an absolute variable.

[Link] ABSOLUTE VARIABLES IN DATA MEMORY

Absolute variables are primarily intended for equating the address of a C identifier with
a special function register, but can be used to place ordinary variables at an absolute
address in data memory.
For example:
volatile unsigned char Portvar @ 0x06;
will declare a variable called Portvar located at 06h in the data memory. The compiler
will reserve storage for this object (if the address falls into general-purpose RAM) and
will equate the variable’s identifier to that address.
The auto variables cannot be made absolute as they are located in a stack. Nor can
you make static local objects absolute. The compiler does not make any checks for
overlap of absolute variables with other absolute variables, so this must be considered
when choosing the variable locations. There is no harm in defining more than one
absolute variable to live at the same address if this is what you require. The compiler
will not locate ordinary variables over the top of absolutes, so there is no overlap
between these objects.

Note: Defining absolute objects can fragment memory and can make it impossi-
ble for the linker to position other objects. Avoid absolute objects if at all
possible. If absolute objects must be defined, try to place them at either end
of a memory bank or page so that the remaining free memory is not
fragmented into smaller chunks.

When defining absolute bit variables (see Section [Link] “Bit Data Types and Vari-
ables”), the address specified must be a bit address. A bit address is obtained by mul-
tiplying the desired byte address by 8, then adding the bit offset within that bit. So, for
example, to place a bit variable called mode at bit position #2 at byte address 0x50,
use the following:
bit mode @ 0x282;
If you wish to place a bit variable over an existing object (typically this will be an SFR
variable or another absolute variable) then you can use the symbol of that object, as in
the following example which places flag at bit position #3 in the char variable
MOT_STATUS:
bit flag @ ((unsigned) &MOT_STATUS)*8 + 3;
When compiling for an enhanced mid-range PIC device, the memory allocated for
some objects can be spread over multiple RAM banks. Such objects will only ever be
accessed indirectly in assembly code, and will use the linear GPR memory imple-
mented on these devices. A linear address (which can be mapped back to the ordinary
banked address) will be used with these objects internally by the compiler.

 2012-2013 Microchip Technology Inc. DS50002053D-page 181

MPLAB® XC8 C Compiler User’s Guide
The address specified for absolute objects on these devices can either be the tradi-
tional banked memory address or the linear address. As the linear addresses start
above the largest banked address, it is clear which address is intended. In the following
example:
int inputBuffer[100] @ 0x2000;
it is clear that inputBuffer should placed at address 0x2000 in the linear address
space, which is address 0x20 in bank 0 RAM in the traditional banked address space.
See the device data sheet for exact details regarding your selected device.
Absolute variables in RAM cannot be initialized when they are defined. Define the
absolute variables, then assign them a value at a suitable point in your main-line code.

[Link] ABSOLUTE OBJECTS IN PROGRAM MEMORY

Non-auto objects qualified const can also be made absolute in the same way,
however, the address will indicate an address in program memory. For example:
const int settings[] @ 0x200 = { 1, 5, 10, 50, 100 };
will place the array settings at address 0x200 in the program memory.
Both initialized and uninitialized const objects can be made absolute. That latter is
useful when you only need to define a label in program memory without making a
contribution to the output file.
Variables can also be placed at specific positions by using the psect pragma, see
Section [Link] “The #pragma psect Directive”. The decision whether variables
should be positioned this way or using absolute variables should be based on the loca-
tion requirements. Using absolute variables is the easiest method, but only allows
placement at an address which must be known prior to compilation. The psect
pragma is more complex, but offers all the flexibility of the linker to position the new
psect into memory. You can, for example, specify that variables reside at a fixed
address, or that they be placed after other psects, or that the they be placed anywhere
in a compiler-defined or user-defined range of address.

DS50002053D-page 182  2012-2013 Microchip Technology Inc.

 C Language Features

5.5.5 Variables in EEPROM

For devices with on-chip EEPROM, the compiler offers several methods of accessing
this memory. You can defined named variables in this memory space, or use
block-access routines to read or write EEPROM. The EEPROM access methods are
described in the following sections.

[Link] EEPROM VARIABLES

When compiling for baseline and mid-range parts, the eeprom qualifier allows you to
create named C variables that reside in the EEPROM space. See
Section [Link] “Eeprom Type Qualifier”.
Variables qualified eeprom are cleared or initialized, just like ordinary RAM-based vari-
ables; however, the initialization process is not carried out by the runtime startup code.
Initial values are placed into the HEX file and are burnt into the EEPROM when you
program the device. Thus, if you modify the EEPROM during program execution and
then reset the device, these variables will not contain the initial values specified in your
code at startup up.
The following example defines two arrays in EEPROM.
eeprom char regNumber[10] = "A93213";
eeprom int lastValues[3];
For both these objects, initial values will be placed into psects and will appear in the
HEX file. Zeros will be used as the initial values for lastValues.
The generated code to access eeprom-qualified variables will be much longer and
slower than code to access RAM-based variables. You should avoid using
eeprom-qualified variables in complicated expressions. Consider copying values from
the EEPROM to regular RAM-based variables and using these in your code.

[Link] EEPROM INITIALIZATION

For those devices that support external programming of their EEPROM data area, the
__EEPROM_DATA() macro can be used to place initial values into the HEX file ready
for programming. The macro is used as follows.
#include <xc.h>
__EEPROM_DATA(0, 1, 2, 3, 4, 5, 6, 7);
The macro has eight parameters, representing eight data values. Each value should be
a byte in size. Unused values should be specified with zero.
The __EEPROM_DATA() macro expands into in-line assembly code. If expressions are
used to evaluate the macro arguments, ensure that any operators or tokens in these
expressions are written in assembly code (see Section 6.4 “MPLAB XC8 Assembly
Language”).
The macro can be called multiple times to define the required amount of EEPROM
data. It is recommended that the macro be placed outside any function definition.
This macro cannot used to write to EEPROM locations during runtime; it is used for
pre-loading EEPROM contents at program time only.
The values defined by this macro share the EEPROM space with any eeprom-qualified
variables. The macro cannot be used to initialize eeprom-qualified variables. The psect
used by this macro to hold the data values is different to those used by eeprom-qual-
ified variables. The link order of these psects can be adjusted, if required (see
Section 4.8.6 “-L-: Adjust Linker Options Directly”).
For convenience, the macro _EEPROMSIZE represents the number of bytes of
EEPROM available on the target device.

 2012-2013 Microchip Technology Inc. DS50002053D-page 183

MPLAB® XC8 C Compiler User’s Guide
[Link] EEPROM ACCESS FUNCTIONS
The library functions eeprom_read() and eeprom_write(), can be called to read
from, and write to, the EEPROM during program execution. On PIC18 devices, these
functions are derived from the peripheral library. The prototypes for these functions are
as below.
#include <xc.h>
unsigned char eeprom_read(unsigned char address);
void eeprom_write(unsigned char address, unsigned char value);
These functions test and wait for any concurrent writes to EEPROM to conclude before
performing the required operation. The eeprom_write() function will initiate the pro-
cess of writing to EEPROM and this process will not have completed by the time that
eeprom_write() returns. The new data written to EEPROM will become valid at a
later time. See your device data sheet for exact information about EEPROM on your
target device.
It can also be convenient to use the preprocessor symbol, _EEPROMSIZE, in conjunc-
tion with some of these access methods. This symbol defines the number of EEPROM
bytes available for the selected chip.

[Link] EEPROM ACCESS MACROS

Macro version of the EEPROM functions are also provided. The PIC18 version of these
macros purely call the function equivalents. Those for other 8-bit PIC devices perform
similar operations to their function counterparts, with the exception of some timing
issues described below. Use the macro forms of these routines for faster execution and