# TMS320C28x Optimizing C/C++ Compiler

TI最新应用解决方案

###### 工业电子汽车电子个人消费电子

DSP相关资料

TMS320C28x Optimizing C/C++ Compiler 是TMS320C28x 的組譯器使用指導書

TMS320C28x Optimizing C/C++ Compiler v17.6.0.STS User's Guide Literature Number: SPRU514N June 2017 Contents Preface....................................................................................................................................... 10 1 Introduction to the Software Development Tools.................................................................... 13 1.1 Software Development Tools Overview ................................................................................. 14 1.2 Compiler Interface.......................................................................................................... 15 1.3 ANSI/ISO Standard ........................................................................................................ 16 1.4 Output Files ................................................................................................................. 16 1.5 Utilities ....................................................................................................................... 16 2 Using the C/C++ Compiler ................................................................................................... 17 2.1 About the Compiler......................................................................................................... 18 2.2 Invoking the C/C++ Compiler ............................................................................................. 18 2.3 Changing the Compiler's Behavior with Options ....................................................................... 19 2.3.1 Linker Options ..................................................................................................... 26 2.3.2 Frequently Used Options......................................................................................... 28 2.3.3 Miscellaneous Useful Options ................................................................................... 29 2.3.4 Run-Time Model Options......................................................................................... 31 2.3.5 Symbolic Debugging and Profiling Options .................................................................... 32 2.3.6 Specifying Filenames ............................................................................................. 33 2.3.7 Changing How the Compiler Interprets Filenames ........................................................... 33 2.3.8 Changing How the Compiler Processes C Files .............................................................. 33 2.3.9 Changing How the Compiler Interprets and Names Extensions ............................................ 34 2.3.10 Specifying Directories............................................................................................ 34 2.3.11 Assembler Options............................................................................................... 34 2.3.12 Dynamic Linking.................................................................................................. 36 2.3.13 Deprecated Options.............................................................................................. 36 2.4 Controlling the Compiler Through Environment Variables ............................................................ 37 2.4.1 Setting Default Compiler Options (C2000_C_OPTION) ..................................................... 37 2.4.2 Naming One or More Alternate Directories (C2000_C_DIR) ................................................ 38 2.5 Controlling the Preprocessor ............................................................................................. 38 2.5.1 Predefined Macro Names ........................................................................................ 39 2.5.2 The Search Path for #include Files ............................................................................. 40 2.5.3 Support for the #warning and #warn Directives ............................................................... 41 2.5.4 Generating a Preprocessed Listing File (--preproc_only Option) ........................................... 41 2.5.5 Continuing Compilation After Preprocessing (--preproc_with_compile Option) ........................... 41 2.5.6 Generating a Preprocessed Listing File with Comments (--preproc_with_comment Option) ........... 42 2.5.7 Generating Preprocessed Listing with Line-Control Details (--preproc_with_line Option) ............... 42 2.5.8 Generating Preprocessed Output for a Make Utility (--preproc_dependency Option) ................... 42 2.5.9 Generating a List of Files Included with #include (--preproc_includes Option) ........................... 42 2.5.10 Generating a List of Macros in a File (--preproc_macros Option) ......................................... 42 2.6 Passing Arguments to main()............................................................................................. 42 2.7 Understanding Diagnostic Messages.................................................................................... 43 2.7.1 Controlling Diagnostic Messages ............................................................................... 44 2.7.2 How You Can Use Diagnostic Suppression Options ......................................................... 45 2.8 Other Messages ............................................................................................................ 46 2.9 Generating Cross-Reference Listing Information (--gen_cross_reference Option) ................................ 46 2.10 Generating a Raw Listing File (--gen_preprocessor_listing Option) ................................................. 47 2 Contents Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com 2.11 Using Inline Function Expansion ......................................................................................... 48 2.11.1 Inlining Intrinsic Operators ...................................................................................... 49 2.11.2 Unguarded Definition-Controlled Inlining...................................................................... 49 2.11.3 Guarded Inlining and the _INLINE Preprocessor Symbol .................................................. 50 2.11.4 Inlining Restrictions .............................................................................................. 51 2.12 Using Interlist ............................................................................................................... 52 2.13 Enabling Entry Hook and Exit Hook Functions......................................................................... 53 3 Optimizing Your Code......................................................................................................... 54 3.1 Invoking Optimization ...................................................................................................... 55 3.2 Performing File-Level Optimization (--opt_level=3 option) ............................................................ 56 3.2.1 Creating an Optimization Information File (--gen_opt_info Option)......................................... 56 3.3 Program-Level Optimization (--program_level_compile and --opt_level=3 options)............................... 56 3.3.1 Controlling Program-Level Optimization (--call_assumptions Option)...................................... 57 3.3.2 Optimization Considerations When Mixing C/C++ and Assembly .......................................... 58 3.4 Link-Time Optimization (--opt_level=4 Option) ......................................................................... 59 3.4.1 Option Handling ................................................................................................... 59 3.4.2 Incompatible Types ............................................................................................... 59 3.5 Using Feedback Directed Optimization.................................................................................. 59 3.5.1 Feedback Directed Optimization ................................................................................ 60 3.5.2 Profile Data Decoder.............................................................................................. 62 3.5.3 Feedback Directed Optimization API ........................................................................... 63 3.5.4 Feedback Directed Optimization Summary .................................................................... 63 3.6 Using Profile Information to Analyze Code Coverage ................................................................. 64 3.6.1 Code Coverage.................................................................................................... 64 3.6.2 Related Features and Capabilities .............................................................................. 65 3.7 Special Considerations When Using Optimization..................................................................... 66 3.7.1 Use Caution With asm Statements in Optimized Code ...................................................... 66 3.7.2 Use the Volatile Keyword for Necessary Memory Accesses ................................................ 66 3.8 Automatic Inline Expansion (--auto_inline Option) ..................................................................... 68 3.9 Using the Interlist Feature With Optimization........................................................................... 68 3.10 Data Page (DP) Pointer Load Optimization............................................................................. 71 3.11 Debugging and Profiling Optimized Code............................................................................... 71 3.11.1 Profiling Optimized Code........................................................................................ 71 3.12 Controlling Code Size Versus Speed ................................................................................... 72 3.13 Increasing Code-Size Optimizations (--opt_for_space Option)....................................................... 73 3.14 Compiler Support for Generating DMAC Instructions ................................................................. 74 3.14.1 Automatic Generation of DMAC Instructions ................................................................. 75 3.14.2 Assertions to Specify Data Address Alignment .............................................................. 75 3.14.3 __dmac Intrinsic .................................................................................................. 76 3.15 What Kind of Optimization Is Being Performed? ....................................................................... 77 3.15.1 Cost-Based Register Allocation ................................................................................ 77 3.15.2 Alias Disambiguation ............................................................................................ 77 3.15.3 Branch Optimizations and Control-Flow Simplification ...................................................... 78 3.15.4 Data Flow Optimizations ........................................................................................ 78 3.15.5 Expression Simplification........................................................................................ 78 3.15.6 Inline Expansion of Functions .................................................................................. 78 3.15.7 Function Symbol Aliasing ....................................................................................... 78 3.15.8 Induction Variables and Strength Reduction ................................................................. 79 3.15.9 Loop-Invariant Code Motion .................................................................................... 79 3.15.10 Loop Rotation ................................................................................................... 79 3.15.11 Instruction Scheduling.......................................................................................... 79 3.15.12 Register Variables .............................................................................................. 79 3.15.13 Register Tracking/Targeting................................................................................... 79 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Contents 3 www.ti.com 3.15.14 Tail Merging ..................................................................................................... 79 3.15.15 Autoincrement Addressing .................................................................................... 79 3.15.16 Removing Comparisons to Zero .............................................................................. 80 3.15.17 RPTB Generation (for FPU Targets Only)................................................................... 80 4 Linking C/C++ Code............................................................................................................ 81 4.1 Invoking the Linker Through the Compiler (-z Option) ................................................................ 82 4.1.1 Invoking the Linker Separately .................................................................................. 82 4.1.2 Invoking the Linker as Part of the Compile Step .............................................................. 83 4.1.3 Disabling the Linker (--compile_only Compiler Option) ...................................................... 83 4.2 Linker Code Optimizations ................................................................................................ 84 4.2.1 Generating Function Subsections (--gen_func_subsections Compiler Option) ........................... 84 4.2.2 Generating Aggregate Data Subsections (--gen_data_subsections Compiler Option) .................. 84 4.3 Controlling the Linking Process .......................................................................................... 84 4.3.1 Including the Run-Time-Support Library ....................................................................... 84 4.3.2 Run-Time Initialization ............................................................................................ 85 4.3.3 Initialization by the Interrupt Vector ............................................................................. 86 4.3.4 Global Object Constructors ...................................................................................... 86 4.3.5 Specifying the Type of Global Variable Initialization.......................................................... 86 4.3.6 Specifying Where to Allocate Sections in Memory ........................................................... 87 4.3.7 A Sample Linker Command File ................................................................................ 88 4.4 Linking C28x and C2XLP Code .......................................................................................... 89 5 Post-Link Optimizer ............................................................................................................ 90 5.1 The Post-Link Optimizer’s Role in the Software Development Flow ................................................ 91 5.2 Removing Redundant DP Loads ......................................................................................... 92 5.3 Tracking DP Values Across Branches................................................................................... 92 5.4 Tracking DP Values Across Function Calls............................................................................. 93 5.5 Other Post-Link Optimizations............................................................................................ 93 5.6 Controlling Post-Link Optimizations...................................................................................... 94 5.6.1 Excluding Files (-ex Option) ..................................................................................... 94 5.6.2 Controlling Post-Link Optimization Within an Assembly File ................................................ 94 5.6.3 Retaining Post-Link Optimizer Output (--keep_asm Option) ................................................ 94 5.6.4 Disable Optimization Across Function Calls (-nf Option ).................................................... 94 5.6.5 Annotating Assembly with Advice (--plink_advice_only option) ............................................. 94 5.7 Restrictions on Using the Post-Link Optimizer ......................................................................... 95 5.8 Naming the Outfile (--output_file Option)................................................................................ 95 6 TMS320C28x C/C++ Language Implementation ...................................................................... 96 6.1 Characteristics of TMS320C28x C....................................................................................... 97 6.1.1 Implementation-Defined Behavior............................................................................... 97 6.2 Characteristics of TMS320C28x C++ .................................................................................. 101 6.3 Using MISRA C 2004 .................................................................................................... 102 6.4 Data Types ................................................................................................................ 103 6.4.1 Size of Enum Types ............................................................................................. 103 6.4.2 Support for 64-Bit Integers ..................................................................................... 103 6.4.3 C28x long double Floating-Point Type Change.............................................................. 104 6.5 Keywords .................................................................................................................. 105 6.5.1 The const Keyword .............................................................................................. 105 6.5.2 The __cregister Keyword ....................................................................................... 106 6.5.3 The __interrupt Keyword........................................................................................ 107 6.5.4 The restrict Keyword ............................................................................................ 108 6.5.5 The volatile Keyword ............................................................................................ 108 6.6 C++ Exception Handling ................................................................................................. 110 6.7 Register Variables and Parameters .................................................................................... 110 6.8 The __asm Statement.................................................................................................... 111 4 Contents Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com 6.9 Pragma Directives ........................................................................................................ 112 6.9.1 The CALLS Pragma ............................................................................................. 113 6.9.2 The CHECK_MISRA Pragma .................................................................................. 113 6.9.3 The CLINK Pragma.............................................................................................. 113 6.9.4 The CODE_ALIGN Pragma .................................................................................... 114 6.9.5 The CODE_SECTION Pragma ................................................................................ 114 6.9.6 The DATA_ALIGN Pragma..................................................................................... 116 6.9.7 The DATA_SECTION Pragma ................................................................................. 117 6.9.8 The Diagnostic Message Pragmas ............................................................................ 118 6.9.9 The FAST_FUNC_CALL Pragma ............................................................................. 118 6.9.10 The FUNC_ALWAYS_INLINE Pragma ...................................................................... 120 6.9.11 The FUNC_CANNOT_INLINE Pragma ...................................................................... 120 6.9.12 The FUNC_EXT_CALLED Pragma .......................................................................... 121 6.9.13 The FUNCTION_OPTIONS Pragma ......................................................................... 121 6.9.14 The INTERRUPT Pragma ..................................................................................... 121 6.9.15 The MUST_ITERATE Pragma ................................................................................ 123 6.9.16 The NO_HOOKS Pragma ..................................................................................... 124 6.9.17 The RESET_MISRA Pragma ................................................................................. 124 6.9.18 The RETAIN Pragma........................................................................................... 124 6.9.19 The SET_CODE_SECTION and SET_DATA_SECTION Pragmas...................................... 125 6.9.20 The UNROLL Pragma.......................................................................................... 126 6.10 The _Pragma Operator .................................................................................................. 126 6.11 Object File Symbol Naming Conventions (Linknames) .............................................................. 127 6.12 Initializing Static and Global Variables ................................................................................ 127 6.12.1 Initializing Static and Global Variables With the Linker .................................................... 127 6.12.2 Initializing Static and Global Variables With the const Type Qualifier ................................... 128 6.13 Changing the ANSI/ISO C/C++ Language Mode..................................................................... 128 6.13.1 Enabling C99 Mode (--c99).................................................................................... 129 6.13.2 Enabling Strict ANSI/ISO Mode and Relaxed ANSI/ISO Mode (--strict_ansi and --relaxed_ansi Options) ........................................................................................................... 130 6.14 GNU Language Extensions ............................................................................................. 131 6.14.1 Extensions ....................................................................................................... 131 6.14.2 Function Attributes.............................................................................................. 132 6.14.3 Variable Attributes .............................................................................................. 133 6.14.4 Type Attributes .................................................................................................. 133 6.14.5 Built-In Functions ............................................................................................... 133 6.14.6 Using the Byte Peripheral Type Attribute.................................................................... 133 6.15 Compiler Limits............................................................................................................ 134 7 Run-Time Environment...................................................................................................... 136 7.1 Memory Model ............................................................................................................ 137 7.1.1 Sections ........................................................................................................... 137 7.1.2 C/C++ System Stack ............................................................................................ 138 7.1.3 Allocating .econst to Program Memory ....................................................................... 139 7.1.4 Dynamic Memory Allocation.................................................................................... 139 7.1.5 Initialization of Variables ....................................................................................... 140 7.1.6 Allocating Memory for Static and Global Variables.......................................................... 140 7.1.7 Field/Structure Alignment ....................................................................................... 140 7.1.8 Character String Constants..................................................................................... 140 7.2 Register Conventions .................................................................................................... 141 7.2.1 TMS320C28x Register Use and Preservation ............................................................... 141 7.2.2 Status Registers ................................................................................................. 142 7.3 Function Structure and Calling Conventions .......................................................................... 143 7.3.1 How a Function Makes a Call .................................................................................. 143 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Contents 5 www.ti.com 7.3.2 How a Called Function Responds ............................................................................. 144 7.3.3 Special Case for a Called Function (Large Frames) ........................................................ 145 7.3.4 Accessing Arguments and Local Variables................................................................... 145 7.3.5 Allocating the Frame and Accessing 32-Bit Values in Memory............................................ 146 7.4 Accessing Linker Symbols in C and C++.............................................................................. 146 7.5 Interfacing C and C++ With Assembly Language .................................................................... 146 7.5.1 Using Assembly Language Modules With C/C++ Code .................................................... 146 7.5.2 Accessing Assembly Language Functions From C/C++ ................................................... 147 7.5.3 Accessing Assembly Language Variables From C/C++ .................................................... 148 7.5.4 Sharing C/C++ Header Files With Assembly Source ....................................................... 149 7.5.5 Using Inline Assembly Language.............................................................................. 150 7.5.6 Using Intrinsics to Access Assembly Language Statements............................................... 151 7.6 Interrupt Handling......................................................................................................... 158 7.6.1 General Points About Interrupts ............................................................................... 158 7.6.2 Using C/C++ Interrupt Routines ............................................................................... 158 7.7 Integer Expression Analysis ............................................................................................. 159 7.7.1 Operations Evaluated With Run-Time-Support Calls ....................................................... 159 7.7.2 C/C++ Code Access to the Upper 16 Bits of 16-Bit Multiply ............................................... 159 7.8 Floating-Point Expression Analysis..................................................................................... 160 7.9 System Initialization ...................................................................................................... 160 7.9.1 Boot Hook Functions for System Pre-Initialization .......................................................... 161 7.9.2 Run-Time Stack .................................................................................................. 161 7.9.3 Automatic Initialization of Variables ........................................................................... 162 8 Using Run-Time-Support Functions and Building Libraries ................................................... 166 8.1 C and C++ Run-Time Support Libraries ............................................................................... 167 8.1.1 Linking Code With the Object Library ......................................................................... 167 8.1.2 Header Files ...................................................................................................... 167 8.1.3 Modifying a Library Function ................................................................................... 167 8.1.4 Support for String Handling..................................................................................... 168 8.1.5 Minimal Support for Internationalization ...................................................................... 168 8.1.6 Allowable Number of Open Files .............................................................................. 168 8.1.7 Library Naming Conventions ................................................................................... 168 8.2 The C I/O Functions ...................................................................................................... 169 8.2.1 High-Level I/O Functions ....................................................................................... 170 8.2.2 Overview of Low-Level I/O Implementation .................................................................. 171 8.2.3 Device-Driver Level I/O Functions............................................................................. 174 8.2.4 Adding a User-Defined Device Driver for C I/O.............................................................. 178 8.2.5 The device Prefix ................................................................................................ 179 8.3 Handling Reentrancy (_register_lock() and _register_unlock() Functions)........................................ 181 8.4 Library-Build Process..................................................................................................... 182 8.4.1 Required Non-Texas Instruments Software .................................................................. 182 8.4.2 Using the Library-Build Process ............................................................................... 182 8.4.3 Extending mklib .................................................................................................. 185 9 C++ Name Demangler........................................................................................................ 186 9.1 Invoking the C++ Name Demangler.................................................................................... 187 9.2 C++ Name Demangler Options ......................................................................................... 187 9.3 Sample Usage of the C++ Name Demangler ......................................................................... 188 10 CLA Compiler................................................................................................................... 190 10.1 How to Invoke the CLA Compiler....................................................................................... 191 10.2 CLA C Language Implementation ...................................................................................... 192 10.2.1 Characteristics .................................................................................................. 192 10.2.2 C Language Restrictions....................................................................................... 193 10.2.3 Memory Model - Sections ..................................................................................... 194 6 Contents Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com 10.2.4 Function Structure and Calling Conventions ................................................................ 194 A Glossary .......................................................................................................................... 195 A.1 Terminology ............................................................................................................... 195 B Revision History ............................................................................................................... 200 B.1 Recent Revisions ......................................................................................................... 200 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Contents 7 www.ti.com List of Figures 1-1. TMS320C28x Software Development Flow............................................................................. 14 5-1. The Post-Link Optimizer in the TMS320C28x Software Development Flow ....................................... 91 7-1. Use of the Stack During a Function Call............................................................................... 143 7-2. Format of Initialization Records in the .cinit Section ................................................................. 162 7-3. Format of Initialization Records in the .pinit Section ................................................................. 163 7-4. Autoinitialization at Run Time ........................................................................................... 164 7-5. Initialization at Load Time ............................................................................................... 165 10-1. CLA Compilation Overview .............................................................................................. 191 List of Tables 2-1. Processor Options ......................................................................................................... 19 2-2. Optimization Options ...................................................................................................... 19 2-3. Advanced Optimization Options ......................................................................................... 20 2-4. Debug Options.............................................................................................................. 20 2-5. Include Options ............................................................................................................ 20 2-6. Control Options ............................................................................................................ 21 2-7. Language Options.......................................................................................................... 21 2-8. Parser Preprocessing Options............................................................................................ 21 2-9. Predefined Symbols Options ............................................................................................. 22 2-10. Diagnostic Message Options ............................................................................................. 22 2-11. Run-Time Model Options.................................................................................................. 23 2-12. Entry/Exit Hook Options ................................................................................................... 23 2-13. Feedback Options ......................................................................................................... 23 2-14. Assembler Options ......................................................................................................... 24 2-15. File Type Specifier Options ............................................................................................... 24 2-16. Directory Specifier Options................................................................................................ 24 2-17. Default File Extensions Options .......................................................................................... 25 2-18. Command Files Options................................................................................................... 25 2-19. MISRA-C 2004 Options ................................................................................................... 25 2-20. Linker Basic Options ....................................................................................................... 26 2-21. File Search Path Options.................................................................................................. 26 2-22. Command File Preprocessing Options .................................................................................. 26 2-23. Diagnostic Message Options ............................................................................................. 26 2-24. Linker Output Options ..................................................................................................... 27 2-25. Symbol Management Options ............................................................................................ 27 2-26. Run-Time Environment Options .......................................................................................... 27 2-27. Link-Time Optimization Options .......................................................................................... 27 2-28. Miscellaneous Options..................................................................................................... 28 2-29. Linker Options For Dynamic Linking..................................................................................... 36 2-30. Predefined C28x Macro Names .......................................................................................... 39 2-31. Raw Listing File Identifiers ................................................................................................ 47 2-32. Raw Listing File Diagnostic Identifiers................................................................................... 47 3-1. Interaction Between Debugging and Optimization Options........................................................... 56 3-2. Options That You Can Use With --opt_level=3......................................................................... 56 3-3. Selecting a Level for the --gen_opt_info Option........................................................................ 56 3-4. Selecting a Level for the --call_assumptions Option................................................................... 57 3-5. Special Considerations When Using the --call_assumptions Option ................................................ 57 8 List of Figures Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com 3-6. Interaction Between Debugging and Optimization Options........................................................... 71 4-1. Initialized Sections Created by the Compiler .......................................................................... 87 4-2. Uninitialized Sections Created by the Compiler ........................................................................ 87 6-1. TMS320C28x C/C++ Data Types....................................................................................... 103 6-2. Valid Control Registers................................................................................................... 106 6-3. GCC Language Extensions ............................................................................................. 131 7-1. Summary of Sections and Memory Placement ....................................................................... 138 7-2. Register Use and Preservation Conventions.......................................................................... 141 7-3. FPU Register Use and Preservation Conventions ................................................................... 141 7-4. Status Register Fields.................................................................................................... 142 7-5. Floating-Point Status Register (STF) Fields For FPU Targets Only ............................................... 142 7-6. TMS320C28x C/C++ Compiler Intrinsics .............................................................................. 151 7-7. C/C++ Compiler Intrinsics for FPU ..................................................................................... 156 7-8. C/C++ Compiler Intrinsics for TMU..................................................................................... 156 8-1. The mklib Program Options ............................................................................................. 184 10-1. C/C++ Compiler Intrinsics for CLA ..................................................................................... 192 B-1. Revision History........................................................................................................... 200 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated List of Tables 9 Preface SPRU514N – June 2017 Read This First About This Manual The TMS320C28x Optimizing C/C++ Compiler User's Guide explains how to use the following Texas Instruments Code Generation compiler tools: • Compiler • Post-link optimizer • Library build utility • C++ name demangler The TI compiler accepts C and C++ code conforming to the International Organization for Standardization (ISO) standards for these languages. The compiler supports both the 1989 and 1999 versions of the C language and the 2003 version of the C++ language. This user's guide discusses the characteristics of the TI C/C++ compiler. It assumes that you already know how to write C/C++ programs. The C Programming Language (second edition), by Brian W. Kernighan and Dennis M. Ritchie, describes C based on the ISO C standard. You can use the Kernighan and Ritchie (hereafter referred to as K&R) book as a supplement to this manual. References to K&R C (as opposed to ISO C) in this manual refer to the C language as defined in the first edition of Kernighan and Ritchie's The C Programming Language. Notational Conventions This document uses the following conventions: • Program listings, program examples, and interactive displays are shown in a special typeface. Interactive displays use a bold version of the special typeface to distinguish commands that you enter from items that the system displays (such as prompts, command output, error messages, etc.). Here is a sample of C code: #include main() { printf("Hello World\n"); } • In syntax descriptions, instructions, commands, and directives arein a bold typeface and parameters are in an italic typeface. Portions of a syntax that are in bold should be entered as shown; portions of a syntax that are in italics describe the type of information that should be entered. • Square brackets ( [ and ] ) identify an optional parameter. If you use an optional parameter, you specify the information within the brackets. Unless the square brackets are in the bold typeface, do not enter the brackets themselves. The following is an example of a command that has an optional parameter: cl2000 [options] [filenames] [--run_linker [link_options] [object files]] • Braces ( { and } ) indicate that you must choose one of the parameters within the braces; you do not enter the braces themselves. This is an example of a command with braces that are not included in the actual syntax but indicate that you must specify either the --rom_model or --ram_model option: cl2000 --run_linker {--rom_model | --ram_model} filenames [--output_file= name.out] --library= libraryname 10 Read This First Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Related Documentation • In assembler syntax statements, the leftmost column is reserved for the first character of a label or symbol. If the label or symbol is optional, it is usually not shown. If a label or symbol is a required parameter, it is shown starting against the left margin of the box, as in the example below. No instruction, command, directive, or parameter, other than a symbol or label, can begin in the leftmost column. symbol .usect "section name", size in bytes[, alignment] • Some directives can have a varying number of parameters. For example, the .byte directive. This syntax is shown as [, ..., parameter]. • The TMS320C2800™ core is referred to as TMS320C28x or C28x. Related Documentation You can use the following books to supplement this user's guide: ANSI X3.159-1989, Programming Language - C (Alternate version of the 1989 C Standard), American National Standards Institute ISO/IEC 9899:1989, International Standard - Programming Languages - C (The 1989 C Standard), International Organization for Standardization ISO/IEC 9899:1999, International Standard - Programming Languages - C (The 1999 C Standard), International Organization for Standardization ISO/IEC 14882-2003, International Standard - Programming Languages - C++ (The 2003 C++ Standard), International Organization for Standardization The C Programming Language (second edition), by Brian W. Kernighan and Dennis M. Ritchie, published by Prentice-Hall, Englewood Cliffs, New Jersey, 1988 The Annotated C++ Reference Manual, Margaret A. Ellis and Bjarne Stroustrup, published by AddisonWesley Publishing Company, Reading, Massachusetts, 1990 C: A Reference Manual (fourth edition), by Samuel P. Harbison, and Guy L. Steele Jr., published by Prentice Hall, Englewood Cliffs, New Jersey Programming Embedded Systems in C and C++, by Michael Barr, Andy Oram (Editor), published by O'Reilly & Associates; ISBN: 1565923545, February 1999 Programming in C, Steve G. Kochan, Hayden Book Company The C++ Programming Language (second edition), Bjarne Stroustrup, published by Addison-Wesley Publishing Company, Reading, Massachusetts, 1990 Tool Interface Standards (TIS) DWARF Debugging Information Format Specification Version 2.0, TIS Committee, 1995 DWARF Debugging Information Format Version 3, DWARF Debugging Information Format Workgroup, Free Standards Group, 2005 (http://dwarfstd.org) DWARF Debugging Information Format Version 4, DWARF Debugging Information Format Workgroup, Free Standards Group, 2010 (http://dwarfstd.org) System V ABI specification (http://www.sco.com/developers/gabi/) SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Read This First 11 Related Documentation From Texas Instruments www.ti.com Related Documentation From Texas Instruments See the following resources for further information about the TI Code Generation Tools: • Texas Instruments Wiki: Compiler topics • Texas Instruments E2E Community: Compiler forum You can use the following documents to supplement this user's guide: SPRAAB5 — The Impact of DWARF on TI Object Files. Describes the Texas Instruments extensions to the DWARF specification. SPRU513— TMS320C28x Assembly Language Tools User's Guide describes the assembly language tools (assembler and other tools used to develop assembly language code), assembler directives, macros, common object file format, and symbolic debugging directives for the TMS320C28x device. SPRU430— TMS320C28x DSP CPU and Instruction Set Reference Guide describes the central processing unit (CPU) and the assembly language instructions of the TMS320C28x fixed-point digital signal processors (DSPs). It also describes emulation features available on these DSPs. SPRU566— TMS320x28xx, 28xxx Peripheral Reference Guide describes the peripheral reference guides of the 28x digital signal processors (DSPs). SPRUEX3— TI SYS/BIOS Real-time Operating System User's Guide. SYS/BIOS gives application developers the ability to develop embedded real-time software. SYS/BIOS is a scalable real-time kernel. It is designed to be used by applications that require real-time scheduling and synchronization or real-time instrumentation. SYS/BIOS provides preemptive multithreading, hardware abstraction, real-time analysis, and configuration tools. Trademarks TMS320C2800, TMS320C28x are trademarks of Texas Instruments. All other trademarks are the property of their respective owners. 12 Read This First Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback Chapter 1 SPRU514N – June 2017 Introduction to the Software Development Tools The TMS320C28x™ is supported by a set of software development tools, which includes an optimizing C/C++ compiler, an assembler, a linker, and assorted utilities. This chapter provides an overview of these tools and introduces the features of the optimizing C/C++ compiler. The assembler and linker are discussed in detail in the TMS320C28x Assembly Language Tools User's Guide. Topic ........................................................................................................................... Page 1.1 Software Development Tools Overview................................................................. 14 1.2 Compiler Interface.............................................................................................. 15 1.3 ANSI/ISO Standard ............................................................................................. 16 1.4 Output Files....................................................................................................... 16 1.5 Utilities ............................................................................................................. 16 SPRU514N – June 2017 Submit Documentation Feedback Introduction to the Software Development Tools 13 Copyright © 2017, Texas Instruments Incorporated Software Development Tools Overview www.ti.com 1.1 Software Development Tools Overview Figure 1-1 illustrates the software development flow. The shaded portion of the figure highlights the most common path of software development for C language programs. The other portions are peripheral functions that enhance the development process. Figure 1-1. TMS320C28x Software Development Flow Macro source files C/C++ source files C/C++ compiler Archiver Assembler source C/C++ name demangling utility Macro library Assembler Archiver Library of object files Object files Linker Library-build utility Run-timesupport library Debugging tools Hex-conversion utility Executable object file Programmer Absolute lister Cross-reference lister Object file utilities C28x The following list describes the tools that are shown in Figure 1-1: • The compiler accepts C/C++ source code and produces C28x assembly language source code. See Chapter 2. • The assembler translates assembly language source files into machine language relocatable object files. See the TMS320C28x Assembly Language Tools User's Guide. • The linker combines relocatable object files into a single absolute executable object file. As it creates the executable file, it performs relocation and resolves external references. The linker accepts relocatable object files and object libraries as input. See Chapter 4 for an overview of the linker. See the TMS320C28x Assembly Language Tools User's Guide for details. 14 Introduction to the Software Development Tools Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Compiler Interface • The archiver allows you to collect a group of files into a single archive file, called a library. The archiver allows you to modify such libraries by deleting, replacing, extracting, or adding members. One of the most useful applications of the archiver is building a library of object files. See the TMS320C28x Assembly Language Tools User's Guide. • The run-time-support libraries contain the standard ISO C and C++ library functions, compiler-utility functions, floating-point arithmetic functions, and C I/O functions that are supported by the compiler. See Chapter 8. The library-build utility automatically builds the run-time-support library if compiler and linker options require a custom version of the library. See Section 8.4. Source code for the standard run-time-support library functions for C and C++ is provided in the lib\src subdirectory of the directory where the compiler is installed. • The hex conversion utility converts an object file into other object formats. You can download the converted file to an EPROM programmer. See the TMS320C28x Assembly Language Tools User's Guide. • The absolute lister accepts linked object files as input and creates .abs files as output. You can assemble these .abs files to produce a listing that contains absolute, rather than relative, addresses. Without the absolute lister, producing such a listing would be tedious and would require many manual operations. See the TMS320C28x Assembly Language Tools User's Guide. • The cross-reference lister uses object files to produce a cross-reference listing showing symbols, their definitions, and their references in the linked source files. See the TMS320C28x Assembly Language Tools User's Guide. • The C++ name demangler is a debugging aid that converts names mangled by the compiler back to their original names as declared in the C++ source code. As shown in Figure 1-1, you can use the C++ name demangler on the assembly file that is output by the compiler; you can also use this utility on the assembler listing file and the linker map file. See Chapter 9. • The post-link optimizer removes or modifies assembly language instructions to generate better code. The post-link optimizer must be run with the compiler -plink option. See Chapter 5. • The disassembler decodes object files to show the assembly instructions that they represent. See the TMS320C28x Assembly Language Tools User's Guide. • The main product of this development process is an executable object file that can be executed in a TMS320C28x device. 1.2 Compiler Interface The compiler is a command-line program named cl2000. This program can compile, optimize, assemble, and link programs in a single step. Within Code Composer Studio, the compiler is run automatically to perform the steps needed to build a project. For more information about compiling a program, see Section 2.1 The compiler has straightforward calling conventions, so you can write assembly and C functions that call each other. For more information about calling conventions, see Chapter 7. SPRU514N – June 2017 Submit Documentation Feedback Introduction to the Software Development Tools 15 Copyright © 2017, Texas Instruments Incorporated ANSI/ISO Standard www.ti.com 1.3 ANSI/ISO Standard The compiler supports both the 1989 and 1999 versions of the C language and the 2003 version of the C++ language. The C and C++ language features in the compiler are implemented in conformance with the following ISO standards: • ISO-standard C The C compiler supports the 1989 and 1999 versions of the C language. – C89. Compiling with the --c89 option causes the compiler to conform to the ISO/IEC 9899:1990 C standard, which was previously ratified as ANSI X3.159-1989. The names "C89" and "C90" refer to the same programming language. "C89" is used in this document. – C99. Compiling with the --c99 option causes the compiler to conform to the ISO/IEC 9899:1999 C standard. This standard supports several features not part of C89, such as inline functions, new data types, and one-line comments beginning with //. The C language is also described in the second edition of Kernighan and Ritchie's The C Programming Language (K&R). • ISO-standard C++ Compiling with the --c++03 option causes the C++ compiler to conform to the C++ Standard ISO/IEC 14882:2003. The language is also described in Ellis and Stroustrup's The Annotated C++ Reference Manual (ARM), but this is not the standard. The compiler also supports embedded C++. For a description of unsupported C++ features, see Section 6.2. • ISO-standard run-time support The compiler tools come with an extensive run-time library. Library functions conform to the ISO C/C++ library standard unless otherwise stated. The library includes functions for standard input and output, string manipulation, dynamic memory allocation, data conversion, timekeeping, trigonometry, and exponential and hyperbolic functions. Functions for signal handling are not included, because these are target-system specific. For more information, see Chapter 8. See Section 6.13 for command line options to select the C or C++ standard your code uses. 1.4 Output Files The following type of output file is created by the compiler: • COFF object files Common object file format (COFF) provides basic modular (separately-compiled) compilation features, such as relocations. 1.5 Utilities These features are compiler utilities: • Library-build utility The library-build utility lets you custom-build object libraries from source for any combination of runtime models. For more information, see Section 8.4. • C++ name demangler The C++ name demangler (dem2000) is a debugging aid that translates each mangled name it detects in compiler-generated assembly code, disassembly output, or compiler diagnostic messages to its original name found in the C++ source code. For more information, see Chapter 9. • Hex conversion utility For stand-alone embedded applications, the compiler has the ability to place all code and initialization data into ROM, allowing C/C++ code to run from reset. The COFF files output by the compiler can be converted to EPROM programmer data files by using the hex conversion utility, as described in the TMS320C28x Assembly Language Tools User's Guide. 16 Introduction to the Software Development Tools Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback Chapter 2 SPRU514N – June 2017 Using the C/C++ Compiler The compiler translates your source program into machine language object code that the TMS320C28x can execute. Source code must be compiled, assembled, and linked to create an executable object file. All of these steps are executed at once by using the compiler. Topic ........................................................................................................................... Page 2.1 About the Compiler ............................................................................................ 18 2.2 Invoking the C/C++ Compiler ............................................................................... 18 2.3 Changing the Compiler's Behavior with Options .................................................... 19 2.4 Controlling the Compiler Through Environment Variables ...................................... 37 2.5 Controlling the Preprocessor............................................................................... 38 2.6 Passing Arguments to main() .............................................................................. 42 2.7 Understanding Diagnostic Messages.................................................................... 43 2.8 Other Messages ................................................................................................. 46 2.9 Generating Cross-Reference Listing Information (--gen_cross_reference Option)...... 46 2.10 Generating a Raw Listing File (--gen_preprocessor_listing Option) .......................... 47 2.11 Using Inline Function Expansion.......................................................................... 48 2.12 Using Interlist .................................................................................................... 52 2.13 Enabling Entry Hook and Exit Hook Functions ...................................................... 53 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 17 About the Compiler www.ti.com 2.1 About the Compiler The compiler lets you compile, optimize, assemble, and optionally link in one step. The compiler performs the following steps on one or more source modules: • The compiler accepts C/C++ source code and assembly code. It produces object code. You can compile C, C++, and assembly files in a single command. The compiler uses the filename extensions to distinguish between different file types. See Section 2.3.9 for more information. • The linker combines object files to create a static executable or dynamic object file. The link step is optional, so you can compile and assemble many modules independently and link them later. See Chapter 4 for information about linking the files. Invoking the Linker NOTE: By default, the compiler does not invoke the linker. You can invoke the linker by using the -- run_linker (-z)compiler option. See Section 4.1.1 for details. For a complete description of the assembler and the linker, see the TMS320C28x Assembly Language Tools User's Guide. 2.2 Invoking the C/C++ Compiler To invoke the compiler, enter: cl2000 [options] [filenames] [--run_linker [link_options] object files]] cl2000 options filenames --run_linker (-z) link_options object files Command that runs the compiler and the assembler. Options that affect the way the compiler processes input files. The options are listed in Table 2-6 through Table 2-28. One or more C/C++ source files and assembly language source files. Option that invokes the linker. The --run_linker option's short form is -z. See Chapter 4 for more information. Options that control the linking process. Names of the object files for the linking process. The arguments to the compiler are of three types: • Compiler options • Link options • Filenames The --run_linker option indicates linking is to be performed. If the --run_linker option is used, any compiler options must precede the --run_linker option, and all link options must follow the --run_linker option. Source code filenames must be placed before the --run_linker option. Additional object file filenames can be placed after the --run_linker option. For example, if you want to compile two files named symtab.c and file.c, assemble a third file named seek.asm, and link to create an executable program called myprogram.out, you will enter: cl2000 symtab.c file.c seek.asm --run_linker --library=lnk.cmd --output_file=myprogram.out 18 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Changing the Compiler's Behavior with Options 2.3 Changing the Compiler's Behavior with Options Options control the operation of the compiler. This section provides a description of option conventions and an option summary table. It also provides detailed descriptions of the most frequently used options, including options used for type-checking and assembling. For a help screen summary of the options, enter cl2000 with no parameters on the command line. The following apply to the compiler options: • There are typically two ways of specifying a given option. The "long form" uses a two hyphen prefix and is usually a more descriptive name. The "short form" uses a single hyphen prefix and a combination of letters and numbers that are not always intuitive. • Options are usually case sensitive. • Individual options cannot be combined. • An option with a parameter should be specified with an equal sign before the parameter to clearly associate the parameter with the option. For example, the option to undefine a constant can be expressed as --undefine=name. Likewise, the option to specify the maximum amount of optimization can be expressed as -O=3. You can also specify a parameter directly after certain options, for example -O3 is the same as -O=3. No space is allowed between the option and the optional parameter, so -O 3 is not accepted. • Files and options except the --run_linker option can occur in any order. The --run_linker option must follow all compiler options and precede any linker options. You can define default options for the compiler by using the C2000_C_OPTION environment variable. For a detailed description of the environment variable, see Section 2.4.1. Table 2-6 through Table 2-28 summarize all options (including link options). Use the references in the tables for more complete descriptions of the options. Option --silicon_version=28 --unified_memory --cla_support[=cla0|cla1|cla2] --float_support={fpu32|softlib} --tmu_support[=tmu0] --vcu_support[=vcu0|vcu2] Alias -v28 -mt Table 2-1. Processor Options Effect Section Specifies TMS320C28x architecture. The default (and only value accepted) is 28. This option is no longer required. Section 2.3.4 Generates code for the unified memory model. Section 2.3.4 Specifies TMS320C28x CLA accelerator support for Type 0, Type 1, Section 2.3.4 or Type 2. Default is cla0. Specifies use of TMS320C28x 32-bit hardware floating-point support. Section 2.3.4 The default is softlib. Enables support for the Trigonometric Math Unit (TMU). Using this option also enables FPU32 support (as with --float_support=fpu32). If this option is used but no value is specified, the default is tmu0. Section 2.3.4 Specifies C28x VCU coprocessor support Type 0 or Type 2. Default Section 2.3.4 is vcu0. Option --opt_level=off --opt_level=n --opt_for_space=n --fp_mode={relaxed|strict} --fp_reassoc={on|off} Alias -On -ms Table 2-2. Optimization Options(1) Effect Disables all optimization (default). Level 0 (-O0) optimizes register usage only. Level 1 (-O1) uses Level 0 optimizations and optimizes locally. Level 2 (-O2) uses Level 1 optimizations and optimizes globally. Level 3 (-O3) uses Level 2 optimizations and optimizes the file. () Level 4 (-O4) uses Level 3 optimizations and performs link-time optimization. () Controls code size on four levels (0, 1, 2, and 3). Enables or disables relaxed floating-point mode. Enables or disables the reassociation of floating-point arithmetic. Section Section 3.1 Section 3.1, Section 3.2, Section 3.4 Section 3.13 Section 2.3.3 Section 2.3.3 (1) Note: Machine-specific options (see Table 2-11) can also affect optimization. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 19 Changing the Compiler's Behavior with Options Option --sat_reassoc={on|off} Table 2-2. Optimization Options(1) (continued) Alias Effect Enables or disables the reassociation of saturating arithmetic. Default is --sat_reassoc=off. www.ti.com Section Section 2.3.3 Table 2-3. Advanced Optimization Options(1) Option --auto_inline=[size] --call_assumptions=n Alias -oi -opn --gen_opt_info=n --opt_for_speed[=n] --optimizer_interlist --program_level_compile --aliased_variables -onn -mf -os -pm -ma Effect Section Sets automatic inlining size (--opt_level=3 only). If size is not specified, the default is 1. Section 3.8 Level 0 (-op0) specifies that the module contains functions and Section 3.3.1 variables that are called or modified from outside the source code provided to the compiler. Level 1 (-op1) specifies that the module contains variables modified from outside the source code provided to the compiler but does not use functions called from outside the source code. Level 2 (-op2) specifies that the module contains no functions or variables that are called or modified from outside the source code provided to the compiler (default). Level 3 (-op3) specifies that the module contains functions that are called from outside the source code provided to the compiler but does not use variables modified from outside the source code. Level 0 (-on0) disables the optimization information file. Level 1 (-on2) produces an optimization information file. Level 2 (-on2) produces a verbose optimization information file. Section 3.2.1 Controls the tradeoff between size and speed (0-5 range). If this option is specified without n, the default value is 4. If this option is not specified, the default setting is 2. Section 3.12 Interlists optimizer comments with assembly statements. Section 3.9 Combines source files to perform program-level optimization. Section 3.3 Notifies the compiler that addresses passed to functions may be modified by an alias in the called function. Section 3.7.2.2 (1) Note: Machine-specific options (see Table 2-11) can also affect optimization. Option --symdebug:dwarf --symdebug:none --symdebug:profile_coff --symdebug:skeletal Alias -g Table 2-4. Debug Options Effect Default behavior. Enables symbolic debugging. The generation of debug information no longer impacts optimization. Therefore, generating debug information is enabled by default. If you explicitly use the -g option but do not specify an optimization level, no optimization is performed. Disables all symbolic debugging. Enables profiling using the alternate STABS debugging format. (Deprecated; has no effect.) Section Section 2.3.5 Section 3.11 Section 2.3.5 Section 3.11 Section 2.3.5 Option --include_path=directory --preinclude=filename Alias -I Table 2-5. Include Options Effect Adds the specified directory to the #include search path. Includes filename at the beginning of compilation. Section Section 2.5.2.1 Section 2.3.3 20 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Option --compile_only --help --run_linker --skip_assembler Alias -c -h -z -n Changing the Compiler's Behavior with Options Table 2-6. Control Options Effect Section Disables linking (negates --run_linker). Section 4.1.3 Prints (on the standard output device) a description of the options understood by the compiler. Section 2.3.2 Causes the linker to be invoked from the compiler command line. Section 2.3.2 Compiles C/C++ source file, producing an assembly language output Section 2.3.2 file. The assembler is not run and no object file is produced. Option --c89 --c99 --c++03 --cla_default --cpp_default --exceptions --float_operations_allowed={none| all|32|64} --gen_cross_reference ---gen_preprocessor_listing --relaxed_ansi Alias -fg -px -pl -pr --rtti -rtti --strict_ansi -ps Table 2-7. Language Options Effect Processes C files according to the ISO C89 standard. Processes C files according to the ISO C99 standard. Processes C++ files according to the ISO C++03 standard. Processes both .c and .cla files as CLA files. Processes all source files with a C extension as C++ source files. Enables C++ exception handling. Restricts the types of floating point operations allowed. Section Section 6.13 Section 6.13 Section 6.13 Section 10.1 Section 2.3.7 Section 6.6 Section 2.3.3 Generates a cross-reference listing file (.crl). Generates a raw listing file (.rl). Enables relaxed mode; ignores strict ISO violations. This is on by default. To disable this mode, use the --strict_ansi option. Enables C++ run-time type information (RTTI). Enables strict ANSI/ISO mode (for C/C++, not for K&R C). In this mode, language extensions that conflict with ANSI/ISO C/C++ are disabled. In strict ANSI/ISO mode, most ANSI/ISO violations are reported as errors. Violations that are considered discretionary may be reported as warnings instead. Section 2.9 Section 2.10 Section 6.13.2 –Section 6.13.2 Option --preproc_dependency[=filename] --preproc_includes[=filename] --preproc_macros[=filename] --preproc_only --preproc_with_comment --preproc_with_compile --preproc_with_line Table 2-8. Parser Preprocessing Options Alias -ppd -ppi -ppm -ppo -ppc -ppa -ppl Effect Performs preprocessing only, but instead of writing preprocessed output, writes a list of dependency lines suitable for input to a standard make utility. Performs preprocessing only, but instead of writing preprocessed output, writes a list of files included with the #include directive. Performs preprocessing only. Writes list of predefined and userdefined macros to a file with the same name as the input but with a .pp extension. Performs preprocessing only. Writes preprocessed output to a file with the same name as the input but with a .pp extension. Performs preprocessing only. Writes preprocessed output, keeping the comments, to a file with the same name as the input but with a .pp extension. Continues compilation after preprocessing with any of the -pp options that normally disable compilation. Performs preprocessing only. Writes preprocessed output with linecontrol information (#line directives) to a file with the same name as the input but with a .pp extension. Section Section 2.5.8 Section 2.5.9 Section 2.5.10 Section 2.5.4 Section 2.5.6 Section 2.5.5 Section 2.5.7 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 21 Changing the Compiler's Behavior with Options Table 2-9. Predefined Symbols Options Option --define=name[=def] --undefine=name Alias -D -U Effect Predefines name. Undefines name. www.ti.com Section Section 2.3.2 Section 2.3.2 Option --advice:performance[=all,none] --compiler_revision --diag_error=num --diag_remark=num --diag_suppress=num --diag_warning=num --diag_wrap={on|off} --display_error_number --emit_warnings_as_errors --issue_remarks --no_warnings --quiet --set_error_limit=num --super_quiet --tool_version --verbose --verbose_diagnostics --write_diagnostics_file Table 2-10. Diagnostic Message Options Alias -pdse -pdsr -pds -pdsw -pden -pdew -pdr -pdw -q -pdel -qq -version -pdv -pdf Effect Section Provides advice on ways to improve performance. Default is all. Section 2.3.3 Prints out the compiler release revision and exits. -- Categorizes the diagnostic identified by num as an error. Section 2.7.1 Categorizes the diagnostic identified by num as a remark. Section 2.7.1 Suppresses the diagnostic identified by num. Section 2.7.1 Categorizes the diagnostic identified by num as a warning. Section 2.7.1 Wrap diagnostic messages (default is on). Note that this commandline option cannot be used within the Code Composer Studio IDE. Displays a diagnostic's identifiers along with its text. Note that this command-line option cannot be used within the Code Composer Studio IDE. Section 2.7.1 Treat warnings as errors. Section 2.7.1 Issues remarks (non-serious warnings). Section 2.7.1 Suppresses diagnostic warnings (errors are still issued). Section 2.7.1 Suppresses progress messages (quiet). -- Sets the error limit to num. The compiler abandons compiling after this number of errors. (The default is 100.) Section 2.7.1 Super quiet mode. -- Displays version number for each tool. -- Display banner and function progress information. -- Provides verbose diagnostic messages that display the original Section 2.7.1 source with line-wrap. Note that this command-line option cannot be used within the Code Composer Studio IDE. Generates a diagnostic message information file. Compiler only Section 2.7.1 option. Note that this command-line option cannot be used within the Code Composer Studio IDE. 22 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Option --disable_dp_load_opt --gen_data_subsections={on|off} --gen_func_subsections={on|off} --no_rpt --printf_support={nofloat|full| minimal} --protect_volatile --ramfunc={on|off} --rpt_threshold=k --symdebug:dwarf_version=2|3|4 Changing the Compiler's Behavior with Options Table 2-11. Run-Time Model Options Alias -md -mo -mi -mv Effect Disables DP load optimizations. Place all aggregate data (arrays, structs, and unions) into subsections. This gives the linker more control over removing unused data during the final link step. The default is on. Puts each function in a separate subsection in the object file. Disables generation of RPT instructions. Enables support for smaller, limited versions of the printf function family (sprintf, fprintf, etc.) and the scanf function family (sscanf, fscanf, etc.) run-time-support functions. Enables volatile reference protection. If set to on, specifies that all functions should be placed in the .TI.ramfunc section, which is placed in RAM. Generates RPT loops that iterate k times or less. (k is a constant between 0 and 256.) Specifies the DWARF format version. Section Section 2.3.4 Section 4.2.2 Section 4.2.1 Section 2.3.4 Section 2.3.3 Section 2.3.4 Section 2.3.4 Section 2.3.4 Section 2.3.5 Option --entry_hook[=name] --entry_parm={none|name| address} --exit_hook[=name] --exit_parm={none|name|address} --remove_hooks_when_inlining Table 2-12. Entry/Exit Hook Options Alias Effect Enables entry hooks. Specifies the parameters to the function to the --entry_hook option. Enables exit hooks. Specifies the parameters to the function to the --exit_hook option. Removes entry/exit hooks for auto-inlined functions. Section Section 2.13 Section 2.13 Section 2.13 Section 2.13 Section 2.13 Option --analyze={codecov|callgraph} --analyze_only --gen_profile_info --use_profile_info=file1[, file2,...] Alias Table 2-13. Feedback Options Effect Generate analysis info from profile data. Only generate analysis. Generates instrumentation code to collect profile information. Specifies the profile information file(s). Section Section 3.6.2.2 Section 3.6.2.2 Section 3.5.1.3 Section 3.5.1.3 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 23 Changing the Compiler's Behavior with Options Table 2-14. Assembler Options www.ti.com Option --keep_asm --asm_listing --c_src_interlist --src_interlist --absolute_listing --asm_define=name[=def] --asm_dependency --asm_includes --issue_remarks --asm_undefine=name --asm_listing_cross_reference --flash_prefetch_warn --include_file=filename --preproc_asm Alias -k -al -ss -s -aa -ad -apd -api -au -ax -ahi -mx Effect Keeps the assembly language (.asm) file. Generates an assembly listing file. Interlists C source and assembly statements. Interlists optimizer comments (if available) and assembly source statements; otherwise interlists C and assembly source statements. Enables absolute listing. Sets the name symbol. Performs preprocessing; lists only assembly dependencies. Performs preprocessing; lists only included #include files. Issues remarks (non-serious warnings), which include additional assembly-time checking. Undefines the predefined constant name. Generates the cross-reference file. Assembler warnings for F281X BF flash prefetch issue. Includes the specified file for the assembly module. Preprocesses assembly source, expands assembly macros. Section Section 2.3.11 Section 2.3.11 Section 2.12 Section 3.9 Section 2.3.2 Section 2.3.11 Section 2.3.11 Section 2.3.11 Section 2.3.11 Section 2.3.11 Section 2.3.11 Section 2.3.11 Section 2.3.11 Section 2.3.11 Section 2.3.11 Option --asm_file=filename --c_file=filename --cpp_file=filename --obj_file=filename Table 2-15. File Type Specifier Options Alias -fa -fc -fp -fo Effect Section Identifies filename as an assembly source file regardless of its Section 2.3.7 extension. By default, the compiler and assembler treat .asm files as assembly source files. Identifies filename as a C source file regardless of its extension. By Section 2.3.7 default, the compiler treats .c files as C source files. Identifies filename as a C++ file, regardless of its extension. By Section 2.3.7 default, the compiler treats .C, .cpp, .cc and .cxx files as a C++ files. Identifies filename as an object code file regardless of its extension. Section 2.3.7 By default, the compiler and linker treat .obj files as object code files. Option --abs_directory=directory --asm_directory=directory --list_directory=directory --obj_directory=directory --output_file=filename --pp_directory=dir --temp_directory=directory Table 2-16. Directory Specifier Options Alias -fb -fs -ff -fr -fe -ft Effect Section Specifies an absolute listing file directory. By default, the compiler uses the .obj directory. Section 2.3.10 Specifies an assembly file directory. By default, the compiler uses the current directory. Section 2.3.10 Specifies an assembly listing file and cross-reference listing file directory By default, the compiler uses the .obj directory. Section 2.3.10 Specifies an object file directory. By default, the compiler uses the current directory. Section 2.3.10 Specifies a compilation output file name; can override -obj_directory. Section 2.3.10 Specifies a preprocessor file directory. By default, the compiler uses Section 2.3.10 the current directory. Specifies a temporary file directory. By default, the compiler uses the Section 2.3.10 current directory. 24 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Option --asm_extension=[.]extension --c_extension=[.]extension --cpp_extension=[.]extension --listing_extension=[.]extension --obj_extension=[.]extension Changing the Compiler's Behavior with Options Table 2-17. Default File Extensions Options Alias -ea -ec -ep -es -eo Effect Sets a default extension for assembly source files. Sets a default extension for C source files. Sets a default extension for C++ source files. Sets a default extension for listing files. Sets a default extension for object files. Section Section 2.3.9 Section 2.3.9 Section 2.3.9 Section 2.3.9 Section 2.3.9 Option --cmd_file=filename Table 2-18. Command Files Options Alias -@ Effect Interprets contents of a file as an extension to the command line. Multiple -@ instances can be used. Section Section 2.3.2 Option --check_misra[={all|required| advisory|none|rulespec}] --misra_advisory={error|warning| remark|suppress} --misra_required={error|warning| remark|suppress} Table 2-19. MISRA-C 2004 Options Alias Effect Enables checking of the specified MISRA-C:2004 rules. Default is all. Sets the diagnostic severity for advisory MISRA-C:2004 rules. Sets the diagnostic severity for required MISRA-C:2004 rules. Section Section 2.3.3 Section 2.3.3 Section 2.3.3 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 25 Changing the Compiler's Behavior with Options www.ti.com 2.3.1 Linker Options The following tables list the linker options. See Chapter 4 of this document and the TMS320C28x Assembly Language Tools User's Guide for details on these options. Option --run_linker --output_file=file --map_file=file --stack_size=size --heap_size=size --warn_sections Table 2-20. Linker Basic Options Alias -z -o -m [-]-stack [-]-heap -w Description Enables linking. Names the executable output file. The default filename is a .out file. Produces a map or listing of the input and output sections, including holes, and places the listing in file. Sets C system stack size to size words and defines a global symbol that specifies the stack size. Default = 1K words. Sets heap size (for the dynamic memory allocation in C) to size words and defines a global symbol that specifies the heap size. Default = 1K words. Displays a message when an undefined output section is created. Option --library=file --search_path=pathname --priority --reread_libs --disable_auto_rts Table 2-21. File Search Path Options Alias -l -I -priority -x Description Names an archive library or link command file as linker input. Alters library-search algorithms to look in a directory named with pathname before looking in the default location. This option must appear before the --library option. Satisfies unresolved references by the first library that contains a definition for that symbol. Forces rereading of libraries, which resolves back references. Disables the automatic selection of a run-time-support library. See Section 4.3.1.1. Option --define=name=value --undefine=name --disable_pp Table 2-22. Command File Preprocessing Options Alias Description Predefines name as a preprocessor macro. Removes the preprocessor macro name. Disables preprocessing for command files. Option --diag_error=num --diag_remark=num --diag_suppress=num --diag_warning=num --display_error_number --emit_warnings_as_errors --issue_remarks --no_demangle --no_warnings --set_error_limit=count --verbose_diagnostics Table 2-23. Diagnostic Message Options Alias -pdew Description Categorizes the diagnostic identified by num as an error. Categorizes the diagnostic identified by num as a remark. Suppresses the diagnostic identified by num. Categorizes the diagnostic identified by num as a warning. Displays a diagnostic's identifiers along with its text. Treat warnings as errors. Issues remarks (non-serious warnings). Disables demangling of symbol names in diagnostic messages. Suppresses diagnostic warnings (errors are still issued). Sets the error limit to count. The linker abandons linking after this number of errors. (The default is 100.) Provides verbose diagnostic messages that display the original source with line-wrap. 26 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Option --absolute_exe --ecc:data_error --ecc:ecc_error --mapfile_contents=attribute --relocatable --rom --run_abs --xml_link_info=file Changing the Compiler's Behavior with Options Table 2-24. Linker Output Options Alias -a -r -abs Description Produces an absolute, executable object file. This is the default; if neither --absolute_exe nor --relocatable is specified, the linker acts as if --absolute_exe were specified. Inject specified errors into the output file for testing. Inject specified errors into the Error Correcting Code (ECC) for testing. Controls the information that appears in the map file. Produces a nonexecutable, relocatable output object file. Creates a ROM object. Produces an absolute listing file. Generates a well-formed XML file containing detailed information about the result of a link. Option --entry_point=symbol --globalize=pattern --hide=pattern --localize=pattern --make_global=symbol --make_static --no_sym_merge --no_symtable --scan_libraries --symbol_map=refname=defname --undef_sym=symbol --unhide=pattern Table 2-25. Symbol Management Options Alias -e -g -h -b -s -scanlibs -u Description Defines a global symbol that specifies the primary entry point for the executable object file. Changes the symbol linkage to global for symbols that match pattern. Hides symbols that match the specified pattern. Make the symbols that match the specified pattern local. Makes symbol global (overrides -h). Makes all global symbols static. Disables merge of symbolic debugging information in COFF object files. Strips symbol table information and line number entries from the executable object file. Scans all libraries for duplicate symbol definitions. Specifies a symbol mapping; references to the refname symbol are replaced with references to the defname symbol. The --symbol_map option is now supported when used with --opt_level=4. Adds symbol to the symbol table as an unresolved symbol. Excludes symbols that match the specified pattern from being hidden. Option --arg_size=size --fill_value=value --ram_model --rom_model Table 2-26. Run-Time Environment Options Alias --args -f -cr -c Description Reserve size bytes for the argc/argv memory area. Sets default fill value for holes within output sections Initializes variables at load time. Autoinitializes variables at run time. Option --keep_asm --no_postlink_across_calls --plink_advice_only --postlink_exclude --postlink_opt (1) See Chapter 5 for details. SPRU514N – June 2017 Submit Documentation Feedback Table 2-27. Link-Time Optimization Options(1) Alias -nf -ex -plink Description Retain any post-link files (.pl) and .absolute listing files (.abs) generated by the −plink option. This allows you to view any changes the post-link optimizer makes. (Requires use of -plink) Disable post-link optimizations across functions. (Requires use of -plink) Annotates assembly code with comments if changes cannot be made safely due to pipeline considerations, such as when float support or VCU support is enabled. (Requires use of -plink) Exclude files from post-link pass. (Requires use of -plink) Post-link optimizations (only after -z). Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 27 Changing the Compiler's Behavior with Options Table 2-28. Miscellaneous Options www.ti.com Option --disable_clink --linker_help --preferred_order=function --strict_compatibility[=off|on] Alias -j [-]-help Description Disables conditional linking of COFF object files. Displays information about syntax and available options. Prioritizes placement of functions. Performs more conservative and rigorous compatibility checking of input object files. Default is on. 2.3.2 Frequently Used Options Following are detailed descriptions of options that you will probably use frequently: --c_src_interlist --cmd_file=filename --compile_only --define=name[=def] --help --include_path=directory Invokes the interlist feature, which interweaves original C/C++ source with compiler-generated assembly language. The interlisted C statements may appear to be out of sequence. You can use the interlist feature with the optimizer by combining the --optimizer_interlist and -c_src_interlist options. See Section 3.9. The --c_src_interlist option can have a negative performance and/or code size impact. Appends the contents of a file to the option set. You can use this option to avoid limitations on command line length or C style comments imposed by the host operating system. Use a # or ; at the beginning of a line in the command file to include comments. You can also include comments by delimiting them with /* and */. To specify options, surround hyphens with quotation marks. For example, "--"quiet. You can use the --cmd_file option multiple times to specify multiple files. For instance, the following indicates that file3 should be compiled as source and file1 and file2 are --cmd_file files: cl2000 --cmd_file=file1 --cmd_file=file2 file3 Suppresses the linker and overrides the --run_linker option, which specifies linking. The --compile_only option's short form is -c. Use this option when you have --run_linker specified in the C2000_C_OPTION environment variable and you do not want to link. See Section 4.1.3. Predefines the constant name for the preprocessor. This is equivalent to inserting #define name def at the top of each C source file. If the optional[=def] is omitted, the name is set to 1. The --define option's short form is -D. If you want to define a quoted string and keep the quotation marks, do one of the following: • For Windows, use --define=name="\"string def\"". For example, -define=car="\"sedan\"" • For UNIX, use --define=name='"string def"'. For example, -define=car='"sedan"' • For Code Composer Studio, enter the definition in a file and include that file with the --cmd_file option. Displays the syntax for invoking the compiler and lists available options. If the --help option is followed by another option or phrase, detailed information about the option or phrase is displayed. For example, to see information about debugging options use --help debug. Adds directory to the list of directories that the compiler searches for #include files. The --include_path option's short form is -I. You can use this option several times to define several directories; be sure to separate the --include_path options with spaces. If you do not specify a directory name, the preprocessor ignores the --include_path option. See Section 2.5.2.1. 28 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com --keep_asm --quiet --run_linker --skip_assembler --src_interlist --tool_version --undefine=name --verbose Changing the Compiler's Behavior with Options Retains the assembly language output from the compiler or assembly optimizer. Normally, the compiler deletes the output assembly language file after assembly is complete. The --keep_asm option's short form is -k. Suppresses banners and progress information from all the tools. Only source filenames and error messages are output. The --quiet option's short form is -q. Runs the linker on the specified object files. The --run_linker option and its parameters follow all other options on the command line. All arguments that follow --run_linker are passed to the linker. The -run_linker option's short form is -z. See Section 4.1. Compiles only. The specified source files are compiled but not assembled or linked. The --skip_assembler option's short form is -n. This option overrides --run_linker. The output is assembly language output from the compiler. Invokes the interlist feature, which interweaves optimizer comments or C/C++ source with assembly source. If the optimizer is invoked (-opt_level=n option), optimizer comments are interlisted with the assembly language output of the compiler, which may rearrange code significantly. If the optimizer is not invoked, C/C++ source statements are interlisted with the assembly language output of the compiler, which allows you to inspect the code generated for each C/C++ statement. The --src_interlist option implies the --keep_asm option. The --src_interlist option's short form is -s. Prints the version number for each tool in the compiler. No compiling occurs. Undefines the predefined constant name. This option overrides any -define options for the specified constant. The --undefine option's short form is -U. Displays progress information and toolset version while compiling. Resets the --quiet option. 2.3.3 Miscellaneous Useful Options Following are detailed descriptions of miscellaneous options: --advice:performance[=all|none] When Trigonometric Math Unit (TMU) support is enabled (-tmu_support) and --fp_mode=strict, advice will be generated if the compiler encounters function calls that could be replaced with TMU hardware instructions under --fp_mode=relaxed. These include floating point division, sqrt, sin, cos, atan, and atan2. --check_misra={all|required| advisory|none|rulespec} Displays the specified amount or type of MISRA-C documentation. This option must be used if you want to enable use of the CHECK_MISRA and RESET_MISRA pragmas within the source code. The rulespec parameter is a comma-separated list of specifiers. See Section 6.3 for details. --float_operations_allowed= {none|all|32|64} Restricts the type of floating point operations allowed in the application. The default is all. If set to none, 32, or 64, the application is checked for operations that will be performed at runtime. For example, if --float_operations_allowed=32 is specified on the command line, the compiler issues an error if a double precision operation will be generated. This can be used to ensure that double precision operations are not accidentally introduced into an application. The checks are performed after relaxed mode optimizations have been performed, so illegal operations that are completely removed result in no diagnostic messages. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 29 Changing the Compiler's Behavior with Options www.ti.com --fp_mode={relaxed|strict} --fp_reassoc={on|off} --misra_advisory={error| warning|remark|suppress} --misra_required={error| warning|remark|suppress} --preinclude=filename --printf_support={full| nofloat|minimal} --sat_reassoc={on|off} The default floating-point mode is strict. To enable relaxed floatingpoint mode use the --fp_mode=relaxed option. Relaxed floating-point mode causes double-precision floating-point computations and storage to be converted to integers where possible. This behavior does not conform with ISO, but it results in faster code, with some loss in accuracy. The following specific changes occur in relaxed mode: • Division by a constant is converted to inverse multiplication. • Certain C standard float functions--such as sqrt, sin, cos, atan, and atan2--are redirected to optimized inline functions where possible. • If the --tmu_support option is used to enable support for the Trigonometric Math Unit (TMU) and relaxed floating-point mode is enabled, RTS library calls are replaced with the corresponding TMU hardware instructions for the following floating-point operations: floating point division, sqrt, sin, cos, atan, and atan2. Note that there are algorithmic differences between the TMU hardware instructions and the library routines, so the results of operations may differ slightly. Enables or disables the reassociation of floating-point arithmetic. The default value is --fp_reassoc=on. Because floating-point values are of limited precision, and because floating-point operations round, floating-point arithmetic is neither associative nor distributive. For instance, (1 + 3e100) - 3e100 is not equal to 1 + (3e100 - 3e100). If strictly following IEEE 754, the compiler cannot, in general, reassociate floating-point operations. Using --fp_reassoc=on allows the compiler to perform the algebraic reassociation, at the cost of a small amount of precision for some operations. When --fp_reassoc=on, RPT MACF32 instructions may be generated. Because the RPT MACF32 instruction computes two partial sums and adds them together afterward to compute the entire accumulation, the result can vary in precision from a serial floating-point multiply accumulate loop. Sets the diagnostic severity for advisory MISRA-C:2004 rules. Sets the diagnostic severity for required MISRA-C:2004 rules. Includes the source code of filename at the beginning of the compilation. This can be used to establish standard macro definitions. The filename is searched for in the directories on the include search list. The files are processed in the order in which they were specified. Enables support for smaller, limited versions of the printf function family (sprintf, fprintf, etc.) and the scanf function family (sscanf, fscanf, etc.) run-time-support functions. The valid values are: • full: Supports all format specifiers. This is the default. • nofloat: Excludes support for printing and scanning floating-point values. Supports all format specifiers except %a, %A, %f, %F, %g, %G, %e, and %E. • minimal: Supports the printing and scanning of integer, char, or string values without width or precision flags. Specifically, only the %%, %d, %o, %c, %s, and %x format specifiers are supported There is no run-time error checking to detect if a format specifier is used for which support is not included. The --printf_support option precedes the --run_linker option, and must be used when performing the final link. Enables or disables the reassociation of saturating arithmetic. 30 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Changing the Compiler's Behavior with Options 2.3.4 Run-Time Model Options These options are specific to the TMS320C28x toolset. See the referenced sections for more information. TMS320C28x-specific assembler options are listed in Section 2.3.11. --cla_support={cla0|cla1|cla2} --disable_dp_load_opt --float_support={fpu32} --no_rpt --protect_volatile=num --ramfunc={on|off} --rpt_threshold=k --silicon_version=28 Specifies TMS320C28x Control Law Accelerator (CLA) Type 0, Type 1, or Type 2 support. This option is used to compile or assemble code written for the CLA. This option does not require any special library support when linking; the libraries used for C28x with/without FPU support should be sufficient. Disables the compiler from optimizing redundant loads of the DP register when using DP direct addressing. See Section 3.10 for more about when to use this option to disable data page blocking. Specifies use of TMS320C28x 32-bit hardware floating-point support. Using --float_support=fpu32 specifies the C28x architecture with 32bit hardware floating-point support. If the --tmu_support option is used to enable support for the Trigonometric Math Unit, the --float_support option is automatically set to fpu32. Prevents the compiler from generating repeat (RPT) instructions. By default, repeat instructions are generated for certain memcpy, division, and multiply-accumulate operations. However, repeat instructions are not interruptible. Enables volatile reference protection. Pipeline conflicts may occur between non-local variables that have been declared volatile. A conflict can occur between a write to one volatile variable that is followed by a read from a different volatile variable. The -protect_volatile option allows at least num instructions to be placed between the two volatile references to ensure the write occurs before the read. The num is optional. If no num is given, the default value is 2. For example, if --protect_volatile=4 is used, volatile writes and volatile reads are protected by at least 4 instructions. The peripheral pipeline protection hardware protects all internal peripherals and XINTF zone 1. If you connect peripherals to Xintf zone 0, 2, 6, 7 then you may need to use the --protect_volatile option. Hardware protection or using this option is not required for memories. If set to on, specifies that all functions should be placed in the .TI.ramfunc section, which is placed in RAM. If set to off, only functions with the ramfunc function attribute are treated this way. See Section 6.14.2. Newer TI linker command files support the --ramfunc option automatically by placing functions in the .TI.ramfunc section. If you have a linker command file that does not include a section specification for the .TI.ramfunc section, you can modify the linker command file to place this section in RAM. See the TMS320C28x Assembly Language Tools User's Guide for details on section placement. Projects or files that were previously compiled without the -no_fast_branch option should use the --ramfunc=on option instead to generate fast branch instructions for all functions. Generates RPT loops that iterate k times or less (k is a constant between 0 and 256). Multiple RPT’s may be generated for the same loop, if iteration count is more than k and if code size does not increase too much. Using this option when optimizing for code size disables RPT loop generation for loops whose iteration count can be greater than k. Generates code for the TMS320C28x architecture. The only value accepted is 28. This is the default, so this option is no longer required on command lines. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 31 Changing the Compiler's Behavior with Options www.ti.com --unified_memory --tmu_support[=tmu0] --vcu_support[=vcu0|vcu2] Use the --unified_memory (-mt) option if your memory map is configured as a single unified space; this allows the compiler to generate RPT PREAD instructions for most memcpy calls and structure assignments. This also allows MAC instructions to be generated. The --unified_memory option also allows more efficient data memory instructions to be used to access switch tables. Even under unified memory, memory for some peripherals and some RAM associated with those peripherals is allocated only in data memory. If –unified_memory is enabled, you can prevent program memory address access to specific symbols by declaring those symbols as volatile. Enables support for the Trigonometric Math Unit (TMU). Using this option automatically enables FPU32 support (as with the -float_support=fpu32 option). When TMU support is enabled, intrinsics are available to perform trigonometric instructions on the TMU. In addition, in relaxed floating point mode, RTS library calls are replaced with the corresponding TMU hardware instructions for the following floating-point operations: floating point division, sqrt, sin, cos, atan, and atan2. Note that there are algorithmic differences between the TMU hardware instructions and the library routines, so the results of operations may differ slightly. Specifies support for Type 0 or Type 2 of the Viterbi, Complex Math and CRC Unit (VCU). This option is useful only if the source is in assembly code, written for the VCU. The option is ignored for C/C++ code. This option does not need any special library support when linking; the libraries used for C28x with/without VCU support should be sufficient. Also note that there is no VCU Type 1. 2.3.5 Symbolic Debugging and Profiling Options The following options are used to select symbolic debugging or profiling: --symdebug:dwarf --symdebug:dwarf_ version={2|3|4} --symdebug:none --symdebug:profile_coff --symdebug:skeletal (Default) Generates directives that are used by the C/C++ source-level debugger and enables assembly source debugging in the assembler. The --symdebug:dwarf option's short form is -g. See Section 3.11. For more information on the DWARF debug format, see The DWARF Debugging Standard. Specifies the DWARF debugging format version (2, 3, or 4) to be generated when --symdebug:dwarf (the default) is specified. By default, the compiler generates DWARF version 3 debug information. DWARF versions 2, 3, and 4 may be intermixed safely. When DWARF 4 is used, type information is placed in the .debug_types section. At link time, duplicate type information will be removed. This method of type merging is superior to those used in DWARF 2 or 3 and will result in a smaller executable. In addition, DWARF 4 reduces the size of intermediate object files in comparison to DWARF 3. For more information on TI extensions to the DWARF language, see The Impact of DWARF on TI Object Files (SPRAAB5). Disables all symbolic debugging output. This option is not recommended; it prevents debugging and most performance analysis capabilities. Adds the necessary debug directives to the object file which are needed by the profiler to allow function level profiling with minimal impact on optimization (when used). This option does not hinder optimization. You can set breakpoints and profile on function-level boundaries in Code Composer Studio, but you cannot single-step through code as with full debug ability. Deprecated. Has no effect. 32 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Changing the Compiler's Behavior with Options 2.3.6 Specifying Filenames The input files that you specify on the command line can be C source files, C++ source files, assembly source files, or object files. The compiler uses filename extensions to determine the file type. Extension .asm, .abs, or .s* (extension begins with s) .c .C .cpp, .cxx, .cc .obj .o* .dll .so File Type Assembly source C source Depends on operating system C++ source Object NOTE: Case Sensitivity in Filename Extensions Case sensitivity in filename extensions is determined by your operating system. If your operating system is not case sensitive, a file with a .C extension is interpreted as a C file. If your operating system is case sensitive, a file with a .C extension is interpreted as a C++ file. For information about how you can alter the way that the compiler interprets individual filenames, see Section 2.3.7. For information about how you can alter the way that the compiler interprets and names the extensions of assembly source and object files, see Section 2.3.10. You can use wildcard characters to compile or assemble multiple files. Wildcard specifications vary by system; use the appropriate form listed in your operating system manual. For example, to compile all of the files in a directory with the extension .cpp, enter the following: cl2000 *.cpp NOTE: No Default Extension for Source Files is Assumed If you list a filename called example on the command line, the compiler assumes that the entire filename is example not example.c. No default extensions are added onto files that do not contain an extension. 2.3.7 Changing How the Compiler Interprets Filenames You can use options to change how the compiler interprets your filenames. If the extensions that you use are different from those recognized by the compiler, you can use the filename options to specify the type of file. You can insert an optional space between the option and the filename. Select the appropriate option for the type of file you want to specify: --asm_file=filename --c_file=filename --cpp_file=filename --obj_file=filename for an assembly language source file for a C source file for a C++ source file for an object file For example, if you have a C source file called file.s and an assembly language source file called assy, use the --asm_file and --c_file options to force the correct interpretation: cl2000 --c_file=file.s --asm_file=assy You cannot use the filename options with wildcard specifications. 2.3.8 Changing How the Compiler Processes C Files The --cpp_default option causes the compiler to process C files as C++ files. By default, the compiler treats files with a .c extension as C files. See Section 2.3.9 for more information about filename extension conventions. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 33 Changing the Compiler's Behavior with Options www.ti.com 2.3.9 Changing How the Compiler Interprets and Names Extensions You can use options to change how the compiler program interprets filename extensions and names the extensions of the files that it creates. The filename extension options must precede the filenames they apply to on the command line. You can use wildcard specifications with these options. An extension can be up to nine characters in length. Select the appropriate option for the type of extension you want to specify: --asm_extension=new extension --c_extension=new extension --cpp_extension=new extension --listing_extension=new extension --obj_extension=new extension for an assembly language file for a C source file for a C++ source file sets default extension for listing files for an object file The following example assembles the file fit.rrr and creates an object file named fit.o: cl2000 --asm_extension=.rrr --obj_extension=.o fit.rrr The period (.) in the extension is optional. You can also write the example above as: cl2000 --asm_extension=rrr --obj_extension=o fit.rrr 2.3.10 Specifying Directories By default, the compiler program places the object, assembly, and temporary files that it creates into the current directory. If you want the compiler program to place these files in different directories, use the following options: --abs_directory=directory --asm_directory=directory --list_directory=directory --obj_directory=directory --output_file=filename --pp_directory=directory --temp_directory=directory Specifies the destination directory for absolute listing files. The default is to use the same directory as the object file directory. For example: cl2000 --abs_directory=d:\abso_list Specifies a directory for assembly files. For example: cl2000 --asm_directory=d:\assembly Specifies the destination directory for assembly listing files and crossreference listing files. The default is to use the same directory as the object file directory. For example: cl2000 --list_directory=d:\listing Specifies a directory for object files. For example: cl2000 --obj_directory=d:\object Specifies a compilation output file name; can override --obj_directory . For example: cl2000 --output_file=transfer Specifies a preprocessor file directory for object files (default is .). For example: cl2000 --pp_directory=d:\preproc Specifies a directory for temporary intermediate files. For example: cl2000 --temp_directory=d:\temp 2.3.11 Assembler Options Following are assembler options that you can use with the compiler. For more information, see the TMS320C28x Assembly Language Tools User's Guide. 34 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Changing the Compiler's Behavior with Options --absolute_listing --asm_define=name[=def] --asm_dependency --asm_includes --asm_listing --issue_remarks --asm_undefine=name -asm_listing_cross_referen ce --flash_prefetch_warn --include_file=filename --preproc_asm Generates a listing with absolute addresses rather than section-relative offsets. Predefines the constant name for the assembler; produces a .set directive for a constant or an .arg directive for a string. If the optional [=def] is omitted, the name is set to 1. If you want to define a quoted string and keep the quotation marks, do one of the following: • For Windows, use --asm_define=name="\"string def\"". For example: -asm_define=car="\"sedan\"" • For UNIX, use --asm_define=name='"string def"'. For example: -asm_define=car='"sedan"' • For Code Composer Studio, enter the definition in a file and include that file with the --cmd_file option. Performs preprocessing for assembly files, but instead of writing preprocessed output, writes a list of dependency lines suitable for input to a standard make utility. The list is written to a file with the same name as the source file but with a .ppa extension. Performs preprocessing for assembly files, but instead of writing preprocessed output, writes a list of files included with the #include directive. The list is written to a file with the same name as the source file but with a .ppa extension. Produces an assembly listing file. Issues remarks (non-serious warnings). For the assembler, enables additional assembly-time checking. A remark is generated if an .ebss allocation size is greater than 64 words, or if a 16-bit immediate operand value resides outside of the -32 768 to 65 535 range. Undefines the predefined constant name. This option overrides any -asm_define options for the specified name. Produces a symbolic cross-reference in the listing file. Enables assembler warnings if a program data access instruction follows within 8 words of a BF or SBF instruction. As outlined in the TMS320C281X/TMS320F281X DSP Silicon Errata (SPRZ193) advisory on the "Flash and OTP Prefetch Buffer Overflow", the flash prefetch buffer may overflow if this instruction sequence is executed from flash or OneTime-Programmable (OTP) memory with the flash prefetch buffer enabled. Whether or not an overflow actually occurs depends on the instruction sequence, flash wait states, and CPU pipeline stalls. If an overflow occurs, it will result in execution of invalid opcodes. Instructions that use program memory addressing include MAC/XMAC, DMAC/XMACD, QMACL, IMACL, PREAD/XPREAD, and PWRITE/XPWRITE. Includes the specified file for the assembly module; acts like an .include directive. The file is included before source file statements. The included file does not appear in the assembly listing files. Expands macros in an assembly file and assembles the expanded file. Expanding macros helps you to debug the assembly file. The -preproc_asm option affects only the assembly file. When --preproc_asm is used, the compiler first invokes the assembler to generate the macroexpanded source .exp file. Then the .exp file is assembled to generate the object file. The debugger uses the .exp file for debugging. The .exp file is an intermediate file and any update to this file will be lost. You need to make any updates to the original assembly file. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 35 Changing the Compiler's Behavior with Options www.ti.com 2.3.12 Dynamic Linking The C28x v?.? Code Generation Tools (CGT) support dynamic linking. For details on dynamic linking with the C28x CGT, see the TMS320C28x Assembly Language Tools User's Guide. Table 2-29 provides a brief summary of the linker options that are related to support for the Dynamic Linking Model in the C28x CGT. Option --dynamic[=exe] --dynamic=lib --export=symbol --fini=symbol --import=symbol --init=symbol --rpath=dir --runpath=dir --soname=string Table 2-29. Linker Options For Dynamic Linking Description Specifies that the result of a link will be a lightweight dynamic executable. Specifies that the result of a link will be a dynamic library. Specifies that symbol is exported by the ELF object that is generated for this link. Specifies the symbol name of the termination code for the output file currently being linked. Specifies that symbol is imported by the ELF object that is generated for this link. Specifies the symbol name of the initialization code for the output file currently being linked. Adds a directory to the beginning of the dynamic library search path. Adds a directory to the end of the dynamic library search path. Specifies shared object name to be used to identify this ELF object to the any downstream ELF object consumers. 2.3.13 Deprecated Options Several compiler options have been deprecated, removed, or renamed. The compiler continues to accept some of the deprecated options, but they are not recommended for use. See the Compiler Option Cleanup wiki page for a list of deprecated and removed options, options that have been removed from CCS, and options that have been renamed. 36 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Controlling the Compiler Through Environment Variables 2.4 Controlling the Compiler Through Environment Variables An environment variable is a system symbol that you define and assign a string to. Setting environment variables is useful when you want to run the compiler repeatedly without re-entering options, input filenames, or pathnames. NOTE: C_OPTION and C_DIR -- The C_OPTION and C_DIR environment variables are deprecated. Use device-specific environment variables instead. 2.4.1 Setting Default Compiler Options (C2000_C_OPTION) You might find it useful to set the compiler, assembler, and linker default options using the C2000_C_OPTION environment variable. If you do this, the compiler uses the default options and/or input filenames that you name C2000_C_OPTION every time you run the compiler. Setting the default options with these environment variables is useful when you want to run the compiler repeatedly with the same set of options and/or input files. After the compiler reads the command line and the input filenames, it looks for the C2000_C_OPTION environment variable and processes it. The table below shows how to set the C2000_C_OPTION environment variable. Select the command for your operating system: Operating System UNIX (Bourne shell) Windows Enter C2000_C_OPTION=" option1 [option2 . . .]"; export C2000_C_OPTION set C2000_C_OPTION= option1 [option2 . . .] Environment variable options are specified in the same way and have the same meaning as they do on the command line. For example, if you want to always run quietly (the --quiet option), enable C/C++ source interlisting (the --src_interlist option), and link (the --run_linker option) for Windows, set up the C2000_C_OPTION environment variable as follows: set C2000_C_OPTION=--quiet --src_interlist --run_linker In the following examples, each time you run the compiler, it runs the linker. Any options following -run_linker on the command line or in C2000_C_OPTION are passed to the linker. Thus, you can use the C2000_C_OPTION environment variable to specify default compiler and linker options and then specify additional compiler and linker options on the command line. If you have set --run_linker in the environment variable and want to compile only, use the compiler --compile_only option. These additional examples assume C2000_C_OPTION is set as shown above: cl2000 *c ; compiles and links cl2000 --compile_only *.c ; only compiles cl2000 *.c --run_linker lnk.cmd ; compiles and links using a command file cl2000 --compile_only *.c --run_linker lnk.cmd ; only compiles (--compile_only overrides --run_linker) For details on compiler options, see Section 2.3. For details on linker options, see the Linker Description chapter in the TMS320C28x Assembly Language Tools User's Guide. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 37 Controlling the Compiler Through Environment Variables www.ti.com 2.4.2 Naming One or More Alternate Directories (C2000_C_DIR) The linker uses the C2000_C_DIR environment variable to name alternate directories that contain object libraries. The command syntaxes for assigning the environment variable are: Operating System UNIX (Bourne shell) Windows Enter C2000_C_DIR=" pathname1 ; pathname2 ;..."; export C2000_C_DIR set C2000_C_DIR= pathname1 ; pathname2 ;... The pathnames are directories that contain input files. The pathnames must follow these constraints: • Pathnames must be separated with a semicolon. • Spaces or tabs at the beginning or end of a path are ignored. For example, the space before and after the semicolon in the following is ignored: set C2000_C_DIR=c:\path\one\to\tools ; c:\path\two\to\tools • Spaces and tabs are allowed within paths to accommodate Windows directories that contain spaces. For example, the pathnames in the following are valid: set C2000_C_DIR=c:\first path\to\tools;d:\second path\to\tools The environment variable remains set until you reboot the system or reset the variable by entering: Operating System UNIX (Bourne shell) Windows Enter unset C2000_C_DIR set C2000_C_DIR= 2.5 Controlling the Preprocessor This section describes features that control the preprocessor, which is part of the parser. A general description of C preprocessing is in section A12 of K&R. The C/C++ compiler includes standard C/C++ preprocessing functions, which are built into the first pass of the compiler. The preprocessor handles: • Macro definitions and expansions • #include files • Conditional compilation • Various preprocessor directives, specified in the source file as lines beginning with the # character The preprocessor produces self-explanatory error messages. The line number and the filename where the error occurred are printed along with a diagnostic message. 38 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Controlling the Preprocessor 2.5.1 Predefined Macro Names The compiler maintains and recognizes the predefined macro names listed in Table 2-30. Table 2-30. Predefined C28x Macro Names Macro Name Description _ _DATE_ _(1) Expands to the compilation date in the form mmm dd yyyy _ _FILE_ _(1) Expands to the current source filename _ _LINE_ _(1) Expands to the current line number _ _little_endian_ _ Always defined _ _PTRDIFF_T_TYPE_ _ Set to the type of ptrdiff_t. _ _SIZE_T_TYPE_ _ Set to the type of size_t. _ _STDC_ _(1) Defined to indicate that compiler conforms to ISO C Standard. See Section 6.1 for exceptions to ISO C conformance. _ _STDC_VERSION_ _ C standard macro _ _TI_COMPILER_VERSION_ _ Defined to a 7-9 digit integer, depending on if X has 1, 2, or 3 digits. The number does not contain a decimal. For example, version 3.2.1 is represented as 3002001. The leading zeros are dropped to prevent the number being interpreted as an octal. _ _TI_GNU_ATTRIBUTE_SUPPORT_ _ Defined if GCC extensions are enabled (which is the default) _ _TI_STRICT_ANSI_MODE__ Defined if strict ANSI/ISO mode is enabled (the --strict_ansi option is used); otherwise, it is undefined. _ _TI_STRICT_FP_MODE_ _ Defined to 1 if --fp_mode=strict is used (or implied); otherwise, it is undefined. _ _TIME_ _(1) Expands to the compilation time in the form "hh:mm:ss" _ _TMS320C2000_ _ Defined for C28x processor _ _TMS320C28XX_ _ Defined if target is C28x __TMS320C28XX_CLA__ Defined to 1 if any --cla_support option was used. __TMS320C28XX_CLA0__ Defined to 1 if the --cla_support=cla0 option was used. __TMS320C28XX_CLA1__ Defined to 1 if the --cla_support=cla1 option was used. __TMS320C28XX_CLA2__ Defined to 1 if the --cla_support=cla2 option was used. _ _TMS320C28XX_FPU32_ _ Expands to 1 (identifies the C28x processor with 32-bit hardware floating-point support). _ _TMS320C28XX_TMU_ _ Defined to 1 if the --tmu_support option was used to enable the Trigonometric Math Unit. _ _WCHAR_T_TYPE_ _ Set to the type of wchar_t. _INLINE Expands to 1 if optimization is used (--opt_level or -O option); undefined otherwise. (1) Specified by the ISO standard You can use the names listed in Table 2-30 in the same manner as any other defined name. For example, printf ( "%s %s" , __TIME__ , __DATE__); translates to a line such as: printf ("%s %s" , "13:58:17", "Jan 14 1997"); SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 39 Controlling the Preprocessor www.ti.com 2.5.2 The Search Path for #include Files The #include preprocessor directive tells the compiler to read source statements from another file. When specifying the file, you can enclose the filename in double quotes or in angle brackets. The filename can be a complete pathname, partial path information, or a filename with no path information. • If you enclose the filename in double quotes (" "), the compiler searches for the file in the following directories in this order: 1. The directory of the file that contains the #include directive and in the directories of any files that contain that file. 2. Directories named with the --include_path option. 3. Directories set with the C2000_C_DIR environment variable. • If you enclose the filename in angle brackets (< >), the compiler searches for the file in the following directories in this order: 1. Directories named with the --include_path option. 2. Directories set with the C2000_C_DIR environment variable. See Section 2.5.2.1 for information on using the --include_path option. See Section 2.4.2 for more information on input file directories. 2.5.2.1 Adding a Directory to the #include File Search Path (--include_path Option) The --include_path option names an alternate directory that contains #include files. The --include_path option's short form is -I. The format of the --include_path option is: --include_path=directory1 [--include_path= directory2 ...] There is no limit to the number of --include_path options per invocation of the compiler; each -include_path option names one directory. In C source, you can use the #include directive without specifying any directory information for the file; instead, you can specify the directory information with the -include_path option. For example, assume that a file called source.c is in the current directory. The file source.c contains the following directive statement: #include "alt.h" Assume that the complete pathname for alt.h is: UNIX Windows /tools/files/alt.h c:\tools\files\alt.h 40 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Controlling the Preprocessor The table below shows how to invoke the compiler. Select the command for your operating system: Operating System UNIX Windows Enter cl2000 --include_path=/tools/files source.c cl2000 --include_path=c:\tools\files source.c NOTE: Specifying Path Information in Angle Brackets If you specify the path information in angle brackets, the compiler applies that information relative to the path information specified with --include_path options and the C2000_C_DIR environment variable. For example, if you set up C2000_C_DIR with the following command: C2000_C_DIR "/usr/include;/usr/ucb"; export C2000_C_DIR or invoke the compiler with the following command: cl2000 --include_path=/usr/include file.c and file.c contains this line: #include the result is that the included file is in the following path: /usr/include/sys/proc.h 2.5.3 Support for the #warning and #warn Directives In strict ANSI mode, the TI preprocessor allows you to use the #warn directive to cause the preprocessor to issue a warning and continue preprocessing. The #warn directive is equivalent to the #warning directive supported by GCC, IAR, and other compilers. If you use the --relaxed_ansi option (on by default), both the #warn and #warning preprocessor directives are supported. 2.5.4 Generating a Preprocessed Listing File (--preproc_only Option) The --preproc_only option allows you to generate a preprocessed version of your source file with an extension of .pp. The compiler's preprocessing functions perform the following operations on the source file: • Each source line ending in a backslash (\) is joined with the following line. • Trigraph sequences are expanded. • Comments are removed. • #include files are copied into the file. • Macro definitions are processed. • All macros are expanded. • All other preprocessing directives, including #line directives and conditional compilation, are expanded. The --preproc_only option is useful when creating a source file for a technical support case or to ask a question about your code. It allows you to reduce the test case to a single source file, because #include files are incorporated when the preprocessor runs. 2.5.5 Continuing Compilation After Preprocessing (--preproc_with_compile Option) If you are preprocessing, the preprocessor performs preprocessing only; it does not compile your source code. To override this feature and continue to compile after your source code is preprocessed, use the -preproc_with_compile option along with the other preprocessing options. For example, use -preproc_with_compile with --preproc_only to perform preprocessing, write preprocessed output to a file with a .pp extension, and compile your source code. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 41 Controlling the Preprocessor www.ti.com 2.5.6 Generating a Preprocessed Listing File with Comments (--preproc_with_comment Option) The --preproc_with_comment option performs all of the preprocessing functions except removing comments and generates a preprocessed version of your source file with a .pp extension. Use the -preproc_with_comment option instead of the --preproc_only option if you want to keep the comments. 2.5.7 Generating Preprocessed Listing with Line-Control Details (--preproc_with_line Option) By default, the preprocessed output file contains no preprocessor directives. To include the #line directives, use the --preproc_with_line option. The --preproc_with_line option performs preprocessing only and writes preprocessed output with line-control information (#line directives) to a file named as the source file but with a .pp extension. 2.5.8 Generating Preprocessed Output for a Make Utility (--preproc_dependency Option) The --preproc_dependency option performs preprocessing only. Instead of writing preprocessed output, it writes a list of dependency lines suitable for input to a standard make utility. If you do not supply an optional filename, the list is written to a file with the same name as the source file but a .pp extension. 2.5.9 Generating a List of Files Included with #include (--preproc_includes Option) The --preproc_includes option performs preprocessing only, but instead of writing preprocessed output, writes a list of files included with the #include directive. If you do not supply an optional filename, the list is written to a file with the same name as the source file but with a .pp extension. 2.5.10 Generating a List of Macros in a File (--preproc_macros Option) The --preproc_macros option generates a list of all predefined and user-defined macros. If you do not supply an optional filename, the list is written to a file with the same name as the source file but with a .pp extension. Predefined macros are listed first and indicated by the comment /* Predefined */. User-defined macros are listed next and indicated by the source filename. 2.6 Passing Arguments to main() Some programs pass arguments to main() via argc and argv. This presents special challenges in an embedded program that is not run from the command line. In general, argc and argv are made available to your program through the .args section. There are various ways to populate the contents of this section for use by your program. To cause the linker to allocate an .args section of the appropriate size, use the --arg_size=size linker option. This option tells the linker to allocate an uninitialized section named .args, which can be used by the loader to pass arguments from the command line of the loader to the program. The size is the number of bytes to be allocated. When you use the --arg_size option, the linker defines the __c_args__ symbol to contain the address of the .args section. It is the responsibility of the loader to populate the .args section. The loader and the target boot code can use the .args section and the __c_args__ symbol to determine whether and how to pass arguments from the host to the target program. The format of the arguments is an array of pointers to char on the target. Due to variations in loaders, it is not specified how the loader determines which arguments to pass to the target. 42 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Understanding Diagnostic Messages If you are using Code Composer Studio to run your application, you can use the Scripting Console tool to populate the .args section. To open this tool, choose View > Scripting Console from the CCS menus. You can use the loadProg command to load an object file and its associated symbol table into memory and pass an array of arguments to main(). These arguments are automatically written to the allocated .args section. The loadProg syntax is as follows, where file is an executable file and args is an object array of arguments. Use JavaScript to declare the array of arguments before using this command. loadProg(file, args) The .args section is loaded with the following data for non-SYS/BIOS-based executables, where each element in the argv[] array contains a string corresponding to that argument: Int argc; Char * argv[0]; Char * argv[1]; ... Char * argv[n]; For SYS/BIOS-based executables, the elements in the .args section are as follows: Int argc; Char ** argv; Char * envp; Char * argv[0]; Char * argv[1]; ... Char * argv[n]; /* points to argv[0] */ /* ignored by loadProg command */ For more details, see the "Scripting Console" topic in the TI Processors Wiki. 2.7 Understanding Diagnostic Messages One of the primary functions of the compiler and linker is to report diagnostic messages for the source program. A diagnostic message indicates that something may be wrong with the program. When the compiler or linker detects a suspect condition, it displays a message in the following format: "file.c", line n : diagnostic severity : diagnostic message "file.c" line n : diagnostic severity diagnostic message The name of the file involved The line number where the diagnostic applies The diagnostic message severity (severity category descriptions follow) The text that describes the problem Diagnostic messages have a severity, as follows: • A fatal error indicates a problem so severe that the compilation cannot continue. Examples of such problems include command-line errors, internal errors, and missing include files. If multiple source files are being compiled, any source files after the current one will not be compiled. • An error indicates a violation of the syntax or semantic rules of the C/C++ language. Compilation may continue, but object code is not generated. • A warning indicates something that is likely to be a problem, but cannot be proven to be an error. For example, the compiler emits a warning for an unused variable. An unused variable does not affect program execution, but its existence suggests that you might have meant to use it. Compilation continues and object code is generated (if no errors are detected). • A remark is less serious than a warning. It may indicate something that is a potential problem in rare cases, or the remark may be strictly informational. Compilation continues and object code is generated (if no errors are detected). By default, remarks are not issued. Use the --issue_remarks compiler option to enable remarks. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 43 Understanding Diagnostic Messages www.ti.com Diagnostic messages are written to standard error with a form like the following example: "test.c", line 5: error: a break statement may only be used within a loop or switch break; ^ By default, the source code line is not printed. Use the --verbose_diagnostics compiler option to display the source line and the error position. The above example makes use of this option. The message identifies the file and line involved in the diagnostic, and the source line itself (with the position indicated by the ^ character) follows the message. If several diagnostic messages apply to one source line, each diagnostic has the form shown; the text of the source line is displayed several times, with an appropriate position indicated each time. Long messages are wrapped to additional lines, when necessary. You can use the --display_error_number command-line option to request that the diagnostic's numeric identifier be included in the diagnostic message. When displayed, the diagnostic identifier also indicates whether the diagnostic can have its severity overridden on the command line. If the severity can be overridden, the diagnostic identifier includes the suffix -D (for discretionary); otherwise, no suffix is present. For example: "Test_name.c", line 7: error #64-D: declaration does not declare anything struct {}; ^ "Test_name.c", line 9: error #77: this declaration has no storage class or type specifier xxxxx; ^ Because errors are determined to be discretionary based on the severity in a specific context, an error can be discretionary in some cases and not in others. All warnings and remarks are discretionary. For some messages, a list of entities (functions, local variables, source files, etc.) is useful; the entities are listed following the initial error message: "test.c", line 4: error: more than one instance of overloaded function "f" matches the argument list: function "f(int)" function "f(float)" argument types are: (double) f(1.5); ^ In some cases, additional context information is provided. Specifically, the context information is useful when the front end issues a diagnostic while doing a template instantiation or while generating a constructor, destructor, or assignment operator function. For example: "test.c", line 7: error: "A::A()" is inaccessible B x; ^ detected during implicit generation of "B::B()" at line 7 Without the context information, it is difficult to determine to what the error refers. 2.7.1 Controlling Diagnostic Messages The C/C++ compiler provides diagnostic options to control compiler- and linker-generated diagnostic messages. The diagnostic options must be specified before the --run_linker option. --diag_error=num Categorizes the diagnostic identified by num as an error. To determine the numeric identifier of a diagnostic message, use the --display_error_number option first in a separate compile. Then use --diag_error=num to recategorize the diagnostic as an error. You can only alter the severity of discretionary diagnostic messages. 44 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Understanding Diagnostic Messages --diag_remark=num Categorizes the diagnostic identified by num as a remark. To determine the numeric identifier of a diagnostic message, use the --display_error_number option first in a separate compile. Then use --diag_remark=num to recategorize the diagnostic as a remark. You can only alter the severity of discretionary diagnostic messages. --diag_suppress=num Suppresses the diagnostic identified by num. To determine the numeric identifier of a diagnostic message, use the --display_error_number option first in a separate compile. Then use --diag_suppress=num to suppress the diagnostic. You can only suppress discretionary diagnostic messages. --diag_warning=num Categorizes the diagnostic identified by num as a warning. To determine the numeric identifier of a diagnostic message, use the --display_error_number option first in a separate compile. Then use --diag_warning=num to recategorize the diagnostic as a warning. You can only alter the severity of discretionary diagnostic messages. --display_error_number Displays a diagnostic's numeric identifier along with its text. Use this option in determining which arguments you need to supply to the diagnostic suppression options (--diag_suppress, --diag_error, --diag_remark, and -diag_warning). This option also indicates whether a diagnostic is discretionary. A discretionary diagnostic is one whose severity can be overridden. A discretionary diagnostic includes the suffix -D; otherwise, no suffix is present. See Section 2.7. --emit_warnings_as_ errors Treats all warnings as errors. This option cannot be used with the -no_warnings option. The --diag_remark option takes precedence over this option. This option takes precedence over the --diag_warning option. --issue_remarks Issues remarks (non-serious warnings), which are suppressed by default. --no_warnings Suppresses diagnostic warnings (errors are still issued). --set_error_limit=num Sets the error limit to num, which can be any decimal value. The compiler abandons compiling after this number of errors. (The default is 100.) --verbose_diagnostics Provides verbose diagnostic messages that display the original source with line-wrap and indicate the position of the error in the source line. Note that this command-line option cannot be used within the Code Composer Studio IDE. --write_diagnostics_file Produces a diagnostic message information file with the same source file name with an .err extension. (The --write_diagnostics_file option is not supported by the linker.) Note that this command-line option cannot be used within the Code Composer Studio IDE. 2.7.2 How You Can Use Diagnostic Suppression Options The following example demonstrates how you can control diagnostic messages issued by the compiler. You control the linker diagnostic messages in a similar manner. int one(); int I; int main() { switch (I){ case 1; return one (); break; default: return 0; break; } } If you invoke the compiler with the --quiet option, this is the result: "err.c", line 9: warning: statement is unreachable "err.c", line 12: warning: statement is unreachable SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 45 Other Messages www.ti.com Because it is standard programming practice to include break statements at the end of each case arm to avoid the fall-through condition, these warnings can be ignored. Using the --display_error_number option, you can find out the diagnostic identifier for these warnings. Here is the result: [err.c] "err.c", line 9: warning #111-D: statement is unreachable "err.c", line 12: warning #111-D: statement is unreachable Next, you can use the diagnostic identifier of 111 as the argument to the --diag_remark option to treat this warning as a remark. This compilation now produces no diagnostic messages (because remarks are disabled by default). NOTE: You can suppress any non-fatal errors, but be careful to make sure you only suppress diagnostic messages that you understand and are known not to affect the correctness of your program. 2.8 Other Messages Other error messages that are unrelated to the source, such as incorrect command-line syntax or inability to find specified files, are usually fatal. They are identified by the symbol >> preceding the message. 2.9 Generating Cross-Reference Listing Information (--gen_cross_reference Option) The --gen_cross_reference option generates a cross-reference listing file that contains reference information for each identifier in the source file. (The --gen_cross_reference option is separate from -asm_listing_cross_reference, which is an assembler rather than a compiler option.) The cross-reference listing file has the same name as the source file with a .crl extension. The information in the cross-reference listing file is displayed in the following format: sym-id name X filename line number column number sym-id name X filename line number column number An integer uniquely assigned to each identifier The identifier name One of the following values: D Definition d Declaration (not a definition) M Modification A Address taken U Used C Changed (used and modified in a single operation) R Any other kind of reference E Error; reference is indeterminate The source file The line number in the source file The column number in the source file 46 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Generating a Raw Listing File (--gen_preprocessor_listing Option) 2.10 Generating a Raw Listing File (--gen_preprocessor_listing Option) The --gen_preprocessor_listing option generates a raw listing file that can help you understand how the compiler is preprocessing your source file. Whereas the preprocessed listing file (generated with the -preproc_only, --preproc_with_comment, --preproc_with_line, and --preproc_dependency preprocessor options) shows a preprocessed version of your source file, a raw listing file provides a comparison between the original source line and the preprocessed output. The raw listing file has the same name as the corresponding source file with an .rl extension. The raw listing file contains the following information: • Each original source line • Transitions into and out of include files • Diagnostic messages • Preprocessed source line if nontrivial processing was performed (comment removal is considered trivial; other preprocessing is nontrivial) Each source line in the raw listing file begins with one of the identifiers listed in Table 2-31. Identifier N X S L Table 2-31. Raw Listing File Identifiers Definition Normal line of source Expanded line of source. It appears immediately following the normal line of source if nontrivial preprocessing occurs. Skipped source line (false #if clause) Change in source position, given in the following format: L line number filename key Where line number is the line number in the source file. The key is present only when the change is due to entry/exit of an include file. Possible values of key are: 1 = entry into an include file 2 = exit from an include file The --gen_preprocessor_listing option also includes diagnostic identifiers as defined in Table 2-32. Table 2-32. Raw Listing File Diagnostic Identifiers Diagnostic Identifier E F R W Definition Error Fatal Remark Warning Diagnostic raw listing information is displayed in the following format: S filename line number column number diagnostic S filename line number column number diagnostic One of the identifiers in Table 2-32 that indicates the severity of the diagnostic The source file The line number in the source file The column number in the source file The message text for the diagnostic Diagnostic messages after the end of file are indicated as the last line of the file with a column number of 0. When diagnostic message text requires more than one line, each subsequent line contains the same file, line, and column information but uses a lowercase version of the diagnostic identifier. For more information about diagnostic messages, see Section 2.7. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 47 Using Inline Function Expansion www.ti.com 2.11 Using Inline Function Expansion When an inline function is called, a copy of the C/C++ source code for the function is inserted at the point of the call. This is known as inline function expansion, commonly called function inlining or just inlining. Inline function expansion can speed up execution by eliminating function call overhead. This is particularly beneficial for very small functions that are called frequently. Function inlining involves a tradeoff between execution speed and code size, because the code is duplicated at each function call site. Large functions that are called frequently are poor candidates for inlining. Function inlining is triggered by the following situations: • The use of built-in intrinsic operations. Intrinsic operations look like function calls, and are inlined automatically, even though no function body exists. • Use of the inline keyword or the equivalent __inline keyword. Functions declared with the inline keyword may be inlined by the compiler if you set --opt_level=3 or greater. The inline keyword is a suggestion from the programmer to the compiler. Even if your optimization level is high, inlining is still optional for the compiler. The compiler decides whether to inline a function based on the length of the function, the number of times it is called, your --opt_for_speed setting, your --opt_for_space setting, and any contents of the function that disqualify it from inlining (see Section 2.11.4). Functions can be inlined at --opt_level=3 if the function body is visible in the same module or if -pm is also used and the function is visible in one of the modules being compiled. Functions may be inlined at link time if the file containing the definition and the call site were both compiled with --opt_level=4. • Use of static inline functions. Functions defined as both static and inline are more likely to be inlined. • When --opt_level=3 or greater is used, the compiler may automatically inline eligible functions even if they are not declared as inline functions. The same list of decision factors listed for functions explicitly defined with the inline keyword is used. For more about automatic function inlining, see Section 3.8 • Definition-controlled inlining with the unguarded inline keyword • Definition-controlled inlining with the guarded inline keyword • The pragma FUNC_ALWAYS_INLINE forces a function to be inlined (where it is legal to do so) unless --opt_level=off. That is, the pragma FUNC_ALWAYS_INLINE forces function inlining even if -opt_level=0 or --opt_level=1. NOTE: Function Inlining Can Greatly Increase Code Size Function inlining increases code size, especially inlining a function that is called in a number of places. Function inlining is optimal for functions that are called only from a small number of places and for small functions. The semantics of the inline keyword in C code follow the C99 standard. The semantics of the inline keyword in C++ code follow the C++ standard. The inline keyword is supported in all C++ modes, in relaxed ANSI mode for all C standards, and in strict ANSI mode for C99. It is disabled in strict ANSI mode for C89, because it is a language extension that could conflict with a strictly conforming program. If you want to define inline functions while in strict ANSI C89 mode, use the alternate keyword _ _inline. Compiler options that affect inlining are: --opt_level, --auto_inline, -remove_hooks_when_inlining, -opt_for_space, and -opt_for_speed. In most cases, inlining will reduce the code size by a small amount. 48 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Using Inline Function Expansion 2.11.1 Inlining Intrinsic Operators The compiler has built-in function-like operations called intrinsics. Intrinsic operations look like function calls, and can be implemented very efficiently with the target's instruction set. There are many intrinsic operators for the C28x. All of them are automatically inlined by the compiler. The inlining happens automatically whether or not you use the optimizer. For details about intrinsics, and a list of the intrinsics, see Section 7.5.6. 2.11.2 Unguarded Definition-Controlled Inlining The inline keyword specifies that a function is expanded inline at the point at which it is called rather than by using standard calling procedures. The compiler performs inline expansion of functions declared with the inline keyword. You must invoke the optimizer with any --opt_level option (--opt_level=0, --opt_level=1, --opt_level=2, or -opt_level=3) to turn on definition-controlled inlining. Automatic inlining is also turned on when using -opt_level=3. Example 2-1 shows usage of the inline keyword, where the function call is replaced by the code in the called function. Example 2-1. Using the Inline Keyword inline float volume_sphere(float r) { return 4.0/3.0 * PI * r * r * r; } int foo(...) { ... volume = volume_sphere(radius); ... } SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 49 Using Inline Function Expansion www.ti.com 2.11.3 Guarded Inlining and the _INLINE Preprocessor Symbol When declaring a function in a header file as static inline, you must follow additional procedures to avoid a potential code size increase the optimizer is not run. To prevent a static inline function in a header file from causing an increase in code size when inlining gets turned off, use the following procedure. This allows external-linkage when inlining is turned off; thus, only one function definition will exist throughout the object files. • Prototype a static inline version of the function. Then, prototype an alternative, nonstatic, externallylinked version of the function. Conditionally preprocess these two prototypes with the _INLINE preprocessor symbol, as shown in Example 2-2. • Create an identical version of the function definition in a .c or .cpp file, as shown in Example 2-3. In the following examples there are two definitions of the strlen function. The first (Example 2-2), in the header file, is an inline definition. This definition is enabled and the prototype is declared as static inline only if _INLINE is true (_INLINE is automatically defined for you when the optimizer is used). The second definition (see Example 2-3) for the library, ensures that the callable version of strlen exists when inlining is disabled. Since this is not an inline function, the _INLINE preprocessor symbol is undefined (#undef) before string.h is included to generate a noninline version of strlen's prototype. Example 2-2. Header File string.h /*****************************************************************************/ /* string.h vx.xx */ /* Copyright (c) 1993-2006 Texas Instruments Incorporated */ /* Excerpted ... */ /*****************************************************************************/ #ifdef _INLINE #define _IDECL static inline #else #define _IDECL extern _CODE_ACCESS #endif _IDECL size_t strlen(const char *_string); #ifdef _INLINE /****************************************************************************/ /* strlen */ /****************************************************************************/ static inline size_t strlen(const char *string) { size_t n = (size_t)-1; const char *s = string - 1; do n++; while (*++s); return n } #endif 50 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Using Inline Function Expansion Example 2-3. Library Definition File /****************************************************************************/ /* strlen */ /****************************************************************************/ #undef _INLINE #include _CODE_ACCESS size_t strlen(cont char * string) { size_t n = (size_t)-1; const char *s = string - 1; do n++; while (*++s); return n; } 2.11.4 Inlining Restrictions The compiler makes decisions about which functions to inline based on the factors mentioned in Section 2.11. In addition, there are several restrictions that can disqualify a function from being inlined by automatic inlining or inline keyword-based inlining. The FUNC_ALWAYS_INLINE pragma overrides these disqualifications, so you should be aware of situations that can cause problems if you are using the FUNC_ALWAYS_INLINE pragma. • Is not defined in the current compilation unit and you are not using -O4 optimization • Never returns • Is a recursive or nonleaf function that exceeds the depth limit • Has a variable-length argument list • Has a different number of arguments than the call site • Has an argument whose type is incompatible with the corresponding call site argument • Has a class, struct, or union parameter • Contains a volatile local variable or argument • Contains local static variables but is not a static inline function. • Is not declared inline and contains an asm() statement that is not a comment • Is the main() function • Is an interrupt function • Is not declared inline and returns void but its return value is needed. • Is not declared inline and will require too much stack space for local array or structure variables. Furthermore, inlining should be used for small functions or functions that are called in a few places (though the compiler does not enforce this). NOTE: Excessive Inlining Can Degrade Performance Excessive inlining can make the compiler dramatically slower and degrade the performance of generated code. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 51 Using Interlist www.ti.com 2.12 Using Interlist The compiler tools include a feature that interlists C/C++ source statements into the assembly language output of the compiler. The interlist feature enables you to inspect the assembly code generated for each C statement. The interlist behaves differently, depending on whether or not the optimizer is used, and depending on which options you specify. The easiest way to invoke the interlist feature is to use the --c_src_interlist option. To compile and run the interlist on a program called function.c, enter: cl2000 --c_src_interlist function The --c_src_interlist option prevents the compiler from deleting the interlisted assembly language output file. The output assembly file, function.asm, is assembled normally. When you invoke the interlist feature without the optimizer, the interlist runs as a separate pass between the code generator and the assembler. It reads both the assembly and C/C++ source files, merges them, and writes the C/C++ statements into the assembly file as comments. Using the --c_src_interlist option can cause performance and/or code size degradation. Example 2-4 shows a typical interlisted assembly file. For more information about using the interlist feature with the optimizer, see Section 3.9. Example 2‑4. An Interlisted Assembly Language File ;---------------------------------------------------------------------; 1 | int main() ;---------------------------------------------------------------------- ;*************************************************************** ;* FNAME: _main FR SIZE: 0 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 0 Auto, 0 SOE * ;*************************************************************** _main: ;---------------------------------------------------------------------- ; 3 | printf("Hello World\n"); ;---------------------------------------------------------------------- MOVL XAR4,#SL1 ; |3| LCR #_printf ; |3| ; call occurs [#_printf] ; |3| ;---------------------------------------------------------------------- ; 4 | return 0; ;---------------------------------------------------------------------- ;*************************************************************** ;* STRINGS * ;*************************************************************** .sect ".econst" SL1: .string "Hello World",10,0 ;*************************************************************** ;* UNDEFINED EXTERNAL REFERENCES * ;*************************************************************** .global _printf 52 Using the C/C++ Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Enabling Entry Hook and Exit Hook Functions 2.13 Enabling Entry Hook and Exit Hook Functions An entry hook is a routine that is called upon entry to each function in the program. An exit hook is a routine that is called upon exit of each function. Applications for hooks include debugging, trace, profiling, and stack overflow checking. Entry and exit hooks are enabled using the following options: --entry_hook[=name] --entry_parm{=name| address|none} --exit_hook[=name] --exit_parm{=name| address|none} Enables entry hooks. If specified, the hook function is called name. Otherwise, the default entry hook function name is __entry_hook. Specify the parameters to the hook function. The name parameter specifies that the name of the calling function is passed to the hook function as an argument. In this case the signature for the hook function is: void hook(const char *name); The address parameter specifies that the address of the calling function is passed to the hook function. In this case the signature for the hook function is: void hook(void (*addr)()); The none parameter specifies that the hook is called with no parameters. This is the default. In this case the signature for the hook function is: void hook(void); Enables exit hooks. If specified, the hook function is called name. Otherwise, the default exit hook function name is __exit_hook. Specify the parameters to the hook function. The name parameter specifies that the name of the calling function is passed to the hook function as an argument. In this case the signature for the hook function is: void hook(const char *name); The address parameter specifies that the address of the calling function is passed to the hook function. In this case the signature for the hook function is: void hook(void (*addr)()); The none parameter specifies that the hook is called with no parameters. This is the default. In this case the signature for the hook function is: void hook(void); The presence of the hook options creates an implicit declaration of the hook function with the given signature. If a declaration or definition of the hook function appears in the compilation unit compiled with the options, it must agree with the signatures listed above. In C++, the hooks are declared extern "C". Thus you can define them in C (or assembly) without being concerned with name mangling. Hooks can be declared inline, in which case the compiler tries to inline them using the same criteria as other inline functions. Entry hooks and exit hooks are independent. You can enable one but not the other, or both. The same function can be used as both the entry and exit hook. You must take care to avoid recursive calls to hook functions. The hook function should not call any function which itself has hook calls inserted. To help prevent this, hooks are not generated for inline functions, or for the hook functions themselves. You can use the --remove_hooks_when_inlining option to remove entry/exit hooks for functions that are auto-inlined by the optimizer. See Section 6.9.16 for information about the NO_HOOKS pragma. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Using the C/C++ Compiler 53 Chapter 3 SPRU514N – June 2017 Optimizing Your Code The compiler tools can perform many optimizations to improve the execution speed and reduce the size of C and C++ programs by simplifying loops, rearranging statements and expressions, and allocating variables into registers. This chapter describes how to invoke different levels of optimization and describes which optimizations are performed at each level. This chapter also describes how you can use the Interlist feature when performing optimization and how you can profile or debug optimized code. Topic ........................................................................................................................... Page 3.1 Invoking Optimization......................................................................................... 55 3.2 Performing File-Level Optimization (--opt_level=3 option) ....................................... 56 3.3 Program-Level Optimization (--program_level_compile and --opt_level=3 options)..... 56 3.4 Link-Time Optimization (--opt_level=4 Option) ....................................................... 59 3.5 Using Feedback Directed Optimization ................................................................. 59 3.6 Using Profile Information to Analyze Code Coverage ............................................. 64 3.7 Special Considerations When Using Optimization.................................................. 66 3.8 Automatic Inline Expansion (--auto_inline Option).................................................. 68 3.9 Using the Interlist Feature With Optimization......................................................... 68 3.10 Data Page (DP) Pointer Load Optimization ............................................................ 71 3.11 Debugging and Profiling Optimized Code.............................................................. 71 3.12 Controlling Code Size Versus Speed ................................................................... 72 3.13 Increasing Code-Size Optimizations (--opt_for_space Option) ................................. 73 3.14 Compiler Support for Generating DMAC Instructions ............................................. 74 3.15 What Kind of Optimization Is Being Performed? .................................................... 77 54 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Invoking Optimization 3.1 Invoking Optimization The C/C++ compiler is able to perform various optimizations. High-level optimizations are performed in the optimizer and low-level, target-specific optimizations occur in the code generator. Use high-level optimization levels, such as --opt_level=2 and --opt_level=3, to achieve optimal code. The easiest way to invoke optimization is to use the compiler program, specifying the --opt_level=n option on the compiler command line. You can use -On to alias the --opt_level option. The n denotes the level of optimization (0, 1, 2, 3, and 4), which controls the type and degree of optimization. • --opt_level=off or -Ooff – Performs no optimization • --opt_level=0 or -O0 – Performs control-flow-graph simplification – Allocates variables to registers – Performs loop rotation – Eliminates unused code – Simplifies expressions and statements – Expands calls to functions declared inline • --opt_level=1 or -O1 Performs all --opt_level=0 (-O0) optimizations, plus: – Performs local copy/constant propagation – Removes unused assignments – Eliminates local common expressions • --opt_level=2 or -O2 Performs all --opt_level=1 (-O1) optimizations, plus: – Performs loop optimizations – Eliminates global common subexpressions – Eliminates global unused assignments – Performs loop unrolling • --opt_level=3 or -O3 Performs all --opt_level=2 (-O2) optimizations, plus: – Removes all functions that are never called – Simplifies functions with return values that are never used – Inlines calls to small functions – Reorders function declarations; the called functions attributes are known when the caller is optimized – Propagates arguments into function bodies when all calls pass the same value in the same argument position – Identifies file-level variable characteristics If you use --opt_level=3 (-O3), see Section 3.2 and Section 3.3 for more information. • --opt_level=4 or -O4 Performs link-time optimization. See Section 3.4 for details. By default, debugging is enabled and the default optimization level is unaffected by the generation of debug information. However, the optimization level used is affected by whether or not the command line includes the -g (--symdebug:dwarf) option and the --opt_level option as shown in the following table: SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 55 Performing File-Level Optimization (--opt_level=3 option) Table 3-1. Interaction Between Debugging and Optimization Options www.ti.com Optimization no --opt_level --opt_level --opt_level=n no -g --opt_level=off --opt_level=2 optimized as specified -g --opt_level=off --opt_level=2 optimized as specified The levels of optimizations described above are performed by the stand-alone optimization pass. The code generator performs several additional optimizations, particularly processor-specific optimizations. It does so regardless of whether you invoke the optimizer. These optimizations are always enabled, although they are more effective when the optimizer is used. 3.2 Performing File-Level Optimization (--opt_level=3 option) The --opt_level=3 option (aliased as the -O3 option) instructs the compiler to perform file-level optimization. You can use the --opt_level=3 option alone to perform general file-level optimization, or you can combine it with other options to perform more specific optimizations. The options listed in Table 3-2 work with --opt_level=3 to perform the indicated optimization: Table 3-2. Options That You Can Use With --opt_level=3 If You ... Want to create an optimization information file Want to compile multiple source files Use this Option --gen_opt_level=n --program_level_compile See Section 3.2.1 Section 3.3 3.2.1 Creating an Optimization Information File (--gen_opt_info Option) When you invoke the compiler with the --opt_level=3 option, you can use the --gen_opt_info option to create an optimization information file that you can read. The number following the option denotes the level (0, 1, or 2). The resulting file has an .nfo extension. Use Table 3-3 to select the appropriate level to append to the option. Table 3-3. Selecting a Level for the --gen_opt_info Option If you... Do not want to produce an information file, but you used the --gen_opt_level=1 or --gen_opt_level=2 option in a command file or an environment variable. The --gen_opt_level=0 option restores the default behavior of the optimizer. Want to produce an optimization information file Want to produce a verbose optimization information file Use this option --gen_opt_info=0 --gen_opt_info=1 --gen_opt_info=2 3.3 Program-Level Optimization (--program_level_compile and --opt_level=3 options) You can specify program-level optimization by using the --program_level_compile option with the -opt_level=3 option (aliased as -O3). (If you use --opt_level=4 (-O4), the --program_level_compile option cannot be used, because link-time optimization provides the same optimization opportunities as program level optimization.) With program-level optimization, all of your source files are compiled into one intermediate file called a module. The module moves to the optimization and code generation passes of the compiler. Because the compiler can see the entire program, it performs several optimizations that are rarely applied during filelevel optimization: • If a particular argument in a function always has the same value, the compiler replaces the argument with the value and passes the value instead of the argument. • If a return value of a function is never used, the compiler deletes the return code in the function. • If a function is not called directly or indirectly by main(), the compiler removes the function. 56 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Program-Level Optimization (--program_level_compile and --opt_level=3 options) The --program_level_compile option requires use of --opt_level=3 or higher in order to perform these optimizations. To see which program-level optimizations the compiler is applying, use the --gen_opt_level=2 option to generate an information file. See Section 3.2.1 for more information. In Code Composer Studio, when the --program_level_compile option is used, C and C++ files that have the same options are compiled together. However, if any file has a file-specific option that is not selected as a project-wide option, that file is compiled separately. For example, if every C and C++ file in your project has a different set of file-specific options, each is compiled separately, even though program-level optimization has been specified. To compile all C and C++ files together, make sure the files do not have file-specific options. Be aware that compiling C and C++ files together may not be safe if previously you used a file-specific option. Compiling Files With the --program_level_compile and --keep_asm Options NOTE: If you compile all files with the --program_level_compile and --keep_asm options, the compiler produces only one .asm file, not one for each corresponding source file. 3.3.1 Controlling Program-Level Optimization (--call_assumptions Option) You can control program-level optimization, which you invoke with --program_level_compile --opt_level=3, by using the --call_assumptions option. Specifically, the --call_assumptions option indicates if functions in other modules can call a module's external functions or modify a module's external variables. The number following --call_assumptions indicates the level you set for the module that you are allowing to be called or modified. The --opt_level=3 option combines this information with its own file-level analysis to decide whether to treat this module's external function and variable declarations as if they had been declared static. Use Table 3-4 to select the appropriate level to append to the --call_assumptions option. Table 3-4. Selecting a Level for the --call_assumptions Option If Your Module … Has functions that are called from other modules and global variables that are modified in other modules Does not have functions that are called by other modules but has global variables that are modified in other modules Does not have functions that are called by other modules or global variables that are modified in other modules Has functions that are called from other modules but does not have global variables that are modified in other modules Use this Option --call_assumptions=0 --call_assumptions=1 --call_assumptions=2 --call_assumptions=3 In certain circumstances, the compiler reverts to a different --call_assumptions level from the one you specified, or it might disable program-level optimization altogether. Table 3-5 lists the combinations of -call_assumptions levels and conditions that cause the compiler to revert to other --call_assumptions levels. Table 3-5. Special Considerations When Using the --call_assumptions Option If --call_assumptions is... Under these Conditions... Not specified The --opt_level=3 optimization level was specified Not specified The compiler sees calls to outside functions under the -opt_level=3 optimization level Not specified Main is not defined --call_assumptions=1 or --call_assumptions=2 No function has main defined as an entry point, and no interrupt functions are defined, and no functions are identified by the FUNC_EXT_CALLED pragma --call_assumptions=1 or --call_assumptions=2 A main function is defined, or, an interrupt function is defined, or a function is identified by the FUNC_EXT_CALLED pragma --call_assumptions=3 Any condition Then the --call_assumptions Level... Defaults to --call_assumptions=2 Reverts to --call_assumptions=0 Reverts to --call_assumptions=0 Reverts to --call_assumptions=0 Remains --call_assumptions=1 or --call_assumptions=2 Remains --call_assumptions=3 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 57 Program-Level Optimization (--program_level_compile and --opt_level=3 options) www.ti.com In some situations when you use --program_level_compile and --opt_level=3, you must use a -call_assumptions option or the FUNC_EXT_CALLED pragma. See Section 3.3.2 for information about these situations. 3.3.2 Optimization Considerations When Mixing C/C++ and Assembly If you have any assembly functions in your program, you need to exercise caution when using the -program_level_compile option. The compiler recognizes only the C/C++ source code and not any assembly code that might be present. Because the compiler does not recognize the assembly code calls and variable modifications to C/C++ functions, the --program_level_compile option optimizes out those C/C++ functions. To keep these functions, place the FUNC_EXT_CALLED pragma (see Section 6.9.12) before any declaration or reference to a function that you want to keep. Another approach you can take when you use assembly functions in your program is to use the -call_assumptions=n option with the --program_level_compile and --opt_level=3 options. See Section 3.3.1 for information about the --call_assumptions=n option. In general, you achieve the best results through judicious use of the FUNC_EXT_CALLED pragma in combination with --program_level_compile --opt_level=3 and --call_assumptions=1 or -call_assumptions=2. If any of the following situations apply to your application, use the suggested solution: Situation — Your application consists of C/C++ source code that calls assembly functions. Those assembly functions do not call any C/C++ functions or modify any C/C++ variables. Solution — Compile with --program_level_compile --opt_level=3 --call_assumptions=2 to tell the compiler that outside functions do not call C/C++ functions or modify C/C++ variables. If you compile with the --program_level_compile --opt_level=3 options only, the compiler reverts from the default optimization level (--call_assumptions=2) to --call_assumptions=0. The compiler uses --call_assumptions=0, because it presumes that the calls to the assembly language functions that have a definition in C/C++ may call other C/C++ functions or modify C/C++ variables. Situation — Your application consists of C/C++ source code that calls assembly functions. The assembly language functions do not call C/C++ functions, but they modify C/C++ variables. Solution — Try both of these solutions and choose the one that works best with your code: • Compile with --program_level_compile --opt_level=3 --call_assumptions=1. • Add the volatile keyword to those variables that may be modified by the assembly functions and compile with --program_level_compile --opt_level=3 --call_assumptions=2. Situation — Your application consists of C/C++ source code and assembly source code. The assembly functions are interrupt service routines that call C/C++ functions; the C/C++ functions that the assembly functions call are never called from C/C++. These C/C++ functions act like main: they function as entry points into C/C++. Solution — Add the volatile keyword to the C/C++ variables that may be modified by the interrupts. Then, you can optimize your code in one of these ways: • You achieve the best optimization by applying the FUNC_EXT_CALLED pragma to all of the entry-point functions called from the assembly language interrupts, and then compiling with -program_level_compile --opt_level=3 --call_assumptions=2. Be sure that you use the pragma with all of the entry-point functions. If you do not, the compiler might remove the entry-point functions that are not preceded by the FUNC_EXT_CALLED pragma. • Compile with --program_level_compile --opt_level=3 --call_assumptions=3. Because you do not use the FUNC_EXT_CALLED pragma, you must use the --call_assumptions=3 option, which is less aggressive than the --call_assumptions=2 option, and your optimization may not be as effective. Keep in mind that if you use --program_level_compile --opt_level=3 without additional options, the compiler removes the C functions that the assembly functions call. Use the FUNC_EXT_CALLED pragma to keep these functions. 58 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Link-Time Optimization (--opt_level=4 Option) 3.4 Link-Time Optimization (--opt_level=4 Option) Link-time optimization is an optimization mode that allows the compiler to have visibility of the entire program. The optimization occurs at link-time instead of compile-time like other optimization levels. Link-time optimization is invoked by using the --opt_level=4 option. This option must be used in both the compilation and linking steps. At compile time, the compiler embeds an intermediate representation of the file being compiled into the resulting object file. At link-time this representation is extracted from every object file which contains it, and is used to optimize the entire program. If you use --opt_level=4 (-O4), the --program_level_compile option cannot also be used, because link-time optimization provides the same optimization opportunities as program level optimization (Section 3.3). Link-time optimization provides the following benefits: • Each source file can be compiled separately. One issue with program-level compilation is that it requires all source files to be passed to the compiler at one time. This often requires significant modification of a customer's build process. With link-time optimization, all files can be compiled separately. • References to C/C++ symbols from assembly are handled automatically. When doing program-level compilation, the compiler has no knowledge of whether a symbol is referenced externally. When performing link-time optimization during a final link, the linker can determine which symbols are referenced externally and prevent eliminating them during optimization. • Third party object files can participate in optimization. If a third party vendor provides object files that were compiled with the --opt_level=4 option, those files participate in optimization along with usergenerated files. This includes object files supplied as part of the TI run-time support. Object files that were not compiled with –opt_level=4 can still be used in a link that is performing link-time optimization. Those files that were not compiled with –opt_level=4 do not participate in the optimization. • Source files can be compiled with different option sets. With program-level compilation, all source files must be compiled with the same option set. With link-time optimization files can be compiled with different options. If the compiler determines that two options are incompatible, it issues an error. 3.4.1 Option Handling When performing link-time optimization, source files can be compiled with different options. When possible, the options that were used during compilation are used during link-time optimization. For options which apply at the program level, --auto_inline for instance, the options used to compile the main function are used. If main is not included in link-time optimization, the option set used for the first object file specified on the command line is used. Some options, --opt_for_speed for instance, can affect a wide range of optimizations. For these options, the program-level behavior is derived from main, and the local optimizations are obtained from the original option set. Some options are incompatible when performing link-time optimization. These are usually options which conflict on the command line as well, but can also be options that cannot be handled during link-time optimization. 3.4.2 Incompatible Types During a normal link, the linker does not check to make sure that each symbol was declared with the same type in different files. This is not necessary during a normal link. When performing link-time optimization, however, the linker must ensure that all symbols are declared with compatible types in different source files. If a symbol is found which has incompatible types, an error is issued. The rules for compatible types are derived from the C and C++ standards. 3.5 Using Feedback Directed Optimization Feedback directed optimization provides a method for finding frequently executed paths in an application using compiler-based instrumentation. This information is fed back to the compiler and is used to perform optimizations. This information is also used to provide you with information about application behavior. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 59 Using Feedback Directed Optimization www.ti.com 3.5.1 Feedback Directed Optimization Feedback directed optimization uses run-time feedback to identify and optimize frequently executed program paths. Feedback directed optimization is a two-phase process. 3.5.1.1 Phase 1 -- Collect Program Profile Information In this phase the compiler is invoked with the option --gen_profile_info, which instructs the compiler to add instrumentation code to collect profile information. The compiler inserts a minimal amount of instrumentation code to determine control flow frequencies. Memory is allocated to store counter information. The instrumented application program is executed on the target using representative input data sets. The input data sets should correlate closely with the way the program is expected to be used in the end product environment. When the program completes, a run-time-support function writes the collected information into a profile data file called a PDAT file. Multiple executions of the program using different input data sets can be performed and in such cases, the run-time-support function appends the collected information into the PDAT file. The resulting PDAT file is post-processed using a tool called the Profile Data Decoder or pdd2000. The pdd2000 tool consolidates multiple data sets and formats the data into a feedback file (PRF file, see Section 3.5.2) for consumption by phase 2 of feedback directed optimization. 3.5.1.2 Phase 2 -- Use Application Profile Information for Optimization In this phase, the compiler is invoked with the --use_profile_info=file.prf option, which reads the specified PRF file generated in phase 1. In phase 2, optimization decisions are made using the data generated during phase 1. The profile feedback file is used to guide program optimization. The compiler optimizes frequently executed program paths more aggressively. The compiler uses data in the profile feedback file to guide certain optimizations of frequently executed program paths. 60 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com 3.5.1.3 Generating and Using Profile Information There are two options that control feedback directed optimization: Using Feedback Directed Optimization --gen_profile_info --use_profile_info tells the compiler to add instrumentation code to collect profile information. When the program executes the run-time-support exit() function, the profile data is written to a PDAT file. This option applies to all the C/C++ source files being compiled on the command-line. If the environment variable TI_PROFDATA on the host is set, the data is written into the specified file. Otherwise, it uses the default filename: pprofout.pdat. The full pathname of the PDAT file (including the directory name) can be specified using the TI_PROFDATA host environment variable. By default, the RTS profile data output routine uses the C I/O mechanism to write data to the PDAT file. You can install a device handler for the PPHNDL device to re-direct the profile data to a custom device driver routine. For example, this could be used to send the profile data to a device that does not use a file system. Feedback directed optimization requires you to turn on at least some debug information when using the --gen_profile_info option. This enables the compiler to output debug information that allows pdd2000 to correlate compiled functions and their associated profile data. specifies the profile information file(s) to use for performing phase 2 of feedback directed optimization. More than one profile information file can be specified on the command line; the compiler uses all input data from multiple information files. The syntax for the option is: --use_profile_info==file1[, file2, ..., filen] If no filename is specified, the compiler looks for a file named pprofout.prf in the directory where the compiler in invoked. 3.5.1.4 Example Use of Feedback Directed Optimization These steps illustrate the creation and use of feedback directed optimization. 1. Generate profile information. cl2000 --opt_level=2 --gen_profile_info foo.c --run_linker --output_file=foo.out --library=lnk.cmd --library=rts2800_ml.lib 2. Execute the application. The execution of the application creates a PDAT file named pprofout.pdat in the current (host) directory. The application can be run on on target hardware connected to a host machine. 3. Process the profile data. After running the application with multiple data-sets, run pdd2000 on the PDAT files to create a profile information (PRF) file to be used with --use_profile_info. pdd2000 -e foo.out -o pprofout.prf pprofout.pdat 4. Re-compile using the profile feedback file. cl2000 --opt_level=2 --use_profile_info=pprofout.prf foo.c --run_linker --output_file=foo.out --library=lnk.cmd --library=rts2800_ml.lib 3.5.1.5 The .ppdata Section The profile information collected in phase 1 is stored in the .ppdata section, which must be allocated into target memory. The .ppdata section contains profiler counters for all functions compiled with -gen_profile_info. The default lnk.cmd file in has directives to place the .ppdata section in data memory. If the link command file has no section directive for allocating .ppdata section, the link step places the .ppdata section in a writable memory range. The .ppdata section must be allocated memory in multiples of 32 bytes. Please refer to the linker command file in the distribution for example usage. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 61 Using Feedback Directed Optimization www.ti.com 3.5.1.6 Feedback Directed Optimization and Code Size Tune Feedback directed optimization is different from the Code Size Tune feature in Code Composer Studio (CCS). The code size tune feature uses CCS profiling to select specific compilation options for each function in order to minimize code size while still maintaining a specific performance point. Code size tune is coarse-grained, since it is selecting an option set for the whole function. Feedback directed optimization selects different optimization goals along specific regions within a function. 3.5.1.7 Instrumented Program Execution Overhead During profile collection, the execution time of the application may increase. The amount of increase depends on the size of the application and the number of files in the application compiled for profiling. The profiling counters increase the code and data size of the application. Consider using the -opt_for_space (-ms) code size options when using profiling to mitigate the code size increase. This has no effect on the accuracy of the profile data being collected. Since profiling only counts execution frequency and not cycle counts, code size optimization flags do not affect profiler measurements. 3.5.1.8 Invalid Profile Data When recompiling with --use_profile_info, the profile information is invalid in the following cases: • The source file name changed between the generation of profile information (gen-profile) and the use of the profile information (use-profile). • The source code was modified since gen-profile. In this case, profile information is invalid for the modified functions. • Certain compiler options used with gen-profile are different from those with used with use-profile. In particular, options that affect parser behavior could invalidate profile data during use-profile. In general, using different optimization options during use-profile should not affect the validity of profile data. 3.5.2 Profile Data Decoder The code generation tools include a tool called the Profile Data Decoder or pdd2000, which is used for post processing profile data (PDAT) files. The pdd2000 tool generates a profile feedback (PRF) file. See Section 3.5.1 for a discussion of where pdd2000 fits in the profiling flow. The pdd2000 tool is invoked with this syntax: pdd2000 -e exec.out -o application.prf filename .pdat -a -e exec.out -o application.prf filename .pdat Computes the average of the data values in the data sets instead of accumulating data values Specifies exec.out is the name of the application executable. Specifies application.prf is the formatted profile feedback file that is used as the argument to --use_profile_info during recompilation. If no output file is specified, the default output filename is pprofout.prf. Is the name of the profile data file generated by the run-time-support function. This is the default name and it can be overridden by using the host environment variable TI_PROFDATA. 62 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Using Feedback Directed Optimization The run-time-support function and pdd2000 append to their respective output files and do not overwrite them. This enables collection of data sets from multiple runs of the application. Profile Data Decoder Requirements NOTE: Your application must be compiled with at least DWARF debug support to enable feedback directed optimization. When compiling for feedback directed optimization, the pdd2000 tool relies on basic debug information about each function in generating the formatted .prf file. The pprofout.pdat file generated by the run-time support is a raw data file of a fixed format understood only by pdd2000. You should not modify this file in any way. 3.5.3 Feedback Directed Optimization API There are two user interfaces to the profiler mechanism. You can start and stop profiling in your application by using the following run-time-support calls. • _TI_start_pprof_collection() This interface informs the run-time support that you wish to start profiling collection from this point on and causes the run-time support to clear all profiling counters in the application (that is, discard old counter values). • _TI_stop_pprof_collection() This interface directs the run-time support to stop profiling collection and output profiling data into the output file (into the default file or one specified by the TI_PROFDATA host environment variable). The run-time support also disables any further output of profile data into the output file during exit(), unless you call _TI_start_pprof_collection() again. 3.5.4 Feedback Directed Optimization Summary Options --gen_profile_info --use_profile_info=file.prf --analyze=codecov --analyze_only Adds instrumentation to the compiled code. Execution of the code results in profile data being emitted to a PDAT file. Uses profile information for optimization and/or generating code coverage information. Generates a code coverage information file and continues with profile-based compilation. Must be used with --use_profile_info. Generates only a code coverage information file. Must be used with -use_profile_info. You must specify both --analyze=codecov and -analyze_only to do code coverage analysis of the instrumented application. Host Environment Variables TI_PROFDATA TI_COVDIR TI_COVDATA Writes profile data into the specified file Creates code coverage files in the specified directory Writes code coverage data into the specified file API _TI_start_pprof_collection() Clears the profile counters to file _TI_stop_pprof_collection() Writes out all profile counters to file PPHDNL Device driver handle for low-level C I/O based driver for writing out profile data from a target program. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 63 Using Profile Information to Analyze Code Coverage Files Created www.ti.com *.pdat *.prf Profile data file, which is created by executing an instrumented program and used as input to the profile data decoder Profiling feedback file, which is created by the profile data decoder and used as input to the re-compilation step 3.6 Using Profile Information to Analyze Code Coverage You can use the analysis information from the Profile Data Decoder to analyze code coverage. 3.6.1 Code Coverage The information collected during feedback directed optimization can be used for generating code coverage reports. As with feedback directed optimization, the program must be compiled with the --gen_profile_info option. Code coverage conveys the execution count of each line of source code in the file being compiled, using data collected during profiling. 3.6.1.1 Phase1 -- Collect Program Profile Information In this phase the compiler is invoked with the option --gen_profile_info, which instructs the compiler to add instrumentation code to collect profile information. The compiler inserts a minimal amount of instrumentation code to determine control flow frequencies. Memory is allocated to store counter information. The instrumented application program is executed on the target using representative input data sets. The input data sets should correlate closely with the way the program is expected to be used in the end product environment. When the program completes, a run-time-support function writes the collected information into a profile data file called a PDAT file. Multiple executions of the program using different input data sets can be performed and in such cases, the run-time-support function appends the collected information into the PDAT file. The resulting PDAT file is post-processed using a tool called the Profile Data Decoder or pdd2000. The pdd2000 tool consolidates multiple data sets and formats the data into a feedback file (PRF file, see Section 3.5.2) for consumption by phase 2 of feedback directed optimization. 3.6.1.2 Phase 2 -- Generate Code Coverage Reports In this phase, the compiler is invoked with the --use_profile_info=file.prf option, which indicates that the compiler should read the specified PRF file generated in phase 1. The application must also be compiled with either the --codecov or --onlycodecov option; the compiler generates a code-coverage info file. The -codecov option directs the compiler to continue compilation after generating code-coverage information, while the --onlycodecov option stops the compiler after generating code-coverage data. For example: cl2000 --opt_level=2 --use_profile_info=pprofout.prf --onlycodecov foo.c You can specify two environment variables to control the destination of the code-coverage information file. • The TI_COVDIR environment variable specifies the directory where the code-coverage file should be generated. The default is the directory where the compiler is invoked. • The TI_COVDATA environment variable specifies the name of the code-coverage data file generated by the compiler. the default is filename.csv where filename is the base-name of the file being compiled. For example, if foo.c is being compiled, the default code-coverage data file name is foo.csv. If the code-coverage data file already exists, the compiler appends the new dataset at the end of the file. Code-coverage data is a comma-separated list of data items that can be conveniently handled by dataprocessing tools and scripting languages. The following is the format of code-coverage data: "filename-with-full-path","funcname",line#,column#,exec-frequency,"comments" 64 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Using Profile Information to Analyze Code Coverage "filename-with-full-path" "funcname" line# column# exec-frequency "comments" Full pathname of the file corresponding to the entry Name of the function Line number of the source line corresponding to frequency data Column number of the source line Execution frequency of the line Intermediate-level representation of the source-code generated by the parser The full filename, function name, and comments appear within quotation marks ("). For example: "/some_dir/zlib/c2000/deflate.c","_deflateInit2_",216,5,1,"( strm->zalloc )" Other tools, such as a spreadsheet program, can be used to format and view the code coverage data. 3.6.2 Related Features and Capabilities The code generation tools provide some features and capabilities that can be used in conjunction with code coverage analysis. The following is a summary: 3.6.2.1 Path Profiler The code generation tools include a path profiling utility, pprof2000, that is run from the compiler, cl2000. The pprof2000 utility is invoked by the compiler when the --gen_profile or the --use_profile command is used from the compiler command line: cl2000 --gen_profile ... file.c cl2000 --use_profile ... file.c For further information about profile-based optimization and a more detailed description of the profiling infrastructure, see Section 3.5. 3.6.2.2 Analysis Options The path profiling utility, pprof2000, appends code coverage information to existing CSV (comma separated values) files that contain the same type of analysis information. The utility checks to make sure that an existing CSV file contains analysis information that is consistent with the type of analysis information it is being asked to generate. Attempts to mix code coverage and other analysis information in the same output CSV file will be detected, and pprof2000 will emit a fatal error and abort. --analyze=codecov --analyze_only Instructs the compiler to generate code coverage analysis information. This option replaces the previous --codecov option. Halts compilation after generation of analysis information is completed. 3.6.2.3 Environment Variables To assist with the management of output CSV analysis files, pprof2000 supports this environment variable: TI_ANALYSIS_DIR Specifies the directory in which the output analysis file will be generated. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 65 Special Considerations When Using Optimization www.ti.com 3.7 Special Considerations When Using Optimization The compiler is designed to improve your ANSI/ISO-conforming C and C++ programs while maintaining their correctness. However, when you write code for optimization, you should note the special considerations discussed in the following sections to ensure that your program performs as you intend. 3.7.1 Use Caution With asm Statements in Optimized Code You must be extremely careful when using asm (inline assembly) statements in optimized code. The compiler rearranges code segments, uses registers freely, and can completely remove variables or expressions. Although the compiler never optimizes out an asm statement (except when it is unreachable), the surrounding environment where the assembly code is inserted can differ significantly from the original C/C++ source code. It is usually safe to use asm statements to manipulate hardware controls such as interrupt masks, but asm statements that attempt to interface with the C/C++ environment or access C/C++ variables can have unexpected results. After compilation, check the assembly output to make sure your asm statements are correct and maintain the integrity of the program. 3.7.2 Use the Volatile Keyword for Necessary Memory Accesses The compiler analyzes data flow to avoid memory accesses whenever possible. If you have code that depends on memory accesses exactly as written in the C/C++ code, you must use the volatile keyword to identify these accesses. The compiler does not optimize out any references to volatile variables. In the following example, the loop waits for a location to be read as 0xFF: unsigned int *ctrl; while (*ctrl !=0xFF); In this example, *ctrl is a loop-invariant expression, so the loop is optimized down to a single memory read. To correct this, declare ctrl as: volatile unsigned int *ctrl 3.7.2.1 Use Caution When Accessing Aliased Variables Aliasing occurs when a single object can be accessed in more than one way, such as when two pointers point to the same object or when a pointer points to a named object. Aliasing can disrupt optimization because any indirect reference can refer to another object. The compiler analyzes the code to determine where aliasing can and cannot occur, then optimizes as much as possible while still preserving the correctness of the program. The compiler behaves conservatively. The compiler assumes that if the address of a local variable is passed to a function, the function might change the local by writing through the pointer, but that it will not make its address available for use elsewhere after returning. For example, the called function cannot assign the local’s address to a global variable or return it. In cases where this assumption is invalid, use the -ma compiler option to force the compiler to assume worst-case aliasing. In worst-case aliasing, any indirect reference (that is, using a pointer) can refer to such a variable. 66 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Special Considerations When Using Optimization 3.7.2.2 Use the --aliased_variables Option to Indicate That the Following Technique Is Used The compiler, when invoked with optimization, assumes that any variable whose address is passed as an argument to a function will not be subsequently modified by an alias set up in the called function. Examples include: • Returning the address from a function • Assigning the address to a global If you use aliases like this in your code, you must use the --aliased_variables option when you are optimizing your code. For example, if your code is similar to this, use the --aliased_variables option: int *glob_ptr; g() { int x = 1; int *p = f(&x); *p = 5; /* p aliases x */ *glob_ptr = 10; /* glob_ptr aliases x */ h(x); } int *f(int *arg) { glob_ptr = arg; return arg; } 3.7.2.3 On FPU Targets Only: Use restrict Keyword to Indicate That Pointers Are Not Aliased On FPU targets, with --opt_level=2, the optimizer performs dependency analysis. To help the compiler determine memory dependencies, you can qualify a pointer, reference, or array with the restrict keyword. The restrict keyword is a type qualifier that can be applied to pointers, references, and arrays. Its use represents a guarantee by the programmer that within the scope of the pointer declaration the object pointed to can be accessed only by that pointer. Any violation of this guarantee renders the program undefined. This practice helps the compiler optimize certain sections of code because aliasing information can be more easily determined. This can improve performance and code size, as more FPU operations can be parallelized. As shown in Example 3-1 and Example 3-2 you can use the restrict keyword to tell the compiler that a and b never point to the same object in foo. Furthermore, the compiler is assured that the objects pointed to by a and b do not overlap in memory. Example 3-1. Use of the restrict Type Qualifier With Pointers void foo(float * restrict a, float * restrict b) { /* foo's code here */ } Example 3-2. Use of the restrict Type Qualifier With Pointers void foo(float c[restrict], float d[restrict]) { /* foo's code here */ } SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 67 Automatic Inline Expansion (--auto_inline Option) www.ti.com 3.8 Automatic Inline Expansion (--auto_inline Option) When optimizing with the --opt_level=3 option (aliased as -O3), the compiler automatically inlines small functions. A command-line option, --auto_inline=size, specifies the size threshold. Any function larger than the size threshold is not automatically inlined. You can use the --auto_inline=size option in the following ways: • If you set the size parameter to 0 (--auto_inline=0), automatic inline expansion is disabled. • If you set the size parameter to a nonzero integer, the compiler uses this size threshold as a limit to the size of the functions it automatically inlines. The compiler multiplies the number of times the function is inlined (plus 1 if the function is externally visible and its declaration cannot be safely removed) by the size of the function. The compiler inlines the function only if the result is less than the size parameter. The compiler measures the size of a function in arbitrary units; however, the optimizer information file (created with the -gen_opt_level=1 or --gen_opt_level=2 option) reports the size of each function in the same units that the -auto_inline option uses. The --auto_inline=size option controls only the inlining of functions that are not explicitly declared as inline. If you do not use the --auto_inline=size option, the compiler inlines very small functions. Optimization Level 3 and Inlining NOTE: In order to turn on automatic inlining, you must use the --opt_level=3 option. If you desire the --opt_level=3 optimizations, but not automatic inlining, use --auto_inline=0 with the -opt_level=3 option. Inlining and Code Size NOTE: Expanding functions inline increases code size, especially inlining a function that is called in a number of places. Function inlining is optimal for functions that are called only from a small number of places and for small functions. To prevent increases in code size because of inlining, use the --auto_inline=0 option. This option causes the compiler to inline intrinsics only. 3.9 Using the Interlist Feature With Optimization You control the output of the interlist feature when compiling with optimization (the --opt_level=n or -On option) with the --optimizer_interlist and --c_src_interlist options. • The --optimizer_interlist option interlists compiler comments with assembly source statements. • The --c_src_interlist and --optimizer_interlist options together interlist the compiler comments and the original C/C++ source with the assembly code. When you use the --optimizer_interlist option with optimization, the interlist feature does not run as a separate pass. Instead, the compiler inserts comments into the code, indicating how the compiler has rearranged and optimized the code. These comments appear in the assembly language file as comments starting with ;**. The C/C++ source code is not interlisted, unless you use the --c_src_interlist option also. The interlist feature can affect optimized code because it might prevent some optimization from crossing C/C++ statement boundaries. Optimization makes normal source interlisting impractical, because the compiler extensively rearranges your program. Therefore, when you use the --optimizer_interlist option, the compiler writes reconstructed C/C++ statements. Example 3-4 shows a function that has been compiled with optimization (--opt_level=2) and the -optimizer_interlist option. The assembly file contains compiler comments interlisted with assembly code. Impact on Performance and Code Size NOTE: The --c_src_interlist option can have a negative effect on performance and code size. 68 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Using the Interlist Feature With Optimization When you use the --c_src_interlist and --optimizer_interlist options with optimization, the compiler inserts its comments and the interlist feature runs before the assembler, merging the original C/C++ source into the assembly file. Example 3-5 shows the function from Example 3-4 compiled with the optimization (--opt_level=2) and the -c_src_interlist and --optimizer_interlist options. The assembly file contains compiler comments and C source interlisted with assembly code. Example 3‑3. C Code for Interlist Illustration int copy (char *str, const char *s, int n) { int i; for (i = 0; i < n; i ++) *str++ = *s++; } Example 3‑4. The Function From Example 3-3 Compiled With the -O2 and --optimizer_interlist Options ;*************************************************************** ;* FNAME: _copy FR SIZE: 0 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 0 Auto, 0 SOE * ;*************************************************************** _copy: ;*** 6 ;*** :*** L1: ;*** ;*** 7 ;*** 7 ;*** ;*** L2: ----------------------- CMPB AL,#0 B L2,LEQ ; branch occurs ; |6| ----------------------- ----------------------- ADDB AL,#-1 MOVZ AR6,AL if ( n <= 0 ) goto g4; ; |6| ; |6| #pragma MUST_ITERATE(1, 4294967295, 1) L$1 = n-1; -----------------------g3: ----------------------- *str++ = *s++; ----------------------- if ( (--L$1) != (-1) ) goto g3; MOV AL,*XAR5++ ; |7| MOV *XAR4++,AL ; |7| BANZ L1,AR6-- ; branch occurs ; |7| -----------------------g4: ----------------------- return; LRETR ; return occurs SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 69 Using the Interlist Feature With Optimization www.ti.com Example 3‑5. The Function From Example 3-3 Compiled with the --opt_level=2, --optimizer_interlist, and -c_src_interlist Options ;---------------------------------------------------------------------; 2 | int copy (char *str, const char *s, int n) ;---------------------------------------------------------------------- ;*************************************************************** ;* FNAME: _copy FR SIZE: 0 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 0 Auto, 0 SOE * ;*************************************************************** _copy ;* AR4 assigned to _str ;* AR5 assigned to _s ;* AL assigned to _n ;* AL assigned to _n ;* AR5 assigned to _s ;* AR4 assigned to _str ;* AR6 assigned to L$1 ;*** 6 ----------------------- if ( n <= 0 ) goto g4; ;---------------------------------------------------------------------- ; 4 | int i; ;---------------------------------------------------------------------- ;---------------------------------------------------------------------- ; 6 | for (i = 0; i < n; i++) ;---------------------------------------------------------------------- CMPB AL,#0 ; |6| B L2,LEQ ; |6| ; branch occurs ; |6| ;*** ----------------------- #pragma MUST_ITERATE(1, 4294967295, 1) :*** ----------------------- L$1 = n-1; ADDB AL,#-1 MOVZ AR6,AL NOP L1: ;*** 7 ----------------------- *str++ = *s++; ;*** 7 ----------------------- if ( (--L1) != (-1) ) goto g3; ;---------------------------------------------------------------------- ; 7 | *str++ = *s++; ;---------------------------------------------------------------------- MOV AL,*XAR5++ ; |7| MOV *XAR4++,AL ; |7| BANZ L1,AR6-- ; branch occurs ; |7| ;*** -----------------------g4: ;*** ----------------------- return; L2: LRETR ; return occurs 70 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Data Page (DP) Pointer Load Optimization 3.10 Data Page (DP) Pointer Load Optimization By default, the compiler performs data page load optimization. The compiler blocks global variables by either forcing them to fit within a 64-word data page or aligning the start to a page boundary. Data page blocking allows the compiler to use the direct addressing mode more often. As a result, it minimizes the need for DP load instructions when accessing global variables known to be stored on a single data page. However, the data page blocking optimization can result in unused holes within data sections. You can weigh your application's need for code size and speed optimization against its need for data size optimization. If you determine that you want to disable data page load optimization, use the -disable_dp_load_opt compiler option. Another technique for managing how global variables are organized in memory include grouping global variables that should be stored together in a structure. The DATA_SECTION and SET_DATA_SECTION pragmas can also be used to manage data pages. For more details about and examples of data blocking, see the Tools Insider blog in TI's E2E community. 3.11 Debugging and Profiling Optimized Code Generating symbolic debugging information no longer affects the ability to optimize code. The same executable code is generated regardless of whether generation of debug information is turned on or off. For this reason, debug information is now generated by default. You do not need to specify the -g option in order to debug your application. If you do not specify the -g option and allow the default generation of debug information to be used, the default level of optimization is used unless you specify some other optimization level. The --symdebug:dwarf option no longer disables optimization, because generation of debug information no longer impacts optimization. If you specify the -g option explicitly but do not specify an optimization level, no optimization is performed. This is because while generating debug information does not affect the ability to optimize code, optimizing code does make it more difficult to debug code. At higher levels of optimization, the compiler's extensive rearrangement of code and the many-to-many allocation of variables to registers often make it difficult to correlate source code with object code for debugging purposes. It is recommended that you perform debugging using the lowest level of optimization possible. If you specify an --opt_level (aliased as -O) option, that optimization level is used no matter what type of debugging information you enabled. The optimization level used if you do not specify the level on the command line is affected by whether or not the command line includes the -g option and the --opt_level option as shown in the following table: Table 3-6. Interaction Between Debugging and Optimization Options Optimization no --opt_level --opt_level --opt_level=n no -g --opt_level=off --opt_level=2 optimized as specified -g --opt_level=off --opt_level=2 optimized as specified Debug information increases the size of object files, but it does not affect the size of code or data on the target. If object file size is a concern and debugging is not needed, use --symdebug:none to disable the generation of debug information. 3.11.1 Profiling Optimized Code To profile optimized code, use optimization (--opt_level=0 through --opt_level=3). SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 71 Controlling Code Size Versus Speed www.ti.com 3.12 Controlling Code Size Versus Speed The latest mechanism for controlling the goal of optimizations in the compiler is represented by the -opt_for_speed=num option. The num denotes the level of optimization (0-5), which controls the type and degree of code size or code speed optimization: • --opt_for_speed=0 Enables optimizations geared towards improving the code size with a high risk of worsening or impacting performance. • --opt_for_speed=1 Enables optimizations geared towards improving the code size with a medium risk of worsening or impacting performance. • --opt_for_speed=2 Enables optimizations geared towards improving the code size with a low risk of worsening or impacting performance. • --opt_for_speed=3 Enables optimizations geared towards improving the code performance/speed with a low risk of worsening or impacting code size. • --opt_for_speed=4 Enables optimizations geared towards improving the code performance/speed with a medium risk of worsening or impacting code size. • --opt_for_speed=5 Enables optimizations geared towards improving the code performance/speed with a high risk of worsening or impacting code size. If you specify the --opt_for_speed option without a parameter, the default setting is --opt_for_speed=4. If you do not specify the --opt_for_speed option, the default setting is 2 The initial mechanism for controlling code space, the --opt_for_space option, has the following equivalences with the --opt_for_speed option: --opt_for_space none =0 =1 =2 =3 --opt_for_speed =4 =3 =2 =1 =0 A fast branch (BF) instruction is generated by default when the --opt_for_speed option is used. When -opt_for_speed is not used, the compiler generates a BF instruction only when the condition code is one of NEQ, EQ, NTC and TC. The reason is that BF with these condition codes can be optimized to SBF. There is a code size penalty to use BF instruction when the condition code is NOT one of NEQ, EQ, NTC and TC. (Fast branch instructions are also generated for functions with the ramfunc function attribute.) The --no_fast_branch option is deprecated and has no effect. 72 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Increasing Code-Size Optimizations (--opt_for_space Option) 3.13 Increasing Code-Size Optimizations (--opt_for_space Option) The --opt_for_space option increases the level of code-size optimizations performed by the compiler. These optimizations are done at the expense of performance. The optimizations include procedural abstraction where common blocks of code are replaced with function calls. For example, prolog and epilog code, certain intrinsics, and other common code sequences, can be replaced with calls to functions that are defined in the run-time library. It is necessary to link with the supplied run-time library when using the -opt_for_space option. It is not necessary to use optimization to invoke the --opt_for_space option. To illustrate how the --opt_for_space option works, the following describes how prolog and epilog code can be replaced. This code is changed to function calls depending on the number of SOE registers, the size of the frame, and whether a frame pointer is used. These functions are defined in each file with the -opt_for_space option, as shown below: _prolog_c28x_1 _prolog_c28x_2 _prolog_c28x_3 _epilog_c28x_1 _epilog_c28x_2 Example 3-6 provides an example of C code to be compiled with the --opt_for_space option. The resulting output is shown in Example 3-7. Example 3‑6. C Code to Show Code-Size Optimizations extern int x, y, *ptr; extern int foo(); int main(int a, int b, int c) { ptr[50] = foo(); y = ptr[50] + x + y + a +b + c; } SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 73 Increasing Code-Size Optimizations (--opt_for_space Option) Example 3‑7. Example 3-6 Compiled With the --opt_for_space Option FP .set XAR2 .global _prolog_c28x_1 .global _prolog_c28x_2 .global _prolog_c28x_3 .global _epilog_c28x_1 .global _epilog_c28x_2 .sect ".text" .global _main ;*************************************************************** ;* FNAME: _main FR SIZE: 6 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 0 Auto, 6 SOE * ;*************************************************************** _main: FFC XAR7,_prolog_c28x_1 MOVZ AR3,AR4 MOVZ AR2,AH MOVZ AR1,AL LCR #_foo ; call occurs [#_foo] ; |6| MOVW DP,#_ptr MOVL XAR6,@_ptr MOVB XAR0,#50 MOVW DP,#_y MOV *+XAR6[AR0],AL MOV AH,@_y MOVW DP,#_x ADD AH,AL ADD AH,@_x ADD AH,AR3 ADD AH,AR1 ADD AH,AR2 MOVB AL,#0 MOVW DP,#_y MOV @_y,AH FFC XAR7,_epilog_c28x_1 LRETR ; return occurs ; |5| ; |5| ; |5| ; |6| ; |6| ; |6| ; |6| ; |7| ; |7| ; |7| ; |7| ; |7| ; |7| ; |7| www.ti.com 3.14 Compiler Support for Generating DMAC Instructions The C28x compiler supports DMAC instructions, which perform multiply-accumulate operations on two adjacent signed integers at the same time, optionally shifting the products. A multiply–accumulate operation multiplies two numbers and adds that product to an accumulator. That is, it computes a = a + (b x c). There are three levels of DMAC support that require different levels of C-source modification: • Generate DMAC instructions automatically from C code (see Section 3.14.1). • Use assertions for data address alignment to enable DMAC instruction generation (see Section 3.14.2). • Use the __dmac intrinsic (see Section 3.14.3). 74 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Compiler Support for Generating DMAC Instructions 3.14.1 Automatic Generation of DMAC Instructions The compiler automatically generates DMAC instructions if the compiler recognizes the C-language statements as a DMAC opportunity and the compiler can verify that the data addresses being operated upon are 32-bit aligned. This is the best scenario, because it requires no code modification aside from data alignment pragmas. The following is an example: int src1[N], src2[N]; #pragma DATA_ALIGN(src1,2); #pragma DATA_ALIGN(src2,2); // int arrays must be 32-bit aligned {...} int i; long res = 0; for (i = 0; i < N; i++) // N must be a known even constant res += (long)src1[i] * src2[i]; // Arrays must be accessed via array indices At optimization levels >= -O2, the compiler generates a RPT || DMAC instruction for the example code above if N is a known even constant. DMAC instructions can also shift the product left by 1 or right by 1 to 6 before accumulation. For example: for (i = 0; i < N; i++) res += (long)src1[i] * src2[i] >> 1; // product shifted right by 1 3.14.2 Assertions to Specify Data Address Alignment In some cases the compiler may recognize a DMAC opportunity in C-language statements but not be able to verify that the data addresses passed to the computation are both 32-bit aligned. In this case, assertions placed in the code can enable the compiler to generate DMAC instructions. The following is an example: int *src1, *src2; // src1 and src2 are pointers to int arrays of at least size N // You must ensure that both are 32-bit aligned addresses {...} int i; long res = 0; _nassert((long)src1 % 2 == 0); _nassert((long)src2 % 2 == 0); for (i = 0; i < N; i++) // N must be a known even constant res += (long)src1[i] * src2[i]; // src1 and src2 must be accessed via array indices At optimization levels >= -O2, the compiler generates a RPT || DMAC instruction for the example code above if N is a known even constant. The _nassert intrinsic generates no code and so is not a compiler intrinsic like those listed in Table 7-6. Instead, it tells the optimizer that the expression declared with the assert function is true. This can be used to give hints to the optimizer about what optimizations might be valid. In this example, _nassert is used to assert that the data addresses represented by the src1 and src2 pointers are 32-bit aligned. You are responsible for ensuring that only 32-bit aligned data addresses are passed via these pointers. The code will result in a run-time failure if the asserted conditions are not true. DMAC instructions can also shift the product left by 1 or right by 1 to 6 before accumulation. For example: for (i = 0; i < N; i++) res += (long)src1[i] * src2[i] >> 1; // product shifted right by 1 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 75 Compiler Support for Generating DMAC Instructions www.ti.com 3.14.3 __dmac Intrinsic You can force the compiler to generate a DMAC instruction by using the __dmac intrinsic. In this case, you are fully responsible for ensuring that the data addresses are 32-bit aligned. void __dmac(long *src1, long *src2, long &accum1, long &accum2, int shift); • Src1 and src2 must be 32-bit aligned addresses that point to int arrays. • Accum1 and accum2 are pass-by-reference longs for the two temporary accumulations. These must be added together after the intrinsic to compute the total sum. • Shift specifies how far to shift the products prior to each accumulation. Valid shift values are -1 for a left shift by 1, 0 for no shift, and 1-6 for a right shift by 1-6, respectively. Note that this argument is required, so you must pass 0 if you want no shift. See Table 7-6 for details about the __dmac intrinsic. Example 1: int src1[2N], src2[2N]; // src1 and src2 are int arrays of at least size 2N // You must ensure that both start on 32-bit // aligned boundaries. {...} int i; long res = 0; long temp = 0; for (i=0; i < N; i++) // N does not have to be a known constant __dmac(((long *)src1)[i], ((long *)src2)[i], res, temp, 0); res += temp; Example 2: int *src1, *src2; {...} int i; long res = 0; long temp = 0; // src1 and src2 are pointers to int arrays of at // least size 2N. User must ensure that both are // 32-bit aligned addresses. long *ls1 = (long *)src1; long *ls2 = (long *)src2; for (i=0; i < N; i++) // N does not have to be a known constant __dmac(*ls1++, *ls2++, res, temp, 0); res += temp; In these examples, res holds the final sum of a multiply-accumulate operation on int arrays of length 2*N, with two computations being performed at a time. Additionally, an optimization level >= -O2 must be used to generate efficient code. Moreover, if there is nothing else in the loop body as in these examples, the compiler generates a RPT || DMAC instruction, further improving performance. 76 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com What Kind of Optimization Is Being Performed? 3.15 What Kind of Optimization Is Being Performed? The TMS320C28x C/C++ compiler uses a variety of optimization techniques to improve the execution speed of your C/C++ programs and to reduce their size. Following are some of the optimizations performed by the compiler: Optimization Cost-based register allocation Alias disambiguation Branch optimizations and control-flow simplification Data flow optimizations • Copy propagation • Common subexpression elimination • Redundant assignment elimination Expression simplification Inline expansion of functions Function Symbol Aliasing Induction variable optimizations and strength reduction Loop-invariant code motion Loop rotation Instruction scheduling See Section 3.15.1 Section 3.15.1 Section 3.15.3 Section 3.15.4 Section 3.15.5 Section 3.15.6 Section 3.15.7 Section 3.15.8 Section 3.15.9 Section 3.15.10 Section 3.15.11 C28x-Specific Optimization Register variables Register tracking/targeting Tail merging Autoincrement addressing Removing comparisons to zero RPTB generation (for FPU targets only) See Section 3.15.12 Section 3.15.13 Section 3.15.14 Section 3.15.15 Section 3.15.16 Section 3.15.17 3.15.1 Cost-Based Register Allocation The compiler, when optimization is enabled, allocates registers to user variables and compiler temporary values according to their type, use, and frequency. Variables used within loops are weighted to have priority over others, and those variables whose uses do not overlap can be allocated to the same register. Induction variable elimination and loop test replacement allow the compiler to recognize the loop as a simple counting loop and unroll or eliminate the loop. Strength reduction turns the array references into efficient pointer references with autoincrements. 3.15.2 Alias Disambiguation C and C++ programs generally use many pointer variables. Frequently, compilers are unable to determine whether or not two or more I values (lowercase L: symbols, pointer references, or structure references) refer to the same memory location. This aliasing of memory locations often prevents the compiler from retaining values in registers because it cannot be sure that the register and memory continue to hold the same values over time. Alias disambiguation is a technique that determines when two pointer expressions cannot point to the same location, allowing the compiler to freely optimize such expressions. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Optimizing Your Code 77 What Kind of Optimization Is Being Performed? www.ti.com 3.15.3 Branch Optimizations and Control-Flow Simplification The compiler analyzes the branching behavior of a program and rearranges the linear sequences of operations (basic blocks) to remove branches or redundant conditions. Unreachable code is deleted, branches to branches are bypassed, and conditional branches over unconditional branches are simplified to a single conditional branch. When the value of a condition is determined at compile time (through copy propagation or other data flow analysis), the compiler can delete a conditional branch. Switch case lists are analyzed in the same way as conditional branches and are sometimes eliminated entirely. Some simple control flow constructs are reduced to conditional instructions, totally eliminating the need for branches. 3.15.4 Data Flow Optimizations Collectively, the following data flow optimizations replace expressions with less costly ones, detect and remove unnecessary assignments, and avoid operations that produce values that are already computed. The compiler with optimization enabled performs these data flow optimizations both locally (within basic blocks) and globally (across entire functions). • Copy propagation. Following an assignment to a variable, the compiler replaces references to the variable with its value. The value can be another variable, a constant, or a common subexpression. This can result in increased opportunities for constant folding, common subexpression elimination, or even total elimination of the variable. • Common subexpression elimination. When two or more expressions produce the same value, the compiler computes the value once, saves it, and reuses it. • Redundant assignment elimination. Often, copy propagation and common subexpression elimination optimizations result in unnecessary assignments to variables (variables with no subsequent reference before another assignment or before the end of the function). The compiler removes these dead assignments. 3.15.5 Expression Simplification For optimal evaluation, the compiler simplifies expressions into equivalent forms, requiring fewer instructions or registers. Operations between constants are folded into single constants. For example, a = (b + 4) - (c + 1) becomes a = b - c + 3. 3.15.6 Inline Expansion of Functions The compiler replaces calls to small functions with inline code, saving the overhead associated with a function call as well as providing increased opportunities to apply other optimizations. 3.15.7 Function Symbol Aliasing The compiler recognizes a function whose definition contains only a call to another function. If the two functions have the same signature (same return value and same number of parameters with the same type, in the same order), then the compiler can make the calling function an alias of the called function. For example, consider the following: int bbb(int arg1, char *arg2); int aaa(int n, char *str) { return bbb(n, str); } For this example, the compiler makes aaa an alias of bbb, so that at link time all calls to function aaa should be redirected to bbb. If the linker can successfully redirect all references to aaa, then the body of function aaa can be removed and the symbol aaa is defined at the same address as bbb. For information about using the GCC function attribute syntax to declare function aliases, see Section 6.14.2 78 Optimizing Your Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com What Kind of Optimization Is Being Performed? 3.15.8 Induction Variables and Strength Reduction Induction variables are variables whose value within a loop is directly related to the number of executions of the loop. Array indices and control variables for loops are often induction variables. Strength reduction is the process of replacing inefficient expressions involving induction variables with more efficient expressions. For example, code that indexes into a sequence of array elements is replaced with code that increments a pointer through the array. Induction variable analysis and strength reduction together often remove all references to your loopcontrol variable, allowing its elimination. 3.15.9 Loop-Invariant Code Motion This optimization identifies expressions within loops that always compute to the same value. The computation is moved in front of the loop, and each occurrence of the expression in the loop is replaced by a reference to the precomputed value. 3.15.10 Loop Rotation The compiler evaluates loop conditionals at the bottom of loops, saving an extra branch out of the loop. In many cases, the initial entry conditional check and the branch are optimized out. 3.15.11 Instruction Scheduling The compiler performs instruction scheduling, which is the rearranging of machine instructions in such a way that improves performance while maintaining the semantics of the original order. Instruction scheduling is used to improve instruction parallelism and hide latencies. It can also be used to reduce code size. 3.15.12 Register Variables The compiler helps maximize the use of registers for storing local variables, parameters, and temporary values. Accessing variables stored in registers is more efficient than accessing variables in memory. Register variables are particularly effective for pointers. 3.15.13 Register Tracking/Targeting The compiler tracks the contents of registers to avoid reloading values if they are used again soon. Variables, constants, and structure references such as (a.b) are tracked through straight-line code. Register targeting also computes expressions directly into specific registers when required, as in the case of assigning to register variables or returning values from functions. 3.15.14 Tail Merging If you are optimizing for code size, tail merging can be very effective for some functions. Tail merging finds basic blocks that end in an identical sequence of instructions and have a common destination. If such a set of blocks is found, the sequence of identical instructions is made into its own block. These instructions are then removed from the set of blocks and replaced with branches to the newly created block. Thus, there is only one copy of the sequence of instructions, rather than one for each block in the set. 3.15.15 Autoincrement Addressing For pointer expressions of the form *p++, the compiler uses efficient C28x autoincrement addressing modes. In many cases, where code steps through an array in a loop such as below, the loop optimizations convert the array references to indirect references through autoincremented register variable pointers. for (I = 0; I remark: linking in "libc.a" remark: linking in "rts2800_ml.lib" in place of "libc.a" 4.3.1.2 Manual Run-Time-Support Library Selection You can bypass automatic library selection by explicitly specifying the desired run-time-support library to use. Use the --library linker option to specify the name of the library. The linker will search the path specified by the --search_path option and then the C2000_C_DIR environment variable for the named library. You can use the --library linker option on the command line or in a command file. cl2000 --run_linker {--rom_model | --ram_model} filenames --library= libraryname 4.3.1.3 Library Order for Searching for Symbols Generally, you should specify the run-time-support library as the last name on the command line because the linker searches libraries for unresolved references in the order that files are specified on the command line. If any object files follow a library, references from those object files to that library are not resolved. You can use the --reread_libs option to force the linker to reread all libraries until references are resolved. Whenever you specify a library as linker input, the linker includes and links only those library members that resolve undefined references. By default, if a library introduces an unresolved reference and multiple libraries have a definition for it, then the definition from the same library that introduced the unresolved reference is used. Use the --priority option if you want the linker to use the definition from the first library on the command line that contains the definition. 4.3.2 Run-Time Initialization You must link all C/C++ programs with code to initialize and execute the program called a bootstrap routine. The bootstrap routine is responsible for the following tasks: 1. Set up status and configuration registers 2. Set up the stack and secondary system stack 3. Process the .cinit run-time initialization table to autoinitialize global variables (when using the -- rom_model option) 4. Call all global object constructors (.pinit) 5. Call the main() function 6. Call exit() when main() returns NOTE: The _c_int00 Symbol If you use the --ram_model or --rom_model link option, _c_int00 is automatically defined as the entry point for the program. Otherwise, an entry point is not automatically selected and you will receive a linker warning. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Linking C/C++ Code 85 Controlling the Linking Process www.ti.com 4.3.3 Initialization by the Interrupt Vector If your C/C++ program begins running at RESET, you must set up the RESET vector to branch to a suitable bootstrap routine, such as _c_int00. You must also make sure the interrupt vectors are included in the project, typically by using the --undef_sym linker option to make a symbol at the start of the interrupt vector a root linker object. A sample interrupt vector is provided in vectors.obj in rts2800_ml.lib. For C28x, the first few lines of the vector are: .def _Reset .ref _c_int00 _Reset: .vec _c_int00, USE_RETA 4.3.4 Global Object Constructors Global C++ variables that have constructors and destructors require their constructors to be called during program initialization and their destructors to be called during program termination. The C++ compiler produces a table of constructors to be called at startup. Constructors for global objects from a single module are invoked in the order declared in the source code, but the relative order of objects from different object files is unspecified. Global constructors are called after initialization of other global variables and before the main() function is called. Global destructors are invoked during the exit run-time support function, similar to functions registered through atexit. Section 7.9.3.4 discusses the format of the global constructor table. 4.3.5 Specifying the Type of Global Variable Initialization The C/C++ compiler produces data tables for initializing global variables. Section 7.9.3.1 discusses the format of these initialization tables. The initialization tables are used in one of the following ways: • Global variables are initialized at run time. Use the --rom_model linker option (see Section 7.9.3.2). • Global variables are initialized at load time. Use the --ram_model linker option (see Section 7.9.3.3). When you link a C/C++ program, you must use either the --rom_model or --ram_model option. These options tell the linker to select initialization at run time or load time. When you compile and link programs, the --rom_model option is the default. If used, the --rom_model option must follow the --run_linker option (see Section 4.1). The following list outlines the linking conventions used with --rom_model or -ram_model: • The symbol _c_int00 is defined as the program entry point; it identifies the beginning of the C/C++ boot routine in boot.obj. When you use --rom_model or --ram_model, _c_int00 is automatically referenced, ensuring that boot.obj is automatically linked in from the run-time-support library. • The initialization output section is padded with a termination record so that the loader (load-time initialization) or the boot routine (run-time initialization) knows when to stop reading the initialization tables. • When initializing at load time (the --ram_model option), the following occur: – The linker sets the initialization table symbol to -1. This indicates that the initialization tables are not in memory, so no initialization is performed at run time. – The STYP_COPY flag is set in the initialization table section header. STYP_COPY is the special attribute that tells the loader to perform autoinitialization directly and not to load the initialization table into memory. The linker does not allocate space in memory for the initialization table. • When autoinitializing at run time (--rom_model option), the linker defines the initialization table symbol as the starting address of the initialization table. The boot routine uses this symbol as the starting point for autoinitialization. 86 Linking C/C++ Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Controlling the Linking Process 4.3.6 Specifying Where to Allocate Sections in Memory The compiler produces relocatable blocks of code and data. These blocks, called sections, are allocated in memory in a variety of ways to conform to a variety of system configurations. See Section 7.1.1 for a complete description of how the compiler uses these sections. The compiler creates two basic kinds of sections: initialized and uninitialized. Table 4-1 summarizes the initialized sections. Table 4-2 summarizes the uninitialized sections. Name .binit .cinit .econst .pinit .switch .text Table 4-1. Initialized Sections Created by the Compiler Contents Boot time copy tables (See the Assembly Language Tools User's Guide for information on BINIT in linker command files.) Tables for explicitly initialized global and static variables. Global and static const variables that are explicitly initialized and contain string literals. String literals are placed in the .econst:.string subsection to enable greater link-time placement control. Table of constructors to be called at startup. Jump tables for large switch statements. Executable code and constants. Restrictions Program Anywhere in data Program Program with -mt option Data without -mt option Program Name .ebss .stack .esysmem Table 4-2. Uninitialized Sections Created by the Compiler Contents Global and static variables Stack Memory for malloc functions (heap) Restrictions Anywhere in data Low 64K Anywhere in data When you link your program, you must specify where to allocate the sections in memory. In general, initialized sections are linked into ROM or RAM; uninitialized sections are linked into RAM. With the exception of .text, the initialized and uninitialized sections created by the compiler cannot be allocated into internal program memory. The linker provides MEMORY and SECTIONS directives for allocating sections. For more information about allocating sections into memory, see the TMS320C28x Assembly Language Tools User's Guide. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Linking C/C++ Code 87 Controlling the Linking Process www.ti.com 4.3.7 A Sample Linker Command File Example 4-2 shows a typical linker command file that links a C program. The command file in this example is named lnk.cmd. It links three object files (a.obj, b.obj, and c.obj) and creates a program (prog.out) and a map file (prog.map). To link the program, enter the following: cl2000 --run_linker lnk.cmd The MEMORY and possibly the SECTIONS directives, might require modification to work with your system. See the TMS320C28x Assembly Language Tools User's Guide for more information on these directives. Example 4-2. Linker Command File a.obj b.obj c.obj --output_file=prog.out --map_file=prog.map --library=rts2800_ml.lib /* Input filenames */ /* Options */ /* Get run-time support */ MEMORY { RAM: ROM: } origin = 100h origin = 01000h /* MEMORY directive */ length = 0100h length = 0100h SECTIONS { .text: > ROM .data: > ROM .ebss: > RAM .pinit: > ROM .cinit: > ROM .switch: > ROM .econst: > RAM .stack: > RAM .esysmem: > RAM /* SECTIONS directive */ 88 Linking C/C++ Code Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Linking C28x and C2XLP Code 4.4 Linking C28x and C2XLP Code The error in the C28x linker to prevent linking code with a 64-word page size (C28x) and a 128-word page size (C2XLP) has been changed to a warning. It is possible to call a C2XLP assembly function from C28x C/C++ code. One possible way is to replace the call to the C2XLP function with a veneer function that correctly sets up the arguments and call stack for the C2XLP code. For example, to make a call to a C2XLP function expecting five integer arguments, change the C28x code to: extern void foo_veneer(int, int, int, int, int); void bar() { /* replace the C2XLP call with a veneer call */ /* foo(1, 2, 3, 4, 5); */ foo_veneer(1, 2, 3, 4, 5); } Example 4-3 illustrates how the veneer function might look: Example 4‑3. Veneer Function for Linking C2xx and C2XLP Code .sect ".text" .global _foo_veneer .global _foo _foo_veneer: ;save registers PUSH AR1:AR0 PUSH AR3:AR2 PUSH AR5:AR4 ;set the size of the C2XLP frame (including args size) ADDB SP,#10 ;push args onto the C2XLP frame MOV *-SP[10],AL ;copy arg 1 MOV *-SP[9],AH ;copy arg 2 MOV *-SP[8],AR4 ;copy arg 3 MOV *-SP[7],AR5 ;copy arg 4 MOV AL,*-SP[19] MOV *-SP[6],AL ;copy arg 5 ;save the return address MOV *-SP[5],#_label ;set AR1,ARP MOV AL,SP SUBB AL,#3 MOV AR1,AL NOP *ARP1 ;jump to C2XLP function LB _foo _label: ;restore register POP AR5:AR4 POP AR3:AR2 POP AR1:AR0LRETR Since the veneer function frame will act as the frame for all C2XLP calls, it is necessary to add sufficient size to the frame for any subsequent calls made by the first C2XLP function. Global variables will be placed in the .ebss sections for C28x C/C++ code. A C2XLP .ebss section is not guaranteed to begin on a 128-word boundary when linked with C28x code. To avoid this problem, define a new section, change the C2XLP globals to the new section, and update the linker command file to ensure this new section begins at a 128-word boundary. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Linking C/C++ Code 89 Chapter 5 SPRU514N – June 2017 Post-Link Optimizer The TMS320C28x post-link optimizer removes or modifies assembly language instructions to generate better code. The post-link optimizer examines the final addresses of symbols determined by linking and uses this information to make code changes. Post-link optimization requires the -plink compiler option. The -plink compiler option invokes added passes of the tools that include running the absolute lister and rerunning the assembler and linker. You must use the -plink option following the --run_linker option. Topic ........................................................................................................................... Page 5.1 The Post-Link Optimizer’s Role in the Software Development Flow.......................... 91 5.2 Removing Redundant DP Loads .......................................................................... 92 5.3 Tracking DP Values Across Branches .................................................................. 92 5.4 Tracking DP Values Across Function Calls ........................................................... 93 5.5 Other Post-Link Optimizations ............................................................................. 93 5.6 Controlling Post-Link Optimizations ..................................................................... 94 5.7 Restrictions on Using the Post-Link Optimizer ...................................................... 95 5.8 Naming the Outfile (--output_file Option)............................................................... 95 90 Post-Link Optimizer Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com The Post-Link Optimizer’s Role in the Software Development Flow 5.1 The Post-Link Optimizer’s Role in the Software Development Flow The post-link optimizer is not part of the normal development flow. Figure 5-1 shows the flow including the post-link optimizer; this flow occurs only when you use the compiler and the -plink option. Figure 5-1. The Post-Link Optimizer in the TMS320C28x Software Development Flow C/C++ source files Macro source files C/C++ compiler Archiver Assembler source C/C++ name demangling utility Macro library Assembler Archiver Library of object files Object files Linker Hex-conversion utility Executable object file EPROM Cross-reference TMS320C28x programmer lister Library-build process Run-timesupport library Absolute lister Assembler source Post-link optimizer Debugging tools As the flow shows, the absolute lister (abs2000) is also part of the post-link optimizing process. The absolute lister outputs the absolute addresses of all globally defined symbols and COFF sections. The post-link optimizer takes .abs files as input and uses these addresses to perform optimizations. The output is a .pl file, which is an optimized version of the original .asm file. The flow then reruns the assembler and linker to produce a final output file. The described flow is supported only when you use the compiler (cl2000) and the -plink option. If you use a batch file to invoke each tool individually, you must adapt the flow to use the compiler instead. In addition, you must use the --output_file option to specify an output file name when using the -plink option. See Section 5.8 for more details. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Post-Link Optimizer 91 Removing Redundant DP Loads www.ti.com For example, replace these lines: cl2000 file1.asm file1.obj cl2000 file2.asm file2.obj cl2000 --run_linker file1.obj file2.obj lnk.cmd --output_file=prog.out with this line: cl2000 file1.asm file2.asm --run_linker lnk.cmd --output_file=prog.out -plink The advice_only mode is the only post-link optimization support provided for FPU and VCU targets. 5.2 Removing Redundant DP Loads Post-link optimization reduces the difficulty of managing the Data Page (DP) register by removing redundant DP loads. It does this by tracking the current value of the DP and determining whether the address in a MOV DP,#address instruction is located on the same 64-word page to which the DP is currently pointing. If the address can be accessed using the current DP value, the instruction is redundant and can be removed. For example, consider the following code segment: MOVZ ADD MOVZ ADD DP,#name1 @name1,#10 DP,#name2 @name2,#10 If name1 and name2 are linked to the same page, the post-link optimizer determines that loading DP with the address of name2 is not necessary, and it comments out the redundant load. MOVZ DP,#name1 ADD @name1,#10 ; <> ADD @name2,#10 MOVZ DP,#name2 This optimization can be used on C files as well. Even though the compiler manages the DP for all global variable references that are defined within a module, it conservatively emits DP loads for any references to global variables that are externally defined. Using the post-link optimizer can help reduce the number of DP loads in these instances. Additionally, the --map_file linker option can be used to generate a listing of symbols sorted by data page to assist with data layout and DP load reduction. For more information, refer to the Linker Description chapter in the TMS320C28x Assembly Language Tools User's Guide. 5.3 Tracking DP Values Across Branches In order to track DP values across branches, the post-link optimizer requires that there are no indirect calls or branches, and all possible branch destinations have labels. If an indirect branch or call is encountered, the post-link optimizer will only track the DP value within a basic block. Branch destinations without labels may cause incorrect output from the post-link optimizer. If the post-link optimizer encounters indirect calls or branches, it issues the following warning: NO POST LINK OPTIMIZATION DONE ACROSS BRANCHES Branch/Call must have labeled destination This warning is issued so that if the file is a hand written assembly file, you can try to change the indirect call/branch to a direct one to obtain the best optimization from the post linker. 92 Post-Link Optimizer Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Tracking DP Values Across Function Calls 5.4 Tracking DP Values Across Function Calls The post-link optimizer optimizes DP loads after a call to a function if the function is defined in the same file scope. For example, consider the following post-link optimized code: _main: LCR #_foo MOVB AL, #0 ;<> MOVZ MOV @_g2, #20 LRETR DP,#_g2 .global _foo _foo: MOVZ DP, #g1 MOV @_g1, #10 LRETR The MOVZ DP after the function call to _foo is removed by the post-link optimizer as the variables _g1 and _g2 are in the same page and the function _foo already set the DP. In order for the post-link optimizer to optimize across the function calls, the functions should have only one return statement. If you are running the post-link optimizer on hand written assembly that has more than one return statement per function, the post-link optimization output can be incorrect. You can turn off the optimization across function calls by specifying the -nf option after the -plink option. 5.5 Other Post-Link Optimizations An externally defined symbol used as a constant operand forces the assembler to choose a 16-bit encoding to hold the immediate value. Since the post-link optimizer has access to the externally defined symbol value, it replaces a 16-bit encoding with an 8-bit encoding when possible. For example: .ref ext_sym ; externally defined to be 4 : : ADD AL, #ext_sym ; assembly will encode ext_sym with 16 bits Since ext_sym is externally defined, the assembler chooses a 16-bit encoding for ext_sym. The post-link optimizer changes the encoding of ext_sym to an 8-bit encoding: .ref ext_sym : : ; << ADD=>ADDB>> ADD AL,#ext_sym ADDB AL, #ext_sym Similarly the post-link optimizer attempts to reduce the following 2-word instructions to 1-word instructions: 2-Word Instructions ADD ACC, #imm ADD AL, #imm AND AL, #imm CMP AL, #imm MOVL XARn, #imm OR AL, #imm SUB ACC, #imm SUB AL, #imm XOR AL, #imm 1-Word Instructions ADDB ACC, #imm ADDB AL, #imm ANDB AL, #imm CMPB AL, #imm MOVB XARn, #imm ORB AL, #imm SUBB ACC, #imm SUBB AL, #imm XORB AL, #imm SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Post-Link Optimizer 93 Controlling Post-Link Optimizations www.ti.com 5.6 Controlling Post-Link Optimizations There are three ways to control post-link optimizations: by excluding files, by inserting specific comments within an assembly file, and by manually editing the post-link optimization file. 5.6.1 Excluding Files (-ex Option) Specific files can be excluded from the post-link optimization process by using the −ex option. The files to be excluded must follow the −ex option and include file extensions. The −ex option must be used after the −plink option and no other option may follow. For example: cl2000 file1.asm file2.asm file3.asm --keep_asm --run_linker lnk.cmd -plink -o=prog.out ex=file3.asm The file3.asm will be excluded from the post-link optimization process. 5.6.2 Controlling Post-Link Optimization Within an Assembly File Within an assembly file, post-link optimizations can be disabled or enabled by using two specially formatted comment statements: ;//NOPLINK// ;//PLINK// Assembly statements following the NOPLINK comment are not optimized. Post-link optimization can be reenabled using the //PLINK// comment. The PLINK and NOPLINK comment format is not case sensitive. There can be white space between the semicolon and PLINK delimiter. The PLINK and NOPLINK comments must appear on a separate line, alone, and must begin in column 1. For example: ; //PLINK// 5.6.3 Retaining Post-Link Optimizer Output (--keep_asm Option) The --keep_asm option allows you to retain any post-link files (.pl) and .absolute listing files (.abs) generated by the −plink option. Using the --keep_asm option lets you view any changes the post-link optimizer makes. The .pl files contain the commented out statement shown with <> or any improvements to instructions, such as <ADDB>>. The .pl files are assembled and linked again to exclude the commented out lines. 5.6.4 Disable Optimization Across Function Calls (-nf Option ) The −nf option disables the post-link optimization across function calls. The post-link optimizer recognizes the function end by the return statement and assumes there is only one return statement per function. In some hand written assembly code, it is possible to have more than one return statement per function. In such cases, the output of the post-link optimization can be incorrect. You can turn off the optimization across function calls by using the− nf option. This option affects all the files. 5.6.5 Annotating Assembly with Advice (--plink_advice_only option) The --plink_advice_only option annotates assembly code with comments if changes cannot be made safely due to pipeline considerations, such as when float support or VCU support is enabled. If you use this option, note that the post-link files (.pl), which contain the generated advice, are retained only if you also use the --keep_asm option. 94 Post-Link Optimizer Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Restrictions on Using the Post-Link Optimizer 5.7 Restrictions on Using the Post-Link Optimizer The following restrictions affect post-link optimization: • The advice_only mode is the only post-link optimization support provided for FPU and VCU targets. • Branches or calls to unlabeled destinations invalidate DP load optimizations. All branch destinations must have labels. • If the position of the data sections depends on the size of the code sections, the data page layout information used to decide which DP load instructions to remove may no longer be valid. For example, consider the following link command file: SECTIONS { .text .mydata > MEM, > MEM,} A change in the size of the .text section after optimizing causes the .mydata section to shift. Ensuring that all output data sections are aligned on a 64-word boundary removes this shifting issue. For example, consider the following link command file: SECTIONS { .text > MEM, .mydata align = 64 > MEM,} 5.8 Naming the Outfile (--output_file Option) When using the -plink option, you must include the --output_file option. If the output filename is specified in a linker command file, the compiler does not have access to the filename to pass it along to other phases of post-link optimization, and the process will fail. For example: cl2000 file1.c file2.asm --run_linker --output_file=prog.out lnk.cmd -plink Because the post-link optimization flow uses the absolute lister, abs2000, it must be included in your path. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Post-Link Optimizer 95 Chapter 6 SPRU514N – June 2017 TMS320C28x C/C++ Language Implementation The C/C++ compiler supports the C/C++ language standard that was developed by a committee of the American National Standards Institute (ANSI) and subsequently adopted by the International Standards Organization (IS0). The C++ language supported by the C28x is defined by the ANSI/ISO/IEC 14882:2003 standard with certain exceptions. Topic ........................................................................................................................... Page 6.1 Characteristics of TMS320C28x C ........................................................................ 97 6.2 Characteristics of TMS320C28x C++ ................................................................... 101 6.3 Using MISRA C 2004 ......................................................................................... 102 6.4 Data Types....................................................................................................... 103 6.5 Keywords ........................................................................................................ 105 6.6 C++ Exception Handling .................................................................................... 110 6.7 Register Variables and Parameters ..................................................................... 110 6.8 The __asm Statement........................................................................................ 111 6.9 Pragma Directives ............................................................................................ 112 6.10 The _Pragma Operator ...................................................................................... 126 6.11 Object File Symbol Naming Conventions (Linknames) .......................................... 127 6.12 Initializing Static and Global Variables ............................................................... 127 6.13 Changing the ANSI/ISO C/C++ Language Mode .................................................... 128 6.14 GNU Language Extensions ................................................................................ 131 6.15 Compiler Limits................................................................................................ 134 96 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Characteristics of TMS320C28x C 6.1 Characteristics of TMS320C28x C The C compiler supports the 1989 and 1999 versions of the C language: • C89. Compiling with the --c89 option causes the compiler to conform to the ISO/IEC 9899:1990 C standard, which was previously ratified as ANSI X3.159-1989. The names "C89" and "C90" refer to the same programming language. "C89" is used in this document. • C99. Compiling with the --c99 option causes the compiler to conform to the ISO/IEC 9899:1999 C standard. This standard supports several features not part of C89, such as inline functions, new data types, and one-line comments beginning with //. The C language is also described in the second edition of Kernighan and Ritchie's The C Programming Language (K&R). The compiler can also accept many of the language extensions found in the GNU C compiler (see Section 6.14). The compiler supports some features of C99 in the default relaxed ANSI mode with C89 support. It supports all language features of C99 in C99 mode. See Section 6.13. The ANSI/ISO standard identifies some features of the C language that may be affected by characteristics of the target processor, run-time environment, or host environment. This set of features can differ among standard compilers. Unsupported features of the C library are: • The run-time library has minimal support for wide and multibyte characters. The type wchar_t is implemented as int. The wide character set is equivalent to the set of values of type char. The library includes the header files and , but does not include all the functions specified in the standard. • The run-time library includes the header file , but with a minimal implementation. The only supported locale is the C locale. That is, library behavior that is specified to vary by locale is hardcoded to the behavior of the C locale, and attempting to install a different locale by way of a call to setlocale() will return NULL. • Some run-time functions and features in the C99 specification are not supported. See Section 6.13. 6.1.1 Implementation-Defined Behavior The C standard requires that conforming implementations provide documentation on how the compiler handles instances of implementation-defined behavior. The TI compiler officially supports a freestanding environment. The C standard does not require a freestanding environment to supply every C feature; in particular the library need not be complete. However, the TI compiler strives to provide most features of a hosted environment. The section numbers in the lists that follow correspond to section numbers in Appendix J of the C99 standard. The numbers in parentheses at the end of each item are sections in the C99 standard that discuss the topic. Certain items listed in Appendix J of the C99 standard have been omitted from this list. J.3.1 Translation • The compiler and related tools emit diagnostic messages with several distinct formats. Diagnostic messages are emitted to stderr; any text on stderr may be assumed to be a diagnostic. If any errors are present, the tool will exit with an exit status indicating failure (non-zero). (3.10, 5.1.1.3) • Nonempty sequences of white-space characters are preserved and are not replaced by a single space character in translation phase 3. (5.1.1.2) J.3.2 Environment • The compiler does not support multibyte characters in identifiers, so there is no mapping from multibyte characters to the source character set. However, the compiler accepts multibyte characters in comments, string literals, and character constants in the physical source file. (5.1.1.2) • The name of the function called at program startup is "main" (5.1.2.1) • Program termination does not affect the environment; there is no way to return an exit code to the environment. By default, the program is known to have halted when execution reaches the special CEXIT label. (5.1.2.1) • In relaxed ANSI mode, the compiler accepts "void main(void)" and "void main(int argc, char *argv[])" as SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 97 Copyright © 2017, Texas Instruments Incorporated Characteristics of TMS320C28x C www.ti.com alternate definitions of main. The alternate definitions are rejected in strict ANSI mode. (5.1.2.2.1) • If space is provided for program arguments at link time with the --args option and the program is run under a system that can populate the .args section (such as CCS), argv[0] will contain the filename of the executable, argv[1] through argv[argc-1] will contain the command-line arguments to the program, and argv[argc] will be NULL. Otherwise, the value of argv and argc are undefined. (5.1.2.2.1) • Interactive devices include stdin, stdout, and stderr (when attached to a system that honors CIO requests). Interactive devices are not limited to those output locations; the program may access hardware peripherals that interact with the external state. (5.1.2.3) • Signals are not supported. The function signal is not supported. (7.14) (7.14.1.1) • The library function getenv is implemented through the CIO interface. If the program is run under a system that supports CIO, the system performs getenv calls on the host system and passes the result back to the program. Otherwise the operation of getenv is undefined. No method of changing the environment from inside the target program is provided. (7.20.4.5) • The system function is not supported. (7.20.4.6). J.3.3. Identifiers • The compiler does not support multibyte characters in identifiers. (6.4.2) • The number of significant initial characters in an identifier is unlimited. (5.2.4.1, 6.4.2) J.3.4 Characters • The number of bits in a byte (CHAR_BIT) is 16. See Section 6.4 for details about data types. (3.6) • The execution character set is the same as the basic execution character set: plain ASCII. (5.2.1) • The values produced for the standard alphabetic escape sequences are as follows: (5.2.2) Escape Sequence \a \b \f \n \r \t \v ASCII Meaning BEL (bell) BS (backspace) FF (form feed) LF (line feed) CR (carriage return) HT (horizontal tab) VT (vertical tab) Integer Value 7 8 12 10 13 9 11 • The value of a char object into which any character other than a member of the basic execution character set has been stored is the ASCII value of that character. (6.2.5) • Plain char is identical to signed char. (6.2.5, 6.3.1.1) • The source character set and execution character set are both plain ASCII, so the mapping between them is one-to-one. The compiler does accept multibyte characters in comments, string literals, and character constants. (6.4.4.4, 5.1.1.2) • The compiler currently supports only one locale, "C". (6.4.4.4). • The compiler currently supports only one locale, "C". (6.4.5). J.3.5 Integers • No extended integer types are provided. (6.2.5) • Integer types are represented as two's complement, and there are no trap representations. (6.2.6.2) • No extended integer types are provided, so there is no change to the integer ranks. (6.3.1.1) • When an integer is converted to a signed integer type which cannot represent the value, the value is truncated (without raising a signal) by discarding the bits which cannot be stored in the destination type; the lowest bits are not modified. (6.3.1.3) • Right shift of a signed integer value performs an arithmetic (signed) shift. The bitwise operations other than right shift operate on the bits in exactly the same way as on an unsigned value. That is, after the usual arithmetic conversions, the bitwise operation is performed without regard to the format of the integer type, in particular the sign bit. (6.5) 98 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Characteristics of TMS320C28x C J.3.6 Floating point • The accuracy of floating-point operations (+ - * /) is bit-exact. The accuracy of library functions that return floating-point results is not specified. (5.2.4.2.2) • The compiler does not provide non-standard values for FLT_ROUNDS (5.2.4.2.2) • The compiler does not provide non-standard negative values of FLT_EVAL_METHOD (5.2.4.2.2) • The rounding direction when an integer is converted to a floating-point number is IEEE-754 "round to even". (6.3.1.4) • The rounding direction when a floating-point number is converted to a narrower floating-point number is IEEE-754 "round to even". (6.3.1.5) • For floating-point constants that are not exactly representable, the implementation uses the nearest representable value. (6.4.4.2) • The compiler does not contract float expressions. (6.5) • The default state for the FENV_ACCESS pragma is off. (7.6.1) • The TI compiler does not define any additional float exceptions (7.6, 7.12) • The default state for the FP_CONTRACT pragma is off. (7.12.2) • The "inexact" floating-point exception cannot be raised if the rounded result equals the mathematical result. (F.9) • The "underflow" and "inexact" floating-point exceptions cannot be raised if the result is tiny but not inexact. (F.9) J.3.7 Arrays and pointers • When converting a pointer to an integer or vice versa, the pointer is considered an unsigned integer of the same size, and the normal integer conversion rules apply. Some pointers are not the same size as any integral type, but the conversion proceeds as if such a type did exist, with the rules implied by normal integer conversion. • When converting a pointer to an integer or vice versa, if the bitwise representation of the destination can hold all of the bits in the bitwise representation of the source, the bits are copied exactly. (6.3.2.3) • The size of the result of subtracting two pointers to elements of the same array is the size of ptrdiff_t, which is defined in Section 6.4. (6.5.6) J.3.8 Hints • When the optimizer is used, the register storage-class specifier is ignored. When the optimizer is not used, the compiler will preferentially place register storage class objects into registers to the extent possible. The compiler reserves the right to place any register storage class object somewhere other than a register. (6.7.1) • The inline function specifier is ignored unless the optimizer is used. For other restrictions on inlining, see Section 2.11.4. (6.7.4) J.3.9 Structures, unions, enumerations, and bit-fields • A "plain" int bit-field is treated as a signed int bit-field. (6.7.2, 6.7.2.1) • In addition to _Bool, signed int, and unsigned int, the compiler allows char, signed char, unsigned char, signed short, unsigned shot, signed long, unsigned long, signed long long, unsigned long long, and enum types as bit-field types. (6.7.2.1) • Bit-fields may not straddle a storage-unit boundary.(6.7.2.1) • Bit-fields are allocated in endianness order within a unit. (6.7.2.1) • Non-bit-field members of structures are aligned as specified in See . (6.7.2.1) • The integer type underlying each enumerated type is described in Section 6.4.1. (6.7.2.2) J.3.10 Qualifiers • The TI compiler does not shrink or grow volatile accesses. It is the user's responsibility to make sure the access size is appropriate for devices that only tolerate accesses of certain widths. The TI compiler does not change the number of accesses to a volatile variable unless absolutely necessary. This is significant for read-modify-write expressions such as += ; for an architecture which does not have a corresponding read-modify-write instruction, the compiler will be forced to use two accesses, one for SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 99 Copyright © 2017, Texas Instruments Incorporated Characteristics of TMS320C28x C www.ti.com the read and one for the write. Even for architectures with such instructions, it is not guaranteed that the compiler will be able to map such expressions to an instruction with a single memory operand. It is not guaranteed that the memory system will lock that memory location for the duration of the instruction. In a multi-core system, some other core may write the location after a RMW instruction reads it, but before it writes the result. The TI compiler will not reorder two volatile accesses, but it may reorder a volatile and a non-volatile access, so volatile cannot be used to create a critical section. Use some sort of lock if you need to create a critical section. (6.7.3) J.3.11 Preprocessing directives • Include directives may have one of two forms, " " or < >. For both forms, the compiler will look for a real file on-disk by that name using the include file search path. See Section 2.5.2. (6.4.7). • The value of a character constant in a constant expression that controls conditional inclusion matches the value of the same character constant in the execution character set (both are ASCII). (6.10.1). • The compiler uses the file search path to search for an included < > delimited header file. See Section 2.5.2. (6.10.2). • he compiler uses the file search path to search for an included " " delimited header file. See Section 2.5.2. (6.10.2). (6.10.2). • There is no arbitrary nesting limit for #include processing. (6.10.2). • See Section 6.9 for a description of the recognized non-standard pragmas. (6.10.6). • The date and time of translation are always available from the host. (6.10.8). J.3.12 Library functions • Almost all of the library functions required for a hosted implementation are provided by the TI library, with exceptions noted in Section 6.13.1. (5.1.2.1). • The format of the diagnostic printed by the assert macro is "Assertion failed, (assertion macro argument), file file, line line". (7.2.1.1). • No strings other than "C" and "" may be passed as the second argument to the setlocale function (7.11.1.1). • No signal handling is supported. (7.14.1.1). • The +INF, -INF, +inf, -inf, NAN, and nan styles can be used to print an infinity or NaN. (7.19.6.1, 7.24.2.1). • The output for %p conversion in the fprintf or fwprintf function is the same as %x of the appropriate size. (7.19.6.1, 7.24.2.1). • The termination status returned to the host environment by the abort, exit, or _Exit function is not returned to the host environment. (7.20.4.1, 7.20.4.3, 7.20.4.4). • The system function is not supported. (7.20.4.6). J.3.13 Architecture • The values or expressions assigned to the macros specified in the headers float.h, limits.h, and stdint.h are described along with the sizes and format of integer types are described in Section 6.4. (5.2.4.2, 7.18.2, 7.18.3) • The number, order, and encoding of bytes in any object are described in . (6.2.6.1) • The value of the result of the sizeof operator is the storage size for each type, in terms of bytes. See . (6.5.3.4) 100 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Characteristics of TMS320C28x C++ 6.2 Characteristics of TMS320C28x C++ The C28x compiler supports C++ as defined in the ANSI/ISO/IEC 14882:2003 standard, including these features: • Complete C++ standard library support, with exceptions noted below. • Templates • Exceptions, which are enabled with the --exceptions option; see Section 6.6. • Run-time type information (RTTI), which can be enabled with the --rtti compiler option. The exceptions to the standard are as follows: • The compiler does not support embedded C++ run-time-support libraries. • The header and its functions are not included in the library. • The library supports wide chars (wchar_t), in that template functions and classes that are defined for char are also available for wchar_t. For example, wide char stream classes wios, wiostream, wstreambuf and so on (corresponding to char classes ios, iostream, streambuf) are implemented. However, there is no low-level file I/O for wide chars. Also, the C library interface to wide char support (through the C++ headers and ) is limited as described above in the C library. • If the definition of an inline function contains a static variable, and it appears in multiple compilation units (usually because it’s a member function of a class defined in a header file), the compiler generates multiple copies of the static variable rather than resolving them to a single definition. The compiler emits a warning (#1369) in such cases. • Two-phase name binding in templates, as described in [tesp.res] and [temp.dep] of the standard, is not implemented. • The export keyword for templates is not implemented. • A typedef of a function type cannot include member function cv-qualifiers. • A partial specialization of a class member template cannot be added outside of the class definition. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 101 Copyright © 2017, Texas Instruments Incorporated Using MISRA C 2004 www.ti.com 6.3 Using MISRA C 2004 MISRA C is a set of software development guidelines for the C programming language. It promotes best practices in developing safety-related electronic systems in road vehicles and other embedded systems. MISRA C was originally launched in 1998 by the Motor Industry Software Reliability Association, and has since been adopted across a wide variety of industries. A subsequent update to the guidelines was publishes as MISRA C:2004 You can alter your code to work with the MISRA C:2004 rules. The following options and pragmas can be used to enable/disable rules: • The --check_misra option enables checking of the specified MISRA C:2004 rules. This compiler option must be used if you want to enable further control over checking using the CHECK_MISRA and RESET_MISRA pragmas. • The CHECK_MISRA pragma enables/disables MISRA C:2004 rules at the source level. See Section 6.9.2. • The RESET_MISRA pragma resets the specified MISRA C:2004 rules to their state before any CHECK_MISRA pragmas were processed. See Section 6.9.17. The syntax of the option and the pragmas is: --check_misra={all|required|advisory|none|rulespec} #pragma CHECK_MISRA ("{all|required|advisory|none|rulespec}") #pragma RESET_MISRA ("{all|required|advisory|rulespec}") The rulespec parameter is a comma-separated list of rule numbers to enable or disable. Example: --check_misra=1.1,1.4,1.5,2.1,2.7,7.1,7.2,8.4 • Enables checking of rules 1.1, 1.4, 1.5, 2.1, 2.7, 7.1, 7.2, and 8.4. Example: #pragma CHECK_MISRA("-7.1,-7.2,-8.4") • Disables checking of rules 7.1, 7.2, and 8.4. A typical use case is to use the --check_misra option on the command line to specify the rules that should be checked in most of your code. Then, use the CHECK_MISRA pragma with a rulespec to activate or deactivate certain rules for a particular region of code. Two options control the severity of certain MISRA C:2004 rules: • The --misra_required option sets the diagnostic severity for required MISRA C:2004 rules. • The --misra_advisory option sets the diagnostic severity for advisory MISRA C:2004 rules. The syntax for these options is: --misra_advisory={error|warning|remark|suppress} --misra_required={error|warning|remark|suppress} 102 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Data Types 6.4 Data Types Table 6-1 lists the size, representation, and range of each scalar data type for the C28x compiler. Many of the range values are available as standard macros in the header file limits.h. Table 6-1. TMS320C28x C/C++ Data Types Range Type Size Representation Minimum Maximum char, signed char 16 bits ASCII -32 768 32 767 unsigned char, _Bool 16 bits ASCII 0 65 535 short 16 bits 2s complement -32 768 32 767 unsigned short 16 bits Binary 0 65 535 int, signed int 16 bits 2s complement -32 768 32 767 unsigned int 16 bits Binary 0 65 535 long, signed long 32 bits 2s complement -2 147 483 648 2 147 483 647 unsigned long 32 bits Binary 0 4 294 967 295 long long, signed long long 64 bits 2s complement -9 223 372 036 854 775 808 9 223 372 036 854 775 807 unsigned long long 64 bits Binary 0 18 446 744 073 709 551 615 enum (1) 16 bits 2s complement -32 768 32 767 float 32 bits IEEE 32-bit 1.19 209 290e-38(2) 3.40 282 35e+38 double 32 bits IEEE 32-bit 1.19 209 290e-38(2) 3.40 282 35e+38 long double 64 bits IEEE 64-bit 2.22 507 385e-308(2) 1.79 769 313e+308 pointers (3) 32 bits Binary 0 0xFFFFFFFF (1) For details about the size of an enum type, see Section 6.4.1. (2) Figures are minimum precision. (3) Even though pointers are 32-bits, the compiler assumes that the addresses of global variables and functions are within 22-bits. NOTE: TMS320C28x Byte is 16 Bits By ANSI/ISO C definition, the sizeof operator yields the number of bytes required to store an object. ANSI/ISO further stipulates that when sizeof is applied to char, the result is 1. Since the TMS320C28x char is 16 bits (to make it separately addressable), a byte is also 16 bits. This yields results you may not expect; for example, size of (int) = = 1 (not 2). TMS320C28x bytes and words are equivalent (16 bits). To access data in increments of 8 bits, use the __byte() and __mov_byte() intrinsics described in Section 7.5.6. 6.4.1 Size of Enum Types An enum type is represented by an underlying integer type. The size of the integer type and whether it is signed is based on the range of values of the enumerated constants. In strict C89 or C99 mode, the compiler allows only enumeration constants with values that will fit in "int" or "unsigned int". For C++ and relaxed C89/C99, the compiler allows enumeration constants up to the largest integral type (64 bits). The default, which is recommended, is for the underlying type to be the first type in the following list in which all the enumerated constant values can be represented: int, unsigned int, long, unsigned long long long, unsigned long long. 6.4.2 Support for 64-Bit Integers The TMS320C28x compiler supports the long long and unsigned long long data types. The range values are available as standard macros in the header file limits.h. The long long data types are stored in register pairs. In memory they are stored as 64-bit objects at word (32-bit) aligned addresses. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 103 Copyright © 2017, Texas Instruments Incorporated Data Types www.ti.com A long long integer constant can have an ll or LL suffix. Without the suffix the value of the constant determines the type of the constant. The formatting rules for long long in C I/O require ll in the format string. For example: printf("%lld", 0x0011223344556677); printf("%llx", 0x0011223344556677); The run-time-support library functions, llabs(), strtoll() and strtoull(), are added. 6.4.3 C28x long double Floating-Point Type Change When compiling C/C++ code for the TMS320C28x only, the long double floating point type is now IEEE 64-bit double precision. No other floating-point types have changed formats. C28x floating point types are: Type float double long double Format IEEE 32-bit single precision IEEE 32-bit single precision IEEE 64-bit double precision When you initialize a long double to a constant, you must use the l or L suffix. The constant is treated as a double type without the suffix and the run-time support double-to-long conversion routine is called for the initialization. This could result in the loss of precision. For example: long double a = 12.34L; /* correctly initializes to double precision */ long double b = 56.78; /* converts single precision value to double precision */ The formatting rules for long doubles in C I/O require a capital ’L’ in the format string. For example: printf("%Lg", 1.23L); printf("%Le", 3.45L); In response to the change in the long double type to 64-bit IEEE double-precision format, the C28x calling conventions have changed. All long double arguments are passed by reference. A long double return value is returned by reference. The first two long double arguments will pass their addresses in XAR4 and XAR5. All other long double arguments will have their addresses passed on the stack. If a function returns a long double, the function making that call will place the return address in XAR6. For example: long double foo(long double a, long double b, long double c) { long double d = a + b + c; return d; } long double a = 1.2L; long double b = 2.2L; long double c = 3.2L; long double d; void bar() { d = foo(a, b, c); } In function bar(), at the call to foo(), the register values are: Register XAR4 XAR5 *−SP[2] XAR6 Equals The address of a The address of b The address of c The address of d 104 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Keywords The run-time-support library has been updated to include the necessary long double arithmetic operations and conversion functions. All C28x floating-point run-time-support routines have had their names updated. For example, a previous call to the floating point add routine was: LCR FADD This has been updated to: LCR FS$$ADD ; single-precision add LCR FD$$ADD ; double-precision add Any C28x routine that calls floating-point arithmetic or conversion routines will need to be recompiled. 6.5 Keywords The C28x C/C++ compiler supports all of the standard C89 keywords, including const, volatile, and register. It also supports all of the standard C99 keywords, including inline and restrict. It also supports TI extension keywords __interrupt, __cregister, and __asm. The compiler supports the restrict keyword for FPU targets; for other targets restrict is ignored. Some keywords are not available in strict ANSI mode. The following keywords may appear in other target documentation and require the same treatment as the interrupt and restrict keywords: • trap • reentrant • cregister 6.5.1 The const Keyword The C/C++ compiler supports the ANSI/ISO standard keyword const in all modes. This keyword gives you greater optimization and control over allocation of storage for certain data objects. You can apply the const qualifier to the definition of any variable or array to ensure that its value is not altered. Global objects qualified as const are placed in the .econst section. The linker allocates the .econst section from ROM or FLASH, which are typically more plentiful than RAM. The const data storage allocation rule has two exceptions: • If the keyword volatile is also specified in the definition of an object (for example, volatile const int x). Volatile keywords are assumed to be allocated to RAM. (The program is not allowed to modify a const volatile object, but something external to the program might.) • If the object has automatic storage (function scope). In both cases, the storage for the object is the same as if the const keyword were not used. The placement of the const keyword within a definition is important. For example, the first statement below defines a constant pointer p to a modifiable int. The second statement defines a modifiable pointer q to a constant int: int * const p = &x; const int * q = &x; Using the const keyword, you can define large constant tables and allocate them into system ROM. For example, to allocate a ROM table, you could use the following definition: const int digits[] = {0,1,2,3,4,5,6,7,8,9}; SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 105 Copyright © 2017, Texas Instruments Incorporated Keywords www.ti.com 6.5.2 The __cregister Keyword The compiler extends the C/C++ language by adding the cregister keyword to allow high level language access to control registers. This keyword is available in normal mode, but not in strict ANSI/ISO mode (using the --strict_ansi compiler option). The alternate keyword, __cregister, provides the same functionality but is available in either strict ANSI/ISO mode or normal mode. When you use the cregister keyword on an object, the compiler compares the name of the object to a list of standard control registers for the C28x (see Table 6-2). If the name matches, the compiler generates the code to reference the control register. If the name does not match, the compiler issues an error. Register IER IFR Table 6-2. Valid Control Registers Description Interrupt enable register Interrupt flag register The cregister keyword can be used only in file scope. The cregister keyword is not allowed on any declaration within the boundaries of a function. It can only be used on objects of type integer or pointer. The cregister keyword is not allowed on objects of any floating-point type or on any structure or union objects. The cregister keyword does not imply that the object is volatile. If the control register being referenced is volatile (that is, can be modified by some external control), then the object must be declared with the volatile keyword also. To use the control registers in Table 6-2, you must declare each register as follows. The c28x.h include file defines all the control registers through this syntax: extern cregister volatile unsigned int register ; Once you have declared the register, you can use the register name directly, though in a limited manner. IFR is read only and can be set only by using the | (OR) operation with an immediate. IFR can be cleared only by using the & (AND) operation with an immediate. For example: IFR |= 0x4; IFR &= 0x0800 The IER register also can be used in an assignment other than OR and AND. Since the C28x architecture has limited instructions to manipulate these registers, the compiler terminates with the following message if illegal use of these registers is found: >>> Illegal use of control register See Example 6-1 for an example that declares and uses control registers. Example 6-1. Define and Use Control Registers extern cregister volatile unsigned int IFR; extern cregister volatile unsigned int IER; extern int x; main() { IER = x; IER |= 0x100; printf("IER = %x\n", IER); IFR &= 0x100; IFR |= 0x100; 106 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Keywords 6.5.3 The __interrupt Keyword The compiler extends the C/C++ language by adding the __interrupt keyword, which specifies that a function is treated as an interrupt function. This keyword is an IRQ interrupt. The alternate keyword, "interrupt", may also be used except in strict ANSI C or C++ modes. Note that the interrupt function attribute described in Section 6.9.14 is the recommended syntax for declaring interrupt functions. Functions that handle interrupts follow special register-saving rules and a special return sequence. The implementation stresses safety. The interrupt routine does not assume that the C run-time conventions for the various CPU register and status bits are in effect; instead, it re-establishes any values assumed by the run-time environment. When C/C++ code is interrupted, the interrupt routine must preserve the contents of all machine registers that are used by the routine or by any function called by the routine. When you use the __interrupt keyword with the definition of the function, the compiler generates register saves based on the rules for interrupt functions and the special return sequence for interrupts. You can only use the __interrupt keyword with a function that is defined to return void and that has no parameters. The body of the interrupt function can have local variables and is free to use the stack or global variables. For example: __interrupt void int_handler() { unsigned int flags; ... } The name c_int00 is the C/C++ entry point. This name is reserved for the system reset interrupt. This special interrupt routine initializes the system and calls the main() function. Because it has no caller, c_int00 does not save any registers. Hwi Objects and the __interrupt Keyword NOTE: The __interrupt keyword must not be used when SYS/BIOS Hwi objects are used in conjunction with C functions. The Hwi_enter/Hwi_exit macros and the Hwi dispatcher already contain this functionality, and the use of the C modifier can cause unwanted conflicts. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 107 Copyright © 2017, Texas Instruments Incorporated Keywords www.ti.com 6.5.4 The restrict Keyword To help the compiler determine memory dependencies, you can qualify a pointer, reference, or array with the restrict keyword. The restrict keyword is a type qualifier that can be applied to pointers, references, and arrays. Its use represents a guarantee by you, the programmer, that within the scope of the pointer declaration the object pointed to can be accessed only by that pointer. Any violation of this guarantee renders the program undefined. This practice helps the compiler optimize certain sections of code because aliasing information can be more easily determined. In Example 6-2, the restrict keyword is used to tell the compiler that the function func1 is never called with the pointers a and b pointing to objects that overlap in memory. You are promising that accesses through a and b will never conflict; therefore, a write through one pointer cannot affect a read from any other pointers. The precise semantics of the restrict keyword are described in the 1999 version of the ANSI/ISO C Standard. The "restrict" keyword is a C99 keyword, and cannot be accepted in strict ANSI C89 mode. Use the "__restrict" keyword if the strict ANSI C89 mode must be used. See Section 6.13. Example 6-2. Use of the restrict Type Qualifier With Pointers void func1(int * restrict a, int * restrict b) { /* func1's code here */ } Example 6-3 illustrates using the restrict keyword when passing arrays to a function. Here, the arrays c and d must not overlap, nor may c and d point to the same array. Example 6-3. Use of the restrict Type Qualifier With Arrays void func2(int c[restrict], int d[restrict]) { int i; for(i = 0; i < 64; i++) { c[i] += d[i]; d[i] += 1; } } At this time the restrict keyword is useful only for FPU targets. For non-FPU targets restrict is ignored. 6.5.5 The volatile Keyword The C/C++ compiler supports the volatile keyword in all modes. In addition, the __volatile keyword is supported in relaxed ANSI mode for C89, C99, and C++. The volatile keyword indicates to the compiler that there is something about how the variable is accessed that requires that the compiler not use overly-clever optimization on expressions involving that variable. For example, the variable may also be accessed by an external program, an interrupt, another thread, or a peripheral device. The compiler eliminates redundant memory accesses whenever possible, using data flow analysis to figure out when it is legal. However, some memory accesses may be special in some way that the compiler cannot see, and in such cases you should use the volatile keyword to prevent the compiler from optimizing away something important. The compiler does not optimize out any accesses to variables declared volatile. The number of volatile reads and writes will be exactly as they appear in the C/C++ code, no more and no less and in the same order. 108 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Keywords Any variable which might be modified by something external to the obvious control flow of the program (such as an interrupt service routine) must be declared volatile. This tells the compiler that an interrupt function might modify the value at any time, so the compiler should not perform optimizations which will change the number or order of accesses of that variable. This is the primary purpose of the volatile keyword. In the following example, the loop intends to wait for a location to be read as 0xFF: unsigned int *ctrl; while (*ctrl !=0xFF); However, in this example, *ctrl is a loop-invariant expression, so the loop is optimized down to a singlememory read. To get the desired result, define ctrl as: volatile unsigned int *ctrl; Here the *ctrl pointer is intended to reference a hardware location, such as an interrupt flag. The volatile keyword must also be used when accessing memory locations that represent memorymapped peripheral devices. Such memory locations might change value in ways that the compiler cannot predict. These locations might change if accessed, or when some other memory location is accessed, or when some signal occurs. Volatile must also be used for local variables in a function which calls setjmp, if the value of the local variables needs to remain valid if a longjmp occurs. Example 6-4. Volatile for Local Variables With setjmp #include jmp_buf context; void function() { volatile int x = 3; switch(setjmp(context)) { case 0: setup(); break; default: { printf("x == %d\n", x); /* We can only reach here if longjmp has occurred; because x's lifetime begins before the setjmp and lasts through the longjmp, the C standard requires x be declared "volatile" */ break; } } } The --unified_memory option can be used if your memory map is configured as a single unified space; this option allows the compiler to generate more efficient instructions for most memcpy calls and structure assignments. Even under unified memory, memory for some peripherals and some RAM associated with those peripherals is allocated only in data memory. If –unified_memory is enabled, you can prevent program memory address access to specific symbols by declaring those symbols as volatile. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 109 Copyright © 2017, Texas Instruments Incorporated C++ Exception Handling www.ti.com 6.6 C++ Exception Handling The compiler supports all the C++ exception handling features as defined by the ANSI/ISO 14882 C++ Standard. More details are discussed in The C++ Programming Language, Third Edition by Bjarne Stroustrup. The compiler --exceptions option enables exception handling. The compiler’s default is no exception handling support. For exceptions to work correctly, all C++ files in the application must be compiled with the --exceptions option, regardless of whether exceptions occur in a particular file. Mixing exception-enabled object files and libraries with object files and libraries that do not have exceptions enabled can lead to undefined behavior. Exception handling requires support in the run-time-support library, which come in exception-enabled and exception-disabled forms; you must link with the correct form. When using automatic library selection (the default), the linker automatically selects the correct library Section 4.3.1.1. If you select the library manually, you must use run-time-support libraries whose name contains _eh if you enable exceptions. Using the --exceptions option causes the compiler to insert exception handling code. This code will increase the size of the programand execution time, even if no exceptions are thrown. See Section 8.1 for details on the run-time libraries. 6.7 Register Variables and Parameters The C/C++ compiler treats register variables (variables defined with the register keyword) differently, depending on whether you use the --opt_level (-O) option. • Compiling with optimization The compiler ignores any register definitions and allocates registers to variables and temporary values by using an algorithm that makes the most efficient use of registers. • Compiling without optimization If you use the register keyword, you can suggest variables as candidates for allocation into registers. The compiler uses the same set of registers for allocating temporary expression results as it uses for allocating register variables. The compiler attempts to honor all register definitions. If the compiler runs out of appropriate registers, it frees a register by moving its contents to memory. If you define too many objects as register variables, you limit the number of registers the compiler has for temporary expression results. This limit causes excessive movement of register contents to memory. Any object with a scalar type (integral, floating point, or pointer) can be defined as a register variable. The register designator is ignored for objects of other types, such as arrays. The register storage class is meaningful for parameters as well as local variables. Normally, in a function, some of the parameters are copied to a location on the stack where they are referenced during the function body. The compiler copies a register parameter to a register instead of the stack, which speeds access to the parameter within the function. For more information about register conventions, see Section 7.2. 110 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com The __asm Statement 6.8 The __asm Statement The C/C++ compiler can embed assembly language instructions or directives directly into the assembly language output of the compiler. This capability is an extension to the C/C++ language implemented through the __asm keyword. The __asm keyword provides access to hardware features that C/C++ cannot provide. The alternate keyword, "asm", may also be used except in strict ANSI C mode. It is available in relaxed C and C++ modes. Using __asm is syntactically performed as a call to a function named __asm, with one string constant argument: __asm(" assembler text "); The compiler copies the argument string directly into your output file. The assembler text must be enclosed in double quotes. All the usual character string escape codes retain their definitions. For example, you can insert a .byte directive that contains quotes as follows: __asm("STR: .byte \"abc\""); The inserted code must be a legal assembly language statement. Like all assembly language statements, the line of code inside the quotes must begin with a label, a blank, a tab, or a comment (asterisk or semicolon). The compiler performs no checking on the string; if there is an error, the assembler detects it. For more information about the assembly language statements, see the TMS320C28x Assembly Language Tools User's Guide. The __asm statements do not follow the syntactic restrictions of normal C/C++ statements. Each can appear as a statement or a declaration, even outside of blocks. This is useful for inserting directives at the very beginning of a compiled module. The __asm statement does not provide any way to refer to local variables. If your assembly code needs to refer to local variables, you will need to write the entire function in assembly code. For more information, refer to Section 7.5.5. NOTE: Avoid Disrupting the C/C++ Environment With asm Statements Be careful not to disrupt the C/C++ environment with __asm statements. The compiler does not check the inserted instructions. Inserting jumps and labels into C/C++ code can cause unpredictable results in variables manipulated in or around the inserted code. Directives that change sections or otherwise affect the assembly environment can also be troublesome. Be especially careful when you use optimization with __asm statements. Although the compiler cannot remove __asm statements, it can significantly rearrange the code order near them and cause undesired results. NOTE: Use Single asm Statement for the RPT Instruction When adding a C28x RPT instruction, do not use a separate asm statement for RPT and the repeated instruction. The compiler could insert debug directives between asm directives and the assembler does not allow any directives between the RPT and the repeated instruction. For example, to insert a RPT MAC instruction, use the following: asm("\tRPT #10\n\t||MAC P, *XAR4++, *XAR7++"); SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 111 Copyright © 2017, Texas Instruments Incorporated Pragma Directives www.ti.com 6.9 Pragma Directives Pragma directives tell the compiler how to treat a certain function, object, or section of code. The C28x C/C++ compiler supports the following pragmas: • CALLS (See Section 6.9.1) • CHECK_MISRA (See Section 6.9.2) • CLINK (See Section 6.9.3) • CODE_ALIGN (See Section 6.9.4) • CODE_SECTION (See Section 6.9.5) • DATA_ALIGN (See Section 6.9.6) • DATA_SECTION (See Section 6.9.7) • diag_suppress, diag_remark, diag_warning, diag_error, diag_default, diag_push, diag_pop (See Section 6.9.8) • FAST_FUNC_CALL (See Section 6.9.9) • FUNC_ALWAYS_INLINE (See Section 6.9.10) • FUNC_CANNOT_INLINE (See Section 6.9.11) • FUNC_EXT_CALLED (See Section 6.9.12) • FUNCTION_OPTIONS (See Section 6.9.13) • INTERRUPT (See Section 6.9.14) • MUST_ITERATE (See Section 6.9.15) • NO_HOOKS (See Section 6.9.16) • RESET_MISRA (See Section 6.9.17) • RETAIN (See Section 6.9.18) • SET_CODE_SECTION (See Section 6.9.19) • SET_DATA_SECTION (See Section 6.9.19) • UNROLL (See Section 6.9.20) The arguments func and symbol cannot be defined or declared inside the body of a function. You must specify the pragma outside the body of a function; and the pragma specification must occur before any declaration, definition, or reference to the func or symbol argument. If you do not follow these rules, the compiler issues a warning and may ignore the pragma. For pragmas that apply to functions or symbols, the syntax differs between C and C++. • In C, you must supply the name of the object or function to which you are applying the pragma as the first argument. Because the entity operated on is specified, a pragma in C can appear some distance way from the definition of that entity. • In C++, pragmas are positional. They do not name the entity on which they operate as an argument. Instead, they always operate on the next entity defined after the pragma. 112 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Pragma Directives 6.9.1 The CALLS Pragma The CALLS pragma specifies a set of functions that can be called indirectly from a specified calling function. The CALLS pragma is used by the compiler to embed debug information about indirect calls in object files. Using the CALLS pragma on functions that make indirect calls enables such indirect calls to be included in calculations for such functions' inclusive stack sizes. For more information on generating function stack usage information, see the -cg option of the Object File Display Utility in the "Invoking the Object File Display Utility" section of the TMS320C28x Assembly Language Tools User's Guide. The CALLS pragma can precede either the calling function's definition or its declaration. In C, the pragma must have at least 2 arguments—the first argument is the calling function, followed by at least one function that will be indirectly called from the calling function. In C++, the pragma applies to the next function declared or defined, and the pragma must have at least one argument. The syntax for the CALLS pragma in C is as follows. This indicates that calling_function can indirectly call function_1 through function_n. #pragma CALLS ( calling_function, function_1, function_2, ..., function_n ) The syntax for the CALLS pragma in C++ is: #pragma CALLS ( function_1_mangled_name, ..., function_n_mangled_name ) Note that in C++, the arguments to the CALLS pragma must be the full mangled names for the functions that can be indirectly called from the calling function. The GCC-style "calls" attribute syntax, which has the same effect as the CALLS pragma, is as follows: __attribute__((calls("function_1","function_2",..., "function_n"))) 6.9.2 The CHECK_MISRA Pragma The CHECK_MISRA pragma enables/disables MISRA C:2004 rules at the source level. The compiler option --check_misra must be used to enable checking in order for this pragma to function at the source level. The syntax of the pragma in C is: #pragma CHECK_MISRA (" {all|required|advisory|none|rulespec} ") The rulespec parameter is a comma-separated list of rule numbers. See Section 6.3 for details. The RESET_MISRA pragma can be used to reset any CHECK_MISRA pragmas; see Section 6.9.17. 6.9.3 The CLINK Pragma The CLINK pragma can be applied to a code or data symbol. It causes a .clink directive to be generated into the section that contains the definition of the symbol. The .clink directive tells the linker that a section is eligible for removal during conditional linking. Thus, if the section is not referenced by any other section in the application being compiled and linked, it will not be included in the resulting output file. The syntax of the pragma in C is: #pragma CLINK (symbol ) The syntax of the pragma in C++ is: #pragma CLINK SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 113 Copyright © 2017, Texas Instruments Incorporated Pragma Directives www.ti.com The RETAIN pragma has the opposite effect of the CLINK pragma. See Section 6.9.18 for more details. 6.9.4 The CODE_ALIGN Pragma The CODE_ALIGN pragma aligns func along the specified alignment. The alignment constant must be a power of 2. The CODE_ALIGN pragma is useful if you have functions that you want to start at a certain boundary. The syntax of the pragma in C is: #pragma CODE_ALIGN ( func, constant ) The syntax of the pragma in C++ is: #pragma CODE_ALIGN ( constant ) 6.9.5 The CODE_SECTION Pragma The CODE_SECTION pragma allocates space for the symbol in C, or the next symbol declared in C++, in a section named section name. The syntax of the pragma in C is: #pragma CODE_SECTION (symbol , "section name ") The syntax of the pragma in C++ is: #pragma CODE_SECTION (" section name ") The CODE_SECTION pragma is useful if you have code objects that you want to link into an area separate from the .text section. The following example demonstrates the use of the CODE_SECTION pragma. Example 6-5. Using the CODE_SECTION Pragma C Source File char bufferA[80]; char bufferB[80]; #pragma CODE_SECTION(funcA, "codeA") char funcA(int i); char funcB(int i); void main() { char c; c = funcA(1); c = funcB(2); } char funcA (int i) { return bufferA[i]; } char funcB (int j) { return bufferB[j]; } 114 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Example 6-6. Generated Assembly Code From Example 6-5 Pragma Directives .sect ".text" .global _main; **************************************************************** ;* FNAME: _main FR SIZE: 2 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 1 Auto, 0 SOE * ;*************************************************************** :_main: ADDB SP,#2 MOVB AL,#1 ; |12| LCR #_funcA ; |12| ; call occurs [#_funcA] ; |12| MOV *-SP[1],AL ; |12| MOVB AL,#1 ; |13| LCR #_funcB ; |13| ; call occurs [#_funcB] ; |13| MOV *-SP[1],AL ; |13| SUBB SP,#2 LRETR ; return occurs .sect "codeA" .global _funcA ;*************************************************************** ;* FNAME: _funcA FR SIZE: 1 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 1 Auto, 0 SOE * ;*************************************************************** _funcA: ADDB SP,#1 MOV *-SP[1],AL MOVZ AR6,*-SP[1] ADD AR6,#_bufferA SUBB SP,#1 MOV AL,*+XAR6[0] LRETR ;return occurs ; |17| ; |18| ; |18| ; |18| ; |18| .sect ".text" .global _funcB; **************************************************************** ;* FNAME: _funcB FR SIZE: 1 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 1 Auto, 0 SOE * ;*************************************************************** _funcB: ADDB MOV MOVZ SP,#1 *-SP[1],AL AR6,*-SP[1] ; |22| ; |23| SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 115 Copyright © 2017, Texas Instruments Incorporated Pragma Directives Example 6-6. Generated Assembly Code From Example 6-5 (continued) ADD AR6,#_bufferB SUBB SP,#1 MOV AL,*+XAR6[0] LRETR ;return occurs ; |23| ; |23| ; |23| www.ti.com 6.9.6 The DATA_ALIGN Pragma The DATA_ALIGN pragma aligns the symbol in C, or the next symbol declared in C++, to an alignment boundary. The alignment boundary is the maximum of the symbol's default alignment value or the value of the constant in bytes. The constant must be a power of 2. The maximum alignment is 32768. The DATA_ALIGN pragma cannot be used to reduce an object's natural alignment. The syntax of the pragma in C is: #pragma DATA_ALIGN ( symbol , constant ) The syntax of the pragma in C++ is: #pragma DATA_ALIGN ( constant ) 116 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Pragma Directives 6.9.7 The DATA_SECTION Pragma The DATA_SECTION pragma allocates space for the symbol in C, or the next symbol declared in C++, in a section named section name. The syntax of the pragma in C is: #pragma DATA_SECTION ( symbol , " section name ") The syntax of the pragma in C++ is: #pragma DATA_SECTION (" section name ") The DATA_SECTION pragma is useful if you have data objects that you want to link into an area separate from the .ebss section. Example 6-7 through Example 6-9 demonstrate the use of the DATA_SECTION pragma. Example 6-7. Using the DATA_SECTION Pragma C Source File #pragma DATA_SECTION(bufferB, "my_sect") char bufferA[512]; char bufferB[512]; Example 6-8. Using the DATA_SECTION Pragma C++ Source File char bufferA[512]; #pragma DATA_SECTION("my_sect") char bufferB[512]; Example 6-9. Using the DATA_SECTION Pragma Assembly Source File _bufferB: .global _bufferA .ebss _bufferA,512,4 .global _bufferB .usect "my_sect",512,4 SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 117 Copyright © 2017, Texas Instruments Incorporated Pragma Directives 6.9.8 The Diagnostic Message Pragmas The following pragmas can be used to control diagnostic messages in the same ways as the corresponding command line options: www.ti.com Pragma diag_suppress num diag_remark num diag_warning num diag_error num diag_default num diag_push diag_pop Option -pds=num[, num2, num3...] -pdsr=num[, num2, num3...] -pdsw=num[, num2, num3...] -pdse=num[, num2, num3...] n/a n/a n/a Description Suppress diagnostic num Treat diagnostic num as a remark Treat diagnostic num as a warning Treat diagnostic num as an error Use default severity of the diagnostic Push the current diagnostics severity state to store it for later use. Pop the most recent diagnostic severity state stored with #pragma diag_push to be the current setting. The syntax of the diag_suppress, diag_remark, diag_warning, and diag_error pragmas in C is: #pragma diag_xxx [=]num[, num2, num3...] Notice that the names of these pragmas are in lowercase. The diagnostic affected (num) is specified using either an error number or an error tag name. The equal sign (=) is optional. Any diagnostic can be overridden to be an error, but only diagnostic messages with a severity of discretionary error or below can have their severity reduced to a warning or below, or be suppressed. The diag_default pragma is used to return the severity of a diagnostic to the one that was in effect before any pragmas were issued (i.e., the normal severity of the message as modified by any command-line options). The diagnostic identifier number is output with the message when you use the -pden command line option. The following example suppresses a diagnostic message and then restores the previous diagnostics severity state: #pragma diag_push #pragma diag_suppress 551 #pragma CHECK_MISRA("-9.1") #pragma diag_pop 6.9.9 The FAST_FUNC_CALL Pragma The FAST_FUNC_CALL pragma, when applied to a function, generates a TMS320C28x FFC instruction to call the function instead of the CALL instruction. Refer to the TMS320C28x DSP CPU and Instruction Set User’s Guide for more details on the FFC instruction. The syntax of the pragma in C is: #pragma FAST_FUNC_CALL ( func ) The syntax of the pragma in C++ is: #pragma FAST_FUNC_CALL ( func ) The FAST_FUNC_CALL pragma should be applied only to a call to an assembly function that returns with the LB *XAR7 instruction. See Section 7.5.1 for information on combining C/C++ and assembly code. Since this pragma should be applied only to assembly functions, if the compiler finds a definition for func in the file scope, it issues a warning and ignores the pragma. 118 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com The following example demonstrates the use of the FAST_FUNC_CALL pragma. Example 6-10. Using the FAST_FUNC_CALL Pragma Assembly Function _add_long: ADD ACC, *-SP[2] LB *XAR7 Example 6-11. Using the FAST_FUNC_CALL Pragma C Source File #pragma FAST_FUNC_CALL (add_long) long add_long(long, long); void foo() { long x, y; x = 0xffff; y = 0xff; y = add_long(x, y); } Example 6-12. Generated Assembly File ;*************************************************************** ;* FNAME: _foo FR SIZE: 6 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 2 Parameter, 4 Auto, 0 SOE * ;*************************************************************** _foo: ADDB SP,#6 MOVB ACC,#255 MOVL XAR6,#65535 MOVL *-SP[6],ACC MOVL *-SP[2],ACC MOVL *-SP[4],XAR6 MOVL ACC,*-SP[4] FFC XAR7,#_add_long ; call occurs [#_add_long] MOVL *-SP[6],ACC SUBB SP,#6 LRETR ; return occurs ; |8| ; |10| ; |8| ; |10| ; |10| ; |10| ; |10| Pragma Directives SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 119 Copyright © 2017, Texas Instruments Incorporated Pragma Directives www.ti.com 6.9.10 The FUNC_ALWAYS_INLINE Pragma The FUNC_ALWAYS_INLINE pragma instructs the compiler to always inline the named function. The compiler only inlines the function if it is legal to inline the function and the compiler is invoked with any level of optimization (--opt_level=0). See Section 2.11 for details about interaction between various types of inlining. This pragma must appear before any declaration or reference to the function that you want to inline. In C, the argument func is the name of the function that will be inlined. In C++, the pragma applies to the next function declared. The syntax of the pragma in C is: #pragma FUNC_ALWAYS_INLINE ( func ) The syntax of the pragma in C++ is: #pragma FUNC_ALWAYS_INLINE The following example uses this pragma: #pragma FUNC_ALWAYS_INLINE(functionThatMustGetInlined) static inline void functionThatMustGetInlined(void) { P1OUT |= 0x01; P1OUT &= ~0x01; } Use Caution with the FUNC_ALWAYS_INLINE Pragma NOTE: The FUNC_ALWAYS_INLINE pragma overrides the compiler's inlining decisions. Overuse of this pragma could result in increased compilation times or memory usage, potentially enough to consume all available memory and result in compilation tool failures. 6.9.11 The FUNC_CANNOT_INLINE Pragma The FUNC_CANNOT_INLINE pragma instructs the compiler that the named function cannot be expanded inline. Any function named with this pragma overrides any inlining you designate in any other way, such as using the inline keyword. Automatic inlining is also overridden with this pragma; see Section 2.11. The pragma must appear before any declaration or reference to the function that you want to keep. In C, the argument func is the name of the function that cannot be inlined. In C++, the pragma applies to the next function declared. The syntax of the pragma in C is: #pragma FUNC_CANNOT_INLINE ( func ) The syntax of the pragma in C++ is: #pragma FUNC_CANNOT_INLINE 120 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Pragma Directives 6.9.12 The FUNC_EXT_CALLED Pragma When you use the --program_level_compile option, the compiler uses program-level optimization. When you use this type of optimization, the compiler removes any function that is not called, directly or indirectly, by main(). You might have C/C++ functions that are called by hand-coded assembly instead of main(). The FUNC_EXT_CALLED pragma specifies that the optimizer should keep these C functions or any functions these C/C++ functions call. These functions act as entry points into C/C++. The pragma must appear before any declaration or reference to the function to keep. In C, the argument func is the name of the function to keep. In C++, the pragma applies to the next function declared. The syntax of the pragma in C is: #pragma FUNC_EXT_CALLED ( func ) The syntax of the pragma in C++ is: #pragma FUNC_EXT_CALLED Except for _c_int00, which is the name reserved for the system reset interrupt for C/C++programs, the name of the interrupt (the func argument) does not need to conform to a naming convention. When you use program-level optimization, you may need to use the FUNC_EXT_CALLED pragma with certain options. See Section 3.3.2. 6.9.13 The FUNCTION_OPTIONS Pragma The FUNCTION_OPTIONS pragma allows you to compile a specific function in a C or C++ file with additional command-line compiler options. The affected function will be compiled as if the specified list of options appeared on the command line after all other compiler options. In C, the pragma is applied to the function specified. In C++, the pragma is applied to the next function. The syntax of the pragma in C is: #pragma FUNCTION_OPTIONS ( func, "additional options" ) The syntax of the pragma in C++ is: #pragma FUNCTION_OPTIONS( "additional options" ) Supported options for this pragma are --opt_level, --auto_inline, --code_state, --opt_for_space, and -opt_for_speed. In order to use --opt_level and --auto_inline with the FUNCTION_OPTIONS pragma, the compiler must be invoked with some optimization level (that is, at least --opt_level=0). 6.9.14 The INTERRUPT Pragma The INTERRUPT pragma enables you to handle interrupts directly with C code. In C, the argument func is the name of a function. In C++, the pragma applies to the next function declared. The syntax of the pragma in C is: #pragma INTERRUPT ( func ) The syntax of the pragma in C++ is: #pragma INTERRUPT void func( void ) The GCC interrupt attribute syntax, which has the same effects as the INTERRUPT pragma, is as follows. Note that the interrupt attribute can precede either the function's definition or its declaration. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 121 Copyright © 2017, Texas Instruments Incorporated Pragma Directives www.ti.com __attribute__((interrupt)) void func( void ) On the FPU, there are two kinds of interrupts - High Priority Interrupt (HPI) and Low Priority Interrupt (LPI). High priority interrupts use a fast context save and cannot be nested. Low priority interrupts behave like normal C28x interrupts and can be nested. The kind of interrupt can be specified by way of the interrupt pragma using an optional second argument. The C syntax of the pragma is: #pragma INTERRUPT ( func , {HPI|LPI} ) The syntax of the pragma in C++ is: #pragma INTERRUPT ( {HPI|LPI} ) The syntax of the GCC attribute, which has the same effects as the INTERRUPT pragma, is: __attribute__((interrupt( "HPI"|"LPI" ))) void func( void ) { ... } On FPU, if no interrupt priority is specified LPI is assumed. Interrupts specified with the interrupt keyword also default to LPI. CLA interrupts and CLA2 background tasks can be created using the interrupt attribute or the INTERRUPT pragma. For example the following could both be used with a CLA interrupt: __attribute__((interrupt)) void interrupt_name(void) {...} #pragma INTERRUPT(interrupt _name); void interrupt _name(void) {...} The following examples both create a CLA2 background task: __attribute__((interrupt("BACKGROUND"))) void task_name(void) {...} #pragma INTERRUPT(task_name, "BACKGROUND"); void task_name(void) {...} Hwi Objects and the INTERRUPT Pragma NOTE: The INTERRUPT pragma must not be used when SYS/BIOS Hwi objects are used in conjunction with C functions. The Hwi_enter/Hwi_exit macros and the Hwi dispatcher contain this functionality, and the use of the C modifier can cause negative results. 122 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Pragma Directives 6.9.15 The MUST_ITERATE Pragma The MUST_ITERATE pragma specifies to the compiler certain properties of a loop. When you use this pragma, you are guaranteeing to the compiler that a loop executes a specific number of times or a number of times within a specified range. Any time the UNROLL pragma is applied to a loop, MUST_ITERATE should be applied to the same loop. For loops the MUST_ITERATE pragma's third argument, multiple, is the most important and should always be specified. Furthermore, the MUST_ITERATE pragma should be applied to any other loops as often as possible. This is because the information provided via the pragma (especially the minimum number of iterations) aids the compiler in choosing the best loops and loop transformations (that is, nested loop transformations). It also helps the compiler reduce code size. No statements are allowed between the MUST_ITERATE pragma and the for, while, or do-while loop to which it applies. However, other pragmas, such as UNROLL, can appear between the MUST_ITERATE pragma and the loop. 6.9.15.1 The MUST_ITERATE Pragma Syntax The syntax of the pragma for C and C++ is: #pragma MUST_ITERATE ( min, max, multiple ) The arguments min and max are programmer-guaranteed minimum and maximum trip counts. The trip count is the number of times a loop iterates. The trip count of the loop must be evenly divisible by multiple. All arguments are optional. For example, if the trip count could be 5 or greater, you can specify the argument list as follows: #pragma MUST_ITERATE(5) However, if the trip count could be any nonzero multiple of 5, the pragma would look like this: #pragma MUST_ITERATE(5, , 5) /* Note the blank field for max */ It is sometimes necessary for you to provide min and multiple in order for the compiler to perform unrolling. This is especially the case when the compiler cannot easily determine how many iterations the loop will perform (that is, the loop has a complex exit condition). When specifying a multiple via the MUST_ITERATE pragma, results of the program are undefined if the trip count is not evenly divisible by multiple. Also, results of the program are undefined if the trip count is less than the minimum or greater than the maximum specified. If no min is specified, zero is used. If no max is specified, the largest possible number is used. If multiple MUST_ITERATE pragmas are specified for the same loop, the smallest max and largest min are used. 6.9.15.2 Using MUST_ITERATE to Expand Compiler Knowledge of Loops Through the use of the MUST_ITERATE pragma, you can guarantee that a loop executes a certain number of times. The example below tells the compiler that the loop is guaranteed to run exactly 10 times: #pragma MUST_ITERATE(10,10) for(i = 0; i < trip_count; i++) { ... In this example, the compiler attempts to generate a loop even without the pragma. However, if MUST_ITERATE is not specified for a loop such as this, the compiler generates code to bypass the loop, to account for the possibility of 0 iterations. With the pragma specification, the compiler knows that the loop iterates at least once and can eliminate the loop-bypassing code. MUST_ITERATE can specify a range for the trip count as well as a factor of the trip count. The following example tells the compiler that the loop executes between 8 and 48 times and the trip_count variable is a multiple of 8 (8, 16, 24, 32, 40, 48). The multiple argument allows the compiler to unroll the loop. #pragma MUST_ITERATE(8, 48, 8) for(i = 0; i < trip_count; i++) { ... SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 123 Copyright © 2017, Texas Instruments Incorporated Pragma Directives www.ti.com You should consider using MUST_ITERATE for loops with complicated bounds. In the following example, the compiler would have to generate a divide function call to determine, at run time, the number of iterations performed. for(i2 = ipos[2]; i2 < 40; i2 += 5) { ... The compiler will not do the above. In this case, using MUST_ITERATE to specify that the loop always executes eight times allows the compiler to attempt to generate a loop: #pragma MUST_ITERATE(8, 8) for(i2 = ipos[2]; i2 < 40; i2 += 5) { ... 6.9.16 The NO_HOOKS Pragma The NO_HOOKS pragma prevents entry and exit hook calls from being generated for a function. The syntax of the pragma in C is: #pragma NO_HOOKS ( func ) The syntax of the pragma in C++ is: #pragma NO_HOOKS See Section 2.13 for details on entry and exit hooks. 6.9.17 The RESET_MISRA Pragma The RESET_MISRA pragma resets the specified MISRA C:2004 rules to the state they were before any CHECK_MISRA pragmas (see Section 6.9.2) were processed. For instance, if a rule was enabled on the command line but disabled in the source, the RESET_MISRA pragma resets it to enabled. This pragma accepts the same format as the --check_misra option, except for the "none" keyword. The --check_misra compiler command-line option must be used to enable MISRA C:2004 rule checking in order for this pragma to function at the source level. The syntax of the pragma in C is: #pragma RESET_MISRA (" {all|required|advisory|rulespec} ") The rulespec parameter is a comma-separated list of rule numbers. See Section 6.3 for details. 6.9.18 The RETAIN Pragma The RETAIN pragma can be applied to a code or data symbol. It causes a .retain directive to be generated into the section that contains the definition of the symbol. The .retain directive indicates to the linker that the section is ineligible for removal during conditional linking. Therefore, regardless whether or not the section is referenced by another section in the application that is being compiled and linked, it will be included in the output file result of the link. The syntax of the pragma in C is: #pragma RETAIN ( symbol ) The syntax of the pragma in C++ is: #pragma RETAIN The CLINK pragma has the opposite effect of the RETAIN pragma. See Section 6.9.3 for more details. 124 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com 6.9.19 The SET_CODE_SECTION and SET_DATA_SECTION Pragmas These pragmas can be used to set the section for all declarations below the pragma. The syntax of the pragmas in C/C++ is: #pragma SET_CODE_SECTION ("section name") Pragma Directives #pragma SET_DATA_SECTION ("section name") In Example 6-13 x and y are put in the section mydata. To reset the current section to the default used by the compiler, a blank parameter should be passed to the pragma. An easy way to think of the pragma is that it is like applying the CODE_SECTION or DATA_SECTION pragma to all symbols below it. Example 6-13. Setting Section With SET_DATA_SECTION Pragma #pragma SET_DATA_SECTION("mydata") int x; int y; #pragma SET_DATA_SECTION() The pragmas apply to both declarations and definitions. If applied to a declaration and not the definition, the pragma that is active at the declaration is used to set the section for that symbol. Here is an example: Example 6-14. Setting a Section With SET_CODE_SECTION Pragma #pragma SET_CODE_SECTION("func1") extern void func1(); #pragma SET_CODE_SECTION() ... void func1() { ... } In Example 6-14 func1 is placed in section func1. If conflicting sections are specified at the declaration and definition, a diagnostic is issued. The current CODE_SECTION and DATA_SECTION pragmas and GCC attributes can be used to override the SET_CODE_SECTION and SET_DATA_SECTION pragmas. For example: Example 6-15. Overriding SET_DATA_SECTION Setting #pragma DATA_SECTION(x, "x_data") #pragma SET_DATA_SECTION("mydata") int x; int y; #pragma SET_DATA_SECTION() In Example 6-15 x is placed in x_data and y is placed in mydata. No diagnostic is issued for this case. The pragmas work for both C and C++. In C++, the pragmas are ignored for templates and for implicitly created objects, such as implicit constructors and virtual function tables. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 125 Copyright © 2017, Texas Instruments Incorporated Pragma Directives www.ti.com 6.9.20 The UNROLL Pragma No statements are allowed between the UNROLL pragma and the for, while, or do-while loop to which it applies. However, other pragmas, such as MUST_ITERATE, can appear between the UNROLL pragma and the loop. The syntax of the pragma for C and C++ is: #pragma UNROLL( n ) If possible, the compiler unrolls the loop so there are n copies of the original loop. The compiler only unrolls if it can determine that unrolling by a factor of n is safe. In order to increase the chances the loop is unrolled, the compiler needs to know certain properties: • The loop iterates a multiple of n times. This information can be specified to the compiler via the multiple argument in the MUST_ITERATE pragma. • The smallest possible number of iterations of the loop • The largest possible number of iterations of the loop The compiler can sometimes obtain this information itself by analyzing the code. However, sometimes the compiler can be overly conservative in its assumptions and therefore generates more code than is necessary when unrolling. This can also lead to not unrolling at all. Furthermore, if the mechanism that determines when the loop should exit is complex, the compiler may not be able to determine these properties of the loop. In these cases, you must tell the compiler the properties of the loop by using the MUST_ITERATE pragma. Specifying #pragma UNROLL(1) asks that the loop not be unrolled. Automatic loop unrolling also is not performed in this case. If multiple UNROLL pragmas are specified for the same loop, it is undefined which pragma is used, if any. 6.10 The _Pragma Operator The C28x C/C++ compiler supports the C99 preprocessor _Pragma() operator. This preprocessor operator is similar to #pragma directives. However, _Pragma can be used in preprocessing macros (#defines). The syntax of the operator is: _Pragma (" string_literal "); The argument string_literal is interpreted in the same way the tokens following a #pragma directive are processed. The string_literal must be enclosed in quotes. A quotation mark that is part of the string_literal must be preceded by a backward slash. You can use the _Pragma operator to express #pragma directives in macros. For example, the DATA_SECTION syntax: #pragma DATA_SECTION( func ," section ") Is represented by the _Pragma() operator syntax: _Pragma ("DATA_SECTION( func ,\" section \")") The following code illustrates using _Pragma to specify the DATA_SECTION pragma in a macro: ... #define EMIT_PRAGMA(x) _Pragma(#x) #define COLLECT_DATA(var) EMIT_PRAGMA(DATA_SECTION(var,"mysection")) COLLECT_DATA(x) int x; ... The EMIT_PRAGMA macro is needed to properly expand the quotes that are required to surround the section argument to the DATA_SECTION pragma. 126 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Object File Symbol Naming Conventions (Linknames) 6.11 Object File Symbol Naming Conventions (Linknames) Each externally visible identifier is assigned a unique symbol name to be used in the object file, a socalled linkname. This name is assigned by the compiler according to an algorithm which depends on the name, type, and source language of the symbol. This algorithm may add a prefix to the identifier (typically an underscore), and it may mangle the name. For COFF, the compiler places an underscore at the beginning of the linknames of C identifiers, so you can safely use identifiers that do not begin with an underscore in your assembly code. Name mangling encodes the types of the parameters of a function in the linkname for a function. Name mangling only occurs for C++ functions which are not declared 'extern "C"'. Mangling allows function overloading, operator overloading, and type-safe linking. Be aware that the return value of the function is not encoded in the mangled name, as C++ functions cannot be overloaded based on the return value. The mangling algorithm used closely follows that described in The Annotated Reference Manual (ARM). For example, the general form of a C++ linkname for a function named func is: _func__F parmcodes Where parmcodes is a sequence of letters that encodes the parameter types of func. For this simple C++ source file: int foo(int i){ } //global C++ function This is the resulting assembly code: _foo__Fi The linkname of foo is _foo__Fi, indicating that foo is a function that takes a single argument of type int. To aid inspection and debugging, a name demangling utility is provided that demangles names into those found in the original C++ source. See Chapter 9 for more information. 6.12 Initializing Static and Global Variables The ANSI/ISO C standard specifies that global (extern) and static variables without explicit initializations must be initialized to 0 before the program begins running. This task is typically done when the program is loaded. Because the loading process is heavily dependent on the specific environment of the target application system, the compiler itself makes no provision for initializing to 0 otherwise uninitialized static storage class variables at run time. It is up to your application to fulfill this requirement. Initialize Global Objects NOTE: You should explicitly initialize all global objects which you expected the compiler would set to zero by default. 6.12.1 Initializing Static and Global Variables With the Linker If your loader does not preinitialize variables, you can use the linker to preinitialize the variables to 0 in the object file. For example, in the linker command file, use a fill value of 0 in the .ebss section: SECTIONS { ... .ebss: {} = 0x00; ... } Because the linker writes a complete load image of the zeroed .ebss section into the output COFF file, this method can have the unwanted effect of significantly increasing the size of the output file (but not the program). If you burn your application into ROM, you should explicitly initialize variables that require initialization. The preceding method initializes .ebss to 0 only at load time, not at system reset or power up. To make these variables 0 at run time, explicitly define them in your code. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 127 Copyright © 2017, Texas Instruments Incorporated Initializing Static and Global Variables www.ti.com For more information about linker command files and the SECTIONS directive, see the linker description information in the TMS320C28x Assembly Language Tools User's Guide. 6.12.2 Initializing Static and Global Variables With the const Type Qualifier Static and global variables of type const without explicit initializations are similar to other static and global variables because they might not be preinitialized to 0 (for the same reasons discussed in Section 6.12). For example: const int zero; /* might not be initialized to 0 */ However, the initialization of const global and static variables is different because these variables are declared and initialized in a section called .econst. For example: const int zero = 0 /* guaranteed to be 0 */ This corresponds to an entry in the .econst section: .sect _zero .word .econst 0 This feature is particularly useful for declaring a large table of constants, because neither time nor space is wasted at system startup to initialize the table. Additionally, the linker can be used to place the .econst section in ROM. You can use the DATA_SECTION pragma to put the variable in a section other than .econst. For example, the following C code: #pragma DATA_SECTION (var, ".mysect") const int zero=0; is compiled into this assembly code: .sect _zero .word .mysect 0 6.13 Changing the ANSI/ISO C/C++ Language Mode The language mode command-line options determine how the compiler interprets your source code. You specify one option to identify which language standard your code follows. You can also specify a separate option to specify how strictly the compiler should expect your code to conform to the standard. Specify one of the following language options to control the language standard that the compiler expects the source to follow. The options are: • ANSI/ISO C89 (--c89, default for C files) • ANSI/ISO C99 (--c99, see Section 6.13.1.) • ISO C++03 (--c++03, default for C++ files) Use one of the following options to specify how strictly the code conforms to the standard: • Relaxed ANSI/ISO (--relaxed_ansi or -pr) This is the default. • Strict ANSI/ISO (--strict_ansi or -ps) The default is relaxed ANSI/ISO mode. Under relaxed ANSI/ISO mode, the compiler accepts language extensions that could potentially conflict with ANSI/ISO C/C++. Under strict ANSI mode, these language extensions are suppressed so that the compiler will accept all strictly conforming programs. (See Section 6.13.2.) 128 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Changing the ANSI/ISO C/C++ Language Mode 6.13.1 Enabling C99 Mode (--c99) The compiler supports the 1999 standard of C as standardized by the ISO. However, the following list of run-time functions and features are not implemented or fully supported: • complex.h • ctype.h – isblank() • fenv.h • float.h – DECIMAL_DIG – FLT_EVAL_METHOD • inttypes.h – wcstoimax() / wcstoumax() • math.h: The math library used by the compiler in C99 mode has been changed. Full C99 math support is now available, including long double (64-bit) and float versions of floating point math routines. See the list of standard math.h C99 routines. • stdarg.h – va_copy macro • stdio.h – %a and %A format specifiers for hexadecimal float – The %e specifier may produce "-0" when "0" is expected by the standard – snprintf() does not properly pad with spaces when writing to a wide character array • stdlib.h – strtof() atof() / strtod() / strtold() do not support hexadecimal float strings – vfscanf() / vscanf() / vsscanf() return value on floating point matching failure is incorrect • tgmath.h • time.h – strftime() • wchar.h – getws() / fputws() – mbrlen() – mbsrtowcs() – wcscat() – wcschr() – wcscmp() / wcsncmp() – wcscpy() / wcsncpy() – wcsftime() – wcsrtombs() – wcsstr() – wcstok() – wcsxfrm() – Wide character print / scan functions – Wide character conversion functions SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 129 Copyright © 2017, Texas Instruments Incorporated Changing the ANSI/ISO C/C++ Language Mode www.ti.com 6.13.2 Enabling Strict ANSI/ISO Mode and Relaxed ANSI/ISO Mode (--strict_ansi and -relaxed_ansi Options) Under relaxed ANSI/ISO mode (the default), the compiler accepts language extensions that could potentially conflict with a strictly conforming ANSI/ISO C/C++ program. Under strict ANSI mode, these language extensions are suppressed so that the compiler will accept all strictly conforming programs. Use the --strict_ansi option when you know your program is a conforming program and it will not compile in relaxed mode. In this mode, language extensions that conflict with ANSI/ISO C/C++ are disabled and the compiler will emit error messages where the standard requires it to do so. Violations that are considered discretionary by the standard may be emitted as warnings instead. Examples: The following is strictly conforming C code, but will not be accepted by the compiler in the default relaxed mode. To get the compiler to accept this code, use strict ANSI mode. The compiler will suppress the interrupt keyword language exception, and interrupt may then be used as an identifier in the code. int main() { int interrupt = 0; return 0; } The following is not strictly conforming code. The compiler will not accept this code in strict ANSI mode. To get the compiler to accept it, use relaxed ANSI mode. The compiler will provide the interrupt keyword extension and will accept the code interrupt void isr(void); int main() { return 0; } The following code is accepted in all modes. The __interrupt keyword does not conflict with the ANSI/ISO C standard, so it is always available as a language extension. __interrupt void isr(void); int main() { return 0; } The default mode is relaxed ANSI. This mode can be selected with the --relaxed_ansi (or -pr) option. Relaxed ANSI mode accepts the broadest range of programs. It accepts all TI language extensions, even those which conflict with ANSI/ISO, and ignores some ANSI/ISO violations for which the compiler can do something reasonable. The GCC language extensions described in Section 6.14 are available in relaxed ANSI/ISO mode. 130 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com GNU Language Extensions 6.14 GNU Language Extensions The GNU compiler collection (GCC) defines a number of language features not found in the ANSI/ISO C and C++ standards. The definition and examples of these extensions (for GCC version 4.7) can be found at the GNU web site, http://gcc.gnu.org/onlinedocs/gcc-4.7.2/gcc/C-Extensions.html#C-Extensions. Most of these extensions are also available for C++ source code. 6.14.1 Extensions Most of the GCC language extensions are available in the TI compiler when compiling in relaxed ANSI mode (--relaxed_ansi). The extensions that the TI compiler supports are listed in Table 6-3, which is based on the list of extensions found at the GNU web site. The shaded rows describe extensions that are not supported. Table 6-3. GCC Language Extensions Extensions Statement expressions Local labels Labels as values Nested functions Constructing calls Naming types(1) typeof operator Generalized lvalues Conditionals long long Hex floats Complex Zero length Variadic macros Variable length Empty structures Subscripting Escaped newlines Multi-line strings(1) Pointer arithmetic Initializers Compound literals Designated initializers Cast to union Case ranges Mixed declarations Function attributes Attribute syntax Function prototypes C++ comments Dollar signs Character escapes Variable attributes Type attributes Descriptions Putting statements and declarations inside expressions (useful for creating smart 'safe' macros) Labels local to a statement expression Pointers to labels and computed gotos As in Algol and Pascal, lexical scoping of functions Dispatching a call to another function Giving a name to the type of an expression typeof referring to the type of an expression Using question mark (?) and comma (,) and casts in lvalues Omitting the middle operand of a ?: expression Double long word integers and long long int type Hexadecimal floating-point constants Data types for complex numbers Zero-length arrays Macros with a variable number of arguments Arrays whose length is computed at run time Structures with no members Any array can be subscripted, even if it is not an lvalue. Slightly looser rules for escaped newlines String literals with embedded newlines Arithmetic on void pointers and function pointers Non-constant initializers Compound literals give structures, unions, or arrays as values Labeling elements of initializers Casting to union type from any member of the union 'Case 1 ... 9' and such Mixing declarations and code Declaring that functions have no side effects, or that they can never return Formal syntax for attributes Prototype declarations and old-style definitions C++ comments are recognized. A dollar sign is allowed in identifiers. The character ESC is represented as \e Specifying the attributes of variables Specifying the attributes of types (1) Feature defined for GCC 3.0; definition and examples at http://gcc.gnu.org/onlinedocs/gcc-4.7.2/gcc/C-Extensions.html#CExtensions SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 131 Copyright © 2017, Texas Instruments Incorporated GNU Language Extensions Extensions Alignment Inline Assembly labels Extended asm Constraints Alternate keywords Explicit reg vars Incomplete enum types Function names Return address Other built-ins Vector extensions Target built-ins Pragmas Unnamed fields Thread-local Binary constants Table 6-3. GCC Language Extensions (continued) Descriptions Inquiring about the alignment of a type or variable Defining inline functions (as fast as macros) Specifying the assembler name to use for a C symbol Assembler instructions with C operands Constraints for asm operands Header files can use __const__, __asm__, etc Defining variables residing in specified registers Define an enum tag without specifying its possible values Printable strings which are the name of the current function Getting the return or frame address of a function (limited support) Other built-in functions (see Section 6.14.5) Using vector instructions through built-in functions Built-in functions specific to particular targets Pragmas accepted by GCC Unnamed struct/union fields within structs/unions Per-thread variables Binary constants using the '0b' prefix. www.ti.com 6.14.2 Function Attributes The following GCC function attributes are supported: • alias • always_inline • const • constructor • deprecated • format • format_arg • interrupt • malloc • noinline • noreturn • pure • ramfunc • section • unused • used • warn_unused_result For example, this function declaration uses the alias attribute to make "my_alias" a function alias for the "myFunc" function: void my_alias() __attribute__((alias("myFunc"))); The format attribute is applied to the declarations of printf, fprintf, sprintf, snprintf, vprintf, vfprintf, vsprintf, vsnprintf, scanf, fscanf, vfscanf, vscanf, vsscanf, and sscanf in stdio.h. Thus when GCC extensions are enabled, the data arguments of these functions are type checked against the format specifiers in the format string argument and warnings are issued when there is a mismatch. These warnings can be suppressed in the usual ways if they are not desired. 132 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com GNU Language Extensions See Section 6.9.14 for more about using the interrupt function attribute. The malloc attribute is applied to the declarations of malloc, calloc, realloc and memalign in stdlib.h. The ramfunc attribute specifies that a function will be placed in and executed from RAM. The ramfunc attribute allows the compiler to optimize functions for RAM execution, as well as to automatically copy functions to RAM on flash-based devices. For example: __attribute__((ramfunc)) void f(void) { ... } The --ramfunc=on option specifies that all functions compiled with this option are placed in and executed from RAM, even if this function attribute is not used. Newer TI linker command files support the ramfunc attribute automatically by placing functions with this attribute in the .TI.ramfunc section. If you have a linker command file that does not include a section specification for the .TI.ramfunc section, you can modify the linker command file to place this section in RAM. See the Placing functions in RAM wiki page for more about the ramfunc attribute and option. See the TMS320C28x Assembly Language Tools User's Guide for details on section placement. Fast branch instructions are generated for ramfunc functions. Regular branch instructions are generated for all other functions. The ramfunc attribute is ignored by the CLA compiler. 6.14.3 Variable Attributes The following variable attributes are supported: aligned, deprecated, mode, section, transparent_union, unused, and used. The used attribute is defined in GCC 4.2 (see http://gcc.gnu.org/onlinedocs/gcc-4.2.4/gcc/VariableAttributes.html#Variable-Attributes). 6.14.4 Type Attributes The following type attributes are supported: aligned, byte_peripheral, deprecated, transparent_union, and unused. You cannot create arrays using the aligned type attribute with an alignment larger than the size of the type. This is because the C language guarantees that the size of an array is the same as the size of each element times the number of elements. This might not be the case if an array used the aligned type attribute, because padding could be added to the array. See Section 6.14.6 for details about the byte_peripheral type attribute. 6.14.5 Built-In Functions The following built-in functions are supported: __builtin_abs, __builtin_classify_type, __builtin_constant_p, __builtin_expect, __builtin_fabs, __builtin_fabsf, __builtin_frame_address, __builtin_labs, __builtin_memcpy, and __builtin_return_address. The __builtin_frame_address function always returns zero. The __builtin_return_address function always returns zero. 6.14.6 Using the Byte Peripheral Type Attribute The C2000 architecture has 16-bit words. Some peripherals, however, are 8-bit byte addressable. The byte peripherals bridge translates addresses between the CPU and byte peripherals by treating an address as a byte address. Therefore, only some C2000 addresses map correctly to the byte peripherals. Even and odd addresses to 16-bit data both map to the same data element on the byte peripheral. The same is true for addresses to 32-bit data. Addresses for 16-bit accesses must be 32-bit aligned and those for 32-bit accesses must be 64-bit aligned. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 133 Copyright © 2017, Texas Instruments Incorporated Compiler Limits www.ti.com C2000 driver libraries and bitfield header files are provided to access peripherals. To support correct accesses to byte peripheral data, the compiler provides the __byte_peripheral_32 intrinsic and the byte_peripheral type attribute. The C2000 driver library accesses byte peripheral data at the correct starting address. However, on the C2000, operations on 32-bit data are often broken up into two operations on 16-bit data because these are more efficient on the architecture. Accesses to 32-bit byte peripheral data cannot be broken up regularly into two 16-bit accesses because the starting offset for the second 16-bits will be incorrect. The __byte_peripheral_32 intrinsic can be used to access a 32-bit byte peripheral data address, preventing these accesses from being broken up. The intrinsic returns a reference to an unsigned long and can be used both to read and write data. See Section 7.5.6 for information about this intrinsic. The byte_peripheral type attribute can be applied as follows to typedefs of unsigned ints and unsigned longs for bitfield support. typedef unsigned int bp_16 __attribute__((byte_peripheral)) ; typedef unsigned long bp_32 __attribute__((byte_peripheral)); The typedef names are not significant. The attributes automatically apply the volatile keyword and handle alignment. All struct members in byte peripheral structs, whether they are bitfields or not, must have these attributes applied via typedefs to ensure proper alignment of the struct accesses. Note that struct layout is different due to differences in alignment, so the bitfields cannot always be accessed via the same container types as in regular structs. For example, the bit positions for byte peripheral bitfield types are compared in the following examples to the positions of regular bitfields. Because 16-bit accesses must be 32-bit aligned, it is not possible to access bits at offsets 16-31 with a 16-bit container. To access these bits, a 32-bit container must be used. In example 1, you could create the same layout as in the regular case by changing the field types to bp_32 for field4 through field6. In example 2, you would change the field types to bp_32 for field2 through field4. struct example1 { bp_16 field1:9; bp_16 field2:6; bp_32 field3:4; bp_16 field4:1; bp_16 field5:5; bp_16 field6:7; }; // regular bits position // 0-8 // 9-14 // 15-18 // 19 // 20-24 // 25-31 //byte peripherals // 0-8 // 9-14 // 15-18 // 32 // 33-37 // 38-44 struct example2{ bp_32 field1:29; bp_16 field2:1; bp_16 field3:1; bp_16 field4:1; }; // regular bits position // 0-28 // 29 // 30 // 31 //byte peripherals // 0-28 // 32 // 33 // 34 Because the alignment will create padding in any objects that are declared, it is recommended that you cast byte peripheral addresses as byte peripheral struct types instead of declaring objects of those struct types. You cannot create arrays using the byte peripheral type attribute. This is because the C language guarantees that the size of an array is the same as the size of each element times the number of elements. This would not be the case in an array of byte peripherals, because there would need to be padding in the array. 6.15 Compiler Limits Due to the variety of host systems supported by the C/C++ compiler and the limitations of some of these systems, the compiler may not be able to successfully compile source files that are excessively large or complex. In general, exceeding such a system limit prevents continued compilation, so the compiler aborts immediately after printing the error message. Simplify the program to avoid exceeding a system limit. Some systems do not allow filenames longer than 500 characters. Make sure your filenames are shorter than 500. 134 TMS320C28x C/C++ Language Implementation Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Compiler Limits The compiler has no arbitrary limits but is limited by the amount of memory available on the host system. On smaller host systems such as PCs, the optimizer may run out of memory. If this occurs, the optimizer terminates and the shell continues compiling the file with the code generator. This results in a file compiled with no optimization. The optimizer compiles one function at a time, so the most likely cause of this is a large or extremely complex function in your source module. To correct the problem, your options are: • Don't optimize the module in question. • Identify the function that caused the problem and break it down into smaller functions. • Extract the function from the module and place it in a separate module that can be compiled without optimization so that the remaining functions can be optimized. SPRU514N – June 2017 Submit Documentation Feedback TMS320C28x C/C++ Language Implementation 135 Copyright © 2017, Texas Instruments Incorporated Chapter 7 SPRU514N – June 2017 Run-Time Environment This chapter describes the TMS320C28x C/C++ run-time environment. To ensure successful execution of C/C++ programs, it is critical that all run-time code maintain this environment. It is also important to follow the guidelines in this chapter if you write assembly language functions that interface with C/C++ code. Topic ........................................................................................................................... Page 7.1 Memory Model.................................................................................................. 137 7.2 Register Conventions........................................................................................ 141 7.3 Function Structure and Calling Conventions........................................................ 143 7.4 Accessing Linker Symbols in C and C++ ............................................................. 146 7.5 Interfacing C and C++ With Assembly Language .................................................. 146 7.6 Interrupt Handling............................................................................................. 158 7.7 Integer Expression Analysis .............................................................................. 159 7.8 Floating-Point Expression Analysis .................................................................... 160 7.9 System Initialization.......................................................................................... 160 136 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Memory Model 7.1 Memory Model The C28x compiler treats memory as two linear blocks of program and data memory: • Program memory contains executable code, initialization records, and switch tables. • Data memory contains external variables, static variables, and the system stack. Blocks of code or data generated by a C/C++ program are placed into contiguous blocks in the appropriate memory space. The Linker Defines the Memory Map NOTE: The linker, not the compiler, defines the memory map and allocates code and data into target memory. The compiler assumes nothing about the types of memory available, about any locations not available for code or data (holes), or about any locations reserved for I/O or control purposes. The compiler produces relocatable code that allows the linker to allocate code and data into the appropriate memory spaces. For example, you can use the linker to allocate global variables into on-chip RAM or to allocate executable code into external ROM. You can allocate each block of code or data individually into memory, but this is not a general practice (an exception to this is memory-mapped I/O, although you can access physical memory locations with C/C++ pointer types). 7.1.1 Sections The compiler produces relocatable blocks of code and data called sections. The sections are allocated into memory in a variety of ways to conform to a variety of system configurations. For more information about sections and allocating them, see the introductory object file information in the TMS320C28x Assembly Language Tools User's Guide. There are two basic types of sections: • Initialized sections contain data or executable code. Initialized sections are usually, but not always, read-only. The C/C++ compiler creates the following initialized sections: – The .binit section contains boot time copy tables. This is a read-only section. For details on BINIT, see the TMS320C28x Assembly Language Tools User's Guide for linker command file information. – The .cinit section and the .pinit section contain tables for initializing variables and constants. These are read-only sections. The C28x .cinit record is limited to 16 bits. This limits initialized objects to 64K. – The .ovly section contains copy tables other than boot time (.binit) copy tables. This is a read-only section. – The .econst section contains string constants, string literals, switch tables, the declaration and initialization of global and static variables, and data defined with the C/C++ qualifier const (provided the constant is not also defined as volatile). This is a read-only section. String literals are placed in the .econst:.string subsection to enable greater link-time placement control. – The .switch section contains tables for switch statements. – The .text section contains all the executable code and compiler-generated constants. This section is usually read-only. – The .TI.crctab section contains CRC checking tables. This is a read-only section. • Uninitialized sections reserve space in memory (usually RAM). A program can use this space at run time to create and store variables. The compiler creates the following uninitialized sections: – The .ebss section reserves space for global and static variables defined. At program startup time, the C/C++ boot routine copies data out of the .cinit section (which can be in ROM) and uses it for initializing variables in the .ebss section. – The .stack section reserves memory for the C/C++ software stack. This memory is used to pass arguments to functions and to allocate space for local variables. – The .esysmem section reserves space for dynamic memory allocation. The reserved space is used by dynamic memory allocation routines, such as malloc, calloc, realloc, or new. If a C/C++ program does not use these functions, the compiler does not create the .esysmem section. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 137 Memory Model www.ti.com The assembler creates the default sections .text, .ebss, and .data. You can instruct the compiler to create additional sections by using the CODE_SECTION and DATA_SECTION pragmas (see Section 6.9.5 and Section 6.9.7). The linker takes the individual sections from different object files and combines sections that have the same name. The resulting output sections and the appropriate placement in memory for each section are listed in Table 7-1. You can place these output sections anywhere in the address space as needed to meet system requirements. Table 7-1. Summary of Sections and Memory Placement Section .cinit .data .ebss .econst .esysmem Type of Memory ROM or RAM RAM RAM ROM or RAM RAM Page 0 1 1 Section .pinit .stack .switch .text Type of Memory ROM or RAM RAM ROM or RAM ROM or RAM Page 0 1 0, 1 0 You can use the SECTIONS directive in the linker command file to customize the section-allocation process. For more information about allocating sections into memory, see the linker description chapter in the TMS320C28x Assembly Language Tools User's Guide. 7.1.2 C/C++ System Stack The C/C++ compiler uses a stack to: • Allocate local variables • Pass arguments to functions • Save the processor status • Save function return addresses • Save temporary results The run-time stack grows up from low addresses to higher addresses. By default, the stack is allocated in the .stack section. (See the run-time-support boot.asm file.) The compiler uses the hardware stack pointer (SP) to manage this stack. Linking the .stack Section NOTE: The .stack section has to be linked into the low 64K of data memory. The SP is a 16-bit register and cannot access addresses beyond 64K. For frames that exceed 63 words in size (the maximum reach of the SP offset addressing mode), the compiler uses XAR2 as a frame pointer (FP). Each function invocation creates a new frame at the top of the stack, from which local and temporary variables are allocated. The FP points at the beginning of this frame to access memory locations that cannot be referenced directly using the SP. The stack size is set by the linker. The linker also creates a global symbol, __STACK_SIZE_, and assigns it a value equal to the size of the stack in bytes. The default stack size is 1K words. You can change the size of the stack at link time by using the --stack_size linker option. Stack Overflow NOTE: The compiler provides no means to check for stack overflow during compilation or at run time. A stack overflow disrupts the run-time environment, causing your program to fail. Be sure to allow enough space for the stack to grow. You can use the --entry_hook option to add code to the beginning of each function to check for stack overflow; see Section 2.13. 138 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Memory Model 7.1.3 Allocating .econst to Program Memory If your system configuration does not support allocating an initialized section such as .econst to data memory, then you have to allocate the .econst section to load in program memory and run in data memory. At boot time, copy the .econst section from program to data memory. The following sequence shows how you can perform this task. 1. Extract boot.asm from the source library: ar2000 -x rts.src boot.asm 2. Edit boot.asm and change the CONST_COPY flag to 1: CONST_COPY .set 1 3. Assemble boot.asm: cl2000 boot.asm 4. Archive the boot routine into the object library: ar2000 -r rts2800_ml.lib boot.obj For an .econst section, link with a linker command file that contains the following entries: SECTIONS { ... .econst : load = PROG PAGE 1, run = DATA PAGE 1 { /* GET RUN ADDRESS */ __econst_run = .; /* MARK LOAD ADDRESS */ *(.ec_mark) /* ALLOCATE .econst */ *(.econst) /* COMPUTE LENGTH */ __econst_length = - .__econst_run; } } In your linker command file, you can substitute the name PROG with the name of a memory area on page 0 and DATA with the name of a memory area on page 1. The rest of the command file must use the names as above. The code in boot.asm that is enabled when you change CONST_COPY to 1 depends on the linker command file using these names in this manner. To change any of the names, you must edit boot.asm and change the names in the same way. 7.1.4 Dynamic Memory Allocation The run-time-support library supplied with the C28x compiler contains several functions (such as malloc, calloc, and realloc) that allow you to allocate memory dynamically for variables at run time. Memory is allocated from a global pool, or heap, that is defined in the .esysmem section. You can set the size of the .esysmem section by using the --heap_size=size option with the linker command. The linker also creates a global symbol, __SYSMEM_SIZE, and assigns it a value equal to the size of the heap in words. The default size is 1K words. For more information on the --heap_size option, see the linker description chapter in the TMS320C28x Assembly Language Tools User's Guide. If you use any C I/O function, the RTS library allocates an I/O buffer for each file you access. This buffer will be a bit larger than BUFSIZ, which is defined in stdio.h and defaults to 256. Make sure you allocate a heap large enough for these buffers or use setvbuf to change the buffer to a statically-allocated buffer. Dynamically allocated objects are not addressed directly (they are always accessed with pointers) and the memory pool is in a separate section (.esysmem); therefore, the dynamic memory pool can have a size limited only by the amount of available memory in your system. To conserve space in the .ebss section, you can allocate large arrays from the heap instead of defining them as global or static. For example, instead of a definition such as: struct big table[100]; Use a pointer and call the malloc function: struct big *table table = (struct big *)malloc(100*sizeof(struct big)); SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 139 Memory Model www.ti.com 7.1.5 Initialization of Variables The C/C++ compiler produces code that is suitable for use as firmware in a ROM-based system. In such a system, the initialization tables in the .cinit section are stored in ROM. At system initialization time, the C/C++ boot routine copies data from these tables (in ROM) to the initialized variables in .ebss (RAM). In situations where a program is loaded directly from an object file into memory and run, you can avoid having the .cinit section occupy space in memory. A loader can read the initialization tables directly from the object file (instead of from ROM) and perform the initialization directly at load time instead of at run time. You can specify this to the linker by using the --ram_model link option. For more information, see Section 7.9. 7.1.6 Allocating Memory for Static and Global Variables A unique, contiguous space is allocated for all static variables declared in a C/C++ program. The linker determines the address of the space. The compiler ensures that space for these variables is allocated in multiples of words so that each variable is aligned on a word boundary. The C/C++ compiler expects global variables to be allocated into data memory. (It reserves space for them in .ebss.) Variables declared in the same module are allocated into a single, contiguous block of memory. 7.1.7 Field/Structure Alignment When the compiler allocates space for a structure, it allocates as many words as are needed to hold all of the structure's members and to comply with alignment constraints for each member. All non-field types are aligned on word boundaries. Fields are allocated as many bits as requested. Adjacent fields are packed into adjacent bits of a word, but they do not overlap words. If a field would overlap into the next word, the entire field is placed into the next word. Fields are packed as they are encountered; the least significant bits of the structure word are filled first. 7.1.8 Character String Constants In C, a character string constant is used in one of the following ways: • To initialize an array of characters. For example: char s[] = "abc"; When a string is used as an initializer, it is simply treated as an initialized array; each character is a separate initializer. For more information about initialization, see Section 7.9. • In an expression. For example: strcpy (s, "abc"); When a string is used in an expression, the string itself is defined in the .econst section with the .string assembler directive, along with a unique label that points to the string; the terminating 0 byte is included. For example, the following lines define the string abc, and the terminating 0 byte (the label SL5 points to the string): .sect ".econst" SL5: .string "abc",0 String labels have the form SLn, where n is a number assigned by the compiler to make the label unique. The number begins at 0 and is increased by 1 for each string defined. All strings used in a source module are defined at the end of the compiled assembly language module. The label SLn represents the address of the string constant. The compiler uses this label to reference the string expression. Because strings are stored in the .econst section (possibly in ROM) and shared, it is bad practice for a program to modify a string constant. The following code is an example of incorrect string use: const char *a = "abc" a[1] = 'x'; /* Incorrect! undefined behavior */ 140 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Register Conventions 7.2 Register Conventions Strict conventions associate specific registers with specific operations in the C/C++ environment. If you plan to interface an assembly language routine to a C/C++ program, you must understand and follow these register conventions. The register conventions dictate how the compiler uses registers and how values are preserved across function calls. There are two types of register variable registers, save on entry and save on call. The distinction between these two types of registers is the method by which they are preserved across function calls. It is the called function’s responsibility to preserve save-on-entry registers, and the calling function’s responsibility to preserve save-on-call registers if you need to preserve that register’s value. 7.2.1 TMS320C28x Register Use and Preservation Table 7-2 summarizes how the compiler uses the TMS320C28x registers and shows which registers are defined to be preserved across function calls. The FPU uses all the C28x registers as well as the registers described in Table 7-3. Table 7-2. Register Use and Preservation Conventions Register Usage Save on Entry AL Expressions, argument passing, and returns 16-bit results from No functions AH Expressions and argument passing No DP Data page pointer (used to access global variables) No PH Multiply expressions and Temp variables No PL Multiply expressions and Temp variables No SP Stack pointer T Multiply and shift expressions No TL Multiply and shift expressions No XAR0 Pointers and expressions No XAR1 Pointers and expressions Yes XAR2 Pointers, expressions, and frame pointing (when needed) Yes XAR3 Pointers and expressions Yes XAR4 Pointers, expressions, argument passing, and returns 32-bit No pointer values from functions XAR5 Pointers, expressions, and arguments No XAR6 Pointers and expressions No XAR7 Pointers, expressions, indirect calls and branches (used to No implement pointers to functions and switch statements) (1) The SP is preserved by the convention that everything pushed on the stack is popped off before returning. Save on Call Yes Yes No Yes Yes (1) (1) Yes Yes Yes No No No Yes Yes Yes Yes Register R0H R1H R2H R3H R4H R5H R6H R7H Table 7-3. FPU Register Use and Preservation Conventions Usage Expressions, argument passing, and returns 32-bit float from functions Expressions and argument passing Expressions and argument passing Expressions and argument passing Expressions Expressions Expressions Expressions Save on Entry No No No No Yes Yes Yes Yes Save on Call Yes Yes Yes Yes No No No No SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 141 Register Conventions www.ti.com 7.2.2 Status Registers Table 7-4 shows all of the status fields used by the compiler. Presumed value is the value the compiler expects in that field upon entry to, or return from, a function; a dash in this column indicates the compiler does not expect a particular value. The modified column indicates whether code generated by the compiler ever modifies this field. Table 7-4. Status Register Fields Field Name Presumed Value Modified ARP Auxiliary Register Pointer - Yes C Carry - Yes N Negative flag - Yes OVM Overflow mode 0 (1) Yes PAGE0 Direct/stack address mode 0 (1) No PM Product shift mode 0 (1) Yes SPA Stack pointer align bit - Yes (in interrupts) SXM Sign extension mode - Yes TC Test/control flag - Yes V Overflow flag - Yes Z Zero flag - Yes (1) The initialization routine that sets up the C run-time environment sets these fields to the presumed value. Table 7-5 shows the additional status fields used by the compiler for FPU Targets. Table 7-5. Floating-Point Status Register (STF(1)) Fields For FPU Targets Only Field Name Presumed Value Modified LVF (2) (3) Latched overflow float flag - Yes LUF(2) (3) Latched underflow float flag - Yes NF (2) Negative float flag - Yes ZF (2) Zero float flag - Yes NI (2) Negative integer flag - Yes ZI (2) Zero integer flag bit - Yes TF (2) Test flag bit - Yes RNDF32 Round F32 mode(4) - Yes RNDF64 Round F64 mode(4) - Yes SHDWS Shadow mode status - Yes (1) Unused STF register bits read 0 and writes are ignored. (2) Specified flags in the STF register can be exported to the ST0 register by way of the MOVST0 instruction. (3) The LVF and LUF flag signals can be connected to the PIE to generate an overflow-and-underflow interrupt. This can be a useful debug tool. (4) If RNDF32 or RNDF64 is 0, mode is round to zero (truncate), otherwise mode is round to nearest (even). All other status register fields are not used and do not affect code generated by the compiler. 142 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Function Structure and Calling Conventions 7.3 Function Structure and Calling Conventions The C/C++ compiler imposes a strict set of rules on function calls. Except for special run-time support functions, any function that calls or is called by a C/C++ function must follow these rules. Failure to adhere to these rules can disrupt the C/C++ environment and cause a program to fail. Figure 7-1 illustrates a typical function call. In this example, parameters that cannot be placed in registers are passed to the function on the stack. The function then allocates local variables and calls another function. This example shows the allocated local frame and argument block for the called function. Functions that have no local variables and do not require an argument block do not allocate a local frame. The term argument block refers to the part of the local frame used to pass arguments to other functions. Parameters are passed to a function by moving them into the argument block rather than pushing them on the stack. The local frame and argument block are allocated at the same time. High Before call Figure 7-1. Use of the Stack During a Function Call Move parameters to argument block and registers. Call function. Allocate new frame and argument block. High High Arguments passed on stack Caller’s Low local vars SP RPC = Return Address Low Previous RPC Value Arguments passed on stack Caller’s local vars SP RPC = Return Address Arguments Accumulator AL AH XAR4 XAR5 Low Local argument block Local frame Previous RPC Value Arguments passed on stack Caller’s local vars SP RPC = Return Address 7.3.1 How a Function Makes a Call A function (parent function) performs the following tasks when it calls another function (child function). 1. Any registers whose values are not necessarily preserved by the function being called (registers that are not save-on-entry (SOE) registers), but will be needed after the function returns are saved on the stack. 2. If the called function returns a structure, the calling function allocates the space for the structure and pass the address of that space to the called function as the first argument. 3. Arguments passed to the called function are placed in registers and, when necessary, placed on the stack. Arguments are placed in registers using the following scheme: (a) If the target is FPU and there are any 32-bit float arguments, the first four float arguments are placed in registers R0H-R3H. (b) If there are any 64-bit floating point arguments (long doubles), they are passed by reference. (c) If there are any 64-bit integer arguments (long long), the first is placed in ACC and P (ACC holds the upper 32 bits and P holds the lower 32 bits). All other 64-bit integer arguments are placed on the stack in reverse order. If the P register is used for argument passing, then prolog/epilog abstraction is disabled for that SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 143 Function Structure and Calling Conventions www.ti.com function. See Section 3.13 for more information on abstraction. (d) If there are any 32-bit arguments (longs or floats), the first is placed in the 32-bit ACC (AH/AL). All other 32-bit arguments are placed on the stack in reverse order. func1(long a, long long b, int c, int* d); stack ACC/P XAR5, XAR4 (e) Pointer arguments are placed in XAR4 and XAR5. All other pointers are placed on the stack. (f) Remaining 16-bit arguments are placed in the order AL, AH, XAR4, XAR5 if they are available. 4. Any remaining arguments not placed in registers are pushed on the stack in reverse order. That is, the leftmost argument that is placed on the stack is pushed on the stack last. All 32-bit arguments are aligned to even addresses on the stack. A structure argument is passed as the address of the structure. The called function must make a local copy. For a function declared with an ellipsis, indicating that it is called with varying numbers of arguments, the convention is slightly modified. The last explicitly declared argument is passed on the stack so that its stack address can act as a reference for accessing the undeclared arguments. 5. The stack pointer (SP) must be even-aligned by the parent function prior to making a call to the child function. This is done by incrementing the stack pointer by 1, if necessary. If needed, the coder should increment the SP before making the call. Some examples of function calls that show where arguments are placed are listed below: func1 (int a, int b. long c) XAR4 XAR5 AH/AL func1 (long a, int b, long c) ; AH/AL XAR4 stack vararg (int a, int b, int c, ...) AL AH stack 6. The caller calls the function using the LCR instruction. The RPC register value is pushed on the stack. The return address is then stored in the RPC register. 7. The stack is aligned at function boundary. 7.3.2 How a Called Function Responds A called function (child function) must perform the following tasks: 1. If the called function modifies XAR1, XAR2, or XAR3, it must save them, since the calling function assumes that the values of these registers are preserved upon return. If the target is FPU, then in addition to the C28x registers, the called function must save registers R4H, R5H, R6H or R7H, if it modifies any of them. Any other registers may be modified without preserving them. 2. The called function allocates enough space on the stack for any local variables, temporary storage area, and arguments to functions that this function might call. This allocation occurs once at the beginning of the function by adding a constant to the SP register. 3. The stack is aligned at function boundary. 4. If the called function expects a structure argument, it receives a pointer to the structure instead. If writes are made to the structure from within the called function, space for a local copy of the structure must be allocated on the stack and the local structure must be copied from the passes pointer to the structure. If no writes are made to the structure, it can be referenced in the called function indirectly through the pointer argument. You must be careful to properly declare functions that accept structure arguments, both at the point where they are called (so that the structure argument is passed as an address) and at the point where they are declared (so the function knows to copy the structure to a local copy). 5. The called function executes the code for the function. 6. The called function returns a value. It is placed in a register using the following convention: 144 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Function Structure and Calling Conventions 16-bit integer value 32-bit integer value 64-bit integer value 32-bit pointer structure reference AL ACC ACC/P XAR4 XAR6 If the target is FPU and a 32-bit float value is returned, the called function places this value in R0H. If the function returns a structure, the caller allocates space for the structure and passes the address of the return space to the called function in XAR6. To return a structure, the called function copies the structure to the memory block pointed by the extra argument. In this way, the caller can be smart about telling the called function where to return the structure. For example, in the statement s= f(x), where S is a structure and F is a function that returns a structure, the caller can actually make the call as f(&s, x). The function f then copies the return structure directly into s, performing the assignment automatically. If the caller does not use the return structure value, an address value of 0 can be passed as the first argument. This directs the called function not to copy the return structure. You must be careful to properly declare functions that return structures both at the point where they are called (so that the extra argument is passed) and at the point where they are declared (so the function knows to copy the result). Returning 64-bit floating-point values (long double) are returned similarly to structures. 7. The called function deallocates the frame by subtracting the value that was added to the SP earlier. 8. The called function restores the values of all registers saved in Step 1. 9. The called function returns using the LRETR instruction. The PC is set to the value in the RPC register. The previous RPC value is popped from the stack and stored in the RPC register. 7.3.3 Special Case for a Called Function (Large Frames) If the space that needs to be allocated on the stack (step 2 in the previous section) is larger than 63 words, additional steps and resources are required to ensure that all local nonregister variables can be accessed. Large frames require using a frame pointer register (XAR2) to reference local non-register variables within the frame. Prior to allocating space on the frame, the frame pointer is set up to point at the first argument on the stack that was passed on to the called function. If no incoming arguments are passed on to the stack, the frame pointer points to the return address of the calling function, which is at the top of the stack upon entry to the called function. Avoid allocating large amounts of local data when possible. For example, do not declare large arrays within functions. 7.3.4 Accessing Arguments and Local Variables A function accesses its local nonregister variables and its stack arguments indirectly through either the SP or the FP (frame pointer, designated to be XAR2). All local and argument data that can be accessed with the SP use the *−SP [offset] addressing mode since the SP always points one past the top of the stack and the stack grows toward larger addresses. The PAGE0 Mode Bit Must Be Reset NOTE: Since the compiler uses the *-SP [offset] addressing mode, the PAGE0 mode bit must be reset (set to 0). The largest offset available using *-SP [offset] is 63. If an object is too far away from the SP to use this mode of access, the compiler uses the FP (XAR2). Since FP points at the bottom of the frame, accesses made with the FP use either *+FP [offset] or *+FP [AR0/AR1] addressing modes. Since large frames require utilizing XAR2 and possibly an index register, extra code and resources are required to make local accesses. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 145 Function Structure and Calling Conventions www.ti.com 7.3.5 Allocating the Frame and Accessing 32-Bit Values in Memory Some TMS320C28x instructions read and write 32 bits of memory at once (MOVL, ADDL, etc.). These instructions require that 32-bit objects be allocated on an even boundary. To ensure that this occurs, the compiler takes these steps: 1. It initializes the SP to an even boundary. 2. Because a call instruction adds 2 to the SP, it assumes that the SP is pointing at an even address. 3. It makes sure that the space allocated on the frame totals an even number, so that the SP points to an even address. 4. It makes sure that 32-bit objects are allocated to even addresses, relative to the known even address in the SP. 5. Because interrupts cannot assume that the SP is odd or even, it aligns the SP to an even address. For more information on how these instructions access memory, see the TMS320C28x Assembly Language Tools User’s Guide. 7.4 Accessing Linker Symbols in C and C++ See the section on "Using Linker Symbols in C/C++ Applications" in the TMS320C28x Assembly Language Tools User's Guide for information about referring to linker symbols in C/C++ code. 7.5 Interfacing C and C++ With Assembly Language The following are ways to use assembly language with C/C++ code: • Use separate modules of assembled code and link them with compiled C/C++ modules (see Section 7.5.1). • Use assembly language variables and constants in C/C++ source (see Section 7.5.3). • Use inline assembly language embedded directly in the C/C++ source (see Section 7.5.5). • Use intrinsics in C/C++ source to directly call an assembly language statement (see Section 7.5.6). 7.5.1 Using Assembly Language Modules With C/C++ Code Interfacing C/C++ with assembly language functions is straightforward if you follow the calling conventions defined in Section 7.3, and the register conventions defined in Section 7.2. C/C++ code can access variables and call functions defined in assembly language, and assembly code can access C/C++ variables and call C/C++ functions. Follow these guidelines to interface assembly language and C: • All functions, whether they are written in C/C++ or assembly language, must follow the register conventions outlined in Section 7.2. • Dedicated registers modified by a function must be preserved. Dedicated registers include: XAR1 XAR2 XAR3 SP R4H (FPU only) R5H (FPU only) R6H (FPU only) R7H (FPU only) If the SP is used normally, it does not need to be preserved explicitly. The assembly function is free to use the stack as long as anything that is pushed on the stack is popped back off before the function returns (thus preserving the SP). Any register that is not dedicated can be used freely without being preserved. • The stack pointer (SP) must be even-aligned by the parent function prior to making a call to the child function. This is done by incrementing the stack pointer by 1, if necessary. If needed, the coder should increment the SP before making the call. • The stack is aligned at function boundary. • Interrupt routines must save all the registers they use. For more information, see Section 7.6. 146 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Interfacing C and C++ With Assembly Language • When you call a C/C++ function from assembly language, load the designated registers with arguments and push the remaining arguments onto the stack as described in Section 7.3.1. When accessing arguments passed in from a C/C++ function, these same conventions apply. • Longs and floats are stored in memory with the least significant word at the lower address. • Structures are returned as described in Section 7.3.2. • No assembly module should use the .cinit section for any purpose other than autoinitialization of global variables. The C/C++ startup routine assumes that the .cinit section consists entirely of initialization tables. Disrupting the tables by putting other information in .cinit can cause unpredictable results. • The compiler prepends an underscore ( _ ) to the beginning of all identifiers. In assembly language modules, you must use the prefix _ for all objects that are to be accessible from C/C++. For example, a C/C++ object named x is called _x in assembly language. For identifiers that are to be used only in an assembly language module or modules, any name that does not begin with an underscore can be safely used without conflicting with a C/C++ identifier. • The compiler assigns linknames to all external objects. Thus, when you write assembly language code, you must use the same linknames as those assigned by the compiler. See Section 6.11 for details. • Any object or function declared in assembly language that is accessed or called from C/C++ must be declared with the .def or .global directive in the assembly language modifier. This declares the symbol as external and allows the linker to resolve references to it. Likewise, to access a C/C++ function or object from assembly language, declare the C/C++ object with the .ref or .global directive in the assembly language module. This creates an undeclared external reference that the linker resolves. • Because compiled code runs with the PAGE0 mode bit reset, if you set the PAGE0 bit to 1 in your assembly language function, you must set it back to 0 before returning to compiled code. • If you define a structure in assembly and access it in C using extern struct, the structure should be blocked. The compiler assumes that structure definitions are blocked to optimize the DP load. So the definition should honor this assumption. You can block the structure by specifying the blocking flag in the .usect directive. See the TMS320C28x Assembly Language Tools User’s Guide for more information on these directives. 7.5.2 Accessing Assembly Language Functions From C/C++ Functions defined in C++ that will be called from assembly should be defined as extern "C" in the C++ file. Functions defined in assembly that will be called from C++ must be prototyped as extern "C" in C++. Example 7-1 illustrates a C++ function called main, which calls an assembly language function called asmfunc, Example 7-2. The asmfunc function takes its single argument, adds it to the C++ global variable called gvar, and returns the result. Example 7-1. Calling an Assembly Language Function From a C/C++ Program extern "C"{ extern int asmfunc(int a); /* declare external asm function */ int gvar = 0; /* define global variable */ } void main() { int i = 5; i = asmfunc(i); /* call function normally */ } SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 147 Interfacing C and C++ With Assembly Language Example 7-2. Assembly Language Program Called by Example 7-1 www.ti.com .global _gvar .global _asmfunc _asmfunc: MOVZ ADDB MOV LRETR DP,#_gvar AL,#5 @_gvar,AL In the C++ program in Example 7-1, the extern “C” declaration tells the compiler to use C naming conventions (i.e., no name mangling). When the linker resolves the .global _asmfunc reference, the corresponding definition in the assembly file will match. The parameter i is passed in register AL. 7.5.3 Accessing Assembly Language Variables From C/C++ It is sometimes useful for a C/C++ program to access variables or constants defined in assembly language. There are several methods that you can use to accomplish this, depending on where and how the item is defined: a variable defined in the .ebss section, a variable not defined in the .ebss section, or a linker symbol. 7.5.3.1 Accessing Assembly Language Global Variables Accessing variables from the .ebss section or a section named with .usect is straightforward: 1. Use the .usect directive to define the variable. 2. Use the .def or .global directive to make the definition external. 3. Use the appropriate linkname in assembly language. 4. In C/C++, declare the variable as extern and access it normally. Example 7-4 and Example 7-3 show how you can access a variable defined in .ebss. Example 7-3. Assembly Language Variable Program * Note the use of underscores in the following lines .ebss _var,1 ; Define the variable .global _var ; Declare it as external Example 7-4. C Program to Access Assembly Language From Example 7-3 extern int var; var = 1; /* External variable */ /* Use the variable */ 148 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Interfacing C and C++ With Assembly Language 7.5.3.2 Accessing Assembly Language Constants You can define global constants in assembly language by using the .set directive in combination with either the .def or .global directive, or you can define them in a linker command file using a linker assignment statement. These constants are accessible from C/C++ only with the use of special operators. For variables defined in C/C++ or assembly language, the symbol table contains the address of the value contained by the variable. When you access an assembly variable by name from C/C++, the compiler gets the value using the address in the symbol table. For assembly constants, however, the symbol table contains the actual value of the constant. The compiler cannot tell which items in the symbol table are addresses and which are values. If you access an assembly (or linker) constant by name, the compiler tries to use the value in the symbol table as an address to fetch a value. To prevent this behavior, you must use the & (address of) operator to get the value (_symval). In other words, if x is an assembly language constant, its value in C/C++ is &x. See the section on "Using Linker Symbols in C/C++ Applications" in the TMS320C28x Assembly Language Tools User's Guide for more examples that use _symval. For more about symbols and the symbol table, refer to the section on "Symbols" in the TMS320C28x Assembly Language Tools User's Guide. You can use casts and #defines to ease the use of these symbols in your program, as in Example 7-5 and Example 7-6. Example 7-5. Accessing an Assembly Language Constant From C extern int table_size; /*external ref */ #define TABLE_SIZE ((int) (&table_size)) . /* use cast to hide address-of */ . . for (I=0; i N > -17: QMPYL ACC, XT, B SETC SXM SFR ACC, #abs(N) If 16 < N < 32: IMPYL P, XT, B QMPYL ACC, XT, B ASR64 ACC:P, # N If N == 32: IMPYL P, XT, B If -16 > N > -33 QMPYL ACC, XT, B SETC SXM SRF ACC, #16 SRF ACC, #abs(N)−16 If 32 < N < 49: IMPYL ACC, XT, B LSL ACC, # N -32 If -32 > N > -49: QMPYL ACC, XT, B SETC SXM SFR ACC, #16 SFR ACC, #16 If 48 < N < 65: IMPYL ACC, XT, B LSL64 ACC:P, #16 LSL64 ACC:P, # N −48 If -48 > N > -65: QMPYL ACC, XT, B SETC SXM SFR ACC, #16 SFR ACC, #16 long long __llmax(long long dst, long long src); MAXL ACC, src.hi32 MAXCUL P, src.lo32 long long __llmin(long long dst, long long src); MINL ACC, src.hi32 MINCUL P, src.lo32 long __lmax(long dst, long src); MAXL ACC, src long __lmin(long dst, long src); MINL ACC, src int __max(int dst, int src); MAX dst, src int __min(int dst, int src); MIN dst, src int __mov_byte( int *src, unsigned int n ); MOVB AX.LSB,*+XARx[ n ] or MOVZ AR0/AR1, @ n MOVB AX.LSB,*XARx[ {AR0|AR1} ] long __mpy( int src1, int src2 ); MPY ACC, src1 , # src2 long __mpyb( int src1, uint src2 ); MPYB {ACC | P}, T, # src2 long __mpy_mov_t( int src1, int src2, int * dst2 ); unsigned long __mpyu(unit src2, unit srt2); MPY ACC, T, src2 MOV @ dst2, T MPYU {ACC | P}, T,src2 long __mpyxu( int src1, uint src2 ); MPYXU ACC, T, {mem|reg} long dst = __norm32(long src, int * shift ); CSB ACC LSLL ACC, T MOV @ shift, T Description The dst is in ACC. The dst is in ACC. The dst is in P. The dst is in P. The dst is in ACC. The dst is in ACC. The dst is in ACC. The dst is in ACC. The dst is in ACC. If src > dst, copy src to dst. If src < dst, copy src to dst If src > dst, copy src to dst. If src < dst, copy src to dst If src > dst, copy src to dst If src < dst, copy src to dst Return the 8-bit nth element of a byte table pointed to by src. This intrinsic is provided for backward compatibility. The intrinsic __byte is preferred as it returns a reference. Nothing can be done with __mov_byte() that cannot be done with __byte(). Move src1 to the T register. Multiply T by a 16bit immediate (src2). The result is in ACC. Multiply src1 (the T register) by an unsigned 8bit immediate (src2). The result is in ACC or P. Multiply src1 (the T register) by src2. The result is in ACC. Move src1 to *dst2. Multiply src1 (the T register) by src2. Both operands are treated as unsigned 16-bit numbers. The result is in ACC or P. The T register is loaded with src1. The src2 is referenced by memory or loaded into a register. The result is in ACC. Normalize src into dst and update *shift with the number of bits shifted. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 153 Interfacing C and C++ With Assembly Language www.ti.com Table 7-6. TMS320C28x C/C++ Compiler Intrinsics (continued) Intrinsic long long dst = __norm64(long long src, int * shift ); void __or( int * m , int b ); Assembly Instruction(s) CSB ACC LSL64 ACC:P, T MOV @ shift , T CSB ACC LSL64 ACC:P, T MOV TMP16, AH MOV AH, T ADD shift, AH MOV AH, TMP16 OR * m , b long __qmpy32( long src32a, long src32b, int q ); CLRC OVM SPM − 1 MOV T, src32a + 1 MPYXU P, T, src32b + 0 MOVP T, src32b + 1 MPYXU P, T, src32a + 0 MPYA P, T, src32a + 1 If q = 31,30: SPM q − 30 SFR ACC, #45 − q ADDL ACC, P If q = 29: SFR ACC, #16 ADDL ACC, P If q = 28 through 24: SPM q - 30 SFR ACC, #16 SFR ACC, #29 - q ADDL ACC, P If q = 23 through 13: SFR ACC, #16 ADDL ACC, P SFR ACC, #29 − q If q = 12 through 0: SFR ACC, #16 ADDL ACC, P SFR ACC, #16 SFR ACC, #13 − q long __qmpy32by16(long src32, int src16, int q); CLRC OVM MOV T, src16 + 0 MPYXU P, T, src32 + 0 MPY P, T, src32 + 1 If q = 31, 30: SPM q − 30 SFR ACC, #46 − q ADDL ACC, P If q = 29 through 14: SPM 0 SFR ACC, #16 ADDL ACC, P SFR ACC, #30 − q If q = 13 through 0: SPM 0 SFR ACC, #16 ADDL ACC, P SFR ACC, #16 SFR ACC, #14 − q void __restore_interrupts(unsigned int val ); PUSH val POP ST1 long __rol( long src ); ROL ACC long __ror( long src ); ROR ACC void * result = __rpt_mov_imm(void * dst , int src , MOV result , dst int count ); MOV ARx, dst RPT # count || MOV *XARx++, # src Description Normalize 64-bit src into dst and update *shift with the number of bits shifted. OR the contents of memory location m to b and store the result in m, in an atomic way. Extended precision DSP Q math. Different code is generated based on the value of q. Extended precision DSP Q math. Different code is generated based on the value of q. Restore interrupts and set the interrupt vector to value val. Rotate ACC left. Rotate ACC right. Move the dst register to the result register. Move the dst register to a temp (ARx) register. Copy the immediate src to the temp register count + 1 times. The src must be a 16-bit immediate. The count can be an immediate from 0 to 255 or a variable. 154 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Interfacing C and C++ With Assembly Language Table 7-6. TMS320C28x C/C++ Compiler Intrinsics (continued) Intrinsic int __rpt_norm_inc( long src, int dst, int count ); Assembly Instruction(s) MOV ARx, dst RPT # count || NORM ACC, ARx++ int __rpt_norm_dec(long src, int dst, int count); MOV ARx, dst RPT # count || NORM ACC, ARx-- long __rpt_rol(long src, int count); RPT # count || ROL ACC long __rpt_ror(long src, int count); RPT # count || ROR ACC long __rpt_subcu(long dst, int src, int count); RPT count || SUBCU ACC, src unsigned long __rpt_subcul(unsigned long num, unsigned long den, unsigned long &remainder, int count); long __sat( long src ); RPT count || SUBCUL ACC, den SAT ACC long __sat32( long src, long limit ); long __sathigh16(long src, int limit); long __satlow16( long src ); long __sbbu( long src1 , uint src2 ); SETC OVM ADDL ACC, {mem|P} SUBL ACC, {mem|P} SUBL ACC, {mem|P} ADDL ACC, {mem|P} CLRC OVM SETC OVM ADDL ACC, {mem|P}<<16 SUBL ACC, {mem|P}<<16 SUBL ACC, {mem|P}<<16 ADDL ACC, {mem|P}<<16 CLRC OVM SFR ACC, rshift SETC OVM MOV T, #0xFFFF CLR SXM ; if necessary ADD ACC, T <<15 SUB ACC, T <<15 SUB ACC, T <<15 ADD ACC, T <<15 CLRC OVM SBBU ACC, src2 void __sub( int * m, int b ); SUB * m , b long __subcu( long src1, int src2 ); unsigned long __subcul(unsigned long num, unsigned long den, unsigned long &remainder); SUBCU ACC, src2 SUBCUL ACC, den Description Repeat the normalize accumulator value count + 1 times. The count can be an immediate from 0 to 255 or a variable. Repeat the normalize accumulator value count + 1 times. The count can be an immediate from 0 to 255 or a variable. Repeat the rotate accumulator left count + 1 times. The result is in ACC. The count can be an immediate from 0 to 255 or a variable. Repeat the rotate accumulator right count + 1 times. The result is in ACC. The count can be an immediate from 0 to 255 or a variable. The src operand is referenced from memory or loaded into a register and used as an operand to the SUBCU instruction. The result is in ACC. The count can be an immediate from 0 to 255 or a variable. The instruction repeats count + 1 times. Performs repeated conditional long subtraction as typically used in unsigned modulus division. Returns the quotient. Load ACC with 32-bit src. The result is in ACC. Saturate a 32-bit value to a 32-bit mask. Load ACC with src. Limit value is either referenced from memory or loaded into the P register. The result is in ACC. Saturate a 32-bit value to 16-bits high. Load ACC with src. The limit value is either referenced from memory or loaded into register. The result is in ACC. The result can be right shifted and stored into an int. For example: ivar=__sathigh16(lvar, mask)>>6; Saturate a 32-bit value to 16-bits low. Load ACC with src. Load T register with #0xFFFF. The result is in ACC. Subtract src2 + logical inverse of C from ACC (src1). The result is in ACC. Subtract b from the contents of memory location m and store the result in m, in an atomic way. Subtract src2 shifted left 15 from ACC (src1). The result is in ACC. Performs a single conditional long subtraction as typically used in unsigned modulus division. Returns the quotient. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 155 Interfacing C and C++ With Assembly Language www.ti.com Table 7-6. TMS320C28x C/C++ Compiler Intrinsics (continued) Intrinsic void __subl( long * m, long b ); void __subr( int * m , int b ); void __subrl( long * m , long b ); if (__tbit( int src , int bit ) ); void __xor( int * m, int b ); Assembly Instruction(s) SUBL * m , b SUBR * m , b SUBRL * m , b TBIT src , # bit XOR * m, b Description Subtract b from the contents of memory location m and store the result in m, in an atomic way. Subtract the contents of memory location m from b and store the result in m, in an atomic way. Subtract the contents of memory location m from b and store the result in m, in an atomic way. SET TC status bit if specified bit of src is 1. XOR the contents of memory location m to b and store the result in m, in an atomic way. Table 7-7. C/C++ Compiler Intrinsics for FPU Intrinsic double __einvf32( double x ); double __eisqrtf32( double x ); void __f32_max_idx( double &dst, double src, double &idx_dst, double idx_src ); void __f32_min_idx( double &dst, double src, double &idx_dst, double idx_src ); int __f32toi16r(double src ); unsigned int __f32toui16r(double src ); float __fmax( float x, float y ); float __fmin( float x, float y ); double __fracf32(double src ); double __fsat(double val, double max, double min ); void __swapf( double &a, double &b ); void __swapff( float &a, float &b ); Assembly Instruction(s) EINVF32 x EISQRTF32 x MAXF32 dst, src || MOV32 idx_dst, idx_src MINF32 dst, src || MOV32 idx_dst, idx_src F32TOI16R dst, src F32TOUI16R dst, src MAXF32 dst, src MINF32 dst, src FRACF32 dst, src MAXF32 dst, src2 MINF32 dst, src1 swapf a, b swapff a, b Description Compute and return 1/x to about 8 bits of precision. Find the square root of 1/x to about 8 bits of precision. If src>dst, copy src to dst, and copy idx_src to idx_dst. If srcdst, copy src to dst If src max. Swap the contents of a and b. Swap the contents of a and b. The following intrinsics perform hardware instructions using the Trigonometric Math Unit (TMU). These intrinsics are enabled if the --tmu_support=tmu0 compiler option is used. Table 7-8. C/C++ Compiler Intrinsics for TMU Intrinsic Assembly Instruction(s) double __divf32( double num , double denom ); DIVF32 dst, num, denom double __sqrt( double src ); double __sin( double src ); SQRTF32 dst, src SINF32 dst, src double __cos( double src ); COSF32 dst, src double __atan( double src ); ATANF32 dst, src double __atan2( double y , double x ); QUADF32 quadrant, ratio, y, x ATANPUF32 atanpu, ratio ADDF32 atan2pu, atanpu MPY2PIF32 atan2, atan2pu Description Return num divided by denom using the TMU hardware instruction for floating point division. Return the square root of src. Return the sine of src radians, where src is provided in radians. Return the cosine of src radians, where src is provided in radians. Return the principal value of the arc tangent of src radians. Return the principal value of the arc tangent plus the quadrant for x, y. The value is returned in radians. 156 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Interfacing C and C++ With Assembly Language Table 7-8. C/C++ Compiler Intrinsics for TMU (continued) Intrinsic double __mpy2pif32( double src ); Assembly Instruction(s) MPY2PIF32 dst, src double __div2pif32( double src ); DIV2PIF32 dst, src double __sinpuf32( double src ); double __cospuf32( double src ); double __atanpuf32( double src ); double __atan2puf32( double x, double y ); double __quadf32( double ratio, double y, double x ); SINPUF32 dst, src COSPUF32 dst, src ATANPUF32 dst, src QUADF32 quadrant, ratio, y, x ATANPUF32 atanpu, ratio ADDF32 dst, atanpu QUADF32 quadrant, ratio, y, x Description Return the result of multiplying src by 2pi. This converts a per unit value to radians. Per unit values are commonly used in control applications to represent normalized radians. Return the result of multiplying src by 1/2pi (effectively dividing by 2pi). This converts a value in radians to a per unit value. Return the sine of src in radians, where src is provided as a per unit value. Return the cosine of src in radians, where src is provided as a per unit value. Return the principal value of the arc tangent of src, which is provided as a per unit value. Return the principal value of the arc tangent plus the quadrant value for y, x. The value is returned as a per unit value. Return the quadrant value (0.0, +/-0.25, or +/0.5) and the ratio of x and y, which are provided as per unit values. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 157 Interrupt Handling www.ti.com 7.6 Interrupt Handling As long as you follow the guidelines in this section, you can interrupt and return to C/C++ code without disrupting the C/C++ environment. When the C/C++ environment is initialized, the startup routine does not enable or disable interrupts. If the system is initialized by way of a hardware reset, interrupts are disabled. If your system uses interrupts, you must handle any required enabling or masking of interrupts. Such operations have no effect on the C/C++ environment and are easily incorporated with asm statements or calling an assembly language function. 7.6.1 General Points About Interrupts An interrupt routine can perform any task performed by any other function, including accessing global variables, allocating local variables, and calling other functions. When you write interrupt routines, keep the following points in mind: • An interrupt handling routine cannot have arguments. If any are declared, they are ignored. • An interrupt handling routine can be called by normal C/C++ code, but it is inefficient to do this because all the registers are saved. • An interrupt handling routine can handle a single interrupt or multiple interrupts. The compiler does not generate code that is specific to a certain interrupt, except for c_int00, which is the system reset interrupt. When you enter this routine, you cannot assume that the run-time stack is set up; therefore, you cannot allocate local variables, and you cannot save any information on the run-time stack. • To associate an interrupt routine with an interrupt, the address of the interrupt function must be placed in the appropriate interrupt vector. You can use the assembler and linker to do this by creating a simple table of interrupt addresses using the .sect assembler directive. • In assembly language, remember to precede the symbol name with an underscore. For example, refer to c_int00 as _c_int00. 7.6.2 Using C/C++ Interrupt Routines If a C/C++ interrupt routine does not call any other functions, only those registers that the interrupt handler uses are saved and restored. However, if a C/C++ interrupt routine does call other functions, these functions can modify unknown registers that the interrupt handler does not use. For this reason, the compiler saves all the save-on-call registers if any other functions are called. A C/C++ interrupt routine is like any other C/C++ function in that it can have local variables and register variables; however, it should be declared with no arguments and should return void. Interrupt handling functions should not be called directly. Interrupts can be handled directly with C/C++ functions by using the INTERRUPT pragma or the __interrupt keyword. For information about the INTERRUPT pragma, see Section 6.9.14. For information about the __interrupt keyword, see Section 6.5.3. 158 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Integer Expression Analysis 7.7 Integer Expression Analysis This section describes some special considerations to keep in mind when evaluating integer expressions. 7.7.1 Operations Evaluated With Run-Time-Support Calls The TMS320C28x does not directly support some C/C++ integer operations. Evaluating these operations is done with calls to run-time-support routines. These routines are hard-coded in assembly language. They are members of the object and source run-time-support libraries (rts2800_ml.lib and rtssrc.zip) in the toolset. The conventions for calling these routines are modeled on the standard C/C++ calling conventions. Operation Type Operations Evaluated With Run-Time-Support Calls 16-bit int Divide (signed) Modulus 32-bit long Divide (signed) Modulus 64-bit long long Multiply (1) Divide Bitwise AND, OR, and XOR Compare (1) 64-bit long long multiplies are inlined if -mf=5 is specified. 7.7.2 C/C++ Code Access to the Upper 16 Bits of 16-Bit Multiply The following methods provide access to the upper 16 bits of a 16-bit multiply in C/C++ language: • Signed-results method: int m1, m2; int result; result = ((long) m1 * (long) m2) >> 16; • Unsigned-results method: unsigned m1, m2; unsigned result; result = ((unsigned long) m1 * (unsigned long) m2) >> 16; Danger of Complicated Expressions NOTE: The compiler must recognize the structure of the expression for it to return the expected results. Avoid complicated expressions such as the following example: ((long)((unsigned)((a*b)+c)<5)*(long)(z*sin(w)>6))>>16 SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 159 Floating-Point Expression Analysis www.ti.com 7.8 Floating-Point Expression Analysis The C28x C/C++ compiler represents float and double floating-point values as IEEE single-precision numbers. Long double floating-point values are represented as IEEE double-precision numbers. Singleprecision floating-point numbers are represented as 32-bit values and double-precision floating-point numbers are represented as 64-bit values. The run-time-support library, rts2800_ml.lib, contains a set of floating-point math functions that support: • Addition, subtraction, multiplication, and division • Comparisons (>, <, >=, <=, ==, !=) • Conversions from integer or long to floating-point and floating-point to integer or long, both signed and unsigned • Standard error handling The conventions for calling these routines are the same as the conventions used to call the integer operation routines. Conversions are unary operations. 7.9 System Initialization Before you can run a C/C++ program, you must create the C/C++ run-time environment. The C/C++ boot routine performs this task using a function called c_int00 (or _c_int00). The run-time-support source library, rts.src, contains the source to this routine in a module named boot.c (or boot.asm). To begin running the system, the c_int00 function can be called by reset hardware. You must link the c_int00 function with the other object files. This occurs automatically when you use the --rom_model or -ram_model link option and include a standard run-time-support library as one of the linker input files. When C/C++ programs are linked, the linker sets the entry point value in the executable output file to the symbol c_int00. The c_int00 function performs the following tasks to initialize the environment: 1. Defines a section called .stack for the system stack and sets up the initial stack pointers 2. Initializes global variables by copying the data from the initialization tables to the storage allocated for the variables in the .ebss section. If you are initializing variables at load time (--ram_model option), a loader performs this step before the program runs (it is not performed by the boot routine). For more information, see Section 7.9.3. 3. Executes the global constructors found in the global constructors table. For more information, see Section 7.9.3.4. 4. Calls the main() function to run the C/C++ program You can replace or modify the boot routine to meet your system requirements. However, the boot routine must perform the operations listed above to correctly initialize the C/C++ environment. 160 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com System Initialization 7.9.1 Boot Hook Functions for System Pre-Initialization Boot hooks are points at which you may insert application functions into the C/C++ boot process. Default boot hook functions are provided with the run-time support (RTS) library. However, you can implement customized versions of these boot hook functions, which override the default boot hook functions in the RTS library if they are linked before the run-time library. Such functions can perform any applicationspecific initialization before continuing with the C/C++ environment setup. Note that the TI-RTOS operating system uses custom versions of the boot hook functions for system setup, so you should be careful about overriding these functions if you are using TI-RTOS. The following boot hook functions are available: _system_pre_init(): This function provides a place to perform application-specific initialization. It is invoked after the stack pointer is initialized but before any C/C++ environment setup is performed. By default, _system_pre_init() should return a non-zero value. The default C/C++ environment setup is bypassed if _system_pre_init() returns 0. _system_post_cinit(): This function is invoked during C/C++ environment setup, after C/C++ global data is initialized but before any C++ constructors are called. This function should not return a value. 7.9.2 Run-Time Stack The run-time stack is allocated in a single continuous block of memory and grows down from low addresses to higher addresses. The SP points to the top of the stack. The code does not check to see if the run-time stack overflows. Stack overflow occurs when the stack grows beyond the limits of the memory space that was allocated for it. Be sure to allocate adequate memory for the stack. The stack size can be changed at link time by using the --stack_size link option on the linker command line and specifying the stack size as a constant directly after the option. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 161 System Initialization www.ti.com 7.9.3 Automatic Initialization of Variables Some global variables must have initial values assigned to them before a C/C++ program starts running. The process of retrieving these variables' data and initializing the variables with the data is called autoinitialization. The compiler builds tables in a special section called .cinit that contains data for initializing global and static variables. Each compiled module contains these initialization tables. The linker combines them into a single table (a single .cinit section). The boot routine or a loader uses this table to initialize all the system variables. Initializing Variables NOTE: In ANSI/ISO C, global and static variables that are not explicitly initialized must be set to 0 before program execution. The C/C++ compiler does not perform any preinitialization of uninitialized variables. Explicitly initialize any variable that must have an initial value of 0. The easiest method is to have a loader clear the .ebss section before the program starts running. Another method is to set a fill value of 0 in the linker control map for the .ebss section. You cannot use these methods with code that is burned into ROM. Global variables are either autoinitialized at run time or at load time; see Section 7.9.3.2 and Section 7.9.3.3. Also see Section 6.12. 7.9.3.1 Initialization Tables The tables in the .cinit section consist of variable-size initialization records. Each variable that must be autoinitialized has a record in the .cinit section. Figure 7-2 shows the format of the .cinit section and the initialization records. Figure 7-2. Format of Initialization Records in the .cinit Section The fields of an initialization record contain the following information: • The first field of an initialization record contains the size (in words) of the initialization data. That is, this field contains the size of the third field. The size is specified as a negative value; this is legacy behavior, and the absolute value of this field is the size of the data. • The second field contains the starting address of the area within the .ebss section where the initialization data must be copied. The second field requires two words to hold the address. • The third field contains the data that is copied into the .ebss section to initialize the variable. The width of this field is variable. Each variable that must be autoinitialized has an initialization record in the .cinit section. 162 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com System Initialization Example 7-7 shows initialized global variables defined in C. Example 7-8 shows the corresponding initialization table. Example 7-7. Initialized Variables Defined in C int i= 23; far int j[2] = { 1,2}; Example 7-8. Initialized Information for Variables Defined in Example 7-7 .global _i .ebss _i,1,1,0 .global _j _j: .usect .ebss,2,1,0 .sect .align .field .field .field ".cinit" 1 1,16 _i+0,16 23,16 IR_1: .sect .align .field .field .field .field .set2 ".cinit" 1 -IR_1,16 _j+0,32 1,16 2,16 ; _i @ 0 ; _j[0] @ 0 ; _j[1] @ 16 The .cinit section must contain only initialization tables in this format. When interfacing assembly language modules, do not use the .cinit section for any other purpose. The table in the .pinit section simply consists of a list of addresses of constructors to be called (see Figure 7-3). The constructors appear in the table after the .cinit initialization. Figure 7-3. Format of Initialization Records in the .pinit Section .pinit section Address of constructor 1 Address of constructor 2 Address of constructor 3  • • Address of constructor n When you use the --rom_model or --ram_model option, the linker combines the .cinit sections from all the C modules and appends a null word to the end of the composite .cinit section. This terminating record appears as a record with a size field of 0 and marks the end of the initialization tables. Likewise, the --rom_model or --ram_model link option causes the linker to combine all of the .pinit sections from all C/C++ modules and append a null word to the end of the composite .pinit section. The boot routine knows the end of the global constructor table when it encounters a null constructor address. The const-qualified variables are initialized differently; see Section 6.5.1. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 163 System Initialization www.ti.com 7.9.3.2 Autoinitialization of Variables at Run Time Autoinitializing variables at run time is the default method of autoinitialization. To use this method, invoke the linker with the --rom_model option. Using this method, the .cinit section is loaded into memory along with all the other initialized sections, and global variables are initialized at run time. The linker defines a special symbol called cinit that points to the beginning of the initialization tables in memory. When the program begins running, the C/C++ boot routine copies data from the tables (pointed to by .cinit) into the specified variables in the .ebss section. This allows initialization data to be stored in ROM and copied to RAM each time the program starts. Figure 7-4 illustrates autoinitialization at run time. Use this method in any system where your application runs from code burned into ROM. Figure 7-4. Autoinitialization at Run Time 7.9.3.3 Initialization of Variables at Load Time Initialization of variables at load time enhances performance by reducing boot time and by saving the memory used by the initialization tables. To use this method, invoke the linker with the --ram_model option. When you use the --ram_model link option, the linker sets the STYP_COPY bit in the .cinit section's header. This tells the loader not to load the .cinit section into memory. (The .cinit section occupies no space in the memory map.) The linker also sets the cinit symbol to -1 (normally, cinit points to the beginning of the initialization tables). This indicates to the boot routine that the initialization tables are not present in memory; accordingly, no run-time initialization is performed at boot time. A loader (which is not part of the compiler package) must be able to perform the following tasks to use initialization at load time: • Detect the presence of the .cinit section in the object file • Determine that STYP_COPY is set in the .cinit section header, so that it knows not to copy the .cinit section into memory • Understand the format of the initialization tables 164 Run-Time Environment Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Figure 7-5 illustrates the initialization of variables at load time. Figure 7-5. Initialization at Load Time System Initialization Regardless of the use of the --rom_model or --ram_model options, the .pinit section is always loaded and processed at run time. 7.9.3.4 Global Constructors All global C++ variables that have constructors must have their constructor called before main(). The compiler builds a table in a section called .pinit of global constructor addresses that must be called, in order, before main(). The linker combines the .pinit section form each input file to form a single table in the .pinit section. The boot routine uses this table to execute the constructors. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated Run-Time Environment 165 Chapter 8 SPRU514N – June 2017 Using Run-Time-Support Functions and Building Libraries Some of the features of C/C++ (such as I/O, dynamic memory allocation, string operations, and trigonometric functions) are provided as an ANSI/ISO C/C++ standard library, rather than as part of the compiler itself. The TI implementation of this library is the run-time-support library (RTS). The C/C++ compiler implements the ISO standard library except for those facilities that handle exception conditions, signal, and locale issues (properties that depend on local language, nationality, or culture). Using the ANSI/ISO standard library ensures a consistent set of functions that provide for greater portability. In addition to the ANSI/ISO-specified functions, the run-time-support library includes routines that give you processor-specific commands and direct C language I/O requests. These are detailed in Section 8.1 and Section 8.2. A library-build utility is provided with the code generation tools that lets you create customized run-timesupport libraries. This process is described in Section 8.4 . Topic ........................................................................................................................... Page 8.1 C and C++ Run-Time Support Libraries ............................................................... 167 8.2 The C I/O Functions .......................................................................................... 169 8.3 Handling Reentrancy (_register_lock() and _register_unlock() Functions) ............... 181 8.4 Library-Build Process ....................................................................................... 182 166 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com C and C++ Run-Time Support Libraries 8.1 C and C++ Run-Time Support Libraries TMS320C28x compiler releases include pre-built run-time support (RTS) libraries that provide all the standard capabilities. Separate libraries are provided for FPU support and C++ exception support. See Section 8.1.7 for information on the library-naming conventions. The run-time-support library contains the following: • ANSI/ISO C/C++ standard library • C I/O library • Low-level support functions that provide I/O to the host operating system • Fundamental arithmetic routines • System startup routine, _c_int00 • Compiler helper functions (to support language features that are not directly efficiently expressible in C/C++) The run-time-support libraries do not contain functions involving signals and locale issues. The C++ library supports wide chars, in that template functions and classes that are defined for char are also available for wide char. For example, wide char stream classes wios, wiostream, wstreambuf and so on (corresponding to char classes ios, iostream, streambuf) are implemented. However, there is no lowlevel file I/O for wide chars. Also, the C library interface to wide char support (through the C++ headers and ) is limited as described in Section 6.1. TI does not provide documentation that covers the functionality of the C++ library. TI suggests referring to one of the following sources: • The Standard C++ Library: A Tutorial and Reference,Nicolai M. Josuttis, Addison-Wesley, ISBN 0-20137926-0 • The C++ Programming Language (Third or Special Editions), Bjarne Stroustrup, Addison-Wesley, ISBN 0-201-88954-4 or 0-201-70073-5 8.1.1 Linking Code With the Object Library When you link your program, you must specify the object library as one of the linker input files so that references to the I/O and run-time-support functions can be resolved. You can either specify the library or allow the compiler to select one for you. See Section 4.3.1 for further information. When a library is linked, the linker includes only those library members required to resolve undefined references. For more information about linking, see the TMS320C28x Assembly Language Tools User's Guide. C, C++, and mixed C and C++ programs can use the same run-time-support library. Run-time-support functions and variables that can be called and referenced from both C and C++ will have the same linkage. 8.1.2 Header Files You must use the header files provided with the compiler run-time support when using functions from C/C++ standard library. Set the C2000_C_DIR environment variable to the specific include directory: "include\lib". 8.1.3 Modifying a Library Function You can inspect or modify library functions by examining the source code in the lib/src subdirectory of the compiler installation. For example, C:\ti\ccsv6\tools\compiler\c2000_#.#.#\lib\src. One you have located the relevant source code, change the specific function file and rebuild the library. You can use this source tree to rebuild the rts2800_ml.lib library or to build a new library. See Section 8.4 for details on building. SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 167 Copyright © 2017, Texas Instruments Incorporated C and C++ Run-Time Support Libraries www.ti.com 8.1.4 Support for String Handling The library includes the header files and , which provide the following functions for string handling beyond those required. • string.h – strdup(), which duplicates a string • strings.h – bcmp(), which is equivalent to memcmp() – bcopy(), which is equivalent to memmove() – bzero(), which replaces memory locations with zero-valued bytes – ffs(), which finds the first bit set and returns the index of that bit – index(), which is equivalent to strchr() – rindex(), which is equivalent to strrchr() – strcasecmp() and strncasecmp(), which perform case-insensitive string comparisons 8.1.5 Minimal Support for Internationalization The library includes the header files , , and , which provide APIs to support non-ASCII character sets and conventions. Our implementation of these APIs is limited in the following ways: • The library has minimal support for wide and multibyte characters. The type wchar_t is implemented as int. The wide character set is equivalent to the set of values of type char. The library includes the header files and but does not include all the functions specified in the standard. So-called multibyte characters are limited to single characters. There are no shift states. The mapping between multibyte characters and wide characters is simple equivalence; that is, each wide character maps to and from exactly a single multibyte character having the same value. • The C library includes the header file but with a minimal implementation. The only supported locale is the C locale. That is, library behavior that is specified to vary by locale is hardcoded to the behavior of the C locale, and attempting to install a different locale via a call to setlocale() will return NULL. 8.1.6 Allowable Number of Open Files In the header file, the value for the macro FOPEN_MAX has the value of the macro _NFILE, which is set to 10. The impact is that you can only have 10 files simultaneously open at one time (including the pre-defined streams - stdin, stdout, stderr). The C standard requires that the minimum value for the FOPEN_MAX macro is 8. The macro determines the maximum number of files that can be opened at one time. The macro is defined in the stdio.h header file and can be modified by changing the value of the _NFILE macro and recompiling the library. 8.1.7 Library Naming Conventions By default, the linker uses automatic library selection to select the correct run-time-support library (see Section 4.3.1.1) for your application. If you select the library manually, you must select the matching library according to the following naming scheme: rts2800_ml.libor rts2800_fpu32.lib rts2800_ml The default C28x support library rts2800_fpu32 Indicates support for FPU targets 168 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com The C I/O Functions 8.2 The C I/O Functions The C I/O functions make it possible to access the host's operating system to perform I/O. The capability to perform I/O on the host gives you more options when debugging and testing code. The I/O functions are logically divided into layers: high level, low level, and device-driver level. With properly written device drivers, the C-standard high-level I/O functions can be used to perform I/O on custom user-defined devices. This provides an easy way to use the sophisticated buffering of the highlevel I/O functions on an arbitrary device. Debugger Required for Default HOST NOTE: For the default HOST device to work, there must be a debugger to handle the C I/O requests; the default HOST device cannot work by itself in an embedded system. To work in an embedded system, you will need to provide an appropriate driver for your system. NOTE: C I/O Mysteriously Fails If there is not enough space on the heap for a C I/O buffer, operations on the file will silently fail. If a call to printf() mysteriously fails, this may be the reason. The heap needs to be at least large enough to allocate a block of size BUFSIZ (defined in stdio.h) for every file on which I/O is performed, including stdout, stdin, and stderr, plus allocations performed by the user's code, plus allocation bookkeeping overhead. Alternately, declare a char array of size BUFSIZ and pass it to setvbuf to avoid dynamic allocation. To set the heap size, use the -heap_size option when linking (refer to the Linker Description chapter in the TMS320C28x Assembly Language Tools User's Guide). NOTE: Open Mysteriously Fails The run-time support limits the total number of open files to a small number relative to general-purpose processors. If you attempt to open more files than the maximum, you may find that the open will mysteriously fail. You can increase the number of open files by extracting the source code from rts.src and editing the constants controlling the size of some of the C I/O data structures. The macro _NFILE controls how many FILE (fopen) objects can be open at one time (stdin, stdout, and stderr count against this total). (See also FOPEN_MAX.) The macro _NSTREAM controls how many low-level file descriptors can be open at one time (the low-level files underlying stdin, stdout, and stderr count against this total). The macro _NDEVICE controls how many device drivers are installed at one time (the HOST device counts against this total). SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 169 Copyright © 2017, Texas Instruments Incorporated The C I/O Functions www.ti.com 8.2.1 High-Level I/O Functions The high-level functions are the standard C library of stream I/O routines (printf, scanf, fopen, getchar, and so on). These functions call one or more low-level I/O functions to carry out the high-level I/O request. The high-level I/O routines operate on FILE pointers, also called streams. Portable applications should use only the high-level I/O functions. To use the high-level I/O functions, include the header file stdio.h, or cstdio for C++ code, for each module that references a C I/O function. For example, given the following C program in a file named main.c: #include void main() { FILE *fid; fid = fopen("myfile","w"); fprintf(fid,"Hello, world\n"); fclose(fid); printf("Hello again, world\n"); } Issuing the following compiler command compiles, links, and creates the file main.out from the run-timesupport library: cl2000 main.c --run_linker --heap_size=400 --library=rts2800_ml.lib --output_file=main.out Executing main.out results in Hello, world being output to a file and Hello again, world being output to your host's stdout window. 170 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com The C I/O Functions 8.2.2 Overview of Low-Level I/O Implementation The low-level functions are comprised of seven basic I/O functions: open, read, write, close, lseek, rename, and unlink. These low-level routines provide the interface between the high-level functions and the device-level drivers that actually perform the I/O command on the specified device. The low-level functions are designed to be appropriate for all I/O methods, even those which are not actually disk files. Abstractly, all I/O channels can be treated as files, although some operations (such as lseek) may not be appropriate. See Section 8.2.3 for more details. The low-level functions are inspired by, but not identical to, the POSIX functions of the same names. The low-level functions operate on file descriptors. A file descriptor is an integer returned by open, representing an opened file. Multiple file descriptors may be associated with a file; each has its own independent file position indicator. open Open File for I/O Syntax Description Return Value #include int open (const char * path , unsigned flags , int file_descriptor ); The open function opens the file specified by path and prepares it for I/O. • The path is the filename of the file to be opened, including an optional directory path and an optional device specifier (see Section 8.2.5). • The flags are attributes that specify how the file is manipulated. The flags are specified using the following symbols: O_RDONLY O_WRONLY O_RDWR O_APPEND O_CREAT O_TRUNC O_BINARY (0x0000) (0x0001) (0x0002) (0x0008) (0x0200) (0x0400) (0x8000) /* open for reading */ /* open for writing */ /* open for read & write */ /* append on each write */ /* open with file create */ /* open with truncation */ /* open in binary mode */ Low-level I/O routines allow or disallow some operations depending on the flags used when the file was opened. Some flags may not be meaningful for some devices, depending on how the device implements files. • The file_descriptor is assigned by open to an opened file. The next available file descriptor is assigned to each new file opened. The function returns one of the following values: non-negative file descriptor if successful -1 on failure SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 171 Copyright © 2017, Texas Instruments Incorporated close — Close File for I/O close Close File for I/O Syntax Description Return Value #include int close (int file_descriptor ); The close function closes the file associated with file_descriptor. The file_descriptor is the number assigned by open to an opened file. The return value is one of the following: 0 if successful -1 on failure www.ti.com read Syntax Description Return Value Read Characters from a File #include int read (int file_descriptor , char * buffer , unsigned count ); The read function reads count characters into the buffer from the file associated with file_descriptor. • The file_descriptor is the number assigned by open to an opened file. • The buffer is where the read characters are placed. • The count is the number of characters to read from the file. The function returns one of the following values: 0 if EOF was encountered before any characters were read # number of characters read (may be less than count) -1 on failure write Syntax Description Return Value Write Characters to a File #include int write (int file_descriptor , const char * buffer , unsigned count ); The write function writes the number of characters specified by count from the buffer to the file associated with file_descriptor. • The file_descriptor is the number assigned by open to an opened file. • The buffer is where the characters to be written are located. • The count is the number of characters to write to the file. The function returns one of the following values: # number of characters written if successful (may be less than count) -1 on failure 172 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com lseek Syntax for C Description Return Value unlink Syntax Description Return Value lseek — Set File Position Indicator Set File Position Indicator #include off_t lseek (int file_descriptor , off_t offset , int origin ); The lseek function sets the file position indicator for the given file to a location relative to the specified origin. The file position indicator measures the position in characters from the beginning of the file. • The file_descriptor is the number assigned by open to an opened file. • The offset indicates the relative offset from the origin in characters. • The origin is used to indicate which of the base locations the offset is measured from. The origin must be one of the following macros: SEEK_SET (0x0000) Beginning of file SEEK_CUR (0x0001) Current value of the file position indicator SEEK_END (0x0002) End of file The return value is one of the following: # new value of the file position indicator if successful (off_t)-1 on failure Delete File #include int unlink (const char * path ); The unlink function deletes the file specified by path. Depending on the device, a deleted file may still remain until all file descriptors which have been opened for that file have been closed. See Section 8.2.3. The path is the filename of the file, including path information and optional device prefix. (See Section 8.2.5.) The function returns one of the following values: 0 if successful -1 on failure SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 173 Copyright © 2017, Texas Instruments Incorporated rename — Rename File rename Rename File Syntax for C Syntax for C++ Description #include { | } int rename (const char * old_name , const char * new_name ); #include { | } int std::rename (const char * old_name , const char * new_name ); The rename function changes the name of a file. • The old_name is the current name of the file. • The new_name is the new name for the file. www.ti.com NOTE: The optional device specified in the new name must match the device of the old name. If they do not match, a file copy would be required to perform the rename, and rename is not capable of this action. Return Value The function returns one of the following values: 0 if successful -1 on failure NOTE: Although rename is a low-level function, it is defined by the C standard and can be used by portable applications. 8.2.3 Device-Driver Level I/O Functions At the next level are the device-level drivers. They map directly to the low-level I/O functions. The default device driver is the HOST device driver, which uses the debugger to perform file operations. The HOST device driver is automatically used for the default C streams stdin, stdout, and stderr. The HOST device driver shares a special protocol with the debugger running on a host system so that the host can perform the C I/O requested by the program. Instructions for C I/O operations that the program wants to perform are encoded in a special buffer named _CIOBUF_ in the .cio section. The debugger halts the program at a special breakpoint (C$$IO$$), reads and decodes the target memory, and performs the requested operation. The result is encoded into _CIOBUF_, the program is resumed, and the target decodes the result. The HOST device is implemented with seven functions, HOSTopen, HOSTclose, HOSTread, HOSTwrite, HOSTlseek, HOSTunlink, and HOSTrename, which perform the encoding. Each function is called from the low-level I/O function with a similar name. A device driver is composed of seven required functions. Not all function need to be meaningful for all devices, but all seven must be defined. Here we show the names of all seven functions as starting with DEV, but you may choose any name except for HOST. 174 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com DEV_open Syntax Description Return Value DEV_open — Open File for I/O Open File for I/O int DEV_open (const char * path , unsigned flags , int llv_fd ); This function finds a file matching path and opens it for I/O as requested by flags. • The path is the filename of the file to be opened. If the name of a file passed to open has a device prefix, the device prefix will be stripped by open, so DEV_open will not see it. (See Section 8.2.5 for details on the device prefix.) • The flags are attributes that specify how the file is manipulated. The flags are specified using the following symbols: O_RDONLY O_WRONLY O_RDWR O_APPEND O_CREAT O_TRUNC O_BINARY (0x0000) (0x0001) (0x0002) (0x0008) (0x0200) (0x0400) (0x8000) /* open for reading */ /* open for writing */ /* open for read & write */ /* append on each write */ /* open with file create */ /* open with truncation */ /* open in binary mode */ See POSIX for further explanation of the flags. • The llv_fd is treated as a suggested low-level file descriptor. This is a historical artifact; newly-defined device drivers should ignore this argument. This differs from the low-level I/O open function. This function must arrange for information to be saved for each file descriptor, typically including a file position indicator and any significant flags. For the HOST version, all the bookkeeping is handled by the debugger running on the host machine. If the device uses an internal buffer, the buffer can be created when a file is opened, or the buffer can be created during a read or write. This function must return -1 to indicate an error if for some reason the file could not be opened; such as the file does not exist, could not be created, or there are too many files open. The value of errno may optionally be set to indicate the exact error (the HOST device does not set errno). Some devices might have special failure conditions; for instance, if a device is read-only, a file cannot be opened O_WRONLY. On success, this function must return a non-negative file descriptor unique among all open files handled by the specific device. The file descriptor need not be unique across devices. The device file descriptor is used only by low-level functions when calling the device-driver-level functions. The low-level function open allocates its own unique file descriptor for the high-level functions to call the low-level functions. Code that uses only high-level I/O functions need not be aware of these file descriptors. SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 175 Copyright © 2017, Texas Instruments Incorporated DEV_close — Close File for I/O www.ti.com DEV_close Syntax Description Return Value Close File for I/O int DEV_close (int dev_fd ); This function closes a valid open file descriptor. On some devices, DEV_close may need to be responsible for checking if this is the last file descriptor pointing to a file that was unlinked. If so, it is responsible for ensuring that the file is actually removed from the device and the resources reclaimed, if appropriate. This function should return -1 to indicate an error if the file descriptor is invalid in some way, such as being out of range or already closed, but this is not required. The user should not call close() with an invalid file descriptor. DEV_read Syntax Description Return Value Read Characters from a File int DEV_read (int dev_fd , char * buf , unsigned count ); The read function reads count bytes from the input file associated with dev_fd. • The dev_fd is the number assigned by open to an opened file. • The buf is where the read characters are placed. • The count is the number of characters to read from the file. This function must return -1 to indicate an error if for some reason no bytes could be read from the file. This could be because of an attempt to read from a O_WRONLY file, or for device-specific reasons. If count is 0, no bytes are read and this function returns 0. This function returns the number of bytes read, from 0 to count. 0 indicates that EOF was reached before any bytes were read. It is not an error to read less than count bytes; this is common if the are not enough bytes left in the file or the request was larger than an internal device buffer size. DEV_write Syntax Description Return Value Write Characters to a File int DEV_write (int dev_fd , const char * buf , unsigned count ); This function writes count bytes to the output file. • The dev_fd is the number assigned by open to an opened file. • The buffer is where the write characters are placed. • The count is the number of characters to write to the file. This function must return -1 to indicate an error if for some reason no bytes could be written to the file. This could be because of an attempt to read from a O_RDONLY file, or for device-specific reasons. 176 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com DEV_lseek Syntax Description Return Value DEV_unlink Syntax Description Return Value DEV_rename Syntax Description Return Value DEV_lseek — Set File Position Indicator Set File Position Indicator off_t lseek (int dev_fd , off_t offset , int origin ); This function sets the file's position indicator for this file descriptor as lseek. If lseek is supported, it should not allow a seek to before the beginning of the file, but it should support seeking past the end of the file. Such seeks do not change the size of the file, but if it is followed by a write, the file size will increase. If successful, this function returns the new value of the file position indicator. This function must return -1 to indicate an error if for some reason no bytes could be written to the file. For many devices, the lseek operation is nonsensical (e.g. a computer monitor). Delete File int DEV_unlink (const char * path ); Remove the association of the pathname with the file. This means that the file may no longer be opened using this name, but the file may not actually be immediately removed. Depending on the device, the file may be immediately removed, but for a device which allows open file descriptors to point to unlinked files, the file will not actually be deleted until the last file descriptor is closed. See Section 8.2.3. This function must return -1 to indicate an error if for some reason the file could not be unlinked (delayed removal does not count as a failure to unlink.) If successful, this function returns 0. Rename File int DEV_rename (const char * old_name , const char * new_name ); This function changes the name associated with the file. • The old_name is the current name of the file. • The new_name is the new name for the file. This function must return -1 to indicate an error if for some reason the file could not be renamed, such as the file doesn't exist, or the new name already exists. NOTE: It is inadvisable to allow renaming a file so that it is on a different device. In general this would require a whole file copy, which may be more expensive than you expect. If successful, this function returns 0. SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 177 Copyright © 2017, Texas Instruments Incorporated The C I/O Functions www.ti.com 8.2.4 Adding a User-Defined Device Driver for C I/O The function add_device allows you to add and use a device. When a device is registered with add_device, the high-level I/O routines can be used for I/O on that device. You can use a different protocol to communicate with any desired device and install that protocol using add_device; however, the HOST functions should not be modified. The default streams stdin, stdout, and stderr can be remapped to a file on a user-defined device instead of HOST by using freopen() as in Example 8-1. If the default streams are reopened in this way, the buffering mode will change to _IOFBF (fully buffered). To restore the default buffering behavior, call setvbuf on each reopened file with the appropriate value (_IOLBF for stdin and stdout, _IONBF for stderr). The default streams stdin, stdout, and stderr can be mapped to a file on a user-defined device instead of HOST by using freopen() as shown in Example 8-1. Each function must set up and maintain its own data structures as needed. Some function definitions perform no action and should just return. Example 8-1. Mapping Default Streams to Device #include #include #include "mydevice.h" void main() { add_device("mydevice", _MSA, MYDEVICE_open, MYDEVICE_close, MYDEVICE_read, MYDEVICE_write, MYDEVICE_lseek, MYDEVICE_unlink, MYDEVICE_rename); /*-----------------------------------------------------------------------*/ /* Re-open stderr as a MYDEVICE file */ /*-----------------------------------------------------------------------*/ if (!freopen("mydevice:stderrfile", "w", stderr)) { puts("Failed to freopen stderr"); exit(EXIT_FAILURE); } /*-----------------------------------------------------------------------*/ /* stderr should not be fully buffered; we want errors to be seen as */ /* soon as possible. Normally stderr is line-buffered, but this example */ /* doesn't buffer stderr at all. This means that there will be one call */ /* to write() for each character in the message. */ /*-----------------------------------------------------------------------*/ if (setvbuf(stderr, NULL, _IONBF, 0)) { puts("Failed to setvbuf stderr"); exit(EXIT_FAILURE); } /*-----------------------------------------------------------------------*/ /* Try it out! */ /*-----------------------------------------------------------------------*/ printf("This goes to stdout\n"); fprintf(stderr, "This goes to stderr\n"); } NOTE: Use Unique Function Names The function names open, read, write, close, lseek, rename, and unlink are used by the lowlevel routines. Use other names for the device-level functions that you write. 178 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com The C I/O Functions Use the low-level function add_device() to add your device to the device_table. The device table is a statically defined array that supports n devices, where n is defined by the macro _NDEVICE found in stdio.h/cstdio. The first entry in the device table is predefined to be the host device on which the debugger is running. The low-level routine add_device() finds the first empty position in the device table and initializes the device fields with the passed-in arguments. For a complete description, see the add_device function. 8.2.5 The device Prefix A file can be opened to a user-defined device driver by using a device prefix in the pathname. The device prefix is the device name used in the call to add_device followed by a colon. For example: FILE *fptr = fopen("mydevice:file1", "r"); int fd = open("mydevice:file2, O_RDONLY, 0); If no device prefix is used, the HOST device will be used to open the file. add_device Add Device to Device Table Syntax for C Defined in Description Return Value #include int add_device(char * name, unsigned flags , int (* dopen )(const char *path, unsigned flags, int llv_fd), int (* dclose )( int dev_fd), int (* dread )(intdev_fd, char *buf, unsigned count), int (* dwrite )(int dev_fd, const char *buf, unsigned count), off_t (* dlseek )(int dev_fd, off_t ioffset, int origin), int (* dunlink )(const char * path), int (* drename )(const char *old_name, const char *new_name)); lowlev.c (in the lib/src subdirectory of the compiler installation) The add_device function adds a device record to the device table allowing that device to be used for I/O from C. The first entry in the device table is predefined to be the HOST device on which the debugger is running. The function add_device() finds the first empty position in the device table and initializes the fields of the structure that represent a device. To open a stream on a newly added device use fopen( ) with a string of the format devicename : filename as the first argument. • The name is a character string denoting the device name. The name is limited to 8 characters. • The flags are device characteristics. The flags are as follows: _SSA Denotes that the device supports only one open stream at a time _MSA Denotes that the device supports multiple open streams More flags can be added by defining them in file.h. • The dopen, dclose, dread, dwrite, dlseek, dunlink, and drename specifiers are function pointers to the functions in the device driver that are called by the low-level functions to perform I/O on the specified device. You must declare these functions with the interface specified in Section 8.2.2. The device driver for the HOST that the TMS320C28x debugger is run on are included in the C I/O library. The function returns one of the following values: 0 if successful -1 on failure SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 179 Copyright © 2017, Texas Instruments Incorporated add_device — Add Device to Device Table www.ti.com Example Example 8-2 does the following: • Adds the device mydevice to the device table • Opens a file named test on that device and associates it with the FILE pointer fid • Writes the string Hello, world into the file • Closes the file Example 8-2 illustrates adding and using a device for C I/O: Example 8-2. Program for C I/O Device #include #include /****************************************************************************/ /* Declarations of the user-defined device drivers */ /****************************************************************************/ extern int MYDEVICE_open(const char *path, unsigned flags, int fno); extern int MYDEVICE_close(int fno); extern int MYDEVICE_read(int fno, char *buffer, unsigned count); extern int MYDEVICE_write(int fno, const char *buffer, unsigned count); extern off_t MYDEVICE_lseek(int fno, off_t offset, int origin); extern int MYDEVICE_unlink(const char *path); extern int MYDEVICE_rename(const char *old_name, char *new_name); main() { FILE *fid; add_device("mydevice", _MSA, MYDEVICE_open, MYDEVICE_close, MYDEVICE_read, MYDEVICE_write, MYDEVICE_lseek, MYDEVICE_unlink, MYDEVICE_rename); fid = fopen("mydevice:test","w"); fprintf(fid,"Hello, world\n"); fclose(fid); } 180 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Handling Reentrancy (_register_lock() and _register_unlock() Functions) 8.3 Handling Reentrancy (_register_lock() and _register_unlock() Functions) The C standard assumes only one thread of execution, with the only exception being extremely narrow support for signal handlers. The issue of reentrancy is avoided by not allowing you to do much of anything in a signal handler. However, SYS/BIOS applications have multiple threads which need to modify the same global program state, such as the CIO buffer, so reentrancy is a concern. Part of the problem of reentrancy remains your responsibility, but the run-time-support environment does provide rudimentary support for multi-threaded reentrancy by providing support for critical sections. This implementation does not protect you from reentrancy issues such as calling run-time-support functions from inside interrupts; this remains your responsibility. The run-time-support environment provides hooks to install critical section primitives. By default, a singlethreaded model is assumed, and the critical section primitives are not employed. In a multi-threaded system such as SYS/BIOS, the kernel arranges to install semaphore lock primitive functions in these hooks, which are then called when the run-time-support enters code that needs to be protected by a critical section. Throughout the run-time-support environment where a global state is accessed, and thus needs to be protected with a critical section, there are calls to the function _lock(). This calls the provided primitive, if installed, and acquires the semaphore before proceeding. Once the critical section is finished, _unlock() is called to release the semaphore. Usually SYS/BIOS is responsible for creating and installing the primitives, so you do not need to take any action. However, this mechanism can be used in multi-threaded applications that do not use the SYS/BIOS locking mechanism. You should not define the functions _lock() and _unlock() functions directly; instead, the installation functions are called to instruct the run-time-support environment to use these new primitives: void _register_lock (void ( *lock)()); void _register_unlock(void (*unlock)()); The arguments to _register_lock() and _register_unlock() should be functions which take no arguments and return no values, and which implement some sort of global semaphore locking: extern volatile sig_atomic_t *sema = SHARED_SEMAPHORE_LOCATION; static int sema_depth = 0; static void my_lock(void) { while (ATOMIC_TEST_AND_SET(sema, MY_UNIQUE_ID) != MY_UNIQUE_ID); sema_depth++; } static void my_unlock(void) { if (!--sema_depth) ATOMIC_CLEAR(sema); } The run-time-support nests calls to _lock(), so the primitives must keep track of the nesting level. SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 181 Copyright © 2017, Texas Instruments Incorporated Library-Build Process www.ti.com 8.4 Library-Build Process When using the C/C++ compiler, you can compile your code under a large number of different configurations and options that are not necessarily compatible with one another. Because it would be infeasible to include all possible run-time-support library variants, compiler releases pre-build only a small number of very commonly-used libraries such as rts2800_ml.lib. To provide maximum flexibility, the run-time-support source code is provided as part of each compiler release. You can build the missing libraries as desired. The linker can also automatically build missing libraries. This is accomplished with a new library build process, the core of which is the executable mklib, which is available beginning with CCS 5.1 8.4.1 Required Non-Texas Instruments Software To use the self-contained run-time-support build process to rebuild a library with custom options, the following are required: • sh (Bourne shell) • gmake (GNU make 3.81 or later) More information is available from GNU at http://www.gnu.org/software/make. GNU make (gmake) is also available in earlier versions of Code Composer Studio. GNU make is also included in some UNIX support packages for Windows, such as the MKS Toolkit, Cygwin, and Interix. The GNU make used on Windows platforms should explicitly report "This program build for Windows32" when the following is executed from the Command Prompt window: gmake -h All three of these programs are provided as a non-optional feature of CCS 5.1. They are also available as part of the optional XDC Tools feature if you are using an earlier version of CCS. The mklib program looks for these executables in the following order: 1. in your PATH 2. in the directory getenv("CCS_UTILS_DIR")/cygwin 3. in the directory getenv("CCS_UTILS_DIR")/bin 4. in the directory getenv("XDCROOT") 5. in the directory getenv("XDCROOT")/bin If you are invoking mklib from the command line, and these executables are not in your path, you must set the environment variable CCS_UTILS_DIR such that getenv("CCS_UTILS_DIR")/bin contains the correct programs. 8.4.2 Using the Library-Build Process You should normally let the linker automatically rebuild libraries as needed. If necessary, you can run mklib directly to populate libraries. See Section 8.4.2.2 for situations when you might want to do this. 8.4.2.1 Automatic Standard Library Rebuilding by the Linker The linker looks for run-time-support libraries primarily through the C2000_C_DIR environment variable. Typically, one of the pathnames in C2000_C_DIR is your install directory/lib, which contains all of the prebuilt libraries, as well as the index library libc.a. The linker looks in C2000_C_DIR to find a library that is the best match for the build attributes of the application. The build attributes are set indirectly according to the command-line options used to build the application. Build attributes include things like CPU revision. If the library name is explicitly specified (e.g. -library=rts2800_ml.lib), run-time support looks for that library exactly. If the library name is not specified, the linker uses the index library libc.a to pick an appropriate library. If the library is specified by path (e.g. –library=/foo/rts2800_ml.lib), it is assumed the library already exists and it will not be built automatically. The index library describes a set of libraries with different build attributes. The linker will compare the build attributes for each potential library with the build attributes of the application and will pick the best fit. For details on the index library, see the archiver chapter in the TMS320C28x Assembly Language Tools User's Guide. 182 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Library-Build Process Now that the linker has decided which library to use, it checks whether the run-time-support library is present in C2000_C_DIR . The library must be in exactly the same directory as the index library libc.a. If the library is not present, the linker invokes mklib to build it. This happens when the library is missing, regardless of whether the user specified the name of the library directly or allowed the linker to pick the best library from the index library. The mklib program builds the requested library and places it in 'lib' directory part of C2000_C_DIR in the same directory as the index library, so it is available for subsequent compilations. Things to watch out for: • The linker invokes mklib and waits for it to finish before finishing the link, so you will experience a onetime delay when an uncommonly-used library is built for the first time. Build times of 1-5 minutes have been observed. This depends on the power of the host (number of CPUs, etc). • In a shared installation, where an installation of the compiler is shared among more than one user, it is possible that two users might cause the linker to rebuild the same library at the same time. The mklib program tries to minimize the race condition, but it is possible one build will corrupt the other. In a shared environment, all libraries which might be needed should be built at install time; see Section 8.4.2.2 for instructions on invoking mklib directly to avoid this problem. • The index library must exist, or the linker is unable to rebuild libraries automatically. • The index library must be in a user-writable directory, or the library is not built. If the compiler installation must be installed read-only (a good practice for shared installation), any missing libraries must be built at installation time by invoking mklib directly. • The mklib program is specific to a certain version of a certain library; you cannot use one compiler version's run-time support's mklib to build a different compiler version's run-time support library. 8.4.2.2 Invoking mklib Manually You may need to invoke mklib directly in special circumstances: • The compiler installation directory is read-only or shared. • You want to build a variant of the run-time-support library that is not pre-configured in the index library libc.a or known to mklib. (e.g. a variant with source-level debugging turned on.) 8.4.2.2.1 Building Standard Libraries You can invoke mklib directly to build any or all of the libraries indexed in the index library libc.a. The libraries are built with the standard options for that library; the library names and the appropriate standard option sets are known to mklib. This is most easily done by changing the working directory to be the compiler run-time-support library directory 'lib' and invoking the mklib executable there: mklib --pattern=rts2800_ml.lib For C28x these libraries can be built: • rts2800_ml.lib (C/C++ run-time object library) • rts2800_fpu32.lib (C/C++ run-time object library for FPU targets) SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 183 Copyright © 2017, Texas Instruments Incorporated Library-Build Process www.ti.com 8.4.2.2.2 Shared or Read-Only Library Directory If the compiler tools are to be installed in shared or read-only directory, mklib cannot build the standard libraries at link time; the libraries must be built before the library directory is made shared or read-only. At installation time, the installing user must build all of the libraries which will be used by any user. To build all possible libraries, change the working directory to be the compiler RTS library directory 'lib' and invoke the mklib executable there: mklib --all Some targets have many libraries, so this step can take a long time. To build a subset of the libraries, invoke mklib individually for each desired library. 8.4.2.2.3 Building Libraries With Custom Options You can build a library with any extra custom options desired. This is useful for building a debugging version of the library, or with silicon exception workarounds enabled. The generated library is not a standard library, and must not be placed in the 'lib' directory. It should be placed in a directory local to the project which needs it. To build a debugging version of the library rts2800_ml.lib, change the working directory to the 'lib' directory and run the command: mklib --pattern=rts2800_ml.lib --name=rts2800_debug.lib --install_to=Project/Debug -extra_options="-g" 8.4.2.2.4 The mklib Program Option Summary Run the following command to see the full list of options. These are described in Table 8-1. mklib --help Option --index=filename --pattern=filename --all --install_to=directory --compiler_bin_dir= directory --name=filename --options='str' --extra_options='str' --list_libraries --log=filename --tmpdir=directory --gmake=filename --parallel=N --query=filename --help or --h --quiet or --q --verbose or --v Table 8-1. The mklib Program Options Effect The index library (libc.a) for this release. Used to find a template library for custom builds, and to find the source files (in the lib/src subdirectory of the compiler installation). REQUIRED. Pattern for building a library. If neither --extra_options nor --options are specified, the library will be the standard library with the standard options for that library. If either --extra_options or --options are specified, the library is a custom library with custom options. REQUIRED unless --all is used. Build all standard libraries at once. The directory into which to write the library. For a standard library, this defaults to the same directory as the index library (libc.a). For a custom library, this option is REQUIRED. The directory where the compiler executables are. When invoking mklib directly, the executables should be in the path, but if they are not, this option must be used to tell mklib where they are. This option is primarily for use when mklib is invoked by the linker. File name for the library with no directory part. Only useful for custom libraries. Options to use when building the library. The default options (see below) are replaced by this string. If this option is used, the library will be a custom library. Options to use when building the library. The default options (see below) are also used. If this option is used, the library will be a custom library. List the libraries this script is capable of building and exit. ordinary system-specific directory. Save the build log as filename. Use directory for scratch space instead of the ordinary system-specific directory. Gmake-compatible program to invoke instead of "gmake" Compile N files at once ("gmake -j N"). Does this script know how to build FILENAME? Display this help. Operate silently. Extra information to debug this executable. 184 Using Run-Time-Support Functions and Building Libraries Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Library-Build Process Examples: To build all standard libraries and place them in the compiler's library directory: mklib --all --index=$C_DIR/lib To build one standard library and place it in the compiler's library directory: mklib --pattern=rts2800_ml.lib --index=$C_DIR/lib To build a custom library that is just like rts2800_ml.lib, but has symbolic debugging support enabled: mklib --pattern=rts2800_ml.lib --extra_options="-g" --index=$C_DIR/lib -install_to=$Project/Debug --name=rts2800_debug.lib 8.4.3 Extending mklib The mklib API is a uniform interface that allows Code Composer Studio to build libraries without needing to know exactly what underlying mechanism is used to build it. Each library vendor (e.g. the TI compiler) provides a library-specific copy of 'mklib' in the library directory that can be invoked, which understands a standardized set of options, and understands how to build the library. This allows the linker to automatically build application-compatible versions of any vendor's library without needing to register the library in advance, as long as the vendor supports mklib. 8.4.3.1 Underlying Mechanism The underlying mechanism can be anything the vendor desires. For the compiler run-time-support libraries, mklib is just a wrapper that knows how to use the files in the lib/src subdirectory of the compiler installation and invoke gmake with the appropriate options to build each library. If necessary, mklib can be bypassed and the Makefile used directly, but this mode of operation is not supported by TI, and you are responsible for any changes to the Makefile. The format of the Makefile and the interface between mklib and the Makefile is subject to change without notice. The mklib program is the forward-compatible path. 8.4.3.2 Libraries From Other Vendors Any vendor who wishes to distribute a library that can be rebuilt automatically by the linker must provide: • An index library (like 'libc.a', but with a different name) • A copy of mklib specific to that library • A copy of the library source code (in whatever format is convenient) These things must be placed together in one directory that is part of the linker's library search path (specified either in C2000_C_DIR or with the linker --search_path option). If mklib needs extra information that is not possible to pass as command-line options to the compiler, the vendor will need to provide some other means of discovering the information (such as a configuration file written by a wizard run from inside CCS). The vendor-supplied mklib must at least accept all of the options listed in Table 8-1 without error, even if they do not do anything. SPRU514N – June 2017 Submit Documentation Feedback Using Run-Time-Support Functions and Building Libraries 185 Copyright © 2017, Texas Instruments Incorporated Chapter 9 SPRU514N – June 2017 C++ Name Demangler The C++ compiler implements function overloading, operator overloading, and type-safe linking by encoding a function's prototype and namespace in its link-level name. The process of encoding the prototype into the linkname is often referred to as name mangling. When you inspect mangled names, such as in assembly files, disassembler output, or compiler or linker diagnostic messages, it can be difficult to associate a mangled name with its corresponding name in the C++ source code. The C++ name demangler is a debugging aid that translates each mangled name it detects to its original name found in the C++ source code. These topics tell you how to invoke and use the C++ name demangler. The C++ name demangler reads in input, looking for mangled names. All unmangled text is copied to output unaltered. All mangled names are demangled before being copied to output. Topic ........................................................................................................................... Page 9.1 Invoking the C++ Name Demangler..................................................................... 187 9.2 C++ Name Demangler Options ........................................................................... 187 9.3 Sample Usage of the C++ Name Demangler ......................................................... 188 186 C++ Name Demangler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com 9.1 Invoking the C++ Name Demangler The syntax for invoking the C++ name demangler is: dem2000 [options ] [filenames] Invoking the C++ Name Demangler dem2000 options filenames Command that invokes the C++ name demangler. Options affect how the name demangler behaves. Options can appear anywhere on the command line. (Options are discussed in Section 9.2.) Text input files, such as the assembly file output by the compiler, the assembler listing file, the disassembly file, and the linker map file. If no filenames are specified on the command line, dem2000 uses standard input. By default, the C++ name demangler outputs to standard output. You can use the -o file option if you want to output to a file. 9.2 C++ Name Demangler Options The following options apply only to the C++ name demangler: --debug --diag_wrap[=on,off] --help --output= file --quiet -u -v Prints debug messages. Sets diagnostic messages to wrap at 79 columns (on, which is the default) or not (off). Prints a help screen that provides an online summary of the C++ name demangler options. Outputs to the specified file rather than to standard out. Reduces the number of messages generated during execution. Specifies that external names do not have a C++ prefix. Enables verbose mode (outputs a banner). SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated C++ Name Demangler 187 Sample Usage of the C++ Name Demangler www.ti.com 9.3 Sample Usage of the C++ Name Demangler The examples in this section illustrate the demangling process. Example 9-1 shows a sample C++ program. Example 9-2 shows the resulting assembly that is output by the compiler. In this example, the linknames of all the functions are mangled; that is, their signature information is encoded into their names. Example 9‑1. C++ Code int compute(int val, int *err); int foo(int val, int *err) { static int last_err = 0; int result = 0 if (last_err == 0) result = compute(val, &last_err); *err = last_err; return result; } Example 9‑2. Resulting Assembly for Example 9-1 ;*************************************************************** ;* FNAME: _foo_FiPi FR SIZE: 4 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 3 Auto, 0 SOE * ;*************************************************************** _foo_FiPi: ADDB SP,#4 MOVZ DP,#_last_err$1 MOV *-SP[1],AL MOV AL,@_last_err$1 MOV *-SP[2],AR4 MOV *-SP[3],#0 BF L1,NEQ ; branch occurs MOVL XAR4,#_last_err$1 MOV AL,*-SP[1] LCR #_compute__FiPi ; call occurs [#_compute__FiPi] MOV *-SP[3],AL L1: MOVZ AR6,*-SP[2] MOV *+XAR6[0],*(0:_last_err$1) MOV AL,*-SP[3] SUBB SP,#4 LRETR ; return occurs Executing the C++ name demangler demangles all names that it believes to be mangled. Enter: dem2000 foo.asm The result is shown in Example 9-3. The linknames in Example 9-2 foo( ) and compute( ) are demangled. 188 C++ Name Demangler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com Example 9‑3. Result After Running the C++ Name Demangler Sample Usage of the C++ Name Demangler ;*************************************************************** ;* FNAME: foo(int, int *) FR SIZE: 4 * ;* * ;* FUNCTION ENVIRONMENT * ;* * ;* FUNCTION PROPERTIES * ;* 0 Parameter, 3 Auto, 0 SOE * ;*************************************************************** foo(int, int *): ADDB SP,#4 MOVZ DP,#_last_err$1 MOV *-SP[1],AL MOV AL,@_last_err$1 MOV *-SP[2],AR4 MOV *-SP[3],#0 BF L1,NEQ ; branch occurs MOVL XAR4,#_last_err$1 MOV AL,*-SP[1] LCR #compute(int, int *) ; call occurs [#compute(int, int *)] MOV *-SP[3],AL L1: MOVZ AR6,*-SP[2] MOV *+XAR6[0],*(0:_last_err$1) MOV AL,*-SP[3] SUBB SP,#4 LRETR ; return occurs SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated C++ Name Demangler 189 Chapter 10 SPRU514N – June 2017 CLA Compiler The TMS320C28x Software Development toolset also includes support for compiling CLA C code. Because of CLA architecture and programming environment constraints, the C language support has some restrictions that are detailed in Section 10.2.2. As when including CLA assembly, compiled CLA code is linked together with C28x code to create a single executable. Topic ........................................................................................................................... Page 10.1 How to Invoke the CLA Compiler........................................................................ 191 10.2 CLA C Language Implementation ....................................................................... 192 190 CLA Compiler Copyright © 2017, Texas Instruments Incorporated SPRU514N – June 2017 Submit Documentation Feedback www.ti.com How to Invoke the CLA Compiler 10.1 How to Invoke the CLA Compiler The Control Law Accelerator (CLA) compiler is also invoked using the cl2000 command. Files that have a .cla extension are recognized by the compiler as CLA C files. The shell invokes separate CLA versions of the compiler passes to generate CLA-specific code. The --cla_support option is also required to assemble the output from the CLA code generator. If you use the --cla_default option, files with an extension of .c are also compiled as CLA files. Support is provided for Type 0, Type 1, and Type 2 CLA. The object files generated by the compile can then be linked with other C28x objects files to create a combined C28x/CLA program. To invoke the CLA compiler, enter: cl2000 --cla_support=[cla0|cla1|cla2] [other options] file.cla x.c y.cla cl2000 ac2000 accla opt2000 optcla cg2000 cgcla asm2000 C28x/CLA linker cmd lnk2000 prog.out Figure 10-1. CLA Compilation Overview Important facts about CLA C files: • Files with .cla extension compile using different parser, optimizer, and code generator. • cl2000 does not support compiling files using both the C and CLA compilers in a single invocation. • C28X/CLA builds requires linker command file changes to accommodate CLA compiler-generated sections and scratchpad area. SPRU514N – June 2017 Submit Documentation Feedback Copyright © 2017, Texas Instruments Incorporated CLA Compiler 191 CLA C Language Implementation 10.2 CLA C Language Implementation The CLA C language implementation requires changes from standard C. www.ti.com 10.2.1 Characteristics CLA implementation supports C only. The compiler produces an error if compiling for C++. These data types are supported: • Type char is 16 bits. • Type short is 16 bits. A char/short should be limited to load/store operations. • Types int, long, and long long are 32 bits. Note that the int size for CLA is 32 bits versus 16 bits on C28x. To avoid ambiguity when sharing data between CLA and C28x, it is strongly recommended to use type declarations that include size information (for example, int32 and uint16). • Types float, double, and long double are 32 bits. • Pointers are 16 bits. Pointer sizes for CLA are always 16-bits. This differs from C28x, which has 32-bit pointers. No 64-bit Types With CLA Compiler NOTE: The CLA compiler does not support 64-bit types. CLA accepts the C28x pragmas except for FAST_FUNC_CALL. The C standard library is not supported. The far and ioport keywords are not recognized. Access to the 'MMOV32 MSTF,mem32' and 'MMOV32 mem32,MSTF' instructions is provided using the cregister keyword. To access these MSTF instructions include the following declaration: extern cregister volatile unsigned int MSTF; The intrinsics listed in Table 10-1 are supported. Additionally, the run-time library functions abs() and fabs() are implemented as intrinsics. Table 10-1. C/C++ Compiler Intrinsics for CLA Intrinsic __mdebugstop() __meallow() __medis() float __meinvf32(float x); float __meisqrtf32(float x); short __mf32toi16r(float src); unsigned short __mf32toui16r(float src); float __mmaxf32( float x, float y); float __mminf32( float x, float y); float __mfracf32(float src); __mnop() Assembly Instruction(s) MDEBUGSTOP MEALLOW MEDIS MEINVF32 x MEISQRTF32 x MF32TOI16R dst, src MF32TOUI16R dst, src MMAXF32 dst, src MMINF32 dst, src MFRACF32 dst, src MNOP Description Debug stop Enable write access to EALLOW registers Disable write access to EALLOW registers Estimate 1/x to about 8 bits of precision. Estimate the square root of 1/x to about 8 bits of precision. Convert double to int and round. Convert double to unsigned int and round. If src>dst, copy src to dst If src

EEWORLD下载中心所有资源均来自网友分享，如有侵权，请发送举报邮件到客服邮箱bbs_service@eeworld.com.cn 或通过站内短信息或QQ：273568022联系管理员 高进，我们会尽快处理。