1 \input texinfo @c -*- texinfo -*-
3 @setfilename tcc-doc.info
4 @settitle Tiny C Compiler Reference Documentation
5 @dircategory Software development
7 * TCC: (tcc-doc). The Tiny C Compiler.
17 @center @titlefont{Tiny C Compiler Reference Documentation}
25 @node Top, Introduction, (dir), (dir)
26 @top Tiny C Compiler Reference Documentation
28 This manual documents version @value{VERSION} of the Tiny C Compiler.
31 * Introduction:: Introduction to tcc.
32 * Invoke:: Invocation of tcc (command line, options).
33 * Clang:: ANSI C and extensions.
34 * asm:: Assembler syntax.
35 * linker:: Output file generation and supported targets.
36 * Bounds:: Automatic bounds-checking of C code.
37 * Libtcc:: The libtcc library.
38 * devel:: Guide for Developers.
45 TinyCC (aka TCC) is a small but hyper fast C compiler. Unlike other C
46 compilers, it is meant to be self-relying: you do not need an
47 external assembler or linker because TCC does that for you.
49 TCC compiles so @emph{fast} that even for big projects @code{Makefile}s may
52 TCC not only supports ANSI C, but also most of the new ISO C99
53 standard and many GNUC extensions including inline assembly.
55 TCC can also be used to make @emph{C scripts}, i.e. pieces of C source
56 that you run as a Perl or Python script. Compilation is so fast that
57 your script will be as fast as if it was an executable.
59 TCC can also automatically generate memory and bound checks
60 (@pxref{Bounds}) while allowing all C pointers operations. TCC can do
61 these checks even if non patched libraries are used.
63 With @code{libtcc}, you can use TCC as a backend for dynamic code
64 generation (@pxref{Libtcc}).
66 TCC mainly supports the i386 target on Linux and Windows. There are alpha
67 ports for the ARM (@code{arm-tcc}) and the TMS320C67xx targets
68 (@code{c67-tcc}). More information about the ARM port is available at
69 @url{http://lists.gnu.org/archive/html/tinycc-devel/2003-10/msg00044.html}.
71 For usage on Windows, see also @url{tcc-win32.txt}.
74 @chapter Command line invocation
80 usage: tcc [options] [@var{infile1} @var{infile2}@dots{}] [@option{-run} @var{infile} @var{args}@dots{}]
85 @c man begin DESCRIPTION
86 TCC options are a very much like gcc options. The main difference is that TCC
87 can also execute directly the resulting program and give it runtime
90 Here are some examples to understand the logic:
93 @item @samp{tcc -run a.c}
94 Compile @file{a.c} and execute it directly
96 @item @samp{tcc -run a.c arg1}
97 Compile a.c and execute it directly. arg1 is given as first argument to
98 the @code{main()} of a.c.
100 @item @samp{tcc a.c -run b.c arg1}
101 Compile @file{a.c} and @file{b.c}, link them together and execute them. arg1 is given
102 as first argument to the @code{main()} of the resulting program.
104 Because multiple C files are specified, @option{--} are necessary to clearly
105 separate the program arguments from the TCC options.
108 @item @samp{tcc -o myprog a.c b.c}
109 Compile @file{a.c} and @file{b.c}, link them and generate the executable @file{myprog}.
111 @item @samp{tcc -o myprog a.o b.o}
112 link @file{a.o} and @file{b.o} together and generate the executable @file{myprog}.
114 @item @samp{tcc -c a.c}
115 Compile @file{a.c} and generate object file @file{a.o}.
117 @item @samp{tcc -c asmfile.S}
118 Preprocess with C preprocess and assemble @file{asmfile.S} and generate
119 object file @file{asmfile.o}.
121 @item @samp{tcc -c asmfile.s}
122 Assemble (but not preprocess) @file{asmfile.s} and generate object file
125 @item @samp{tcc -r -o ab.o a.c b.c}
126 Compile @file{a.c} and @file{b.c}, link them together and generate the object file @file{ab.o}.
132 TCC can be invoked from @emph{scripts}, just as shell scripts. You just
133 need to add @code{#!/usr/local/bin/tcc -run} at the start of your C source:
136 #!/usr/local/bin/tcc -run
141 printf("Hello World\n");
146 TCC can read C source code from @emph{standard input} when @option{-} is used in
147 place of @option{infile}. Example:
150 echo 'main()@{puts("hello");@}' | tcc -run -
154 @section Option summary
161 Generate an object file.
164 Put object file, executable, or dll into output file @file{outfile}.
166 @item -run source [args...]
167 Compile file @var{source} and run it with the command line arguments
168 @var{args}. In order to be able to give more than one argument to a
169 script, several TCC options can be given @emph{after} the
170 @option{-run} option, separated by spaces:
172 tcc "-run -L/usr/X11R6/lib -lX11" ex4.c
174 In a script, it gives the following header:
176 #!/usr/local/bin/tcc -run -L/usr/X11R6/lib -lX11
183 Show included files. As sole argument, print search dirs. -vvv shows tries too.
186 Display compilation statistics.
190 Preprocessor options:
194 Specify an additional include path. Include paths are searched in the
195 order they are specified.
197 System include paths are always searched after. The default system
198 include paths are: @file{/usr/local/include}, @file{/usr/include}
199 and @file{PREFIX/lib/tcc/include}. (@file{PREFIX} is usually
200 @file{/usr} or @file{/usr/local}).
203 Define preprocessor symbol @samp{sym} to
204 val. If val is not present, its value is @samp{1}. Function-like macros can
205 also be defined: @option{-DF(a)=a+1}
208 Undefine preprocessor symbol @samp{sym}.
211 Preprocess only, to stdout or file (with -o).
217 Note: each of the following options has a negative form beginning with
221 @item -funsigned-char
222 Let the @code{char} type be unsigned.
225 Let the @code{char} type be signed.
228 Do not generate common symbols for uninitialized data.
230 @item -fleading-underscore
231 Add a leading underscore at the beginning of each C symbol.
233 @item -fms-extensions
234 Allow a MS C compiler extensions to the language. Currently this
235 assumes a nested named structure declaration without an identifier
236 behaves like an unnamed one.
238 @item -fdollars-in-identifiers
239 Allow dollar signs in identifiers
247 Disable all warnings.
251 Note: each of the following warning options has a negative form beginning with
255 @item -Wimplicit-function-declaration
256 Warn about implicit function declaration.
259 Warn about unsupported GCC features that are ignored by TCC.
261 @item -Wwrite-strings
262 Make string constants be of type @code{const char *} instead of @code{char
266 Abort compilation if warnings are issued.
269 Activate all warnings, except @option{-Werror}, @option{-Wunusupported} and
270 @option{-Wwrite-strings}.
278 Specify an additional static library path for the @option{-l} option. The
279 default library paths are @file{/usr/local/lib}, @file{/usr/lib} and @file{/lib}.
282 Link your program with dynamic library libxxx.so or static library
283 libxxx.a. The library is searched in the paths specified by the
284 @option{-L} option and @env{LIBRARY_PATH} variable.
287 Set the path where the tcc internal libraries (and include files) can be
288 found (default is @file{PREFIX/lib/tcc}).
291 Generate a shared library instead of an executable.
294 set name for shared library to be used at runtime
297 Generate a statically linked executable (default is a shared linked
301 Export global symbols to the dynamic linker. It is useful when a library
302 opened with @code{dlopen()} needs to access executable symbols.
305 Generate an object file combining all input files.
307 @item -Wl,-rpath=path
308 Put custom search path for dynamic libraries into executable.
310 @item -Wl,--oformat=fmt
311 Use @var{fmt} as output format. The supported output formats are:
314 ELF output format (default)
316 Binary image (only for executable output)
318 COFF output format (only for executable output for TMS320C67xx target)
321 @item -Wl,-subsystem=console/gui/wince/...
322 Set type for PE (Windows) executables.
324 @item -Wl,-[Ttext=# | section-alignment=# | file-alignment=# | image-base=# | stack=#]
325 Modify executable layout.
330 @item -Wl,-(no-)whole-archive
331 Turn on/off linking of all objects in archives.
339 Generate run time debug information so that you get clear run time
340 error messages: @code{ test.c:68: in function 'test5()': dereferencing
341 invalid pointer} instead of the laconic @code{Segmentation
345 Generate additional support code to check
346 memory allocations and array/pointer bounds. @option{-g} is implied. Note
347 that the generated code is slower and bigger in this case.
349 Note: @option{-b} is only available on i386 when using libtcc for the moment.
352 Display N callers in stack traces. This is useful with @option{-g} or
361 Generate makefile fragment with dependencies.
364 Use @file{depfile} as output for -MD.
366 @item -print-search-dirs
367 Print the configured installation directory and a list of library
368 and include directories tcc will search.
375 Target specific options:
379 Use an algorithm for bitfield alignment consistent with MSVC. Default is
382 @item -mfloat-abi (ARM only)
383 Select the float ABI. Possible values: @code{softfp} and @code{hard}
386 Do not use sse registers on x86-64
389 Pass command line to the i386/x86-64 cross compiler.
393 Note: GCC options @option{-Ox}, @option{-fx} and @option{-mx} are
397 @c man begin ENVIRONMENT
398 Environment variables that affect how tcc operates.
404 A colon-separated list of directories searched for include files,
405 directories given with @option{-I} are searched first.
408 A colon-separated list of directories searched for libraries for the
409 @option{-l} option, directories given with @option{-L} are searched first.
418 @settitle Tiny C Compiler
432 @chapter C language support
436 TCC implements all the ANSI C standard, including structure bit fields
437 and floating point numbers (@code{long double}, @code{double}, and
438 @code{float} fully supported).
440 @section ISOC99 extensions
442 TCC implements many features of the new C standard: ISO C99. Currently
443 missing items are: complex and imaginary numbers.
445 Currently implemented ISOC99 features:
449 @item variable length arrays.
451 @item 64 bit @code{long long} types are fully supported.
453 @item The boolean type @code{_Bool} is supported.
455 @item @code{__func__} is a string variable containing the current
458 @item Variadic macros: @code{__VA_ARGS__} can be used for
459 function-like macros:
461 #define dprintf(level, __VA_ARGS__) printf(__VA_ARGS__)
465 @code{dprintf} can then be used with a variable number of parameters.
467 @item Declarations can appear anywhere in a block (as in C++).
469 @item Array and struct/union elements can be initialized in any order by
472 struct @{ int x, y; @} st[10] = @{ [0].x = 1, [0].y = 2 @};
474 int tab[10] = @{ 1, 2, [5] = 5, [9] = 9@};
477 @item Compound initializers are supported:
479 int *p = (int [])@{ 1, 2, 3 @};
481 to initialize a pointer pointing to an initialized array. The same
482 works for structures and strings.
484 @item Hexadecimal floating point constants are supported:
486 double d = 0x1234p10;
490 is the same as writing
492 double d = 4771840.0;
495 @item @code{inline} keyword is ignored.
497 @item @code{restrict} keyword is ignored.
500 @section GNU C extensions
502 TCC implements some GNU C extensions:
506 @item array designators can be used without '=':
508 int a[10] = @{ [0] 1, [5] 2, 3, 4 @};
511 @item Structure field designators can be a label:
513 struct @{ int x, y; @} st = @{ x: 1, y: 1@};
517 struct @{ int x, y; @} st = @{ .x = 1, .y = 1@};
520 @item @code{\e} is ASCII character 27.
522 @item case ranges : ranges can be used in @code{case}s:
526 printf("range 1 to 9\n");
529 printf("unexpected\n");
534 @cindex aligned attribute
535 @cindex packed attribute
536 @cindex section attribute
537 @cindex unused attribute
538 @cindex cdecl attribute
539 @cindex stdcall attribute
540 @cindex regparm attribute
541 @cindex dllexport attribute
543 @item The keyword @code{__attribute__} is handled to specify variable or
544 function attributes. The following attributes are supported:
547 @item @code{aligned(n)}: align a variable or a structure field to n bytes
548 (must be a power of two).
550 @item @code{packed}: force alignment of a variable or a structure field to
553 @item @code{section(name)}: generate function or data in assembly section
554 name (name is a string containing the section name) instead of the default
557 @item @code{unused}: specify that the variable or the function is unused.
559 @item @code{cdecl}: use standard C calling convention (default).
561 @item @code{stdcall}: use Pascal-like calling convention.
563 @item @code{regparm(n)}: use fast i386 calling convention. @var{n} must be
564 between 1 and 3. The first @var{n} function parameters are respectively put in
565 registers @code{%eax}, @code{%edx} and @code{%ecx}.
567 @item @code{dllexport}: export function from dll/executable (win32 only)
571 Here are some examples:
573 int a __attribute__ ((aligned(8), section(".mysection")));
577 align variable @code{a} to 8 bytes and put it in section @code{.mysection}.
580 int my_add(int a, int b) __attribute__ ((section(".mycodesection")))
587 generate function @code{my_add} in section @code{.mycodesection}.
589 @item GNU style variadic macros:
591 #define dprintf(fmt, args@dots{}) printf(fmt, ## args)
594 dprintf("one arg %d\n", 1);
597 @item @code{__FUNCTION__} is interpreted as C99 @code{__func__}
598 (so it has not exactly the same semantics as string literal GNUC
599 where it is a string literal).
601 @item The @code{__alignof__} keyword can be used as @code{sizeof}
602 to get the alignment of a type or an expression.
604 @item The @code{typeof(x)} returns the type of @code{x}.
605 @code{x} is an expression or a type.
607 @item Computed gotos: @code{&&label} returns a pointer of type
608 @code{void *} on the goto label @code{label}. @code{goto *expr} can be
609 used to jump on the pointer resulting from @code{expr}.
611 @item Inline assembly with asm instruction:
612 @cindex inline assembly
613 @cindex assembly, inline
616 static inline void * my_memcpy(void * to, const void * from, size_t n)
619 __asm__ __volatile__(
624 "1:\ttestb $1,%b4\n\t"
628 : "=&c" (d0), "=&D" (d1), "=&S" (d2)
629 :"0" (n/4), "q" (n),"1" ((long) to),"2" ((long) from)
637 TCC includes its own x86 inline assembler with a @code{gas}-like (GNU
638 assembler) syntax. No intermediate files are generated. GCC 3.x named
639 operands are supported.
641 @item @code{__builtin_types_compatible_p()} and @code{__builtin_constant_p()}
644 @item @code{#pragma pack} is supported for win32 compatibility.
648 @section TinyCC extensions
652 @item @code{__TINYC__} is a predefined macro to indicate that you use TCC.
654 @item @code{#!} at the start of a line is ignored to allow scripting.
656 @item Binary digits can be entered (@code{0b101} instead of
659 @item @code{__BOUNDS_CHECKING_ON} is defined if bound checking is activated.
664 @chapter TinyCC Assembler
666 Since version 0.9.16, TinyCC integrates its own assembler. TinyCC
667 assembler supports a gas-like syntax (GNU assembler). You can
668 desactivate assembler support if you want a smaller TinyCC executable
669 (the C compiler does not rely on the assembler).
671 TinyCC Assembler is used to handle files with @file{.S} (C
672 preprocessed assembler) and @file{.s} extensions. It is also used to
673 handle the GNU inline assembler with the @code{asm} keyword.
677 TinyCC Assembler supports most of the gas syntax. The tokens are the
682 @item C and C++ comments are supported.
684 @item Identifiers are the same as C, so you cannot use '.' or '$'.
686 @item Only 32 bit integer numbers are supported.
694 @item Integers in decimal, octal and hexa are supported.
696 @item Unary operators: +, -, ~.
698 @item Binary operators in decreasing priority order:
706 @item A value is either an absolute number or a label plus an offset.
707 All operators accept absolute values except '+' and '-'. '+' or '-' can be
708 used to add an offset to a label. '-' supports two labels only if they
709 are the same or if they are both defined and in the same section.
717 @item All labels are considered as local, except undefined ones.
719 @item Numeric labels can be used as local @code{gas}-like labels.
720 They can be defined several times in the same source. Use 'b'
721 (backward) or 'f' (forward) as suffix to reference them:
725 jmp 1b /* jump to '1' label before */
726 jmp 1f /* jump to '1' label after */
733 @cindex assembler directives
734 @cindex directives, assembler
735 @cindex align directive
736 @cindex skip directive
737 @cindex space directive
738 @cindex byte directive
739 @cindex word directive
740 @cindex short directive
741 @cindex int directive
742 @cindex long directive
743 @cindex quad directive
744 @cindex globl directive
745 @cindex global directive
746 @cindex section directive
747 @cindex text directive
748 @cindex data directive
749 @cindex bss directive
750 @cindex fill directive
751 @cindex org directive
752 @cindex previous directive
753 @cindex string directive
754 @cindex asciz directive
755 @cindex ascii directive
757 All directives are preceded by a '.'. The following directives are
761 @item .align n[,value]
762 @item .skip n[,value]
763 @item .space n[,value]
764 @item .byte value1[,...]
765 @item .word value1[,...]
766 @item .short value1[,...]
767 @item .int value1[,...]
768 @item .long value1[,...]
769 @item .quad immediate_value1[,...]
772 @item .section section
776 @item .fill repeat[,size[,value]]
779 @item .string string[,...]
780 @item .asciz string[,...]
781 @item .ascii string[,...]
784 @section X86 Assembler
787 All X86 opcodes are supported. Only ATT syntax is supported (source
788 then destination operand order). If no size suffix is given, TinyCC
789 tries to guess it from the operand sizes.
791 Currently, MMX opcodes are supported but not SSE ones.
794 @chapter TinyCC Linker
797 @section ELF file generation
800 TCC can directly output relocatable ELF files (object files),
801 executable ELF files and dynamic ELF libraries without relying on an
804 Dynamic ELF libraries can be output but the C compiler does not generate
805 position independent code (PIC). It means that the dynamic library
806 code generated by TCC cannot be factorized among processes yet.
808 TCC linker eliminates unreferenced object code in libraries. A single pass is
809 done on the object and library list, so the order in which object files and
810 libraries are specified is important (same constraint as GNU ld). No grouping
811 options (@option{--start-group} and @option{--end-group}) are supported.
813 @section ELF file loader
815 TCC can load ELF object files, archives (.a files) and dynamic
818 @section PE-i386 file generation
821 TCC for Windows supports the native Win32 executable file format (PE-i386). It
822 generates EXE files (console and gui) and DLL files.
824 For usage on Windows, see also tcc-win32.txt.
826 @section GNU Linker Scripts
827 @cindex scripts, linker
828 @cindex linker scripts
829 @cindex GROUP, linker command
830 @cindex FILE, linker command
831 @cindex OUTPUT_FORMAT, linker command
832 @cindex TARGET, linker command
834 Because on many Linux systems some dynamic libraries (such as
835 @file{/usr/lib/libc.so}) are in fact GNU ld link scripts (horrible!),
836 the TCC linker also supports a subset of GNU ld scripts.
838 The @code{GROUP} and @code{FILE} commands are supported. @code{OUTPUT_FORMAT}
839 and @code{TARGET} are ignored.
841 Example from @file{/usr/lib/libc.so}:
844 Use the shared library, but some functions are only in
845 the static library, so try that secondarily. */
846 GROUP ( /lib/libc.so.6 /usr/lib/libc_nonshared.a )
850 @chapter TinyCC Memory and Bound checks
852 @cindex memory checks
854 This feature is activated with the @option{-b} (@pxref{Invoke}).
856 Note that pointer size is @emph{unchanged} and that code generated
857 with bound checks is @emph{fully compatible} with unchecked
858 code. When a pointer comes from unchecked code, it is assumed to be
859 valid. Even very obscure C code with casts should work correctly.
861 For more information about the ideas behind this method, see
862 @url{http://www.doc.ic.ac.uk/~phjk/BoundsChecking.html}.
864 Here are some examples of caught errors:
868 @item Invalid range with standard string function:
876 @item Out of bounds-error in global or local arrays:
886 @item Out of bounds-error in malloc'ed data:
890 tab = malloc(20 * sizeof(int));
898 @item Access of freed memory:
902 tab = malloc(20 * sizeof(int));
914 tab = malloc(20 * sizeof(int));
923 @chapter The @code{libtcc} library
925 The @code{libtcc} library enables you to use TCC as a backend for
926 dynamic code generation.
928 Read the @file{libtcc.h} to have an overview of the API. Read
929 @file{libtcc_test.c} to have a very simple example.
931 The idea consists in giving a C string containing the program you want
932 to compile directly to @code{libtcc}. Then you can access to any global
933 symbol (function or variable) defined.
936 @chapter Developer's guide
938 This chapter gives some hints to understand how TCC works. You can skip
939 it if you do not intend to modify the TCC code.
941 @section File reading
943 The @code{BufferedFile} structure contains the context needed to read a
944 file, including the current line number. @code{tcc_open()} opens a new
945 file and @code{tcc_close()} closes it. @code{inp()} returns the next
950 @code{next()} reads the next token in the current
951 file. @code{next_nomacro()} reads the next token without macro
954 @code{tok} contains the current token (see @code{TOK_xxx})
955 constants. Identifiers and keywords are also keywords. @code{tokc}
956 contains additional infos about the token (for example a constant value
957 if number or string token).
961 The parser is hardcoded (yacc is not necessary). It does only one pass,
966 @item For initialized arrays with unknown size, a first pass
967 is done to count the number of elements.
969 @item For architectures where arguments are evaluated in
970 reverse order, a first pass is done to reverse the argument order.
976 The types are stored in a single 'int' variable. It was chosen in the
977 first stages of development when tcc was much simpler. Now, it may not
978 be the best solution.
981 #define VT_INT 0 /* integer type */
982 #define VT_BYTE 1 /* signed byte type */
983 #define VT_SHORT 2 /* short type */
984 #define VT_VOID 3 /* void type */
985 #define VT_PTR 4 /* pointer */
986 #define VT_ENUM 5 /* enum definition */
987 #define VT_FUNC 6 /* function type */
988 #define VT_STRUCT 7 /* struct/union definition */
989 #define VT_FLOAT 8 /* IEEE float */
990 #define VT_DOUBLE 9 /* IEEE double */
991 #define VT_LDOUBLE 10 /* IEEE long double */
992 #define VT_BOOL 11 /* ISOC99 boolean type */
993 #define VT_LLONG 12 /* 64 bit integer */
994 #define VT_LONG 13 /* long integer (NEVER USED as type, only
996 #define VT_BTYPE 0x000f /* mask for basic type */
997 #define VT_UNSIGNED 0x0010 /* unsigned type */
998 #define VT_ARRAY 0x0020 /* array type (also has VT_PTR) */
999 #define VT_VLA 0x20000 /* VLA type (also has VT_PTR and VT_ARRAY) */
1000 #define VT_BITFIELD 0x0040 /* bitfield modifier */
1001 #define VT_CONSTANT 0x0800 /* const modifier */
1002 #define VT_VOLATILE 0x1000 /* volatile modifier */
1003 #define VT_DEFSIGN 0x2000 /* signed type */
1005 #define VT_STRUCT_SHIFT 18 /* structure/enum name shift (14 bits left) */
1008 When a reference to another type is needed (for pointers, functions and
1009 structures), the @code{32 - VT_STRUCT_SHIFT} high order bits are used to
1010 store an identifier reference.
1012 The @code{VT_UNSIGNED} flag can be set for chars, shorts, ints and long
1015 Arrays are considered as pointers @code{VT_PTR} with the flag
1016 @code{VT_ARRAY} set. Variable length arrays are considered as special
1017 arrays and have flag @code{VT_VLA} set instead of @code{VT_ARRAY}.
1019 The @code{VT_BITFIELD} flag can be set for chars, shorts, ints and long
1020 longs. If it is set, then the bitfield position is stored from bits
1021 VT_STRUCT_SHIFT to VT_STRUCT_SHIFT + 5 and the bit field size is stored
1022 from bits VT_STRUCT_SHIFT + 6 to VT_STRUCT_SHIFT + 11.
1024 @code{VT_LONG} is never used except during parsing.
1026 During parsing, the storage of an object is also stored in the type
1030 #define VT_EXTERN 0x00000080 /* extern definition */
1031 #define VT_STATIC 0x00000100 /* static variable */
1032 #define VT_TYPEDEF 0x00000200 /* typedef definition */
1033 #define VT_INLINE 0x00000400 /* inline definition */
1034 #define VT_IMPORT 0x00004000 /* win32: extern data imported from dll */
1035 #define VT_EXPORT 0x00008000 /* win32: data exported from dll */
1036 #define VT_WEAK 0x00010000 /* win32: data exported from dll */
1041 All symbols are stored in hashed symbol stacks. Each symbol stack
1042 contains @code{Sym} structures.
1044 @code{Sym.v} contains the symbol name (remember
1045 an idenfier is also a token, so a string is never necessary to store
1046 it). @code{Sym.t} gives the type of the symbol. @code{Sym.r} is usually
1047 the register in which the corresponding variable is stored. @code{Sym.c} is
1048 usually a constant associated to the symbol like its address for normal
1049 symbols, and the number of entries for symbols representing arrays.
1050 Variable length array types use @code{Sym.c} as a location on the stack
1051 which holds the runtime sizeof for the type.
1053 Four main symbol stacks are defined:
1058 for the macros (@code{#define}s).
1061 for the global variables, functions and types.
1064 for the local variables, functions and types.
1066 @item global_label_stack
1067 for the local labels (for @code{goto}).
1070 for GCC block local labels (see the @code{__label__} keyword).
1074 @code{sym_push()} is used to add a new symbol in the local symbol
1075 stack. If no local symbol stack is active, it is added in the global
1078 @code{sym_pop(st,b)} pops symbols from the symbol stack @var{st} until
1079 the symbol @var{b} is on the top of stack. If @var{b} is NULL, the stack
1082 @code{sym_find(v)} return the symbol associated to the identifier
1083 @var{v}. The local stack is searched first from top to bottom, then the
1088 The generated code and datas are written in sections. The structure
1089 @code{Section} contains all the necessary information for a given
1090 section. @code{new_section()} creates a new section. ELF file semantics
1091 is assumed for each section.
1093 The following sections are predefined:
1098 is the section containing the generated code. @var{ind} contains the
1099 current position in the code section.
1102 contains initialized data
1105 contains uninitialized data
1107 @item bounds_section
1108 @itemx lbounds_section
1109 are used when bound checking is activated
1112 @itemx stabstr_section
1113 are used when debugging is active to store debug information
1115 @item symtab_section
1116 @itemx strtab_section
1117 contain the exported symbols (currently only used for debugging).
1121 @section Code generation
1122 @cindex code generation
1124 @subsection Introduction
1126 The TCC code generator directly generates linked binary code in one
1127 pass. It is rather unusual these days (see gcc for example which
1128 generates text assembly), but it can be very fast and surprisingly
1131 The TCC code generator is register based. Optimization is only done at
1132 the expression level. No intermediate representation of expression is
1133 kept except the current values stored in the @emph{value stack}.
1135 On x86, three temporary registers are used. When more registers are
1136 needed, one register is spilled into a new temporary variable on the stack.
1138 @subsection The value stack
1139 @cindex value stack, introduction
1141 When an expression is parsed, its value is pushed on the value stack
1142 (@var{vstack}). The top of the value stack is @var{vtop}. Each value
1143 stack entry is the structure @code{SValue}.
1145 @code{SValue.t} is the type. @code{SValue.r} indicates how the value is
1146 currently stored in the generated code. It is usually a CPU register
1147 index (@code{REG_xxx} constants), but additional values and flags are
1151 #define VT_CONST 0x00f0
1152 #define VT_LLOCAL 0x00f1
1153 #define VT_LOCAL 0x00f2
1154 #define VT_CMP 0x00f3
1155 #define VT_JMP 0x00f4
1156 #define VT_JMPI 0x00f5
1157 #define VT_LVAL 0x0100
1158 #define VT_SYM 0x0200
1159 #define VT_MUSTCAST 0x0400
1160 #define VT_MUSTBOUND 0x0800
1161 #define VT_BOUNDED 0x8000
1162 #define VT_LVAL_BYTE 0x1000
1163 #define VT_LVAL_SHORT 0x2000
1164 #define VT_LVAL_UNSIGNED 0x4000
1165 #define VT_LVAL_TYPE (VT_LVAL_BYTE | VT_LVAL_SHORT | VT_LVAL_UNSIGNED)
1171 indicates that the value is a constant. It is stored in the union
1172 @code{SValue.c}, depending on its type.
1175 indicates a local variable pointer at offset @code{SValue.c.i} in the
1179 indicates that the value is actually stored in the CPU flags (i.e. the
1180 value is the consequence of a test). The value is either 0 or 1. The
1181 actual CPU flags used is indicated in @code{SValue.c.i}.
1183 If any code is generated which destroys the CPU flags, this value MUST be
1184 put in a normal register.
1188 indicates that the value is the consequence of a conditional jump. For VT_JMP,
1189 it is 1 if the jump is taken, 0 otherwise. For VT_JMPI it is inverted.
1191 These values are used to compile the @code{||} and @code{&&} logical
1194 If any code is generated, this value MUST be put in a normal
1195 register. Otherwise, the generated code won't be executed if the jump is
1199 is a flag indicating that the value is actually an lvalue (left value of
1200 an assignment). It means that the value stored is actually a pointer to
1203 Understanding the use @code{VT_LVAL} is very important if you want to
1204 understand how TCC works.
1207 @itemx VT_LVAL_SHORT
1208 @itemx VT_LVAL_UNSIGNED
1209 if the lvalue has an integer type, then these flags give its real
1210 type. The type alone is not enough in case of cast optimisations.
1213 is a saved lvalue on the stack. @code{VT_LVAL} must also be set with
1214 @code{VT_LLOCAL}. @code{VT_LLOCAL} can arise when a @code{VT_LVAL} in
1215 a register has to be saved to the stack, or it can come from an
1216 architecture-specific calling convention.
1219 indicates that a cast to the value type must be performed if the value
1220 is used (lazy casting).
1223 indicates that the symbol @code{SValue.sym} must be added to the constant.
1227 are only used for optional bound checking.
1231 @subsection Manipulating the value stack
1234 @code{vsetc()} and @code{vset()} pushes a new value on the value
1235 stack. If the previous @var{vtop} was stored in a very unsafe place(for
1236 example in the CPU flags), then some code is generated to put the
1237 previous @var{vtop} in a safe storage.
1239 @code{vpop()} pops @var{vtop}. In some cases, it also generates cleanup
1240 code (for example if stacked floating point registers are used as on
1243 The @code{gv(rc)} function generates code to evaluate @var{vtop} (the
1244 top value of the stack) into registers. @var{rc} selects in which
1245 register class the value should be put. @code{gv()} is the @emph{most
1246 important function} of the code generator.
1248 @code{gv2()} is the same as @code{gv()} but for the top two stack
1251 @subsection CPU dependent code generation
1252 @cindex CPU dependent
1253 See the @file{i386-gen.c} file to have an example.
1258 must generate the code needed to load a stack value into a register.
1261 must generate the code needed to store a register into a stack value
1265 @itemx gfunc_param()
1267 should generate a function call
1269 @item gfunc_prolog()
1270 @itemx gfunc_epilog()
1271 should generate a function prolog/epilog.
1274 must generate the binary integer operation @var{op} on the two top
1275 entries of the stack which are guaranted to contain integer types.
1277 The result value should be put on the stack.
1280 same as @code{gen_opi()} for floating point operations. The two top
1281 entries of the stack are guaranted to contain floating point values of
1284 @item gen_cvt_itof()
1285 integer to floating point conversion.
1287 @item gen_cvt_ftoi()
1288 floating point to integer conversion.
1290 @item gen_cvt_ftof()
1291 floating point to floating point of different size conversion.
1293 @item gen_bounded_ptr_add()
1294 @item gen_bounded_ptr_deref()
1295 are only used for bounds checking.
1299 @section Optimizations done
1300 @cindex optimizations
1301 @cindex constant propagation
1302 @cindex strength reduction
1303 @cindex comparison operators
1304 @cindex caching processor flags
1305 @cindex flags, caching
1306 @cindex jump optimization
1307 Constant propagation is done for all operations. Multiplications and
1308 divisions are optimized to shifts when appropriate. Comparison
1309 operators are optimized by maintaining a special cache for the
1310 processor flags. &&, || and ! are optimized by maintaining a special
1311 'jump target' value. No other jump optimization is currently performed
1312 because it would require to store the code in a more abstract fashion.
1314 @unnumbered Concept Index
1321 @c texinfo-column-for-description: 32