Paze-C-CPP-Compiler/pcc-doc.texi

\input texinfo @c -*- texinfo -*-
@c %**start of header
@setfilename pcc-doc.info
@settitle Paze C Compiler Reference Documentation
@dircategory Software development
@direntry
* PCC: (pcc-doc).               The Paze C Compiler.
@end direntry
@c %**end of header

@include config.texi

@iftex
@titlepage
@afourpaper
@sp 7
@center @titlefont{Paze C Compiler Reference Documentation}
@sp 3
@end titlepage
@headings double
@end iftex

@contents

@node Top, Introduction, (dir), (dir)
@top Paze C Compiler Reference Documentation

This manual documents version @value{VERSION} of the Paze C Compiler.

@menu
* Introduction::                Introduction to pcc.
* Invoke::                      Invocation of pcc (command line, options).
* Clang::                       ANSI C and extensions.
* asm::                         Assembler syntax.
* linker::                      Output file generation and supported targets.
* Bounds::                      Automatic bounds-checking of C code.
* Libtcc::                      The libpcc library.
* devel::                       Guide for Developers.
@end menu


@node Introduction
@chapter Introduction

PazeCC (aka PCC) is a small but hyper fast C compiler. Unlike other C
compilers, it is meant to be self-relying: you do not need an
external assembler or linker because PCC does that for you.

PCC compiles so @emph{fast} that even for big projects @code{Makefile}s may
not be necessary.

PCC not only supports ANSI C, but also most of the ISO C99 standard,
many ISO C11 features, and many GNUC extensions including inline assembly.

PCC can also be used to make @emph{C scripts}, i.e. pieces of C source
that you run as a Perl or Python script. Compilation is so fast that
your script will be as fast as if it was an executable.

PCC can also automatically generate memory and bound checks
(@pxref{Bounds}) while allowing all C pointers operations. PCC can do
these checks even if non patched libraries are used.

With @code{libpcc}, you can use PCC as a backend for dynamic code
generation (@pxref{Libtcc}).

PCC supports the following target and platform combinations:

@multitable @columnfractions .15 .10 .12 .11 .11 .18 .09
@item @strong{OS} @tab @strong{i386} @tab @strong{x86-64} @tab @strong{ARM} @tab @strong{ARM64} @tab @strong{RISC-V 64} @tab @strong{C67}
@item Linux     @tab yes @tab yes @tab yes   @tab yes @tab yes @tab yes
@item macOS     @tab     @tab yes @tab       @tab yes @tab     @tab
@item Windows   @tab yes @tab yes @tab WinCE @tab yes @tab     @tab
@item Android   @tab yes @tab yes @tab yes   @tab yes @tab     @tab
@item FreeBSD   @tab     @tab yes @tab       @tab yes @tab     @tab
@item NetBSD    @tab     @tab yes @tab yes   @tab yes @tab     @tab
@item OpenBSD   @tab yes @tab yes @tab       @tab yes @tab yes @tab
@item DragonFly @tab     @tab yes @tab       @tab     @tab     @tab
@end multitable

The TMS320C67xx (C67) target is a cross compiler for the digital signal
processors of that family. It outputs COFF instead of ELF and has neither
@file{libpcc1.a} nor bound checking.

For usage on Windows, see also @url{pcc-win32.txt}.

@node Invoke
@chapter Command line invocation

@section Quick start

@example
@c man begin SYNOPSIS
usage: pcc [options] [@var{infile1} @var{infile2}@dots{}] [@option{-run} @var{infile} @var{args}@dots{}]
@c man end
@end example

@noindent
@c man begin DESCRIPTION
PCC options are very much like gcc options. The main difference is that PCC
can also execute directly the resulting program and give it runtime
arguments.

Here are some examples to understand the logic:

@table @code
@item @samp{pcc -run a.c}
Compile @file{a.c} and execute it directly

@item @samp{pcc -run a.c arg1}
Compile a.c and execute it directly. arg1 is given as first argument to
the @code{main()} of a.c.

@item @samp{pcc a.c -run b.c arg1}
Compile @file{a.c} and @file{b.c}, link them together and execute them. arg1 is given
as first argument to the @code{main()} of the resulting program.
@ignore
Because multiple C files are specified, @option{--} are necessary to clearly
separate the program arguments from the PCC options.
@end ignore

@item @samp{pcc -o myprog a.c b.c}
Compile @file{a.c} and @file{b.c}, link them and generate the executable @file{myprog}.

@item @samp{pcc -o myprog a.o b.o}
link @file{a.o} and @file{b.o} together and generate the executable @file{myprog}.

@item @samp{pcc -c a.c}
Compile @file{a.c} and generate object file @file{a.o}.

@item @samp{pcc -c asmfile.S}
Preprocess with C preprocess and assemble @file{asmfile.S} and generate
object file @file{asmfile.o}.

@item @samp{pcc -c asmfile.s}
Assemble (but not preprocess) @file{asmfile.s} and generate object file
@file{asmfile.o}.

@item @samp{pcc -r -o ab.o a.c b.c}
Compile @file{a.c} and @file{b.c}, link them together and generate the object file @file{ab.o}.

@end table

Scripting:

PCC can be invoked from @emph{scripts}, just as shell scripts. You just
need to add @code{#!/usr/local/bin/pcc -run} at the start of your C source:

@example
#!/usr/local/bin/pcc -run
#include <stdio.h>

int main()
@{
    printf("Hello World\n");
    return 0;
@}
@end example

PCC can read C source code from @emph{standard input} when @option{-} is used in
place of @option{infile}. Example:

@example
echo 'main()@{puts("hello");@}' | pcc -run -
@end example
@c man end

@section Option summary

@c man begin OPTIONS
@subheading General Options

@table @option
@item -c
Generate an object file.

@item -o outfile
Put object file, executable, or dll into output file @file{outfile}.

@item -run source [args...]
Compile file @var{source} and run it with the command line arguments
@var{args}. In order to be able to give more than one argument to a
script, several PCC options can be given @emph{after} the
@option{-run} option, separated by spaces:
@example
pcc "-run -L/usr/X11R6/lib -lX11" ex4.c
@end example
In a script, it gives the following header:
@example
#!/usr/local/bin/pcc -run -L/usr/X11R6/lib -lX11
@end example

@item -rstdin file
With @option{-run}: reopen standard input from @file{file} before executing
the program. This is mainly useful when the C source itself was read from
standard input (see @option{-} above).

@item -v
Display PCC version.

@item -vv
Show included files.  As sole argument, print search dirs.  -vvv shows tries too.

@item -bench
Display compilation statistics.

@item -dumpmachine
Print target machine architecture triplet.

@item -dumpversion
Print PCC compiler version.

@end table

@subheading Preprocessor Options

@table @option
@item -Idir
Specify an additional include path. Include paths are searched in the
order they are specified.

System include paths are always searched after. The defaults are
@file{@var{tccdir}/include} and then @file{/usr/include}, where
@var{tccdir} is the pcc private directory (@file{PREFIX/lib/pcc}, see
@option{-B}; @file{PREFIX} is usually @file{/usr} or @file{/usr/local}).
On Windows the defaults are @file{@var{tccdir}/include} and
@file{@var{tccdir}/include/winapi}. Use @option{-vv} as sole argument to
print the paths of the current build.

@item -isystem dir
Specify a system include path to be added to the defaults.

@item -nostdinc
Do not search the default system include paths; only search include paths
provided on the command line.

@item -include file
Include @option{file} above each input file.

@item -Dsym[=val]
Define preprocessor symbol @samp{sym} to
val. If val is not present, its value is @samp{1}. Function-like macros can
also be defined: @option{-DF(a)=a+1}

@item -Usym
Undefine preprocessor symbol @samp{sym}.

@item -E
Preprocess only, to stdout or file (with -o).

@item -P
Do not output @code{#line} directives.

@item -P1
Output alternative @code{#line} directives.

@item -dD, -dM
Output @code{#define} directives.

@item -Wp,-opt
Same as @option{-opt}.

@end table

@subheading Compilation Flags

Note: each of the following options has a negative form beginning with
@option{-fno-}.

@table @option
@item -funsigned-char
Let the @code{char} type be unsigned.

@item -fsigned-char
Let the @code{char} type be signed.

@item -fcommon
Generate common symbols for uninitialized data. The default is
@option{-fno-common}, which puts tentative definitions directly in the
bss section.

@item -fleading-underscore
Add a leading underscore at the beginning of each C symbol.

@item -fms-extensions
Allow a MS C compiler extensions to the language. Currently this
assumes a nested named structure declaration without an identifier
behaves like an unnamed one.

@item -fdollars-in-identifiers
Allow dollar signs in identifiers

@item -freverse-funcargs
Evaluate function arguments right to left.

@item -fgnu89-inline
@code{extern inline} is like @code{static inline}.

@item -fasynchronous-unwind-tables
Create eh_frame section [on]

@item -ftest-coverage
Create code coverage code. After running the resulting code an executable.tcov
or sofile.tcov file is generated with code coverage.

@end table

@subheading Warning Options

@table @option
@item -w
Disable all warnings.

@end table

Note: each of the following warning options has a negative form beginning with
@option{-Wno-}.

@table @option
@item -Wimplicit-function-declaration
Warn about implicit function declaration (missing prototype).

@item -Wdiscarded-qualifiers
Warn when const is dropped.

@item -Wunsupported
Warn about unsupported GCC features that are ignored by PCC.

@item -Wwrite-strings
Make string constants be of type @code{const char *} instead of @code{char
*}.

@item -Werror
Abort compilation if a warning is issued. Can be given an option to enable
the specified warning and turn it into an error, for example
@option{-Werror=unsupported}.

@item -Wall
Activate some useful warnings (@option{-Wimplicit-function-declaration},
@option{-Wdiscarded-qualifiers}).

@end table

@subheading Linker Options

@table @option
@item -Ldir
Specify an additional static library path for the @option{-l} option. The
defaults are the pcc private directory (see @option{-B}) and
@file{/usr/lib}, which the build may replace by the library directory of
the host (@file{/usr/lib64} or a multiarch triplet directory). On Windows
the default is @file{@var{tccdir}/lib}. Use @option{-vv} as sole argument
to print the paths of the current build.

@item -lxxx
Link your program with dynamic library libxxx.so or static library
libxxx.a. The library is searched in the paths specified by the
@option{-L} option and @env{LIBRARY_PATH} variable.

@item -Bdir
Set the path where the pcc internal libraries (and include files) can be
found (default is @file{PREFIX/lib/pcc}).

@item -shared
Generate a shared library instead of an executable.

@item -soname name
set name for shared library to be used at runtime

@item -static
Generate a statically linked executable (default is a shared linked
executable).

@item -rdynamic
Export global symbols to the dynamic linker. It is useful when a library
opened with @code{dlopen()} needs to access executable symbols.

@item -pthread
Preprocess with @option{-D_REENTRANT} and link with @option{-lpthread}.

@item -r
Generate an object file combining all input files.

@item -nostdlib
Don't implicitly link with libc, the C runtime files, and libpcc1.

@item -Wl,-nostdlib
Don't search the default library paths (see @option{-L}). Only the paths
specified with @option{-L} and @env{LIBRARY_PATH} are searched.

@item -Wl,-rpath=path
Put custom search path for dynamic libraries into executable.

@item -Wl,-Ipath
@item -Wl,--dynamic-linker=path
Set the ELF interpreter (dynamic linker). This defaults to the value of the
environment variable @env{LD_SO} if set, or a compiled-in default.

@item -Wl,--enable-new-dtags
When putting a custom search path for dynamic libraries into the executable,
create the new ELF dynamic tag DT_RUNPATH instead of the old legacy DT_RPATH.

@item -Wl,--oformat=fmt
Use @var{fmt} as output format. The supported output formats are:
@table @code
@item elf32-i386
ELF output format (default)
@item binary
Binary image (only for executable output)
@item coff
COFF output format (only for executable output for TMS320C67xx target)
@end table

@item -Wl,--export-all-symbols
@item -Wl,--export-dynamic
Export global symbols to the dynamic linker. It is useful when a library
opened with @code{dlopen()} needs to access executable symbols.

@item -Wl,-subsystem=console/gui/wince/...
Set type for PE (Windows) executables.

@item -Wl,-[Ttext=# | section-alignment=# | file-alignment=# | image-base=# | stack=#]
Modify executable layout.

@item -Wl,-(no-|disable-)[dynamicbase | nxcompat | high-entropy-va | tsaware]
Set or clear PE (Windows) executable header hardening flags.  The
@option{-Wl,-high-entropy-va} option is supported on x86-64 and ARM64 PE
targets and implies @option{-Wl,-dynamicbase}.  Clearing dynamicbase also
clears high-entropy-va.  When @option{-Wl,-dynamicbase} is used for an
executable, PCC also enables base relocation emission for Windows ASLR.

@item -Wl,-Bsymbolic
Set DT_SYMBOLIC tag.

@item -Wl,-(no-)whole-archive
Turn on/off linking of all objects in archives.

@end table

@subheading Debugger Options

@table @option
@item -g
Generate run time stab debug information so that you get clear run time
error messages: @code{ test.c:68: in function 'test5()': dereferencing
invalid pointer} instead of the laconic @code{Segmentation
fault}.

@item -gdwarf[-x]
Generate run time dwarf debug information instead of stab debug information.

@item -b
Generate additional support code to check memory allocations and array/pointer
bounds (@pxref{Bounds}). @option{-g} is implied.

@item -bt[N]
Display N callers in stack traces. This is useful with @option{-g} or @option{-b}.
When activated, @code{__TCC_BACKTRACE__} is defined.

With executables, additional support for stack traces is included. A function
    @code{ int pcc_backtrace(const char *fmt, ...); }
is provided to trigger a stack trace with a message on demand.

@end table

@subheading Misc Options

@table @option

@item -std=version
Define @code{__STDC_VERSION__} to @code{201112} if @option{version} is
c11 or gnu11; @code{199901} otherwise.

@item -x[c|a|b|n]
Specify content of next input file: respectively C, assembly, binary, or none.

@item -O[n]
Same as @option{-D__OPTIMIZE__} except for -O0.  @option{-Os} is treated
the same as @option{-O1}.

@item -M
Just output makefile fragment with dependencies

@item -MM
Like -M except mention only user header files, not system header files.

@item -MD
Generate makefile fragment with dependencies.

@item -MMD
Like -MD except mention only user header files, not system header files.

@item -MF depfile
Use @file{depfile} as output for -MD.

@item -MP
Mention all dependencies as targets too.

@item -print-search-dirs
Print the configured installation directory and a list of library
and include directories pcc will search.

@item -dt
With @option{-run}/@option{-E}: auto-define 'test_...' macros

@end table

@subheading Target Specific Options

@table @option
@item -mms-bitfields
Use an algorithm for bitfield alignment consistent with MSVC. Default is
gcc's algorithm.

@item -mfloat-abi (ARM only)
Select the float ABI. Possible values: @code{softfp} and @code{hard}

@item -mno-sse
Do not use sse registers on x86_64

@item -m32, -m64
Pass command line to the i386/x86_64 cross compiler.

@end table

@subheading macOS Specific Options (Mach-O Targets Only)

@table @option
@item -dynamiclib
Generate a dynamic library instead of an executable.

@item -install_name name
Set the install name for a dynamic library.

@item -flat_namespace
Use a flat namespace (ignored, accepted for compatibility).

@item -two_levelnamespace
Use a two-level namespace (default, accepted for compatibility).

@item -undefined @var{treatment}
Specify how undefined symbols are treated (accepted for compatibility).

@item -compatibility_version version
Set the compatibility version for a dynamic library.

@item -current_version version
Set the current version for a dynamic library.

@end table

@subheading Tool Modes

@table @option
@item pcc -ar [crstvx] lib [files]
Create a static library archive.  PCC can function as an @command{ar}
replacement without requiring an external archiver tool.  The
@code{[abdiopN]} keys are not supported.

@item pcc -impdef lib.dll [-v] [-o lib.def] (Windows only)
Create a @file{.def} definition file from a DLL.

@end table

Note: GCC options @option{-fx}, @option{-mx}, @option{-arch}, @option{-C},
@option{--param}, @option{-pedantic}, @option{-pie}, @option{-no-pie},
@option{-pipe}, @option{-s}, and @option{-traditional} are ignored.
@option{-Wunsupported} makes PCC warn about them.
@c man end

@c man begin ENVIRONMENT
Environment variables that affect how pcc operates.

@table @option

@item CPATH
@item C_INCLUDE_PATH
A colon-separated list of directories searched for include files,
directories given with @option{-I} are searched first.

@item LIBRARY_PATH
A colon-separated list of directories searched for libraries for the
@option{-l} option, directories given with @option{-L} are searched first.

@end table

@c man end

@ignore

@setfilename pcc
@settitle Paze C Compiler

@c man begin SEEALSO
cpp(1),
gcc(1)
@c man end

@c man begin AUTHOR
Fabrice Bellard
@c man end

@end ignore

@node Clang
@chapter C language support

@section ANSI C

PCC implements all the ANSI C standard, including structure bit fields
and floating point numbers (@code{long double}, @code{double}, and
@code{float} fully supported).

@section ISOC99 extensions

PCC implements many features of the new C standard: ISO C99. Currently
missing items are: complex and imaginary numbers.

Currently implemented ISOC99 features:

@itemize

@item variable length arrays.

@item 64 bit @code{long long} types are fully supported.

@item The boolean type @code{_Bool} is supported.

@item @code{__func__} is a string variable containing the current
function name.

@item Variadic macros: @code{__VA_ARGS__} can be used for
   function-like macros:
@example
    #define dprintf(level, __VA_ARGS__) printf(__VA_ARGS__)
@end example

@noindent
@code{dprintf} can then be used with a variable number of parameters.

@item Declarations can appear anywhere in a block (as in C++).

@item Array and struct/union elements can be initialized in any order by
  using designators:
@example
    struct @{ int x, y; @} st[10] = @{ [0].x = 1, [0].y = 2 @};

    int tab[10] = @{ 1, 2, [5] = 5, [9] = 9@};
@end example

@item Compound initializers are supported:
@example
    int *p = (int [])@{ 1, 2, 3 @};
@end example
to initialize a pointer pointing to an initialized array. The same
works for structures and strings.

@item Hexadecimal floating point constants are supported:
@example
          double d = 0x1234p10;
@end example

@noindent
is the same as writing
@example
          double d = 4771840.0;
@end example

@item @code{inline} keyword is supported.  Static inline functions are
emitted at the end of the compilation unit only if used.  The behavior
of @code{extern inline} depends on @option{-fgnu89-inline} (@pxref{Invoke}).

@item @code{restrict} keyword is ignored.
@end itemize

@section ISO C11 extensions

PCC implements several features of the ISO C11 standard:

@itemize

@item @code{_Generic} keyword for type-generic expressions:
@example
    #define print_type(x) _Generic((x), \
        int: "int", \
        float: "float", \
        default: "other")
@end example

@item @code{_Static_assert(expr, msg)} for compile-time assertions.
The single-argument form @code{_Static_assert(expr)} (C23) is also
supported.

@item @code{_Atomic} type qualifier for atomic types.

@item @code{_Thread_local} storage-class specifier for thread-local
storage.  The GCC extension @code{__thread} is also supported.

@item @code{_Noreturn} function specifier (equivalent to the
@code{__attribute__((noreturn))} GNU extension).

@item @code{_Alignas(n)} specifier to request alignment of a variable
or structure field.

@item @code{_Alignof} operator to query the alignment of a type or
expression.

@end itemize

@section GNU C extensions

PCC implements some GNU C extensions:

@itemize

@item array designators can be used without '=':
@example
    int a[10] = @{ [0] 1, [5] 2, 3, 4 @};
@end example

@item Structure field designators can be a label:
@example
    struct @{ int x, y; @} st = @{ x: 1, y: 1@};
@end example
instead of
@example
    struct @{ int x, y; @} st = @{ .x = 1, .y = 1@};
@end example

@item @code{\e} is ASCII character 27.

@item case ranges : ranges can be used in @code{case}s:
@example
    switch(a) @{
    case 1 @dots{} 9:
          printf("range 1 to 9\n");
          break;
    default:
          printf("unexpected\n");
          break;
    @}
@end example

@cindex aligned attribute
@cindex packed attribute
@cindex section attribute
@cindex unused attribute
@cindex cleanup attribute
@cindex constructor attribute
@cindex destructor attribute
@cindex always_inline attribute
@cindex alias attribute
@cindex visibility attribute
@cindex weak attribute
@cindex noreturn attribute
@cindex cdecl attribute
@cindex stdcall attribute
@cindex fastcall attribute
@cindex thiscall attribute
@cindex regparm attribute
@cindex mode attribute
@cindex dllexport attribute
@cindex dllimport attribute
@cindex nodecorate attribute
@cindex format attribute
@cindex used attribute
@cindex nodebug attribute

@item The keyword @code{__attribute__} is handled to specify variable or
function attributes. The following attributes are supported:
  @itemize

  @item @code{aligned(n)}: align a variable or a structure field to n bytes
(must be a power of two).

  @item @code{packed}: force alignment of a variable or a structure field to
  1.

  @item @code{section(name)}: generate function or data in assembly section
name (name is a string containing the section name) instead of the default
section.

  @item @code{unused}: specify that the variable or the function is unused.
Accepted but currently ignored.

  @item @code{cleanup(func)}: specify a function to be called automatically
when the variable goes out of scope.  The cleanup function must take one
parameter, a pointer to a type compatible with the variable, and return
@code{void}.

  @item @code{constructor}: specify that the function is to be executed automatically before @code{main()} is called.

  @item @code{destructor}: specify that the function is to be executed automatically after @code{main()} returns or @code{exit()} is called.

  @item @code{always_inline}: force inline expansion of the function.

  @item @code{alias("target")}: declare the function or variable as an alias for symbol @code{target} (a string containing the target symbol name).

  @item @code{visibility("mode")}: set symbol visibility. Supported modes are @code{"default"}, @code{"hidden"}, @code{"internal"}, and @code{"protected"}.

  @item @code{weak}: declare the symbol as a weak symbol.

  @item @code{noreturn}: specify that the function does not return to its caller.

  @item @code{cdecl}: use standard C calling convention (default).

  @item @code{stdcall}: use Pascal-like calling convention.

  @item @code{fastcall}: use fastcall calling convention (i386 only).

  @item @code{thiscall}: use C++-style thiscall calling convention (i386 only).

  @item @code{regparm(n)}: use fast i386 calling convention. @var{n} must be
between 1 and 3. The first @var{n} function parameters are respectively put in
registers @code{%eax}, @code{%edx} and @code{%ecx}.

  @item @code{__mode__(mode)}: specify the data type size mode for integer
variables or typedefs (supported modes: @code{__QI__}, @code{__HI__},
@code{__SI__}, @code{__DI__}, @code{__word__}). The attribute is only
taken into account when it is given with the declaration specifiers, that
is before the declarator:
@example
    typedef __attribute__ ((__mode__(__QI__))) int int8_type;
@end example

  @item @code{dllexport}: export function from dll/executable (win32 only).

  @item @code{dllimport}: import function from dll/executable (win32 only).

  @item @code{nodecorate}: do not apply any decorations that would otherwise be applied when exporting function from dll/executable (win32 only).

  @item @code{nodebug}: suppress debug information for the symbol.

  @item @code{format(archetype, fmt_idx, chk_idx)}: marks functions that
take @code{printf}- or @code{scanf}-style format strings.  Accepted but
currently ignored (not checked by PCC).

  @item @code{used}: indicates that the symbol must be retained even if it
appears unreferenced.  Accepted but currently ignored.

  @end itemize

Here are some examples:
@example
    int a __attribute__ ((aligned(8), section(".mysection")));
@end example

@noindent
align variable @code{a} to 8 bytes and put it in section @code{.mysection}.

@example
    int my_add(int a, int b) __attribute__ ((section(".mycodesection")))
    @{
        return a + b;
    @}
@end example

@noindent
generate function @code{my_add} in section @code{.mycodesection}.

@example
    void free_ptr(void *p) @{ free(*(void **)p); @}
    void foo(void) @{
        char *buf __attribute__((cleanup(free_ptr))) = malloc(1024);
        /* buf is freed automatically when leaving foo */
    @}
@end example

@noindent
automatically call @code{free_ptr(&buf)} when variable @code{buf} goes out of scope.

@item GNU style variadic macros:
@example
    #define dprintf(fmt, args@dots{}) printf(fmt, ## args)

    dprintf("no arg\n");
    dprintf("one arg %d\n", 1);
@end example

@item @code{__FUNCTION__} is interpreted as C99 @code{__func__}
(so it has not exactly the same semantics as string literal GNUC
where it is a string literal).

@item The @code{__alignof__} keyword can be used as @code{sizeof}
to get the alignment of a type or an expression.

@item The @code{typeof(x)} returns the type of @code{x}.
@code{x} is an expression or a type.

@item Computed gotos: @code{&&label} returns a pointer of type
@code{void *} on the goto label @code{label}. @code{goto *expr} can be
used to jump on the pointer resulting from @code{expr}.

@item Statement expressions: a compound statement enclosed in
parentheses may appear as an expression.  The value of the expression is
the value of the last statement.  This is commonly used in macros:
@example
    #define max(a,b) (@{ int _a = (a), _b = (b); _a > _b ? _a : _b; @})
@end example

@item The @code{__label__} keyword declares local labels within a block:
@example
    @{ __label__ done; /* ... */ done: ; @}
@end example

@item The @code{__extension__} keyword suppresses warnings for GNU
extensions in strict compilation modes.

@item Inline assembly with asm instruction:
@cindex inline assembly
@cindex assembly, inline
@cindex __asm__
@example
static inline void * my_memcpy(void * to, const void * from, size_t n)
@{
int d0, d1, d2;
__asm__ __volatile__(
        "rep ; movsl\n\t"
        "testb $2,%b4\n\t"
        "je 1f\n\t"
        "movsw\n"
        "1:\ttestb $1,%b4\n\t"
        "je 2f\n\t"
        "movsb\n"
        "2:"
        : "=&c" (d0), "=&D" (d1), "=&S" (d2)
        :"0" (n/4), "q" (n),"1" ((long) to),"2" ((long) from)
        : "memory");
return (to);
@}
@end example

@noindent
@cindex gas
PCC includes its own x86 inline assembler with a @code{gas}-like (GNU
assembler) syntax. No intermediate files are generated. GCC 3.x named
operands are supported.  @code{asm goto} is also supported for
control-flow transfer to C labels:
@example
  asm goto ("jmp %l[label]" : : : : label);
@end example

@item The following GCC built-in functions are supported:
  @itemize
  @item @code{__builtin_types_compatible_p(type1, type2)}: returns 1 if @code{type1} is compatible with @code{type2}, 0 otherwise.
  @item @code{__builtin_constant_p(exp)}: returns 1 if @code{exp} evaluates to a compile-time constant, 0 otherwise.
  @item @code{__builtin_choose_expr(const_expr, exp1, exp2)}: evaluates to @code{exp1} if @code{const_expr} is non-zero, otherwise @code{exp2}.
  @item @code{__builtin_expect(expr, expected)}: branch prediction hint (evaluates to @code{expr}).
  @item @code{__builtin_unreachable()}: informs the compiler that the code location cannot be reached.
  @item @code{__builtin_frame_address(level)}: returns the frame pointer address for stack level @code{level}.
  @item @code{__builtin_return_address(level)}: returns the return address for stack level @code{level}.
  @item @code{__builtin_offsetof(type, field)}: returns the byte offset of @code{field} within structure @code{type}.
  @item Bit manipulation built-ins: @code{__builtin_clz},
@code{__builtin_ctz}, @code{__builtin_ffs}, @code{__builtin_popcount},
@code{__builtin_parity}, @code{__builtin_clrsb} (and their @code{l} /
@code{ll} type variants).
  @item Variadic argument handling built-ins: @code{__builtin_va_start}, @code{__builtin_va_arg}, @code{__builtin_va_copy}, @code{__builtin_va_end}.
  @item Memory and string built-ins (mapped to libc):
@code{__builtin_memcpy}, @code{__builtin_memmove},
@code{__builtin_memset}, @code{__builtin_memcmp},
@code{__builtin_strlen}, @code{__builtin_strcpy},
@code{__builtin_strncpy}, @code{__builtin_strcmp},
@code{__builtin_strncmp}, @code{__builtin_strcat},
@code{__builtin_strncat}, @code{__builtin_strchr},
@code{__builtin_strrchr}, @code{__builtin_strdup},
@code{__builtin_malloc}, @code{__builtin_realloc},
@code{__builtin_calloc}, @code{__builtin_memalign},
@code{__builtin_free}, @code{__builtin_alloca},
@code{__builtin_abort}.
  @item Atomic built-ins: @code{__atomic_store},
@code{__atomic_load}, @code{__atomic_exchange},
@code{__atomic_compare_exchange}, @code{__atomic_fetch_add},
@code{__atomic_fetch_sub}, @code{__atomic_fetch_or},
@code{__atomic_fetch_xor}, @code{__atomic_fetch_and},
@code{__atomic_fetch_nand}, and their reverse
@code{__atomic_add_fetch} variants.
  @end itemize

@item Pragmas supported by PazeCC:
  @itemize
  @item @code{#pragma pack(n)}, @code{#pragma pack()}, @code{#pragma pack(push)}, @code{#pragma pack(push, n)}, @code{#pragma pack(pop)}: set structure alignment or manipulate the packing alignment stack.
  @item @code{#pragma push_macro("MACRO")} and @code{#pragma pop_macro("MACRO")}: push and pop macro definitions to/from a macro stack.
  @item @code{#pragma once}: ensure the header file is included only once during compilation.
  @item @code{#pragma comment(lib, "libname")}: specify a library to link automatically.
  @item @code{#pragma comment(option, "flags")}: pass command line options directly within source code.
  @end itemize

@item Preprocessor features:
  @itemize
  @item @code{#include_next <file>}: include the next file of the same name in the header search path order.
  @item @code{#warning "message"}: emit a preprocessor warning message.
  @item GNU empty variadic macro argument pasting: @code{, ##__VA_ARGS__} strips the leading comma when @code{__VA_ARGS__} is empty.
  @item @code{__has_include(<file>)} and @code{__has_include("file")}: test whether a header file exists.  Can be used in @code{#if} / @code{#elif} preprocessor conditions.
  @item @code{__has_include_next(<file>)}: like @code{__has_include} but searches in the next include path, for use in wrapper headers.
  @end itemize

@end itemize

@section PazeCC extensions

@itemize

@item @code{__TINYC__} is a predefined macro to indicate that you use PCC.

@item @code{#!} at the start of a line is ignored to allow scripting.

@item Binary digits can be entered (@code{0b101} instead of
@code{5}).

@end itemize

@node asm
@chapter PazeCC Assembler

Since version 0.9.16, PazeCC integrates its own assembler. PazeCC
assembler supports a gas-like syntax (GNU assembler). You can
deactivate assembler support if you want a smaller PazeCC executable
(the C compiler does not rely on the assembler).

PazeCC Assembler is used to handle files with @file{.S} (C
preprocessed assembler) and @file{.s} extensions. It is also used to
handle the GNU inline assembler with the @code{asm} keyword.

@section Syntax

PazeCC Assembler supports most of the gas syntax. The tokens are the
same as C.

@itemize

@item C and C++ comments are supported.

@item Identifiers are mostly the same as C.  The dot (@code{.}) is
accepted as an identifier character in assembler code.  The dollar
sign (@code{$}) is accepted with @option{-fdollars-in-identifiers},
except in inline assembly on targets other than x86-64 and RISC-V 64.

@item 64-bit integer numbers are supported.

@end itemize

@section Expressions

@itemize

@item Integers in decimal, octal and hexa are supported.

@item Unary operators: +, -, ~.

@item Binary operators in decreasing priority order:

@enumerate
@item *, /, %, <<, >>
@item &, |, ^
@item +, -
@end enumerate

@item A value is either an absolute number or a label plus an offset.
Only the @code{+} and @code{-} operators can be used with labels (to
add or subtract an offset).  All other operators require absolute
values.  @code{-} supports two labels only if they are the same or if
they are both defined and in the same section.

@end itemize

@section Labels

@itemize

@item All labels are considered as local, except undefined ones.

@item Numeric labels can be used as local @code{gas}-like labels.
They can be defined several times in the same source. Use 'b'
(backward) or 'f' (forward) as suffix to reference them:

@example
 1:
      jmp 1b /* jump to '1' label before */
      jmp 1f /* jump to '1' label after */
 1:
@end example

@end itemize

@section Directives
@cindex assembler directives
@cindex directives, assembler
@cindex align directive
@cindex skip directive
@cindex space directive
@cindex byte directive
@cindex word directive
@cindex short directive
@cindex int directive
@cindex long directive
@cindex quad directive
@cindex globl directive
@cindex global directive
@cindex section directive
@cindex text directive
@cindex data directive
@cindex bss directive
@cindex fill directive
@cindex org directive
@cindex previous directive
@cindex string directive
@cindex asciz directive
@cindex ascii directive
@cindex p2align directive
@cindex balign directive
@cindex pushsection directive
@cindex popsection directive
@cindex set directive
@cindex weak directive
@cindex hidden directive
@cindex type directive

All directives are preceded by a '.'. The following directives are
supported:

@itemize
@item .align n[,value]
@item .skip n[,value]
@item .space n[,value]
@item .byte value1[,...]
@item .word value1[,...]
@item .short value1[,...]
@item .int value1[,...]
@item .long value1[,...]
@item .quad immediate_value1[,...]
@item .globl symbol
@item .global symbol
@item .section section
@item .text
@item .data
@item .bss
@item .fill repeat[,size[,value]]
@item .org n
@item .previous
@item .string string[,...]
@item .asciz string[,...]
@item .ascii string[,...]
@item .p2align n
@item .balign n[,value]
@item .pushsection section
@item .popsection
@item .set symbol, expr
@item .weak symbol
@item .hidden symbol
@item .type symbol, type
@end itemize

@section X86 Assembler
@cindex assembler

All i386 opcodes are supported.  Most common x86_64 opcodes are
supported, including MMX and a subset of the SSE instruction set; SSE2
support is limited to a couple of instructions.  Only AT&T syntax is
supported (source then destination operand order).  If no size suffix is
given, PazeCC tries to guess it from the operand sizes.  Note that more
recent extensions such as AVX, AVX2, and AVX-512 are @emph{not}
supported.

x86_64 adds 64-bit operand support (the @code{q} suffix), additional
registers (@code{%r8}--@code{%r15}), and instructions such as
@code{cqto}, @code{pushfq}, @code{popfq}, @code{bswapq},
@code{cmpxchg16b}, @code{movnti}, @code{prefetch*}, @code{lfence},
@code{mfence}, @code{sfence}, and @code{endbr64} (Control-Flow
Enforcement).

@section ARM, ARM64 and RISC-V Assemblers
@cindex assembler

PCC also includes assemblers for the ARM (32-bit), ARM64 (AArch64) and
RISC-V 64 targets.  These handle the same @file{.S} and @file{.s} file
extensions and support the same inline assembly mechanism as the x86
assemblers.

@node linker
@chapter PazeCC Linker
@cindex linker

PCC includes its own linker and can directly output executables,
shared libraries, and object files without relying on an external
linker.

@section ELF file generation
@cindex ELF

PCC can directly output relocatable ELF files (object files),
executable ELF files and dynamic ELF libraries without relying on an
external linker.

Dynamic ELF libraries can be output but the C compiler does not generate
position independent code (PIC). It means that the dynamic library
code generated by PCC cannot be factorized among processes yet.

PCC linker eliminates unreferenced object code in libraries. A single pass is
done on the object and library list, so the order in which object files and
libraries are specified is important (same constraint as GNU ld). No grouping
options (@option{--start-group} and @option{--end-group}) are supported.

PCC generates standard ELF sections including @code{.text}, @code{.data},
@code{.data.ro} (read only data), @code{.bss}, @code{.plt}, @code{.got},
and @code{.eh_frame}.
Shared libraries additionally get @code{.gnu.hash}, @code{.gnu.version},
@code{.dynamic}, and relocation sections (@code{.rela.got},
@code{.rela.plt}).

@cindex section boundary symbols
@cindex __start_SEC
@cindex __stop_SEC
PCC automatically defines @emph{section boundary symbols}
@code{__start_SEC} and @code{__stop_SEC} for every allocated section
whose name, without the leading dot, is a valid C identifier.  These are
useful for iterating over section contents from C code:

@example
extern char __start_text[], __stop_text[];
size_t text_size = __stop_text - __start_text;
@end example

@section ELF file loader
@cindex ELF loader

PCC can load ELF object files, archives (@code{.a} files) and dynamic
libraries (@code{.so}).

@section PE file generation
@cindex PE
@cindex PE-i386
@cindex PE32+

PCC for Windows supports the native Win32 executable file format
(PE-i386).  It generates PE32 (i386) and PE32+ (x86-64) executables,
as well as DLL files.  PCC also supports PE with ARM64 machine type,
including @code{.pdata} unwind information for structured exception
handling.

PE executables can target several subsystems (see @ref{Invoke,, the
@option{-Wl,-subsystem} option}); which ones are available depends on
the target.  Standard PE sections are emitted (@code{.text},
@code{.data}, @code{.rdata}, @code{.bss}, @code{.rsrc}, @code{.pdata},
@code{.idata}, @code{.reloc}); the export directory goes to
@code{.rdata}.

@section Mach-O file generation
@cindex Mach-O
@cindex macOS

PCC on macOS produces Mach-O executables and dynamic libraries
(@code{.dylib}) for x86-64 and ARM64 (AArch64).  The files PCC writes
always contain a single architecture.

PCC can load Mach-O dynamic libraries and text-based stub files
(@code{.tbd}) for linking against system frameworks.  Fat (universal)
libraries are accepted, the slice matching the target being selected.

Mach-O output uses dyld chained fixups on both x86-64 and ARM64.

@section Binary and COFF output
@cindex binary output
@cindex COFF

In addition to the platform-native container formats, PCC supports:

@table @code
@item Binary image
Raw binary output with no container headers.  Selected via
@code{-Wl,--oformat=binary} (@pxref{Invoke}).  Useful for bootloaders,
kernels, and embedded targets.  Only valid for executable output.

@item COFF
The TMS320C67xx (C67) target outputs COFF executables. Object files are
always written as ELF.
@end table

@section GNU Linker Scripts
@cindex scripts, linker
@cindex linker scripts
@cindex GROUP, linker command
@cindex FILE, linker command
@cindex OUTPUT_FORMAT, linker command
@cindex TARGET, linker command

Because on many Linux systems some dynamic libraries (such as
@file{/usr/lib/libc.so}) are in fact GNU ld link scripts (horrible!),
the PCC linker also supports a subset of GNU ld scripts.

The @code{GROUP} and @code{FILE} commands are supported. @code{OUTPUT_FORMAT}
and @code{TARGET} are ignored.

Example from @file{/usr/lib/libc.so}:
@example
/* GNU ld script
   Use the shared library, but some functions are only in
   the static library, so try that secondarily.  */
GROUP ( /lib/libc.so.6 /usr/lib/libc_nonshared.a )
@end example

@node Bounds
@chapter PazeCC Memory and Bound checks
@cindex bound checks
@cindex memory checks

This feature is activated with the @option{-b} option (@pxref{Invoke}).
Here are some examples of caught errors:

@table @asis

@item Invalid range with standard string function:
@example
@{
    char tab[10];
    memset(tab, 0, 11);
@}
@end example

@item Out of bounds-error in global or local arrays:
@example
@{
    int tab[10];
    for(i=0;i<11;i++) @{
        sum += tab[i];
    @}
@}
@end example

@item Out of bounds-error in malloc'ed data:
@example
@{
    int *tab;
    tab = malloc(20 * sizeof(int));
    for(i=0;i<21;i++) @{
        sum += tab[i];
    @}
    free(tab);
@}
@end example

@item Out of bounds-error in alloca'ed data:
@example
@{
    char *p = alloca(10);
    memset(p, 'a', 11);
@}
@end example

@item Out of bounds-error in variable length arrays:
@example
@{
    int tab[n];
    for(i=0;i<n+1;i++) @{
        sum += tab[i];
    @}
@}
@end example

@item Access of freed memory:
@example
@{
    int *tab;
    tab = malloc(20 * sizeof(int));
    free(tab);
    for(i=0;i<20;i++) @{
        sum += tab[i];
    @}
@}
@end example

@item Double free:
@example
@{
    int *tab;
    tab = malloc(20 * sizeof(int));
    free(tab);
    free(tab);
@}
@end example

@item Overlapping regions in memory and string functions:
@example
@{
    char tab[10];
    memcpy(tab, tab + 1, 5);
@}
@end example
@end table

The checks are not limited to the code generated by PCC: the bound
checking runtime also replaces the memory allocation functions
(@code{malloc}, @code{calloc}, @code{realloc}, @code{memalign},
@code{free}) and the most common memory and string functions
(@code{memcpy}, @code{memmove}, @code{memset}, @code{memcmp},
@code{strlen}, @code{strcpy}, @code{strncpy}, @code{strcmp},
@code{strncmp}, @code{strcat}, @code{strncat}, @code{strchr},
@code{strrchr}, @code{strdup}) by versions checking their arguments.
Regions created by @code{alloca}, by variable length arrays and (except
on Windows) by @code{mmap} are tracked as well.  @code{setjmp} and
@code{longjmp} (and @code{sigsetjmp} and @code{siglongjmp}) are handled
so that the regions of the abandoned stack frames are released.

PCC defines @code{__TCC_BCHECK__} if activated.

There are five environment variables that can be used to control the behavior:
@itemize
@item PCC_BOUNDS_WARN_POINTER_ADD
- Print warning when pointer add creates an illegal pointer.
@item PCC_BOUNDS_PRINT_CALLS
- Print bound checking calls. Can be used for debugging.
@item PCC_BOUNDS_PRINT_HEAP
- Print heap objects that are not freed at exit of program.
@item PCC_BOUNDS_PRINT_STATISTIC
- Print statistic information at exit of program.
@item PCC_BOUNDS_NEVER_FATAL
- Try to continue in case of a bound checking error.
@end itemize

Also, a function @code{__bounds_checking(x)} can be used to turn off/on bounds
checking from usercode (see below).

Notes:
@itemize
@item Available for all targets and platforms supported by PCC, except
the TMS320C67xx target. The checking code is only built for the native
compiler, and can be left out with @option{--config-bcheck=no} at
configure time.
@item The generated code is slower and bigger.
@item The bound checking code is not included in shared libraries. The main
executable should always be compiled with the @option{-b}.
@item Pointer size is @emph{unchanged} and code generated with bound checks is
@emph{fully compatible} with unchecked code. When a pointer comes from
unchecked code, it is assumed to be valid. Even very obscure C code with
casts should work correctly.
@item Signal handlers are not compatible with bounds checking. The
bounds checking code disables checking in signal/sigaction handlers.
The fork() function call in a multi threaded application is also a problem.
The bound checking code fixes this for the child process.
@item The reason that signals and fork have problems is that we use locking
inside the bounds checking code.
Inside a signal handler we can not use locks. Also in a multi threaded
application after a fork the child process can have the lock set
by another thread.
@item The BOUNDS_CHECKING_OFF and BOUNDS_CHECKING_ON can also be used to
disable bounds checking for some code.
@item The __bounds_checking call adds a value to a thread local value.
The value starts at 0. If the value is not 0 the code is not checked
for bounds checking errors.
@end itemize

@example
#ifdef __TCC_BCHECK__
extern void __bounds_checking (int x);
# define BOUNDS_CHECKING_OFF __bounds_checking(1)
# define BOUNDS_CHECKING_ON  __bounds_checking(-1)
#else
# define BOUNDS_CHECKING_OFF
# define BOUNDS_CHECKING_ON
#endif
@end example

For more information about the ideas behind this method, see
@url{http://www.doc.ic.ac.uk/~phjk/BoundsChecking.html}.

@node Libtcc
@chapter The @code{libpcc} library

The @code{libpcc} library enables you to use PCC as a backend for
dynamic code generation.

Read the @file{libpcc.h} to have an overview of the API. Read
@file{tests/libpcc_test.c} to have a very simple example.

The idea consists in giving a C string containing the program you want
to compile directly to @code{libpcc}. Then you can access to any global
symbol (function or variable) defined.

A typical session creates a compilation state with @code{pcc_new()},
selects the output type with @code{pcc_set_output_type()} (this must be
done before any compilation), passes command line options with
@code{pcc_set_options()} and adds input with @code{pcc_compile_string()}
or @code{pcc_add_file()}.  The result is then either written to a file
with @code{pcc_output_file()}, run with @code{pcc_run()}, or relocated in
memory with @code{pcc_relocate()} so that the compiled functions and
variables can be retrieved with @code{pcc_get_symbol()}.
@code{pcc_output_file()} and @code{pcc_run()} do the relocation
themselves, so @code{pcc_relocate()} must not be called before them.
Symbols of the host program are made visible to the compiled code with
@code{pcc_add_symbol()}.  The state and everything it allocated is
released by @code{pcc_delete()}.

Errors and warnings go to @code{stderr} unless a callback is installed
with @code{pcc_set_error_func()}. Several states may be used at the same
time from different threads; see @file{tests/libpcc_test_mt.c}, which
also shows how @code{pcc_setjmp()} catches the runtime exceptions of code
compiled with @option{-b} or @option{-bt}.

@node devel
@chapter Developer's guide

This chapter gives some hints to understand how PCC works. You can skip
it if you do not intend to modify the PCC code.

@section File reading

The @code{BufferedFile} structure contains the context needed to read a
file, including the current line number. @code{pcc_open()} opens a new
file and @code{pcc_close()} closes it. @code{pcc_open_bf()} builds the
same context for a memory buffer instead of a file; it is used for the
@option{-D} options of the command line and for token pasting.

The contexts are stacked through the @code{BufferedFile.prev} field, so
the current file is always the innermost @code{#include}.
@code{next_c()} returns the next character and calls @code{handle_eob()}
to refill the buffer when it is exhausted.

@section Lexer

@code{next()} reads the next token in the current
file. @code{next_nomacro()} reads the next token without macro
expansion.

@code{tok} contains the current token (see @code{TOK_xxx}
constants). Identifiers and keywords are also tokens: they are entered
in the @code{table_ident} array of @code{TokenSym} structures, so a
string is never needed to designate them. @code{tokc} contains
additional infos about the token (for example a constant value if number
or string token).

@section Parser

The parser is hardcoded (yacc is not necessary). It does only one pass,
except:

@itemize

@item For initialized arrays with unknown size, a first pass
is done to count the number of elements.

@item For architectures where arguments are evaluated in
reverse order, a first pass is done to reverse the argument order.

@item For inline functions, the body is recorded as a token string and
parsed again where the function is really needed.

@end itemize

@section Types

The types are stored in a @code{CType} structure. @code{CType.t} is a
single 'int' holding the basic type, the type modifiers and, during
parsing, the storage class. @code{CType.ref} points to the @code{Sym}
describing the referenced type: the pointed type for pointers, the
element type for arrays, the return type and the parameters for
functions, and the fields for structures and unions.

@example
#define VT_BTYPE       0x000f  /* mask for basic type */
#define VT_VOID             0  /* void type */
#define VT_BYTE             1  /* signed byte type */
#define VT_SHORT            2  /* short type */
#define VT_INT              3  /* integer type */
#define VT_LLONG            4  /* 64 bit integer */
#define VT_PTR              5  /* pointer */
#define VT_FUNC             6  /* function type */
#define VT_STRUCT           7  /* struct/union definition */
#define VT_FLOAT            8  /* IEEE float */
#define VT_DOUBLE           9  /* IEEE double */
#define VT_LDOUBLE         10  /* IEEE long double */
#define VT_BOOL            11  /* ISOC99 boolean type */
#define VT_QLONG           13  /* 128-bit integer, only for the x86-64 ABI */
#define VT_QFLOAT          14  /* 128-bit float, only for the x86-64 ABI */

#define VT_UNSIGNED    0x0010  /* unsigned type */
#define VT_DEFSIGN     0x0020  /* explicitly signed or unsigned */
#define VT_ARRAY       0x0040  /* array type (also has VT_PTR) */
#define VT_BITFIELD    0x0080  /* bitfield modifier */
#define VT_CONSTANT    0x0100  /* const modifier */
#define VT_VOLATILE    0x0200  /* volatile modifier */
#define VT_VLA         0x0400  /* VLA type (also has VT_PTR and VT_ARRAY) */
#define VT_LONG        0x0800  /* long type (also has VT_INT rsp. VT_LLONG) */

#define VT_STRUCT_SHIFT 20     /* shift for bitfield shift values */
@end example

The @code{VT_UNSIGNED} flag can be set for chars, shorts, ints and long
longs. @code{VT_DEFSIGN} tells whether the signedness was written
explicitly; @code{compare_types()} only looks at it for @code{char}, the
one type whose default signedness differs from an explicit one.

Arrays are considered as pointers @code{VT_PTR} with the flag
@code{VT_ARRAY} set. Variable length arrays have @code{VT_VLA} set in
addition to @code{VT_PTR} and @code{VT_ARRAY}.

@code{VT_LONG} is a modifier, not a basic type: @code{long} is
@code{VT_LONG | VT_INT} or @code{VT_LONG | VT_LLONG} depending on the
target. It does not change the generated code, but it is part of the type
so that @code{long} stays distinct from @code{int} rsp. @code{long long}
for type compatibility, for @code{_Generic} and in diagnostics.

If @code{VT_BITFIELD} is set, then the bitfield position is stored in the
6 bits starting at @code{VT_STRUCT_SHIFT} and the bitfield size in the 6
bits above it (macros @code{BIT_POS()} and @code{BIT_SIZE()}).

The same high order bits also distinguish types which share a basic
type: @code{VT_UNION} is a @code{VT_STRUCT} with 1 there, @code{VT_ENUM}
and @code{VT_ENUM_VAL} mark an integral type which is really an enum or
an enum constant, and @code{VT_ASM} marks a symbol which was created by
the assembler.

During parsing, the storage of an object is also stored in the type
integer:

@example
#define VT_EXTERN  0x00001000  /* extern definition */
#define VT_STATIC  0x00002000  /* static variable */
#define VT_TYPEDEF 0x00004000  /* typedef definition */
#define VT_INLINE  0x00008000  /* inline definition */
#define VT_TLS     0x00010000  /* thread-local storage */
@end example

The declaration properties which do not fit in the type integer
(alignment, packing, visibility, weak, dllimport/dllexport, calling
convention, @dots{}) are kept in the @code{SymAttr} and @code{FuncAttr}
bitfields of the @code{Sym} structure.

@section Symbols

All symbols are stored in symbol stacks, chained through the
@code{Sym.prev} field. Each symbol stack contains @code{Sym} structures.

@code{Sym.v} contains the symbol name (remember
an identifier is also a token, so a string is never necessary to store
it). @code{Sym.type} gives the type of the symbol. @code{Sym.r} is usually
the register in which the corresponding variable is stored. @code{Sym.c} is
usually a constant associated to the symbol like its address for normal
symbols, and the number of entries for symbols representing arrays.
Variable length array types use @code{Sym.c} as a location on the stack
which holds the runtime sizeof for the type.

Five main symbol stacks are defined:

@table @code

@item define_stack
for the macros (@code{#define}s).

@item global_stack
for the global variables, functions and types.

@item local_stack
for the local variables, functions and types.

@item global_label_stack
for the function local labels (for @code{goto}).

@item local_label_stack
for GCC block local labels (see the @code{__label__} keyword).

@end table

@code{sym_push()} is used to add a new symbol in the local symbol
stack. If no local symbol stack is active, it is added in the global
symbol stack.

@code{sym_pop(st,b,keep)} pops symbols from the symbol stack @var{st}
until the symbol @var{b} is on the top of stack. If @var{b} is NULL, the
stack is emptied. The symbols are always made invisible to the parser,
but if @var{keep} is non zero they are not freed, which is needed as long
as the recorded body of an inline function may still refer to them.

@code{sym_find(v)} returns the symbol associated to the identifier
@var{v}. No stack is searched: each @code{TokenSym} of
@code{table_ident} directly holds the innermost visible @code{Sym} for
its identifier, and @code{sym_pop()} restores the previous one. Tags of
structures, unions and enums live in their own namespace and are looked
up with @code{struct_find()}, labels with @code{label_find()}.

@section Sections

The generated code and data are written in sections. The structure
@code{Section} contains all the necessary information for a given
section. @code{new_section()} creates a new section. ELF file semantics
is assumed for each section.

The following sections are predefined:

@table @code

@item text_section
is the section containing the generated code. @var{ind} contains the
current position in the code section. @var{cur_text_section} is the
section the code of the current function goes to, which is another
section when the function has a @code{section} attribute.

@item data_section
contains initialized data

@item rodata_section
contains read only data: the literal strings and the objects declared
@code{const}

@item bss_section
contains uninitialized data

@item common_section
receives the tentative definitions when @option{-fcommon} is given. Its
symbols are @code{SHN_COMMON}, with the alignment in the symbol value,
and @code{resolve_common_syms()} allocates them in @code{bss_section} at
link time. By default PCC does not use common symbols and puts tentative
definitions directly in @code{bss_section}

@item bounds_section
@itemx lbounds_section
are used when bound checking is activated

@item stab_section
@itemx stabstr_section
are used when debugging is active and stabs debug information is selected

@item dwarf_info_section
@itemx dwarf_abbrev_section
@itemx dwarf_line_section
@itemx dwarf_aranges_section
@itemx dwarf_str_section
@itemx dwarf_line_str_section
are used when debugging is active and dwarf debug information is
selected, which is the default on macOS and Android and can be enabled
elsewhere with the @option{--config-dwarf} option of @command{configure}

@item tcov_section
contains the counters used by @option{-ftest-coverage}

@item symtab_section
contains the symbols and their names (in @code{symtab_section->link}).
It is needed for linking, not only for debugging.

@end table

@section Code generation
@cindex code generation

@subsection Introduction

The PCC code generator directly generates linked binary code in one
pass. It is rather unusual these days (see gcc for example which
generates text assembly), but it can be very fast and surprisingly
little complicated.

The PCC code generator is register based. Optimization is only done at
the expression level. No intermediate representation of expression is
kept except the current values stored in the @emph{value stack}.

Each target declares the registers the code generator may use in
@code{reg_classes[]}, and @code{NB_REGS} gives their number: five on
i386 (@code{eax}, @code{ecx}, @code{edx}, @code{ebx}, which is only
usable when @code{USE_EBX} is set, and the top of the x87 stack), 25 on
x86-64, 28 on arm64. When more registers are needed, one register is
spilled into a new temporary variable on the stack.

@subsection The value stack
@cindex value stack, introduction

When an expression is parsed, its value is pushed on the value stack
(@var{vstack}). The top of the value stack is @var{vtop}. Each value
stack entry is the structure @code{SValue}.

@code{SValue.type} is the type. @code{SValue.r} indicates how the value is
currently stored in the generated code. It is usually a CPU register
index (@code{REG_xxx} constants), but additional values and flags are
defined:

@example
#define VT_VALMASK   0x003f  /* mask for value location, register or: */
#define VT_CONST     0x0030  /* constant in vc */
#define VT_LLOCAL    0x0031  /* lvalue, offset on stack */
#define VT_LOCAL     0x0032  /* offset on stack */
#define VT_CMP       0x0033  /* the value is stored in processor flags */
#define VT_JMP       0x0034  /* value is the consequence of jmp true (even) */
#define VT_JMPI      0x0035  /* value is the consequence of jmp false (odd) */
#define VT_LVAL      0x0100  /* var is an lvalue */
#define VT_SYM       0x0200  /* a symbol value is added */
#define VT_MUSTCAST  0x0C00  /* value must be casted to be correct */
#define VT_NONCONST  0x1000  /* VT_CONST, but not an (C standard) integer
                                constant expression */
#define VT_MUSTBOUND 0x4000  /* bound checking must be done before
                                dereferencing value */
#define VT_BOUNDED   0x8000  /* value is bounded */
@end example

@code{SValue.r2} holds the second register when a value needs two of
them, as for a @code{long long} on a 32 bit target. It is set to
@code{VT_CONST} when it is not used.

@table @code

@item VT_CONST
indicates that the value is a constant. It is stored in the union
@code{SValue.c}, depending on its type.

@item VT_LOCAL
indicates a local variable pointer at offset @code{SValue.c.i} in the
stack.

@item VT_CMP
indicates that the value is actually stored in the CPU flags (i.e. the
value is the consequence of a test). The value is either 0 or 1. The
comparison which set the flags is kept in @code{SValue.cmp_op};
@code{SValue.cmp_r} is left to the code generator, which uses it to
remember where the compared values are on the targets that have no flags
register.

If any code is generated which destroys the CPU flags, this value MUST be
put in a normal register. @code{vcheck_cmp()} does it, and the functions
which manipulate the value stack call it before generating code.

@item VT_JMP
@itemx VT_JMPI
indicates that the value is the consequence of a conditional jump. For VT_JMP,
it is 1 if the jump is taken, 0 otherwise. For VT_JMPI it is inverted.

These values are used to compile the @code{||} and @code{&&} logical
operators. The jumps which still have to be resolved are chained in
@code{SValue.jtrue} and @code{SValue.jfalse}; @code{gvtst()} adds to
these lists and generates the final test.

If any code is generated, this value MUST be put in a normal
register. Otherwise, the generated code won't be executed if the jump is
taken.

@item VT_LVAL
is a flag indicating that the value is actually an lvalue (left value of
an assignment). It means that the value stored is actually a pointer to
the wanted value.

Understanding the use @code{VT_LVAL} is very important if you want to
understand how PCC works.

@item VT_LLOCAL
is a saved lvalue on the stack. @code{VT_LVAL} must also be set with
@code{VT_LLOCAL}. @code{VT_LLOCAL} can arise when a @code{VT_LVAL} in
a register has to be saved to the stack, or it can come from an
architecture-specific calling convention.

@item VT_MUSTCAST
indicates that a cast to the value type must be performed if the value
is used (lazy casting), as for a @code{char} or a @code{short} kept in an
integer register. It is a two bit field, so that the width the value has
to be casted from (@code{int} or @code{long long}) is remembered too.

@item VT_SYM
indicates that the symbol @code{SValue.sym} must be added to the constant.

@item VT_NONCONST
marks a @code{VT_CONST} value which only became constant through
optimization. Such a value is folded like any other constant, but it is
not a constant expression in the sense of the C standard: it is rejected
where the language requires one (@code{expr_const64()}) and it is not a
null pointer constant.

@item VT_MUSTBOUND
@itemx VT_BOUNDED
are only used for optional bound checking.

@end table

@subsection Manipulating the value stack
@cindex value stack

@code{vsetc()} and @code{vset()} pushes a new value on the value
stack. If the previous @var{vtop} was stored in a very unsafe place(for
example in the CPU flags), then some code is generated to put the
previous @var{vtop} in a safe storage.

@code{vpop()} pops @var{vtop}. In some cases, it also generates cleanup
code (for example if stacked floating point registers are used as on
x86).

@code{vdup()}, @code{vswap()}, @code{vrotb()} and @code{vrott()}
duplicate, exchange and rotate the topmost entries. They all go through
@code{vcheck_cmp()} first, because a @code{VT_CMP} value may only stay on
top of the stack.

The @code{gv(rc)} function generates code to evaluate @var{vtop} (the
top value of the stack) into registers. @var{rc} selects in which
register class the value should be put. @code{gv()} is the @emph{most
important function} of the code generator.

@code{gv2()} is the same as @code{gv()} but for the top two stack
entries.

@code{get_reg(rc)} returns a free register of the class @var{rc}. When
they are all busy it spills the one held by the oldest value stack entry,
starting from the bottom of the stack so that the registers of the
operation being generated are never taken away. The spill itself is done
by @code{save_reg()}, which copies to the stack every value living in a
given register.

@subsection CPU dependent code generation
@cindex CPU dependent
See the @file{i386-gen.c} file to have an example.

@table @code

@item load()
must generate the code needed to load a stack value into a register.

@item store()
must generate the code needed to store a register into a stack value
lvalue.

@item gfunc_call(nb_args)
should generate a function call, the arguments and the function address
being the @code{nb_args + 1} topmost entries of the value stack.
@code{gfunc_sret()} tells the parser how a structure return value is
passed for the target ABI.

@item gfunc_prolog()
@itemx gfunc_epilog()
should generate a function prolog/epilog.

@item gen_opi(op)
must generate the binary integer operation @var{op} on the two top
entries of the stack which are guaranteed to contain integer types.

The result value should be put on the stack.

@item gen_opf(op)
same as @code{gen_opi()} for floating point operations. The two top
entries of the stack are guaranteed to contain floating point values of
same types.

@item gen_cvt_itof()
integer to floating point conversion.

@item gen_cvt_ftoi()
floating point to integer conversion.

@item gen_cvt_ftof()
floating point to floating point of different size conversion.

@item gen_cvt_csti()
performs the delayed cast of a @code{char} or a @code{short} held in an
integer register (see @code{VT_MUSTCAST}).

@item gjmp()
@itemx gjmp_addr()
@itemx gjmp_cond()
@itemx gsym_addr()
generate unconditional and conditional jumps, and patch the jumps whose
target was not known yet.

@item gen_vla_alloc()
@itemx gen_vla_sp_save()
@itemx gen_vla_sp_restore()
allocate a variable length array on the stack, and save and restore the
stack pointer around it.

@end table

@section Optimizations done
@cindex optimizations
@cindex constant propagation
@cindex strength reduction
@cindex comparison operators
@cindex caching processor flags
@cindex flags, caching
@cindex jump optimization
Constant propagation is done for all operations. Multiplications and
divisions by a power of two are optimized to shifts, and operations
which are a no-operation for the given constant (@code{x*1},
@code{x-0}, @code{x&-1}, @dots{}) are dropped. Constants added to a
symbol or to a local variable address are folded into the address.
Comparison operators are optimized by maintaining a special cache for
the processor flags. &&, || and ! are optimized by maintaining a special
'jump target' value. No other jump optimization is currently performed
because it would require to store the code in a more abstract fashion.

@unnumbered Concept Index
@printindex cp

@bye

@c Local variables:
@c fill-column: 78
@c texinfo-column-for-description: 32
@c End: