.\" This code is derived from software copyrighted by the Free Software
.\" Modified 1991 by Donn Seeley at UUNET Technologies, Inc.
.\" @(#)cc.1 6.5 (Berkeley) 5/9/91
.TH CC 1 "May 9, 1991" "Version 1.36"
cc \- GNU project C Compiler
is a version of the GNU C compiler.
It accepts a dialect of ANSI C with extensions;
this dialect is different from the dialect used
in 4.3 BSD and earlier distributions.
flag causes the compiler to accept
a dialect of extended Classic C,
much like the C of these earlier distributions.
If you are not already familiar with
ANSI C and its new features,
you will want to build your software with
Most older C compiler flags are supported by
to generate symbol tables for the unsupported
for single precision floating point
which is now the default; and
for alternate compiler passes.
The differences between ANSI C and Classic C dialects
are too numerous to describe here in detail.
The following quick summary is meant to make
users aware of potential subtle problems when
converting Classic C code to ANSI C.
The most obvious change is the pervasive use of
.IR "function prototypes" .
Under the ANSI C dialect, the number and type of arguments
to C library functions are checked by the compiler
when standard header files are included;
calls that fail to match will yield errors.
A subtle consequence of adding prototype declarations is that
user code that inadvertently redefines a C library function
may break; for example it is no longer possible to write an
function that takes different parameters or returns a different value
when including standard header files.
Another issue with prototypes is that functions
that take different parameter types no longer have the same type;
function pointers now differ by parameter types as well as return types.
Variable argument lists are handled differently;
which unfortunately is not completely compatible.
A subtle change in type promotion can be confusing: small
unsigned types are now widened into signed types
rather than unsigned types.
A similar problem can occur with the
operator, which now yields an unsigned type rather than a signed type.
One common problem is due to a change in scoping: external declarations
are now scoped to the block they occur in,
so a declaration for (say)
inside one block will no longer
declare it in all subsequent blocks.
The syntax for braces in structure initializations is now a bit stricter,
and it is sometimes necessary to add braces to please the compiler.
Two very subtle and sometimes very annoying features
apply to constant strings and to the
Constant strings in the ANSI dialect are read-only;
attempts to alter them cause protection violations.
This ANSI feature permits the compiler to coalesce
identical strings in the same source file,
and saves space when multiple copies of a binary are running
at the same time, since the read-only part of a binary
The most common difficulty with read-only strings
function, which in the past often altered a constant string argument.
It is now necessary to copy a constant string
before it may be altered.
function may now destroy any register
stack variable in the function that made
the corresponding call to the
to protect a local variable, the new ANSI
This often leads to confusing situations
The compiler has extended warning flags for
dealing with read-only strings and
but these are not very effective.
If your code has problems with any of these ANSI features,
you will probably want to use
there are some differences between
this dialect of Classic C and the dialect supported on older distributions.
There are at least two differences that are a consequence
uses an ANSI C style grammar for both
traditional and ANSI modes.
The old C dialect permitted a typedef
to replace a simple type in the idiom ``unsigned
treats such forms as syntax errors.
The old C dialect also permitted formal parameters
to have the same names as typedef types;
the current dialect does not.
Some questionable or illegal practices
that were supported in the old C dialect are
non-comment text at the end of a ``#include''
preprocessor control line is an error, not ignored;
compound assignment operators must not contain white space,
``*\0='' is not the same as ``*='';
the last member declaration in a structure or union
must be terminated by a semicolon;
it is not possible to ``switch'' on function pointers;
more than one occurrence of ``#else'' at the same level
in a preprocessor ``#if'' clause is an error, not ignored.
Some truly ancient C practices are no longer supported.
The idiom of declaring an anonymous structure and
using its members to extract fields from other structures
or even non-structures is illegal.
Integers are not automatically converted to pointers
when they are dereferenced.
dialect does not retain the so-called ``old-fashioned'' assignment operators
(with the ``='' preceding rather than following the operator)
or initializations (with no ``='' between initializer and initializee).
This rest of man page is an extract of the documentation of the
and is limited to the meaning of the options.
.BR "It is not kept up to date" .
If you want to be certain of the information
below, check it in the manual "Using and Porting GCC". Refer to the Info file
which are made from the Texinfo source file
uses a command syntax much like the Unix C compiler.
program accepts options and file names as operands.
Multiple single-letter options may
When you invoke GNU CC, it normally does preprocessing, compilation,
are taken as C source to be preprocessed and compiled;
are taken as preprocessor output to be compiled;
compiler output files plus any input files with names ending in
then the resulting object files, plus any other input files,
are linked together to produce an executable.
Command options allow you to stop this process at an intermediate stage.
option says not to run the linker.
Then the output consists of object files output by the assembler.
Other command options are passed on to one stage of processing.
Some options control the preprocessor and others the compiler itself.
Yet other options control the assembler and linker;
these are not documented here, but you rarely need to use any of them.
Here are the options to control the overall compilation process,
including those that say whether to link, whether to assemble, and so on.
This applies regardless to whatever sort of output is being produced,
whether it be an executable file, an object file,
an assembler file or preprocessed C code.
is not specified, the default is to put an executable file in
and preprocessed C on standard output.
Compile or assemble the source files, but do not link.
Produce object files with names made by replacing
at the end of the input file names.
Do nothing at all for object files specified as input.
Compile into assembler code but do not assemble.
The assembler output file name is made by replacing
at the end of the input file name.
Do nothing at all for assembler source files or
object files specified as input.
Run only the C preprocessor.
Preprocess all the C source files specified and output
the results to standard output.
Compiler driver program prints the commands it executes as it runs
the preprocessor, compiler proper, assembler and linker.
Some of these are directed to print their own version numbers.
Use pipes rather than temporary files for communication between the
various stages of compilation.
This fails to work on some systems where the assembler is unable
to read from a pipe; but the GNU assembler has no trouble.
Compiler driver program tries
as a prefix for each program it tries to run.
For each subprogram to be run, the compiler driver first tries the
If that name is not found, or if
was not specified, the driver tries a standard prefix, which currently is
If this does not result in a file name that is found, the
unmodified program name is searched for using the directories
You can get a similar result from the environment variable
if it is defined, its value is used as a prefix in the same way.
variable are present, the
option is used first and the environment variable value second.
is used as a second prefix for the compiler executables and libraries.
This prefix is optional: the compiler tries each file first with it,
This prefix follows the prefix specified with
in the presence of environment variable
causes GNU CC to try the following file names for the preprocessor executable:
These options control the details of C compilation itself.
Support all ANSI standard C programs.
This turns off certain features of GNU C that are incompatible with
keywords, and predefined macros such as
that identify the type of system you are using.
It also enables the undesirable and rarely used ANSI trigraph feature.
You would not want to use them in an ANSI C program, of course,
but it useful to put them in header files that might be included
in compilations done with
Alternate predefined macros such as
are also available, with or without
option does not cause non-ANSI programs to be rejected gratuitously.
is required in addition to
Some header files may notice this macro and refrain from declaring
certain functions or defining certain macros that the ANSI standard
doesn't call for; this is to avoid interfering with any programs
that might use these names for other things.
Attempt to support some aspects of traditional C compilers.
declarations take effect globally even if they are
written inside of a function definition.
This includes implicit declarations of functions.
Comparisons between pointers and integers are always allowed.
Out-of-range floating point literals are not an error.
All automatic variables not declared
Ordinarily, GNU C follows ANSI C: automatic variables not declared
In the preprocessor, comments convert to nothing at all,
This allows traditional token concatenation.
In the preprocessor, macro arguments are recognized within string
constants in a macro definition (and their values are stringified, though
without additional quote marks, when they appear in such a context).
The preprocessor always considers a string constant to end at a newline.
is not defined when you use
is (since the GNU extensions which
indicates are not affected by
If you need to write header files that work differently depending on whether
is in use, by testing both of these predefined macros you can distinguish
four situations: GNU C, traditional GNU C, other ANSI C compilers, and
Optimizing compilation takes somewhat more time,
and a lot more memory for a large function.
the compiler's goal is to reduce the cost of compilation and
to make debugging produce the expected results.
Statements are independent: if you stop the program with a breakpoint
between statements, you can then assign a new value to any variable or
change the program counter to any other statement in the function and
get exactly the results you would expect from the source code.
are allocated in registers.
The resulting compiled code is a little worse than produced by PCC without
the compiler tries to reduce code size and execution time.
options described below turn specific kinds of optimization on or off.
Produce debugging information in the operating system's
native format (for DBX or SDB).
GDB also can work with this debugging information.
Unlike most other C compilers, GNU CC allows you to use
The shortcuts taken by optimized code may occasionally
produce surprising results: some variables you declared may not exist
at all; flow of control may briefly move where you did not expect it;
some statements may not be executed because they compute constant
results or their values were already at hand; some statements may
execute in different places because they were moved out of loops.
Nevertheless it proves possible to debug optimized output.
This makes it reasonable to use the optimizer for programs
Inhibit all warning messages.
Print extra warning messages for these events:
An automatic variable is used without first being initialized.
These warnings are possible only in optimizing compilation,
because they require data flow information that is computed only
you simply won't get these warnings.
These warnings occur only for variables that are candidates for
Therefore, they do not occur for a variable that is declared
or whose address is taken, or whose size is other than 1, 2, 4 or 8 bytes.
Also, they do not occur for structures, unions or arrays, even when
Note that there may be no warning about a variable that is used only
to compute a value that itself is never used, because such
computations may be deleted by data flow analysis before the warnings
These warnings are made optional because GNU CC is not smart
enough to see all the reasons why the code might be correct
despite appearing to have an error.
Here is one example of how this can happen:
is always 1, 2 or 3, then
is always initialized, but GNU CC doesn't know this.
Here is another common case:
\ \ if (change_y) save_y = y, y = new_y;
\ \ if (change_y) y = save_y;
is used only if it is set.
Some spurious warnings can be avoided if you declare as
all the functions you use that never return.
A nonvolatile automatic variable might be changed by a call to
These warnings as well are possible only in optimizing compilation.
The compiler sees only the calls to
will be called; in fact, a signal handler could
call it at any point in the code.
As a result, you may get a warning even when there is
in fact no problem because
cannot in fact be called at the place which would cause a problem.
A function can return either with or without a value.
(Falling off the end of the function body is considered returning without
For example, this function would evoke such a warning:
Spurious warnings can occur because GNU CC does not realize that
certain functions (including
An expression-statement contains no side effects.
In the future, other useful warnings may also be enabled by this option.
Warn whenever a function is implicitly declared.
Warn whenever a function is defined with a return-type that defaults to
statement with no return-value in a function whose return-type is not
Warn whenever a local variable is unused aside from its declaration,
and whenever a function is declared static but never defined.
statement has an index of enumeral type and lacks a
for one or more of the named codes of that enumeration.
label prevents this warning.)
labels outside the enumeration range also provoke
warnings when this option is used.
Warn whenever a comment-start sequence
Warn if any trigraphs are encountered (assuming they are enabled).
These are all the options which pertain to usage that we do not recommend and
that we believe is always easy to avoid, even in conjunction with macros.
options below are not implied by
because certain kinds of useful macros are almost impossible to write
without causing those warnings.
Warn whenever a local variable shadows another local variable.
Warn whenever two distinct identifiers match in the first
This may help you prepare a program that will compile with certain obsolete,
Warn about anything that depends on the size of a function type or of
GNU C assigns these types a size of 1, for convenience in calculations with
pointers and pointers to functions.
Warn whenever a pointer is cast so as to remove a type qualifier from
Give string constants the type
.B const char[\fIlength\fB]
so that copying the address of one into a
.RB non- "const char \(**"
pointer will get a warning.
These warnings will help you find at compile time
code that can try to write into a string constant,
but only if you have been very careful about using
in declarations and prototypes.
Otherwise, it will just be a nuisance; this is why we did not make
Generate extra code to write profile information suitable
Generate extra code to write profile information suitable for the
Generate extra code to write profile information for basic blocks,
suitable for the analysis program
should be extended to process this data.
Search a standard list of directories for a library named
which is actually a file named
The linker uses this file as if it had been specified precisely by name.
The directories searched include several standard system directories
plus any that you specify with
Normally the files found this way are library files--archive files
whose members are object files.
The linker handles an archive file by scanning through it for members
which define symbols that have so far been referenced but not defined.
But if the file that is found is an ordinary object file, it is linked
The only difference between using an
option and specifying a file name is that
searches several directories.
to the list of directories to be searched for
Don't use the standard system libraries and startup files when linking.
Only the files you specify (plus
will be passed to the linker.
Machine-dependent option specifying something about the type of target machine.
These options are defined by the macro
in the machine description.
The default for the options is also defined by that macro,
which enables you to change the defaults.
options defined in the 68000 machine description:
Generate output for a 68020 (rather than a 68000).
This is the default if you use the unmodified sources.
Generate output for a 68000 (rather than a 68020).
Generate output containing 68881 instructions for floating point.
This is the default if you use the unmodified sources.
Generate output containing Sun FPA instructions for floating point.
Generate output containing library calls for floating point.
Do not use the bit-field instructions.
Do use the bit-field instructions.
This is the default if you use the unmodified sources.
Use a different function-calling convention, in which functions
that take a fixed number of arguments return with the
instruction, which pops their arguments while returning.
This saves one instruction in the caller since there is no need to pop
This calling convention is incompatible with the one normally
used on Unix, so you cannot use it if you need to call libraries
compiled with the Unix compiler.
Also, you must provide function prototypes for all functions that
take variable numbers of arguments (including
otherwise incorrect code will be generated for calls to those functions.
In addition, seriously incorrect code will result if you call a
function with too many arguments.
(Normally, extra arguments are harmlessly ignored.)
instruction is supported by the 68010 and 68020 processors,
options are defined in the Vax machine description:
Do not output certain jump instructions
and so on) that the Unix assembler for the Vax
cannot handle across long ranges.
Do output those jump instructions, on the assumption that you
will assemble with the GNU assembler.
Output code for g-format floating point numbers instead of d-format.
switches are supported on the Sparc:
Generate output containing floating point instructions.
This is the default if you use the unmodified sources.
Generate output containing library calls for floating point.
Generate separate return instructions for
This has both advantages and disadvantages; I don't recall what they are.
options are defined in the Convex machine description:
Generate output for a C1.
This is the default when the compiler is configured for a C1.
Generate output for a C2.
This is the default when the compiler is configured for a C2.
Generate code which puts an argument count in the word preceding each
Some nonportable Convex and Vax programs need this word.
(Debuggers don't; this info is in the symbol table.)
Omit the argument count word.
This is the default if you use the unmodified sources.
Specify machine-independent flags.
Most flags have both positive and negative forms; the negative form of
In the table below, only one of the forms is listed--the one which
You can figure out the other form by either removing
Use the same convention for returning
values that is used by the usual C compiler on your system.
This convention is less efficient for small structures, and on many
machines it fails to be reentrant; but it has the advantage of allowing
intercallability between GCC-compiled code and PCC-compiled code.
Do not store floating-point variables in registers.
This prevents undesirable excess precision on machines such as the
68000 where the floating registers (of the 68881) keep more
For most programs, the excess precision does only good, but a few
programs rely on the precise definition of IEEE floating point.
These words may then be used as identifiers.
Always pop the arguments to each function call as soon as that
Normally the compiler (when optimizing) lets arguments accumulate
on the stack for several function calls and pops them all at once.
Perform the optimizations of loop strength reduction and
elimination of iteration variables.
Allow the combine pass to combine an instruction that copies one
This might or might not produce better code when used in addition to
I am interested in hearing about the difference this makes.
Force memory operands to be copied into registers before doing
This may produce better code by making all memory references
potential common subexpressions.
When they are not common subexpressions, instruction combination should
eliminate the separate register-load.
I am interested in hearing about the difference this makes.
Force memory address constants to be copied into registers before
doing arithmetic on them.
This may produce better code just as
I am interested in hearing about the difference this makes.
Don't keep the frame pointer in a register for functions that
This avoids the instructions to save, set up and restore frame pointers;
it also makes an extra register available in many functions.
.B "It also makes debugging impossible."
On some machines, such as the Vax, this flag has no effect, because
the standard calling sequence automatically handles the frame pointer
and nothing is saved by pretending it doesn't exist.
The machine-description macro
.B FRAME_POINTER_REQUIRED
controls whether a target machine supports this flag.
Integrate all simple functions into their callers.
The compiler heuristically decides which functions are simple
enough to be worth integrating in this way.
If all calls to a given function are integrated, and the function is declared
then the function is normally not output as assembler code in its own right.
Enable values to be allocated in registers that will be clobbered by
function calls, by emitting extra instructions to save and restore the
registers around such calls.
Such allocation is done only when it seems to result in better code than
would otherwise be produced.
This option is enabled by default on certain machines, usually those
which have no call-preserved registers to use instead.
.B \-fkeep-inline-functions
Even if all calls to a given function are integrated, and the function is
nevertheless output a separate run-time callable version of the function.
Store string constants in the writable data segment and don't uniquize them.
This is for compatibility with old programs which assume they can write
Writing into string constants is a very bad idea;
constants should be constant.
Allow conditional expressions with mismatched types in the second and
The value of such an expression is void.
Do not put function addresses in registers; make each instruction that
calls a constant function contain the function's address explicitly.
This option results in less efficient code, but some strange hacks that
alter the assembler output may be confused by the optimizations performed
when this option is not used.
Consider all memory references through pointers to be volatile.
Requests that the data and
variables of this compilation be shared data rather than private data.
The distinction makes sense only on certain operating systems, where
shared data is shared between processes running the same program, while
private data exists in one copy per process.
Each kind of machine has a default for what
(Actually, at present, the default is always signed.)
is always a distinct type from either
even though its behavior is always just like one of those two.
Note that this is equivalent to
which is the negative form of
Note that this is equivalent to
.BR \-fno-unsigned-char ,
which is the negative form of
If supported for the target machine, attempt to reorder instructions to
exploit instruction slots available after delayed branch instructions.
as a fixed register; generated code should never refer to it
(except perhaps as a stack pointer, frame pointer or in some other fixed role).
must be the name of a register.
The register names accepted are machine-specific and are defined in the
macro in the machine description macro file.
This flag does not have a negative form, because it specifies a
as an allocatable register that is clobbered by function calls.
It may be allocated for temporaries or variables that do not live
Functions compiled this way will not save and restore the register REG.
Use of this flag for a register that has a fixed pervasive role
in the machine's execution model, such as the stack pointer or
frame pointer, will produce disastrous results.
This flag does not have a negative form, because it specifies a
as an allocatable register saved by functions.
It may be allocated even for temporaries or variables that live across a call.
Functions compiled this way will save and restore the register
Use of this flag for a register that has a fixed pervasive role
in the machine's execution model, such as the stack pointer or
frame pointer, will produce disastrous results.
A different sort of disaster will result from the use of this
flag for a register in which function values may be returned.
This flag does not have a negative form, because it specifies a
Says to make debugging dumps at times specified by
Here are the possible letters:
Dump after RTL generation.
Dump after first jump optimization.
Dump after last jump optimization.
Dump after CSE (including the jump optimization that sometimes follows CSE).
Dump after loop optimization.
Dump after flow analysis.
Dump after instruction combination.
Dump after local register allocation.
Dump after global register allocation.
Dump after delayed branch scheduling.
Print statistics on memory usage, at the end of the run.
Issue all the warnings demanded by strict ANSI standard C; reject
all programs that use forbidden extensions.
Valid ANSI standard C programs should compile properly with or without
this option (though a rare few will require
However, without this option, certain GNU extensions and traditional C
features are supported as well.
With this option, they are rejected.
There is no reason to use this option; it exists only to satisfy pedants.
does not cause warning messages for use of the alternate keywords whose
On Suns running version 4, this prevents linking with the shared
These options control the C preprocessor, which is run on each C source
file before actual compilation. If you use the `-E' option, nothing
is done except C preprocessing. Some of these options make sense only
together with `-E' because they request preprocessor output that is
not suitable for actual compilation.
Tell the preprocessor not to discard comments.
Any directories specified with
option are searched only for the case of
\fB"\fIfile\fB"\fR; they are not searched for
.BR "#include <\fIfile\fB>" .
If additional directories are specified with
these directories are searched for all
directories are used this way.)
option inhibits the use of the current directory as the first
Therefore, the current directory is searched only if it is requested
allows you to control precisely which directories are searched before
the current one and which are searched after.
Do not search the standard system directories for header files.
Only the directories you have specified with
options (and the current directory, if appropriate) are searched.
you can eliminate all directories from the search path
except those you specify.
Tell the preprocessor to output a rule suitable for
describing the dependencies of each source file.
For each source file, the preprocessor outputs one
whose target is the object file name for that source file and whose
dependencies are all the files
This rule may be a single line or may be continued with
but the output mentions only the user-header files included with
System header files included with
.B "#include <\fIfile\fB>"
with the empty string as its definition.
Support ANSI C trigraphs.
You don't want to know about this brain-damage.
option also has this effect.
.ta \w'/usr/lib/libgnulib.a 'u
file.s assembly language file
/tmp/cc\(** temporary files
/usr/libexec/cpp preprocessor
/usr/libexec/cc1 compiler
/usr/lib/libgnulib.a library needed by GCC on some machines
/usr/lib/crt0.o start-up routine
/usr/lib/libc.a standard C library, see
/usr/include standard directory for
as(1), ld(1), adb(1), dbx(1), gdb(1).
Bugs should be reported to
.BR bug-gcc@prep.ai.mit.edu .
Bugs tend actually to be fixed if they can be isolated, so it is in your
interest to report them in such a way that they can be easily reproduced.
Copyright (c) 1988 Free Software Foundation, Inc.
Permission is granted to make and distribute verbatim copies of
this manual provided the copyright notice and this permission notice
are preserved on all copies.
Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided that the
entire resulting derived work is distributed under the terms of a
permission notice identical to this one.
Permission is granted to copy and distribute translations of this
manual into another language, under the above conditions for modified
versions, except that this permission notice may be included in
translations approved by the Free Software Foundation instead of in
See the GNU CC Manual for the contributors to GNU CC.