<html lang="en">
<head>
<title>Debugging Options - Using the GNU Compiler Collection (GCC)</title>
<meta http-equiv="Content-Type" content="text/html">
<meta name="description" content="Using the GNU Compiler Collection (GCC)">
<meta name="generator" content="makeinfo 4.13">
<link title="Top" rel="start" href="index.html#Top">
<link rel="up" href="Invoking-GCC.html#Invoking-GCC" title="Invoking GCC">
<link rel="prev" href="Warning-Options.html#Warning-Options" title="Warning Options">
<link rel="next" href="Optimize-Options.html#Optimize-Options" title="Optimize Options">
<link href="http://www.gnu.org/software/texinfo/" rel="generator-home" title="Texinfo Homepage">
<!--
Copyright (C) 1988-2015 Free Software Foundation, Inc.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with the
Invariant Sections being ``Funding Free Software'', the Front-Cover
Texts being (a) (see below), and with the Back-Cover Texts being (b)
(see below). A copy of the license is included in the section entitled
``GNU Free Documentation License''.
(a) The FSF's Front-Cover Text is:
A GNU Manual
(b) The FSF's Back-Cover Text is:
You have freedom to copy and modify this GNU Manual, like GNU
software. Copies published by the Free Software Foundation raise
funds for GNU development.-->
<meta http-equiv="Content-Style-Type" content="text/css">
<style type="text/css"><!--
pre.display { font-family:inherit }
pre.format { font-family:inherit }
pre.smalldisplay { font-family:inherit; font-size:smaller }
pre.smallformat { font-family:inherit; font-size:smaller }
pre.smallexample { font-size:smaller }
pre.smalllisp { font-size:smaller }
span.sc { font-variant:small-caps }
span.roman { font-family:serif; font-weight:normal; }
span.sansserif { font-family:sans-serif; font-weight:normal; }
--></style>
</head>
<body>
<div class="node">
<a name="Debugging-Options"></a>
<p>
Next: <a rel="next" accesskey="n" href="Optimize-Options.html#Optimize-Options">Optimize Options</a>,
Previous: <a rel="previous" accesskey="p" href="Warning-Options.html#Warning-Options">Warning Options</a>,
Up: <a rel="up" accesskey="u" href="Invoking-GCC.html#Invoking-GCC">Invoking GCC</a>
<hr>
</div>
<h3 class="section">3.9 Options for Debugging Your Program or GCC</h3>
<p><a name="index-options_002c-debugging-624"></a><a name="index-debugging-information-options-625"></a>
GCC has various special options that are used for debugging
either your program or GCC:
<dl>
<dt><code>-g</code><dd><a name="index-g-626"></a>Produce debugging information in the operating system's native format
(stabs, COFF, XCOFF, or DWARF 2). GDB can work with this debugging
information.
<p>On most systems that use stabs format, <samp><span class="option">-g</span></samp> enables use of extra
debugging information that only GDB can use; this extra information
makes debugging work better in GDB but probably makes other debuggers
crash or
refuse to read the program. If you want to control for certain whether
to generate the extra information, use <samp><span class="option">-gstabs+</span></samp>, <samp><span class="option">-gstabs</span></samp>,
<samp><span class="option">-gxcoff+</span></samp>, <samp><span class="option">-gxcoff</span></samp>, or <samp><span class="option">-gvms</span></samp> (see below).
<p>GCC allows you to use <samp><span class="option">-g</span></samp> with
<samp><span class="option">-O</span></samp>. 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 are already at hand; some statements may
execute in different places because they have been moved out of loops.
<p>Nevertheless it proves possible to debug optimized output. This makes
it reasonable to use the optimizer for programs that might have bugs.
<p>The following options are useful when GCC is generated with the
capability for more than one debugging format.
<br><dt><code>-gsplit-dwarf</code><dd><a name="index-gsplit_002ddwarf-627"></a>Separate as much dwarf debugging information as possible into a
separate output file with the extension .dwo. This option allows
the build system to avoid linking files with debug information. To
be useful, this option requires a debugger capable of reading .dwo
files.
<br><dt><code>-ggdb</code><dd><a name="index-ggdb-628"></a>Produce debugging information for use by GDB. This means to use the
most expressive format available (DWARF 2, stabs, or the native format
if neither of those are supported), including GDB extensions if at all
possible.
<br><dt><code>-gpubnames</code><dd><a name="index-gpubnames-629"></a>Generate dwarf .debug_pubnames and .debug_pubtypes sections.
<br><dt><code>-ggnu-pubnames</code><dd><a name="index-ggnu_002dpubnames-630"></a>Generate .debug_pubnames and .debug_pubtypes sections in a format
suitable for conversion into a GDB index. This option is only useful
with a linker that can produce GDB index version 7.
<br><dt><code>-gstabs</code><dd><a name="index-gstabs-631"></a>Produce debugging information in stabs format (if that is supported),
without GDB extensions. This is the format used by DBX on most BSD
systems. On MIPS, Alpha and System V Release 4 systems this option
produces stabs debugging output that is not understood by DBX or SDB.
On System V Release 4 systems this option requires the GNU assembler.
<br><dt><code>-feliminate-unused-debug-symbols</code><dd><a name="index-feliminate_002dunused_002ddebug_002dsymbols-632"></a>Produce debugging information in stabs format (if that is supported),
for only symbols that are actually used.
<br><dt><code>-femit-class-debug-always</code><dd><a name="index-femit_002dclass_002ddebug_002dalways-633"></a>Instead of emitting debugging information for a C++ class in only one
object file, emit it in all object files using the class. This option
should be used only with debuggers that are unable to handle the way GCC
normally emits debugging information for classes because using this
option increases the size of debugging information by as much as a
factor of two.
<br><dt><code>-fdebug-types-section</code><dd><a name="index-fdebug_002dtypes_002dsection-634"></a><a name="index-fno_002ddebug_002dtypes_002dsection-635"></a>When using DWARF Version 4 or higher, type DIEs can be put into
their own <code>.debug_types</code> section instead of making them part of the
<code>.debug_info</code> section. It is more efficient to put them in a separate
comdat sections since the linker can then remove duplicates.
But not all DWARF consumers support <code>.debug_types</code> sections yet
and on some objects <code>.debug_types</code> produces larger instead of smaller
debugging information.
<br><dt><code>-gstabs+</code><dd><a name="index-gstabs_002b-636"></a>Produce debugging information in stabs format (if that is supported),
using GNU extensions understood only by the GNU debugger (GDB). The
use of these extensions is likely to make other debuggers crash or
refuse to read the program.
<br><dt><code>-gcoff</code><dd><a name="index-gcoff-637"></a>Produce debugging information in COFF format (if that is supported).
This is the format used by SDB on most System V systems prior to
System V Release 4.
<br><dt><code>-gxcoff</code><dd><a name="index-gxcoff-638"></a>Produce debugging information in XCOFF format (if that is supported).
This is the format used by the DBX debugger on IBM RS/6000 systems.
<br><dt><code>-gxcoff+</code><dd><a name="index-gxcoff_002b-639"></a>Produce debugging information in XCOFF format (if that is supported),
using GNU extensions understood only by the GNU debugger (GDB). The
use of these extensions is likely to make other debuggers crash or
refuse to read the program, and may cause assemblers other than the GNU
assembler (GAS) to fail with an error.
<br><dt><code>-gdwarf-</code><var>version</var><dd><a name="index-gdwarf_002d_0040var_007bversion_007d-640"></a>Produce debugging information in DWARF format (if that is supported).
The value of <var>version</var> may be either 2, 3, 4 or 5; the default version
for most targets is 4. DWARF Version 5 is only experimental.
<p>Note that with DWARF Version 2, some ports require and always
use some non-conflicting DWARF 3 extensions in the unwind tables.
<p>Version 4 may require GDB 7.0 and <samp><span class="option">-fvar-tracking-assignments</span></samp>
for maximum benefit.
<br><dt><code>-grecord-gcc-switches</code><dd><a name="index-grecord_002dgcc_002dswitches-641"></a>This switch causes the command-line options used to invoke the
compiler that may affect code generation to be appended to the
DW_AT_producer attribute in DWARF debugging information. The options
are concatenated with spaces separating them from each other and from
the compiler version. See also <samp><span class="option">-frecord-gcc-switches</span></samp> for another
way of storing compiler options into the object file. This is the default.
<br><dt><code>-gno-record-gcc-switches</code><dd><a name="index-gno_002drecord_002dgcc_002dswitches-642"></a>Disallow appending command-line options to the DW_AT_producer attribute
in DWARF debugging information.
<br><dt><code>-gstrict-dwarf</code><dd><a name="index-gstrict_002ddwarf-643"></a>Disallow using extensions of later DWARF standard version than selected
with <samp><span class="option">-gdwarf-</span><var>version</var></samp>. On most targets using non-conflicting
DWARF extensions from later standard versions is allowed.
<br><dt><code>-gno-strict-dwarf</code><dd><a name="index-gno_002dstrict_002ddwarf-644"></a>Allow using extensions of later DWARF standard version than selected with
<samp><span class="option">-gdwarf-</span><var>version</var></samp>.
<br><dt><code>-gz</code><span class="roman">[</span><code>=</code><var>type</var><span class="roman">]</span><dd><a name="index-gz-645"></a>Produce compressed debug sections in DWARF format, if that is supported.
If <var>type</var> is not given, the default type depends on the capabilities
of the assembler and linker used. <var>type</var> may be one of
‘<samp><span class="samp">none</span></samp>’ (don't compress debug sections), ‘<samp><span class="samp">zlib</span></samp>’ (use zlib
compression in ELF gABI format), or ‘<samp><span class="samp">zlib-gnu</span></samp>’ (use zlib
compression in traditional GNU format). If the linker doesn't support
writing compressed debug sections, the option is rejected. Otherwise,
if the assembler does not support them, <samp><span class="option">-gz</span></samp> is silently ignored
when producing object files.
<br><dt><code>-gvms</code><dd><a name="index-gvms-646"></a>Produce debugging information in Alpha/VMS debug format (if that is
supported). This is the format used by DEBUG on Alpha/VMS systems.
<br><dt><code>-g</code><var>level</var><dt><code>-ggdb</code><var>level</var><dt><code>-gstabs</code><var>level</var><dt><code>-gcoff</code><var>level</var><dt><code>-gxcoff</code><var>level</var><dt><code>-gvms</code><var>level</var><dd>Request debugging information and also use <var>level</var> to specify how
much information. The default level is 2.
<p>Level 0 produces no debug information at all. Thus, <samp><span class="option">-g0</span></samp> negates
<samp><span class="option">-g</span></samp>.
<p>Level 1 produces minimal information, enough for making backtraces in
parts of the program that you don't plan to debug. This includes
descriptions of functions and external variables, and line number
tables, but no information about local variables.
<p>Level 3 includes extra information, such as all the macro definitions
present in the program. Some debuggers support macro expansion when
you use <samp><span class="option">-g3</span></samp>.
<p><samp><span class="option">-gdwarf-2</span></samp> does not accept a concatenated debug level, because
GCC used to support an option <samp><span class="option">-gdwarf</span></samp> that meant to generate
debug information in version 1 of the DWARF format (which is very
different from version 2), and it would have been too confusing. That
debug format is long obsolete, but the option cannot be changed now.
Instead use an additional <samp><span class="option">-g</span><var>level</var></samp> option to change the
debug level for DWARF.
<br><dt><code>-gtoggle</code><dd><a name="index-gtoggle-647"></a>Turn off generation of debug info, if leaving out this option
generates it, or turn it on at level 2 otherwise. The position of this
argument in the command line does not matter; it takes effect after all
other options are processed, and it does so only once, no matter how
many times it is given. This is mainly intended to be used with
<samp><span class="option">-fcompare-debug</span></samp>.
<br><dt><code>-fsanitize=address</code><dd><a name="index-fsanitize_003daddress-648"></a>Enable AddressSanitizer, a fast memory error detector.
Memory access instructions are instrumented to detect
out-of-bounds and use-after-free bugs.
See <a href="https://github.com/google/sanitizers/wiki/AddressSanitizer">https://github.com/google/sanitizers/wiki/AddressSanitizer</a> for
more details. The run-time behavior can be influenced using the
<samp><span class="env">ASAN_OPTIONS</span></samp> environment variable. When set to <code>help=1</code>,
the available options are shown at startup of the instrumended program. See
<a href="https://github.com/google/sanitizers/wiki/AddressSanitizerFlags#run-time-flags">https://github.com/google/sanitizers/wiki/AddressSanitizerFlags#run-time-flags</a>
for a list of supported options.
<br><dt><code>-fsanitize=kernel-address</code><dd><a name="index-fsanitize_003dkernel_002daddress-649"></a>Enable AddressSanitizer for Linux kernel.
See <a href="https://github.com/google/kasan/wiki">https://github.com/google/kasan/wiki</a> for more details.
<br><dt><code>-fsanitize=thread</code><dd><a name="index-fsanitize_003dthread-650"></a>Enable ThreadSanitizer, a fast data race detector.
Memory access instructions are instrumented to detect
data race bugs. See <a href="https://github.com/google/sanitizers/wiki#threadsanitizer">https://github.com/google/sanitizers/wiki#threadsanitizer</a> for more
details. The run-time behavior can be influenced using the <samp><span class="env">TSAN_OPTIONS</span></samp>
environment variable; see
<a href="https://github.com/google/sanitizers/wiki/ThreadSanitizerFlags">https://github.com/google/sanitizers/wiki/ThreadSanitizerFlags</a> for a list of
supported options.
<br><dt><code>-fsanitize=leak</code><dd><a name="index-fsanitize_003dleak-651"></a>Enable LeakSanitizer, a memory leak detector.
This option only matters for linking of executables and if neither
<samp><span class="option">-fsanitize=address</span></samp> nor <samp><span class="option">-fsanitize=thread</span></samp> is used. In that
case the executable is linked against a library that overrides <code>malloc</code>
and other allocator functions. See
<a href="https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer">https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer</a> for more
details. The run-time behavior can be influenced using the
<samp><span class="env">LSAN_OPTIONS</span></samp> environment variable.
<br><dt><code>-fsanitize=undefined</code><dd><a name="index-fsanitize_003dundefined-652"></a>Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector.
Various computations are instrumented to detect undefined behavior
at runtime. Current suboptions are:
<dl>
<dt><code>-fsanitize=shift</code><dd><a name="index-fsanitize_003dshift-653"></a>This option enables checking that the result of a shift operation is
not undefined. Note that what exactly is considered undefined differs
slightly between C and C++, as well as between ISO C90 and C99, etc.
<br><dt><code>-fsanitize=integer-divide-by-zero</code><dd><a name="index-fsanitize_003dinteger_002ddivide_002dby_002dzero-654"></a>Detect integer division by zero as well as <code>INT_MIN / -1</code> division.
<br><dt><code>-fsanitize=unreachable</code><dd><a name="index-fsanitize_003dunreachable-655"></a>With this option, the compiler turns the <code>__builtin_unreachable</code>
call into a diagnostics message call instead. When reaching the
<code>__builtin_unreachable</code> call, the behavior is undefined.
<br><dt><code>-fsanitize=vla-bound</code><dd><a name="index-fsanitize_003dvla_002dbound-656"></a>This option instructs the compiler to check that the size of a variable
length array is positive.
<br><dt><code>-fsanitize=null</code><dd><a name="index-fsanitize_003dnull-657"></a>This option enables pointer checking. Particularly, the application
built with this option turned on will issue an error message when it
tries to dereference a NULL pointer, or if a reference (possibly an
rvalue reference) is bound to a NULL pointer, or if a method is invoked
on an object pointed by a NULL pointer.
<br><dt><code>-fsanitize=return</code><dd><a name="index-fsanitize_003dreturn-658"></a>This option enables return statement checking. Programs
built with this option turned on will issue an error message
when the end of a non-void function is reached without actually
returning a value. This option works in C++ only.
<br><dt><code>-fsanitize=signed-integer-overflow</code><dd><a name="index-fsanitize_003dsigned_002dinteger_002doverflow-659"></a>This option enables signed integer overflow checking. We check that
the result of <code>+</code>, <code>*</code>, and both unary and binary <code>-</code>
does not overflow in the signed arithmetics. Note, integer promotion
rules must be taken into account. That is, the following is not an
overflow:
<pre class="smallexample"> signed char a = SCHAR_MAX;
a++;
</pre>
<br><dt><code>-fsanitize=bounds</code><dd><a name="index-fsanitize_003dbounds-660"></a>This option enables instrumentation of array bounds. Various out of bounds
accesses are detected. Flexible array members, flexible array member-like
arrays, and initializers of variables with static storage are not instrumented.
<br><dt><code>-fsanitize=alignment</code><dd><a name="index-fsanitize_003dalignment-661"></a>
This option enables checking of alignment of pointers when they are
dereferenced, or when a reference is bound to insufficiently aligned target,
or when a method or constructor is invoked on insufficiently aligned object.
<br><dt><code>-fsanitize=object-size</code><dd><a name="index-fsanitize_003dobject_002dsize-662"></a>This option enables instrumentation of memory references using the
<code>__builtin_object_size</code> function. Various out of bounds pointer
accesses are detected.
<br><dt><code>-fsanitize=float-divide-by-zero</code><dd><a name="index-fsanitize_003dfloat_002ddivide_002dby_002dzero-663"></a>Detect floating-point division by zero. Unlike other similar options,
<samp><span class="option">-fsanitize=float-divide-by-zero</span></samp> is not enabled by
<samp><span class="option">-fsanitize=undefined</span></samp>, since floating-point division by zero can
be a legitimate way of obtaining infinities and NaNs.
<br><dt><code>-fsanitize=float-cast-overflow</code><dd><a name="index-fsanitize_003dfloat_002dcast_002doverflow-664"></a>This option enables floating-point type to integer conversion checking.
We check that the result of the conversion does not overflow.
Unlike other similar options, <samp><span class="option">-fsanitize=float-cast-overflow</span></samp> is
not enabled by <samp><span class="option">-fsanitize=undefined</span></samp>.
This option does not work well with <code>FE_INVALID</code> exceptions enabled.
<br><dt><code>-fsanitize=nonnull-attribute</code><dd><a name="index-fsanitize_003dnonnull_002dattribute-665"></a>
This option enables instrumentation of calls, checking whether null values
are not passed to arguments marked as requiring a non-null value by the
<code>nonnull</code> function attribute.
<br><dt><code>-fsanitize=returns-nonnull-attribute</code><dd><a name="index-fsanitize_003dreturns_002dnonnull_002dattribute-666"></a>
This option enables instrumentation of return statements in functions
marked with <code>returns_nonnull</code> function attribute, to detect returning
of null values from such functions.
<br><dt><code>-fsanitize=bool</code><dd><a name="index-fsanitize_003dbool-667"></a>
This option enables instrumentation of loads from bool. If a value other
than 0/1 is loaded, a run-time error is issued.
<br><dt><code>-fsanitize=enum</code><dd><a name="index-fsanitize_003denum-668"></a>
This option enables instrumentation of loads from an enum type. If
a value outside the range of values for the enum type is loaded,
a run-time error is issued.
<br><dt><code>-fsanitize=vptr</code><dd><a name="index-fsanitize_003dvptr-669"></a>
This option enables instrumentation of C++ member function calls, member
accesses and some conversions between pointers to base and derived classes,
to verify the referenced object has the correct dynamic type.
</dl>
<p>While <samp><span class="option">-ftrapv</span></samp> causes traps for signed overflows to be emitted,
<samp><span class="option">-fsanitize=undefined</span></samp> gives a diagnostic message.
This currently works only for the C family of languages.
<br><dt><code>-fno-sanitize=all</code><dd><a name="index-fno_002dsanitize_003dall-670"></a>
This option disables all previously enabled sanitizers.
<samp><span class="option">-fsanitize=all</span></samp> is not allowed, as some sanitizers cannot be used
together.
<br><dt><code>-fasan-shadow-offset=</code><var>number</var><dd><a name="index-fasan_002dshadow_002doffset-671"></a>This option forces GCC to use custom shadow offset in AddressSanitizer checks.
It is useful for experimenting with different shadow memory layouts in
Kernel AddressSanitizer.
<br><dt><code>-fsanitize-recover</code><span class="roman">[</span><code>=</code><var>opts</var><span class="roman">]</span><dd><a name="index-fsanitize_002drecover-672"></a><a name="index-fno_002dsanitize_002drecover-673"></a><samp><span class="option">-fsanitize-recover=</span></samp> controls error recovery mode for sanitizers
mentioned in comma-separated list of <var>opts</var>. Enabling this option
for a sanitizer component causes it to attempt to continue
running the program as if no error happened. This means multiple
runtime errors can be reported in a single program run, and the exit
code of the program may indicate success even when errors
have been reported. The <samp><span class="option">-fno-sanitize-recover=</span></samp> option
can be used to alter
this behavior: only the first detected error is reported
and program then exits with a non-zero exit code.
<p>Currently this feature only works for <samp><span class="option">-fsanitize=undefined</span></samp> (and its suboptions
except for <samp><span class="option">-fsanitize=unreachable</span></samp> and <samp><span class="option">-fsanitize=return</span></samp>),
<samp><span class="option">-fsanitize=float-cast-overflow</span></samp>, <samp><span class="option">-fsanitize=float-divide-by-zero</span></samp> and
<samp><span class="option">-fsanitize=kernel-address</span></samp>. For these sanitizers error recovery is turned on by default.
<samp><span class="option">-fsanitize-recover=all</span></samp> and <samp><span class="option">-fno-sanitize-recover=all</span></samp> is also
accepted, the former enables recovery for all sanitizers that support it,
the latter disables recovery for all sanitizers that support it.
<p>Syntax without explicit <var>opts</var> parameter is deprecated. It is equivalent to
<pre class="smallexample"> -fsanitize-recover=undefined,float-cast-overflow,float-divide-by-zero
</pre>
<p class="noindent">Similarly <samp><span class="option">-fno-sanitize-recover</span></samp> is equivalent to
<pre class="smallexample"> -fno-sanitize-recover=undefined,float-cast-overflow,float-divide-by-zero
</pre>
<br><dt><code>-fsanitize-undefined-trap-on-error</code><dd><a name="index-fsanitize_002dundefined_002dtrap_002don_002derror-674"></a>The <samp><span class="option">-fsanitize-undefined-trap-on-error</span></samp> option instructs the compiler to
report undefined behavior using <code>__builtin_trap</code> rather than
a <code>libubsan</code> library routine. The advantage of this is that the
<code>libubsan</code> library is not needed and is not linked in, so this
is usable even in freestanding environments.
<br><dt><code>-fcheck-pointer-bounds</code><dd><a name="index-fcheck_002dpointer_002dbounds-675"></a><a name="index-fno_002dcheck_002dpointer_002dbounds-676"></a><a name="index-Pointer-Bounds-Checker-options-677"></a>Enable Pointer Bounds Checker instrumentation. Each memory reference
is instrumented with checks of the pointer used for memory access against
bounds associated with that pointer.
<p>Currently there
is only an implementation for Intel MPX available, thus x86 target
and <samp><span class="option">-mmpx</span></samp> are required to enable this feature.
MPX-based instrumentation requires
a runtime library to enable MPX in hardware and handle bounds
violation signals. By default when <samp><span class="option">-fcheck-pointer-bounds</span></samp>
and <samp><span class="option">-mmpx</span></samp> options are used to link a program, the GCC driver
links against the <samp><span class="file">libmpx</span></samp> runtime library and <samp><span class="file">libmpxwrappers</span></samp>
library. It also passes '-z bndplt' to a linker in case it supports this
option (which is checked on libmpx configuration). Note that old versions
of linker may ignore option. Gold linker doesn't support '-z bndplt'
option. With no '-z bndplt' support in linker all calls to dynamic libraries
lose passed bounds reducing overall protection level. It's highly
recommended to use linker with '-z bndplt' support. In case such linker
is not available it is adviced to always use <samp><span class="option">-static-libmpxwrappers</span></samp>
for better protection level or use <samp><span class="option">-static</span></samp> to completely avoid
external calls to dynamic libraries. MPX-based instrumentation
may be used for debugging and also may be included in production code
to increase program security. Depending on usage, you may
have different requirements for the runtime library. The current version
of the MPX runtime library is more oriented for use as a debugging
tool. MPX runtime library usage implies <samp><span class="option">-lpthread</span></samp>. See
also <samp><span class="option">-static-libmpx</span></samp>. The runtime library behavior can be
influenced using various <samp><span class="env">CHKP_RT_*</span></samp> environment variables. See
<a href="https://gcc.gnu.org/wiki/Intel%20MPX%20support%20in%20the%20GCC%20compiler">https://gcc.gnu.org/wiki/Intel%20MPX%20support%20in%20the%20GCC%20compiler</a>
for more details.
<p>Generated instrumentation may be controlled by various
<samp><span class="option">-fchkp-*</span></samp> options and by the <code>bnd_variable_size</code>
structure field attribute (see <a href="Type-Attributes.html#Type-Attributes">Type Attributes</a>) and
<code>bnd_legacy</code>, and <code>bnd_instrument</code> function attributes
(see <a href="Function-Attributes.html#Function-Attributes">Function Attributes</a>). GCC also provides a number of built-in
functions for controlling the Pointer Bounds Checker. See <a href="Pointer-Bounds-Checker-builtins.html#Pointer-Bounds-Checker-builtins">Pointer Bounds Checker builtins</a>, for more information.
<br><dt><code>-fchkp-check-incomplete-type</code><dd><a name="index-fchkp_002dcheck_002dincomplete_002dtype-678"></a><a name="index-fno_002dchkp_002dcheck_002dincomplete_002dtype-679"></a>Generate pointer bounds checks for variables with incomplete type.
Enabled by default.
<br><dt><code>-fchkp-narrow-bounds</code><dd><a name="index-fchkp_002dnarrow_002dbounds-680"></a><a name="index-fno_002dchkp_002dnarrow_002dbounds-681"></a>Controls bounds used by Pointer Bounds Checker for pointers to object
fields. If narrowing is enabled then field bounds are used. Otherwise
object bounds are used. See also <samp><span class="option">-fchkp-narrow-to-innermost-array</span></samp>
and <samp><span class="option">-fchkp-first-field-has-own-bounds</span></samp>. Enabled by default.
<br><dt><code>-fchkp-first-field-has-own-bounds</code><dd><a name="index-fchkp_002dfirst_002dfield_002dhas_002down_002dbounds-682"></a><a name="index-fno_002dchkp_002dfirst_002dfield_002dhas_002down_002dbounds-683"></a>Forces Pointer Bounds Checker to use narrowed bounds for the address of the
first field in the structure. By default a pointer to the first field has
the same bounds as a pointer to the whole structure.
<br><dt><code>-fchkp-narrow-to-innermost-array</code><dd><a name="index-fchkp_002dnarrow_002dto_002dinnermost_002darray-684"></a><a name="index-fno_002dchkp_002dnarrow_002dto_002dinnermost_002darray-685"></a>Forces Pointer Bounds Checker to use bounds of the innermost arrays in
case of nested static array access. By default this option is disabled and
bounds of the outermost array are used.
<br><dt><code>-fchkp-optimize</code><dd><a name="index-fchkp_002doptimize-686"></a><a name="index-fno_002dchkp_002doptimize-687"></a>Enables Pointer Bounds Checker optimizations. Enabled by default at
optimization levels <samp><span class="option">-O</span></samp>, <samp><span class="option">-O2</span></samp>, <samp><span class="option">-O3</span></samp>.
<br><dt><code>-fchkp-use-fast-string-functions</code><dd><a name="index-fchkp_002duse_002dfast_002dstring_002dfunctions-688"></a><a name="index-fno_002dchkp_002duse_002dfast_002dstring_002dfunctions-689"></a>Enables use of <code>*_nobnd</code> versions of string functions (not copying bounds)
by Pointer Bounds Checker. Disabled by default.
<br><dt><code>-fchkp-use-nochk-string-functions</code><dd><a name="index-fchkp_002duse_002dnochk_002dstring_002dfunctions-690"></a><a name="index-fno_002dchkp_002duse_002dnochk_002dstring_002dfunctions-691"></a>Enables use of <code>*_nochk</code> versions of string functions (not checking bounds)
by Pointer Bounds Checker. Disabled by default.
<br><dt><code>-fchkp-use-static-bounds</code><dd><a name="index-fchkp_002duse_002dstatic_002dbounds-692"></a><a name="index-fno_002dchkp_002duse_002dstatic_002dbounds-693"></a>Allow Pointer Bounds Checker to generate static bounds holding
bounds of static variables. Enabled by default.
<br><dt><code>-fchkp-use-static-const-bounds</code><dd><a name="index-fchkp_002duse_002dstatic_002dconst_002dbounds-694"></a><a name="index-fno_002dchkp_002duse_002dstatic_002dconst_002dbounds-695"></a>Use statically-initialized bounds for constant bounds instead of
generating them each time they are required. By default enabled when
<samp><span class="option">-fchkp-use-static-bounds</span></samp> is enabled.
<br><dt><code>-fchkp-treat-zero-dynamic-size-as-infinite</code><dd><a name="index-fchkp_002dtreat_002dzero_002ddynamic_002dsize_002das_002dinfinite-696"></a><a name="index-fno_002dchkp_002dtreat_002dzero_002ddynamic_002dsize_002das_002dinfinite-697"></a>With this option, objects with incomplete type whose
dynamically-obtained size is zero are treated as having infinite size
instead by Pointer Bounds
Checker. This option may be helpful if a program is linked with a library
missing size information for some symbols. Disabled by default.
<br><dt><code>-fchkp-check-read</code><dd><a name="index-fchkp_002dcheck_002dread-698"></a><a name="index-fno_002dchkp_002dcheck_002dread-699"></a>Instructs Pointer Bounds Checker to generate checks for all read
accesses to memory. Enabled by default.
<br><dt><code>-fchkp-check-write</code><dd><a name="index-fchkp_002dcheck_002dwrite-700"></a><a name="index-fno_002dchkp_002dcheck_002dwrite-701"></a>Instructs Pointer Bounds Checker to generate checks for all write
accesses to memory. Enabled by default.
<br><dt><code>-fchkp-store-bounds</code><dd><a name="index-fchkp_002dstore_002dbounds-702"></a><a name="index-fno_002dchkp_002dstore_002dbounds-703"></a>Instructs Pointer Bounds Checker to generate bounds stores for
pointer writes. Enabled by default.
<br><dt><code>-fchkp-instrument-calls</code><dd><a name="index-fchkp_002dinstrument_002dcalls-704"></a><a name="index-fno_002dchkp_002dinstrument_002dcalls-705"></a>Instructs Pointer Bounds Checker to pass pointer bounds to calls.
Enabled by default.
<br><dt><code>-fchkp-instrument-marked-only</code><dd><a name="index-fchkp_002dinstrument_002dmarked_002donly-706"></a><a name="index-fno_002dchkp_002dinstrument_002dmarked_002donly-707"></a>Instructs Pointer Bounds Checker to instrument only functions
marked with the <code>bnd_instrument</code> attribute
(see <a href="Function-Attributes.html#Function-Attributes">Function Attributes</a>). Disabled by default.
<br><dt><code>-fchkp-use-wrappers</code><dd><a name="index-fchkp_002duse_002dwrappers-708"></a><a name="index-fno_002dchkp_002duse_002dwrappers-709"></a>Allows Pointer Bounds Checker to replace calls to built-in functions
with calls to wrapper functions. When <samp><span class="option">-fchkp-use-wrappers</span></samp>
is used to link a program, the GCC driver automatically links
against <samp><span class="file">libmpxwrappers</span></samp>. See also <samp><span class="option">-static-libmpxwrappers</span></samp>.
Enabled by default.
<br><dt><code>-fdump-final-insns</code><span class="roman">[</span><code>=</code><var>file</var><span class="roman">]</span><dd><a name="index-fdump_002dfinal_002dinsns-710"></a>Dump the final internal representation (RTL) to <var>file</var>. If the
optional argument is omitted (or if <var>file</var> is <code>.</code>), the name
of the dump file is determined by appending <code>.gkd</code> to the
compilation output file name.
<br><dt><code>-fcompare-debug</code><span class="roman">[</span><code>=</code><var>opts</var><span class="roman">]</span><dd><a name="index-fcompare_002ddebug-711"></a><a name="index-fno_002dcompare_002ddebug-712"></a>If no error occurs during compilation, run the compiler a second time,
adding <var>opts</var> and <samp><span class="option">-fcompare-debug-second</span></samp> to the arguments
passed to the second compilation. Dump the final internal
representation in both compilations, and print an error if they differ.
<p>If the equal sign is omitted, the default <samp><span class="option">-gtoggle</span></samp> is used.
<p>The environment variable <samp><span class="env">GCC_COMPARE_DEBUG</span></samp>, if defined, non-empty
and nonzero, implicitly enables <samp><span class="option">-fcompare-debug</span></samp>. If
<samp><span class="env">GCC_COMPARE_DEBUG</span></samp> is defined to a string starting with a dash,
then it is used for <var>opts</var>, otherwise the default <samp><span class="option">-gtoggle</span></samp>
is used.
<p><samp><span class="option">-fcompare-debug=</span></samp>, with the equal sign but without <var>opts</var>,
is equivalent to <samp><span class="option">-fno-compare-debug</span></samp>, which disables the dumping
of the final representation and the second compilation, preventing even
<samp><span class="env">GCC_COMPARE_DEBUG</span></samp> from taking effect.
<p>To verify full coverage during <samp><span class="option">-fcompare-debug</span></samp> testing, set
<samp><span class="env">GCC_COMPARE_DEBUG</span></samp> to say <samp><span class="option">-fcompare-debug-not-overridden</span></samp>,
which GCC rejects as an invalid option in any actual compilation
(rather than preprocessing, assembly or linking). To get just a
warning, setting <samp><span class="env">GCC_COMPARE_DEBUG</span></samp> to ‘<samp><span class="samp">-w%n-fcompare-debug
not overridden</span></samp>’ will do.
<br><dt><code>-fcompare-debug-second</code><dd><a name="index-fcompare_002ddebug_002dsecond-713"></a>This option is implicitly passed to the compiler for the second
compilation requested by <samp><span class="option">-fcompare-debug</span></samp>, along with options to
silence warnings, and omitting other options that would cause
side-effect compiler outputs to files or to the standard output. Dump
files and preserved temporary files are renamed so as to contain the
<code>.gk</code> additional extension during the second compilation, to avoid
overwriting those generated by the first.
<p>When this option is passed to the compiler driver, it causes the
<em>first</em> compilation to be skipped, which makes it useful for little
other than debugging the compiler proper.
<br><dt><code>-feliminate-dwarf2-dups</code><dd><a name="index-feliminate_002ddwarf2_002ddups-714"></a>Compress DWARF 2 debugging information by eliminating duplicated
information about each symbol. This option only makes sense when
generating DWARF 2 debugging information with <samp><span class="option">-gdwarf-2</span></samp>.
<br><dt><code>-femit-struct-debug-baseonly</code><dd><a name="index-femit_002dstruct_002ddebug_002dbaseonly-715"></a>Emit debug information for struct-like types
only when the base name of the compilation source file
matches the base name of file in which the struct is defined.
<p>This option substantially reduces the size of debugging information,
but at significant potential loss in type information to the debugger.
See <samp><span class="option">-femit-struct-debug-reduced</span></samp> for a less aggressive option.
See <samp><span class="option">-femit-struct-debug-detailed</span></samp> for more detailed control.
<p>This option works only with DWARF 2.
<br><dt><code>-femit-struct-debug-reduced</code><dd><a name="index-femit_002dstruct_002ddebug_002dreduced-716"></a>Emit debug information for struct-like types
only when the base name of the compilation source file
matches the base name of file in which the type is defined,
unless the struct is a template or defined in a system header.
<p>This option significantly reduces the size of debugging information,
with some potential loss in type information to the debugger.
See <samp><span class="option">-femit-struct-debug-baseonly</span></samp> for a more aggressive option.
See <samp><span class="option">-femit-struct-debug-detailed</span></samp> for more detailed control.
<p>This option works only with DWARF 2.
<br><dt><code>-femit-struct-debug-detailed</code><span class="roman">[</span><code>=</code><var>spec-list</var><span class="roman">]</span><dd><a name="index-femit_002dstruct_002ddebug_002ddetailed-717"></a>Specify the struct-like types
for which the compiler generates debug information.
The intent is to reduce duplicate struct debug information
between different object files within the same program.
<p>This option is a detailed version of
<samp><span class="option">-femit-struct-debug-reduced</span></samp> and <samp><span class="option">-femit-struct-debug-baseonly</span></samp>,
which serves for most needs.
<p>A specification has the syntax<br>
[‘<samp><span class="samp">dir:</span></samp>’|‘<samp><span class="samp">ind:</span></samp>’][‘<samp><span class="samp">ord:</span></samp>’|‘<samp><span class="samp">gen:</span></samp>’](‘<samp><span class="samp">any</span></samp>’|‘<samp><span class="samp">sys</span></samp>’|‘<samp><span class="samp">base</span></samp>’|‘<samp><span class="samp">none</span></samp>’)
<p>The optional first word limits the specification to
structs that are used directly (‘<samp><span class="samp">dir:</span></samp>’) or used indirectly (‘<samp><span class="samp">ind:</span></samp>’).
A struct type is used directly when it is the type of a variable, member.
Indirect uses arise through pointers to structs.
That is, when use of an incomplete struct is valid, the use is indirect.
An example is
‘<samp><span class="samp">struct one direct; struct two * indirect;</span></samp>’.
<p>The optional second word limits the specification to
ordinary structs (‘<samp><span class="samp">ord:</span></samp>’) or generic structs (‘<samp><span class="samp">gen:</span></samp>’).
Generic structs are a bit complicated to explain.
For C++, these are non-explicit specializations of template classes,
or non-template classes within the above.
Other programming languages have generics,
but <samp><span class="option">-femit-struct-debug-detailed</span></samp> does not yet implement them.
<p>The third word specifies the source files for those
structs for which the compiler should emit debug information.
The values ‘<samp><span class="samp">none</span></samp>’ and ‘<samp><span class="samp">any</span></samp>’ have the normal meaning.
The value ‘<samp><span class="samp">base</span></samp>’ means that
the base of name of the file in which the type declaration appears
must match the base of the name of the main compilation file.
In practice, this means that when compiling <samp><span class="file">foo.c</span></samp>, debug information
is generated for types declared in that file and <samp><span class="file">foo.h</span></samp>,
but not other header files.
The value ‘<samp><span class="samp">sys</span></samp>’ means those types satisfying ‘<samp><span class="samp">base</span></samp>’
or declared in system or compiler headers.
<p>You may need to experiment to determine the best settings for your application.
<p>The default is <samp><span class="option">-femit-struct-debug-detailed=all</span></samp>.
<p>This option works only with DWARF 2.
<br><dt><code>-fno-merge-debug-strings</code><dd><a name="index-fmerge_002ddebug_002dstrings-718"></a><a name="index-fno_002dmerge_002ddebug_002dstrings-719"></a>Direct the linker to not merge together strings in the debugging
information that are identical in different object files. Merging is
not supported by all assemblers or linkers. Merging decreases the size
of the debug information in the output file at the cost of increasing
link processing time. Merging is enabled by default.
<br><dt><code>-fdebug-prefix-map=</code><var>old</var><code>=</code><var>new</var><dd><a name="index-fdebug_002dprefix_002dmap-720"></a>When compiling files in directory <samp><var>old</var></samp>, record debugging
information describing them as in <samp><var>new</var></samp> instead.
<br><dt><code>-fno-dwarf2-cfi-asm</code><dd><a name="index-fdwarf2_002dcfi_002dasm-721"></a><a name="index-fno_002ddwarf2_002dcfi_002dasm-722"></a>Emit DWARF 2 unwind info as compiler generated <code>.eh_frame</code> section
instead of using GAS <code>.cfi_*</code> directives.
<p><a name="index-g_t_0040command_007bprof_007d-723"></a><br><dt><code>-p</code><dd><a name="index-p-724"></a>Generate extra code to write profile information suitable for the
analysis program <samp><span class="command">prof</span></samp>. You must use this option when compiling
the source files you want data about, and you must also use it when
linking.
<p><a name="index-g_t_0040command_007bgprof_007d-725"></a><br><dt><code>-pg</code><dd><a name="index-pg-726"></a>Generate extra code to write profile information suitable for the
analysis program <samp><span class="command">gprof</span></samp>. You must use this option when compiling
the source files you want data about, and you must also use it when
linking.
<br><dt><code>-Q</code><dd><a name="index-Q-727"></a>Makes the compiler print out each function name as it is compiled, and
print some statistics about each pass when it finishes.
<br><dt><code>-ftime-report</code><dd><a name="index-ftime_002dreport-728"></a>Makes the compiler print some statistics about the time consumed by each
pass when it finishes.
<br><dt><code>-fmem-report</code><dd><a name="index-fmem_002dreport-729"></a>Makes the compiler print some statistics about permanent memory
allocation when it finishes.
<br><dt><code>-fmem-report-wpa</code><dd><a name="index-fmem_002dreport_002dwpa-730"></a>Makes the compiler print some statistics about permanent memory
allocation for the WPA phase only.
<br><dt><code>-fpre-ipa-mem-report</code><dd><a name="index-fpre_002dipa_002dmem_002dreport-731"></a><br><dt><code>-fpost-ipa-mem-report</code><dd><a name="index-fpost_002dipa_002dmem_002dreport-732"></a>Makes the compiler print some statistics about permanent memory
allocation before or after interprocedural optimization.
<br><dt><code>-fprofile-report</code><dd><a name="index-fprofile_002dreport-733"></a>Makes the compiler print some statistics about consistency of the
(estimated) profile and effect of individual passes.
<br><dt><code>-fstack-usage</code><dd><a name="index-fstack_002dusage-734"></a>Makes the compiler output stack usage information for the program, on a
per-function basis. The filename for the dump is made by appending
<samp><span class="file">.su</span></samp> to the <var>auxname</var>. <var>auxname</var> is generated from the name of
the output file, if explicitly specified and it is not an executable,
otherwise it is the basename of the source file. An entry is made up
of three fields:
<ul>
<li>The name of the function.
<li>A number of bytes.
<li>One or more qualifiers: <code>static</code>, <code>dynamic</code>, <code>bounded</code>.
</ul>
<p>The qualifier <code>static</code> means that the function manipulates the stack
statically: a fixed number of bytes are allocated for the frame on function
entry and released on function exit; no stack adjustments are otherwise made
in the function. The second field is this fixed number of bytes.
<p>The qualifier <code>dynamic</code> means that the function manipulates the stack
dynamically: in addition to the static allocation described above, stack
adjustments are made in the body of the function, for example to push/pop
arguments around function calls. If the qualifier <code>bounded</code> is also
present, the amount of these adjustments is bounded at compile time and
the second field is an upper bound of the total amount of stack used by
the function. If it is not present, the amount of these adjustments is
not bounded at compile time and the second field only represents the
bounded part.
<br><dt><code>-fprofile-arcs</code><dd><a name="index-fprofile_002darcs-735"></a>Add code so that program flow <dfn>arcs</dfn> are instrumented. During
execution the program records how many times each branch and call is
executed and how many times it is taken or returns. When the compiled
program exits it saves this data to a file called
<samp><var>auxname</var><span class="file">.gcda</span></samp> for each source file. The data may be used for
profile-directed optimizations (<samp><span class="option">-fbranch-probabilities</span></samp>), or for
test coverage analysis (<samp><span class="option">-ftest-coverage</span></samp>). Each object file's
<var>auxname</var> is generated from the name of the output file, if
explicitly specified and it is not the final executable, otherwise it is
the basename of the source file. In both cases any suffix is removed
(e.g. <samp><span class="file">foo.gcda</span></samp> for input file <samp><span class="file">dir/foo.c</span></samp>, or
<samp><span class="file">dir/foo.gcda</span></samp> for output file specified as <samp><span class="option">-o dir/foo.o</span></samp>).
See <a href="Cross_002dprofiling.html#Cross_002dprofiling">Cross-profiling</a>.
<p><a name="index-g_t_0040command_007bgcov_007d-736"></a><br><dt><code>--coverage</code><dd><a name="index-coverage-737"></a>
This option is used to compile and link code instrumented for coverage
analysis. The option is a synonym for <samp><span class="option">-fprofile-arcs</span></samp>
<samp><span class="option">-ftest-coverage</span></samp> (when compiling) and <samp><span class="option">-lgcov</span></samp> (when
linking). See the documentation for those options for more details.
<ul>
<li>Compile the source files with <samp><span class="option">-fprofile-arcs</span></samp> plus optimization
and code generation options. For test coverage analysis, use the
additional <samp><span class="option">-ftest-coverage</span></samp> option. You do not need to profile
every source file in a program.
<li>Link your object files with <samp><span class="option">-lgcov</span></samp> or <samp><span class="option">-fprofile-arcs</span></samp>
(the latter implies the former).
<li>Run the program on a representative workload to generate the arc profile
information. This may be repeated any number of times. You can run
concurrent instances of your program, and provided that the file system
supports locking, the data files will be correctly updated. Also
<code>fork</code> calls are detected and correctly handled (double counting
will not happen).
<li>For profile-directed optimizations, compile the source files again with
the same optimization and code generation options plus
<samp><span class="option">-fbranch-probabilities</span></samp> (see <a href="Optimize-Options.html#Optimize-Options">Options that Control Optimization</a>).
<li>For test coverage analysis, use <samp><span class="command">gcov</span></samp> to produce human readable
information from the <samp><span class="file">.gcno</span></samp> and <samp><span class="file">.gcda</span></samp> files. Refer to the
<samp><span class="command">gcov</span></samp> documentation for further information.
</ul>
<p>With <samp><span class="option">-fprofile-arcs</span></samp>, for each function of your program GCC
creates a program flow graph, then finds a spanning tree for the graph.
Only arcs that are not on the spanning tree have to be instrumented: the
compiler adds code to count the number of times that these arcs are
executed. When an arc is the only exit or only entrance to a block, the
instrumentation code can be added to the block; otherwise, a new basic
block must be created to hold the instrumentation code.
<br><dt><code>-ftest-coverage</code><dd><a name="index-ftest_002dcoverage-738"></a>Produce a notes file that the <samp><span class="command">gcov</span></samp> code-coverage utility
(see <a href="Gcov.html#Gcov"><samp><span class="command">gcov</span></samp>—a Test Coverage Program</a>) can use to
show program coverage. Each source file's note file is called
<samp><var>auxname</var><span class="file">.gcno</span></samp>. Refer to the <samp><span class="option">-fprofile-arcs</span></samp> option
above for a description of <var>auxname</var> and instructions on how to
generate test coverage data. Coverage data matches the source files
more closely if you do not optimize.
<br><dt><code>-fdbg-cnt-list</code><dd><a name="index-fdbg_002dcnt_002dlist-739"></a>Print the name and the counter upper bound for all debug counters.
<br><dt><code>-fdbg-cnt=</code><var>counter-value-list</var><dd><a name="index-fdbg_002dcnt-740"></a>Set the internal debug counter upper bound. <var>counter-value-list</var>
is a comma-separated list of <var>name</var>:<var>value</var> pairs
which sets the upper bound of each debug counter <var>name</var> to <var>value</var>.
All debug counters have the initial upper bound of <code>UINT_MAX</code>;
thus <code>dbg_cnt</code> returns true always unless the upper bound
is set by this option.
For example, with <samp><span class="option">-fdbg-cnt=dce:10,tail_call:0</span></samp>,
<code>dbg_cnt(dce)</code> returns true only for first 10 invocations.
<br><dt><code>-fenable-</code><var>kind</var><code>-</code><var>pass</var><dt><code>-fdisable-</code><var>kind</var><code>-</code><var>pass</var><code>=</code><var>range-list</var><dd><a name="index-fdisable_002d-741"></a><a name="index-fenable_002d-742"></a>
This is a set of options that are used to explicitly disable/enable
optimization passes. These options are intended for use for debugging GCC.
Compiler users should use regular options for enabling/disabling
passes instead.
<dl>
<dt><code>-fdisable-ipa-</code><var>pass</var><dd>Disable IPA pass <var>pass</var>. <var>pass</var> is the pass name. If the same pass is
statically invoked in the compiler multiple times, the pass name should be
appended with a sequential number starting from 1.
<br><dt><code>-fdisable-rtl-</code><var>pass</var><dt><code>-fdisable-rtl-</code><var>pass</var><code>=</code><var>range-list</var><dd>Disable RTL pass <var>pass</var>. <var>pass</var> is the pass name. If the same pass is
statically invoked in the compiler multiple times, the pass name should be
appended with a sequential number starting from 1. <var>range-list</var> is a
comma-separated list of function ranges or assembler names. Each range is a number
pair separated by a colon. The range is inclusive in both ends. If the range
is trivial, the number pair can be simplified as a single number. If the
function's call graph node's <var>uid</var> falls within one of the specified ranges,
the <var>pass</var> is disabled for that function. The <var>uid</var> is shown in the
function header of a dump file, and the pass names can be dumped by using
option <samp><span class="option">-fdump-passes</span></samp>.
<br><dt><code>-fdisable-tree-</code><var>pass</var><dt><code>-fdisable-tree-</code><var>pass</var><code>=</code><var>range-list</var><dd>Disable tree pass <var>pass</var>. See <samp><span class="option">-fdisable-rtl</span></samp> for the description of
option arguments.
<br><dt><code>-fenable-ipa-</code><var>pass</var><dd>Enable IPA pass <var>pass</var>. <var>pass</var> is the pass name. If the same pass is
statically invoked in the compiler multiple times, the pass name should be
appended with a sequential number starting from 1.
<br><dt><code>-fenable-rtl-</code><var>pass</var><dt><code>-fenable-rtl-</code><var>pass</var><code>=</code><var>range-list</var><dd>Enable RTL pass <var>pass</var>. See <samp><span class="option">-fdisable-rtl</span></samp> for option argument
description and examples.
<br><dt><code>-fenable-tree-</code><var>pass</var><dt><code>-fenable-tree-</code><var>pass</var><code>=</code><var>range-list</var><dd>Enable tree pass <var>pass</var>. See <samp><span class="option">-fdisable-rtl</span></samp> for the description
of option arguments.
</dl>
<p>Here are some examples showing uses of these options.
<pre class="smallexample">
# disable ccp1 for all functions
-fdisable-tree-ccp1
# disable complete unroll for function whose cgraph node uid is 1
-fenable-tree-cunroll=1
# disable gcse2 for functions at the following ranges [1,1],
# [300,400], and [400,1000]
# disable gcse2 for functions foo and foo2
-fdisable-rtl-gcse2=foo,foo2
# disable early inlining
-fdisable-tree-einline
# disable ipa inlining
-fdisable-ipa-inline
# enable tree full unroll
-fenable-tree-unroll
</pre>
<br><dt><code>-d</code><var>letters</var><dt><code>-fdump-rtl-</code><var>pass</var><dt><code>-fdump-rtl-</code><var>pass</var><code>=</code><var>filename</var><dd><a name="index-d-743"></a><a name="index-fdump_002drtl_002d_0040var_007bpass_007d-744"></a>Says to make debugging dumps during compilation at times specified by
<var>letters</var>. This is used for debugging the RTL-based passes of the
compiler. The file names for most of the dumps are made by appending
a pass number and a word to the <var>dumpname</var>, and the files are
created in the directory of the output file. In case of
<samp><span class="option">=</span><var>filename</var></samp> option, the dump is output on the given file
instead of the pass numbered dump files. Note that the pass number is
computed statically as passes get registered into the pass manager.
Thus the numbering is not related to the dynamic order of execution of
passes. In particular, a pass installed by a plugin could have a
number over 200 even if it executed quite early. <var>dumpname</var> is
generated from the name of the output file, if explicitly specified
and it is not an executable, otherwise it is the basename of the
source file. These switches may have different effects when
<samp><span class="option">-E</span></samp> is used for preprocessing.
<p>Debug dumps can be enabled with a <samp><span class="option">-fdump-rtl</span></samp> switch or some
<samp><span class="option">-d</span></samp> option <var>letters</var>. Here are the possible
letters for use in <var>pass</var> and <var>letters</var>, and their meanings:
<dl>
<dt><code>-fdump-rtl-alignments</code><dd><a name="index-fdump_002drtl_002dalignments-745"></a>Dump after branch alignments have been computed.
<br><dt><code>-fdump-rtl-asmcons</code><dd><a name="index-fdump_002drtl_002dasmcons-746"></a>Dump after fixing rtl statements that have unsatisfied in/out constraints.
<br><dt><code>-fdump-rtl-auto_inc_dec</code><dd><a name="index-fdump_002drtl_002dauto_005finc_005fdec-747"></a>Dump after auto-inc-dec discovery. This pass is only run on
architectures that have auto inc or auto dec instructions.
<br><dt><code>-fdump-rtl-barriers</code><dd><a name="index-fdump_002drtl_002dbarriers-748"></a>Dump after cleaning up the barrier instructions.
<br><dt><code>-fdump-rtl-bbpart</code><dd><a name="index-fdump_002drtl_002dbbpart-749"></a>Dump after partitioning hot and cold basic blocks.
<br><dt><code>-fdump-rtl-bbro</code><dd><a name="index-fdump_002drtl_002dbbro-750"></a>Dump after block reordering.
<br><dt><code>-fdump-rtl-btl1</code><dt><code>-fdump-rtl-btl2</code><dd><a name="index-fdump_002drtl_002dbtl2-751"></a><a name="index-fdump_002drtl_002dbtl2-752"></a><samp><span class="option">-fdump-rtl-btl1</span></samp> and <samp><span class="option">-fdump-rtl-btl2</span></samp> enable dumping
after the two branch
target load optimization passes.
<br><dt><code>-fdump-rtl-bypass</code><dd><a name="index-fdump_002drtl_002dbypass-753"></a>Dump after jump bypassing and control flow optimizations.
<br><dt><code>-fdump-rtl-combine</code><dd><a name="index-fdump_002drtl_002dcombine-754"></a>Dump after the RTL instruction combination pass.
<br><dt><code>-fdump-rtl-compgotos</code><dd><a name="index-fdump_002drtl_002dcompgotos-755"></a>Dump after duplicating the computed gotos.
<br><dt><code>-fdump-rtl-ce1</code><dt><code>-fdump-rtl-ce2</code><dt><code>-fdump-rtl-ce3</code><dd><a name="index-fdump_002drtl_002dce1-756"></a><a name="index-fdump_002drtl_002dce2-757"></a><a name="index-fdump_002drtl_002dce3-758"></a><samp><span class="option">-fdump-rtl-ce1</span></samp>, <samp><span class="option">-fdump-rtl-ce2</span></samp>, and
<samp><span class="option">-fdump-rtl-ce3</span></samp> enable dumping after the three
if conversion passes.
<br><dt><code>-fdump-rtl-cprop_hardreg</code><dd><a name="index-fdump_002drtl_002dcprop_005fhardreg-759"></a>Dump after hard register copy propagation.
<br><dt><code>-fdump-rtl-csa</code><dd><a name="index-fdump_002drtl_002dcsa-760"></a>Dump after combining stack adjustments.
<br><dt><code>-fdump-rtl-cse1</code><dt><code>-fdump-rtl-cse2</code><dd><a name="index-fdump_002drtl_002dcse1-761"></a><a name="index-fdump_002drtl_002dcse2-762"></a><samp><span class="option">-fdump-rtl-cse1</span></samp> and <samp><span class="option">-fdump-rtl-cse2</span></samp> enable dumping after
the two common subexpression elimination passes.
<br><dt><code>-fdump-rtl-dce</code><dd><a name="index-fdump_002drtl_002ddce-763"></a>Dump after the standalone dead code elimination passes.
<br><dt><code>-fdump-rtl-dbr</code><dd><a name="index-fdump_002drtl_002ddbr-764"></a>Dump after delayed branch scheduling.
<br><dt><code>-fdump-rtl-dce1</code><dt><code>-fdump-rtl-dce2</code><dd><a name="index-fdump_002drtl_002ddce1-765"></a><a name="index-fdump_002drtl_002ddce2-766"></a><samp><span class="option">-fdump-rtl-dce1</span></samp> and <samp><span class="option">-fdump-rtl-dce2</span></samp> enable dumping after
the two dead store elimination passes.
<br><dt><code>-fdump-rtl-eh</code><dd><a name="index-fdump_002drtl_002deh-767"></a>Dump after finalization of EH handling code.
<br><dt><code>-fdump-rtl-eh_ranges</code><dd><a name="index-fdump_002drtl_002deh_005franges-768"></a>Dump after conversion of EH handling range regions.
<br><dt><code>-fdump-rtl-expand</code><dd><a name="index-fdump_002drtl_002dexpand-769"></a>Dump after RTL generation.
<br><dt><code>-fdump-rtl-fwprop1</code><dt><code>-fdump-rtl-fwprop2</code><dd><a name="index-fdump_002drtl_002dfwprop1-770"></a><a name="index-fdump_002drtl_002dfwprop2-771"></a><samp><span class="option">-fdump-rtl-fwprop1</span></samp> and <samp><span class="option">-fdump-rtl-fwprop2</span></samp> enable
dumping after the two forward propagation passes.
<br><dt><code>-fdump-rtl-gcse1</code><dt><code>-fdump-rtl-gcse2</code><dd><a name="index-fdump_002drtl_002dgcse1-772"></a><a name="index-fdump_002drtl_002dgcse2-773"></a><samp><span class="option">-fdump-rtl-gcse1</span></samp> and <samp><span class="option">-fdump-rtl-gcse2</span></samp> enable dumping
after global common subexpression elimination.
<br><dt><code>-fdump-rtl-init-regs</code><dd><a name="index-fdump_002drtl_002dinit_002dregs-774"></a>Dump after the initialization of the registers.
<br><dt><code>-fdump-rtl-initvals</code><dd><a name="index-fdump_002drtl_002dinitvals-775"></a>Dump after the computation of the initial value sets.
<br><dt><code>-fdump-rtl-into_cfglayout</code><dd><a name="index-fdump_002drtl_002dinto_005fcfglayout-776"></a>Dump after converting to cfglayout mode.
<br><dt><code>-fdump-rtl-ira</code><dd><a name="index-fdump_002drtl_002dira-777"></a>Dump after iterated register allocation.
<br><dt><code>-fdump-rtl-jump</code><dd><a name="index-fdump_002drtl_002djump-778"></a>Dump after the second jump optimization.
<br><dt><code>-fdump-rtl-loop2</code><dd><a name="index-fdump_002drtl_002dloop2-779"></a><samp><span class="option">-fdump-rtl-loop2</span></samp> enables dumping after the rtl
loop optimization passes.
<br><dt><code>-fdump-rtl-mach</code><dd><a name="index-fdump_002drtl_002dmach-780"></a>Dump after performing the machine dependent reorganization pass, if that
pass exists.
<br><dt><code>-fdump-rtl-mode_sw</code><dd><a name="index-fdump_002drtl_002dmode_005fsw-781"></a>Dump after removing redundant mode switches.
<br><dt><code>-fdump-rtl-rnreg</code><dd><a name="index-fdump_002drtl_002drnreg-782"></a>Dump after register renumbering.
<br><dt><code>-fdump-rtl-outof_cfglayout</code><dd><a name="index-fdump_002drtl_002doutof_005fcfglayout-783"></a>Dump after converting from cfglayout mode.
<br><dt><code>-fdump-rtl-peephole2</code><dd><a name="index-fdump_002drtl_002dpeephole2-784"></a>Dump after the peephole pass.
<br><dt><code>-fdump-rtl-postreload</code><dd><a name="index-fdump_002drtl_002dpostreload-785"></a>Dump after post-reload optimizations.
<br><dt><code>-fdump-rtl-pro_and_epilogue</code><dd><a name="index-fdump_002drtl_002dpro_005fand_005fepilogue-786"></a>Dump after generating the function prologues and epilogues.
<br><dt><code>-fdump-rtl-sched1</code><dt><code>-fdump-rtl-sched2</code><dd><a name="index-fdump_002drtl_002dsched1-787"></a><a name="index-fdump_002drtl_002dsched2-788"></a><samp><span class="option">-fdump-rtl-sched1</span></samp> and <samp><span class="option">-fdump-rtl-sched2</span></samp> enable dumping
after the basic block scheduling passes.
<br><dt><code>-fdump-rtl-ree</code><dd><a name="index-fdump_002drtl_002dree-789"></a>Dump after sign/zero extension elimination.
<br><dt><code>-fdump-rtl-seqabstr</code><dd><a name="index-fdump_002drtl_002dseqabstr-790"></a>Dump after common sequence discovery.
<br><dt><code>-fdump-rtl-shorten</code><dd><a name="index-fdump_002drtl_002dshorten-791"></a>Dump after shortening branches.
<br><dt><code>-fdump-rtl-sibling</code><dd><a name="index-fdump_002drtl_002dsibling-792"></a>Dump after sibling call optimizations.
<br><dt><code>-fdump-rtl-split1</code><dt><code>-fdump-rtl-split2</code><dt><code>-fdump-rtl-split3</code><dt><code>-fdump-rtl-split4</code><dt><code>-fdump-rtl-split5</code><dd><a name="index-fdump_002drtl_002dsplit1-793"></a><a name="index-fdump_002drtl_002dsplit2-794"></a><a name="index-fdump_002drtl_002dsplit3-795"></a><a name="index-fdump_002drtl_002dsplit4-796"></a><a name="index-fdump_002drtl_002dsplit5-797"></a>These options enable dumping after five rounds of
instruction splitting.
<br><dt><code>-fdump-rtl-sms</code><dd><a name="index-fdump_002drtl_002dsms-798"></a>Dump after modulo scheduling. This pass is only run on some
architectures.
<br><dt><code>-fdump-rtl-stack</code><dd><a name="index-fdump_002drtl_002dstack-799"></a>Dump after conversion from GCC's “flat register file” registers to the
x87's stack-like registers. This pass is only run on x86 variants.
<br><dt><code>-fdump-rtl-subreg1</code><dt><code>-fdump-rtl-subreg2</code><dd><a name="index-fdump_002drtl_002dsubreg1-800"></a><a name="index-fdump_002drtl_002dsubreg2-801"></a><samp><span class="option">-fdump-rtl-subreg1</span></samp> and <samp><span class="option">-fdump-rtl-subreg2</span></samp> enable dumping after
the two subreg expansion passes.
<br><dt><code>-fdump-rtl-unshare</code><dd><a name="index-fdump_002drtl_002dunshare-802"></a>Dump after all rtl has been unshared.
<br><dt><code>-fdump-rtl-vartrack</code><dd><a name="index-fdump_002drtl_002dvartrack-803"></a>Dump after variable tracking.
<br><dt><code>-fdump-rtl-vregs</code><dd><a name="index-fdump_002drtl_002dvregs-804"></a>Dump after converting virtual registers to hard registers.
<br><dt><code>-fdump-rtl-web</code><dd><a name="index-fdump_002drtl_002dweb-805"></a>Dump after live range splitting.
<br><dt><code>-fdump-rtl-regclass</code><dt><code>-fdump-rtl-subregs_of_mode_init</code><dt><code>-fdump-rtl-subregs_of_mode_finish</code><dt><code>-fdump-rtl-dfinit</code><dt><code>-fdump-rtl-dfinish</code><dd><a name="index-fdump_002drtl_002dregclass-806"></a><a name="index-fdump_002drtl_002dsubregs_005fof_005fmode_005finit-807"></a><a name="index-fdump_002drtl_002dsubregs_005fof_005fmode_005ffinish-808"></a><a name="index-fdump_002drtl_002ddfinit-809"></a><a name="index-fdump_002drtl_002ddfinish-810"></a>These dumps are defined but always produce empty files.
<br><dt><code>-da</code><dt><code>-fdump-rtl-all</code><dd><a name="index-da-811"></a><a name="index-fdump_002drtl_002dall-812"></a>Produce all the dumps listed above.
<br><dt><code>-dA</code><dd><a name="index-dA-813"></a>Annotate the assembler output with miscellaneous debugging information.
<br><dt><code>-dD</code><dd><a name="index-dD-814"></a>Dump all macro definitions, at the end of preprocessing, in addition to
normal output.
<br><dt><code>-dH</code><dd><a name="index-dH-815"></a>Produce a core dump whenever an error occurs.
<br><dt><code>-dp</code><dd><a name="index-dp-816"></a>Annotate the assembler output with a comment indicating which
pattern and alternative is used. The length of each instruction is
also printed.
<br><dt><code>-dP</code><dd><a name="index-dP-817"></a>Dump the RTL in the assembler output as a comment before each instruction.
Also turns on <samp><span class="option">-dp</span></samp> annotation.
<br><dt><code>-dx</code><dd><a name="index-dx-818"></a>Just generate RTL for a function instead of compiling it. Usually used
with <samp><span class="option">-fdump-rtl-expand</span></samp>.
</dl>
<br><dt><code>-fdump-noaddr</code><dd><a name="index-fdump_002dnoaddr-819"></a>When doing debugging dumps, suppress address output. This makes it more
feasible to use diff on debugging dumps for compiler invocations with
different compiler binaries and/or different
text / bss / data / heap / stack / dso start locations.
<br><dt><code>-freport-bug</code><dd><a name="index-freport_002dbug-820"></a>Collect and dump debug information into temporary file if ICE in C/C++
compiler occured.
<br><dt><code>-fdump-unnumbered</code><dd><a name="index-fdump_002dunnumbered-821"></a>When doing debugging dumps, suppress instruction numbers and address output.
This makes it more feasible to use diff on debugging dumps for compiler
invocations with different options, in particular with and without
<samp><span class="option">-g</span></samp>.
<br><dt><code>-fdump-unnumbered-links</code><dd><a name="index-fdump_002dunnumbered_002dlinks-822"></a>When doing debugging dumps (see <samp><span class="option">-d</span></samp> option above), suppress
instruction numbers for the links to the previous and next instructions
in a sequence.
<br><dt><code>-fdump-translation-unit </code><span class="roman">(C++ only)</span><dt><code>-fdump-translation-unit-</code><var>options</var> <span class="roman">(C++ only)</span><dd><a name="index-fdump_002dtranslation_002dunit-823"></a>Dump a representation of the tree structure for the entire translation
unit to a file. The file name is made by appending <samp><span class="file">.tu</span></samp> to the
source file name, and the file is created in the same directory as the
output file. If the ‘<samp><span class="samp">-</span><var>options</var></samp>’ form is used, <var>options</var>
controls the details of the dump as described for the
<samp><span class="option">-fdump-tree</span></samp> options.
<br><dt><code>-fdump-class-hierarchy </code><span class="roman">(C++ only)</span><dt><code>-fdump-class-hierarchy-</code><var>options</var> <span class="roman">(C++ only)</span><dd><a name="index-fdump_002dclass_002dhierarchy-824"></a>Dump a representation of each class's hierarchy and virtual function
table layout to a file. The file name is made by appending
<samp><span class="file">.class</span></samp> to the source file name, and the file is created in the
same directory as the output file. If the ‘<samp><span class="samp">-</span><var>options</var></samp>’ form
is used, <var>options</var> controls the details of the dump as described
for the <samp><span class="option">-fdump-tree</span></samp> options.
<br><dt><code>-fdump-ipa-</code><var>switch</var><dd><a name="index-fdump_002dipa-825"></a>Control the dumping at various stages of inter-procedural analysis
language tree to a file. The file name is generated by appending a
switch specific suffix to the source file name, and the file is created
in the same directory as the output file. The following dumps are
possible:
<dl>
<dt>‘<samp><span class="samp">all</span></samp>’<dd>Enables all inter-procedural analysis dumps.
<br><dt>‘<samp><span class="samp">cgraph</span></samp>’<dd>Dumps information about call-graph optimization, unused function removal,
and inlining decisions.
<br><dt>‘<samp><span class="samp">inline</span></samp>’<dd>Dump after function inlining.
</dl>
<br><dt><code>-fdump-passes</code><dd><a name="index-fdump_002dpasses-826"></a>Dump the list of optimization passes that are turned on and off by
the current command-line options.
<br><dt><code>-fdump-statistics-</code><var>option</var><dd><a name="index-fdump_002dstatistics-827"></a>Enable and control dumping of pass statistics in a separate file. The
file name is generated by appending a suffix ending in
‘<samp><span class="samp">.statistics</span></samp>’ to the source file name, and the file is created in
the same directory as the output file. If the ‘<samp><span class="samp">-</span><var>option</var></samp>’
form is used, ‘<samp><span class="samp">-stats</span></samp>’ causes counters to be summed over the
whole compilation unit while ‘<samp><span class="samp">-details</span></samp>’ dumps every event as
the passes generate them. The default with no option is to sum
counters for each function compiled.
<br><dt><code>-fdump-tree-</code><var>switch</var><dt><code>-fdump-tree-</code><var>switch</var><code>-</code><var>options</var><dt><code>-fdump-tree-</code><var>switch</var><code>-</code><var>options</var><code>=</code><var>filename</var><dd><a name="index-fdump_002dtree-828"></a>Control the dumping at various stages of processing the intermediate
language tree to a file. The file name is generated by appending a
switch-specific suffix to the source file name, and the file is
created in the same directory as the output file. In case of
<samp><span class="option">=</span><var>filename</var></samp> option, the dump is output on the given file
instead of the auto named dump files. If the ‘<samp><span class="samp">-</span><var>options</var></samp>’
form is used, <var>options</var> is a list of ‘<samp><span class="samp">-</span></samp>’ separated options
which control the details of the dump. Not all options are applicable
to all dumps; those that are not meaningful are ignored. The
following options are available
<dl>
<dt>‘<samp><span class="samp">address</span></samp>’<dd>Print the address of each node. Usually this is not meaningful as it
changes according to the environment and source file. Its primary use
is for tying up a dump file with a debug environment.
<br><dt>‘<samp><span class="samp">asmname</span></samp>’<dd>If <code>DECL_ASSEMBLER_NAME</code> has been set for a given decl, use that
in the dump instead of <code>DECL_NAME</code>. Its primary use is ease of
use working backward from mangled names in the assembly file.
<br><dt>‘<samp><span class="samp">slim</span></samp>’<dd>When dumping front-end intermediate representations, inhibit dumping
of members of a scope or body of a function merely because that scope
has been reached. Only dump such items when they are directly reachable
by some other path.
<p>When dumping pretty-printed trees, this option inhibits dumping the
bodies of control structures.
<p>When dumping RTL, print the RTL in slim (condensed) form instead of
the default LISP-like representation.
<br><dt>‘<samp><span class="samp">raw</span></samp>’<dd>Print a raw representation of the tree. By default, trees are
pretty-printed into a C-like representation.
<br><dt>‘<samp><span class="samp">details</span></samp>’<dd>Enable more detailed dumps (not honored by every dump option). Also
include information from the optimization passes.
<br><dt>‘<samp><span class="samp">stats</span></samp>’<dd>Enable dumping various statistics about the pass (not honored by every dump
option).
<br><dt>‘<samp><span class="samp">blocks</span></samp>’<dd>Enable showing basic block boundaries (disabled in raw dumps).
<br><dt>‘<samp><span class="samp">graph</span></samp>’<dd>For each of the other indicated dump files (<samp><span class="option">-fdump-rtl-</span><var>pass</var></samp>),
dump a representation of the control flow graph suitable for viewing with
GraphViz to <samp><var>file</var><span class="file">.</span><var>passid</var><span class="file">.</span><var>pass</var><span class="file">.dot</span></samp>. Each function in
the file is pretty-printed as a subgraph, so that GraphViz can render them
all in a single plot.
<p>This option currently only works for RTL dumps, and the RTL is always
dumped in slim form.
<br><dt>‘<samp><span class="samp">vops</span></samp>’<dd>Enable showing virtual operands for every statement.
<br><dt>‘<samp><span class="samp">lineno</span></samp>’<dd>Enable showing line numbers for statements.
<br><dt>‘<samp><span class="samp">uid</span></samp>’<dd>Enable showing the unique ID (<code>DECL_UID</code>) for each variable.
<br><dt>‘<samp><span class="samp">verbose</span></samp>’<dd>Enable showing the tree dump for each statement.
<br><dt>‘<samp><span class="samp">eh</span></samp>’<dd>Enable showing the EH region number holding each statement.
<br><dt>‘<samp><span class="samp">scev</span></samp>’<dd>Enable showing scalar evolution analysis details.
<br><dt>‘<samp><span class="samp">optimized</span></samp>’<dd>Enable showing optimization information (only available in certain
passes).
<br><dt>‘<samp><span class="samp">missed</span></samp>’<dd>Enable showing missed optimization information (only available in certain
passes).
<br><dt>‘<samp><span class="samp">note</span></samp>’<dd>Enable other detailed optimization information (only available in
certain passes).
<br><dt>‘<samp><span class="samp">=</span><var>filename</var></samp>’<dd>Instead of an auto named dump file, output into the given file
name. The file names <samp><span class="file">stdout</span></samp> and <samp><span class="file">stderr</span></samp> are treated
specially and are considered already open standard streams. For
example,
<pre class="smallexample"> gcc -O2 -ftree-vectorize -fdump-tree-vect-blocks=foo.dump
-fdump-tree-pre=stderr file.c
</pre>
<p>outputs vectorizer dump into <samp><span class="file">foo.dump</span></samp>, while the PRE dump is
output on to <samp><span class="file">stderr</span></samp>. If two conflicting dump filenames are
given for the same pass, then the latter option overrides the earlier
one.
<br><dt>‘<samp><span class="samp">all</span></samp>’<dd>Turn on all options, except <samp><span class="option">raw</span></samp>, <samp><span class="option">slim</span></samp>, <samp><span class="option">verbose</span></samp>
and <samp><span class="option">lineno</span></samp>.
<br><dt>‘<samp><span class="samp">optall</span></samp>’<dd>Turn on all optimization options, i.e., <samp><span class="option">optimized</span></samp>,
<samp><span class="option">missed</span></samp>, and <samp><span class="option">note</span></samp>.
</dl>
<p>The following tree dumps are possible:
<dl>
<dt>‘<samp><span class="samp">original</span></samp>’<dd><a name="index-fdump_002dtree_002doriginal-829"></a>Dump before any tree based optimization, to <samp><var>file</var><span class="file">.original</span></samp>.
<br><dt>‘<samp><span class="samp">optimized</span></samp>’<dd><a name="index-fdump_002dtree_002doptimized-830"></a>Dump after all tree based optimization, to <samp><var>file</var><span class="file">.optimized</span></samp>.
<br><dt>‘<samp><span class="samp">gimple</span></samp>’<dd><a name="index-fdump_002dtree_002dgimple-831"></a>Dump each function before and after the gimplification pass to a file. The
file name is made by appending <samp><span class="file">.gimple</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">cfg</span></samp>’<dd><a name="index-fdump_002dtree_002dcfg-832"></a>Dump the control flow graph of each function to a file. The file name is
made by appending <samp><span class="file">.cfg</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">ch</span></samp>’<dd><a name="index-fdump_002dtree_002dch-833"></a>Dump each function after copying loop headers. The file name is made by
appending <samp><span class="file">.ch</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">ssa</span></samp>’<dd><a name="index-fdump_002dtree_002dssa-834"></a>Dump SSA related information to a file. The file name is made by appending
<samp><span class="file">.ssa</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">alias</span></samp>’<dd><a name="index-fdump_002dtree_002dalias-835"></a>Dump aliasing information for each function. The file name is made by
appending <samp><span class="file">.alias</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">ccp</span></samp>’<dd><a name="index-fdump_002dtree_002dccp-836"></a>Dump each function after CCP. The file name is made by appending
<samp><span class="file">.ccp</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">storeccp</span></samp>’<dd><a name="index-fdump_002dtree_002dstoreccp-837"></a>Dump each function after STORE-CCP. The file name is made by appending
<samp><span class="file">.storeccp</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">pre</span></samp>’<dd><a name="index-fdump_002dtree_002dpre-838"></a>Dump trees after partial redundancy elimination. The file name is made
by appending <samp><span class="file">.pre</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">fre</span></samp>’<dd><a name="index-fdump_002dtree_002dfre-839"></a>Dump trees after full redundancy elimination. The file name is made
by appending <samp><span class="file">.fre</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">copyprop</span></samp>’<dd><a name="index-fdump_002dtree_002dcopyprop-840"></a>Dump trees after copy propagation. The file name is made
by appending <samp><span class="file">.copyprop</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">store_copyprop</span></samp>’<dd><a name="index-fdump_002dtree_002dstore_005fcopyprop-841"></a>Dump trees after store copy-propagation. The file name is made
by appending <samp><span class="file">.store_copyprop</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">dce</span></samp>’<dd><a name="index-fdump_002dtree_002ddce-842"></a>Dump each function after dead code elimination. The file name is made by
appending <samp><span class="file">.dce</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">sra</span></samp>’<dd><a name="index-fdump_002dtree_002dsra-843"></a>Dump each function after performing scalar replacement of aggregates. The
file name is made by appending <samp><span class="file">.sra</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">sink</span></samp>’<dd><a name="index-fdump_002dtree_002dsink-844"></a>Dump each function after performing code sinking. The file name is made
by appending <samp><span class="file">.sink</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">dom</span></samp>’<dd><a name="index-fdump_002dtree_002ddom-845"></a>Dump each function after applying dominator tree optimizations. The file
name is made by appending <samp><span class="file">.dom</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">dse</span></samp>’<dd><a name="index-fdump_002dtree_002ddse-846"></a>Dump each function after applying dead store elimination. The file
name is made by appending <samp><span class="file">.dse</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">phiopt</span></samp>’<dd><a name="index-fdump_002dtree_002dphiopt-847"></a>Dump each function after optimizing PHI nodes into straightline code. The file
name is made by appending <samp><span class="file">.phiopt</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">forwprop</span></samp>’<dd><a name="index-fdump_002dtree_002dforwprop-848"></a>Dump each function after forward propagating single use variables. The file
name is made by appending <samp><span class="file">.forwprop</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">copyrename</span></samp>’<dd><a name="index-fdump_002dtree_002dcopyrename-849"></a>Dump each function after applying the copy rename optimization. The file
name is made by appending <samp><span class="file">.copyrename</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">nrv</span></samp>’<dd><a name="index-fdump_002dtree_002dnrv-850"></a>Dump each function after applying the named return value optimization on
generic trees. The file name is made by appending <samp><span class="file">.nrv</span></samp> to the source
file name.
<br><dt>‘<samp><span class="samp">vect</span></samp>’<dd><a name="index-fdump_002dtree_002dvect-851"></a>Dump each function after applying vectorization of loops. The file name is
made by appending <samp><span class="file">.vect</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">slp</span></samp>’<dd><a name="index-fdump_002dtree_002dslp-852"></a>Dump each function after applying vectorization of basic blocks. The file name
is made by appending <samp><span class="file">.slp</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">vrp</span></samp>’<dd><a name="index-fdump_002dtree_002dvrp-853"></a>Dump each function after Value Range Propagation (VRP). The file name
is made by appending <samp><span class="file">.vrp</span></samp> to the source file name.
<br><dt>‘<samp><span class="samp">all</span></samp>’<dd><a name="index-fdump_002dtree_002dall-854"></a>Enable all the available tree dumps with the flags provided in this option.
</dl>
<br><dt><code>-fopt-info</code><dt><code>-fopt-info-</code><var>options</var><dt><code>-fopt-info-</code><var>options</var><code>=</code><var>filename</var><dd><a name="index-fopt_002dinfo-855"></a>Controls optimization dumps from various optimization passes. If the
‘<samp><span class="samp">-</span><var>options</var></samp>’ form is used, <var>options</var> is a list of
‘<samp><span class="samp">-</span></samp>’ separated option keywords to select the dump details and
optimizations.
<p>The <var>options</var> can be divided into two groups: options describing the
verbosity of the dump, and options describing which optimizations
should be included. The options from both the groups can be freely
mixed as they are non-overlapping. However, in case of any conflicts,
the later options override the earlier options on the command
line.
<p>The following options control the dump verbosity:
<dl>
<dt>‘<samp><span class="samp">optimized</span></samp>’<dd>Print information when an optimization is successfully applied. It is
up to a pass to decide which information is relevant. For example, the
vectorizer passes print the source location of loops which are
successfully vectorized.
<br><dt>‘<samp><span class="samp">missed</span></samp>’<dd>Print information about missed optimizations. Individual passes
control which information to include in the output.
<br><dt>‘<samp><span class="samp">note</span></samp>’<dd>Print verbose information about optimizations, such as certain
transformations, more detailed messages about decisions etc.
<br><dt>‘<samp><span class="samp">all</span></samp>’<dd>Print detailed optimization information. This includes
‘<samp><span class="samp">optimized</span></samp>’, ‘<samp><span class="samp">missed</span></samp>’, and ‘<samp><span class="samp">note</span></samp>’.
</dl>
<p>One or more of the following option keywords can be used to describe a
group of optimizations:
<dl>
<dt>‘<samp><span class="samp">ipa</span></samp>’<dd>Enable dumps from all interprocedural optimizations.
<br><dt>‘<samp><span class="samp">loop</span></samp>’<dd>Enable dumps from all loop optimizations.
<br><dt>‘<samp><span class="samp">inline</span></samp>’<dd>Enable dumps from all inlining optimizations.
<br><dt>‘<samp><span class="samp">vec</span></samp>’<dd>Enable dumps from all vectorization optimizations.
<br><dt>‘<samp><span class="samp">optall</span></samp>’<dd>Enable dumps from all optimizations. This is a superset of
the optimization groups listed above.
</dl>
<p>If <var>options</var> is
omitted, it defaults to ‘<samp><span class="samp">optimized-optall</span></samp>’, which means to dump all
info about successful optimizations from all the passes.
<p>If the <var>filename</var> is provided, then the dumps from all the
applicable optimizations are concatenated into the <var>filename</var>.
Otherwise the dump is output onto <samp><span class="file">stderr</span></samp>. Though multiple
<samp><span class="option">-fopt-info</span></samp> options are accepted, only one of them can include
a <var>filename</var>. If other filenames are provided then all but the
first such option are ignored.
<p>Note that the output <var>filename</var> is overwritten
in case of multiple translation units. If a combined output from
multiple translation units is desired, <samp><span class="file">stderr</span></samp> should be used
instead.
<p>In the following example, the optimization info is output to
<samp><span class="file">stderr</span></samp>:
<pre class="smallexample"> gcc -O3 -fopt-info
</pre>
<p>This example:
<pre class="smallexample"> gcc -O3 -fopt-info-missed=missed.all
</pre>
<p class="noindent">outputs missed optimization report from all the passes into
<samp><span class="file">missed.all</span></samp>, and this one:
<pre class="smallexample"> gcc -O2 -ftree-vectorize -fopt-info-vec-missed
</pre>
<p class="noindent">prints information about missed optimization opportunities from
vectorization passes on <samp><span class="file">stderr</span></samp>.
Note that <samp><span class="option">-fopt-info-vec-missed</span></samp> is equivalent to
<samp><span class="option">-fopt-info-missed-vec</span></samp>.
<p>As another example,
<pre class="smallexample"> gcc -O3 -fopt-info-inline-optimized-missed=inline.txt
</pre>
<p class="noindent">outputs information about missed optimizations as well as
optimized locations from all the inlining passes into
<samp><span class="file">inline.txt</span></samp>.
<p>Finally, consider:
<pre class="smallexample"> gcc -fopt-info-vec-missed=vec.miss -fopt-info-loop-optimized=loop.opt
</pre>
<p class="noindent">Here the two output filenames <samp><span class="file">vec.miss</span></samp> and <samp><span class="file">loop.opt</span></samp> are
in conflict since only one output file is allowed. In this case, only
the first option takes effect and the subsequent options are
ignored. Thus only <samp><span class="file">vec.miss</span></samp> is produced which contains
dumps from the vectorizer about missed opportunities.
<br><dt><code>-frandom-seed=</code><var>string</var><dd><a name="index-frandom_002dseed-856"></a>This option provides a seed that GCC uses in place of
random numbers in generating certain symbol names
that have to be different in every compiled file. It is also used to
place unique stamps in coverage data files and the object files that
produce them. You can use the <samp><span class="option">-frandom-seed</span></samp> option to produce
reproducibly identical object files.
<p>The <var>string</var> can either be a number (decimal, octal or hex) or an
arbitrary string (in which case it's converted to a number by
computing CRC32).
<p>The <var>string</var> should be different for every file you compile.
<br><dt><code>-fsched-verbose=</code><var>n</var><dd><a name="index-fsched_002dverbose-857"></a>On targets that use instruction scheduling, this option controls the
amount of debugging output the scheduler prints. This information is
written to standard error, unless <samp><span class="option">-fdump-rtl-sched1</span></samp> or
<samp><span class="option">-fdump-rtl-sched2</span></samp> is specified, in which case it is output
to the usual dump listing file, <samp><span class="file">.sched1</span></samp> or <samp><span class="file">.sched2</span></samp>
respectively. However for <var>n</var> greater than nine, the output is
always printed to standard error.
<p>For <var>n</var> greater than zero, <samp><span class="option">-fsched-verbose</span></samp> outputs the
same information as <samp><span class="option">-fdump-rtl-sched1</span></samp> and <samp><span class="option">-fdump-rtl-sched2</span></samp>.
For <var>n</var> greater than one, it also output basic block probabilities,
detailed ready list information and unit/insn info. For <var>n</var> greater
than two, it includes RTL at abort point, control-flow and regions info.
And for <var>n</var> over four, <samp><span class="option">-fsched-verbose</span></samp> also includes
dependence info.
<br><dt><code>-save-temps</code><dt><code>-save-temps=cwd</code><dd><a name="index-save_002dtemps-858"></a>Store the usual “temporary” intermediate files permanently; place them
in the current directory and name them based on the source file. Thus,
compiling <samp><span class="file">foo.c</span></samp> with <samp><span class="option">-c -save-temps</span></samp> produces files
<samp><span class="file">foo.i</span></samp> and <samp><span class="file">foo.s</span></samp>, as well as <samp><span class="file">foo.o</span></samp>. This creates a
preprocessed <samp><span class="file">foo.i</span></samp> output file even though the compiler now
normally uses an integrated preprocessor.
<p>When used in combination with the <samp><span class="option">-x</span></samp> command-line option,
<samp><span class="option">-save-temps</span></samp> is sensible enough to avoid over writing an
input source file with the same extension as an intermediate file.
The corresponding intermediate file may be obtained by renaming the
source file before using <samp><span class="option">-save-temps</span></samp>.
<p>If you invoke GCC in parallel, compiling several different source
files that share a common base name in different subdirectories or the
same source file compiled for multiple output destinations, it is
likely that the different parallel compilers will interfere with each
other, and overwrite the temporary files. For instance:
<pre class="smallexample"> gcc -save-temps -o outdir1/foo.o indir1/foo.c&
gcc -save-temps -o outdir2/foo.o indir2/foo.c&
</pre>
<p>may result in <samp><span class="file">foo.i</span></samp> and <samp><span class="file">foo.o</span></samp> being written to
simultaneously by both compilers.
<br><dt><code>-save-temps=obj</code><dd><a name="index-save_002dtemps_003dobj-859"></a>Store the usual “temporary” intermediate files permanently. If the
<samp><span class="option">-o</span></samp> option is used, the temporary files are based on the
object file. If the <samp><span class="option">-o</span></samp> option is not used, the
<samp><span class="option">-save-temps=obj</span></samp> switch behaves like <samp><span class="option">-save-temps</span></samp>.
<p>For example:
<pre class="smallexample"> gcc -save-temps=obj -c foo.c
gcc -save-temps=obj -c bar.c -o dir/xbar.o
gcc -save-temps=obj foobar.c -o dir2/yfoobar
</pre>
<p class="noindent">creates <samp><span class="file">foo.i</span></samp>, <samp><span class="file">foo.s</span></samp>, <samp><span class="file">dir/xbar.i</span></samp>,
<samp><span class="file">dir/xbar.s</span></samp>, <samp><span class="file">dir2/yfoobar.i</span></samp>, <samp><span class="file">dir2/yfoobar.s</span></samp>, and
<samp><span class="file">dir2/yfoobar.o</span></samp>.
<br><dt><code>-time</code><span class="roman">[</span><code>=</code><var>file</var><span class="roman">]</span><dd><a name="index-time-860"></a>Report the CPU time taken by each subprocess in the compilation
sequence. For C source files, this is the compiler proper and assembler
(plus the linker if linking is done).
<p>Without the specification of an output file, the output looks like this:
<pre class="smallexample"> # cc1 0.12 0.01
# as 0.00 0.01
</pre>
<p>The first number on each line is the “user time”, that is time spent
executing the program itself. The second number is “system time”,
time spent executing operating system routines on behalf of the program.
Both numbers are in seconds.
<p>With the specification of an output file, the output is appended to the
named file, and it looks like this:
<pre class="smallexample"> 0.12 0.01 cc1 <var>options</var>
0.00 0.01 as <var>options</var>
</pre>
<p>The “user time” and the “system time” are moved before the program
name, and the options passed to the program are displayed, so that one
can later tell what file was being compiled, and with which options.
<br><dt><code>-fvar-tracking</code><dd><a name="index-fvar_002dtracking-861"></a>Run variable tracking pass. It computes where variables are stored at each
position in code. Better debugging information is then generated
(if the debugging information format supports this information).
<p>It is enabled by default when compiling with optimization (<samp><span class="option">-Os</span></samp>,
<samp><span class="option">-O</span></samp>, <samp><span class="option">-O2</span></samp>, <small class="dots">...</small>), debugging information (<samp><span class="option">-g</span></samp>) and
the debug info format supports it.
<br><dt><code>-fvar-tracking-assignments</code><dd><a name="index-fvar_002dtracking_002dassignments-862"></a><a name="index-fno_002dvar_002dtracking_002dassignments-863"></a>Annotate assignments to user variables early in the compilation and
attempt to carry the annotations over throughout the compilation all the
way to the end, in an attempt to improve debug information while
optimizing. Use of <samp><span class="option">-gdwarf-4</span></samp> is recommended along with it.
<p>It can be enabled even if var-tracking is disabled, in which case
annotations are created and maintained, but discarded at the end.
By default, this flag is enabled together with <samp><span class="option">-fvar-tracking</span></samp>,
except when selective scheduling is enabled.
<br><dt><code>-fvar-tracking-assignments-toggle</code><dd><a name="index-fvar_002dtracking_002dassignments_002dtoggle-864"></a><a name="index-fno_002dvar_002dtracking_002dassignments_002dtoggle-865"></a>Toggle <samp><span class="option">-fvar-tracking-assignments</span></samp>, in the same way that
<samp><span class="option">-gtoggle</span></samp> toggles <samp><span class="option">-g</span></samp>.
<br><dt><code>-print-file-name=</code><var>library</var><dd><a name="index-print_002dfile_002dname-866"></a>Print the full absolute name of the library file <var>library</var> that
would be used when linking—and don't do anything else. With this
option, GCC does not compile or link anything; it just prints the
file name.
<br><dt><code>-print-multi-directory</code><dd><a name="index-print_002dmulti_002ddirectory-867"></a>Print the directory name corresponding to the multilib selected by any
other switches present in the command line. This directory is supposed
to exist in <samp><span class="env">GCC_EXEC_PREFIX</span></samp>.
<br><dt><code>-print-multi-lib</code><dd><a name="index-print_002dmulti_002dlib-868"></a>Print the mapping from multilib directory names to compiler switches
that enable them. The directory name is separated from the switches by
‘<samp><span class="samp">;</span></samp>’, and each switch starts with an ‘<samp><span class="samp">@</span></samp>’ instead of the
‘<samp><span class="samp">-</span></samp>’, without spaces between multiple switches. This is supposed to
ease shell processing.
<br><dt><code>-print-multi-os-directory</code><dd><a name="index-print_002dmulti_002dos_002ddirectory-869"></a>Print the path to OS libraries for the selected
multilib, relative to some <samp><span class="file">lib</span></samp> subdirectory. If OS libraries are
present in the <samp><span class="file">lib</span></samp> subdirectory and no multilibs are used, this is
usually just <samp><span class="file">.</span></samp>, if OS libraries are present in <samp><span class="file">lib</span><var>suffix</var></samp>
sibling directories this prints e.g. <samp><span class="file">../lib64</span></samp>, <samp><span class="file">../lib</span></samp> or
<samp><span class="file">../lib32</span></samp>, or if OS libraries are present in <samp><span class="file">lib/</span><var>subdir</var></samp>
subdirectories it prints e.g. <samp><span class="file">amd64</span></samp>, <samp><span class="file">sparcv9</span></samp> or <samp><span class="file">ev6</span></samp>.
<br><dt><code>-print-multiarch</code><dd><a name="index-print_002dmultiarch-870"></a>Print the path to OS libraries for the selected multiarch,
relative to some <samp><span class="file">lib</span></samp> subdirectory.
<br><dt><code>-print-prog-name=</code><var>program</var><dd><a name="index-print_002dprog_002dname-871"></a>Like <samp><span class="option">-print-file-name</span></samp>, but searches for a program such as <samp><span class="command">cpp</span></samp>.
<br><dt><code>-print-libgcc-file-name</code><dd><a name="index-print_002dlibgcc_002dfile_002dname-872"></a>Same as <samp><span class="option">-print-file-name=libgcc.a</span></samp>.
<p>This is useful when you use <samp><span class="option">-nostdlib</span></samp> or <samp><span class="option">-nodefaultlibs</span></samp>
but you do want to link with <samp><span class="file">libgcc.a</span></samp>. You can do:
<pre class="smallexample"> gcc -nostdlib <var>files</var>... `gcc -print-libgcc-file-name`
</pre>
<br><dt><code>-print-search-dirs</code><dd><a name="index-print_002dsearch_002ddirs-873"></a>Print the name of the configured installation directory and a list of
program and library directories <samp><span class="command">gcc</span></samp> searches—and don't do anything else.
<p>This is useful when <samp><span class="command">gcc</span></samp> prints the error message
‘<samp><span class="samp">installation problem, cannot exec cpp0: No such file or directory</span></samp>’.
To resolve this you either need to put <samp><span class="file">cpp0</span></samp> and the other compiler
components where <samp><span class="command">gcc</span></samp> expects to find them, or you can set the environment
variable <samp><span class="env">GCC_EXEC_PREFIX</span></samp> to the directory where you installed them.
Don't forget the trailing ‘<samp><span class="samp">/</span></samp>’.
See <a href="Environment-Variables.html#Environment-Variables">Environment Variables</a>.
<br><dt><code>-print-sysroot</code><dd><a name="index-print_002dsysroot-874"></a>Print the target sysroot directory that is used during
compilation. This is the target sysroot specified either at configure
time or using the <samp><span class="option">--sysroot</span></samp> option, possibly with an extra
suffix that depends on compilation options. If no target sysroot is
specified, the option prints nothing.
<br><dt><code>-print-sysroot-headers-suffix</code><dd><a name="index-print_002dsysroot_002dheaders_002dsuffix-875"></a>Print the suffix added to the target sysroot when searching for
headers, or give an error if the compiler is not configured with such
a suffix—and don't do anything else.
<br><dt><code>-dumpmachine</code><dd><a name="index-dumpmachine-876"></a>Print the compiler's target machine (for example,
‘<samp><span class="samp">i686-pc-linux-gnu</span></samp>’)—and don't do anything else.
<br><dt><code>-dumpversion</code><dd><a name="index-dumpversion-877"></a>Print the compiler version (for example, <code>3.0</code>)—and don't do
anything else.
<br><dt><code>-dumpspecs</code><dd><a name="index-dumpspecs-878"></a>Print the compiler's built-in specs—and don't do anything else. (This
is used when GCC itself is being built.) See <a href="Spec-Files.html#Spec-Files">Spec Files</a>.
<br><dt><code>-fno-eliminate-unused-debug-types</code><dd><a name="index-feliminate_002dunused_002ddebug_002dtypes-879"></a><a name="index-fno_002deliminate_002dunused_002ddebug_002dtypes-880"></a>Normally, when producing DWARF 2 output, GCC avoids producing debug symbol
output for types that are nowhere used in the source file being compiled.
Sometimes it is useful to have GCC emit debugging
information for all types declared in a compilation
unit, regardless of whether or not they are actually used
in that compilation unit, for example
if, in the debugger, you want to cast a value to a type that is
not actually used in your program (but is declared). More often,
however, this results in a significant amount of wasted space.
</dl>
</body></html>