aarch64-linux-gnu-4.9/share/doc/gcc/Debugging-Options.html - toolchains/linux-x86/gcc - Git at Google

 <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
 <html>
 <!-- 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. -->
 <!-- Created by GNU Texinfo 5.2, http://www.gnu.org/software/texinfo/ -->
 <head>
 <title>Using the GNU Compiler Collection (GCC): Debugging Options</title>

 <meta name="description" content="Using the GNU Compiler Collection (GCC): Debugging Options">
 <meta name="keywords" content="Using the GNU Compiler Collection (GCC): Debugging Options">
 <meta name="resource-type" content="document">
 <meta name="distribution" content="global">
 <meta name="Generator" content="makeinfo">
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
 <link href="index.html#Top" rel="start" title="Top">
 <link href="Option-Index.html#Option-Index" rel="index" title="Option Index">
 <link href="index.html#SEC_Contents" rel="contents" title="Table of Contents">
 <link href="Invoking-GCC.html#Invoking-GCC" rel="up" title="Invoking GCC">
 <link href="Optimize-Options.html#Optimize-Options" rel="next" title="Optimize Options">
 <link href="Warning-Options.html#Warning-Options" rel="prev" title="Warning Options">
 <style type="text/css">
 <!--
 a.summary-letter {text-decoration: none}
 blockquote.smallquotation {font-size: smaller}
 div.display {margin-left: 3.2em}
 div.example {margin-left: 3.2em}
 div.indentedblock {margin-left: 3.2em}
 div.lisp {margin-left: 3.2em}
 div.smalldisplay {margin-left: 3.2em}
 div.smallexample {margin-left: 3.2em}
 div.smallindentedblock {margin-left: 3.2em; font-size: smaller}
 div.smalllisp {margin-left: 3.2em}
 kbd {font-style:oblique}
 pre.display {font-family: inherit}
 pre.format {font-family: inherit}
 pre.menu-comment {font-family: serif}
 pre.menu-preformatted {font-family: serif}
 pre.smalldisplay {font-family: inherit; font-size: smaller}
 pre.smallexample {font-size: smaller}
 pre.smallformat {font-family: inherit; font-size: smaller}
 pre.smalllisp {font-size: smaller}
 span.nocodebreak {white-space:nowrap}
 span.nolinebreak {white-space:nowrap}
 span.roman {font-family:serif; font-weight:normal}
 span.sansserif {font-family:sans-serif; font-weight:normal}
 ul.no-bullet {list-style: none}
 -->
 </style>


 </head>

 <body lang="en" bgcolor="#FFFFFF" text="#000000" link="#0000FF" vlink="#800080" alink="#FF0000">
 <a name="Debugging-Options"></a>
 <div class="header">
 <p>
 Next: <a href="Optimize-Options.html#Optimize-Options" accesskey="n" rel="next">Optimize Options</a>, Previous: <a href="Warning-Options.html#Warning-Options" accesskey="p" rel="prev">Warning Options</a>, Up: <a href="Invoking-GCC.html#Invoking-GCC" accesskey="u" rel="up">Invoking GCC</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Option-Index.html#Option-Index" title="Index" rel="index">Index</a>]</p>
 </div>
 <hr>
 <a name="Options-for-Debugging-Your-Program-or-GCC"></a>
 <h3 class="section">3.9 Options for Debugging Your Program or GCC</h3>
 <a name="index-options_002c-debugging"></a>
 <a name="index-debugging-information-options"></a>

 <p>GCC has various special options that are used for debugging
 either your program or GCC:
 </p>
 <dl compact="compact">
 <dt><code>-g</code></dt>
 <dd><a name="index-g"></a>
 <p>Produce debugging information in the operating system&rsquo;s native format
 (stabs, COFF, XCOFF, or DWARF 2).  GDB can work with this debugging
 information.
 </p>
 <p>On most systems that use stabs format, <samp>-g</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>-gstabs+</samp>, <samp>-gstabs</samp>,
 <samp>-gxcoff+</samp>, <samp>-gxcoff</samp>, or <samp>-gvms</samp> (see below).
 </p>
 <p>GCC allows you to use <samp>-g</samp> with
 <samp>-O</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>
 <p>Nevertheless it proves possible to debug optimized output.  This makes
 it reasonable to use the optimizer for programs that might have bugs.
 </p>
 <p>The following options are useful when GCC is generated with the
 capability for more than one debugging format.
 </p>
 </dd>
 <dt><code>-gsplit-dwarf</code></dt>
 <dd><a name="index-gsplit_002ddwarf"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-ggdb</code></dt>
 <dd><a name="index-ggdb"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-gpubnames</code></dt>
 <dd><a name="index-gpubnames"></a>
 <p>Generate dwarf .debug_pubnames and .debug_pubtypes sections.
 </p>
 </dd>
 <dt><code>-ggnu-pubnames</code></dt>
 <dd><a name="index-ggnu_002dpubnames"></a>
 <p>Generate .debug_pubnames and .debug_pubtypes sections in a format
 suitable for conversion into a GDB&nbsp;index.  This option is only useful
 with a linker that can produce GDB&nbsp;index version 7.
 </p>
 </dd>
 <dt><code>-gstabs</code></dt>
 <dd><a name="index-gstabs"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-feliminate-unused-debug-symbols</code></dt>
 <dd><a name="index-feliminate_002dunused_002ddebug_002dsymbols"></a>
 <p>Produce debugging information in stabs format (if that is supported),
 for only symbols that are actually used.
 </p>
 </dd>
 <dt><code>-femit-class-debug-always</code></dt>
 <dd><p>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.
 </p>
 </dd>
 <dt><code>-fdebug-types-section</code></dt>
 <dd><a name="index-fdebug_002dtypes_002dsection"></a>
 <a name="index-fno_002ddebug_002dtypes_002dsection"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-gstabs+</code></dt>
 <dd><a name="index-gstabs_002b"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-gcoff</code></dt>
 <dd><a name="index-gcoff"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-gxcoff</code></dt>
 <dd><a name="index-gxcoff"></a>
 <p>Produce debugging information in XCOFF format (if that is supported).
 This is the format used by the DBX debugger on IBM RS/6000 systems.
 </p>
 </dd>
 <dt><code>-gxcoff+</code></dt>
 <dd><a name="index-gxcoff_002b"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-gdwarf-<var>version</var></code></dt>
 <dd><a name="index-gdwarf_002dversion"></a>
 <p>Produce debugging information in DWARF format (if that is supported).
 The value of <var>version</var> may be either 2, 3 or 4; the default version
 for most targets is 4.
 </p>
 <p>Note that with DWARF Version 2, some ports require and always
 use some non-conflicting DWARF 3 extensions in the unwind tables.
 </p>
 <p>Version 4 may require GDB 7.0 and <samp>-fvar-tracking-assignments</samp>
 for maximum benefit.
 </p>
 </dd>
 <dt><code>-grecord-gcc-switches</code></dt>
 <dd><a name="index-grecord_002dgcc_002dswitches"></a>
 <p>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>-frecord-gcc-switches</samp> for another
 way of storing compiler options into the object file.  This is the default.
 </p>
 </dd>
 <dt><code>-gno-record-gcc-switches</code></dt>
 <dd><a name="index-gno_002drecord_002dgcc_002dswitches"></a>
 <p>Disallow appending command-line options to the DW_AT_producer attribute
 in DWARF debugging information.
 </p>
 </dd>
 <dt><code>-gstrict-dwarf</code></dt>
 <dd><a name="index-gstrict_002ddwarf"></a>
 <p>Disallow using extensions of later DWARF standard version than selected
 with <samp>-gdwarf-<var>version</var></samp>.  On most targets using non-conflicting
 DWARF extensions from later standard versions is allowed.
 </p>
 </dd>
 <dt><code>-gno-strict-dwarf</code></dt>
 <dd><a name="index-gno_002dstrict_002ddwarf"></a>
 <p>Allow using extensions of later DWARF standard version than selected with
 <samp>-gdwarf-<var>version</var></samp>.
 </p>
 </dd>
 <dt><code>-gvms</code></dt>
 <dd><a name="index-gvms"></a>
 <p>Produce debugging information in Alpha/VMS debug format (if that is
 supported).  This is the format used by DEBUG on Alpha/VMS systems.
 </p>
 </dd>
 <dt><code>-g<var>level</var></code></dt>
 <dt><code>-ggdb<var>level</var></code></dt>
 <dt><code>-gstabs<var>level</var></code></dt>
 <dt><code>-gcoff<var>level</var></code></dt>
 <dt><code>-gxcoff<var>level</var></code></dt>
 <dt><code>-gvms<var>level</var></code></dt>
 <dd><p>Request debugging information and also use <var>level</var> to specify how
 much information.  The default level is 2.
 </p>
 <p>Level 0 produces no debug information at all.  Thus, <samp>-g0</samp> negates
 <samp>-g</samp>.
 </p>
 <p>Level 1 produces minimal information, enough for making backtraces in
 parts of the program that you don&rsquo;t plan to debug.  This includes
 descriptions of functions and external variables, and line number
 tables, but no information about local variables.
 </p>
 <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>-g3</samp>.
 </p>
 <p><samp>-gdwarf-2</samp> does not accept a concatenated debug level, because
 GCC used to support an option <samp>-gdwarf</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>-g<var>level</var></samp> option to change the
 debug level for DWARF.
 </p>
 </dd>
 <dt><code>-gtoggle</code></dt>
 <dd><a name="index-gtoggle"></a>
 <p>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>-fcompare-debug</samp>.
 </p>
 </dd>
 <dt><code>-fsanitize=address</code></dt>
 <dd><a name="index-fsanitize_003daddress"></a>
 <p>Enable AddressSanitizer, a fast memory error detector.
 Memory access instructions will be 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
 <code>ASAN_OPTIONS</code> 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.
 </p>
 </dd>
 <dt><code>-fsanitize=kernel-address</code></dt>
 <dd><a name="index-fsanitize_003dkernel_002daddress"></a>
 <p>Enable AddressSanitizer for Linux kernel.
 See <a href="https://github.com/google/sanitizers/wiki#threadsanitizer">https://github.com/google/sanitizers/wiki#threadsanitizer</a> for more details.
 </p>
 </dd>
 <dt><code>-fsanitize=thread</code></dt>
 <dd><a name="index-fsanitize_003dthread"></a>
 <p>Enable ThreadSanitizer, a fast data race detector.
 Memory access instructions will be instrumented to detect
 data race bugs.  See <a href="https://github.com/google/sanitizers/wiki/ThreadSanitizerFlags">https://github.com/google/sanitizers/wiki/ThreadSanitizerFlags</a> for more
 details. The run-time behavior can be influenced using the <code>TSAN_OPTIONS</code>
 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.
 </p>
 </dd>
 <dt><code>-fsanitize=leak</code></dt>
 <dd><a name="index-fsanitize_003dleak"></a>
 <p>Enable LeakSanitizer, a memory leak detector.
 This option only matters for linking of executables and if neither
 <samp>-fsanitize=address</samp> nor <samp>-fsanitize=thread</samp> is used.  In that
 case it will link the executable 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
 <code>LSAN_OPTIONS</code> environment variable.
 </p>
 </dd>
 <dt><code>-fsanitize=undefined</code></dt>
 <dd><a name="index-fsanitize_003dundefined"></a>
 <p>Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector.
 Various computations will be instrumented to detect undefined behavior
 at runtime.  Current suboptions are:
 </p>
 <dl compact="compact">
 <dt><code>-fsanitize=shift</code></dt>
 <dd><a name="index-fsanitize_003dshift"></a>

 <p>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.
 </p>
 </dd>
 <dt><code>-fsanitize=integer-divide-by-zero</code></dt>
 <dd><a name="index-fsanitize_003dinteger_002ddivide_002dby_002dzero"></a>

 <p>Detect integer division by zero as well as <code>INT_MIN / -1</code> division.
 </p>
 </dd>
 <dt><code>-fsanitize=unreachable</code></dt>
 <dd><a name="index-fsanitize_003dunreachable"></a>

 <p>With this option, the compiler will turn the <code>__builtin_unreachable</code>
 call into a diagnostics message call instead.  When reaching the
 <code>__builtin_unreachable</code> call, the behavior is undefined.
 </p>
 </dd>
 <dt><code>-fsanitize=vla-bound</code></dt>
 <dd><a name="index-fsanitize_003dvla_002dbound"></a>

 <p>This option instructs the compiler to check that the size of a variable
 length array is positive.  This option does not have any effect in
 <samp>-std=c++1y</samp> mode, as the standard requires the exception be thrown
 instead.
 </p>
 </dd>
 <dt><code>-fsanitize=null</code></dt>
 <dd><a name="index-fsanitize_003dnull"></a>

 <p>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.
 </p>
 </dd>
 <dt><code>-fsanitize=return</code></dt>
 <dd><a name="index-fsanitize_003dreturn"></a>

 <p>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.
 </p>
 </dd>
 <dt><code>-fsanitize=signed-integer-overflow</code></dt>
 <dd><a name="index-fsanitize_003dsigned_002dinteger_002doverflow"></a>

 <p>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:
 </p><div class="smallexample">
 <pre class="smallexample">signed char a = SCHAR_MAX;
 a++;
 </pre></div>

 </dd>
 </dl>

 <p>While <samp>-ftrapv</samp> causes traps for signed overflows to be emitted,
 <samp>-fsanitize=undefined</samp> gives a diagnostic message.
 This currently works only for the C family of languages.
 </p>
 </dd>
 <dt><code>-fdump-final-insns<span class="roman">[</span>=<var>file</var><span class="roman">]</span></code></dt>
 <dd><a name="index-fdump_002dfinal_002dinsns"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-fcompare-debug<span class="roman">[</span>=<var>opts</var><span class="roman">]</span></code></dt>
 <dd><a name="index-fcompare_002ddebug"></a>
 <a name="index-fno_002dcompare_002ddebug"></a>
 <p>If no error occurs during compilation, run the compiler a second time,
 adding <var>opts</var> and <samp>-fcompare-debug-second</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>
 <p>If the equal sign is omitted, the default <samp>-gtoggle</samp> is used.
 </p>
 <p>The environment variable <code>GCC_COMPARE_DEBUG</code>, if defined, non-empty
 and nonzero, implicitly enables <samp>-fcompare-debug</samp>.  If
 <code>GCC_COMPARE_DEBUG</code> is defined to a string starting with a dash,
 then it is used for <var>opts</var>, otherwise the default <samp>-gtoggle</samp>
 is used.
 </p>
 <p><samp>-fcompare-debug=</samp>, with the equal sign but without <var>opts</var>,
 is equivalent to <samp>-fno-compare-debug</samp>, which disables the dumping
 of the final representation and the second compilation, preventing even
 <code>GCC_COMPARE_DEBUG</code> from taking effect.
 </p>
 <p>To verify full coverage during <samp>-fcompare-debug</samp> testing, set
 <code>GCC_COMPARE_DEBUG</code> to say &lsquo;<samp>-fcompare-debug-not-overridden</samp>&rsquo;,
 which GCC rejects as an invalid option in any actual compilation
 (rather than preprocessing, assembly or linking).  To get just a
 warning, setting <code>GCC_COMPARE_DEBUG</code> to &lsquo;<samp>-w%n-fcompare-debug
 not overridden</samp>&rsquo; will do.
 </p>
 </dd>
 <dt><code>-fcompare-debug-second</code></dt>
 <dd><a name="index-fcompare_002ddebug_002dsecond"></a>
 <p>This option is implicitly passed to the compiler for the second
 compilation requested by <samp>-fcompare-debug</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>
 <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.
 </p>
 </dd>
 <dt><code>-feliminate-dwarf2-dups</code></dt>
 <dd><a name="index-feliminate_002ddwarf2_002ddups"></a>
 <p>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>-gdwarf-2</samp>.
 </p>
 </dd>
 <dt><code>-femit-struct-debug-baseonly</code></dt>
 <dd><a name="index-femit_002dstruct_002ddebug_002dbaseonly"></a>
 <p>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>
 <p>This option substantially reduces the size of debugging information,
 but at significant potential loss in type information to the debugger.
 See <samp>-femit-struct-debug-reduced</samp> for a less aggressive option.
 See <samp>-femit-struct-debug-detailed</samp> for more detailed control.
 </p>
 <p>This option works only with DWARF 2.
 </p>
 </dd>
 <dt><code>-femit-struct-debug-reduced</code></dt>
 <dd><a name="index-femit_002dstruct_002ddebug_002dreduced"></a>
 <p>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>
 <p>This option significantly reduces the size of debugging information,
 with some potential loss in type information to the debugger.
 See <samp>-femit-struct-debug-baseonly</samp> for a more aggressive option.
 See <samp>-femit-struct-debug-detailed</samp> for more detailed control.
 </p>
 <p>This option works only with DWARF 2.
 </p>
 </dd>
 <dt><code>-femit-struct-debug-detailed<span class="roman">[</span>=<var>spec-list</var><span class="roman">]</span></code></dt>
 <dd><p>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>
 <p>This option is a detailed version of
 <samp>-femit-struct-debug-reduced</samp> and <samp>-femit-struct-debug-baseonly</samp>,
 which serves for most needs.
 </p>
 <p>A specification has the syntax<br>
 [&lsquo;<samp>dir:</samp>&rsquo;|&lsquo;<samp>ind:</samp>&rsquo;][&lsquo;<samp>ord:</samp>&rsquo;|&lsquo;<samp>gen:</samp>&rsquo;](&lsquo;<samp>any</samp>&rsquo;|&lsquo;<samp>sys</samp>&rsquo;|&lsquo;<samp>base</samp>&rsquo;|&lsquo;<samp>none</samp>&rsquo;)
 </p>
 <p>The optional first word limits the specification to
 structs that are used directly (&lsquo;<samp>dir:</samp>&rsquo;) or used indirectly (&lsquo;<samp>ind:</samp>&rsquo;).
 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
 &lsquo;<samp>struct one direct; struct two * indirect;</samp>&rsquo;.
 </p>
 <p>The optional second word limits the specification to
 ordinary structs (&lsquo;<samp>ord:</samp>&rsquo;) or generic structs (&lsquo;<samp>gen:</samp>&rsquo;).
 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>-femit-struct-debug-detailed</samp> does not yet implement them.
 </p>
 <p>The third word specifies the source files for those
 structs for which the compiler should emit debug information.
 The values &lsquo;<samp>none</samp>&rsquo; and &lsquo;<samp>any</samp>&rsquo; have the normal meaning.
 The value &lsquo;<samp>base</samp>&rsquo; 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>foo.c</samp>, debug information
 is generated for types declared in that file and <samp>foo.h</samp>,
 but not other header files.
 The value &lsquo;<samp>sys</samp>&rsquo; means those types satisfying &lsquo;<samp>base</samp>&rsquo;
 or declared in system or compiler headers.
 </p>
 <p>You may need to experiment to determine the best settings for your application.
 </p>
 <p>The default is <samp>-femit-struct-debug-detailed=all</samp>.
 </p>
 <p>This option works only with DWARF 2.
 </p>
 </dd>
 <dt><code>-fno-merge-debug-strings</code></dt>
 <dd><a name="index-fmerge_002ddebug_002dstrings"></a>
 <a name="index-fno_002dmerge_002ddebug_002dstrings"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-fdebug-prefix-map=<var>old</var>=<var>new</var></code></dt>
 <dd><a name="index-fdebug_002dprefix_002dmap"></a>
 <p>When compiling files in directory <samp><var>old</var></samp>, record debugging
 information describing them as in <samp><var>new</var></samp> instead.
 </p>
 </dd>
 <dt><code>-fno-dwarf2-cfi-asm</code></dt>
 <dd><a name="index-fdwarf2_002dcfi_002dasm"></a>
 <a name="index-fno_002ddwarf2_002dcfi_002dasm"></a>
 <p>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-prof"></a>
 </dd>
 <dt><code>-p</code></dt>
 <dd><a name="index-p"></a>
 <p>Generate extra code to write profile information suitable for the
 analysis program <code>prof</code>.  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-gprof"></a>
 </dd>
 <dt><code>-pg</code></dt>
 <dd><a name="index-pg"></a>
 <p>Generate extra code to write profile information suitable for the
 analysis program <code>gprof</code>.  You must use this option when compiling
 the source files you want data about, and you must also use it when
 linking.
 </p>
 </dd>
 <dt><code>-Q</code></dt>
 <dd><a name="index-Q"></a>
 <p>Makes the compiler print out each function name as it is compiled, and
 print some statistics about each pass when it finishes.
 </p>
 </dd>
 <dt><code>-ftime-report</code></dt>
 <dd><a name="index-ftime_002dreport"></a>
 <p>Makes the compiler print some statistics about the time consumed by each
 pass when it finishes.
 </p>
 </dd>
 <dt><code>-fmem-report</code></dt>
 <dd><a name="index-fmem_002dreport"></a>
 <p>Makes the compiler print some statistics about permanent memory
 allocation when it finishes.
 </p>
 </dd>
 <dt><code>-fmem-report-wpa</code></dt>
 <dd><a name="index-fmem_002dreport_002dwpa"></a>
 <p>Makes the compiler print some statistics about permanent memory
 allocation for the WPA phase only.
 </p>
 </dd>
 <dt><code>-fpre-ipa-mem-report</code></dt>
 <dd><a name="index-fpre_002dipa_002dmem_002dreport"></a>
 </dd>
 <dt><code>-fpost-ipa-mem-report</code></dt>
 <dd><a name="index-fpost_002dipa_002dmem_002dreport"></a>
 <p>Makes the compiler print some statistics about permanent memory
 allocation before or after interprocedural optimization.
 </p>
 </dd>
 <dt><code>-fprofile-report</code></dt>
 <dd><a name="index-fprofile_002dreport"></a>
 <p>Makes the compiler print some statistics about consistency of the
 (estimated) profile and effect of individual passes.
 </p>
 </dd>
 <dt><code>-fstack-usage</code></dt>
 <dd><a name="index-fstack_002dusage"></a>
 <p>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>.su</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:
 </p>
 <ul>
 <li> The name of the function.
 </li><li> A number of bytes.
 </li><li> One or more qualifiers: <code>static</code>, <code>dynamic</code>, <code>bounded</code>.
 </li></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>
 <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.
 </p>
 </dd>
 <dt><code>-fprofile-arcs</code></dt>
 <dd><a name="index-fprofile_002darcs"></a>
 <p>Add code so that program flow <em>arcs</em> 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>.gcda</samp> for each source file.  The data may be used for
 profile-directed optimizations (<samp>-fbranch-probabilities</samp>), or for
 test coverage analysis (<samp>-ftest-coverage</samp>).  Each object file&rsquo;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>foo.gcda</samp> for input file <samp>dir/foo.c</samp>, or
 <samp>dir/foo.gcda</samp> for output file specified as <samp>-o dir/foo.o</samp>).
 See <a href="Cross_002dprofiling.html#Cross_002dprofiling">Cross-profiling</a>.
 </p>
 <a name="index-gcov"></a>
 </dd>
 <dt><code>--coverage</code></dt>
 <dd><a name="index-coverage"></a>

 <p>This option is used to compile and link code instrumented for coverage
 analysis.  The option is a synonym for <samp>-fprofile-arcs</samp>
 <samp>-ftest-coverage</samp> (when compiling) and <samp>-lgcov</samp> (when
 linking).  See the documentation for those options for more details.
 </p>
 <ul>
 <li> Compile the source files with <samp>-fprofile-arcs</samp> plus optimization
 and code generation options.  For test coverage analysis, use the
 additional <samp>-ftest-coverage</samp> option.  You do not need to profile
 every source file in a program.

 </li><li> Link your object files with <samp>-lgcov</samp> or <samp>-fprofile-arcs</samp>
 (the latter implies the former).

 </li><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><li> For profile-directed optimizations, compile the source files again with
 the same optimization and code generation options plus
 <samp>-fbranch-probabilities</samp> (see <a href="Optimize-Options.html#Optimize-Options">Options that
 Control Optimization</a>).

 </li><li> For test coverage analysis, use <code>gcov</code> to produce human readable
 information from the <samp>.gcno</samp> and <samp>.gcda</samp> files.  Refer to the
 <code>gcov</code> documentation for further information.

 </li></ul>

 <p>With <samp>-fprofile-arcs</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.
 </p>
 </dd>
 <dt><code>-ftest-coverage</code></dt>
 <dd><a name="index-ftest_002dcoverage"></a>
 <p>Produce a notes file that the <code>gcov</code> code-coverage utility
 (see <a href="Gcov.html#Gcov"><code>gcov</code>&mdash;a Test Coverage Program</a>) can use to
 show program coverage.  Each source file&rsquo;s note file is called
 <samp><var>auxname</var>.gcno</samp>.  Refer to the <samp>-fprofile-arcs</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.
 </p>
 </dd>
 <dt><code>-fdbg-cnt-list</code></dt>
 <dd><a name="index-fdbg_002dcnt_002dlist"></a>
 <p>Print the name and the counter upper bound for all debug counters.
 </p>

 </dd>
 <dt><code>-fdbg-cnt=<var>counter-value-list</var></code></dt>
 <dd><a name="index-fdbg_002dcnt"></a>
 <p>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>-fdbg-cnt=dce:10,tail_call:0</samp>,
 <code>dbg_cnt(dce)</code> returns true only for first 10 invocations.
 </p>
 </dd>
 <dt><code>-fenable-<var>kind</var>-<var>pass</var></code></dt>
 <dt><code>-fdisable-<var>kind</var>-<var>pass</var>=<var>range-list</var></code></dt>
 <dd><a name="index-fdisable_002d"></a>
 <a name="index-fenable_002d"></a>

 <p>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.
 </p>
 <dl compact="compact">
 <dt><code>-fdisable-ipa-<var>pass</var></code></dt>
 <dd><p>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.
 </p>
 </dd>
 <dt><code>-fdisable-rtl-<var>pass</var></code></dt>
 <dt><code>-fdisable-rtl-<var>pass</var>=<var>range-list</var></code></dt>
 <dd><p>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&rsquo;s call graph node&rsquo;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>-fdump-passes</samp>.
 </p>
 </dd>
 <dt><code>-fdisable-tree-<var>pass</var></code></dt>
 <dt><code>-fdisable-tree-<var>pass</var>=<var>range-list</var></code></dt>
 <dd><p>Disable tree pass <var>pass</var>.  See <samp>-fdisable-rtl</samp> for the description of
 option arguments.
 </p>
 </dd>
 <dt><code>-fenable-ipa-<var>pass</var></code></dt>
 <dd><p>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.
 </p>
 </dd>
 <dt><code>-fenable-rtl-<var>pass</var></code></dt>
 <dt><code>-fenable-rtl-<var>pass</var>=<var>range-list</var></code></dt>
 <dd><p>Enable RTL pass <var>pass</var>.  See <samp>-fdisable-rtl</samp> for option argument
 description and examples.
 </p>
 </dd>
 <dt><code>-fenable-tree-<var>pass</var></code></dt>
 <dt><code>-fenable-tree-<var>pass</var>=<var>range-list</var></code></dt>
 <dd><p>Enable tree pass <var>pass</var>.  See <samp>-fdisable-rtl</samp> for the description
 of option arguments.
 </p>
 </dd>
 </dl>

 <p>Here are some examples showing uses of these options.
 </p>
 <div class="smallexample">
 <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></div>

 </dd>
 <dt><code>-d<var>letters</var></code></dt>
 <dt><code>-fdump-rtl-<var>pass</var></code></dt>
 <dt><code>-fdump-rtl-<var>pass</var>=<var>filename</var></code></dt>
 <dd><a name="index-d"></a>
 <a name="index-fdump_002drtl_002dpass"></a>
 <p>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>=<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>-E</samp> is used for preprocessing.
 </p>
 <p>Debug dumps can be enabled with a <samp>-fdump-rtl</samp> switch or some
 <samp>-d</samp> option <var>letters</var>.  Here are the possible
 letters for use in <var>pass</var> and <var>letters</var>, and their meanings:
 </p>
 <dl compact="compact">
 <dt><code>-fdump-rtl-alignments</code></dt>
 <dd><a name="index-fdump_002drtl_002dalignments"></a>
 <p>Dump after branch alignments have been computed.
 </p>
 </dd>
 <dt><code>-fdump-rtl-asmcons</code></dt>
 <dd><a name="index-fdump_002drtl_002dasmcons"></a>
 <p>Dump after fixing rtl statements that have unsatisfied in/out constraints.
 </p>
 </dd>
 <dt><code>-fdump-rtl-auto_inc_dec</code></dt>
 <dd><a name="index-fdump_002drtl_002dauto_005finc_005fdec"></a>
 <p>Dump after auto-inc-dec discovery.  This pass is only run on
 architectures that have auto inc or auto dec instructions.
 </p>
 </dd>
 <dt><code>-fdump-rtl-barriers</code></dt>
 <dd><a name="index-fdump_002drtl_002dbarriers"></a>
 <p>Dump after cleaning up the barrier instructions.
 </p>
 </dd>
 <dt><code>-fdump-rtl-bbpart</code></dt>
 <dd><a name="index-fdump_002drtl_002dbbpart"></a>
 <p>Dump after partitioning hot and cold basic blocks.
 </p>
 </dd>
 <dt><code>-fdump-rtl-bbro</code></dt>
 <dd><a name="index-fdump_002drtl_002dbbro"></a>
 <p>Dump after block reordering.
 </p>
 </dd>
 <dt><code>-fdump-rtl-btl1</code></dt>
 <dt><code>-fdump-rtl-btl2</code></dt>
 <dd><a name="index-fdump_002drtl_002dbtl2"></a>
 <a name="index-fdump_002drtl_002dbtl2-1"></a>
 <p><samp>-fdump-rtl-btl1</samp> and <samp>-fdump-rtl-btl2</samp> enable dumping
 after the two branch
 target load optimization passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-bypass</code></dt>
 <dd><a name="index-fdump_002drtl_002dbypass"></a>
 <p>Dump after jump bypassing and control flow optimizations.
 </p>
 </dd>
 <dt><code>-fdump-rtl-combine</code></dt>
 <dd><a name="index-fdump_002drtl_002dcombine"></a>
 <p>Dump after the RTL instruction combination pass.
 </p>
 </dd>
 <dt><code>-fdump-rtl-compgotos</code></dt>
 <dd><a name="index-fdump_002drtl_002dcompgotos"></a>
 <p>Dump after duplicating the computed gotos.
 </p>
 </dd>
 <dt><code>-fdump-rtl-ce1</code></dt>
 <dt><code>-fdump-rtl-ce2</code></dt>
 <dt><code>-fdump-rtl-ce3</code></dt>
 <dd><a name="index-fdump_002drtl_002dce1"></a>
 <a name="index-fdump_002drtl_002dce2"></a>
 <a name="index-fdump_002drtl_002dce3"></a>
 <p><samp>-fdump-rtl-ce1</samp>, <samp>-fdump-rtl-ce2</samp>, and
 <samp>-fdump-rtl-ce3</samp> enable dumping after the three
 if conversion passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-cprop_hardreg</code></dt>
 <dd><a name="index-fdump_002drtl_002dcprop_005fhardreg"></a>
 <p>Dump after hard register copy propagation.
 </p>
 </dd>
 <dt><code>-fdump-rtl-csa</code></dt>
 <dd><a name="index-fdump_002drtl_002dcsa"></a>
 <p>Dump after combining stack adjustments.
 </p>
 </dd>
 <dt><code>-fdump-rtl-cse1</code></dt>
 <dt><code>-fdump-rtl-cse2</code></dt>
 <dd><a name="index-fdump_002drtl_002dcse1"></a>
 <a name="index-fdump_002drtl_002dcse2"></a>
 <p><samp>-fdump-rtl-cse1</samp> and <samp>-fdump-rtl-cse2</samp> enable dumping after
 the two common subexpression elimination passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-dce</code></dt>
 <dd><a name="index-fdump_002drtl_002ddce"></a>
 <p>Dump after the standalone dead code elimination passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-dbr</code></dt>
 <dd><a name="index-fdump_002drtl_002ddbr"></a>
 <p>Dump after delayed branch scheduling.
 </p>
 </dd>
 <dt><code>-fdump-rtl-dce1</code></dt>
 <dt><code>-fdump-rtl-dce2</code></dt>
 <dd><a name="index-fdump_002drtl_002ddce1"></a>
 <a name="index-fdump_002drtl_002ddce2"></a>
 <p><samp>-fdump-rtl-dce1</samp> and <samp>-fdump-rtl-dce2</samp> enable dumping after
 the two dead store elimination passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-eh</code></dt>
 <dd><a name="index-fdump_002drtl_002deh"></a>
 <p>Dump after finalization of EH handling code.
 </p>
 </dd>
 <dt><code>-fdump-rtl-eh_ranges</code></dt>
 <dd><a name="index-fdump_002drtl_002deh_005franges"></a>
 <p>Dump after conversion of EH handling range regions.
 </p>
 </dd>
 <dt><code>-fdump-rtl-expand</code></dt>
 <dd><a name="index-fdump_002drtl_002dexpand"></a>
 <p>Dump after RTL generation.
 </p>
 </dd>
 <dt><code>-fdump-rtl-fwprop1</code></dt>
 <dt><code>-fdump-rtl-fwprop2</code></dt>
 <dd><a name="index-fdump_002drtl_002dfwprop1"></a>
 <a name="index-fdump_002drtl_002dfwprop2"></a>
 <p><samp>-fdump-rtl-fwprop1</samp> and <samp>-fdump-rtl-fwprop2</samp> enable
 dumping after the two forward propagation passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-gcse1</code></dt>
 <dt><code>-fdump-rtl-gcse2</code></dt>
 <dd><a name="index-fdump_002drtl_002dgcse1"></a>
 <a name="index-fdump_002drtl_002dgcse2"></a>
 <p><samp>-fdump-rtl-gcse1</samp> and <samp>-fdump-rtl-gcse2</samp> enable dumping
 after global common subexpression elimination.
 </p>
 </dd>
 <dt><code>-fdump-rtl-init-regs</code></dt>
 <dd><a name="index-fdump_002drtl_002dinit_002dregs"></a>
 <p>Dump after the initialization of the registers.
 </p>
 </dd>
 <dt><code>-fdump-rtl-initvals</code></dt>
 <dd><a name="index-fdump_002drtl_002dinitvals"></a>
 <p>Dump after the computation of the initial value sets.
 </p>
 </dd>
 <dt><code>-fdump-rtl-into_cfglayout</code></dt>
 <dd><a name="index-fdump_002drtl_002dinto_005fcfglayout"></a>
 <p>Dump after converting to cfglayout mode.
 </p>
 </dd>
 <dt><code>-fdump-rtl-ira</code></dt>
 <dd><a name="index-fdump_002drtl_002dira"></a>
 <p>Dump after iterated register allocation.
 </p>
 </dd>
 <dt><code>-fdump-rtl-jump</code></dt>
 <dd><a name="index-fdump_002drtl_002djump"></a>
 <p>Dump after the second jump optimization.
 </p>
 </dd>
 <dt><code>-fdump-rtl-loop2</code></dt>
 <dd><a name="index-fdump_002drtl_002dloop2"></a>
 <p><samp>-fdump-rtl-loop2</samp> enables dumping after the rtl
 loop optimization passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-mach</code></dt>
 <dd><a name="index-fdump_002drtl_002dmach"></a>
 <p>Dump after performing the machine dependent reorganization pass, if that
 pass exists.
 </p>
 </dd>
 <dt><code>-fdump-rtl-mode_sw</code></dt>
 <dd><a name="index-fdump_002drtl_002dmode_005fsw"></a>
 <p>Dump after removing redundant mode switches.
 </p>
 </dd>
 <dt><code>-fdump-rtl-rnreg</code></dt>
 <dd><a name="index-fdump_002drtl_002drnreg"></a>
 <p>Dump after register renumbering.
 </p>
 </dd>
 <dt><code>-fdump-rtl-outof_cfglayout</code></dt>
 <dd><a name="index-fdump_002drtl_002doutof_005fcfglayout"></a>
 <p>Dump after converting from cfglayout mode.
 </p>
 </dd>
 <dt><code>-fdump-rtl-peephole2</code></dt>
 <dd><a name="index-fdump_002drtl_002dpeephole2"></a>
 <p>Dump after the peephole pass.
 </p>
 </dd>
 <dt><code>-fdump-rtl-postreload</code></dt>
 <dd><a name="index-fdump_002drtl_002dpostreload"></a>
 <p>Dump after post-reload optimizations.
 </p>
 </dd>
 <dt><code>-fdump-rtl-pro_and_epilogue</code></dt>
 <dd><a name="index-fdump_002drtl_002dpro_005fand_005fepilogue"></a>
 <p>Dump after generating the function prologues and epilogues.
 </p>
 </dd>
 <dt><code>-fdump-rtl-sched1</code></dt>
 <dt><code>-fdump-rtl-sched2</code></dt>
 <dd><a name="index-fdump_002drtl_002dsched1"></a>
 <a name="index-fdump_002drtl_002dsched2"></a>
 <p><samp>-fdump-rtl-sched1</samp> and <samp>-fdump-rtl-sched2</samp> enable dumping
 after the basic block scheduling passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-ree</code></dt>
 <dd><a name="index-fdump_002drtl_002dree"></a>
 <p>Dump after sign/zero extension elimination.
 </p>
 </dd>
 <dt><code>-fdump-rtl-seqabstr</code></dt>
 <dd><a name="index-fdump_002drtl_002dseqabstr"></a>
 <p>Dump after common sequence discovery.
 </p>
 </dd>
 <dt><code>-fdump-rtl-shorten</code></dt>
 <dd><a name="index-fdump_002drtl_002dshorten"></a>
 <p>Dump after shortening branches.
 </p>
 </dd>
 <dt><code>-fdump-rtl-sibling</code></dt>
 <dd><a name="index-fdump_002drtl_002dsibling"></a>
 <p>Dump after sibling call optimizations.
 </p>
 </dd>
 <dt><code>-fdump-rtl-split1</code></dt>
 <dt><code>-fdump-rtl-split2</code></dt>
 <dt><code>-fdump-rtl-split3</code></dt>
 <dt><code>-fdump-rtl-split4</code></dt>
 <dt><code>-fdump-rtl-split5</code></dt>
 <dd><a name="index-fdump_002drtl_002dsplit1"></a>
 <a name="index-fdump_002drtl_002dsplit2"></a>
 <a name="index-fdump_002drtl_002dsplit3"></a>
 <a name="index-fdump_002drtl_002dsplit4"></a>
 <a name="index-fdump_002drtl_002dsplit5"></a>
 <p><samp>-fdump-rtl-split1</samp>, <samp>-fdump-rtl-split2</samp>,
 <samp>-fdump-rtl-split3</samp>, <samp>-fdump-rtl-split4</samp> and
 <samp>-fdump-rtl-split5</samp> enable dumping after five rounds of
 instruction splitting.
 </p>
 </dd>
 <dt><code>-fdump-rtl-sms</code></dt>
 <dd><a name="index-fdump_002drtl_002dsms"></a>
 <p>Dump after modulo scheduling.  This pass is only run on some
 architectures.
 </p>
 </dd>
 <dt><code>-fdump-rtl-stack</code></dt>
 <dd><a name="index-fdump_002drtl_002dstack"></a>
 <p>Dump after conversion from GCC&rsquo;s &ldquo;flat register file&rdquo; registers to the
 x87&rsquo;s stack-like registers.  This pass is only run on x86 variants.
 </p>
 </dd>
 <dt><code>-fdump-rtl-subreg1</code></dt>
 <dt><code>-fdump-rtl-subreg2</code></dt>
 <dd><a name="index-fdump_002drtl_002dsubreg1"></a>
 <a name="index-fdump_002drtl_002dsubreg2"></a>
 <p><samp>-fdump-rtl-subreg1</samp> and <samp>-fdump-rtl-subreg2</samp> enable dumping after
 the two subreg expansion passes.
 </p>
 </dd>
 <dt><code>-fdump-rtl-unshare</code></dt>
 <dd><a name="index-fdump_002drtl_002dunshare"></a>
 <p>Dump after all rtl has been unshared.
 </p>
 </dd>
 <dt><code>-fdump-rtl-vartrack</code></dt>
 <dd><a name="index-fdump_002drtl_002dvartrack"></a>
 <p>Dump after variable tracking.
 </p>
 </dd>
 <dt><code>-fdump-rtl-vregs</code></dt>
 <dd><a name="index-fdump_002drtl_002dvregs"></a>
 <p>Dump after converting virtual registers to hard registers.
 </p>
 </dd>
 <dt><code>-fdump-rtl-web</code></dt>
 <dd><a name="index-fdump_002drtl_002dweb"></a>
 <p>Dump after live range splitting.
 </p>
 </dd>
 <dt><code>-fdump-rtl-regclass</code></dt>
 <dt><code>-fdump-rtl-subregs_of_mode_init</code></dt>
 <dt><code>-fdump-rtl-subregs_of_mode_finish</code></dt>
 <dt><code>-fdump-rtl-dfinit</code></dt>
 <dt><code>-fdump-rtl-dfinish</code></dt>
 <dd><a name="index-fdump_002drtl_002dregclass"></a>
 <a name="index-fdump_002drtl_002dsubregs_005fof_005fmode_005finit"></a>
 <a name="index-fdump_002drtl_002dsubregs_005fof_005fmode_005ffinish"></a>
 <a name="index-fdump_002drtl_002ddfinit"></a>
 <a name="index-fdump_002drtl_002ddfinish"></a>
 <p>These dumps are defined but always produce empty files.
 </p>
 </dd>
 <dt><code>-da</code></dt>
 <dt><code>-fdump-rtl-all</code></dt>
 <dd><a name="index-da"></a>
 <a name="index-fdump_002drtl_002dall"></a>
 <p>Produce all the dumps listed above.
 </p>
 </dd>
 <dt><code>-dA</code></dt>
 <dd><a name="index-dA"></a>
 <p>Annotate the assembler output with miscellaneous debugging information.
 </p>
 </dd>
 <dt><code>-dD</code></dt>
 <dd><a name="index-dD"></a>
 <p>Dump all macro definitions, at the end of preprocessing, in addition to
 normal output.
 </p>
 </dd>
 <dt><code>-dH</code></dt>
 <dd><a name="index-dH"></a>
 <p>Produce a core dump whenever an error occurs.
 </p>
 </dd>
 <dt><code>-dp</code></dt>
 <dd><a name="index-dp"></a>
 <p>Annotate the assembler output with a comment indicating which
 pattern and alternative is used.  The length of each instruction is
 also printed.
 </p>
 </dd>
 <dt><code>-dP</code></dt>
 <dd><a name="index-dP"></a>
 <p>Dump the RTL in the assembler output as a comment before each instruction.
 Also turns on <samp>-dp</samp> annotation.
 </p>
 </dd>
 <dt><code>-dx</code></dt>
 <dd><a name="index-dx"></a>
 <p>Just generate RTL for a function instead of compiling it.  Usually used
 with <samp>-fdump-rtl-expand</samp>.
 </p></dd>
 </dl>

 </dd>
 <dt><code>-fdump-noaddr</code></dt>
 <dd><a name="index-fdump_002dnoaddr"></a>
 <p>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.
 </p>
 </dd>
 <dt><code>-fdump-unnumbered</code></dt>
 <dd><a name="index-fdump_002dunnumbered"></a>
 <p>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>-g</samp>.
 </p>
 </dd>
 <dt><code>-fdump-unnumbered-links</code></dt>
 <dd><a name="index-fdump_002dunnumbered_002dlinks"></a>
 <p>When doing debugging dumps (see <samp>-d</samp> option above), suppress
 instruction numbers for the links to the previous and next instructions
 in a sequence.
 </p>
 </dd>
 <dt><code>-fdump-translation-unit <span class="roman">(C++ only)</span></code></dt>
 <dt><code>-fdump-translation-unit-<var>options</var> <span class="roman">(C++ only)</span></code></dt>
 <dd><a name="index-fdump_002dtranslation_002dunit"></a>
 <p>Dump a representation of the tree structure for the entire translation
 unit to a file.  The file name is made by appending <samp>.tu</samp> to the
 source file name, and the file is created in the same directory as the
 output file.  If the &lsquo;<samp>-<var>options</var></samp>&rsquo; form is used, <var>options</var>
 controls the details of the dump as described for the
 <samp>-fdump-tree</samp> options.
 </p>
 </dd>
 <dt><code>-fdump-class-hierarchy <span class="roman">(C++ only)</span></code></dt>
 <dt><code>-fdump-class-hierarchy-<var>options</var> <span class="roman">(C++ only)</span></code></dt>
 <dd><a name="index-fdump_002dclass_002dhierarchy"></a>
 <p>Dump a representation of each class&rsquo;s hierarchy and virtual function
 table layout to a file.  The file name is made by appending
 <samp>.class</samp> to the source file name, and the file is created in the
 same directory as the output file.  If the &lsquo;<samp>-<var>options</var></samp>&rsquo; form
 is used, <var>options</var> controls the details of the dump as described
 for the <samp>-fdump-tree</samp> options.
 </p>
 </dd>
 <dt><code>-fdump-ipa-<var>switch</var></code></dt>
 <dd><a name="index-fdump_002dipa"></a>
 <p>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:
 </p>
 <dl compact="compact">
 <dt>&lsquo;<samp>all</samp>&rsquo;</dt>
 <dd><p>Enables all inter-procedural analysis dumps.
 </p>
 </dd>
 <dt>&lsquo;<samp>cgraph</samp>&rsquo;</dt>
 <dd><p>Dumps information about call-graph optimization, unused function removal,
 and inlining decisions.
 </p>
 </dd>
 <dt>&lsquo;<samp>inline</samp>&rsquo;</dt>
 <dd><p>Dump after function inlining.
 </p>
 </dd>
 </dl>

 </dd>
 <dt><code>-fdump-passes</code></dt>
 <dd><a name="index-fdump_002dpasses"></a>
 <p>Dump the list of optimization passes that are turned on and off by
 the current command-line options.
 </p>
 </dd>
 <dt><code>-fdump-statistics-<var>option</var></code></dt>
 <dd><a name="index-fdump_002dstatistics"></a>
 <p>Enable and control dumping of pass statistics in a separate file.  The
 file name is generated by appending a suffix ending in
 &lsquo;<samp>.statistics</samp>&rsquo; to the source file name, and the file is created in
 the same directory as the output file.  If the &lsquo;<samp>-<var>option</var></samp>&rsquo;
 form is used, &lsquo;<samp>-stats</samp>&rsquo; causes counters to be summed over the
 whole compilation unit while &lsquo;<samp>-details</samp>&rsquo; dumps every event as
 the passes generate them.  The default with no option is to sum
 counters for each function compiled.
 </p>
 </dd>
 <dt><code>-fdump-tree-<var>switch</var></code></dt>
 <dt><code>-fdump-tree-<var>switch</var>-<var>options</var></code></dt>
 <dt><code>-fdump-tree-<var>switch</var>-<var>options</var>=<var>filename</var></code></dt>
 <dd><a name="index-fdump_002dtree"></a>
 <p>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>=<var>filename</var></samp> option, the dump is output on the given file
 instead of the auto named dump files.  If the &lsquo;<samp>-<var>options</var></samp>&rsquo;
 form is used, <var>options</var> is a list of &lsquo;<samp>-</samp>&rsquo; 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
 </p>
 <dl compact="compact">
 <dt>&lsquo;<samp>address</samp>&rsquo;</dt>
 <dd><p>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.
 </p></dd>
 <dt>&lsquo;<samp>asmname</samp>&rsquo;</dt>
 <dd><p>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.
 </p></dd>
 <dt>&lsquo;<samp>slim</samp>&rsquo;</dt>
 <dd><p>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>
 <p>When dumping pretty-printed trees, this option inhibits dumping the
 bodies of control structures.
 </p>
 <p>When dumping RTL, print the RTL in slim (condensed) form instead of
 the default LISP-like representation.
 </p></dd>
 <dt>&lsquo;<samp>raw</samp>&rsquo;</dt>
 <dd><p>Print a raw representation of the tree.  By default, trees are
 pretty-printed into a C-like representation.
 </p></dd>
 <dt>&lsquo;<samp>details</samp>&rsquo;</dt>
 <dd><p>Enable more detailed dumps (not honored by every dump option). Also
 include information from the optimization passes.
 </p></dd>
 <dt>&lsquo;<samp>stats</samp>&rsquo;</dt>
 <dd><p>Enable dumping various statistics about the pass (not honored by every dump
 option).
 </p></dd>
 <dt>&lsquo;<samp>blocks</samp>&rsquo;</dt>
 <dd><p>Enable showing basic block boundaries (disabled in raw dumps).
 </p></dd>
 <dt>&lsquo;<samp>graph</samp>&rsquo;</dt>
 <dd><p>For each of the other indicated dump files (<samp>-fdump-rtl-<var>pass</var></samp>),
 dump a representation of the control flow graph suitable for viewing with
 GraphViz to <samp><var>file</var>.<var>passid</var>.<var>pass</var>.dot</samp>.  Each function in
 the file is pretty-printed as a subgraph, so that GraphViz can render them
 all in a single plot.
 </p>
 <p>This option currently only works for RTL dumps, and the RTL is always
 dumped in slim form.
 </p></dd>
 <dt>&lsquo;<samp>vops</samp>&rsquo;</dt>
 <dd><p>Enable showing virtual operands for every statement.
 </p></dd>
 <dt>&lsquo;<samp>lineno</samp>&rsquo;</dt>
 <dd><p>Enable showing line numbers for statements.
 </p></dd>
 <dt>&lsquo;<samp>uid</samp>&rsquo;</dt>
 <dd><p>Enable showing the unique ID (<code>DECL_UID</code>) for each variable.
 </p></dd>
 <dt>&lsquo;<samp>verbose</samp>&rsquo;</dt>
 <dd><p>Enable showing the tree dump for each statement.
 </p></dd>
 <dt>&lsquo;<samp>eh</samp>&rsquo;</dt>
 <dd><p>Enable showing the EH region number holding each statement.
 </p></dd>
 <dt>&lsquo;<samp>scev</samp>&rsquo;</dt>
 <dd><p>Enable showing scalar evolution analysis details.
 </p></dd>
 <dt>&lsquo;<samp>optimized</samp>&rsquo;</dt>
 <dd><p>Enable showing optimization information (only available in certain
 passes).
 </p></dd>
 <dt>&lsquo;<samp>missed</samp>&rsquo;</dt>
 <dd><p>Enable showing missed optimization information (only available in certain
 passes).
 </p></dd>
 <dt>&lsquo;<samp>notes</samp>&rsquo;</dt>
 <dd><p>Enable other detailed optimization information (only available in
 certain passes).
 </p></dd>
 <dt>&lsquo;<samp>=<var>filename</var></samp>&rsquo;</dt>
 <dd><p>Instead of an auto named dump file, output into the given file
 name. The file names <samp>stdout</samp> and <samp>stderr</samp> are treated
 specially and are considered already open standard streams. For
 example,
 </p>
 <div class="smallexample">
 <pre class="smallexample">gcc -O2 -ftree-vectorize -fdump-tree-vect-blocks=foo.dump
      -fdump-tree-pre=stderr file.c
 </pre></div>

 <p>outputs vectorizer dump into <samp>foo.dump</samp>, while the PRE dump is
 output on to <samp>stderr</samp>. If two conflicting dump filenames are
 given for the same pass, then the latter option overrides the earlier
 one.
 </p>
 </dd>
 <dt>&lsquo;<samp>all</samp>&rsquo;</dt>
 <dd><p>Turn on all options, except <samp>raw</samp>, <samp>slim</samp>, <samp>verbose</samp>
 and <samp>lineno</samp>.
 </p>
 </dd>
 <dt>&lsquo;<samp>optall</samp>&rsquo;</dt>
 <dd><p>Turn on all optimization options, i.e., <samp>optimized</samp>,
 <samp>missed</samp>, and <samp>note</samp>.
 </p></dd>
 </dl>

 <p>The following tree dumps are possible:
 </p><dl compact="compact">
 <dt>&lsquo;<samp>original</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002doriginal"></a>
 <p>Dump before any tree based optimization, to <samp><var>file</var>.original</samp>.
 </p>
 </dd>
 <dt>&lsquo;<samp>optimized</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002doptimized"></a>
 <p>Dump after all tree based optimization, to <samp><var>file</var>.optimized</samp>.
 </p>
 </dd>
 <dt>&lsquo;<samp>gimple</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dgimple"></a>
 <p>Dump each function before and after the gimplification pass to a file.  The
 file name is made by appending <samp>.gimple</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>cfg</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dcfg"></a>
 <p>Dump the control flow graph of each function to a file.  The file name is
 made by appending <samp>.cfg</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>ch</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dch"></a>
 <p>Dump each function after copying loop headers.  The file name is made by
 appending <samp>.ch</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>ssa</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dssa"></a>
 <p>Dump SSA related information to a file.  The file name is made by appending
 <samp>.ssa</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>alias</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dalias"></a>
 <p>Dump aliasing information for each function.  The file name is made by
 appending <samp>.alias</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>ccp</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dccp"></a>
 <p>Dump each function after CCP.  The file name is made by appending
 <samp>.ccp</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>storeccp</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dstoreccp"></a>
 <p>Dump each function after STORE-CCP.  The file name is made by appending
 <samp>.storeccp</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>pre</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dpre"></a>
 <p>Dump trees after partial redundancy elimination.  The file name is made
 by appending <samp>.pre</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>fre</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dfre"></a>
 <p>Dump trees after full redundancy elimination.  The file name is made
 by appending <samp>.fre</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>copyprop</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dcopyprop"></a>
 <p>Dump trees after copy propagation.  The file name is made
 by appending <samp>.copyprop</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>store_copyprop</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dstore_005fcopyprop"></a>
 <p>Dump trees after store copy-propagation.  The file name is made
 by appending <samp>.store_copyprop</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>dce</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002ddce"></a>
 <p>Dump each function after dead code elimination.  The file name is made by
 appending <samp>.dce</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>sra</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dsra"></a>
 <p>Dump each function after performing scalar replacement of aggregates.  The
 file name is made by appending <samp>.sra</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>sink</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dsink"></a>
 <p>Dump each function after performing code sinking.  The file name is made
 by appending <samp>.sink</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>dom</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002ddom"></a>
 <p>Dump each function after applying dominator tree optimizations.  The file
 name is made by appending <samp>.dom</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>dse</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002ddse"></a>
 <p>Dump each function after applying dead store elimination.  The file
 name is made by appending <samp>.dse</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>phiopt</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dphiopt"></a>
 <p>Dump each function after optimizing PHI nodes into straightline code.  The file
 name is made by appending <samp>.phiopt</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>forwprop</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dforwprop"></a>
 <p>Dump each function after forward propagating single use variables.  The file
 name is made by appending <samp>.forwprop</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>copyrename</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dcopyrename"></a>
 <p>Dump each function after applying the copy rename optimization.  The file
 name is made by appending <samp>.copyrename</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>nrv</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dnrv"></a>
 <p>Dump each function after applying the named return value optimization on
 generic trees.  The file name is made by appending <samp>.nrv</samp> to the source
 file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>vect</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dvect"></a>
 <p>Dump each function after applying vectorization of loops.  The file name is
 made by appending <samp>.vect</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>slp</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dslp"></a>
 <p>Dump each function after applying vectorization of basic blocks.  The file name
 is made by appending <samp>.slp</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>vrp</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dvrp"></a>
 <p>Dump each function after Value Range Propagation (VRP).  The file name
 is made by appending <samp>.vrp</samp> to the source file name.
 </p>
 </dd>
 <dt>&lsquo;<samp>all</samp>&rsquo;</dt>
 <dd><a name="index-fdump_002dtree_002dall"></a>
 <p>Enable all the available tree dumps with the flags provided in this option.
 </p></dd>
 </dl>

 </dd>
 <dt><code>-fopt-info</code></dt>
 <dt><code>-fopt-info-<var>options</var></code></dt>
 <dt><code>-fopt-info-<var>options</var>=<var>filename</var></code></dt>
 <dd><a name="index-fopt_002dinfo"></a>
 <p>Controls optimization dumps from various optimization passes. If the
 &lsquo;<samp>-<var>options</var></samp>&rsquo; form is used, <var>options</var> is a list of
 &lsquo;<samp>-</samp>&rsquo; separated options to select the dump details and
 optimizations.  If <var>options</var> is not specified, it defaults to
 <samp>optimized</samp> for details and <samp>optall</samp> for optimization
 groups. If the <var>filename</var> is not specified, it defaults to
 <samp>stderr</samp>. Note that the output <var>filename</var> will be overwritten
 in case of multiple translation units. If a combined output from
 multiple translation units is desired, <samp>stderr</samp> should be used
 instead.
 </p>
 <p>The options can be divided into two groups, 1) options describing the
 verbosity of the dump, and 2) 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 latter options override the earlier options on the command
 line. Though multiple -fopt-info options are accepted, only one of
 them can have <samp>=filename</samp>. If other filenames are provided then
 all but the first one are ignored.
 </p>
 <p>The dump verbosity has the following options
 </p>
 <dl compact="compact">
 <dt>&lsquo;<samp>optimized</samp>&rsquo;</dt>
 <dd><p>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 got
 successfully vectorized.
 </p></dd>
 <dt>&lsquo;<samp>missed</samp>&rsquo;</dt>
 <dd><p>Print information about missed optimizations. Individual passes
 control which information to include in the output. For example,
 </p>
 <div class="smallexample">
 <pre class="smallexample">gcc -O2 -ftree-vectorize -fopt-info-vec-missed
 </pre></div>

 <p>will print information about missed optimization opportunities from
 vectorization passes on stderr.
 </p></dd>
 <dt>&lsquo;<samp>note</samp>&rsquo;</dt>
 <dd><p>Print verbose information about optimizations, such as certain
 transformations, more detailed messages about decisions etc.
 </p></dd>
 <dt>&lsquo;<samp>all</samp>&rsquo;</dt>
 <dd><p>Print detailed optimization information. This includes
 <var>optimized</var>, <var>missed</var>, and <var>note</var>.
 </p></dd>
 </dl>

 <p>The second set of options describes a group of optimizations and may
 include one or more of the following.
 </p>
 <dl compact="compact">
 <dt>&lsquo;<samp>ipa</samp>&rsquo;</dt>
 <dd><p>Enable dumps from all interprocedural optimizations.
 </p></dd>
 <dt>&lsquo;<samp>loop</samp>&rsquo;</dt>
 <dd><p>Enable dumps from all loop optimizations.
 </p></dd>
 <dt>&lsquo;<samp>inline</samp>&rsquo;</dt>
 <dd><p>Enable dumps from all inlining optimizations.
 </p></dd>
 <dt>&lsquo;<samp>vec</samp>&rsquo;</dt>
 <dd><p>Enable dumps from all vectorization optimizations.
 </p></dd>
 <dt>&lsquo;<samp>optall</samp>&rsquo;</dt>
 <dd><p>Enable dumps from all optimizations. This is a superset of
 the optimization groups listed above.
 </p></dd>
 </dl>

 <p>For example,
 </p><div class="smallexample">
 <pre class="smallexample">gcc -O3 -fopt-info-missed=missed.all
 </pre></div>

 <p>outputs missed optimization report from all the passes into
 <samp>missed.all</samp>.
 </p>
 <p>As another example,
 </p><div class="smallexample">
 <pre class="smallexample">gcc -O3 -fopt-info-inline-optimized-missed=inline.txt
 </pre></div>

 <p>will output information about missed optimizations as well as
 optimized locations from all the inlining passes into
 <samp>inline.txt</samp>.
 </p>
 <p>If the <var>filename</var> is provided, then the dumps from all the
 applicable optimizations are concatenated into the <samp>filename</samp>.
 Otherwise the dump is output onto <samp>stderr</samp>. If <var>options</var> is
 omitted, it defaults to <samp>all-optall</samp>, which means dump all
 available optimization info from all the passes. In the following
 example, all optimization info is output on to <samp>stderr</samp>.
 </p>
 <div class="smallexample">
 <pre class="smallexample">gcc -O3 -fopt-info
 </pre></div>

 <p>Note that <samp>-fopt-info-vec-missed</samp> behaves the same as
 <samp>-fopt-info-missed-vec</samp>.
 </p>
 <p>As another example, consider
 </p>
 <div class="smallexample">
 <pre class="smallexample">gcc -fopt-info-vec-missed=vec.miss -fopt-info-loop-optimized=loop.opt
 </pre></div>

 <p>Here the two output filenames <samp>vec.miss</samp> and <samp>loop.opt</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 the <samp>vec.miss</samp> is produced which contains
 dumps from the vectorizer about missed opportunities.
 </p>
 </dd>
 <dt><code>-frandom-seed=<var>string</var></code></dt>
 <dd><a name="index-frandom_002dseed"></a>
 <p>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>-frandom-seed</samp> option to produce
 reproducibly identical object files.
 </p>
 <p>The <var>string</var> should be different for every file you compile.
 </p>
 </dd>
 <dt><code>-fsched-verbose=<var>n</var></code></dt>
 <dd><a name="index-fsched_002dverbose"></a>
 <p>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>-fdump-rtl-sched1</samp> or
 <samp>-fdump-rtl-sched2</samp> is specified, in which case it is output
 to the usual dump listing file, <samp>.sched1</samp> or <samp>.sched2</samp>
 respectively.  However for <var>n</var> greater than nine, the output is
 always printed to standard error.
 </p>
 <p>For <var>n</var> greater than zero, <samp>-fsched-verbose</samp> outputs the
 same information as <samp>-fdump-rtl-sched1</samp> and <samp>-fdump-rtl-sched2</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>-fsched-verbose</samp> also includes
 dependence info.
 </p>
 </dd>
 <dt><code>-save-temps</code></dt>
 <dt><code>-save-temps=cwd</code></dt>
 <dd><a name="index-save_002dtemps"></a>
 <p>Store the usual &ldquo;temporary&rdquo; intermediate files permanently; place them
 in the current directory and name them based on the source file.  Thus,
 compiling <samp>foo.c</samp> with <samp>-c -save-temps</samp> produces files
 <samp>foo.i</samp> and <samp>foo.s</samp>, as well as <samp>foo.o</samp>.  This creates a
 preprocessed <samp>foo.i</samp> output file even though the compiler now
 normally uses an integrated preprocessor.
 </p>
 <p>When used in combination with the <samp>-x</samp> command-line option,
 <samp>-save-temps</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>-save-temps</samp>.
 </p>
 <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:
 </p>
 <div class="smallexample">
 <pre class="smallexample">gcc -save-temps -o outdir1/foo.o indir1/foo.c&amp;
 gcc -save-temps -o outdir2/foo.o indir2/foo.c&amp;
 </pre></div>

 <p>may result in <samp>foo.i</samp> and <samp>foo.o</samp> being written to
 simultaneously by both compilers.
 </p>
 </dd>
 <dt><code>-save-temps=obj</code></dt>
 <dd><a name="index-save_002dtemps_003dobj"></a>
 <p>Store the usual &ldquo;temporary&rdquo; intermediate files permanently.  If the
 <samp>-o</samp> option is used, the temporary files are based on the
 object file.  If the <samp>-o</samp> option is not used, the
 <samp>-save-temps=obj</samp> switch behaves like <samp>-save-temps</samp>.
 </p>
 <p>For example:
 </p>
 <div class="smallexample">
 <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></div>

 <p>creates <samp>foo.i</samp>, <samp>foo.s</samp>, <samp>dir/xbar.i</samp>,
 <samp>dir/xbar.s</samp>, <samp>dir2/yfoobar.i</samp>, <samp>dir2/yfoobar.s</samp>, and
 <samp>dir2/yfoobar.o</samp>.
 </p>
 </dd>
 <dt><code>-time<span class="roman">[</span>=<var>file</var><span class="roman">]</span></code></dt>
 <dd><a name="index-time"></a>
 <p>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>
 <p>Without the specification of an output file, the output looks like this:
 </p>
 <div class="smallexample">
 <pre class="smallexample"># cc1 0.12 0.01
 # as 0.00 0.01
 </pre></div>

 <p>The first number on each line is the &ldquo;user time&rdquo;, that is time spent
 executing the program itself.  The second number is &ldquo;system time&rdquo;,
 time spent executing operating system routines on behalf of the program.
 Both numbers are in seconds.
 </p>
 <p>With the specification of an output file, the output is appended to the
 named file, and it looks like this:
 </p>
 <div class="smallexample">
 <pre class="smallexample">0.12 0.01 cc1 <var>options</var>
 0.00 0.01 as <var>options</var>
 </pre></div>

 <p>The &ldquo;user time&rdquo; and the &ldquo;system time&rdquo; 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.
 </p>
 </dd>
 <dt><code>-fvar-tracking</code></dt>
 <dd><a name="index-fvar_002dtracking"></a>
 <p>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>
 <p>It is enabled by default when compiling with optimization (<samp>-Os</samp>,
 <samp>-O</samp>, <samp>-O2</samp>, &hellip;), debugging information (<samp>-g</samp>) and
 the debug info format supports it.
 </p>
 </dd>
 <dt><code>-fvar-tracking-assignments</code></dt>
 <dd><a name="index-fvar_002dtracking_002dassignments"></a>
 <a name="index-fno_002dvar_002dtracking_002dassignments"></a>
 <p>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>-gdwarf-4</samp> is recommended along with it.
 </p>
 <p>It can be enabled even if var-tracking is disabled, in which case
 annotations are created and maintained, but discarded at the end.
 </p>
 </dd>
 <dt><code>-fvar-tracking-assignments-toggle</code></dt>
 <dd><a name="index-fvar_002dtracking_002dassignments_002dtoggle"></a>
 <a name="index-fno_002dvar_002dtracking_002dassignments_002dtoggle"></a>
 <p>Toggle <samp>-fvar-tracking-assignments</samp>, in the same way that
 <samp>-gtoggle</samp> toggles <samp>-g</samp>.
 </p>
 </dd>
 <dt><code>-print-file-name=<var>library</var></code></dt>
 <dd><a name="index-print_002dfile_002dname"></a>
 <p>Print the full absolute name of the library file <var>library</var> that
 would be used when linking&mdash;and don&rsquo;t do anything else.  With this
 option, GCC does not compile or link anything; it just prints the
 file name.
 </p>
 </dd>
 <dt><code>-print-multi-directory</code></dt>
 <dd><a name="index-print_002dmulti_002ddirectory"></a>
 <p>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 <code>GCC_EXEC_PREFIX</code>.
 </p>
 </dd>
 <dt><code>-print-multi-lib</code></dt>
 <dd><a name="index-print_002dmulti_002dlib"></a>
 <p>Print the mapping from multilib directory names to compiler switches
 that enable them.  The directory name is separated from the switches by
 &lsquo;<samp>;</samp>&rsquo;, and each switch starts with an &lsquo;<samp>@</samp>&rsquo; instead of the
 &lsquo;<samp>-</samp>&rsquo;, without spaces between multiple switches.  This is supposed to
 ease shell processing.
 </p>
 </dd>
 <dt><code>-print-multi-os-directory</code></dt>
 <dd><a name="index-print_002dmulti_002dos_002ddirectory"></a>
 <p>Print the path to OS libraries for the selected
 multilib, relative to some <samp>lib</samp> subdirectory.  If OS libraries are
 present in the <samp>lib</samp> subdirectory and no multilibs are used, this is
 usually just <samp>.</samp>, if OS libraries are present in <samp>lib<var>suffix</var></samp>
 sibling directories this prints e.g. <samp>../lib64</samp>, <samp>../lib</samp> or
 <samp>../lib32</samp>, or if OS libraries are present in <samp>lib/<var>subdir</var></samp>
 subdirectories it prints e.g. <samp>amd64</samp>, <samp>sparcv9</samp> or <samp>ev6</samp>.
 </p>
 </dd>
 <dt><code>-print-multiarch</code></dt>
 <dd><a name="index-print_002dmultiarch"></a>
 <p>Print the path to OS libraries for the selected multiarch,
 relative to some <samp>lib</samp> subdirectory.
 </p>
 </dd>
 <dt><code>-print-prog-name=<var>program</var></code></dt>
 <dd><a name="index-print_002dprog_002dname"></a>
 <p>Like <samp>-print-file-name</samp>, but searches for a program such as &lsquo;<samp>cpp</samp>&rsquo;.
 </p>
 </dd>
 <dt><code>-print-libgcc-file-name</code></dt>
 <dd><a name="index-print_002dlibgcc_002dfile_002dname"></a>
 <p>Same as <samp>-print-file-name=libgcc.a</samp>.
 </p>
 <p>This is useful when you use <samp>-nostdlib</samp> or <samp>-nodefaultlibs</samp>
 but you do want to link with <samp>libgcc.a</samp>.  You can do:
 </p>
 <div class="smallexample">
 <pre class="smallexample">gcc -nostdlib <var>files</var>&hellip; `gcc -print-libgcc-file-name`
 </pre></div>

 </dd>
 <dt><code>-print-search-dirs</code></dt>
 <dd><a name="index-print_002dsearch_002ddirs"></a>
 <p>Print the name of the configured installation directory and a list of
 program and library directories <code>gcc</code> searches&mdash;and don&rsquo;t do anything else.
 </p>
 <p>This is useful when <code>gcc</code> prints the error message
 &lsquo;<samp>installation problem, cannot exec cpp0: No such file or directory</samp>&rsquo;.
 To resolve this you either need to put <samp>cpp0</samp> and the other compiler
 components where <code>gcc</code> expects to find them, or you can set the environment
 variable <code>GCC_EXEC_PREFIX</code> to the directory where you installed them.
 Don&rsquo;t forget the trailing &lsquo;<samp>/</samp>&rsquo;.
 See <a href="Environment-Variables.html#Environment-Variables">Environment Variables</a>.
 </p>
 </dd>
 <dt><code>-print-sysroot</code></dt>
 <dd><a name="index-print_002dsysroot"></a>
 <p>Print the target sysroot directory that is used during
 compilation.  This is the target sysroot specified either at configure
 time or using the <samp>--sysroot</samp> option, possibly with an extra
 suffix that depends on compilation options.  If no target sysroot is
 specified, the option prints nothing.
 </p>
 </dd>
 <dt><code>-print-sysroot-headers-suffix</code></dt>
 <dd><a name="index-print_002dsysroot_002dheaders_002dsuffix"></a>
 <p>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&mdash;and don&rsquo;t do anything else.
 </p>
 </dd>
 <dt><code>-dumpmachine</code></dt>
 <dd><a name="index-dumpmachine"></a>
 <p>Print the compiler&rsquo;s target machine (for example,
 &lsquo;<samp>i686-pc-linux-gnu</samp>&rsquo;)&mdash;and don&rsquo;t do anything else.
 </p>
 </dd>
 <dt><code>-dumpversion</code></dt>
 <dd><a name="index-dumpversion"></a>
 <p>Print the compiler version (for example, &lsquo;<samp>3.0</samp>&rsquo;)&mdash;and don&rsquo;t do
 anything else.
 </p>
 </dd>
 <dt><code>-dumpspecs</code></dt>
 <dd><a name="index-dumpspecs"></a>
 <p>Print the compiler&rsquo;s built-in specs&mdash;and don&rsquo;t do anything else.  (This
 is used when GCC itself is being built.)  See <a href="Spec-Files.html#Spec-Files">Spec Files</a>.
 </p>
 </dd>
 <dt><code>-fno-eliminate-unused-debug-types</code></dt>
 <dd><a name="index-feliminate_002dunused_002ddebug_002dtypes"></a>
 <a name="index-fno_002deliminate_002dunused_002ddebug_002dtypes"></a>
 <p>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.
 </p></dd>
 </dl>

 <hr>
 <div class="header">
 <p>
 Next: <a href="Optimize-Options.html#Optimize-Options" accesskey="n" rel="next">Optimize Options</a>, Previous: <a href="Warning-Options.html#Warning-Options" accesskey="p" rel="prev">Warning Options</a>, Up: <a href="Invoking-GCC.html#Invoking-GCC" accesskey="u" rel="up">Invoking GCC</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Option-Index.html#Option-Index" title="Index" rel="index">Index</a>]</p>
 </div>


 </body>
 </html>