visus
/
toolchain-gcc-arm-embedded

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<!-- Copyright (C) 1988-2020 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 6.5, http://www.gnu.org/software/texinfo/ -->
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>Instrumentation Options (Using the GNU Compiler Collection (GCC))</title>

<meta name="description" content="Instrumentation Options (Using the GNU Compiler Collection (GCC))">
<meta name="keywords" content="Instrumentation Options (Using the GNU Compiler Collection (GCC))">
<meta name="resource-type" content="document">
<meta name="distribution" content="global">
<meta name="Generator" content="makeinfo">
<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="Preprocessor-Options.html#Preprocessor-Options" rel="next" title="Preprocessor Options">
<link href="Optimize-Options.html#Optimize-Options" rel="prev" title="Optimize Options">
<style type="text/css">
<!--
a.summary-letter {text-decoration: none}
blockquote.indentedblock {margin-right: 0em}
blockquote.smallindentedblock {margin-right: 0em; font-size: smaller}
blockquote.smallquotation {font-size: smaller}
div.display {margin-left: 3.2em}
div.example {margin-left: 3.2em}
div.lisp {margin-left: 3.2em}
div.smalldisplay {margin-left: 3.2em}
div.smallexample {margin-left: 3.2em}
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.nolinebreak {white-space: nowrap}
span.roman {font-family: initial; font-weight: normal}
span.sansserif {font-family: sans-serif; font-weight: normal}
ul.no-bullet {list-style: none}
-->
</style>


</head>

<body lang="en">
<a name="Instrumentation-Options"></a>
<div class="header">
<p>
Next: <a href="Preprocessor-Options.html#Preprocessor-Options" accesskey="n" rel="next">Preprocessor Options</a>, Previous: <a href="Optimize-Options.html#Optimize-Options" accesskey="p" rel="prev">Optimize 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="Program-Instrumentation-Options"></a>
<h3 class="section">3.12 Program Instrumentation Options</h3>
<a name="index-instrumentation-options"></a>
<a name="index-program-instrumentation-options"></a>
<a name="index-run_002dtime-error-checking-options"></a>
<a name="index-profiling-options"></a>
<a name="index-options_002c-program-instrumentation"></a>
<a name="index-options_002c-run_002dtime-error-checking"></a>
<a name="index-options_002c-profiling"></a>

<p>GCC supports a number of command-line options that control adding
run-time instrumentation to the code it normally generates.  
For example, one purpose of instrumentation is collect profiling
statistics for use in finding program hot spots, code coverage
analysis, or profile-guided optimizations.
Another class of program instrumentation is adding run-time checking 
to detect programming errors like invalid pointer
dereferences or out-of-bounds array accesses, as well as deliberately
hostile attacks such as stack smashing or C++ vtable hijacking.
There is also a general hook which can be used to implement other
forms of tracing or function-level instrumentation for debug or
program analysis purposes.
</p>
<dl compact="compact">
<dd><a name="index-prof"></a>
<a name="index-gprof"></a>
</dd>
<dt><code>-p</code></dt>
<dt><code>-pg</code></dt>
<dd><a name="index-p"></a>
<a name="index-pg"></a>
<p>Generate extra code to write profile information suitable for the
analysis program <code>prof</code> (for <samp>-p</samp>) or <code>gprof</code>
(for <samp>-pg</samp>).  You must use this option when compiling
the source files you want data about, and you must also use it when
linking.
</p>
<p>You can use the function attribute <code>no_instrument_function</code> to
suppress profiling of individual functions when compiling with these options.
See <a href="Common-Function-Attributes.html#Common-Function-Attributes">Common Function Attributes</a>.
</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.  On targets that support
constructors with priority support, profiling properly handles constructors,
destructors and C++ constructors (and destructors) of classes which are used
as a type of a global variable.
</p>
<p>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> Compile the source files additionally with <samp>-fprofile-abs-path</samp>
to create absolute path names in the <samp>.gcno</samp> files.  This allows
<code>gcov</code> to find the correct sources in projects where compilations
occur with different working directories.

</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.  Unless
a strict ISO C dialect option is in effect, <code>fork</code> calls are
detected and correctly handled without double counting.

</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>-fprofile-abs-path</code></dt>
<dd><a name="index-fprofile_002dabs_002dpath"></a>
<p>Automatically convert relative source file names to absolute path names
in the <samp>.gcno</samp> files.  This allows <code>gcov</code> to find the correct
sources in projects where compilations occur with different working
directories.
</p>
</dd>
<dt><code>-fprofile-dir=<var>path</var></code></dt>
<dd><a name="index-fprofile_002ddir"></a>

<p>Set the directory to search for the profile data files in to <var>path</var>.
This option affects only the profile data generated by
<samp>-fprofile-generate</samp>, <samp>-ftest-coverage</samp>, <samp>-fprofile-arcs</samp>
and used by <samp>-fprofile-use</samp> and <samp>-fbranch-probabilities</samp>
and its related options.  Both absolute and relative paths can be used.
By default, GCC uses the current directory as <var>path</var>, thus the
profile data file appears in the same directory as the object file.
In order to prevent the file name clashing, if the object file name is
not an absolute path, we mangle the absolute path of the
<samp><var>sourcename</var>.gcda</samp> file and use it as the file name of a
<samp>.gcda</samp> file.  See similar option <samp>-fprofile-note</samp>.
</p>
<p>When an executable is run in a massive parallel environment, it is recommended
to save profile to different folders.  That can be done with variables
in <var>path</var> that are exported during run-time:
</p>
<dl compact="compact">
<dt><code>%p</code></dt>
<dd><p>process ID.
</p>
</dd>
<dt><code>%q{VAR}</code></dt>
<dd><p>value of environment variable <var>VAR</var>
</p>
</dd>
</dl>

</dd>
<dt><code>-fprofile-generate</code></dt>
<dt><code>-fprofile-generate=<var>path</var></code></dt>
<dd><a name="index-fprofile_002dgenerate"></a>

<p>Enable options usually used for instrumenting application to produce
profile useful for later recompilation with profile feedback based
optimization.  You must use <samp>-fprofile-generate</samp> both when
compiling and when linking your program.
</p>
<p>The following options are enabled:
<samp>-fprofile-arcs</samp>, <samp>-fprofile-values</samp>,
<samp>-finline-functions</samp>, and <samp>-fipa-bit-cp</samp>.
</p>
<p>If <var>path</var> is specified, GCC looks at the <var>path</var> to find
the profile feedback data files. See <samp>-fprofile-dir</samp>.
</p>
<p>To optimize the program based on the collected profile information, use
<samp>-fprofile-use</samp>.  See <a href="Optimize-Options.html#Optimize-Options">Optimize Options</a>, for more information.
</p>
</dd>
<dt><code>-fprofile-note=<var>path</var></code></dt>
<dd><a name="index-fprofile_002dnote"></a>

<p>If <var>path</var> is specified, GCC saves <samp>.gcno</samp> file into <var>path</var>
location.  If you combine the option with multiple source files,
the <samp>.gcno</samp> file will be overwritten.
</p>
</dd>
<dt><code>-fprofile-prefix-path=<var>path</var></code></dt>
<dd><a name="index-fprofile_002dprefix_002dpath"></a>

<p>This option can be used in combination with
<samp>profile-generate=</samp><var>profile_dir</var> and
<samp>profile-use=</samp><var>profile_dir</var> to inform GCC where is the base
directory of built source tree.  By default <var>profile_dir</var> will contain
files with mangled absolute paths of all object files in the built project.
This is not desirable when directory used to build the instrumented binary
differs from the directory used to build the binary optimized with profile
feedback because the profile data will not be found during the optimized build.
In such setups <samp>-fprofile-prefix-path=</samp><var>path</var> with <var>path</var>
pointing to the base directory of the build can be used to strip the irrelevant
part of the path and keep all file names relative to the main build directory.
</p>
</dd>
<dt><code>-fprofile-update=<var>method</var></code></dt>
<dd><a name="index-fprofile_002dupdate"></a>

<p>Alter the update method for an application instrumented for profile
feedback based optimization.  The <var>method</var> argument should be one of
&lsquo;<samp>single</samp>&rsquo;, &lsquo;<samp>atomic</samp>&rsquo; or &lsquo;<samp>prefer-atomic</samp>&rsquo;.
The first one is useful for single-threaded applications,
while the second one prevents profile corruption by emitting thread-safe code.
</p>
<p><strong>Warning:</strong> When an application does not properly join all threads
(or creates an detached thread), a profile file can be still corrupted.
</p>
<p>Using &lsquo;<samp>prefer-atomic</samp>&rsquo; would be transformed either to &lsquo;<samp>atomic</samp>&rsquo;,
when supported by a target, or to &lsquo;<samp>single</samp>&rsquo; otherwise.  The GCC driver
automatically selects &lsquo;<samp>prefer-atomic</samp>&rsquo; when <samp>-pthread</samp>
is present in the command line.
</p>
</dd>
<dt><code>-fprofile-filter-files=<var>regex</var></code></dt>
<dd><a name="index-fprofile_002dfilter_002dfiles"></a>

<p>Instrument only functions from files where names match
any regular expression (separated by a semi-colon).
</p>
<p>For example, <samp>-fprofile-filter-files=main.c;module.*.c</samp> will instrument
only <samp>main.c</samp> and all C files starting with &rsquo;module&rsquo;.
</p>
</dd>
<dt><code>-fprofile-exclude-files=<var>regex</var></code></dt>
<dd><a name="index-fprofile_002dexclude_002dfiles"></a>

<p>Instrument only functions from files where names do not match
all the regular expressions (separated by a semi-colon).
</p>
<p>For example, <samp>-fprofile-exclude-files=/usr/*</samp> will prevent instrumentation
of all files that are located in <samp>/usr/</samp> folder.
</p>
</dd>
<dt><code>-fprofile-reproducible=<span class="roman">[</span>multithreaded<span class="roman">|</span>parallel-runs<span class="roman">|</span>serial<span class="roman">]</span></code></dt>
<dd><a name="index-fprofile_002dreproducible"></a>
<p>Control level of reproducibility of profile gathered by
<code>-fprofile-generate</code>.  This makes it possible to rebuild program
with same outcome which is useful, for example, for distribution
packages.
</p>
<p>With <samp>-fprofile-reproducible=serial</samp> the profile gathered by
<samp>-fprofile-generate</samp> is reproducible provided the trained program
behaves the same at each invocation of the train run, it is not
multi-threaded and profile data streaming is always done in the same
order.  Note that profile streaming happens at the end of program run but
also before <code>fork</code> function is invoked.
</p>
<p>Note that it is quite common that execution counts of some part of
programs depends, for example, on length of temporary file names or
memory space randomization (that may affect hash-table collision rate).
Such non-reproducible part of programs may be annotated by
<code>no_instrument_function</code> function attribute. <code>gcov-dump</code> with
<samp>-l</samp> can be used to dump gathered data and verify that they are
indeed reproducible.
</p>
<p>With <samp>-fprofile-reproducible=parallel-runs</samp> collected profile
stays reproducible regardless the order of streaming of the data into
gcda files.  This setting makes it possible to run multiple instances of
instrumented program in parallel (such as with <code>make -j</code>). This
reduces quality of gathered data, in particular of indirect call
profiling.
</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 are instrumented to detect
out-of-bounds and use-after-free bugs.
The option enables <samp>-fsanitize-address-use-after-scope</samp>.
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 instrumented 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.
The option cannot be combined with <samp>-fsanitize=thread</samp>.
</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/kasan/wiki">https://github.com/google/kasan/wiki</a> for more details.
</p>
</dd>
<dt><code>-fsanitize=pointer-compare</code></dt>
<dd><a name="index-fsanitize_003dpointer_002dcompare"></a>
<p>Instrument comparison operation (&lt;, &lt;=, &gt;, &gt;=) with pointer operands.
The option must be combined with either <samp>-fsanitize=kernel-address</samp> or
<samp>-fsanitize=address</samp>
The option cannot be combined with <samp>-fsanitize=thread</samp>.
Note: By default the check is disabled at run time.  To enable it,
add <code>detect_invalid_pointer_pairs=2</code> to the environment variable
<code>ASAN_OPTIONS</code>. Using <code>detect_invalid_pointer_pairs=1</code> detects
invalid operation only when both pointers are non-null.
</p>
</dd>
<dt><code>-fsanitize=pointer-subtract</code></dt>
<dd><a name="index-fsanitize_003dpointer_002dsubtract"></a>
<p>Instrument subtraction with pointer operands.
The option must be combined with either <samp>-fsanitize=kernel-address</samp> or
<samp>-fsanitize=address</samp>
The option cannot be combined with <samp>-fsanitize=thread</samp>.
Note: By default the check is disabled at run time.  To enable it,
add <code>detect_invalid_pointer_pairs=2</code> to the environment variable
<code>ASAN_OPTIONS</code>. Using <code>detect_invalid_pointer_pairs=1</code> detects
invalid operation only when both pointers are non-null.
</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 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 <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.
The option cannot be combined with <samp>-fsanitize=address</samp>,
<samp>-fsanitize=leak</samp>.
</p>
<p>Note that sanitized atomic builtins cannot throw exceptions when
operating on invalid memory addresses with non-call exceptions
(<samp>-fnon-call-exceptions</samp>).
</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
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
<code>LSAN_OPTIONS</code> environment variable.
The option cannot be combined with <samp>-fsanitize=thread</samp>.
</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 are 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.
This option has two suboptions, <samp>-fsanitize=shift-base</samp> and
<samp>-fsanitize=shift-exponent</samp>.
</p>
</dd>
<dt><code>-fsanitize=shift-exponent</code></dt>
<dd><a name="index-fsanitize_003dshift_002dexponent"></a>
<p>This option enables checking that the second argument of a shift operation
is not negative and is smaller than the precision of the promoted first
argument.
</p>
</dd>
<dt><code>-fsanitize=shift-base</code></dt>
<dd><a name="index-fsanitize_003dshift_002dbase"></a>
<p>If the second argument of a shift operation is within range, check 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 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.
</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.
</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, or if a method is invoked
on an object pointed by 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>
<dt><code>-fsanitize=bounds</code></dt>
<dd><a name="index-fsanitize_003dbounds"></a>
<p>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.
</p>
</dd>
<dt><code>-fsanitize=bounds-strict</code></dt>
<dd><a name="index-fsanitize_003dbounds_002dstrict"></a>
<p>This option enables strict instrumentation of array bounds.  Most out of bounds
accesses are detected, including flexible array members and flexible array
member-like arrays.  Initializers of variables with static storage are not
instrumented.
</p>
</dd>
<dt><code>-fsanitize=alignment</code></dt>
<dd><a name="index-fsanitize_003dalignment"></a>

<p>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.
</p>
</dd>
<dt><code>-fsanitize=object-size</code></dt>
<dd><a name="index-fsanitize_003dobject_002dsize"></a>
<p>This option enables instrumentation of memory references using the
<code>__builtin_object_size</code> function.  Various out of bounds pointer
accesses are detected.
</p>
</dd>
<dt><code>-fsanitize=float-divide-by-zero</code></dt>
<dd><a name="index-fsanitize_003dfloat_002ddivide_002dby_002dzero"></a>
<p>Detect floating-point division by zero.  Unlike other similar options,
<samp>-fsanitize=float-divide-by-zero</samp> is not enabled by
<samp>-fsanitize=undefined</samp>, since floating-point division by zero can
be a legitimate way of obtaining infinities and NaNs.
</p>
</dd>
<dt><code>-fsanitize=float-cast-overflow</code></dt>
<dd><a name="index-fsanitize_003dfloat_002dcast_002doverflow"></a>
<p>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>-fsanitize=float-cast-overflow</samp> is
not enabled by <samp>-fsanitize=undefined</samp>.
This option does not work well with <code>FE_INVALID</code> exceptions enabled.
</p>
</dd>
<dt><code>-fsanitize=nonnull-attribute</code></dt>
<dd><a name="index-fsanitize_003dnonnull_002dattribute"></a>

<p>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.
</p>
</dd>
<dt><code>-fsanitize=returns-nonnull-attribute</code></dt>
<dd><a name="index-fsanitize_003dreturns_002dnonnull_002dattribute"></a>

<p>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.
</p>
</dd>
<dt><code>-fsanitize=bool</code></dt>
<dd><a name="index-fsanitize_003dbool"></a>

<p>This option enables instrumentation of loads from bool.  If a value other
than 0/1 is loaded, a run-time error is issued.
</p>
</dd>
<dt><code>-fsanitize=enum</code></dt>
<dd><a name="index-fsanitize_003denum"></a>

<p>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.
</p>
</dd>
<dt><code>-fsanitize=vptr</code></dt>
<dd><a name="index-fsanitize_003dvptr"></a>

<p>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.
</p>
</dd>
<dt><code>-fsanitize=pointer-overflow</code></dt>
<dd><a name="index-fsanitize_003dpointer_002doverflow"></a>

<p>This option enables instrumentation of pointer arithmetics.  If the pointer
arithmetics overflows, a run-time error is issued.
</p>
</dd>
<dt><code>-fsanitize=builtin</code></dt>
<dd><a name="index-fsanitize_003dbuiltin"></a>

<p>This option enables instrumentation of arguments to selected builtin
functions.  If an invalid value is passed to such arguments, a run-time
error is issued.  E.g.&nbsp;passing 0 as the argument to <code>__builtin_ctz</code>
or <code>__builtin_clz</code> invokes undefined behavior and is diagnosed
by this option.
</p>
</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>-fno-sanitize=all</code></dt>
<dd><a name="index-fno_002dsanitize_003dall"></a>

<p>This option disables all previously enabled sanitizers.
<samp>-fsanitize=all</samp> is not allowed, as some sanitizers cannot be used
together.
</p>
</dd>
<dt><code>-fasan-shadow-offset=<var>number</var></code></dt>
<dd><a name="index-fasan_002dshadow_002doffset"></a>
<p>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.
</p>
</dd>
<dt><code>-fsanitize-sections=<var>s1</var>,<var>s2</var>,...</code></dt>
<dd><a name="index-fsanitize_002dsections"></a>
<p>Sanitize global variables in selected user-defined sections.  <var>si</var> may
contain wildcards.
</p>
</dd>
<dt><code>-fsanitize-recover<span class="roman">[</span>=<var>opts</var><span class="roman">]</span></code></dt>
<dd><a name="index-fsanitize_002drecover"></a>
<a name="index-fno_002dsanitize_002drecover"></a>
<p><samp>-fsanitize-recover=</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>-fno-sanitize-recover=</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>
<p>Currently this feature only works for <samp>-fsanitize=undefined</samp> (and its suboptions
except for <samp>-fsanitize=unreachable</samp> and <samp>-fsanitize=return</samp>),
<samp>-fsanitize=float-cast-overflow</samp>, <samp>-fsanitize=float-divide-by-zero</samp>,
<samp>-fsanitize=bounds-strict</samp>,
<samp>-fsanitize=kernel-address</samp> and <samp>-fsanitize=address</samp>.
For these sanitizers error recovery is turned on by default,
except <samp>-fsanitize=address</samp>, for which this feature is experimental.
<samp>-fsanitize-recover=all</samp> and <samp>-fno-sanitize-recover=all</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>
<p>Even if a recovery mode is turned on the compiler side, it needs to be also
enabled on the runtime library side, otherwise the failures are still fatal.
The runtime library defaults to <code>halt_on_error=0</code> for
ThreadSanitizer and UndefinedBehaviorSanitizer, while default value for
AddressSanitizer is <code>halt_on_error=1</code>. This can be overridden through
setting the <code>halt_on_error</code> flag in the corresponding environment variable.
</p>
<p>Syntax without an explicit <var>opts</var> parameter is deprecated.  It is
equivalent to specifying an <var>opts</var> list of:
</p>
<div class="smallexample">
<pre class="smallexample">undefined,float-cast-overflow,float-divide-by-zero,bounds-strict
</pre></div>

</dd>
<dt><code>-fsanitize-address-use-after-scope</code></dt>
<dd><a name="index-fsanitize_002daddress_002duse_002dafter_002dscope"></a>
<p>Enable sanitization of local variables to detect use-after-scope bugs.
The option sets <samp>-fstack-reuse</samp> to &lsquo;<samp>none</samp>&rsquo;.
</p>
</dd>
<dt><code>-fsanitize-undefined-trap-on-error</code></dt>
<dd><a name="index-fsanitize_002dundefined_002dtrap_002don_002derror"></a>
<p>The <samp>-fsanitize-undefined-trap-on-error</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.
</p>
</dd>
<dt><code>-fsanitize-coverage=trace-pc</code></dt>
<dd><a name="index-fsanitize_002dcoverage_003dtrace_002dpc"></a>
<p>Enable coverage-guided fuzzing code instrumentation.
Inserts a call to <code>__sanitizer_cov_trace_pc</code> into every basic block.
</p>
</dd>
<dt><code>-fsanitize-coverage=trace-cmp</code></dt>
<dd><a name="index-fsanitize_002dcoverage_003dtrace_002dcmp"></a>
<p>Enable dataflow guided fuzzing code instrumentation.
Inserts a call to <code>__sanitizer_cov_trace_cmp1</code>,
<code>__sanitizer_cov_trace_cmp2</code>, <code>__sanitizer_cov_trace_cmp4</code> or
<code>__sanitizer_cov_trace_cmp8</code> for integral comparison with both operands
variable or <code>__sanitizer_cov_trace_const_cmp1</code>,
<code>__sanitizer_cov_trace_const_cmp2</code>,
<code>__sanitizer_cov_trace_const_cmp4</code> or
<code>__sanitizer_cov_trace_const_cmp8</code> for integral comparison with one
operand constant, <code>__sanitizer_cov_trace_cmpf</code> or
<code>__sanitizer_cov_trace_cmpd</code> for float or double comparisons and
<code>__sanitizer_cov_trace_switch</code> for switch statements.
</p>
</dd>
<dt><code>-fcf-protection=<span class="roman">[</span>full<span class="roman">|</span>branch<span class="roman">|</span>return<span class="roman">|</span>none<span class="roman">|</span>check<span class="roman">]</span></code></dt>
<dd><a name="index-fcf_002dprotection"></a>
<p>Enable code instrumentation of control-flow transfers to increase
program security by checking that target addresses of control-flow
transfer instructions (such as indirect function call, function return,
indirect jump) are valid.  This prevents diverting the flow of control
to an unexpected target.  This is intended to protect against such
threats as Return-oriented Programming (ROP), and similarly
call/jmp-oriented programming (COP/JOP).
</p>
<p>The value <code>branch</code> tells the compiler to implement checking of
validity of control-flow transfer at the point of indirect branch
instructions, i.e. call/jmp instructions.  The value <code>return</code>
implements checking of validity at the point of returning from a
function.  The value <code>full</code> is an alias for specifying both
<code>branch</code> and <code>return</code>. The value <code>none</code> turns off
instrumentation.
</p>
<p>The value <code>check</code> is used for the final link with link-time
optimization (LTO).  An error is issued if LTO object files are
compiled with different <samp>-fcf-protection</samp> values.  The
value <code>check</code> is ignored at the compile time.
</p>
<p>The macro <code>__CET__</code> is defined when <samp>-fcf-protection</samp> is
used.  The first bit of <code>__CET__</code> is set to 1 for the value
<code>branch</code> and the second bit of <code>__CET__</code> is set to 1 for
the <code>return</code>.
</p>
<p>You can also use the <code>nocf_check</code> attribute to identify
which functions and calls should be skipped from instrumentation
(see <a href="Function-Attributes.html#Function-Attributes">Function Attributes</a>).
</p>
<p>Currently the x86 GNU/Linux target provides an implementation based
on Intel Control-flow Enforcement Technology (CET).
</p>
</dd>
<dt><code>-fstack-protector</code></dt>
<dd><a name="index-fstack_002dprotector"></a>
<p>Emit extra code to check for buffer overflows, such as stack smashing
attacks.  This is done by adding a guard variable to functions with
vulnerable objects.  This includes functions that call <code>alloca</code>, and
functions with buffers larger than or equal to 8 bytes.  The guards are
initialized when a function is entered and then checked when the function
exits.  If a guard check fails, an error message is printed and the program
exits.  Only variables that are actually allocated on the stack are
considered, optimized away variables or variables allocated in registers
don&rsquo;t count.
</p>
</dd>
<dt><code>-fstack-protector-all</code></dt>
<dd><a name="index-fstack_002dprotector_002dall"></a>
<p>Like <samp>-fstack-protector</samp> except that all functions are protected.
</p>
</dd>
<dt><code>-fstack-protector-strong</code></dt>
<dd><a name="index-fstack_002dprotector_002dstrong"></a>
<p>Like <samp>-fstack-protector</samp> but includes additional functions to
be protected &mdash; those that have local array definitions, or have
references to local frame addresses.  Only variables that are actually
allocated on the stack are considered, optimized away variables or variables
allocated in registers don&rsquo;t count.
</p>
</dd>
<dt><code>-fstack-protector-explicit</code></dt>
<dd><a name="index-fstack_002dprotector_002dexplicit"></a>
<p>Like <samp>-fstack-protector</samp> but only protects those functions which
have the <code>stack_protect</code> attribute.
</p>
</dd>
<dt><code>-fstack-check</code></dt>
<dd><a name="index-fstack_002dcheck"></a>
<p>Generate code to verify that you do not go beyond the boundary of the
stack.  You should specify this flag if you are running in an
environment with multiple threads, but you only rarely need to specify it in
a single-threaded environment since stack overflow is automatically
detected on nearly all systems if there is only one stack.
</p>
<p>Note that this switch does not actually cause checking to be done; the
operating system or the language runtime must do that.  The switch causes
generation of code to ensure that they see the stack being extended.
</p>
<p>You can additionally specify a string parameter: &lsquo;<samp>no</samp>&rsquo; means no
checking, &lsquo;<samp>generic</samp>&rsquo; means force the use of old-style checking,
&lsquo;<samp>specific</samp>&rsquo; means use the best checking method and is equivalent
to bare <samp>-fstack-check</samp>.
</p>
<p>Old-style checking is a generic mechanism that requires no specific
target support in the compiler but comes with the following drawbacks:
</p>
<ol>
<li> Modified allocation strategy for large objects: they are always
allocated dynamically if their size exceeds a fixed threshold.  Note this
may change the semantics of some code.

</li><li> Fixed limit on the size of the static frame of functions: when it is
topped by a particular function, stack checking is not reliable and
a warning is issued by the compiler.

</li><li> Inefficiency: because of both the modified allocation strategy and the
generic implementation, code performance is hampered.
</li></ol>

<p>Note that old-style stack checking is also the fallback method for
&lsquo;<samp>specific</samp>&rsquo; if no target support has been added in the compiler.
</p>
<p>&lsquo;<samp>-fstack-check=</samp>&rsquo; is designed for Ada&rsquo;s needs to detect infinite recursion
and stack overflows.  &lsquo;<samp>specific</samp>&rsquo; is an excellent choice when compiling
Ada code.  It is not generally sufficient to protect against stack-clash
attacks.  To protect against those you want &lsquo;<samp>-fstack-clash-protection</samp>&rsquo;.
</p>
</dd>
<dt><code>-fstack-clash-protection</code></dt>
<dd><a name="index-fstack_002dclash_002dprotection"></a>
<p>Generate code to prevent stack clash style attacks.  When this option is
enabled, the compiler will only allocate one page of stack space at a time
and each page is accessed immediately after allocation.  Thus, it prevents
allocations from jumping over any stack guard page provided by the
operating system.
</p>
<p>Most targets do not fully support stack clash protection.  However, on
those targets <samp>-fstack-clash-protection</samp> will protect dynamic stack
allocations.  <samp>-fstack-clash-protection</samp> may also provide limited
protection for static stack allocations if the target supports
<samp>-fstack-check=specific</samp>.
</p>
</dd>
<dt><code>-fstack-limit-register=<var>reg</var></code></dt>
<dt><code>-fstack-limit-symbol=<var>sym</var></code></dt>
<dt><code>-fno-stack-limit</code></dt>
<dd><a name="index-fstack_002dlimit_002dregister"></a>
<a name="index-fstack_002dlimit_002dsymbol"></a>
<a name="index-fno_002dstack_002dlimit"></a>
<p>Generate code to ensure that the stack does not grow beyond a certain value,
either the value of a register or the address of a symbol.  If a larger
stack is required, a signal is raised at run time.  For most targets,
the signal is raised before the stack overruns the boundary, so
it is possible to catch the signal without taking special precautions.
</p>
<p>For instance, if the stack starts at absolute address &lsquo;<samp>0x80000000</samp>&rsquo;
and grows downwards, you can use the flags
<samp>-fstack-limit-symbol=__stack_limit</samp> and
<samp>-Wl,--defsym,__stack_limit=0x7ffe0000</samp> to enforce a stack limit
of 128KB.  Note that this may only work with the GNU linker.
</p>
<p>You can locally override stack limit checking by using the
<code>no_stack_limit</code> function attribute (see <a href="Function-Attributes.html#Function-Attributes">Function Attributes</a>).
</p>
</dd>
<dt><code>-fsplit-stack</code></dt>
<dd><a name="index-fsplit_002dstack"></a>
<p>Generate code to automatically split the stack before it overflows.
The resulting program has a discontiguous stack which can only
overflow if the program is unable to allocate any more memory.  This
is most useful when running threaded programs, as it is no longer
necessary to calculate a good stack size to use for each thread.  This
is currently only implemented for the x86 targets running
GNU/Linux.
</p>
<p>When code compiled with <samp>-fsplit-stack</samp> calls code compiled
without <samp>-fsplit-stack</samp>, there may not be much stack space
available for the latter code to run.  If compiling all code,
including library code, with <samp>-fsplit-stack</samp> is not an option,
then the linker can fix up these calls so that the code compiled
without <samp>-fsplit-stack</samp> always has a large stack.  Support for
this is implemented in the gold linker in GNU binutils release 2.21
and later.
</p>
</dd>
<dt><code>-fvtable-verify=<span class="roman">[</span>std<span class="roman">|</span>preinit<span class="roman">|</span>none<span class="roman">]</span></code></dt>
<dd><a name="index-fvtable_002dverify"></a>
<p>This option is only available when compiling C++ code.
It turns on (or off, if using <samp>-fvtable-verify=none</samp>) the security
feature that verifies at run time, for every virtual call, that
the vtable pointer through which the call is made is valid for the type of
the object, and has not been corrupted or overwritten.  If an invalid vtable
pointer is detected at run time, an error is reported and execution of the
program is immediately halted.
</p>
<p>This option causes run-time data structures to be built at program startup,
which are used for verifying the vtable pointers.  
The options &lsquo;<samp>std</samp>&rsquo; and &lsquo;<samp>preinit</samp>&rsquo;
control the timing of when these data structures are built.  In both cases the
data structures are built before execution reaches <code>main</code>.  Using
<samp>-fvtable-verify=std</samp> causes the data structures to be built after
shared libraries have been loaded and initialized.
<samp>-fvtable-verify=preinit</samp> causes them to be built before shared
libraries have been loaded and initialized.
</p>
<p>If this option appears multiple times in the command line with different
values specified, &lsquo;<samp>none</samp>&rsquo; takes highest priority over both &lsquo;<samp>std</samp>&rsquo; and
&lsquo;<samp>preinit</samp>&rsquo;; &lsquo;<samp>preinit</samp>&rsquo; takes priority over &lsquo;<samp>std</samp>&rsquo;.
</p>
</dd>
<dt><code>-fvtv-debug</code></dt>
<dd><a name="index-fvtv_002ddebug"></a>
<p>When used in conjunction with <samp>-fvtable-verify=std</samp> or 
<samp>-fvtable-verify=preinit</samp>, causes debug versions of the 
runtime functions for the vtable verification feature to be called.  
This flag also causes the compiler to log information about which 
vtable pointers it finds for each class.
This information is written to a file named <samp>vtv_set_ptr_data.log</samp> 
in the directory named by the environment variable <code>VTV_LOGS_DIR</code> 
if that is defined or the current working directory otherwise.
</p>
<p>Note:  This feature <em>appends</em> data to the log file. If you want a fresh log
file, be sure to delete any existing one.
</p>
</dd>
<dt><code>-fvtv-counts</code></dt>
<dd><a name="index-fvtv_002dcounts"></a>
<p>This is a debugging flag.  When used in conjunction with
<samp>-fvtable-verify=std</samp> or <samp>-fvtable-verify=preinit</samp>, this
causes the compiler to keep track of the total number of virtual calls
it encounters and the number of verifications it inserts.  It also
counts the number of calls to certain run-time library functions
that it inserts and logs this information for each compilation unit.
The compiler writes this information to a file named
<samp>vtv_count_data.log</samp> in the directory named by the environment
variable <code>VTV_LOGS_DIR</code> if that is defined or the current working
directory otherwise.  It also counts the size of the vtable pointer sets
for each class, and writes this information to <samp>vtv_class_set_sizes.log</samp>
in the same directory.
</p>
<p>Note:  This feature <em>appends</em> data to the log files.  To get fresh log
files, be sure to delete any existing ones.
</p>
</dd>
<dt><code>-finstrument-functions</code></dt>
<dd><a name="index-finstrument_002dfunctions"></a>
<p>Generate instrumentation calls for entry and exit to functions.  Just
after function entry and just before function exit, the following
profiling functions are called with the address of the current
function and its call site.  (On some platforms,
<code>__builtin_return_address</code> does not work beyond the current
function, so the call site information may not be available to the
profiling functions otherwise.)
</p>
<div class="smallexample">
<pre class="smallexample">void __cyg_profile_func_enter (void *this_fn,
                               void *call_site);
void __cyg_profile_func_exit  (void *this_fn,
                               void *call_site);
</pre></div>

<p>The first argument is the address of the start of the current function,
which may be looked up exactly in the symbol table.
</p>
<p>This instrumentation is also done for functions expanded inline in other
functions.  The profiling calls indicate where, conceptually, the
inline function is entered and exited.  This means that addressable
versions of such functions must be available.  If all your uses of a
function are expanded inline, this may mean an additional expansion of
code size.  If you use <code>extern inline</code> in your C code, an
addressable version of such functions must be provided.  (This is
normally the case anyway, but if you get lucky and the optimizer always
expands the functions inline, you might have gotten away without
providing static copies.)
</p>
<p>A function may be given the attribute <code>no_instrument_function</code>, in
which case this instrumentation is not done.  This can be used, for
example, for the profiling functions listed above, high-priority
interrupt routines, and any functions from which the profiling functions
cannot safely be called (perhaps signal handlers, if the profiling
routines generate output or allocate memory).
See <a href="Common-Function-Attributes.html#Common-Function-Attributes">Common Function Attributes</a>.
</p>
</dd>
<dt><code>-finstrument-functions-exclude-file-list=<var>file</var>,<var>file</var>,&hellip;</code></dt>
<dd><a name="index-finstrument_002dfunctions_002dexclude_002dfile_002dlist"></a>

<p>Set the list of functions that are excluded from instrumentation (see
the description of <samp>-finstrument-functions</samp>).  If the file that
contains a function definition matches with one of <var>file</var>, then
that function is not instrumented.  The match is done on substrings:
if the <var>file</var> parameter is a substring of the file name, it is
considered to be a match.
</p>
<p>For example:
</p>
<div class="smallexample">
<pre class="smallexample">-finstrument-functions-exclude-file-list=/bits/stl,include/sys
</pre></div>

<p>excludes any inline function defined in files whose pathnames
contain <samp>/bits/stl</samp> or <samp>include/sys</samp>.
</p>
<p>If, for some reason, you want to include letter &lsquo;<samp>,</samp>&rsquo; in one of
<var>sym</var>, write &lsquo;<samp>\,</samp>&rsquo;. For example,
<samp>-finstrument-functions-exclude-file-list='\,\,tmp'</samp>
(note the single quote surrounding the option).
</p>
</dd>
<dt><code>-finstrument-functions-exclude-function-list=<var>sym</var>,<var>sym</var>,&hellip;</code></dt>
<dd><a name="index-finstrument_002dfunctions_002dexclude_002dfunction_002dlist"></a>

<p>This is similar to <samp>-finstrument-functions-exclude-file-list</samp>,
but this option sets the list of function names to be excluded from
instrumentation.  The function name to be matched is its user-visible
name, such as <code>vector&lt;int&gt; blah(const vector&lt;int&gt; &amp;)</code>, not the
internal mangled name (e.g., <code>_Z4blahRSt6vectorIiSaIiEE</code>).  The
match is done on substrings: if the <var>sym</var> parameter is a substring
of the function name, it is considered to be a match.  For C99 and C++
extended identifiers, the function name must be given in UTF-8, not
using universal character names.
</p>
</dd>
<dt><code>-fpatchable-function-entry=<var>N</var>[,<var>M</var>]</code></dt>
<dd><a name="index-fpatchable_002dfunction_002dentry"></a>
<p>Generate <var>N</var> NOPs right at the beginning
of each function, with the function entry point before the <var>M</var>th NOP.
If <var>M</var> is omitted, it defaults to <code>0</code> so the
function entry points to the address just at the first NOP.
The NOP instructions reserve extra space which can be used to patch in
any desired instrumentation at run time, provided that the code segment
is writable.  The amount of space is controllable indirectly via
the number of NOPs; the NOP instruction used corresponds to the instruction
emitted by the internal GCC back-end interface <code>gen_nop</code>.  This behavior
is target-specific and may also depend on the architecture variant and/or
other compilation options.
</p>
<p>For run-time identification, the starting addresses of these areas,
which correspond to their respective function entries minus <var>M</var>,
are additionally collected in the <code>__patchable_function_entries</code>
section of the resulting binary.
</p>
<p>Note that the value of <code>__attribute__ ((patchable_function_entry
(N,M)))</code> takes precedence over command-line option
<samp>-fpatchable-function-entry=N,M</samp>.  This can be used to increase
the area size or to remove it completely on a single function.
If <code>N=0</code>, no pad location is recorded.
</p>
<p>The NOP instructions are inserted at&mdash;and maybe before, depending on
<var>M</var>&mdash;the function entry address, even before the prologue.
</p>
</dd>
</dl>


<hr>
<div class="header">
<p>
Next: <a href="Preprocessor-Options.html#Preprocessor-Options" accesskey="n" rel="next">Preprocessor Options</a>, Previous: <a href="Optimize-Options.html#Optimize-Options" accesskey="p" rel="prev">Optimize 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>