\input texinfo @c -*-texinfo-*-
@setfilename gprof.info
+@c Copyright (C) 1988-2020 Free Software Foundation, Inc.
@settitle GNU gprof
@setchapternewpage odd
-@ifinfo
+@c man begin INCLUDE
+@include bfdver.texi
+@c man end
+
+@ifnottex
@c This is a dir.info fragment to support semi-automated addition of
@c manuals to an info tree. zoo@cygnus.com is developing this facility.
-@format
-START-INFO-DIR-ENTRY
+@dircategory Software development
+@direntry
* gprof: (gprof). Profiling your program's execution
-END-INFO-DIR-ENTRY
-@end format
-@end ifinfo
+@end direntry
+@end ifnottex
-@ifinfo
+@copying
This file documents the gprof profiler of the GNU system.
-Copyright (C) 1988, 92, 97, 98, 99, 2000 Free Software Foundation, Inc.
-
-Permission is granted to make and distribute verbatim copies of
-this manual provided the copyright notice and this permission notice
-are preserved on all copies.
+@c man begin COPYRIGHT
+Copyright @copyright{} 1988-2020 Free Software Foundation, Inc.
-@ignore
-Permission is granted to process this file through Tex and print the
-results, provided the printed document carries copying permission
-notice identical to this one except for the removal of this paragraph
-(this paragraph not being relevant to the printed manual).
+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 no Invariant Sections, with no Front-Cover Texts, and with no
+Back-Cover Texts. A copy of the license is included in the
+section entitled ``GNU Free Documentation License''.
-@end ignore
-Permission is granted to copy and distribute modified versions of this
-manual under the conditions for verbatim copying, provided that the entire
-resulting derived work is distributed under the terms of a permission
-notice identical to this one.
-
-Permission is granted to copy and distribute translations of this manual
-into another language, under the above conditions for modified versions.
-@end ifinfo
+@c man end
+@end copying
@finalout
@smallbook
@titlepage
@title GNU gprof
-@subtitle The @sc{gnu} Profiler
+@subtitle The @sc{gnu} Profiler
+@ifset VERSION_PACKAGE
+@subtitle @value{VERSION_PACKAGE}
+@end ifset
+@subtitle Version @value{VERSION}
@author Jay Fenlason and Richard Stallman
@page
can use it to determine which parts of a program are taking most of the
execution time. We assume that you know how to write, compile, and
execute programs. @sc{gnu} @code{gprof} was written by Jay Fenlason.
+Eric S. Raymond made some minor corrections and additions in 2003.
@vskip 0pt plus 1filll
-Copyright @copyright{} 1988, 92, 97, 98, 99, 2000 Free Software Foundation, Inc.
-
-Permission is granted to make and distribute verbatim copies of
-this manual provided the copyright notice and this permission notice
-are preserved on all copies.
+Copyright @copyright{} 1988-2020 Free Software Foundation, Inc.
-@ignore
-Permission is granted to process this file through TeX and print the
-results, provided the printed document carries copying permission
-notice identical to this one except for the removal of this paragraph
-(this paragraph not being relevant to the printed manual).
-
-@end ignore
-Permission is granted to copy and distribute modified versions of this
-manual under the conditions for verbatim copying, provided that the entire
-resulting derived work is distributed under the terms of a permission
-notice identical to this one.
-
-Permission is granted to copy and distribute translations of this manual
-into another language, under the same conditions as for modified versions.
+ 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 no Invariant Sections, with no Front-Cover Texts, and with no
+ Back-Cover Texts. A copy of the license is included in the
+ section entitled ``GNU Free Documentation License''.
@end titlepage
+@contents
-@ifinfo
+@ifnottex
@node Top
@top Profiling a Program: Where Does It Spend Its Time?
execution time. We assume that you know how to write, compile, and
execute programs. @sc{gnu} @code{gprof} was written by Jay Fenlason.
+This manual is for @code{gprof}
+@ifset VERSION_PACKAGE
+@value{VERSION_PACKAGE}
+@end ifset
+version @value{VERSION}.
+
+This document is distributed under the terms of the GNU Free
+Documentation License version 1.3. A copy of the license is included
+in the section entitled ``GNU Free Documentation License''.
+
@menu
* Introduction:: What profiling means, and why it is useful.
* Executing:: Executing your program to generate profile data
* Invoking:: How to run @code{gprof}, and its options
-* Output:: Interpreting @code{gprof}'s output
+* Output:: Interpreting @code{gprof}'s output
* Inaccuracy:: Potential problems you should be aware of
* How do I?:: Answers to common questions
* Incompatibilities:: (between @sc{gnu} @code{gprof} and Unix @code{gprof}.)
* Details:: Details of how profiling is done
+* GNU Free Documentation License:: GNU Free Documentation License
@end menu
-@end ifinfo
+@end ifnottex
@node Introduction
@chapter Introduction to Profiling
+@ifset man
+@c man title gprof display call graph profile data
+
+@smallexample
+@c man begin SYNOPSIS
+gprof [ -[abcDhilLrsTvwxyz] ] [ -[ACeEfFJnNOpPqQRStZ][@var{name}] ]
+ [ -I @var{dirs} ] [ -d[@var{num}] ] [ -k @var{from/to} ]
+ [ -m @var{min-count} ] [ -R @var{map_file} ] [ -t @var{table-length} ]
+ [ --[no-]annotated-source[=@var{name}] ]
+ [ --[no-]exec-counts[=@var{name}] ]
+ [ --[no-]flat-profile[=@var{name}] ] [ --[no-]graph[=@var{name}] ]
+ [ --[no-]time=@var{name}] [ --all-lines ] [ --brief ]
+ [ --debug[=@var{level}] ] [ --function-ordering ]
+ [ --file-ordering @var{map_file} ] [ --directory-path=@var{dirs} ]
+ [ --display-unused-functions ] [ --file-format=@var{name} ]
+ [ --file-info ] [ --help ] [ --line ] [ --inline-file-names ]
+ [ --min-count=@var{n} ] [ --no-static ] [ --print-path ]
+ [ --separate-files ] [ --static-call-graph ] [ --sum ]
+ [ --table-length=@var{len} ] [ --traditional ] [ --version ]
+ [ --width=@var{n} ] [ --ignore-non-functions ]
+ [ --demangle[=@var{STYLE}] ] [ --no-demangle ]
+ [--external-symbol-table=name]
+ [ @var{image-file} ] [ @var{profile-file} @dots{} ]
+@c man end
+@end smallexample
+
+@c man begin DESCRIPTION
+@code{gprof} produces an execution profile of C, Pascal, or Fortran77
+programs. The effect of called routines is incorporated in the profile
+of each caller. The profile data is taken from the call graph profile file
+(@file{gmon.out} default) which is created by programs
+that are compiled with the @samp{-pg} option of
+@code{cc}, @code{pc}, and @code{f77}.
+The @samp{-pg} option also links in versions of the library routines
+that are compiled for profiling. @code{Gprof} reads the given object
+file (the default is @code{a.out}) and establishes the relation between
+its symbol table and the call graph profile from @file{gmon.out}.
+If more than one profile file is specified, the @code{gprof}
+output shows the sum of the profile information in the given profile files.
+
+@code{Gprof} calculates the amount of time spent in each routine.
+Next, these times are propagated along the edges of the call graph.
+Cycles are discovered, and calls into a cycle are made to share the time
+of the cycle.
+
+@c man end
+
+@c man begin BUGS
+The granularity of the sampling is shown, but remains
+statistical at best.
+We assume that the time for each execution of a function
+can be expressed by the total time for the function divided
+by the number of times the function is called.
+Thus the time propagated along the call graph arcs to the function's
+parents is directly proportional to the number of times that
+arc is traversed.
+
+Parents that are not themselves profiled will have the time of
+their profiled children propagated to them, but they will appear
+to be spontaneously invoked in the call graph listing, and will
+not have their time propagated further.
+Similarly, signal catchers, even though profiled, will appear
+to be spontaneous (although for more obscure reasons).
+Any profiled children of signal catchers should have their times
+propagated properly, unless the signal catcher was invoked during
+the execution of the profiling routine, in which case all is lost.
+
+The profiled program must call @code{exit}(2)
+or return normally for the profiling information to be saved
+in the @file{gmon.out} file.
+@c man end
+
+@c man begin FILES
+@table @code
+@item @file{a.out}
+the namelist and text space.
+@item @file{gmon.out}
+dynamic call graph and profile.
+@item @file{gmon.sum}
+summarized dynamic call graph and profile.
+@end table
+@c man end
+
+@c man begin SEEALSO
+monitor(3), profil(2), cc(1), prof(1), and the Info entry for @file{gprof}.
+
+``An Execution Profiler for Modular Programs'',
+by S. Graham, P. Kessler, M. McKusick;
+Software - Practice and Experience,
+Vol. 13, pp. 671-685, 1983.
+
+``gprof: A Call Graph Execution Profiler'',
+by S. Graham, P. Kessler, M. McKusick;
+Proceedings of the SIGPLAN '82 Symposium on Compiler Construction,
+SIGPLAN Notices, Vol. 17, No 6, pp. 120-126, June 1982.
+@c man end
+@end ifset
+
Profiling allows you to learn where your program spent its time and which
functions called which other functions while it was executing. This
information can show you which pieces of your program are slower than you
@itemize @bullet
@item
You must compile and link your program with profiling enabled.
-@xref{Compiling}.
+@xref{Compiling, ,Compiling a Program for Profiling}.
@item
You must execute your program to generate a profile data file.
-@xref{Executing}.
+@xref{Executing, ,Executing the Program}.
@item
You must run @code{gprof} to analyze the profile data.
-@xref{Invoking}.
+@xref{Invoking, ,@code{gprof} Command Summary}.
@end itemize
The next three chapters explain these steps in greater detail.
+@c man begin DESCRIPTION
+
Several forms of output are available from the analysis.
The @dfn{flat profile} shows how much time your program spent in each function,
and how many times that function was called. If you simply want to know
which functions burn most of the cycles, it is stated concisely here.
-@xref{Flat Profile}.
+@xref{Flat Profile, ,The Flat Profile}.
The @dfn{call graph} shows, for each function, which functions called it, which
other functions it called, and how many times. There is also an estimate
of how much time was spent in the subroutines of each function. This can
suggest places where you might try to eliminate function calls that use a
-lot of time. @xref{Call Graph}.
+lot of time. @xref{Call Graph, ,The Call Graph}.
The @dfn{annotated source} listing is a copy of the program's
source code, labeled with the number of times each line of the
-program was executed. @xref{Annotated Source}.
+program was executed. @xref{Annotated Source, ,The Annotated Source
+Listing}.
+@c man end
To better understand how profiling works, you may wish to read
a description of its implementation.
-@xref{Implementation}.
+@xref{Implementation, ,Implementation of Profiling}.
@node Compiling
@chapter Compiling a Program for Profiling
cc -o myprog myprog.c utils.c -g -pg
@end example
+Note: The @samp{-pg} option must be part of your compilation options
+as well as your link options. If it is not then no call-graph data
+will be gathered and when you run @code{gprof} you will get an error
+message like this:
+
+@example
+gprof: gmon.out file is missing call-graph data
+@end example
+
+If you add the @samp{-Q} switch to suppress the printing of the call
+graph data you will still be able to see the time samples:
+
+@example
+Flat profile:
+
+Each sample counts as 0.01 seconds.
+ % cumulative self self total
+ time seconds seconds calls Ts/call Ts/call name
+ 44.12 0.07 0.07 zazLoop
+ 35.29 0.14 0.06 main
+ 20.59 0.17 0.04 bazMillion
+@end example
+
If you run the linker @code{ld} directly instead of through a compiler
such as @code{cc}, you may have to specify a profiling startup file
@file{gcrt0.o} as the first input file instead of the usual startup
ld -o myprog /lib/gcrt0.o myprog.o utils.o -lc_p
@end example
+If you are running the program on a system which supports shared
+libraries you may run into problems with the profiling support code in
+a shared library being called before that library has been fully
+initialised. This is usually detected by the program encountering a
+segmentation fault as soon as it is run. The solution is to link
+against a static version of the library containing the profiling
+support code, which for @code{gcc} users can be done via the
+@samp{-static} or @samp{-static-libgcc} command-line option. For
+example:
+
+@example
+gcc -g -pg -static-libgcc myprog.c utils.c -o myprog
+@end example
+
If you compile only some of the modules of the program with @samp{-pg}, you
can still profile the program, but you won't get complete information about
the modules that were compiled without @samp{-pg}. The only information
the functions will be blank), but will greatly reduce the usefulness of the
call graph.
-If you wish to perform line-by-line profiling,
-you will also need to specify the @samp{-g} option,
-instructing the compiler to insert debugging symbols into the program
-that match program addresses to source code lines.
-@xref{Line-by-line}.
-
-In addition to the @samp{-pg} and @samp{-g} options,
-you may also wish to specify the @samp{-a} option when compiling.
-This will instrument
-the program to perform basic-block counting. As the program runs,
-it will count how many times it executed each branch of each @samp{if}
-statement, each iteration of each @samp{do} loop, etc. This will
-enable @code{gprof} to construct an annotated source code
-listing showing how many times each line of code was executed.
+If you wish to perform line-by-line profiling you should use the
+@code{gcov} tool instead of @code{gprof}. See that tool's manual or
+info pages for more details of how to do this.
+
+Note, older versions of @code{gcc} produce line-by-line profiling
+information that works with @code{gprof} rather than @code{gcov} so
+there is still support for displaying this kind of information in
+@code{gprof}. @xref{Line-by-line, ,Line-by-line Profiling}.
+
+It also worth noting that @code{gcc} implements a
+@samp{-finstrument-functions} command-line option which will insert
+calls to special user supplied instrumentation routines at the entry
+and exit of every function in their program. This can be used to
+implement an alternative profiling scheme.
@node Executing
@chapter Executing the Program
generate the information that @code{gprof} needs. Simply run the program
as usual, using the normal arguments, file names, etc. The program should
run normally, producing the same output as usual. It will, however, run
-somewhat slower than normal because of the time spent collecting and the
+somewhat slower than normal because of the time spent collecting and
writing the profile data.
The way you run the program---the arguments and input that you give
just before exiting. If there is already a file called @file{gmon.out},
its contents are overwritten. There is currently no way to tell the
program to write the profile data under a different name, but you can rename
-the file afterward if you are concerned that it may be overwritten.
+the file afterwards if you are concerned that it may be overwritten.
In order to write the @file{gmon.out} file properly, your program must exit
normally: by returning from @code{main} or by calling @code{exit}. Calling
counts didn't get written into @file{gmon.out}.
The Perl script @code{bbconv.pl}, included with the @code{gprof}
source distribution, will convert a @file{bb.out} file into
-a format readable by @code{gprof}.
+a format readable by @code{gprof}. Invoke it like this:
+
+@smallexample
+bbconv.pl < bb.out > @var{bh-data}
+@end smallexample
+
+This translates the information in @file{bb.out} into a form that
+@code{gprof} can understand. But you still need to tell @code{gprof}
+about the existence of this translated information. To do that, include
+@var{bb-data} on the @code{gprof} command line, @emph{along with
+@file{gmon.out}}, like this:
+
+@smallexample
+gprof @var{options} @var{executable-file} gmon.out @var{bb-data} [@var{yet-more-profile-data-files}@dots{}] [> @var{outfile}]
+@end smallexample
@node Invoking
@chapter @code{gprof} Command Summary
@menu
* Output Options:: Controlling @code{gprof}'s output style
-* Analysis Options:: Controlling how @code{gprof} analyses its data
+* Analysis Options:: Controlling how @code{gprof} analyzes its data
* Miscellaneous Options::
* Deprecated Options:: Options you no longer need to use, but which
have been retained for compatibility
* Symspecs:: Specifying functions to include or exclude
@end menu
-@node Output Options,Analysis Options,,Invoking
+@node Output Options
@section Output Options
+@c man begin OPTIONS
These options specify which of several output formats
@code{gprof} should produce.
Many of these options take an optional @dfn{symspec} to specify
functions to be included or excluded. These options can be
specified multiple times, with different symspecs, to include
-or exclude sets of symbols. @xref{Symspecs}.
+or exclude sets of symbols. @xref{Symspecs, ,Symspecs}.
Specifying any of these options overrides the default (@samp{-p -q}),
which prints a flat profile and call graph analysis
@itemx --annotated-source[=@var{symspec}]
The @samp{-A} option causes @code{gprof} to print annotated source code.
If @var{symspec} is specified, print output only for matching symbols.
-@xref{Annotated Source}.
+@xref{Annotated Source, ,The Annotated Source Listing}.
@item -b
@itemx --brief
@itemx --flat-profile[=@var{symspec}]
The @samp{-p} option causes @code{gprof} to print a flat profile.
If @var{symspec} is specified, print flat profile only for matching symbols.
-@xref{Flat Profile}.
+@xref{Flat Profile, ,The Flat Profile}.
@item -P[@var{symspec}]
@itemx --no-flat-profile[=@var{symspec}]
The @samp{-q} option causes @code{gprof} to print the call graph analysis.
If @var{symspec} is specified, print call graph only for matching symbols
and their children.
-@xref{Call Graph}.
+@xref{Call Graph, ,The Call Graph}.
@item -Q[@var{symspec}]
@itemx --no-graph[=@var{symspec}]
If @var{symspec} is specified, @code{gprof} prints a call graph,
but excludes matching symbols.
+@item -t
+@itemx --table-length=@var{num}
+The @samp{-t} option causes the @var{num} most active source lines in
+each source file to be listed when source annotation is enabled. The
+default is 10.
+
@item -y
@itemx --separate-files
This option affects annotated source output only.
to standard-output. If this option is specified,
annotated source for a file named @file{path/@var{filename}}
is generated in the file @file{@var{filename}-ann}. If the underlying
-filesystem would truncate @file{@var{filename}-ann} so that it
+file system would truncate @file{@var{filename}-ann} so that it
overwrites the original @file{@var{filename}}, @code{gprof} generates
annotated source in the file @file{@var{filename}.ann} instead (if the
original file name has an extension, that extension is @emph{replaced}
print a tally of functions and the number of times each was called.
If @var{symspec} is specified, print tally, but exclude matching symbols.
-@item --function-ordering
+@item -r
+@itemx --function-ordering
The @samp{--function-ordering} option causes @code{gprof} to print a
suggested function ordering for the program based on profiling data.
This option suggests an ordering which may improve paging, tlb and
in a particular order is system dependent and out of the scope of this
manual.
-@item --file-ordering @var{map_file}
+@item -R @var{map_file}
+@itemx --file-ordering @var{map_file}
The @samp{--file-ordering} option causes @code{gprof} to print a
suggested .o link line ordering for the program based on profiling data.
This option suggests an ordering which may improve paging, tlb and
@itemx --no-demangle
These options control whether C++ symbol names should be demangled when
printing output. The default is to demangle symbols. The
-@code{--no-demangle} option may be used to turn off demangling. Different
-compilers have different mangling styles. The optional demangling style
-argument can be used to choose an appropriate demangling style for your
+@code{--no-demangle} option may be used to turn off demangling. Different
+compilers have different mangling styles. The optional demangling style
+argument can be used to choose an appropriate demangling style for your
compiler.
@end table
-@node Analysis Options,Miscellaneous Options,Output Options,Invoking
+@node Analysis Options
@section Analysis Options
@table @code
statically declared (private) functions. (These are functions whose
names are not listed as global, and which are not visible outside the
file/function/block where they were defined.) Time spent in these
-functions, calls to/from them, etc, will all be attributed to the
+functions, calls to/from them, etc., will all be attributed to the
function that was loaded directly before it in the executable file.
-@c This is compatible with Unix @code{gprof}, but a bad idea.
+@c This is compatible with Unix @code{gprof}, but a bad idea.
This option affects both the flat profile and the call graph.
@item -c
@itemx --line
The @samp{-l} option enables line-by-line profiling, which causes
histogram hits to be charged to individual source code lines,
-instead of functions.
+instead of functions. This feature only works with programs compiled
+by older versions of the @code{gcc} compiler. Newer versions of
+@code{gcc} are designed to work with the @code{gcov} tool instead.
+
If the program was compiled with basic-block counting enabled,
this option will also identify how many times each line of
code was executed.
a program is spending its time, it also significantly increases
the running time of @code{gprof}, and magnifies statistical
inaccuracies.
-@xref{Sampling Error}.
+@xref{Sampling Error, ,Statistical Sampling Error}.
+
+@item --inline-file-names
+This option causes @code{gprof} to print the source file after each
+symbol in both the flat profile and the call graph. The full path to the
+file is printed if used with the @samp{-L} option.
@item -m @var{num}
@itemx --min-count=@var{num}
This option affects execution count output only.
Symbols that are executed less than @var{num} times are suppressed.
-@item -n[@var{symspec}]
-@itemx --time[=@var{symspec}]
+@item -n@var{symspec}
+@itemx --time=@var{symspec}
The @samp{-n} option causes @code{gprof}, in its call graph analysis,
to only propagate times for symbols matching @var{symspec}.
-@item -N[@var{symspec}]
-@itemx --no-time[=@var{symspec}]
+@item -N@var{symspec}
+@itemx --no-time=@var{symspec}
The @samp{-n} option causes @code{gprof}, in its call graph analysis,
not to propagate times for symbols matching @var{symspec}.
+@item -S@var{filename}
+@itemx --external-symbol-table=@var{filename}
+The @samp{-S} option causes @code{gprof} to read an external symbol table
+file, such as @file{/proc/kallsyms}, rather than read the symbol table
+from the given object file (the default is @code{a.out}). This is useful
+for profiling kernel modules.
+
@item -z
@itemx --display-unused-functions
If you give the @samp{-z} option, @code{gprof} will mention all
@end table
-@node Miscellaneous Options,Deprecated Options,Analysis Options,Invoking
+@node Miscellaneous Options
@section Miscellaneous Options
@table @code
@itemx --debug[=@var{num}]
The @samp{-d @var{num}} option specifies debugging options.
If @var{num} is not specified, enable all debugging.
-@xref{Debugging}.
+@xref{Debugging, ,Debugging @code{gprof}}.
+
+@item -h
+@itemx --help
+The @samp{-h} option prints command line usage.
@item -O@var{name}
@itemx --file-format=@var{name}
@end table
-@node Deprecated Options,Symspecs,Miscellaneous Options,Invoking
+@node Deprecated Options
@section Deprecated Options
-@table @code
-
These options have been replaced with newer versions that use symspecs.
+@table @code
+
@item -e @var{function_name}
The @samp{-e @var{function}} option tells @code{gprof} to not print
information about the function @var{function_name} (and its
as a child of any functions that call it, but its index number will be
shown as @samp{[not printed]}. More than one @samp{-e} option may be
given; only one @var{function_name} may be indicated with each @samp{-e}
-option.
+option.
@item -E @var{function_name}
The @code{-E @var{function}} option works like the @code{-e} option, but
call graph to the function @var{function_name} and its children (and
their children@dots{}). More than one @samp{-f} option may be given;
only one @var{function_name} may be indicated with each @samp{-f}
-option.
+option.
@item -F @var{function_name}
The @samp{-F @var{function}} option works like the @code{-f} option, but
@end table
+@c man end
+
Note that only one function can be specified with each @code{-e},
@code{-E}, @code{-f} or @code{-F} option. To specify more than one
function, use multiple options. For example, this command:
lists in the call graph all functions that were reached from either
@code{foo} or @code{bar} and were not reachable from @code{boring}.
-@node Symspecs,,Deprecated Options,Invoking
+@node Symspecs
@section Symspecs
Many of the output options allow functions to be included or excluded
styles (file information, execution count, and function and file ordering)
are not described here, but are documented with the respective options
that trigger them.
-@xref{Output Options}.
+@xref{Output Options, ,Output Options}.
@menu
* Flat Profile:: The flat profile shows how much time was spent
@end menu
-@node Flat Profile,Call Graph,,Output
+@node Flat Profile
@section The Flat Profile
@cindex flat profile
Flat profile:
Each sample counts as 0.01 seconds.
- % cumulative self self total
- time seconds seconds calls ms/call ms/call name
+ % cumulative self self total
+ time seconds seconds calls ms/call ms/call name
33.34 0.02 0.02 7208 0.00 0.00 open
16.67 0.03 0.01 244 0.04 0.12 offtime
16.67 0.04 0.01 8 1.25 1.25 memccpy
@end smallexample
@noindent
-The functions are sorted by first by decreasing run-time spent in them,
+The functions are sorted first by decreasing run-time spent in them,
then by decreasing number of calls, then alphabetically by name. The
functions @samp{mcount} and @samp{profil} are part of the profiling
apparatus and appear in every flat profile; their time gives a measure of
In another run,
the @samp{self seconds} field for
@samp{mcount} might well be @samp{0.00} or @samp{0.02}.
-@xref{Sampling Error}, for a complete discussion.
+@xref{Sampling Error, ,Statistical Sampling Error},
+for a complete discussion.
The remaining functions in the listing (those whose
@samp{self seconds} field is @samp{0.00}) didn't appear
fields are sorted.
@end table
-@node Call Graph,Line-by-line,Flat Profile,Output
+@node Call Graph
@section The Call Graph
@cindex call graph
Here is a sample call from a small program. This call came from the
same @code{gprof} run as the flat profile example in the previous
-chapter.
+section.
@smallexample
@group
0.00 0.00 8/8 chewtime [24]
0.00 0.00 8/16 skipspace [44]
-----------------------------------------------
-[4] 59.8 0.01 0.02 8+472 <cycle 2 as a whole> [4]
+[4] 59.8 0.01 0.02 8+472 <cycle 2 as a whole> [4]
0.01 0.02 244+260 offtime <cycle 2> [7]
0.00 0.00 236+1 tzset <cycle 2> [26]
-----------------------------------------------
The entries are sorted by time spent in the function and its subroutines.
-The internal profiling function @code{mcount} (@pxref{Flat Profile})
-is never mentioned in the call graph.
+The internal profiling function @code{mcount} (@pxref{Flat Profile, ,The
+Flat Profile}) is never mentioned in the call graph.
@menu
* Primary:: Details of the primary line's contents.
If the function is part of a cycle of recursion, the cycle number is
printed between the function's name and the index number
-(@pxref{Cycles}). For example, if function @code{gnurr} is part of
+(@pxref{Cycles, ,How Mutually Recursive Functions Are Described}).
+For example, if function @code{gnurr} is part of
cycle number one, and has index number twelve, its primary line would
be end like this:
@end example
@end table
-@node Callers, Subroutines, Primary, Call Graph
+@node Callers
@subsection Lines for a Function's Callers
A function's entry has a line for each function it was called by.
@c What if some calls have determinable callers' names but not all?
@c FIXME - still relevant?
-@node Subroutines, Cycles, Callers, Call Graph
+@node Subroutines
@subsection Lines for a Function's Subroutines
A function's entry has a line for each of its subroutines---in other
followed by the total number of non-recursive calls to @code{report}.
This ratio is used to determine how much of @code{report}'s @code{self}
and @code{children} time gets credited to @code{main}.
-@xref{Assumptions}.
+@xref{Assumptions, ,Estimating @code{children} Times}.
@item name
The name of the subroutine of @code{main} to which this line applies,
printed between the name and the index number.
@end table
-@node Cycles,, Subroutines, Call Graph
+@node Cycles
@subsection How Mutually Recursive Functions Are Described
@cindex cycle
@cindex recursion cycle
the amount of time spent @emph{in the whole cycle}, and its other
subroutines, on the times when that caller called a function in the cycle.
-The @code{calls} field in the primary line for the cycle has two numbers:
+The @code{called} field in the primary line for the cycle has two numbers:
first, the number of times functions in the cycle were called by functions
outside the cycle; second, the number of times they were called by
functions in the cycle (including times when a function in the cycle calls
itself). This is a generalization of the usual split into non-recursive and
recursive calls.
-The @code{calls} field of a subroutine-line for a cycle member in the
+The @code{called} field of a subroutine-line for a cycle member in the
cycle's entry says how many time that function was called from functions in
the cycle. The total of all these is the second number in the primary line's
-@code{calls} field.
+@code{called} field.
In the individual entry for a function in a cycle, the other functions in
the same cycle can appear as subroutines and as callers. These lines show
lines are blank because of the difficulty of defining meanings for them
when recursion is going on.
-@node Line-by-line,Annotated Source,Call Graph,Output
+@node Line-by-line
@section Line-by-line Profiling
@code{gprof}'s @samp{-l} option causes the program to perform
@dfn{line-by-line} profiling. In this mode, histogram
samples are assigned not to functions, but to individual
-lines of source code. The program usually must be compiled
-with a @samp{-g} option, in addition to @samp{-pg}, in order
+lines of source code. This only works with programs compiled with
+older versions of the @code{gcc} compiler. Newer versions of @code{gcc}
+use a different program - @code{gcov} - to display line-by-line
+profiling information.
+
+With the older versions of @code{gcc} the program usually has to be
+compiled with a @samp{-g} option, in addition to @samp{-pg}, in order
to generate debugging symbols for tracking source code lines.
+Note, in much older versions of @code{gcc} the program had to be
+compiled with the @samp{-a} command-line option as well.
The flat profile is the most useful output table
in line-by-line mode.
Flat profile:
Each sample counts as 0.01 seconds.
- % cumulative self self total
- time seconds seconds calls us/call us/call name
+ % cumulative self self total
+ time seconds seconds calls us/call us/call name
30.77 0.13 0.04 6335 6.31 6.31 ct_init
Now let's look at some of @code{gprof}'s output from the same program run,
this time with line-by-line profiling enabled. Note that @code{ct_init}'s
-four histogram hits are broken down into four lines of source code - one hit
+four histogram hits are broken down into four lines of source code---one hit
occurred on each of lines 349, 351, 382 and 385. In the call graph,
note how
@code{ct_init}'s 13327 calls to @code{init_block} are broken down
Flat profile:
Each sample counts as 0.01 seconds.
- % cumulative self
- time seconds seconds calls name
+ % cumulative self
+ time seconds seconds calls name
7.69 0.10 0.01 ct_init (trees.c:349)
7.69 0.11 0.01 ct_init (trees.c:351)
7.69 0.12 0.01 ct_init (trees.c:382)
@end smallexample
-@node Annotated Source,,Line-by-line,Output
+@node Annotated Source
@section The Annotated Source Listing
@code{gprof}'s @samp{-A} option triggers an annotated source listing,
number of times it was called. You may also need to specify the
@samp{-I} option, if @code{gprof} can't find the source code files.
-Compiling with @samp{gcc @dots{} -g -pg -a} augments your program
-with basic-block counting code, in addition to function counting code.
-This enables @code{gprof} to determine how many times each line
-of code was executed.
+With older versions of @code{gcc} compiling with @samp{gcc @dots{} -g
+-pg -a} augments your program with basic-block counting code, in
+addition to function counting code. This enables @code{gprof} to
+determine how many times each line of code was executed. With newer
+versions of @code{gcc} support for displaying basic-block counts is
+provided by the @code{gcov} program.
+
For example, consider the following function, taken from gzip,
with line numbers added:
basic-blocks to handle various special cases.
A program augmented for basic-block counting can be analyzed with
-@samp{gprof -l -A}. I also suggest use of the @samp{-x} option,
-which ensures that each line of code is labeled at least once.
+@samp{gprof -l -A}.
+The @samp{-x} option is also helpful,
+to ensure that each line of code is labeled at least once.
Here is @code{updcrc}'s
annotated source listing for a sample @code{gzip} run:
unsigned n;
2 ->@{
register ulg c;
-
+
static ulg crc = (ulg)0xffffffffL;
-
+
2 -> if (s == NULL) @{
- 1 -> c = 0xffffffffL;
+ 1 -> c = 0xffffffffL;
1 -> @} else @{
- 1 -> c = crc;
+ 1 -> c = crc;
1 -> if (n) do @{
26312 -> c = crc_32_tab[...];
26312,1,26311 -> @} while (--n);
* Assumptions:: Estimating children times
@end menu
-@node Sampling Error,Assumptions,,Inaccuracy
+@node Sampling Error
@section Statistical Sampling Error
The run-time figures that @code{gprof} gives you are based on a sampling
ought to catch that function in the act only once, there is a pretty good
chance it will actually find that function zero times, or twice.
-By contrast, the number-of-calls and basic-block figures
-are derived by counting, not
-sampling. They are completely accurate and will not vary from run to run
-if your program is deterministic.
+By contrast, the number-of-calls and basic-block figures are derived
+by counting, not sampling. They are completely accurate and will not
+vary from run to run if your program is deterministic and single
+threaded. In multi-threaded applications, or single threaded
+applications that link with multi-threaded libraries, the counts are
+only deterministic if the counting function is thread-safe. (Note:
+beware that the mcount counting function in glibc is @emph{not}
+thread-safe). @xref{Implementation, ,Implementation of Profiling}.
The @dfn{sampling period} that is printed at the beginning of the flat
profile says how often samples are taken. The rule of thumb is that a
@end example
@end enumerate
-@node Assumptions,,Sampling Error,Inaccuracy
+@node Assumptions
@section Estimating @code{children} Times
Some of the figures in the call graph are estimates---for example, the
-@code{children} time values and all the the time figures in caller and
+@code{children} time values and all the time figures in caller and
subroutine lines.
There is no direct information about these measurements in the profile
@c FIXME - has this been fixed?
We hope some day to put more complete data into @file{gmon.out}, so that
this assumption is no longer needed, if we can figure out how. For the
-nonce, the estimated figures are usually more useful than misleading.
+novice, the estimated figures are usually more useful than misleading.
@node How do I?
@chapter Answers to Common Questions
@table @asis
-@item How do I find which lines in my program were executed the most times?
+@item How can I get more exact information about hot spots in my program?
-Compile your program with basic-block counting enabled, run it, then
-use the following pipeline:
+Looking at the per-line call counts only tells part of the story.
+Because @code{gprof} can only report call times and counts by function,
+the best way to get finer-grained information on where the program
+is spending its time is to re-factor large functions into sequences
+of calls to smaller ones. Beware however that this can introduce
+artificial hot spots since compiling with @samp{-pg} adds a significant
+overhead to function calls. An alternative solution is to use a
+non-intrusive profiler, e.g.@: oprofile.
-@example
-gprof -l -C @var{objfile} | sort -k 3 -n -r
-@end example
+@item How do I find which lines in my program were executed the most times?
-This listing will show you the lines in your code executed most often,
-but not necessarily those that consumed the most time.
+Use the @code{gcov} program.
@item How do I find which lines in my program called a particular function?
@end example
If your program is completely deterministic, all the call counts
-will be simple multiples of 100 (i.e. a function called once in
+will be simple multiples of 100 (i.e., a function called once in
each run will appear with a call count of 100).
@end table
for basic-block execution counts and non-realtime histograms. A magic
cookie and version number allows @code{gprof} to easily identify
new style files. Old BSD-style files can still be read.
-@xref{File Format}.
+@xref{File Format, ,Profiling Data File Format}.
@item
For a recursive function, Unix @code{gprof} lists the function as a
* Debugging:: Using @code{gprof}'s @samp{-d} option
@end menu
-@node Implementation,File Format,,Details
+@node Implementation
@section Implementation of Profiling
Profiling works by changing how every function in your program is compiled
itself is typically a short assembly-language stub routine
that extracts the required
information, and then calls @code{__mcount_internal}
-(a normal C function) with two arguments - @code{frompc} and @code{selfpc}.
+(a normal C function) with two arguments---@code{frompc} and @code{selfpc}.
@code{__mcount_internal} is responsible for maintaining
the in-memory call graph, which records @code{frompc}, @code{selfpc},
and the number of times each of these call arcs was traversed.
added delay required to deliver the signal, this method is
less accurate as well.
-A special startup routine allocates memory for the histogram and
+A special startup routine allocates memory for the histogram and
either calls @code{profil()} or sets up
a clock signal handler.
This routine (@code{monstartup}) can be invoked in several ways.
is also enabled. Each object file is then compiled with a static array
of counts, initially zero.
In the executable code, every time a new basic-block begins
-(i.e. when an @code{if} statement appears), an extra instruction
+(i.e., when an @code{if} statement appears), an extra instruction
is inserted to increment the corresponding count in the array.
At compile time, a paired array was constructed that recorded
the starting address of each basic-block. Taken together,
the other hand, sampling by run time has the advantage that the amount of
load due to other users won't directly affect the output you get.
-@node File Format,Internals,Implementation,Details
+@node File Format
@section Profiling Data File Format
The old BSD-derived file format used for profile data does not contain a
The new file format is defined in header file @file{gmon_out.h}. It
consists of a header containing the magic cookie and a version number,
as well as some spare bytes available for future extensions. All data
-in a profile data file is in the native format of the host on which
-the profile was collected. @sc{gnu} @code{gprof} adapts automatically to the
-byte-order in use.
+in a profile data file is in the native format of the target for which
+the profile was collected. @sc{gnu} @code{gprof} adapts automatically
+to the byte-order in use.
In the new file format, the header is followed by a sequence of
records. Currently, there are three different record types: histogram
physical dimension is specified in two parts: a long name of up to 15
characters and a single character abbreviation. For example, a
histogram representing real-time would specify the long name as
-"seconds" and the abbreviation as "s". This feature is useful for
+``seconds'' and the abbreviation as ``s''. This feature is useful for
architectures that support performance monitor hardware (which,
fortunately, is becoming increasingly common). For example, under DEC
-OSF/1, the "uprofile" command can be used to produce a histogram of,
+OSF/1, the ``uprofile'' command can be used to produce a histogram of,
say, instruction cache misses. In this case, the dimension in the
-histogram header could be set to "i-cache misses" and the abbreviation
-could be set to "1" (because it is simply a count, not a physical
+histogram header could be set to ``i-cache misses'' and the abbreviation
+could be set to ``1'' (because it is simply a count, not a physical
dimension). Also, the profiling rate would have to be set to 1 in
this case.
that basic-block was executed. Any address within the basic-address can
be used.
-@node Internals,Debugging,File Format,Details
+@node Internals
@section @code{gprof}'s Internal Operation
Like most programs, @code{gprof} begins by processing its options.
During this stage, it may building its symspec list
-(@code{sym_ids.c:sym_id_add}), if
+(@code{sym_ids.c:@-sym_id_add}), if
options are specified which use symspecs.
@code{gprof} maintains a single linked list of symspecs,
which will eventually get turned into 12 symbol tables,
-organized into six include/exclude pairs - one
+organized into six include/exclude pairs---one
pair each for the flat profile (INCL_FLAT/EXCL_FLAT),
the call graph arcs (INCL_ARCS/EXCL_ARCS),
printing in the call graph (INCL_GRAPH/EXCL_GRAPH),
Next, the BFD library is called to open the object file,
verify that it is an object file,
-and read its symbol table (@code{core.c:core_init}),
+and read its symbol table (@code{core.c:@-core_init}),
using @code{bfd_canonicalize_symtab} after mallocing
an appropriately sized array of symbols. At this point,
function mappings are read (if the @samp{--file-ordering} option
text space address is examined, and a new symbol table entry
gets created every time the line number changes.
In either case, two passes are made through the symbol
-table - one to count the size of the symbol table required,
+table---one to count the size of the symbol table required,
and the other to actually read the symbols. In between the
two passes, a single array of type @code{Sym} is created of
the appropriate length.
-Finally, @code{symtab.c:symtab_finalize}
+Finally, @code{symtab.c:@-symtab_finalize}
is called to sort the symbol table and remove duplicate entries
(entries with the same memory address).
The symbol table must be a contiguous array for two reasons.
First, the @code{qsort} library function (which sorts an array)
will be used to sort the symbol table.
-Also, the symbol lookup routine (@code{symtab.c:sym_lookup}),
+Also, the symbol lookup routine (@code{symtab.c:@-sym_lookup}),
which finds symbols
based on memory address, uses a binary search algorithm
which requires the symbol table to be a sorted array.
to indicate that it is a local symbol.
With the symbol table read, the symspecs can now be translated
-into Syms (@code{sym_ids.c:sym_id_parse}). Remember that a single
+into Syms (@code{sym_ids.c:@-sym_id_parse}). Remember that a single
symspec can match multiple symbols.
An array of symbol tables
(@code{syms}) is created, each entry of which is a symbol table
in the @code{syms} array.
Now the profile data file(s) themselves are read
-(@code{gmon_io.c:gmon_out_read}),
+(@code{gmon_io.c:@-gmon_out_read}),
first by checking for a new-style @samp{gmon.out} header,
then assuming this is an old-style BSD @samp{gmon.out}
if the magic number test failed.
-New-style histogram records are read by @code{hist.c:hist_read_rec}.
+New-style histogram records are read by @code{hist.c:@-hist_read_rec}.
For the first histogram record, allocate a memory array to hold
all the bins, and read them in.
When multiple profile data files (or files with multiple histogram
-records) are read, the starting address, ending address, number
-of bins and sampling rate must match between the various histograms,
-or a fatal error will result.
-If everything matches, just sum the additional histograms into
-the existing in-memory array.
-
-As each call graph record is read (@code{call_graph.c:cg_read_rec}),
+records) are read, the memory ranges of each pair of histogram records
+must be either equal, or non-overlapping. For each pair of histogram
+records, the resolution (memory region size divided by the number of
+bins) must be the same. The time unit must be the same for all
+histogram records. If the above containts are met, all histograms
+for the same memory range are merged.
+
+As each call graph record is read (@code{call_graph.c:@-cg_read_rec}),
the parent and child addresses
are matched to symbol table entries, and a call graph arc is
-created by @code{cg_arcs.c:arc_add}, unless the arc fails a symspec
+created by @code{cg_arcs.c:@-arc_add}, unless the arc fails a symspec
check against INCL_ARCS/EXCL_ARCS. As each arc is added,
a linked list is maintained of the parent's child arcs, and of the child's
parent arcs.
Both the child's call count and the arc's call count are
incremented by the record's call count.
-Basic-block records are read (@code{basic_blocks.c:bb_read_rec}),
+Basic-block records are read (@code{basic_blocks.c:@-bb_read_rec}),
but only if line-by-line profiling has been selected.
Each basic-block address is matched to a corresponding line
symbol in the symbol table, and an entry made in the symbol's
records are present for the same address, the call counts
are cumulative.
-A gmon.sum file is dumped, if requested (@code{gmon_io.c:gmon_out_write}).
+A gmon.sum file is dumped, if requested (@code{gmon_io.c:@-gmon_out_write}).
If histograms were present in the data files, assign them to symbols
-(@code{hist.c:hist_assign_samples}) by iterating over all the sample
+(@code{hist.c:@-hist_assign_samples}) by iterating over all the sample
bins and assigning them to symbols. Since the symbol table
is sorted in order of ascending memory addresses, we can
simple follow along in the symbol table as we make our pass
cause each of two adjacent lines to be credited with half
a hit, for example.
-If call graph data is present, @code{cg_arcs.c:cg_assemble} is called.
+If call graph data is present, @code{cg_arcs.c:@-cg_assemble} is called.
First, if @samp{-c} was specified, a machine-dependent
routine (@code{find_call}) scans through each symbol's machine code,
looking for subroutine call instructions, and adding them
to the call graph with a zero call count.
A topological sort is performed by depth-first numbering
-all the symbols (@code{cg_dfn.c:cg_dfn}), so that
+all the symbols (@code{cg_dfn.c:@-cg_dfn}), so that
children are always numbered less than their parents,
then making a array of pointers into the symbol table and sorting it into
numerical order, which is reverse topological
new array of pointers is assembled, this time sorted by propagated time.
Finally, print the various outputs the user requested, which is now fairly
-straightforward. The call graph (@code{cg_print.c:cg_print}) and
-flat profile (@code{hist.c:hist_print}) are regurgitations of values
+straightforward. The call graph (@code{cg_print.c:@-cg_print}) and
+flat profile (@code{hist.c:@-hist_print}) are regurgitations of values
already computed. The annotated source listing
-(@code{basic_blocks.c:print_annotated_source}) uses basic-block
+(@code{basic_blocks.c:@-print_annotated_source}) uses basic-block
information, if present, to label each line of code with call counts,
otherwise only the function call counts are presented.
followed by lower use functions, followed by unused functions
at the end.
-@node Debugging,,Internals,Details
-@subsection Debugging @code{gprof}
+@node Debugging
+@section Debugging @code{gprof}
If @code{gprof} was compiled with debugging enabled,
the @samp{-d} option triggers debugging output
Tracks operation of @samp{-A} option
@end table
-@contents
+@node GNU Free Documentation License
+@appendix GNU Free Documentation License
+@include fdl.texi
+
@bye
NEEDS AN INDEX
projects / deliverable / binutils-gdb.git / blobdiff