Merge branch 'vendor/GCC47'
[dragonfly.git] / contrib / gcc-4.7 / gcc / doc / gcov.1
CommitLineData
5ce9237c 1.\" Automatically generated by Pod::Man 2.16 (Pod::Simple 3.05)
e4b17023
JM
2.\"
3.\" Standard preamble:
4.\" ========================================================================
5ce9237c
JM
5.de Sh \" Subsection heading
6.br
7.if t .Sp
8.ne 5
9.PP
10\fB\\$1\fR
11.PP
12..
e4b17023
JM
13.de Sp \" Vertical space (when we can't use .PP)
14.if t .sp .5v
15.if n .sp
16..
17.de Vb \" Begin verbatim text
18.ft CW
19.nf
20.ne \\$1
21..
22.de Ve \" End verbatim text
23.ft R
24.fi
25..
26.\" Set up some character translations and predefined strings. \*(-- will
27.\" give an unbreakable dash, \*(PI will give pi, \*(L" will give a left
28.\" double quote, and \*(R" will give a right double quote. \*(C+ will
29.\" give a nicer C++. Capital omega is used to do unbreakable dashes and
30.\" therefore won't be available. \*(C` and \*(C' expand to `' in nroff,
31.\" nothing in troff, for use with C<>.
32.tr \(*W-
33.ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p'
34.ie n \{\
35. ds -- \(*W-
36. ds PI pi
37. if (\n(.H=4u)&(1m=24u) .ds -- \(*W\h'-12u'\(*W\h'-12u'-\" diablo 10 pitch
38. if (\n(.H=4u)&(1m=20u) .ds -- \(*W\h'-12u'\(*W\h'-8u'-\" diablo 12 pitch
39. ds L" ""
40. ds R" ""
41. ds C` ""
42. ds C' ""
43'br\}
44.el\{\
45. ds -- \|\(em\|
46. ds PI \(*p
47. ds L" ``
48. ds R" ''
49'br\}
50.\"
51.\" Escape single quotes in literal strings from groff's Unicode transform.
52.ie \n(.g .ds Aq \(aq
53.el .ds Aq '
54.\"
55.\" If the F register is turned on, we'll generate index entries on stderr for
5ce9237c 56.\" titles (.TH), headers (.SH), subsections (.Sh), items (.Ip), and index
e4b17023
JM
57.\" entries marked with X<> in POD. Of course, you'll have to process the
58.\" output yourself in some meaningful fashion.
59.ie \nF \{\
60. de IX
61. tm Index:\\$1\t\\n%\t"\\$2"
62..
63. nr % 0
64. rr F
65.\}
66.el \{\
67. de IX
68..
69.\}
70.\"
71.\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2).
72.\" Fear. Run. Save yourself. No user-serviceable parts.
73. \" fudge factors for nroff and troff
74.if n \{\
75. ds #H 0
76. ds #V .8m
77. ds #F .3m
78. ds #[ \f1
79. ds #] \fP
80.\}
81.if t \{\
82. ds #H ((1u-(\\\\n(.fu%2u))*.13m)
83. ds #V .6m
84. ds #F 0
85. ds #[ \&
86. ds #] \&
87.\}
88. \" simple accents for nroff and troff
89.if n \{\
90. ds ' \&
91. ds ` \&
92. ds ^ \&
93. ds , \&
94. ds ~ ~
95. ds /
96.\}
97.if t \{\
98. ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u"
99. ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u'
100. ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u'
101. ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u'
102. ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u'
103. ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u'
104.\}
105. \" troff and (daisy-wheel) nroff accents
106.ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V'
107.ds 8 \h'\*(#H'\(*b\h'-\*(#H'
108.ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#]
109.ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H'
110.ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u'
111.ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#]
112.ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#]
113.ds ae a\h'-(\w'a'u*4/10)'e
114.ds Ae A\h'-(\w'A'u*4/10)'E
115. \" corrections for vroff
116.if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u'
117.if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u'
118. \" for low resolution devices (crt and lpr)
119.if \n(.H>23 .if \n(.V>19 \
120\{\
121. ds : e
122. ds 8 ss
123. ds o a
124. ds d- d\h'-1'\(ga
125. ds D- D\h'-1'\(hy
126. ds th \o'bp'
127. ds Th \o'LP'
128. ds ae ae
129. ds Ae AE
130.\}
131.rm #[ #] #H #V #F C
132.\" ========================================================================
133.\"
134.IX Title "GCOV 1"
95d28233 135.TH GCOV 1 "2014-06-12" "gcc-4.7.4" "GNU"
e4b17023
JM
136.\" For nroff, turn off justification. Always turn off hyphenation; it makes
137.\" way too many mistakes in technical documents.
138.if n .ad l
139.nh
140.SH "NAME"
141gcov \- coverage testing tool
142.SH "SYNOPSIS"
143.IX Header "SYNOPSIS"
144gcov [\fB\-v\fR|\fB\-\-version\fR] [\fB\-h\fR|\fB\-\-help\fR]
145 [\fB\-a\fR|\fB\-\-all\-blocks\fR]
146 [\fB\-b\fR|\fB\-\-branch\-probabilities\fR]
147 [\fB\-c\fR|\fB\-\-branch\-counts\fR]
148 [\fB\-u\fR|\fB\-\-unconditional\-branches\fR]
149 [\fB\-n\fR|\fB\-\-no\-output\fR]
150 [\fB\-l\fR|\fB\-\-long\-file\-names\fR]
151 [\fB\-p\fR|\fB\-\-preserve\-paths\fR]
152 [\fB\-r\fR|\fB\-\-relative\-only\fR]
153 [\fB\-f\fR|\fB\-\-function\-summaries\fR]
154 [\fB\-o\fR|\fB\-\-object\-directory\fR \fIdirectory|file\fR]
155 [\fB\-s\fR|\fB\-\-source\-prefix\fR \fIdirectory\fR]
156 [\fB\-d\fR|\fB\-\-display\-progress\fR]
157 \fIfiles\fR
158.SH "DESCRIPTION"
159.IX Header "DESCRIPTION"
160\&\fBgcov\fR is a test coverage program. Use it in concert with \s-1GCC\s0
161to analyze your programs to help create more efficient, faster running
162code and to discover untested parts of your program. You can use
163\&\fBgcov\fR as a profiling tool to help discover where your
164optimization efforts will best affect your code. You can also use
165\&\fBgcov\fR along with the other profiling tool, \fBgprof\fR, to
166assess which parts of your code use the greatest amount of computing
167time.
168.PP
169Profiling tools help you analyze your code's performance. Using a
170profiler such as \fBgcov\fR or \fBgprof\fR, you can find out some
171basic performance statistics, such as:
172.IP "\(bu" 4
173how often each line of code executes
174.IP "\(bu" 4
175what lines of code are actually executed
176.IP "\(bu" 4
177how much computing time each section of code uses
178.PP
179Once you know these things about how your code works when compiled, you
180can look at each module to see which modules should be optimized.
181\&\fBgcov\fR helps you determine where to work on optimization.
182.PP
183Software developers also use coverage testing in concert with
184testsuites, to make sure software is actually good enough for a release.
185Testsuites can verify that a program works as expected; a coverage
186program tests to see how much of the program is exercised by the
187testsuite. Developers can then determine what kinds of test cases need
188to be added to the testsuites to create both better testing and a better
189final product.
190.PP
191You should compile your code without optimization if you plan to use
192\&\fBgcov\fR because the optimization, by combining some lines of code
193into one function, may not give you as much information as you need to
194look for `hot spots' where the code is using a great deal of computer
195time. Likewise, because \fBgcov\fR accumulates statistics by line (at
196the lowest resolution), it works best with a programming style that
197places only one statement on each line. If you use complicated macros
198that expand to loops or to other control structures, the statistics are
199less helpful\-\-\-they only report on the line where the macro call
200appears. If your complex macros behave like functions, you can replace
201them with inline functions to solve this problem.
202.PP
203\&\fBgcov\fR creates a logfile called \fI\fIsourcefile\fI.gcov\fR which
204indicates how many times each line of a source file \fI\fIsourcefile\fI.c\fR
205has executed. You can use these logfiles along with \fBgprof\fR to aid
206in fine-tuning the performance of your programs. \fBgprof\fR gives
207timing information you can use along with the information you get from
208\&\fBgcov\fR.
209.PP
210\&\fBgcov\fR works only on code compiled with \s-1GCC\s0. It is not
211compatible with any other profiling or test coverage mechanism.
212.SH "OPTIONS"
213.IX Header "OPTIONS"
214.IP "\fB\-h\fR" 4
215.IX Item "-h"
216.PD 0
217.IP "\fB\-\-help\fR" 4
218.IX Item "--help"
219.PD
220Display help about using \fBgcov\fR (on the standard output), and
221exit without doing any further processing.
222.IP "\fB\-v\fR" 4
223.IX Item "-v"
224.PD 0
225.IP "\fB\-\-version\fR" 4
226.IX Item "--version"
227.PD
228Display the \fBgcov\fR version number (on the standard output),
229and exit without doing any further processing.
230.IP "\fB\-a\fR" 4
231.IX Item "-a"
232.PD 0
233.IP "\fB\-\-all\-blocks\fR" 4
234.IX Item "--all-blocks"
235.PD
236Write individual execution counts for every basic block. Normally gcov
237outputs execution counts only for the main blocks of a line. With this
238option you can determine if blocks within a single line are not being
239executed.
240.IP "\fB\-b\fR" 4
241.IX Item "-b"
242.PD 0
243.IP "\fB\-\-branch\-probabilities\fR" 4
244.IX Item "--branch-probabilities"
245.PD
246Write branch frequencies to the output file, and write branch summary
247info to the standard output. This option allows you to see how often
248each branch in your program was taken. Unconditional branches will not
249be shown, unless the \fB\-u\fR option is given.
250.IP "\fB\-c\fR" 4
251.IX Item "-c"
252.PD 0
253.IP "\fB\-\-branch\-counts\fR" 4
254.IX Item "--branch-counts"
255.PD
256Write branch frequencies as the number of branches taken, rather than
257the percentage of branches taken.
258.IP "\fB\-n\fR" 4
259.IX Item "-n"
260.PD 0
261.IP "\fB\-\-no\-output\fR" 4
262.IX Item "--no-output"
263.PD
264Do not create the \fBgcov\fR output file.
265.IP "\fB\-l\fR" 4
266.IX Item "-l"
267.PD 0
268.IP "\fB\-\-long\-file\-names\fR" 4
269.IX Item "--long-file-names"
270.PD
271Create long file names for included source files. For example, if the
272header file \fIx.h\fR contains code, and was included in the file
273\&\fIa.c\fR, then running \fBgcov\fR on the file \fIa.c\fR will
274produce an output file called \fIa.c##x.h.gcov\fR instead of
275\&\fIx.h.gcov\fR. This can be useful if \fIx.h\fR is included in
276multiple source files and you want to see the individual
277contributions. If you use the \fB\-p\fR option, both the including
278and included file names will be complete path names.
279.IP "\fB\-p\fR" 4
280.IX Item "-p"
281.PD 0
282.IP "\fB\-\-preserve\-paths\fR" 4
283.IX Item "--preserve-paths"
284.PD
285Preserve complete path information in the names of generated
286\&\fI.gcov\fR files. Without this option, just the filename component is
287used. With this option, all directories are used, with \fB/\fR characters
288translated to \fB#\fR characters, \fI.\fR directory components
289removed and unremoveable \fI..\fR
290components renamed to \fB^\fR. This is useful if sourcefiles are in several
291different directories.
292.IP "\fB\-r\fR" 4
293.IX Item "-r"
294.PD 0
295.IP "\fB\-\-relative\-only\fR" 4
296.IX Item "--relative-only"
297.PD
298Only output information about source files with a relative pathname
299(after source prefix elision). Absolute paths are usually system
300header files and coverage of any inline functions therein is normally
301uninteresting.
302.IP "\fB\-f\fR" 4
303.IX Item "-f"
304.PD 0
305.IP "\fB\-\-function\-summaries\fR" 4
306.IX Item "--function-summaries"
307.PD
308Output summaries for each function in addition to the file level summary.
309.IP "\fB\-o\fR \fIdirectory|file\fR" 4
310.IX Item "-o directory|file"
311.PD 0
312.IP "\fB\-\-object\-directory\fR \fIdirectory\fR" 4
313.IX Item "--object-directory directory"
314.IP "\fB\-\-object\-file\fR \fIfile\fR" 4
315.IX Item "--object-file file"
316.PD
317Specify either the directory containing the gcov data files, or the
318object path name. The \fI.gcno\fR, and
319\&\fI.gcda\fR data files are searched for using this option. If a directory
320is specified, the data files are in that directory and named after the
321input file name, without its extension. If a file is specified here,
322the data files are named after that file, without its extension.
323.IP "\fB\-s\fR \fIdirectory\fR" 4
324.IX Item "-s directory"
325.PD 0
326.IP "\fB\-\-source\-prefix\fR \fIdirectory\fR" 4
327.IX Item "--source-prefix directory"
328.PD
329A prefix for source file names to remove when generating the output
330coverage files. This option is useful when building in a separate
331directory, and the pathname to the source directory is not wanted when
332determining the output file names. Note that this prefix detection is
333applied before determining whether the source file is absolute.
334.IP "\fB\-u\fR" 4
335.IX Item "-u"
336.PD 0
337.IP "\fB\-\-unconditional\-branches\fR" 4
338.IX Item "--unconditional-branches"
339.PD
340When branch probabilities are given, include those of unconditional branches.
341Unconditional branches are normally not interesting.
342.IP "\fB\-d\fR" 4
343.IX Item "-d"
344.PD 0
345.IP "\fB\-\-display\-progress\fR" 4
346.IX Item "--display-progress"
347.PD
348Display the progress on the standard output.
349.PP
350\&\fBgcov\fR should be run with the current directory the same as that
351when you invoked the compiler. Otherwise it will not be able to locate
352the source files. \fBgcov\fR produces files called
353\&\fI\fImangledname\fI.gcov\fR in the current directory. These contain
354the coverage information of the source file they correspond to.
355One \fI.gcov\fR file is produced for each source (or header) file
356containing code,
357which was compiled to produce the data files. The \fImangledname\fR part
358of the output file name is usually simply the source file name, but can
359be something more complicated if the \fB\-l\fR or \fB\-p\fR options are
360given. Refer to those options for details.
361.PP
362If you invoke \fBgcov\fR with multiple input files, the
363contributions from each input file are summed. Typically you would
364invoke it with the same list of files as the final link of your executable.
365.PP
366The \fI.gcov\fR files contain the \fB:\fR separated fields along with
367program source code. The format is
368.PP
369.Vb 1
370\& <execution_count>:<line_number>:<source line text>
371.Ve
372.PP
373Additional block information may succeed each line, when requested by
374command line option. The \fIexecution_count\fR is \fB\-\fR for lines
375containing no code. Unexecuted lines are marked \fB#####\fR or
376\&\fB====\fR, depending on whether they are reachable by
377non-exceptional paths or only exceptional paths such as \*(C+ exception
378handlers, respectively.
379.PP
380Some lines of information at the start have \fIline_number\fR of zero.
381These preamble lines are of the form
382.PP
383.Vb 1
384\& \-:0:<tag>:<value>
385.Ve
386.PP
387The ordering and number of these preamble lines will be augmented as
388\&\fBgcov\fR development progresses \-\-\- do not rely on them remaining
389unchanged. Use \fItag\fR to locate a particular preamble line.
390.PP
391The additional block information is of the form
392.PP
393.Vb 1
394\& <tag> <information>
395.Ve
396.PP
397The \fIinformation\fR is human readable, but designed to be simple
398enough for machine parsing too.
399.PP
400When printing percentages, 0% and 100% are only printed when the values
401are \fIexactly\fR 0% and 100% respectively. Other values which would
402conventionally be rounded to 0% or 100% are instead printed as the
403nearest non-boundary value.
404.PP
405When using \fBgcov\fR, you must first compile your program with two
406special \s-1GCC\s0 options: \fB\-fprofile\-arcs \-ftest\-coverage\fR.
407This tells the compiler to generate additional information needed by
408gcov (basically a flow graph of the program) and also includes
409additional code in the object files for generating the extra profiling
410information needed by gcov. These additional files are placed in the
411directory where the object file is located.
412.PP
413Running the program will cause profile output to be generated. For each
414source file compiled with \fB\-fprofile\-arcs\fR, an accompanying
415\&\fI.gcda\fR file will be placed in the object file directory.
416.PP
417Running \fBgcov\fR with your program's source file names as arguments
418will now produce a listing of the code along with frequency of execution
419for each line. For example, if your program is called \fItmp.c\fR, this
420is what you see when you use the basic \fBgcov\fR facility:
421.PP
422.Vb 5
423\& $ gcc \-fprofile\-arcs \-ftest\-coverage tmp.c
424\& $ a.out
425\& $ gcov tmp.c
426\& 90.00% of 10 source lines executed in file tmp.c
427\& Creating tmp.c.gcov.
428.Ve
429.PP
430The file \fItmp.c.gcov\fR contains output from \fBgcov\fR.
431Here is a sample:
432.PP
433.Vb 10
434\& \-: 0:Source:tmp.c
435\& \-: 0:Graph:tmp.gcno
436\& \-: 0:Data:tmp.gcda
437\& \-: 0:Runs:1
438\& \-: 0:Programs:1
439\& \-: 1:#include <stdio.h>
440\& \-: 2:
441\& \-: 3:int main (void)
442\& 1: 4:{
443\& 1: 5: int i, total;
444\& \-: 6:
445\& 1: 7: total = 0;
446\& \-: 8:
447\& 11: 9: for (i = 0; i < 10; i++)
448\& 10: 10: total += i;
449\& \-: 11:
450\& 1: 12: if (total != 45)
451\& #####: 13: printf ("Failure\en");
452\& \-: 14: else
453\& 1: 15: printf ("Success\en");
454\& 1: 16: return 0;
455\& \-: 17:}
456.Ve
457.PP
458When you use the \fB\-a\fR option, you will get individual block
459counts, and the output looks like this:
460.PP
461.Vb 10
462\& \-: 0:Source:tmp.c
463\& \-: 0:Graph:tmp.gcno
464\& \-: 0:Data:tmp.gcda
465\& \-: 0:Runs:1
466\& \-: 0:Programs:1
467\& \-: 1:#include <stdio.h>
468\& \-: 2:
469\& \-: 3:int main (void)
470\& 1: 4:{
471\& 1: 4\-block 0
472\& 1: 5: int i, total;
473\& \-: 6:
474\& 1: 7: total = 0;
475\& \-: 8:
476\& 11: 9: for (i = 0; i < 10; i++)
477\& 11: 9\-block 0
478\& 10: 10: total += i;
479\& 10: 10\-block 0
480\& \-: 11:
481\& 1: 12: if (total != 45)
482\& 1: 12\-block 0
483\& #####: 13: printf ("Failure\en");
484\& $$$$$: 13\-block 0
485\& \-: 14: else
486\& 1: 15: printf ("Success\en");
487\& 1: 15\-block 0
488\& 1: 16: return 0;
489\& 1: 16\-block 0
490\& \-: 17:}
491.Ve
492.PP
493In this mode, each basic block is only shown on one line \*(-- the last
494line of the block. A multi-line block will only contribute to the
495execution count of that last line, and other lines will not be shown
496to contain code, unless previous blocks end on those lines.
497The total execution count of a line is shown and subsequent lines show
498the execution counts for individual blocks that end on that line. After each
499block, the branch and call counts of the block will be shown, if the
500\&\fB\-b\fR option is given.
501.PP
502Because of the way \s-1GCC\s0 instruments calls, a call count can be shown
503after a line with no individual blocks.
504As you can see, line 13 contains a basic block that was not executed.
505.PP
506When you use the \fB\-b\fR option, your output looks like this:
507.PP
508.Vb 6
509\& $ gcov \-b tmp.c
510\& 90.00% of 10 source lines executed in file tmp.c
511\& 80.00% of 5 branches executed in file tmp.c
512\& 80.00% of 5 branches taken at least once in file tmp.c
513\& 50.00% of 2 calls executed in file tmp.c
514\& Creating tmp.c.gcov.
515.Ve
516.PP
517Here is a sample of a resulting \fItmp.c.gcov\fR file:
518.PP
519.Vb 10
520\& \-: 0:Source:tmp.c
521\& \-: 0:Graph:tmp.gcno
522\& \-: 0:Data:tmp.gcda
523\& \-: 0:Runs:1
524\& \-: 0:Programs:1
525\& \-: 1:#include <stdio.h>
526\& \-: 2:
527\& \-: 3:int main (void)
528\& function main called 1 returned 1 blocks executed 75%
529\& 1: 4:{
530\& 1: 5: int i, total;
531\& \-: 6:
532\& 1: 7: total = 0;
533\& \-: 8:
534\& 11: 9: for (i = 0; i < 10; i++)
535\& branch 0 taken 91% (fallthrough)
536\& branch 1 taken 9%
537\& 10: 10: total += i;
538\& \-: 11:
539\& 1: 12: if (total != 45)
540\& branch 0 taken 0% (fallthrough)
541\& branch 1 taken 100%
542\& #####: 13: printf ("Failure\en");
543\& call 0 never executed
544\& \-: 14: else
545\& 1: 15: printf ("Success\en");
546\& call 0 called 1 returned 100%
547\& 1: 16: return 0;
548\& \-: 17:}
549.Ve
550.PP
551For each function, a line is printed showing how many times the function
552is called, how many times it returns and what percentage of the
553function's blocks were executed.
554.PP
555For each basic block, a line is printed after the last line of the basic
556block describing the branch or call that ends the basic block. There can
557be multiple branches and calls listed for a single source line if there
558are multiple basic blocks that end on that line. In this case, the
559branches and calls are each given a number. There is no simple way to map
560these branches and calls back to source constructs. In general, though,
561the lowest numbered branch or call will correspond to the leftmost construct
562on the source line.
563.PP
564For a branch, if it was executed at least once, then a percentage
565indicating the number of times the branch was taken divided by the
566number of times the branch was executed will be printed. Otherwise, the
567message \*(L"never executed\*(R" is printed.
568.PP
569For a call, if it was executed at least once, then a percentage
570indicating the number of times the call returned divided by the number
571of times the call was executed will be printed. This will usually be
572100%, but may be less for functions that call \f(CW\*(C`exit\*(C'\fR or \f(CW\*(C`longjmp\*(C'\fR,
573and thus may not return every time they are called.
574.PP
575The execution counts are cumulative. If the example program were
576executed again without removing the \fI.gcda\fR file, the count for the
577number of times each line in the source was executed would be added to
578the results of the previous run(s). This is potentially useful in
579several ways. For example, it could be used to accumulate data over a
580number of program runs as part of a test verification suite, or to
581provide more accurate long-term information over a large number of
582program runs.
583.PP
584The data in the \fI.gcda\fR files is saved immediately before the program
585exits. For each source file compiled with \fB\-fprofile\-arcs\fR, the
586profiling code first attempts to read in an existing \fI.gcda\fR file; if
587the file doesn't match the executable (differing number of basic block
588counts) it will ignore the contents of the file. It then adds in the
589new execution counts and finally writes the data to the file.
5ce9237c 590.Sh "Using \fBgcov\fP with \s-1GCC\s0 Optimization"
e4b17023
JM
591.IX Subsection "Using gcov with GCC Optimization"
592If you plan to use \fBgcov\fR to help optimize your code, you must
593first compile your program with two special \s-1GCC\s0 options:
594\&\fB\-fprofile\-arcs \-ftest\-coverage\fR. Aside from that, you can use any
595other \s-1GCC\s0 options; but if you want to prove that every single line
596in your program was executed, you should not compile with optimization
597at the same time. On some machines the optimizer can eliminate some
598simple code lines by combining them with other lines. For example, code
599like this:
600.PP
601.Vb 4
602\& if (a != b)
603\& c = 1;
604\& else
605\& c = 0;
606.Ve
607.PP
608can be compiled into one instruction on some machines. In this case,
609there is no way for \fBgcov\fR to calculate separate execution counts
610for each line because there isn't separate code for each line. Hence
611the \fBgcov\fR output looks like this if you compiled the program with
612optimization:
613.PP
614.Vb 4
615\& 100: 12:if (a != b)
616\& 100: 13: c = 1;
617\& 100: 14:else
618\& 100: 15: c = 0;
619.Ve
620.PP
621The output shows that this block of code, combined by optimization,
622executed 100 times. In one sense this result is correct, because there
623was only one instruction representing all four of these lines. However,
624the output does not indicate how many times the result was 0 and how
625many times the result was 1.
626.PP
627Inlineable functions can create unexpected line counts. Line counts are
628shown for the source code of the inlineable function, but what is shown
629depends on where the function is inlined, or if it is not inlined at all.
630.PP
631If the function is not inlined, the compiler must emit an out of line
632copy of the function, in any object file that needs it. If
633\&\fIfileA.o\fR and \fIfileB.o\fR both contain out of line bodies of a
634particular inlineable function, they will also both contain coverage
635counts for that function. When \fIfileA.o\fR and \fIfileB.o\fR are
636linked together, the linker will, on many systems, select one of those
637out of line bodies for all calls to that function, and remove or ignore
638the other. Unfortunately, it will not remove the coverage counters for
639the unused function body. Hence when instrumented, all but one use of
640that function will show zero counts.
641.PP
642If the function is inlined in several places, the block structure in
643each location might not be the same. For instance, a condition might
644now be calculable at compile time in some instances. Because the
645coverage of all the uses of the inline function will be shown for the
646same source lines, the line counts themselves might seem inconsistent.
647.SH "SEE ALSO"
648.IX Header "SEE ALSO"
649\&\fIgpl\fR\|(7), \fIgfdl\fR\|(7), \fIfsf\-funding\fR\|(7), \fIgcc\fR\|(1) and the Info entry for \fIgcc\fR.
650.SH "COPYRIGHT"
651.IX Header "COPYRIGHT"
652Copyright (c) 1996, 1997, 1999, 2000, 2001, 2002, 2003, 2004,
6532005, 2008, 2010 Free Software Foundation, Inc.
654.PP
655Permission is granted to copy, distribute and/or modify this document
656under the terms of the \s-1GNU\s0 Free Documentation License, Version 1.3 or
657any later version published by the Free Software Foundation; with the
658Invariant Sections being \*(L"\s-1GNU\s0 General Public License\*(R" and \*(L"Funding
659Free Software\*(R", the Front-Cover texts being (a) (see below), and with
660the Back-Cover Texts being (b) (see below). A copy of the license is
661included in the \fIgfdl\fR\|(7) man page.
662.PP
663(a) The \s-1FSF\s0's Front-Cover Text is:
664.PP
665.Vb 1
666\& A GNU Manual
667.Ve
668.PP
669(b) The \s-1FSF\s0's Back-Cover Text is:
670.PP
671.Vb 3
672\& You have freedom to copy and modify this GNU Manual, like GNU
673\& software. Copies published by the Free Software Foundation raise
674\& funds for GNU development.
675.Ve