| blob | 0fa2367a110bbdedf0d22c024739a5dd3c1142d3 |
1 .\" Automatically generated by Pod::Man 2.25 (Pod::Simple 3.20)
2 .\"
3 .\" Standard preamble:
4 .\" ========================================================================
5 .de Sp \" Vertical space (when we can't use .PP)
6 .if t .sp .5v
7 .if n .sp
8 ..
9 .de Vb \" Begin verbatim text
10 .ft CW
11 .nf
12 .ne \\$1
13 ..
14 .de Ve \" End verbatim text
15 .ft R
16 .fi
17 ..
18 .\" Set up some character translations and predefined strings. \*(-- will
19 .\" give an unbreakable dash, \*(PI will give pi, \*(L" will give a left
20 .\" double quote, and \*(R" will give a right double quote. \*(C+ will
21 .\" give a nicer C++. Capital omega is used to do unbreakable dashes and
22 .\" therefore won't be available. \*(C` and \*(C' expand to `' in nroff,
23 .\" nothing in troff, for use with C<>.
24 .tr \(*W-
25 .ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p'
26 .ie n \{\
27 . ds -- \(*W-
28 . ds PI pi
29 . if (\n(.H=4u)&(1m=24u) .ds -- \(*W\h'-12u'\(*W\h'-12u'-\" diablo 10 pitch
30 . if (\n(.H=4u)&(1m=20u) .ds -- \(*W\h'-12u'\(*W\h'-8u'-\" diablo 12 pitch
31 . ds L" ""
32 . ds R" ""
33 . ds C` ""
34 . ds C' ""
35 'br\}
36 .el\{\
37 . ds -- \|\(em\|
38 . ds PI \(*p
39 . ds L" ``
40 . ds R" ''
41 'br\}
42 .\"
43 .\" Escape single quotes in literal strings from groff's Unicode transform.
44 .ie \n(.g .ds Aq \(aq
45 .el .ds Aq '
46 .\"
47 .\" If the F register is turned on, we'll generate index entries on stderr for
48 .\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index
49 .\" entries marked with X<> in POD. Of course, you'll have to process the
50 .\" output yourself in some meaningful fashion.
51 .ie \nF \{\
52 . de IX
53 . tm Index:\\$1\t\\n%\t"\\$2"
54 ..
55 . nr % 0
56 . rr F
57 .\}
58 .el \{\
59 . de IX
60 ..
61 .\}
62 .\"
63 .\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2).
64 .\" Fear. Run. Save yourself. No user-serviceable parts.
65 . \" fudge factors for nroff and troff
66 .if n \{\
67 . ds #H 0
68 . ds #V .8m
69 . ds #F .3m
70 . ds #[ \f1
71 . ds #] \fP
72 .\}
73 .if t \{\
74 . ds #H ((1u-(\\\\n(.fu%2u))*.13m)
75 . ds #V .6m
76 . ds #F 0
77 . ds #[ \&
78 . ds #] \&
79 .\}
80 . \" simple accents for nroff and troff
81 .if n \{\
82 . ds ' \&
83 . ds ` \&
84 . ds ^ \&
85 . ds , \&
86 . ds ~ ~
87 . ds /
88 .\}
89 .if t \{\
90 . ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u"
91 . ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u'
92 . ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u'
93 . ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u'
94 . ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u'
95 . ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u'
96 .\}
97 . \" troff and (daisy-wheel) nroff accents
98 .ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V'
99 .ds 8 \h'\*(#H'\(*b\h'-\*(#H'
100 .ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#]
101 .ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H'
102 .ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u'
103 .ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#]
104 .ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#]
105 .ds ae a\h'-(\w'a'u*4/10)'e
106 .ds Ae A\h'-(\w'A'u*4/10)'E
107 . \" corrections for vroff
108 .if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u'
109 .if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u'
110 . \" for low resolution devices (crt and lpr)
111 .if \n(.H>23 .if \n(.V>19 \
112 \{\
113 . ds : e
114 . ds 8 ss
115 . ds o a
116 . ds d- d\h'-1'\(ga
117 . ds D- D\h'-1'\(hy
118 . ds th \o'bp'
119 . ds Th \o'LP'
120 . ds ae ae
121 . ds Ae AE
122 .\}
123 .rm #[ #] #H #V #F C
124 .\" ========================================================================
125 .\"
126 .IX Title "GCOV 1"
127 .TH GCOV 1 "2014-01-16" "gcc-4.8.3" "GNU"
128 .\" For nroff, turn off justification. Always turn off hyphenation; it makes
129 .\" way too many mistakes in technical documents.
130 .if n .ad l
131 .nh
132 .SH "NAME"
133 gcov \- coverage testing tool
134 .SH "SYNOPSIS"
135 .IX Header "SYNOPSIS"
136 gcov [\fB\-v\fR|\fB\-\-version\fR] [\fB\-h\fR|\fB\-\-help\fR]
137 [\fB\-a\fR|\fB\-\-all\-blocks\fR]
138 [\fB\-b\fR|\fB\-\-branch\-probabilities\fR]
139 [\fB\-c\fR|\fB\-\-branch\-counts\fR]
140 [\fB\-u\fR|\fB\-\-unconditional\-branches\fR]
141 [\fB\-n\fR|\fB\-\-no\-output\fR]
142 [\fB\-l\fR|\fB\-\-long\-file\-names\fR]
143 [\fB\-p\fR|\fB\-\-preserve\-paths\fR]
144 [\fB\-r\fR|\fB\-\-relative\-only\fR]
145 [\fB\-f\fR|\fB\-\-function\-summaries\fR]
146 [\fB\-o\fR|\fB\-\-object\-directory\fR \fIdirectory|file\fR]
147 [\fB\-s\fR|\fB\-\-source\-prefix\fR \fIdirectory\fR]
148 [\fB\-d\fR|\fB\-\-display\-progress\fR]
149 \fIfiles\fR
150 .SH "DESCRIPTION"
151 .IX Header "DESCRIPTION"
152 \&\fBgcov\fR is a test coverage program. Use it in concert with \s-1GCC\s0
153 to analyze your programs to help create more efficient, faster running
154 code and to discover untested parts of your program. You can use
155 \&\fBgcov\fR as a profiling tool to help discover where your
156 optimization efforts will best affect your code. You can also use
157 \&\fBgcov\fR along with the other profiling tool, \fBgprof\fR, to
158 assess which parts of your code use the greatest amount of computing
159 time.
160 .PP
161 Profiling tools help you analyze your code's performance. Using a
162 profiler such as \fBgcov\fR or \fBgprof\fR, you can find out some
163 basic performance statistics, such as:
164 .IP "\(bu" 4
165 how often each line of code executes
166 .IP "\(bu" 4
167 what lines of code are actually executed
168 .IP "\(bu" 4
169 how much computing time each section of code uses
170 .PP
171 Once you know these things about how your code works when compiled, you
172 can look at each module to see which modules should be optimized.
173 \&\fBgcov\fR helps you determine where to work on optimization.
174 .PP
175 Software developers also use coverage testing in concert with
176 testsuites, to make sure software is actually good enough for a release.
177 Testsuites can verify that a program works as expected; a coverage
178 program tests to see how much of the program is exercised by the
179 testsuite. Developers can then determine what kinds of test cases need
180 to be added to the testsuites to create both better testing and a better
181 final product.
182 .PP
183 You should compile your code without optimization if you plan to use
184 \&\fBgcov\fR because the optimization, by combining some lines of code
185 into one function, may not give you as much information as you need to
186 look for `hot spots' where the code is using a great deal of computer
187 time. Likewise, because \fBgcov\fR accumulates statistics by line (at
188 the lowest resolution), it works best with a programming style that
189 places only one statement on each line. If you use complicated macros
190 that expand to loops or to other control structures, the statistics are
191 less helpful\-\-\-they only report on the line where the macro call
192 appears. If your complex macros behave like functions, you can replace
193 them with inline functions to solve this problem.
194 .PP
195 \&\fBgcov\fR creates a logfile called \fI\fIsourcefile\fI.gcov\fR which
196 indicates how many times each line of a source file \fI\fIsourcefile\fI.c\fR
197 has executed. You can use these logfiles along with \fBgprof\fR to aid
198 in fine-tuning the performance of your programs. \fBgprof\fR gives
199 timing information you can use along with the information you get from
200 \&\fBgcov\fR.
201 .PP
202 \&\fBgcov\fR works only on code compiled with \s-1GCC\s0. It is not
203 compatible with any other profiling or test coverage mechanism.
204 .SH "OPTIONS"
205 .IX Header "OPTIONS"
206 .IP "\fB\-h\fR" 4
207 .IX Item "-h"
208 .PD 0
209 .IP "\fB\-\-help\fR" 4
210 .IX Item "--help"
211 .PD
212 Display help about using \fBgcov\fR (on the standard output), and
213 exit without doing any further processing.
214 .IP "\fB\-v\fR" 4
215 .IX Item "-v"
216 .PD 0
217 .IP "\fB\-\-version\fR" 4
218 .IX Item "--version"
219 .PD
220 Display the \fBgcov\fR version number (on the standard output),
221 and exit without doing any further processing.
222 .IP "\fB\-a\fR" 4
223 .IX Item "-a"
224 .PD 0
225 .IP "\fB\-\-all\-blocks\fR" 4
226 .IX Item "--all-blocks"
227 .PD
228 Write individual execution counts for every basic block. Normally gcov
229 outputs execution counts only for the main blocks of a line. With this
230 option you can determine if blocks within a single line are not being
231 executed.
232 .IP "\fB\-b\fR" 4
233 .IX Item "-b"
234 .PD 0
235 .IP "\fB\-\-branch\-probabilities\fR" 4
236 .IX Item "--branch-probabilities"
237 .PD
238 Write branch frequencies to the output file, and write branch summary
239 info to the standard output. This option allows you to see how often
240 each branch in your program was taken. Unconditional branches will not
241 be shown, unless the \fB\-u\fR option is given.
242 .IP "\fB\-c\fR" 4
243 .IX Item "-c"
244 .PD 0
245 .IP "\fB\-\-branch\-counts\fR" 4
246 .IX Item "--branch-counts"
247 .PD
248 Write branch frequencies as the number of branches taken, rather than
249 the percentage of branches taken.
250 .IP "\fB\-n\fR" 4
251 .IX Item "-n"
252 .PD 0
253 .IP "\fB\-\-no\-output\fR" 4
254 .IX Item "--no-output"
255 .PD
256 Do not create the \fBgcov\fR output file.
257 .IP "\fB\-l\fR" 4
258 .IX Item "-l"
259 .PD 0
260 .IP "\fB\-\-long\-file\-names\fR" 4
261 .IX Item "--long-file-names"
262 .PD
263 Create long file names for included source files. For example, if the
264 header file \fIx.h\fR contains code, and was included in the file
265 \&\fIa.c\fR, then running \fBgcov\fR on the file \fIa.c\fR will
266 produce an output file called \fIa.c##x.h.gcov\fR instead of
267 \&\fIx.h.gcov\fR. This can be useful if \fIx.h\fR is included in
268 multiple source files and you want to see the individual
269 contributions. If you use the \fB\-p\fR option, both the including
270 and included file names will be complete path names.
271 .IP "\fB\-p\fR" 4
272 .IX Item "-p"
273 .PD 0
274 .IP "\fB\-\-preserve\-paths\fR" 4
275 .IX Item "--preserve-paths"
276 .PD
277 Preserve complete path information in the names of generated
278 \&\fI.gcov\fR files. Without this option, just the filename component is
279 used. With this option, all directories are used, with \fB/\fR characters
280 translated to \fB#\fR characters, \fI.\fR directory components
281 removed and unremoveable \fI..\fR
282 components renamed to \fB^\fR. This is useful if sourcefiles are in several
283 different directories.
284 .IP "\fB\-r\fR" 4
285 .IX Item "-r"
286 .PD 0
287 .IP "\fB\-\-relative\-only\fR" 4
288 .IX Item "--relative-only"
289 .PD
290 Only output information about source files with a relative pathname
291 (after source prefix elision). Absolute paths are usually system
292 header files and coverage of any inline functions therein is normally
293 uninteresting.
294 .IP "\fB\-f\fR" 4
295 .IX Item "-f"
296 .PD 0
297 .IP "\fB\-\-function\-summaries\fR" 4
298 .IX Item "--function-summaries"
299 .PD
300 Output summaries for each function in addition to the file level summary.
301 .IP "\fB\-o\fR \fIdirectory|file\fR" 4
302 .IX Item "-o directory|file"
303 .PD 0
304 .IP "\fB\-\-object\-directory\fR \fIdirectory\fR" 4
305 .IX Item "--object-directory directory"
306 .IP "\fB\-\-object\-file\fR \fIfile\fR" 4
307 .IX Item "--object-file file"
308 .PD
309 Specify either the directory containing the gcov data files, or the
310 object path name. The \fI.gcno\fR, and
311 \&\fI.gcda\fR data files are searched for using this option. If a directory
312 is specified, the data files are in that directory and named after the
313 input file name, without its extension. If a file is specified here,
314 the data files are named after that file, without its extension.
315 .IP "\fB\-s\fR \fIdirectory\fR" 4
316 .IX Item "-s directory"
317 .PD 0
318 .IP "\fB\-\-source\-prefix\fR \fIdirectory\fR" 4
319 .IX Item "--source-prefix directory"
320 .PD
321 A prefix for source file names to remove when generating the output
322 coverage files. This option is useful when building in a separate
323 directory, and the pathname to the source directory is not wanted when
324 determining the output file names. Note that this prefix detection is
325 applied before determining whether the source file is absolute.
326 .IP "\fB\-u\fR" 4
327 .IX Item "-u"
328 .PD 0
329 .IP "\fB\-\-unconditional\-branches\fR" 4
330 .IX Item "--unconditional-branches"
331 .PD
332 When branch probabilities are given, include those of unconditional branches.
333 Unconditional branches are normally not interesting.
334 .IP "\fB\-d\fR" 4
335 .IX Item "-d"
336 .PD 0
337 .IP "\fB\-\-display\-progress\fR" 4
338 .IX Item "--display-progress"
339 .PD
340 Display the progress on the standard output.
341 .PP
342 \&\fBgcov\fR should be run with the current directory the same as that
343 when you invoked the compiler. Otherwise it will not be able to locate
344 the source files. \fBgcov\fR produces files called
345 \&\fI\fImangledname\fI.gcov\fR in the current directory. These contain
346 the coverage information of the source file they correspond to.
347 One \fI.gcov\fR file is produced for each source (or header) file
348 containing code,
349 which was compiled to produce the data files. The \fImangledname\fR part
350 of the output file name is usually simply the source file name, but can
351 be something more complicated if the \fB\-l\fR or \fB\-p\fR options are
352 given. Refer to those options for details.
353 .PP
354 If you invoke \fBgcov\fR with multiple input files, the
355 contributions from each input file are summed. Typically you would
356 invoke it with the same list of files as the final link of your executable.
357 .PP
358 The \fI.gcov\fR files contain the \fB:\fR separated fields along with
359 program source code. The format is
360 .PP
361 .Vb 1
362 \& <execution_count>:<line_number>:<source line text>
363 .Ve
364 .PP
365 Additional block information may succeed each line, when requested by
366 command line option. The \fIexecution_count\fR is \fB\-\fR for lines
367 containing no code. Unexecuted lines are marked \fB#####\fR or
368 \&\fB====\fR, depending on whether they are reachable by
369 non-exceptional paths or only exceptional paths such as \*(C+ exception
370 handlers, respectively.
371 .PP
372 Some lines of information at the start have \fIline_number\fR of zero.
373 These preamble lines are of the form
374 .PP
375 .Vb 1
376 \& \-:0:<tag>:<value>
377 .Ve
378 .PP
379 The ordering and number of these preamble lines will be augmented as
380 \&\fBgcov\fR development progresses \-\-\- do not rely on them remaining
381 unchanged. Use \fItag\fR to locate a particular preamble line.
382 .PP
383 The additional block information is of the form
384 .PP
385 .Vb 1
386 \& <tag> <information>
387 .Ve
388 .PP
389 The \fIinformation\fR is human readable, but designed to be simple
390 enough for machine parsing too.
391 .PP
392 When printing percentages, 0% and 100% are only printed when the values
393 are \fIexactly\fR 0% and 100% respectively. Other values which would
394 conventionally be rounded to 0% or 100% are instead printed as the
395 nearest non-boundary value.
396 .PP
397 When using \fBgcov\fR, you must first compile your program with two
398 special \s-1GCC\s0 options: \fB\-fprofile\-arcs \-ftest\-coverage\fR.
399 This tells the compiler to generate additional information needed by
400 gcov (basically a flow graph of the program) and also includes
401 additional code in the object files for generating the extra profiling
402 information needed by gcov. These additional files are placed in the
403 directory where the object file is located.
404 .PP
405 Running the program will cause profile output to be generated. For each
406 source file compiled with \fB\-fprofile\-arcs\fR, an accompanying
407 \&\fI.gcda\fR file will be placed in the object file directory.
408 .PP
409 Running \fBgcov\fR with your program's source file names as arguments
410 will now produce a listing of the code along with frequency of execution
411 for each line. For example, if your program is called \fItmp.c\fR, this
412 is what you see when you use the basic \fBgcov\fR facility:
413 .PP
414 .Vb 5
415 \& $ gcc \-fprofile\-arcs \-ftest\-coverage tmp.c
416 \& $ a.out
417 \& $ gcov tmp.c
418 \& 90.00% of 10 source lines executed in file tmp.c
419 \& Creating tmp.c.gcov.
420 .Ve
421 .PP
422 The file \fItmp.c.gcov\fR contains output from \fBgcov\fR.
423 Here is a sample:
424 .PP
425 .Vb 10
426 \& \-: 0:Source:tmp.c
427 \& \-: 0:Graph:tmp.gcno
428 \& \-: 0:Data:tmp.gcda
429 \& \-: 0:Runs:1
430 \& \-: 0:Programs:1
431 \& \-: 1:#include <stdio.h>
432 \& \-: 2:
433 \& \-: 3:int main (void)
434 \& 1: 4:{
435 \& 1: 5: int i, total;
436 \& \-: 6:
437 \& 1: 7: total = 0;
438 \& \-: 8:
439 \& 11: 9: for (i = 0; i < 10; i++)
440 \& 10: 10: total += i;
441 \& \-: 11:
442 \& 1: 12: if (total != 45)
443 \& #####: 13: printf ("Failure\en");
444 \& \-: 14: else
445 \& 1: 15: printf ("Success\en");
446 \& 1: 16: return 0;
447 \& \-: 17:}
448 .Ve
449 .PP
450 When you use the \fB\-a\fR option, you will get individual block
451 counts, and the output looks like this:
452 .PP
453 .Vb 10
454 \& \-: 0:Source:tmp.c
455 \& \-: 0:Graph:tmp.gcno
456 \& \-: 0:Data:tmp.gcda
457 \& \-: 0:Runs:1
458 \& \-: 0:Programs:1
459 \& \-: 1:#include <stdio.h>
460 \& \-: 2:
461 \& \-: 3:int main (void)
462 \& 1: 4:{
463 \& 1: 4\-block 0
464 \& 1: 5: int i, total;
465 \& \-: 6:
466 \& 1: 7: total = 0;
467 \& \-: 8:
468 \& 11: 9: for (i = 0; i < 10; i++)
469 \& 11: 9\-block 0
470 \& 10: 10: total += i;
471 \& 10: 10\-block 0
472 \& \-: 11:
473 \& 1: 12: if (total != 45)
474 \& 1: 12\-block 0
475 \& #####: 13: printf ("Failure\en");
476 \& $$$$$: 13\-block 0
477 \& \-: 14: else
478 \& 1: 15: printf ("Success\en");
479 \& 1: 15\-block 0
480 \& 1: 16: return 0;
481 \& 1: 16\-block 0
482 \& \-: 17:}
483 .Ve
484 .PP
485 In this mode, each basic block is only shown on one line \*(-- the last
486 line of the block. A multi-line block will only contribute to the
487 execution count of that last line, and other lines will not be shown
488 to contain code, unless previous blocks end on those lines.
489 The total execution count of a line is shown and subsequent lines show
490 the execution counts for individual blocks that end on that line. After each
491 block, the branch and call counts of the block will be shown, if the
492 \&\fB\-b\fR option is given.
493 .PP
494 Because of the way \s-1GCC\s0 instruments calls, a call count can be shown
495 after a line with no individual blocks.
496 As you can see, line 13 contains a basic block that was not executed.
497 .PP
498 When you use the \fB\-b\fR option, your output looks like this:
499 .PP
500 .Vb 6
501 \& $ gcov \-b tmp.c
502 \& 90.00% of 10 source lines executed in file tmp.c
503 \& 80.00% of 5 branches executed in file tmp.c
504 \& 80.00% of 5 branches taken at least once in file tmp.c
505 \& 50.00% of 2 calls executed in file tmp.c
506 \& Creating tmp.c.gcov.
507 .Ve
508 .PP
509 Here is a sample of a resulting \fItmp.c.gcov\fR file:
510 .PP
511 .Vb 10
512 \& \-: 0:Source:tmp.c
513 \& \-: 0:Graph:tmp.gcno
514 \& \-: 0:Data:tmp.gcda
515 \& \-: 0:Runs:1
516 \& \-: 0:Programs:1
517 \& \-: 1:#include <stdio.h>
518 \& \-: 2:
519 \& \-: 3:int main (void)
520 \& function main called 1 returned 1 blocks executed 75%
521 \& 1: 4:{
522 \& 1: 5: int i, total;
523 \& \-: 6:
524 \& 1: 7: total = 0;
525 \& \-: 8:
526 \& 11: 9: for (i = 0; i < 10; i++)
527 \& branch 0 taken 91% (fallthrough)
528 \& branch 1 taken 9%
529 \& 10: 10: total += i;
530 \& \-: 11:
531 \& 1: 12: if (total != 45)
532 \& branch 0 taken 0% (fallthrough)
533 \& branch 1 taken 100%
534 \& #####: 13: printf ("Failure\en");
535 \& call 0 never executed
536 \& \-: 14: else
537 \& 1: 15: printf ("Success\en");
538 \& call 0 called 1 returned 100%
539 \& 1: 16: return 0;
540 \& \-: 17:}
541 .Ve
542 .PP
543 For each function, a line is printed showing how many times the function
544 is called, how many times it returns and what percentage of the
545 function's blocks were executed.
546 .PP
547 For each basic block, a line is printed after the last line of the basic
548 block describing the branch or call that ends the basic block. There can
549 be multiple branches and calls listed for a single source line if there
550 are multiple basic blocks that end on that line. In this case, the
551 branches and calls are each given a number. There is no simple way to map
552 these branches and calls back to source constructs. In general, though,
553 the lowest numbered branch or call will correspond to the leftmost construct
554 on the source line.
555 .PP
556 For a branch, if it was executed at least once, then a percentage
557 indicating the number of times the branch was taken divided by the
558 number of times the branch was executed will be printed. Otherwise, the
559 message \*(L"never executed\*(R" is printed.
560 .PP
561 For a call, if it was executed at least once, then a percentage
562 indicating the number of times the call returned divided by the number
563 of times the call was executed will be printed. This will usually be
564 100%, but may be less for functions that call \f(CW\*(C`exit\*(C'\fR or \f(CW\*(C`longjmp\*(C'\fR,
565 and thus may not return every time they are called.
566 .PP
567 The execution counts are cumulative. If the example program were
568 executed again without removing the \fI.gcda\fR file, the count for the
569 number of times each line in the source was executed would be added to
570 the results of the previous run(s). This is potentially useful in
571 several ways. For example, it could be used to accumulate data over a
572 number of program runs as part of a test verification suite, or to
573 provide more accurate long-term information over a large number of
574 program runs.
575 .PP
576 The data in the \fI.gcda\fR files is saved immediately before the program
577 exits. For each source file compiled with \fB\-fprofile\-arcs\fR, the
578 profiling code first attempts to read in an existing \fI.gcda\fR file; if
579 the file doesn't match the executable (differing number of basic block
580 counts) it will ignore the contents of the file. It then adds in the
581 new execution counts and finally writes the data to the file.
582 .SS "Using \fBgcov\fP with \s-1GCC\s0 Optimization"
583 .IX Subsection "Using gcov with GCC Optimization"
584 If you plan to use \fBgcov\fR to help optimize your code, you must
585 first compile your program with two special \s-1GCC\s0 options:
586 \&\fB\-fprofile\-arcs \-ftest\-coverage\fR. Aside from that, you can use any
587 other \s-1GCC\s0 options; but if you want to prove that every single line
588 in your program was executed, you should not compile with optimization
589 at the same time. On some machines the optimizer can eliminate some
590 simple code lines by combining them with other lines. For example, code
591 like this:
592 .PP
593 .Vb 4
594 \& if (a != b)
595 \& c = 1;
596 \& else
597 \& c = 0;
598 .Ve
599 .PP
600 can be compiled into one instruction on some machines. In this case,
601 there is no way for \fBgcov\fR to calculate separate execution counts
602 for each line because there isn't separate code for each line. Hence
603 the \fBgcov\fR output looks like this if you compiled the program with
604 optimization:
605 .PP
606 .Vb 4
607 \& 100: 12:if (a != b)
608 \& 100: 13: c = 1;
609 \& 100: 14:else
610 \& 100: 15: c = 0;
611 .Ve
612 .PP
613 The output shows that this block of code, combined by optimization,
614 executed 100 times. In one sense this result is correct, because there
615 was only one instruction representing all four of these lines. However,
616 the output does not indicate how many times the result was 0 and how
617 many times the result was 1.
618 .PP
619 Inlineable functions can create unexpected line counts. Line counts are
620 shown for the source code of the inlineable function, but what is shown
621 depends on where the function is inlined, or if it is not inlined at all.
622 .PP
623 If the function is not inlined, the compiler must emit an out of line
624 copy of the function, in any object file that needs it. If
625 \&\fIfileA.o\fR and \fIfileB.o\fR both contain out of line bodies of a
626 particular inlineable function, they will also both contain coverage
627 counts for that function. When \fIfileA.o\fR and \fIfileB.o\fR are
628 linked together, the linker will, on many systems, select one of those
629 out of line bodies for all calls to that function, and remove or ignore
630 the other. Unfortunately, it will not remove the coverage counters for
631 the unused function body. Hence when instrumented, all but one use of
632 that function will show zero counts.
633 .PP
634 If the function is inlined in several places, the block structure in
635 each location might not be the same. For instance, a condition might
636 now be calculable at compile time in some instances. Because the
637 coverage of all the uses of the inline function will be shown for the
638 same source lines, the line counts themselves might seem inconsistent.
639 .PP
640 Long-running applications can use the \f(CW\*(C`_gcov_reset\*(C'\fR and \f(CW\*(C`_gcov_dump\*(C'\fR
641 facilities to restrict profile collection to the program region of
642 interest. Calling \f(CW\*(C`_gcov_reset(void)\*(C'\fR will clear all profile counters
643 to zero, and calling \f(CW\*(C`_gcov_dump(void)\*(C'\fR will cause the profile information
644 collected at that point to be dumped to \fI.gcda\fR output files.
645 .SH "SEE ALSO"
646 .IX Header "SEE ALSO"
647 \&\fIgpl\fR\|(7), \fIgfdl\fR\|(7), \fIfsf\-funding\fR\|(7), \fIgcc\fR\|(1) and the Info entry for \fIgcc\fR.
648 .SH "COPYRIGHT"
649 .IX Header "COPYRIGHT"
650 Copyright (c) 1996\-2013 Free Software Foundation, Inc.
651 .PP
652 Permission is granted to copy, distribute and/or modify this document
653 under the terms of the \s-1GNU\s0 Free Documentation License, Version 1.3 or
654 any later version published by the Free Software Foundation; with the
655 Invariant Sections being \*(L"\s-1GNU\s0 General Public License\*(R" and \*(L"Funding
656 Free Software\*(R", the Front-Cover texts being (a) (see below), and with
657 the Back-Cover Texts being (b) (see below). A copy of the license is
658 included in the \fIgfdl\fR\|(7) man page.
659 .PP
660 (a) The \s-1FSF\s0's Front-Cover Text is:
661 .PP
662 .Vb 1
663 \& A GNU Manual
664 .Ve
665 .PP
666 (b) The \s-1FSF\s0's Back-Cover Text is:
667 .PP
668 .Vb 3
669 \& You have freedom to copy and modify this GNU Manual, like GNU
670 \& software. Copies published by the Free Software Foundation raise
671 \& funds for GNU development.
672 .Ve