Merge branch 'master' of ssh://
[dragonfly.git] / contrib / gcc-3.4 / gcc / doc / gcov.1
1.\" Automatically generated by Pod::Man v1.37, Pod::Parser v1.14
3.\" Standard preamble:
4.\" ======================================================================== Sh \" Subsection heading
7.if t .Sp 5
12.. Sp \" Vertical space (when we can't use .PP)
14.if t .sp .5v
15.if n .sp
16.. Vb \" Begin verbatim text
18.ft CW \\$1
21.. Ve \" End verbatim text
23.ft R
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. | will give a
29.\" real vertical bar. \*(C+ will give a nicer C++. Capital omega is used to
30.\" do unbreakable dashes and therefore won't be available. \*(C` and \*(C'
31.\" expand to `' in nroff, nothing in troff, for use with C<>. \(*W-|\(bv\*(Tr
33.ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p' 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' ""
45. ds -- \|\(em\|
46. ds PI \(*p
47. ds L" ``
48. ds R" ''
51.\" If the F register is turned on, we'll generate index entries on stderr for
52.\" titles (.TH), headers (.SH), subsections (.Sh), items (.Ip), and index
53.\" entries marked with X<> in POD. Of course, you'll have to process the
54.\" output yourself in some meaningful fashion.
55.if \nF \{\
56. de IX
57. tm Index:\\$1\t\\n%\t"\\$2"
59. nr % 0
60. rr F
63.\" For nroff, turn off justification. Always turn off hyphenation; it makes
64.\" way too many mistakes in technical documents.
65.hy 0
66.if n .na
68.\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2).
69.\" Fear. Run. Save yourself. No user-serviceable parts.
70. \" fudge factors for nroff and troff
71.if n \{\
72. ds #H 0
73. ds #V .8m
74. ds #F .3m
75. ds #[ \f1
76. ds #] \fP
78.if t \{\
79. ds #H ((1u-(\\\\n(.fu%2u))*.13m)
80. ds #V .6m
81. ds #F 0
82. ds #[ \&
83. ds #] \&
85. \" simple accents for nroff and troff
86.if n \{\
87. ds ' \&
88. ds ` \&
89. ds ^ \&
90. ds , \&
91. ds ~ ~
92. ds /
94.if t \{\
95. ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u"
96. ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u'
97. ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u'
98. ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u'
99. ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u'
100. ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u'
102. \" troff and (daisy-wheel) nroff accents
103.ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V'
104.ds 8 \h'\*(#H'\(*b\h'-\*(#H'
105.ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#]
106.ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H'
107.ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u'
108.ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#]
109.ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#]
110.ds ae a\h'-(\w'a'u*4/10)'e
111.ds Ae A\h'-(\w'A'u*4/10)'E
112. \" corrections for vroff
113.if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u'
114.if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u'
115. \" for low resolution devices (crt and lpr)
116.if \n(.H>23 .if \n(.V>19 \
118. ds : e
119. ds 8 ss
120. ds o a
121. ds d- d\h'-1'\(ga
122. ds D- D\h'-1'\(hy
123. ds th \o'bp'
124. ds Th \o'LP'
125. ds ae ae
126. ds Ae AE
128.rm #[ #] #H #V #F C
129.\" ========================================================================
131.IX Title "GCOV 1"
132.TH GCOV 1 "2006-03-06" "gcc-3.4.6" "GNU"
133.SH "NAME"
134gcov \- coverage testing tool
136.IX Header "SYNOPSIS"
137gcov [\fB\-v\fR|\fB\-\-version\fR] [\fB\-h\fR|\fB\-\-help\fR]
138 [\fB\-a\fR|\fB\-\-all\-blocks\fR]
139 [\fB\-b\fR|\fB\-\-branch\-probabilities\fR]
140 [\fB\-c\fR|\fB\-\-branch\-counts\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\-f\fR|\fB\-\-function\-summaries\fR]
145 [\fB\-o\fR|\fB\-\-object\-directory\fR \fIdirectory|file\fR] \fIsourcefile\fR
146 [\fB\-u\fR|\fB\-\-unconditional\-branches\fR]
149\&\fBgcov\fR is a test coverage program. Use it in concert with \s-1GCC\s0
150to analyze your programs to help create more efficient, faster running
151code and to discover untested parts of your program. You can use
152\&\fBgcov\fR as a profiling tool to help discover where your
153optimization efforts will best affect your code. You can also use
154\&\fBgcov\fR along with the other profiling tool, \fBgprof\fR, to
155assess which parts of your code use the greatest amount of computing
158Profiling tools help you analyze your code's performance. Using a
159profiler such as \fBgcov\fR or \fBgprof\fR, you can find out some
160basic performance statistics, such as:
161.IP "\(bu" 4
162how often each line of code executes
163.IP "\(bu" 4
164what lines of code are actually executed
165.IP "\(bu" 4
166how much computing time each section of code uses
168Once you know these things about how your code works when compiled, you
169can look at each module to see which modules should be optimized.
170\&\fBgcov\fR helps you determine where to work on optimization.
172Software developers also use coverage testing in concert with
173testsuites, to make sure software is actually good enough for a release.
174Testsuites can verify that a program works as expected; a coverage
175program tests to see how much of the program is exercised by the
176testsuite. Developers can then determine what kinds of test cases need
177to be added to the testsuites to create both better testing and a better
178final product.
180You should compile your code without optimization if you plan to use
181\&\fBgcov\fR because the optimization, by combining some lines of code
182into one function, may not give you as much information as you need to
183look for `hot spots' where the code is using a great deal of computer
184time. Likewise, because \fBgcov\fR accumulates statistics by line (at
185the lowest resolution), it works best with a programming style that
186places only one statement on each line. If you use complicated macros
187that expand to loops or to other control structures, the statistics are
188less helpful\-\-\-they only report on the line where the macro call
189appears. If your complex macros behave like functions, you can replace
190them with inline functions to solve this problem.
192\&\fBgcov\fR creates a logfile called \fI\fIsourcefile\fI.gcov\fR which
193indicates how many times each line of a source file \fI\fIsourcefile\fI.c\fR
194has executed. You can use these logfiles along with \fBgprof\fR to aid
195in fine-tuning the performance of your programs. \fBgprof\fR gives
196timing information you can use along with the information you get from
199\&\fBgcov\fR works only on code compiled with \s-1GCC\s0. It is not
200compatible with any other profiling or test coverage mechanism.
202.IX Header "OPTIONS"
203.IP "\fB\-h\fR" 4
204.IX Item "-h"
205.PD 0
206.IP "\fB\-\-help\fR" 4
207.IX Item "--help"
209Display help about using \fBgcov\fR (on the standard output), and
210exit without doing any further processing.
211.IP "\fB\-v\fR" 4
212.IX Item "-v"
213.PD 0
214.IP "\fB\-\-version\fR" 4
215.IX Item "--version"
217Display the \fBgcov\fR version number (on the standard output),
218and exit without doing any further processing.
219.IP "\fB\-a\fR" 4
220.IX Item "-a"
221.PD 0
222.IP "\fB\-\-all\-blocks\fR" 4
223.IX Item "--all-blocks"
225Write individual execution counts for every basic block. Normally gcov
226outputs execution counts only for the main blocks of a line. With this
227option you can determine if blocks within a single line are not being
229.IP "\fB\-b\fR" 4
230.IX Item "-b"
231.PD 0
232.IP "\fB\-\-branch\-probabilities\fR" 4
233.IX Item "--branch-probabilities"
235Write branch frequencies to the output file, and write branch summary
236info to the standard output. This option allows you to see how often
237each branch in your program was taken. Unconditional branches will not
238be shown, unless the \fB\-u\fR option is given.
239.IP "\fB\-c\fR" 4
240.IX Item "-c"
241.PD 0
242.IP "\fB\-\-branch\-counts\fR" 4
243.IX Item "--branch-counts"
245Write branch frequencies as the number of branches taken, rather than
246the percentage of branches taken.
247.IP "\fB\-n\fR" 4
248.IX Item "-n"
249.PD 0
250.IP "\fB\-\-no\-output\fR" 4
251.IX Item "--no-output"
253Do not create the \fBgcov\fR output file.
254.IP "\fB\-l\fR" 4
255.IX Item "-l"
256.PD 0
257.IP "\fB\-\-long\-file\-names\fR" 4
258.IX Item "--long-file-names"
260Create long file names for included source files. For example, if the
261header file \fIx.h\fR contains code, and was included in the file
262\&\fIa.c\fR, then running \fBgcov\fR on the file \fIa.c\fR will produce
263an output file called \fIa.c##x.h.gcov\fR instead of \fIx.h.gcov\fR.
264This can be useful if \fIx.h\fR is included in multiple source
265files. If you uses the \fB\-p\fR option, both the including and
266included file names will be complete path names.
267.IP "\fB\-p\fR" 4
268.IX Item "-p"
269.PD 0
270.IP "\fB\-\-preserve\-paths\fR" 4
271.IX Item "--preserve-paths"
273Preserve complete path information in the names of generated
274\&\fI.gcov\fR files. Without this option, just the filename component is
275used. With this option, all directories are used, with '/' characters
276translated to '#' characters, '.' directory components removed and '..'
277components renamed to '^'. This is useful if sourcefiles are in several
278different directories. It also affects the \fB\-l\fR option.
279.IP "\fB\-f\fR" 4
280.IX Item "-f"
281.PD 0
282.IP "\fB\-\-function\-summaries\fR" 4
283.IX Item "--function-summaries"
285Output summaries for each function in addition to the file level summary.
286.IP "\fB\-o\fR \fIdirectory|file\fR" 4
287.IX Item "-o directory|file"
288.PD 0
289.IP "\fB\-\-object\-directory\fR \fIdirectory\fR" 4
290.IX Item "--object-directory directory"
291.IP "\fB\-\-object\-file\fR \fIfile\fR" 4
292.IX Item "--object-file file"
294Specify either the directory containing the gcov data files, or the
295object path name. The \fI.gcno\fR, and
296\&\fI.gcda\fR data files are searched for using this option. If a directory
297is specified, the data files are in that directory and named after the
298source file name, without its extension. If a file is specified here,
299the data files are named after that file, without its extension. If this
300option is not supplied, it defaults to the current directory.
301.IP "\fB\-u\fR" 4
302.IX Item "-u"
303.PD 0
304.IP "\fB\-\-unconditional\-branches\fR" 4
305.IX Item "--unconditional-branches"
307When branch counts are given, include those of unconditional branches.
308Unconditional branches are normally not interesting.
310\&\fBgcov\fR should be run with the current directory the same as that
311when you invoked the compiler. Otherwise it will not be able to locate
312the source files. \fBgcov\fR produces files called
313\&\fI\fImangledname\fI.gcov\fR in the current directory. These contain
314the coverage information of the source file they correspond to.
315One \fI.gcov\fR file is produced for each source file containing code,
316which was compiled to produce the data files. The \fImangledname\fR part
317of the output file name is usually simply the source file name, but can
318be something more complicated if the \fB\-l\fR or \fB\-p\fR options are
319given. Refer to those options for details.
321The \fI.gcov\fR files contain the ':' separated fields along with
322program source code. The format is
324.Vb 1
325\& <execution_count>:<line_number>:<source line text>
328Additional block information may succeed each line, when requested by
329command line option. The \fIexecution_count\fR is \fB\-\fR for lines
330containing no code and \fB#####\fR for lines which were never
331executed. Some lines of information at the start have \fIline_number\fR
332of zero.
334When printing percentages, 0% and 100% are only printed when the values
335are \fIexactly\fR 0% and 100% respectively. Other values which would
336conventionally be rounded to 0% or 100% are instead printed as the
337nearest non-boundary value.
339When using \fBgcov\fR, you must first compile your program with two
340special \s-1GCC\s0 options: \fB\-fprofile\-arcs \-ftest\-coverage\fR.
341This tells the compiler to generate additional information needed by
342gcov (basically a flow graph of the program) and also includes
343additional code in the object files for generating the extra profiling
344information needed by gcov. These additional files are placed in the
345directory where the object file is located.
347Running the program will cause profile output to be generated. For each
348source file compiled with \fB\-fprofile\-arcs\fR, an accompanying
349\&\fI.gcda\fR file will be placed in the object file directory.
351Running \fBgcov\fR with your program's source file names as arguments
352will now produce a listing of the code along with frequency of execution
353for each line. For example, if your program is called \fItmp.c\fR, this
354is what you see when you use the basic \fBgcov\fR facility:
356.Vb 5
357\& $ gcc -fprofile-arcs -ftest-coverage tmp.c
358\& $ a.out
359\& $ gcov tmp.c
360\& 90.00% of 10 source lines executed in file tmp.c
361\& Creating tmp.c.gcov.
364The file \fItmp.c.gcov\fR contains output from \fBgcov\fR.
365Here is a sample:
367.Vb 23
368\& -: 0:Source:tmp.c
369\& -: 0:Graph:tmp.gcno
370\& -: 0:Data:tmp.gcda
371\& -: 0:Runs:1
372\& -: 0:Programs:1
373\& -: 1:#include <stdio.h>
374\& -: 2:
375\& -: 3:int main (void)
376\& function main called 1 returned 1 blocks executed 75%
377\& 1: 4:{
378\& 1: 5: int i, total;
379\& -: 6:
380\& 1: 7: total = 0;
381\& -: 8:
382\& 11: 9: for (i = 0; i < 10; i++)
383\& 10: 10: total += i;
384\& -: 11:
385\& 1: 12: if (total != 45)
386\& #####: 13: printf ("Failure\en");
387\& -: 14: else
388\& 1: 15: printf ("Success\en");
389\& 1: 16: return 0;
390\& -: 17:}
393When you use the \fB\-a\fR option, you will get individual block
394counts, and the output looks like this:
396.Vb 30
397\& -: 0:Source:tmp.c
398\& -: 0:Graph:tmp.gcno
399\& -: 0:Data:tmp.gcda
400\& -: 0:Runs:1
401\& -: 0:Programs:1
402\& -: 1:#include <stdio.h>
403\& -: 2:
404\& -: 3:int main (void)
405\& function main called 1 returned 1 blocks executed 75%
406\& 1: 4:{
407\& 1: 4-block 0
408\& 1: 5: int i, total;
409\& -: 6:
410\& 1: 7: total = 0;
411\& -: 8:
412\& 11: 9: for (i = 0; i < 10; i++)
413\& 11: 9-block 0
414\& 10: 10: total += i;
415\& 10: 10-block 0
416\& -: 11:
417\& 1: 12: if (total != 45)
418\& 1: 12-block 0
419\& #####: 13: printf ("Failure\en");
420\& $$$$$: 13-block 0
421\& -: 14: else
422\& 1: 15: printf ("Success\en");
423\& 1: 15-block 0
424\& 1: 16: return 0;
425\& 1: 16-block 0
426\& -: 17:}
429In this mode, each basic block is only shown on one line \*(-- the last
430line of the block. A multi-line block will only contribute to the
431execution count of that last line, and other lines will not be shown
432to contain code, unless previous blocks end on those lines.
433The total execution count of a line is shown and subsequent lines show
434the execution counts for individual blocks that end on that line. After each
435block, the branch and call counts of the block will be shown, if the
436\&\fB\-b\fR option is given.
438Because of the way \s-1GCC\s0 instruments calls, a call count can be shown
439after a line with no individual blocks.
440As you can see, line 13 contains a basic block that was not executed.
442When you use the \fB\-b\fR option, your output looks like this:
444.Vb 6
445\& $ gcov -b tmp.c
446\& 90.00% of 10 source lines executed in file tmp.c
447\& 80.00% of 5 branches executed in file tmp.c
448\& 80.00% of 5 branches taken at least once in file tmp.c
449\& 50.00% of 2 calls executed in file tmp.c
450\& Creating tmp.c.gcov.
453Here is a sample of a resulting \fItmp.c.gcov\fR file:
455.Vb 29
456\& -: 0:Source:tmp.c
457\& -: 0:Graph:tmp.gcno
458\& -: 0:Data:tmp.gcda
459\& -: 0:Runs:1
460\& -: 0:Programs:1
461\& -: 1:#include <stdio.h>
462\& -: 2:
463\& -: 3:int main (void)
464\& function main called 1 returned 1 blocks executed 75%
465\& 1: 4:{
466\& 1: 5: int i, total;
467\& -: 6:
468\& 1: 7: total = 0;
469\& -: 8:
470\& 11: 9: for (i = 0; i < 10; i++)
471\& branch 0 taken 91% (fallthrough)
472\& branch 1 taken 9%
473\& 10: 10: total += i;
474\& -: 11:
475\& 1: 12: if (total != 45)
476\& branch 0 taken 0% (fallthrough)
477\& branch 1 taken 100%
478\& #####: 13: printf ("Failure\en");
479\& call 0 never executed
480\& -: 14: else
481\& 1: 15: printf ("Success\en");
482\& call 0 called 1 returned 100%
483\& 1: 16: return 0;
484\& -: 17:}
487For each basic block, a line is printed after the last line of the basic
488block describing the branch or call that ends the basic block. There can
489be multiple branches and calls listed for a single source line if there
490are multiple basic blocks that end on that line. In this case, the
491branches and calls are each given a number. There is no simple way to map
492these branches and calls back to source constructs. In general, though,
493the lowest numbered branch or call will correspond to the leftmost construct
494on the source line.
496For a branch, if it was executed at least once, then a percentage
497indicating the number of times the branch was taken divided by the
498number of times the branch was executed will be printed. Otherwise, the
499message ``never executed'' is printed.
501For a call, if it was executed at least once, then a percentage
502indicating the number of times the call returned divided by the number
503of times the call was executed will be printed. This will usually be
504100%, but may be less for functions call \f(CW\*(C`exit\*(C'\fR or \f(CW\*(C`longjmp\*(C'\fR,
505and thus may not return every time they are called.
507The execution counts are cumulative. If the example program were
508executed again without removing the \fI.gcda\fR file, the count for the
509number of times each line in the source was executed would be added to
510the results of the previous run(s). This is potentially useful in
511several ways. For example, it could be used to accumulate data over a
512number of program runs as part of a test verification suite, or to
513provide more accurate long-term information over a large number of
514program runs.
516The data in the \fI.gcda\fR files is saved immediately before the program
517exits. For each source file compiled with \fB\-fprofile\-arcs\fR, the
518profiling code first attempts to read in an existing \fI.gcda\fR file; if
519the file doesn't match the executable (differing number of basic block
520counts) it will ignore the contents of the file. It then adds in the
521new execution counts and finally writes the data to the file.
522.Sh "Using \fBgcov\fP with \s-1GCC\s0 Optimization"
523.IX Subsection "Using gcov with GCC Optimization"
524If you plan to use \fBgcov\fR to help optimize your code, you must
525first compile your program with two special \s-1GCC\s0 options:
526\&\fB\-fprofile\-arcs \-ftest\-coverage\fR. Aside from that, you can use any
527other \s-1GCC\s0 options; but if you want to prove that every single line
528in your program was executed, you should not compile with optimization
529at the same time. On some machines the optimizer can eliminate some
530simple code lines by combining them with other lines. For example, code
531like this:
533.Vb 4
534\& if (a != b)
535\& c = 1;
536\& else
537\& c = 0;
540can be compiled into one instruction on some machines. In this case,
541there is no way for \fBgcov\fR to calculate separate execution counts
542for each line because there isn't separate code for each line. Hence
543the \fBgcov\fR output looks like this if you compiled the program with
546.Vb 4
547\& 100: 12:if (a != b)
548\& 100: 13: c = 1;
549\& 100: 14:else
550\& 100: 15: c = 0;
553The output shows that this block of code, combined by optimization,
554executed 100 times. In one sense this result is correct, because there
555was only one instruction representing all four of these lines. However,
556the output does not indicate how many times the result was 0 and how
557many times the result was 1.
559Inlineable functions can create unexpected line counts. Line counts are
560shown for the source code of the inlineable function, but what is shown
561depends on where the function is inlined, or if it is not inlined at all.
563If the function is not inlined, the compiler must emit an out of line
564copy of the function, in any object file that needs it. If
565\&\fIfileA.o\fR and \fIfileB.o\fR both contain out of line bodies of a
566particular inlineable function, they will also both contain coverage
567counts for that function. When \fIfileA.o\fR and \fIfileB.o\fR are
568linked together, the linker will, on many systems, select one of those
569out of line bodies for all calls to that function, and remove or ignore
570the other. Unfortunately, it will not remove the coverage counters for
571the unused function body. Hence when instrumented, all but one use of
572that function will show zero counts.
574If the function is inlined in several places, the block structure in
575each location might not be the same. For instance, a condition might
576now be calculable at compile time in some instances. Because the
577coverage of all the uses of the inline function will be shown for the
578same source lines, the line counts themselves might seem inconsistent.
580.IX Header "SEE ALSO"
581\&\fIgpl\fR\|(7), \fIgfdl\fR\|(7), \fIfsf\-funding\fR\|(7), \fIgcc\fR\|(1) and the Info entry for \fIgcc\fR.
583.IX Header "COPYRIGHT"
584Copyright (c) 1996, 1997, 1999, 2000, 2001, 2002, 2003
585Free Software Foundation, Inc.
587Permission is granted to copy, distribute and/or modify this document
588under the terms of the \s-1GNU\s0 Free Documentation License, Version 1.2 or
589any later version published by the Free Software Foundation; with the
590Invariant Sections being ``\s-1GNU\s0 General Public License'' and ``Funding
591Free Software'', the Front-Cover texts being (a) (see below), and with
592the Back-Cover Texts being (b) (see below). A copy of the license is
593included in the \fIgfdl\fR\|(7) man page.
595(a) The \s-1FSF\s0's Front-Cover Text is:
597.Vb 1
598\& A GNU Manual
601(b) The \s-1FSF\s0's Back-Cover Text is:
603.Vb 3
604\& You have freedom to copy and modify this GNU Manual, like GNU
605\& software. Copies published by the Free Software Foundation raise
606\& funds for GNU development.