1@c Copyright (C) 1996, 1997, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2@c 2008, 2010 Free Software Foundation, Inc. 3@c This is part of the GCC manual. 4@c For copying conditions, see the file gcc.texi. 5 6@ignore 7@c man begin COPYRIGHT 8Copyright @copyright{} 1996, 1997, 1999, 2000, 2001, 2002, 2003, 2004, 92005, 2008, 2010 Free Software Foundation, Inc. 10 11Permission is granted to copy, distribute and/or modify this document 12under the terms of the GNU Free Documentation License, Version 1.3 or 13any later version published by the Free Software Foundation; with the 14Invariant Sections being ``GNU General Public License'' and ``Funding 15Free Software'', the Front-Cover texts being (a) (see below), and with 16the Back-Cover Texts being (b) (see below). A copy of the license is 17included in the gfdl(7) man page. 18 19(a) The FSF's Front-Cover Text is: 20 21 A GNU Manual 22 23(b) The FSF's Back-Cover Text is: 24 25 You have freedom to copy and modify this GNU Manual, like GNU 26 software. Copies published by the Free Software Foundation raise 27 funds for GNU development. 28@c man end 29@c Set file name and title for the man page. 30@setfilename gcov 31@settitle coverage testing tool 32@end ignore 33 34@node Gcov 35@chapter @command{gcov}---a Test Coverage Program 36 37@command{gcov} is a tool you can use in conjunction with GCC to 38test code coverage in your programs. 39 40@menu 41* Gcov Intro:: Introduction to gcov. 42* Invoking Gcov:: How to use gcov. 43* Gcov and Optimization:: Using gcov with GCC optimization. 44* Gcov Data Files:: The files used by gcov. 45* Cross-profiling:: Data file relocation. 46@end menu 47 48@node Gcov Intro 49@section Introduction to @command{gcov} 50@c man begin DESCRIPTION 51 52@command{gcov} is a test coverage program. Use it in concert with GCC 53to analyze your programs to help create more efficient, faster running 54code and to discover untested parts of your program. You can use 55@command{gcov} as a profiling tool to help discover where your 56optimization efforts will best affect your code. You can also use 57@command{gcov} along with the other profiling tool, @command{gprof}, to 58assess which parts of your code use the greatest amount of computing 59time. 60 61Profiling tools help you analyze your code's performance. Using a 62profiler such as @command{gcov} or @command{gprof}, you can find out some 63basic performance statistics, such as: 64 65@itemize @bullet 66@item 67how often each line of code executes 68 69@item 70what lines of code are actually executed 71 72@item 73how much computing time each section of code uses 74@end itemize 75 76Once you know these things about how your code works when compiled, you 77can look at each module to see which modules should be optimized. 78@command{gcov} helps you determine where to work on optimization. 79 80Software developers also use coverage testing in concert with 81testsuites, to make sure software is actually good enough for a release. 82Testsuites can verify that a program works as expected; a coverage 83program tests to see how much of the program is exercised by the 84testsuite. Developers can then determine what kinds of test cases need 85to be added to the testsuites to create both better testing and a better 86final product. 87 88You should compile your code without optimization if you plan to use 89@command{gcov} because the optimization, by combining some lines of code 90into one function, may not give you as much information as you need to 91look for `hot spots' where the code is using a great deal of computer 92time. Likewise, because @command{gcov} accumulates statistics by line (at 93the lowest resolution), it works best with a programming style that 94places only one statement on each line. If you use complicated macros 95that expand to loops or to other control structures, the statistics are 96less helpful---they only report on the line where the macro call 97appears. If your complex macros behave like functions, you can replace 98them with inline functions to solve this problem. 99 100@command{gcov} creates a logfile called @file{@var{sourcefile}.gcov} which 101indicates how many times each line of a source file @file{@var{sourcefile}.c} 102has executed. You can use these logfiles along with @command{gprof} to aid 103in fine-tuning the performance of your programs. @command{gprof} gives 104timing information you can use along with the information you get from 105@command{gcov}. 106 107@command{gcov} works only on code compiled with GCC@. It is not 108compatible with any other profiling or test coverage mechanism. 109 110@c man end 111 112@node Invoking Gcov 113@section Invoking @command{gcov} 114 115@smallexample 116gcov @r{[}@var{options}@r{]} @var{files} 117@end smallexample 118 119@command{gcov} accepts the following options: 120 121@ignore 122@c man begin SYNOPSIS 123gcov [@option{-v}|@option{--version}] [@option{-h}|@option{--help}] 124 [@option{-a}|@option{--all-blocks}] 125 [@option{-b}|@option{--branch-probabilities}] 126 [@option{-c}|@option{--branch-counts}] 127 [@option{-u}|@option{--unconditional-branches}] 128 [@option{-n}|@option{--no-output}] 129 [@option{-l}|@option{--long-file-names}] 130 [@option{-p}|@option{--preserve-paths}] 131 [@option{-r}|@option{--relative-only}] 132 [@option{-f}|@option{--function-summaries}] 133 [@option{-o}|@option{--object-directory} @var{directory|file}] 134 [@option{-s}|@option{--source-prefix} @var{directory}] 135 [@option{-d}|@option{--display-progress}] 136 @var{files} 137@c man end 138@c man begin SEEALSO 139gpl(7), gfdl(7), fsf-funding(7), gcc(1) and the Info entry for @file{gcc}. 140@c man end 141@end ignore 142 143@c man begin OPTIONS 144@table @gcctabopt 145@item -h 146@itemx --help 147Display help about using @command{gcov} (on the standard output), and 148exit without doing any further processing. 149 150@item -v 151@itemx --version 152Display the @command{gcov} version number (on the standard output), 153and exit without doing any further processing. 154 155@item -a 156@itemx --all-blocks 157Write individual execution counts for every basic block. Normally gcov 158outputs execution counts only for the main blocks of a line. With this 159option you can determine if blocks within a single line are not being 160executed. 161 162@item -b 163@itemx --branch-probabilities 164Write branch frequencies to the output file, and write branch summary 165info to the standard output. This option allows you to see how often 166each branch in your program was taken. Unconditional branches will not 167be shown, unless the @option{-u} option is given. 168 169@item -c 170@itemx --branch-counts 171Write branch frequencies as the number of branches taken, rather than 172the percentage of branches taken. 173 174@item -n 175@itemx --no-output 176Do not create the @command{gcov} output file. 177 178@item -l 179@itemx --long-file-names 180Create long file names for included source files. For example, if the 181header file @file{x.h} contains code, and was included in the file 182@file{a.c}, then running @command{gcov} on the file @file{a.c} will 183produce an output file called @file{a.c##x.h.gcov} instead of 184@file{x.h.gcov}. This can be useful if @file{x.h} is included in 185multiple source files and you want to see the individual 186contributions. If you use the @samp{-p} option, both the including 187and included file names will be complete path names. 188 189@item -p 190@itemx --preserve-paths 191Preserve complete path information in the names of generated 192@file{.gcov} files. Without this option, just the filename component is 193used. With this option, all directories are used, with @samp{/} characters 194translated to @samp{#} characters, @file{.} directory components 195removed and unremoveable @file{..} 196components renamed to @samp{^}. This is useful if sourcefiles are in several 197different directories. 198 199@item -r 200@itemx --relative-only 201Only output information about source files with a relative pathname 202(after source prefix elision). Absolute paths are usually system 203header files and coverage of any inline functions therein is normally 204uninteresting. 205 206@item -f 207@itemx --function-summaries 208Output summaries for each function in addition to the file level summary. 209 210@item -o @var{directory|file} 211@itemx --object-directory @var{directory} 212@itemx --object-file @var{file} 213Specify either the directory containing the gcov data files, or the 214object path name. The @file{.gcno}, and 215@file{.gcda} data files are searched for using this option. If a directory 216is specified, the data files are in that directory and named after the 217input file name, without its extension. If a file is specified here, 218the data files are named after that file, without its extension. 219 220@item -s @var{directory} 221@itemx --source-prefix @var{directory} 222A prefix for source file names to remove when generating the output 223coverage files. This option is useful when building in a separate 224directory, and the pathname to the source directory is not wanted when 225determining the output file names. Note that this prefix detection is 226applied before determining whether the source file is absolute. 227 228@item -u 229@itemx --unconditional-branches 230When branch probabilities are given, include those of unconditional branches. 231Unconditional branches are normally not interesting. 232 233@item -d 234@itemx --display-progress 235Display the progress on the standard output. 236 237@end table 238 239@command{gcov} should be run with the current directory the same as that 240when you invoked the compiler. Otherwise it will not be able to locate 241the source files. @command{gcov} produces files called 242@file{@var{mangledname}.gcov} in the current directory. These contain 243the coverage information of the source file they correspond to. 244One @file{.gcov} file is produced for each source (or header) file 245containing code, 246which was compiled to produce the data files. The @var{mangledname} part 247of the output file name is usually simply the source file name, but can 248be something more complicated if the @samp{-l} or @samp{-p} options are 249given. Refer to those options for details. 250 251If you invoke @command{gcov} with multiple input files, the 252contributions from each input file are summed. Typically you would 253invoke it with the same list of files as the final link of your executable. 254 255The @file{.gcov} files contain the @samp{:} separated fields along with 256program source code. The format is 257 258@smallexample 259@var{execution_count}:@var{line_number}:@var{source line text} 260@end smallexample 261 262Additional block information may succeed each line, when requested by 263command line option. The @var{execution_count} is @samp{-} for lines 264containing no code. Unexecuted lines are marked @samp{#####} or 265@samp{====}, depending on whether they are reachable by 266non-exceptional paths or only exceptional paths such as C++ exception 267handlers, respectively. 268 269Some lines of information at the start have @var{line_number} of zero. 270These preamble lines are of the form 271 272@smallexample 273-:0:@var{tag}:@var{value} 274@end smallexample 275 276The ordering and number of these preamble lines will be augmented as 277@command{gcov} development progresses --- do not rely on them remaining 278unchanged. Use @var{tag} to locate a particular preamble line. 279 280The additional block information is of the form 281 282@smallexample 283@var{tag} @var{information} 284@end smallexample 285 286The @var{information} is human readable, but designed to be simple 287enough for machine parsing too. 288 289When printing percentages, 0% and 100% are only printed when the values 290are @emph{exactly} 0% and 100% respectively. Other values which would 291conventionally be rounded to 0% or 100% are instead printed as the 292nearest non-boundary value. 293 294When using @command{gcov}, you must first compile your program with two 295special GCC options: @samp{-fprofile-arcs -ftest-coverage}. 296This tells the compiler to generate additional information needed by 297gcov (basically a flow graph of the program) and also includes 298additional code in the object files for generating the extra profiling 299information needed by gcov. These additional files are placed in the 300directory where the object file is located. 301 302Running the program will cause profile output to be generated. For each 303source file compiled with @option{-fprofile-arcs}, an accompanying 304@file{.gcda} file will be placed in the object file directory. 305 306Running @command{gcov} with your program's source file names as arguments 307will now produce a listing of the code along with frequency of execution 308for each line. For example, if your program is called @file{tmp.c}, this 309is what you see when you use the basic @command{gcov} facility: 310 311@smallexample 312$ gcc -fprofile-arcs -ftest-coverage tmp.c 313$ a.out 314$ gcov tmp.c 31590.00% of 10 source lines executed in file tmp.c 316Creating tmp.c.gcov. 317@end smallexample 318 319The file @file{tmp.c.gcov} contains output from @command{gcov}. 320Here is a sample: 321 322@smallexample 323 -: 0:Source:tmp.c 324 -: 0:Graph:tmp.gcno 325 -: 0:Data:tmp.gcda 326 -: 0:Runs:1 327 -: 0:Programs:1 328 -: 1:#include <stdio.h> 329 -: 2: 330 -: 3:int main (void) 331 1: 4:@{ 332 1: 5: int i, total; 333 -: 6: 334 1: 7: total = 0; 335 -: 8: 336 11: 9: for (i = 0; i < 10; i++) 337 10: 10: total += i; 338 -: 11: 339 1: 12: if (total != 45) 340 #####: 13: printf ("Failure\n"); 341 -: 14: else 342 1: 15: printf ("Success\n"); 343 1: 16: return 0; 344 -: 17:@} 345@end smallexample 346 347When you use the @option{-a} option, you will get individual block 348counts, and the output looks like this: 349 350@smallexample 351 -: 0:Source:tmp.c 352 -: 0:Graph:tmp.gcno 353 -: 0:Data:tmp.gcda 354 -: 0:Runs:1 355 -: 0:Programs:1 356 -: 1:#include <stdio.h> 357 -: 2: 358 -: 3:int main (void) 359 1: 4:@{ 360 1: 4-block 0 361 1: 5: int i, total; 362 -: 6: 363 1: 7: total = 0; 364 -: 8: 365 11: 9: for (i = 0; i < 10; i++) 366 11: 9-block 0 367 10: 10: total += i; 368 10: 10-block 0 369 -: 11: 370 1: 12: if (total != 45) 371 1: 12-block 0 372 #####: 13: printf ("Failure\n"); 373 $$$$$: 13-block 0 374 -: 14: else 375 1: 15: printf ("Success\n"); 376 1: 15-block 0 377 1: 16: return 0; 378 1: 16-block 0 379 -: 17:@} 380@end smallexample 381 382In this mode, each basic block is only shown on one line -- the last 383line of the block. A multi-line block will only contribute to the 384execution count of that last line, and other lines will not be shown 385to contain code, unless previous blocks end on those lines. 386The total execution count of a line is shown and subsequent lines show 387the execution counts for individual blocks that end on that line. After each 388block, the branch and call counts of the block will be shown, if the 389@option{-b} option is given. 390 391Because of the way GCC instruments calls, a call count can be shown 392after a line with no individual blocks. 393As you can see, line 13 contains a basic block that was not executed. 394 395@need 450 396When you use the @option{-b} option, your output looks like this: 397 398@smallexample 399$ gcov -b tmp.c 40090.00% of 10 source lines executed in file tmp.c 40180.00% of 5 branches executed in file tmp.c 40280.00% of 5 branches taken at least once in file tmp.c 40350.00% of 2 calls executed in file tmp.c 404Creating tmp.c.gcov. 405@end smallexample 406 407Here is a sample of a resulting @file{tmp.c.gcov} file: 408 409@smallexample 410 -: 0:Source:tmp.c 411 -: 0:Graph:tmp.gcno 412 -: 0:Data:tmp.gcda 413 -: 0:Runs:1 414 -: 0:Programs:1 415 -: 1:#include <stdio.h> 416 -: 2: 417 -: 3:int main (void) 418function main called 1 returned 1 blocks executed 75% 419 1: 4:@{ 420 1: 5: int i, total; 421 -: 6: 422 1: 7: total = 0; 423 -: 8: 424 11: 9: for (i = 0; i < 10; i++) 425branch 0 taken 91% (fallthrough) 426branch 1 taken 9% 427 10: 10: total += i; 428 -: 11: 429 1: 12: if (total != 45) 430branch 0 taken 0% (fallthrough) 431branch 1 taken 100% 432 #####: 13: printf ("Failure\n"); 433call 0 never executed 434 -: 14: else 435 1: 15: printf ("Success\n"); 436call 0 called 1 returned 100% 437 1: 16: return 0; 438 -: 17:@} 439@end smallexample 440 441For each function, a line is printed showing how many times the function 442is called, how many times it returns and what percentage of the 443function's blocks were executed. 444 445For each basic block, a line is printed after the last line of the basic 446block describing the branch or call that ends the basic block. There can 447be multiple branches and calls listed for a single source line if there 448are multiple basic blocks that end on that line. In this case, the 449branches and calls are each given a number. There is no simple way to map 450these branches and calls back to source constructs. In general, though, 451the lowest numbered branch or call will correspond to the leftmost construct 452on the source line. 453 454For a branch, if it was executed at least once, then a percentage 455indicating the number of times the branch was taken divided by the 456number of times the branch was executed will be printed. Otherwise, the 457message ``never executed'' is printed. 458 459For a call, if it was executed at least once, then a percentage 460indicating the number of times the call returned divided by the number 461of times the call was executed will be printed. This will usually be 462100%, but may be less for functions that call @code{exit} or @code{longjmp}, 463and thus may not return every time they are called. 464 465The execution counts are cumulative. If the example program were 466executed again without removing the @file{.gcda} file, the count for the 467number of times each line in the source was executed would be added to 468the results of the previous run(s). This is potentially useful in 469several ways. For example, it could be used to accumulate data over a 470number of program runs as part of a test verification suite, or to 471provide more accurate long-term information over a large number of 472program runs. 473 474The data in the @file{.gcda} files is saved immediately before the program 475exits. For each source file compiled with @option{-fprofile-arcs}, the 476profiling code first attempts to read in an existing @file{.gcda} file; if 477the file doesn't match the executable (differing number of basic block 478counts) it will ignore the contents of the file. It then adds in the 479new execution counts and finally writes the data to the file. 480 481@node Gcov and Optimization 482@section Using @command{gcov} with GCC Optimization 483 484If you plan to use @command{gcov} to help optimize your code, you must 485first compile your program with two special GCC options: 486@samp{-fprofile-arcs -ftest-coverage}. Aside from that, you can use any 487other GCC options; but if you want to prove that every single line 488in your program was executed, you should not compile with optimization 489at the same time. On some machines the optimizer can eliminate some 490simple code lines by combining them with other lines. For example, code 491like this: 492 493@smallexample 494if (a != b) 495 c = 1; 496else 497 c = 0; 498@end smallexample 499 500@noindent 501can be compiled into one instruction on some machines. In this case, 502there is no way for @command{gcov} to calculate separate execution counts 503for each line because there isn't separate code for each line. Hence 504the @command{gcov} output looks like this if you compiled the program with 505optimization: 506 507@smallexample 508 100: 12:if (a != b) 509 100: 13: c = 1; 510 100: 14:else 511 100: 15: c = 0; 512@end smallexample 513 514The output shows that this block of code, combined by optimization, 515executed 100 times. In one sense this result is correct, because there 516was only one instruction representing all four of these lines. However, 517the output does not indicate how many times the result was 0 and how 518many times the result was 1. 519 520Inlineable functions can create unexpected line counts. Line counts are 521shown for the source code of the inlineable function, but what is shown 522depends on where the function is inlined, or if it is not inlined at all. 523 524If the function is not inlined, the compiler must emit an out of line 525copy of the function, in any object file that needs it. If 526@file{fileA.o} and @file{fileB.o} both contain out of line bodies of a 527particular inlineable function, they will also both contain coverage 528counts for that function. When @file{fileA.o} and @file{fileB.o} are 529linked together, the linker will, on many systems, select one of those 530out of line bodies for all calls to that function, and remove or ignore 531the other. Unfortunately, it will not remove the coverage counters for 532the unused function body. Hence when instrumented, all but one use of 533that function will show zero counts. 534 535If the function is inlined in several places, the block structure in 536each location might not be the same. For instance, a condition might 537now be calculable at compile time in some instances. Because the 538coverage of all the uses of the inline function will be shown for the 539same source lines, the line counts themselves might seem inconsistent. 540 541@c man end 542 543@node Gcov Data Files 544@section Brief description of @command{gcov} data files 545 546@command{gcov} uses two files for profiling. The names of these files 547are derived from the original @emph{object} file by substituting the 548file suffix with either @file{.gcno}, or @file{.gcda}. All of these files 549are placed in the same directory as the object file, and contain data 550stored in a platform-independent format. 551 552The @file{.gcno} file is generated when the source file is compiled with 553the GCC @option{-ftest-coverage} option. It contains information to 554reconstruct the basic block graphs and assign source line numbers to 555blocks. 556 557The @file{.gcda} file is generated when a program containing object files 558built with the GCC @option{-fprofile-arcs} option is executed. A 559separate @file{.gcda} file is created for each object file compiled with 560this option. It contains arc transition counts, and some summary 561information. 562 563The full details of the file format is specified in @file{gcov-io.h}, 564and functions provided in that header file should be used to access the 565coverage files. 566 567@node Cross-profiling 568@section Data file relocation to support cross-profiling 569 570Running the program will cause profile output to be generated. For each 571source file compiled with @option{-fprofile-arcs}, an accompanying @file{.gcda} 572file will be placed in the object file directory. That implicitly requires 573running the program on the same system as it was built or having the same 574absolute directory structure on the target system. The program will try 575to create the needed directory structure, if it is not already present. 576 577To support cross-profiling, a program compiled with @option{-fprofile-arcs} 578can relocate the data files based on two environment variables: 579 580@itemize @bullet 581@item 582GCOV_PREFIX contains the prefix to add to the absolute paths 583in the object file. Prefix can be absolute, or relative. The 584default is no prefix. 585 586@item 587GCOV_PREFIX_STRIP indicates the how many initial directory names to strip off 588the hardwired absolute paths. Default value is 0. 589 590@emph{Note:} If GCOV_PREFIX_STRIP is set without GCOV_PREFIX is undefined, 591 then a relative path is made out of the hardwired absolute paths. 592@end itemize 593 594For example, if the object file @file{/user/build/foo.o} was built with 595@option{-fprofile-arcs}, the final executable will try to create the data file 596@file{/user/build/foo.gcda} when running on the target system. This will 597fail if the corresponding directory does not exist and it is unable to create 598it. This can be overcome by, for example, setting the environment as 599@samp{GCOV_PREFIX=/target/run} and @samp{GCOV_PREFIX_STRIP=1}. Such a 600setting will name the data file @file{/target/run/build/foo.gcda}. 601 602You must move the data files to the expected directory tree in order to 603use them for profile directed optimizations (@option{--use-profile}), or to 604use the @command{gcov} tool. 605