Source file src/cmd/cgo/doc.go
1 // Copyright 2009 The Go Authors. All rights reserved. 2 // Use of this source code is governed by a BSD-style 3 // license that can be found in the LICENSE file. 4 5 /* 6 Cgo enables the creation of Go packages that call C code. 7 8 # Using cgo with the go command 9 10 To use cgo write normal Go code that imports a pseudo-package "C". 11 The Go code can then refer to types such as C.size_t, variables such 12 as C.stdout, or functions such as C.putchar. 13 14 If the import of "C" is immediately preceded by a comment, that 15 comment, called the preamble, is used as a header when compiling 16 the C parts of the package. For example: 17 18 // #include <stdio.h> 19 // #include <errno.h> 20 import "C" 21 22 The preamble may contain any C code, including function and variable 23 declarations and definitions. These may then be referred to from Go 24 code as though they were defined in the package "C". All names 25 declared in the preamble may be used, even if they start with a 26 lower-case letter. Exception: static variables in the preamble may 27 not be referenced from Go code; static functions are permitted. 28 29 See $GOROOT/cmd/cgo/internal/teststdio and $GOROOT/misc/cgo/gmp for examples. See 30 "C? Go? Cgo!" for an introduction to using cgo: 31 https://golang.org/doc/articles/c_go_cgo.html. 32 33 CFLAGS, CPPFLAGS, CXXFLAGS, FFLAGS and LDFLAGS may be defined with pseudo 34 #cgo directives within these comments to tweak the behavior of the C, C++ 35 or Fortran compiler. Values defined in multiple directives are concatenated 36 together. The directive can include a list of build constraints limiting its 37 effect to systems satisfying one of the constraints 38 (see https://golang.org/pkg/go/build/#hdr-Build_Constraints for details about the constraint syntax). 39 For example: 40 41 // #cgo CFLAGS: -DPNG_DEBUG=1 42 // #cgo amd64 386 CFLAGS: -DX86=1 43 // #cgo LDFLAGS: -lpng 44 // #include <png.h> 45 import "C" 46 47 Alternatively, CPPFLAGS and LDFLAGS may be obtained via the pkg-config tool 48 using a '#cgo pkg-config:' directive followed by the package names. 49 For example: 50 51 // #cgo pkg-config: png cairo 52 // #include <png.h> 53 import "C" 54 55 The default pkg-config tool may be changed by setting the PKG_CONFIG environment variable. 56 57 For security reasons, only a limited set of flags are allowed, notably -D, -U, -I, and -l. 58 To allow additional flags, set CGO_CFLAGS_ALLOW to a regular expression 59 matching the new flags. To disallow flags that would otherwise be allowed, 60 set CGO_CFLAGS_DISALLOW to a regular expression matching arguments 61 that must be disallowed. In both cases the regular expression must match 62 a full argument: to allow -mfoo=bar, use CGO_CFLAGS_ALLOW='-mfoo.*', 63 not just CGO_CFLAGS_ALLOW='-mfoo'. Similarly named variables control 64 the allowed CPPFLAGS, CXXFLAGS, FFLAGS, and LDFLAGS. 65 66 Also for security reasons, only a limited set of characters are 67 permitted, notably alphanumeric characters and a few symbols, such as 68 '.', that will not be interpreted in unexpected ways. Attempts to use 69 forbidden characters will get a "malformed #cgo argument" error. 70 71 When building, the CGO_CFLAGS, CGO_CPPFLAGS, CGO_CXXFLAGS, CGO_FFLAGS and 72 CGO_LDFLAGS environment variables are added to the flags derived from 73 these directives. Package-specific flags should be set using the 74 directives, not the environment variables, so that builds work in 75 unmodified environments. Flags obtained from environment variables 76 are not subject to the security limitations described above. 77 78 All the cgo CPPFLAGS and CFLAGS directives in a package are concatenated and 79 used to compile C files in that package. All the CPPFLAGS and CXXFLAGS 80 directives in a package are concatenated and used to compile C++ files in that 81 package. All the CPPFLAGS and FFLAGS directives in a package are concatenated 82 and used to compile Fortran files in that package. All the LDFLAGS directives 83 in any package in the program are concatenated and used at link time. All the 84 pkg-config directives are concatenated and sent to pkg-config simultaneously 85 to add to each appropriate set of command-line flags. 86 87 When the cgo directives are parsed, any occurrence of the string ${SRCDIR} 88 will be replaced by the absolute path to the directory containing the source 89 file. This allows pre-compiled static libraries to be included in the package 90 directory and linked properly. 91 For example if package foo is in the directory /go/src/foo: 92 93 // #cgo LDFLAGS: -L${SRCDIR}/libs -lfoo 94 95 Will be expanded to: 96 97 // #cgo LDFLAGS: -L/go/src/foo/libs -lfoo 98 99 When the Go tool sees that one or more Go files use the special import 100 "C", it will look for other non-Go files in the directory and compile 101 them as part of the Go package. Any .c, .s, .S or .sx files will be 102 compiled with the C compiler. Any .cc, .cpp, or .cxx files will be 103 compiled with the C++ compiler. Any .f, .F, .for or .f90 files will be 104 compiled with the fortran compiler. Any .h, .hh, .hpp, or .hxx files will 105 not be compiled separately, but, if these header files are changed, 106 the package (including its non-Go source files) will be recompiled. 107 Note that changes to files in other directories do not cause the package 108 to be recompiled, so all non-Go source code for the package should be 109 stored in the package directory, not in subdirectories. 110 The default C and C++ compilers may be changed by the CC and CXX 111 environment variables, respectively; those environment variables 112 may include command line options. 113 114 The cgo tool will always invoke the C compiler with the source file's 115 directory in the include path; i.e. -I${SRCDIR} is always implied. This 116 means that if a header file foo/bar.h exists both in the source 117 directory and also in the system include directory (or some other place 118 specified by a -I flag), then "#include <foo/bar.h>" will always find the 119 local version in preference to any other version. 120 121 The cgo tool is enabled by default for native builds on systems where 122 it is expected to work. It is disabled by default when cross-compiling 123 as well as when the CC environment variable is unset and the default 124 C compiler (typically gcc or clang) cannot be found on the system PATH. 125 You can override the default by setting the CGO_ENABLED 126 environment variable when running the go tool: set it to 1 to enable 127 the use of cgo, and to 0 to disable it. The go tool will set the 128 build constraint "cgo" if cgo is enabled. The special import "C" 129 implies the "cgo" build constraint, as though the file also said 130 "//go:build cgo". Therefore, if cgo is disabled, files that import 131 "C" will not be built by the go tool. (For more about build constraints 132 see https://golang.org/pkg/go/build/#hdr-Build_Constraints). 133 134 When cross-compiling, you must specify a C cross-compiler for cgo to 135 use. You can do this by setting the generic CC_FOR_TARGET or the 136 more specific CC_FOR_${GOOS}_${GOARCH} (for example, CC_FOR_linux_arm) 137 environment variable when building the toolchain using make.bash, 138 or you can set the CC environment variable any time you run the go tool. 139 140 The CXX_FOR_TARGET, CXX_FOR_${GOOS}_${GOARCH}, and CXX 141 environment variables work in a similar way for C++ code. 142 143 # Go references to C 144 145 Within the Go file, C's struct field names that are keywords in Go 146 can be accessed by prefixing them with an underscore: if x points at a C 147 struct with a field named "type", x._type accesses the field. 148 C struct fields that cannot be expressed in Go, such as bit fields 149 or misaligned data, are omitted in the Go struct, replaced by 150 appropriate padding to reach the next field or the end of the struct. 151 152 The standard C numeric types are available under the names 153 C.char, C.schar (signed char), C.uchar (unsigned char), 154 C.short, C.ushort (unsigned short), C.int, C.uint (unsigned int), 155 C.long, C.ulong (unsigned long), C.longlong (long long), 156 C.ulonglong (unsigned long long), C.float, C.double, 157 C.complexfloat (complex float), and C.complexdouble (complex double). 158 The C type void* is represented by Go's unsafe.Pointer. 159 The C types __int128_t and __uint128_t are represented by [16]byte. 160 161 A few special C types which would normally be represented by a pointer 162 type in Go are instead represented by a uintptr. See the Special 163 cases section below. 164 165 To access a struct, union, or enum type directly, prefix it with 166 struct_, union_, or enum_, as in C.struct_stat. 167 168 The size of any C type T is available as C.sizeof_T, as in 169 C.sizeof_struct_stat. 170 171 A C function may be declared in the Go file with a parameter type of 172 the special name _GoString_. This function may be called with an 173 ordinary Go string value. The string length, and a pointer to the 174 string contents, may be accessed by calling the C functions 175 176 size_t _GoStringLen(_GoString_ s); 177 const char *_GoStringPtr(_GoString_ s); 178 179 These functions are only available in the preamble, not in other C 180 files. The C code must not modify the contents of the pointer returned 181 by _GoStringPtr. Note that the string contents may not have a trailing 182 NUL byte. 183 184 As Go doesn't have support for C's union type in the general case, 185 C's union types are represented as a Go byte array with the same length. 186 187 Go structs cannot embed fields with C types. 188 189 Go code cannot refer to zero-sized fields that occur at the end of 190 non-empty C structs. To get the address of such a field (which is the 191 only operation you can do with a zero-sized field) you must take the 192 address of the struct and add the size of the struct. 193 194 Cgo translates C types into equivalent unexported Go types. 195 Because the translations are unexported, a Go package should not 196 expose C types in its exported API: a C type used in one Go package 197 is different from the same C type used in another. 198 199 Any C function (even void functions) may be called in a multiple 200 assignment context to retrieve both the return value (if any) and the 201 C errno variable as an error (use _ to skip the result value if the 202 function returns void). For example: 203 204 n, err = C.sqrt(-1) 205 _, err := C.voidFunc() 206 var n, err = C.sqrt(1) 207 208 Calling C function pointers is currently not supported, however you can 209 declare Go variables which hold C function pointers and pass them 210 back and forth between Go and C. C code may call function pointers 211 received from Go. For example: 212 213 package main 214 215 // typedef int (*intFunc) (); 216 // 217 // int 218 // bridge_int_func(intFunc f) 219 // { 220 // return f(); 221 // } 222 // 223 // int fortytwo() 224 // { 225 // return 42; 226 // } 227 import "C" 228 import "fmt" 229 230 func main() { 231 f := C.intFunc(C.fortytwo) 232 fmt.Println(int(C.bridge_int_func(f))) 233 // Output: 42 234 } 235 236 In C, a function argument written as a fixed size array 237 actually requires a pointer to the first element of the array. 238 C compilers are aware of this calling convention and adjust 239 the call accordingly, but Go cannot. In Go, you must pass 240 the pointer to the first element explicitly: C.f(&C.x[0]). 241 242 Calling variadic C functions is not supported. It is possible to 243 circumvent this by using a C function wrapper. For example: 244 245 package main 246 247 // #include <stdio.h> 248 // #include <stdlib.h> 249 // 250 // static void myprint(char* s) { 251 // printf("%s\n", s); 252 // } 253 import "C" 254 import "unsafe" 255 256 func main() { 257 cs := C.CString("Hello from stdio") 258 C.myprint(cs) 259 C.free(unsafe.Pointer(cs)) 260 } 261 262 A few special functions convert between Go and C types 263 by making copies of the data. In pseudo-Go definitions: 264 265 // Go string to C string 266 // The C string is allocated in the C heap using malloc. 267 // It is the caller's responsibility to arrange for it to be 268 // freed, such as by calling C.free (be sure to include stdlib.h 269 // if C.free is needed). 270 func C.CString(string) *C.char 271 272 // Go []byte slice to C array 273 // The C array is allocated in the C heap using malloc. 274 // It is the caller's responsibility to arrange for it to be 275 // freed, such as by calling C.free (be sure to include stdlib.h 276 // if C.free is needed). 277 func C.CBytes([]byte) unsafe.Pointer 278 279 // C string to Go string 280 func C.GoString(*C.char) string 281 282 // C data with explicit length to Go string 283 func C.GoStringN(*C.char, C.int) string 284 285 // C data with explicit length to Go []byte 286 func C.GoBytes(unsafe.Pointer, C.int) []byte 287 288 As a special case, C.malloc does not call the C library malloc directly 289 but instead calls a Go helper function that wraps the C library malloc 290 but guarantees never to return nil. If C's malloc indicates out of memory, 291 the helper function crashes the program, like when Go itself runs out 292 of memory. Because C.malloc cannot fail, it has no two-result form 293 that returns errno. 294 295 # C references to Go 296 297 Go functions can be exported for use by C code in the following way: 298 299 //export MyFunction 300 func MyFunction(arg1, arg2 int, arg3 string) int64 {...} 301 302 //export MyFunction2 303 func MyFunction2(arg1, arg2 int, arg3 string) (int64, *C.char) {...} 304 305 They will be available in the C code as: 306 307 extern GoInt64 MyFunction(int arg1, int arg2, GoString arg3); 308 extern struct MyFunction2_return MyFunction2(int arg1, int arg2, GoString arg3); 309 310 found in the _cgo_export.h generated header, after any preambles 311 copied from the cgo input files. Functions with multiple 312 return values are mapped to functions returning a struct. 313 314 Not all Go types can be mapped to C types in a useful way. 315 Go struct types are not supported; use a C struct type. 316 Go array types are not supported; use a C pointer. 317 318 Go functions that take arguments of type string may be called with the 319 C type _GoString_, described above. The _GoString_ type will be 320 automatically defined in the preamble. Note that there is no way for C 321 code to create a value of this type; this is only useful for passing 322 string values from Go to C and back to Go. 323 324 Using //export in a file places a restriction on the preamble: 325 since it is copied into two different C output files, it must not 326 contain any definitions, only declarations. If a file contains both 327 definitions and declarations, then the two output files will produce 328 duplicate symbols and the linker will fail. To avoid this, definitions 329 must be placed in preambles in other files, or in C source files. 330 331 # Passing pointers 332 333 Go is a garbage collected language, and the garbage collector needs to 334 know the location of every pointer to Go memory. Because of this, 335 there are restrictions on passing pointers between Go and C. 336 337 In this section the term Go pointer means a pointer to memory 338 allocated by Go (such as by using the & operator or calling the 339 predefined new function) and the term C pointer means a pointer to 340 memory allocated by C (such as by a call to C.malloc). Whether a 341 pointer is a Go pointer or a C pointer is a dynamic property 342 determined by how the memory was allocated; it has nothing to do with 343 the type of the pointer. 344 345 Note that values of some Go types, other than the type's zero value, 346 always include Go pointers. This is true of string, slice, interface, 347 channel, map, and function types. A pointer type may hold a Go pointer 348 or a C pointer. Array and struct types may or may not include Go 349 pointers, depending on the element types. All the discussion below 350 about Go pointers applies not just to pointer types, but also to other 351 types that include Go pointers. 352 353 All Go pointers passed to C must point to pinned Go memory. Go pointers 354 passed as function arguments to C functions have the memory they point to 355 implicitly pinned for the duration of the call. Go memory reachable from 356 these function arguments must be pinned as long as the C code has access 357 to it. Whether Go memory is pinned is a dynamic property of that memory 358 region; it has nothing to do with the type of the pointer. 359 360 Go values created by calling new, by taking the address of a composite 361 literal, or by taking the address of a local variable may also have their 362 memory pinned using [runtime.Pinner]. This type may be used to manage 363 the duration of the memory's pinned status, potentially beyond the 364 duration of a C function call. Memory may be pinned more than once and 365 must be unpinned exactly the same number of times it has been pinned. 366 367 Go code may pass a Go pointer to C provided the memory to which it 368 points does not contain any Go pointers to memory that is unpinned. When 369 passing a pointer to a field in a struct, the Go memory in question is 370 the memory occupied by the field, not the entire struct. When passing a 371 pointer to an element in an array or slice, the Go memory in question is 372 the entire array or the entire backing array of the slice. 373 374 C code may keep a copy of a Go pointer only as long as the memory it 375 points to is pinned. 376 377 C code may not keep a copy of a Go pointer after the call returns, 378 unless the memory it points to is pinned with [runtime.Pinner] and the 379 Pinner is not unpinned while the Go pointer is stored in C memory. 380 This implies that C code may not keep a copy of a string, slice, 381 channel, and so forth, because they cannot be pinned with 382 [runtime.Pinner]. 383 384 The _GoString_ type also may not be pinned with [runtime.Pinner]. 385 Because it includes a Go pointer, the memory it points to is only pinned 386 for the duration of the call; _GoString_ values may not be retained by C 387 code. 388 389 A Go function called by C code may return a Go pointer to pinned memory 390 (which implies that it may not return a string, slice, channel, and so 391 forth). A Go function called by C code may take C pointers as arguments, 392 and it may store non-pointer data, C pointers, or Go pointers to pinned 393 memory through those pointers. It may not store a Go pointer to unpinned 394 memory in memory pointed to by a C pointer (which again, implies that it 395 may not store a string, slice, channel, and so forth). A Go function 396 called by C code may take a Go pointer but it must preserve the property 397 that the Go memory to which it points (and the Go memory to which that 398 memory points, and so on) is pinned. 399 400 These rules are checked dynamically at runtime. The checking is 401 controlled by the cgocheck setting of the GODEBUG environment 402 variable. The default setting is GODEBUG=cgocheck=1, which implements 403 reasonably cheap dynamic checks. These checks may be disabled 404 entirely using GODEBUG=cgocheck=0. Complete checking of pointer 405 handling, at some cost in run time, is available by setting 406 GOEXPERIMENT=cgocheck2 at build time. 407 408 It is possible to defeat this enforcement by using the unsafe package, 409 and of course there is nothing stopping the C code from doing anything 410 it likes. However, programs that break these rules are likely to fail 411 in unexpected and unpredictable ways. 412 413 The runtime/cgo.Handle type can be used to safely pass Go values 414 between Go and C. See the runtime/cgo package documentation for details. 415 416 Note: the current implementation has a bug. While Go code is permitted 417 to write nil or a C pointer (but not a Go pointer) to C memory, the 418 current implementation may sometimes cause a runtime error if the 419 contents of the C memory appear to be a Go pointer. Therefore, avoid 420 passing uninitialized C memory to Go code if the Go code is going to 421 store pointer values in it. Zero out the memory in C before passing it 422 to Go. 423 424 # Special cases 425 426 A few special C types which would normally be represented by a pointer 427 type in Go are instead represented by a uintptr. Those include: 428 429 1. The *Ref types on Darwin, rooted at CoreFoundation's CFTypeRef type. 430 431 2. The object types from Java's JNI interface: 432 433 jobject 434 jclass 435 jthrowable 436 jstring 437 jarray 438 jbooleanArray 439 jbyteArray 440 jcharArray 441 jshortArray 442 jintArray 443 jlongArray 444 jfloatArray 445 jdoubleArray 446 jobjectArray 447 jweak 448 449 3. The EGLDisplay and EGLConfig types from the EGL API. 450 451 These types are uintptr on the Go side because they would otherwise 452 confuse the Go garbage collector; they are sometimes not really 453 pointers but data structures encoded in a pointer type. All operations 454 on these types must happen in C. The proper constant to initialize an 455 empty such reference is 0, not nil. 456 457 These special cases were introduced in Go 1.10. For auto-updating code 458 from Go 1.9 and earlier, use the cftype or jni rewrites in the Go fix tool: 459 460 go tool fix -r cftype <pkg> 461 go tool fix -r jni <pkg> 462 463 It will replace nil with 0 in the appropriate places. 464 465 The EGLDisplay case was introduced in Go 1.12. Use the egl rewrite 466 to auto-update code from Go 1.11 and earlier: 467 468 go tool fix -r egl <pkg> 469 470 The EGLConfig case was introduced in Go 1.15. Use the eglconf rewrite 471 to auto-update code from Go 1.14 and earlier: 472 473 go tool fix -r eglconf <pkg> 474 475 # Using cgo directly 476 477 Usage: 478 479 go tool cgo [cgo options] [-- compiler options] gofiles... 480 481 Cgo transforms the specified input Go source files into several output 482 Go and C source files. 483 484 The compiler options are passed through uninterpreted when 485 invoking the C compiler to compile the C parts of the package. 486 487 The following options are available when running cgo directly: 488 489 -V 490 Print cgo version and exit. 491 -debug-define 492 Debugging option. Print #defines. 493 -debug-gcc 494 Debugging option. Trace C compiler execution and output. 495 -dynimport file 496 Write list of symbols imported by file. Write to 497 -dynout argument or to standard output. Used by go 498 build when building a cgo package. 499 -dynlinker 500 Write dynamic linker as part of -dynimport output. 501 -dynout file 502 Write -dynimport output to file. 503 -dynpackage package 504 Set Go package for -dynimport output. 505 -exportheader file 506 If there are any exported functions, write the 507 generated export declarations to file. 508 C code can #include this to see the declarations. 509 -importpath string 510 The import path for the Go package. Optional; used for 511 nicer comments in the generated files. 512 -import_runtime_cgo 513 If set (which it is by default) import runtime/cgo in 514 generated output. 515 -import_syscall 516 If set (which it is by default) import syscall in 517 generated output. 518 -gccgo 519 Generate output for the gccgo compiler rather than the 520 gc compiler. 521 -gccgoprefix prefix 522 The -fgo-prefix option to be used with gccgo. 523 -gccgopkgpath path 524 The -fgo-pkgpath option to be used with gccgo. 525 -gccgo_define_cgoincomplete 526 Define cgo.Incomplete locally rather than importing it from 527 the "runtime/cgo" package. Used for old gccgo versions. 528 -godefs 529 Write out input file in Go syntax replacing C package 530 names with real values. Used to generate files in the 531 syscall package when bootstrapping a new target. 532 -objdir directory 533 Put all generated files in directory. 534 -srcdir directory 535 */ 536 package main 537 538 /* 539 Implementation details. 540 541 Cgo provides a way for Go programs to call C code linked into the same 542 address space. This comment explains the operation of cgo. 543 544 Cgo reads a set of Go source files and looks for statements saying 545 import "C". If the import has a doc comment, that comment is 546 taken as literal C code to be used as a preamble to any C code 547 generated by cgo. A typical preamble #includes necessary definitions: 548 549 // #include <stdio.h> 550 import "C" 551 552 For more details about the usage of cgo, see the documentation 553 comment at the top of this file. 554 555 Understanding C 556 557 Cgo scans the Go source files that import "C" for uses of that 558 package, such as C.puts. It collects all such identifiers. The next 559 step is to determine each kind of name. In C.xxx the xxx might refer 560 to a type, a function, a constant, or a global variable. Cgo must 561 decide which. 562 563 The obvious thing for cgo to do is to process the preamble, expanding 564 #includes and processing the corresponding C code. That would require 565 a full C parser and type checker that was also aware of any extensions 566 known to the system compiler (for example, all the GNU C extensions) as 567 well as the system-specific header locations and system-specific 568 pre-#defined macros. This is certainly possible to do, but it is an 569 enormous amount of work. 570 571 Cgo takes a different approach. It determines the meaning of C 572 identifiers not by parsing C code but by feeding carefully constructed 573 programs into the system C compiler and interpreting the generated 574 error messages, debug information, and object files. In practice, 575 parsing these is significantly less work and more robust than parsing 576 C source. 577 578 Cgo first invokes gcc -E -dM on the preamble, in order to find out 579 about simple #defines for constants and the like. These are recorded 580 for later use. 581 582 Next, cgo needs to identify the kinds for each identifier. For the 583 identifiers C.foo, cgo generates this C program: 584 585 <preamble> 586 #line 1 "not-declared" 587 void __cgo_f_1_1(void) { __typeof__(foo) *__cgo_undefined__1; } 588 #line 1 "not-type" 589 void __cgo_f_1_2(void) { foo *__cgo_undefined__2; } 590 #line 1 "not-int-const" 591 void __cgo_f_1_3(void) { enum { __cgo_undefined__3 = (foo)*1 }; } 592 #line 1 "not-num-const" 593 void __cgo_f_1_4(void) { static const double __cgo_undefined__4 = (foo); } 594 #line 1 "not-str-lit" 595 void __cgo_f_1_5(void) { static const char __cgo_undefined__5[] = (foo); } 596 597 This program will not compile, but cgo can use the presence or absence 598 of an error message on a given line to deduce the information it 599 needs. The program is syntactically valid regardless of whether each 600 name is a type or an ordinary identifier, so there will be no syntax 601 errors that might stop parsing early. 602 603 An error on not-declared:1 indicates that foo is undeclared. 604 An error on not-type:1 indicates that foo is not a type (if declared at all, it is an identifier). 605 An error on not-int-const:1 indicates that foo is not an integer constant. 606 An error on not-num-const:1 indicates that foo is not a number constant. 607 An error on not-str-lit:1 indicates that foo is not a string literal. 608 An error on not-signed-int-const:1 indicates that foo is not a signed integer constant. 609 610 The line number specifies the name involved. In the example, 1 is foo. 611 612 Next, cgo must learn the details of each type, variable, function, or 613 constant. It can do this by reading object files. If cgo has decided 614 that t1 is a type, v2 and v3 are variables or functions, and i4, i5 615 are integer constants, u6 is an unsigned integer constant, and f7 and f8 616 are float constants, and s9 and s10 are string constants, it generates: 617 618 <preamble> 619 __typeof__(t1) *__cgo__1; 620 __typeof__(v2) *__cgo__2; 621 __typeof__(v3) *__cgo__3; 622 __typeof__(i4) *__cgo__4; 623 enum { __cgo_enum__4 = i4 }; 624 __typeof__(i5) *__cgo__5; 625 enum { __cgo_enum__5 = i5 }; 626 __typeof__(u6) *__cgo__6; 627 enum { __cgo_enum__6 = u6 }; 628 __typeof__(f7) *__cgo__7; 629 __typeof__(f8) *__cgo__8; 630 __typeof__(s9) *__cgo__9; 631 __typeof__(s10) *__cgo__10; 632 633 long long __cgodebug_ints[] = { 634 0, // t1 635 0, // v2 636 0, // v3 637 i4, 638 i5, 639 u6, 640 0, // f7 641 0, // f8 642 0, // s9 643 0, // s10 644 1 645 }; 646 647 double __cgodebug_floats[] = { 648 0, // t1 649 0, // v2 650 0, // v3 651 0, // i4 652 0, // i5 653 0, // u6 654 f7, 655 f8, 656 0, // s9 657 0, // s10 658 1 659 }; 660 661 const char __cgodebug_str__9[] = s9; 662 const unsigned long long __cgodebug_strlen__9 = sizeof(s9)-1; 663 const char __cgodebug_str__10[] = s10; 664 const unsigned long long __cgodebug_strlen__10 = sizeof(s10)-1; 665 666 and again invokes the system C compiler, to produce an object file 667 containing debug information. Cgo parses the DWARF debug information 668 for __cgo__N to learn the type of each identifier. (The types also 669 distinguish functions from global variables.) Cgo reads the constant 670 values from the __cgodebug_* from the object file's data segment. 671 672 At this point cgo knows the meaning of each C.xxx well enough to start 673 the translation process. 674 675 Translating Go 676 677 Given the input Go files x.go and y.go, cgo generates these source 678 files: 679 680 x.cgo1.go # for gc (cmd/compile) 681 y.cgo1.go # for gc 682 _cgo_gotypes.go # for gc 683 _cgo_import.go # for gc (if -dynout _cgo_import.go) 684 x.cgo2.c # for gcc 685 y.cgo2.c # for gcc 686 _cgo_defun.c # for gcc (if -gccgo) 687 _cgo_export.c # for gcc 688 _cgo_export.h # for gcc 689 _cgo_main.c # for gcc 690 _cgo_flags # for build tool (if -gccgo) 691 692 The file x.cgo1.go is a copy of x.go with the import "C" removed and 693 references to C.xxx replaced with names like _Cfunc_xxx or _Ctype_xxx. 694 The definitions of those identifiers, written as Go functions, types, 695 or variables, are provided in _cgo_gotypes.go. 696 697 Here is a _cgo_gotypes.go containing definitions for needed C types: 698 699 type _Ctype_char int8 700 type _Ctype_int int32 701 type _Ctype_void [0]byte 702 703 The _cgo_gotypes.go file also contains the definitions of the 704 functions. They all have similar bodies that invoke runtime·cgocall 705 to make a switch from the Go runtime world to the system C (GCC-based) 706 world. 707 708 For example, here is the definition of _Cfunc_puts: 709 710 //go:cgo_import_static _cgo_be59f0f25121_Cfunc_puts 711 //go:linkname __cgofn__cgo_be59f0f25121_Cfunc_puts _cgo_be59f0f25121_Cfunc_puts 712 var __cgofn__cgo_be59f0f25121_Cfunc_puts byte 713 var _cgo_be59f0f25121_Cfunc_puts = unsafe.Pointer(&__cgofn__cgo_be59f0f25121_Cfunc_puts) 714 715 func _Cfunc_puts(p0 *_Ctype_char) (r1 _Ctype_int) { 716 _cgo_runtime_cgocall(_cgo_be59f0f25121_Cfunc_puts, uintptr(unsafe.Pointer(&p0))) 717 return 718 } 719 720 The hexadecimal number is a hash of cgo's input, chosen to be 721 deterministic yet unlikely to collide with other uses. The actual 722 function _cgo_be59f0f25121_Cfunc_puts is implemented in a C source 723 file compiled by gcc, the file x.cgo2.c: 724 725 void 726 _cgo_be59f0f25121_Cfunc_puts(void *v) 727 { 728 struct { 729 char* p0; 730 int r; 731 char __pad12[4]; 732 } __attribute__((__packed__, __gcc_struct__)) *a = v; 733 a->r = puts((void*)a->p0); 734 } 735 736 It extracts the arguments from the pointer to _Cfunc_puts's argument 737 frame, invokes the system C function (in this case, puts), stores the 738 result in the frame, and returns. 739 740 Linking 741 742 Once the _cgo_export.c and *.cgo2.c files have been compiled with gcc, 743 they need to be linked into the final binary, along with the libraries 744 they might depend on (in the case of puts, stdio). cmd/link has been 745 extended to understand basic ELF files, but it does not understand ELF 746 in the full complexity that modern C libraries embrace, so it cannot 747 in general generate direct references to the system libraries. 748 749 Instead, the build process generates an object file using dynamic 750 linkage to the desired libraries. The main function is provided by 751 _cgo_main.c: 752 753 int main() { return 0; } 754 void crosscall2(void(*fn)(void*), void *a, int c, uintptr_t ctxt) { } 755 uintptr_t _cgo_wait_runtime_init_done(void) { return 0; } 756 void _cgo_release_context(uintptr_t ctxt) { } 757 char* _cgo_topofstack(void) { return (char*)0; } 758 void _cgo_allocate(void *a, int c) { } 759 void _cgo_panic(void *a, int c) { } 760 void _cgo_reginit(void) { } 761 762 The extra functions here are stubs to satisfy the references in the C 763 code generated for gcc. The build process links this stub, along with 764 _cgo_export.c and *.cgo2.c, into a dynamic executable and then lets 765 cgo examine the executable. Cgo records the list of shared library 766 references and resolved names and writes them into a new file 767 _cgo_import.go, which looks like: 768 769 //go:cgo_dynamic_linker "/lib64/ld-linux-x86-64.so.2" 770 //go:cgo_import_dynamic puts puts#GLIBC_2.2.5 "libc.so.6" 771 //go:cgo_import_dynamic __libc_start_main __libc_start_main#GLIBC_2.2.5 "libc.so.6" 772 //go:cgo_import_dynamic stdout stdout#GLIBC_2.2.5 "libc.so.6" 773 //go:cgo_import_dynamic fflush fflush#GLIBC_2.2.5 "libc.so.6" 774 //go:cgo_import_dynamic _ _ "libpthread.so.0" 775 //go:cgo_import_dynamic _ _ "libc.so.6" 776 777 In the end, the compiled Go package, which will eventually be 778 presented to cmd/link as part of a larger program, contains: 779 780 _go_.o # gc-compiled object for _cgo_gotypes.go, _cgo_import.go, *.cgo1.go 781 _all.o # gcc-compiled object for _cgo_export.c, *.cgo2.c 782 783 If there is an error generating the _cgo_import.go file, then, instead 784 of adding _cgo_import.go to the package, the go tool adds an empty 785 file named dynimportfail. The _cgo_import.go file is only needed when 786 using internal linking mode, which is not the default when linking 787 programs that use cgo (as described below). If the linker sees a file 788 named dynimportfail it reports an error if it has been told to use 789 internal linking mode. This approach is taken because generating 790 _cgo_import.go requires doing a full C link of the package, which can 791 fail for reasons that are irrelevant when using external linking mode. 792 793 The final program will be a dynamic executable, so that cmd/link can avoid 794 needing to process arbitrary .o files. It only needs to process the .o 795 files generated from C files that cgo writes, and those are much more 796 limited in the ELF or other features that they use. 797 798 In essence, the _cgo_import.o file includes the extra linking 799 directives that cmd/link is not sophisticated enough to derive from _all.o 800 on its own. Similarly, the _all.o uses dynamic references to real 801 system object code because cmd/link is not sophisticated enough to process 802 the real code. 803 804 The main benefits of this system are that cmd/link remains relatively simple 805 (it does not need to implement a complete ELF and Mach-O linker) and 806 that gcc is not needed after the package is compiled. For example, 807 package net uses cgo for access to name resolution functions provided 808 by libc. Although gcc is needed to compile package net, gcc is not 809 needed to link programs that import package net. 810 811 Runtime 812 813 When using cgo, Go must not assume that it owns all details of the 814 process. In particular it needs to coordinate with C in the use of 815 threads and thread-local storage. The runtime package declares a few 816 variables: 817 818 var ( 819 iscgo bool 820 _cgo_init unsafe.Pointer 821 _cgo_thread_start unsafe.Pointer 822 ) 823 824 Any package using cgo imports "runtime/cgo", which provides 825 initializations for these variables. It sets iscgo to true, _cgo_init 826 to a gcc-compiled function that can be called early during program 827 startup, and _cgo_thread_start to a gcc-compiled function that can be 828 used to create a new thread, in place of the runtime's usual direct 829 system calls. 830 831 Internal and External Linking 832 833 The text above describes "internal" linking, in which cmd/link parses and 834 links host object files (ELF, Mach-O, PE, and so on) into the final 835 executable itself. Keeping cmd/link simple means we cannot possibly 836 implement the full semantics of the host linker, so the kinds of 837 objects that can be linked directly into the binary is limited (other 838 code can only be used as a dynamic library). On the other hand, when 839 using internal linking, cmd/link can generate Go binaries by itself. 840 841 In order to allow linking arbitrary object files without requiring 842 dynamic libraries, cgo supports an "external" linking mode too. In 843 external linking mode, cmd/link does not process any host object files. 844 Instead, it collects all the Go code and writes a single go.o object 845 file containing it. Then it invokes the host linker (usually gcc) to 846 combine the go.o object file and any supporting non-Go code into a 847 final executable. External linking avoids the dynamic library 848 requirement but introduces a requirement that the host linker be 849 present to create such a binary. 850 851 Most builds both compile source code and invoke the linker to create a 852 binary. When cgo is involved, the compile step already requires gcc, so 853 it is not problematic for the link step to require gcc too. 854 855 An important exception is builds using a pre-compiled copy of the 856 standard library. In particular, package net uses cgo on most systems, 857 and we want to preserve the ability to compile pure Go code that 858 imports net without requiring gcc to be present at link time. (In this 859 case, the dynamic library requirement is less significant, because the 860 only library involved is libc.so, which can usually be assumed 861 present.) 862 863 This conflict between functionality and the gcc requirement means we 864 must support both internal and external linking, depending on the 865 circumstances: if net is the only cgo-using package, then internal 866 linking is probably fine, but if other packages are involved, so that there 867 are dependencies on libraries beyond libc, external linking is likely 868 to work better. The compilation of a package records the relevant 869 information to support both linking modes, leaving the decision 870 to be made when linking the final binary. 871 872 Linking Directives 873 874 In either linking mode, package-specific directives must be passed 875 through to cmd/link. These are communicated by writing //go: directives in a 876 Go source file compiled by gc. The directives are copied into the .o 877 object file and then processed by the linker. 878 879 The directives are: 880 881 //go:cgo_import_dynamic <local> [<remote> ["<library>"]] 882 883 In internal linking mode, allow an unresolved reference to 884 <local>, assuming it will be resolved by a dynamic library 885 symbol. The optional <remote> specifies the symbol's name and 886 possibly version in the dynamic library, and the optional "<library>" 887 names the specific library where the symbol should be found. 888 889 On AIX, the library pattern is slightly different. It must be 890 "lib.a/obj.o" with obj.o the member of this library exporting 891 this symbol. 892 893 In the <remote>, # or @ can be used to introduce a symbol version. 894 895 Examples: 896 //go:cgo_import_dynamic puts 897 //go:cgo_import_dynamic puts puts#GLIBC_2.2.5 898 //go:cgo_import_dynamic puts puts#GLIBC_2.2.5 "libc.so.6" 899 900 A side effect of the cgo_import_dynamic directive with a 901 library is to make the final binary depend on that dynamic 902 library. To get the dependency without importing any specific 903 symbols, use _ for local and remote. 904 905 Example: 906 //go:cgo_import_dynamic _ _ "libc.so.6" 907 908 For compatibility with current versions of SWIG, 909 #pragma dynimport is an alias for //go:cgo_import_dynamic. 910 911 //go:cgo_dynamic_linker "<path>" 912 913 In internal linking mode, use "<path>" as the dynamic linker 914 in the final binary. This directive is only needed from one 915 package when constructing a binary; by convention it is 916 supplied by runtime/cgo. 917 918 Example: 919 //go:cgo_dynamic_linker "/lib/ld-linux.so.2" 920 921 //go:cgo_export_dynamic <local> <remote> 922 923 In internal linking mode, put the Go symbol 924 named <local> into the program's exported symbol table as 925 <remote>, so that C code can refer to it by that name. This 926 mechanism makes it possible for C code to call back into Go or 927 to share Go's data. 928 929 For compatibility with current versions of SWIG, 930 #pragma dynexport is an alias for //go:cgo_export_dynamic. 931 932 //go:cgo_import_static <local> 933 934 In external linking mode, allow unresolved references to 935 <local> in the go.o object file prepared for the host linker, 936 under the assumption that <local> will be supplied by the 937 other object files that will be linked with go.o. 938 939 Example: 940 //go:cgo_import_static puts_wrapper 941 942 //go:cgo_export_static <local> <remote> 943 944 In external linking mode, put the Go symbol 945 named <local> into the program's exported symbol table as 946 <remote>, so that C code can refer to it by that name. This 947 mechanism makes it possible for C code to call back into Go or 948 to share Go's data. 949 950 //go:cgo_ldflag "<arg>" 951 952 In external linking mode, invoke the host linker (usually gcc) 953 with "<arg>" as a command-line argument following the .o files. 954 Note that the arguments are for "gcc", not "ld". 955 956 Example: 957 //go:cgo_ldflag "-lpthread" 958 //go:cgo_ldflag "-L/usr/local/sqlite3/lib" 959 960 A package compiled with cgo will include directives for both 961 internal and external linking; the linker will select the appropriate 962 subset for the chosen linking mode. 963 964 Example 965 966 As a simple example, consider a package that uses cgo to call C.sin. 967 The following code will be generated by cgo: 968 969 // compiled by gc 970 971 //go:cgo_ldflag "-lm" 972 973 type _Ctype_double float64 974 975 //go:cgo_import_static _cgo_gcc_Cfunc_sin 976 //go:linkname __cgo_gcc_Cfunc_sin _cgo_gcc_Cfunc_sin 977 var __cgo_gcc_Cfunc_sin byte 978 var _cgo_gcc_Cfunc_sin = unsafe.Pointer(&__cgo_gcc_Cfunc_sin) 979 980 func _Cfunc_sin(p0 _Ctype_double) (r1 _Ctype_double) { 981 _cgo_runtime_cgocall(_cgo_gcc_Cfunc_sin, uintptr(unsafe.Pointer(&p0))) 982 return 983 } 984 985 // compiled by gcc, into foo.cgo2.o 986 987 void 988 _cgo_gcc_Cfunc_sin(void *v) 989 { 990 struct { 991 double p0; 992 double r; 993 } __attribute__((__packed__)) *a = v; 994 a->r = sin(a->p0); 995 } 996 997 What happens at link time depends on whether the final binary is linked 998 using the internal or external mode. If other packages are compiled in 999 "external only" mode, then the final link will be an external one. 1000 Otherwise the link will be an internal one. 1001 1002 The linking directives are used according to the kind of final link 1003 used. 1004 1005 In internal mode, cmd/link itself processes all the host object files, in 1006 particular foo.cgo2.o. To do so, it uses the cgo_import_dynamic and 1007 cgo_dynamic_linker directives to learn that the otherwise undefined 1008 reference to sin in foo.cgo2.o should be rewritten to refer to the 1009 symbol sin with version GLIBC_2.2.5 from the dynamic library 1010 "libm.so.6", and the binary should request "/lib/ld-linux.so.2" as its 1011 runtime dynamic linker. 1012 1013 In external mode, cmd/link does not process any host object files, in 1014 particular foo.cgo2.o. It links together the gc-generated object 1015 files, along with any other Go code, into a go.o file. While doing 1016 that, cmd/link will discover that there is no definition for 1017 _cgo_gcc_Cfunc_sin, referred to by the gc-compiled source file. This 1018 is okay, because cmd/link also processes the cgo_import_static directive and 1019 knows that _cgo_gcc_Cfunc_sin is expected to be supplied by a host 1020 object file, so cmd/link does not treat the missing symbol as an error when 1021 creating go.o. Indeed, the definition for _cgo_gcc_Cfunc_sin will be 1022 provided to the host linker by foo2.cgo.o, which in turn will need the 1023 symbol 'sin'. cmd/link also processes the cgo_ldflag directives, so that it 1024 knows that the eventual host link command must include the -lm 1025 argument, so that the host linker will be able to find 'sin' in the 1026 math library. 1027 1028 cmd/link Command Line Interface 1029 1030 The go command and any other Go-aware build systems invoke cmd/link 1031 to link a collection of packages into a single binary. By default, cmd/link will 1032 present the same interface it does today: 1033 1034 cmd/link main.a 1035 1036 produces a file named a.out, even if cmd/link does so by invoking the host 1037 linker in external linking mode. 1038 1039 By default, cmd/link will decide the linking mode as follows: if the only 1040 packages using cgo are those on a list of known standard library 1041 packages (net, os/user, runtime/cgo), cmd/link will use internal linking 1042 mode. Otherwise, there are non-standard cgo packages involved, and cmd/link 1043 will use external linking mode. The first rule means that a build of 1044 the godoc binary, which uses net but no other cgo, can run without 1045 needing gcc available. The second rule means that a build of a 1046 cgo-wrapped library like sqlite3 can generate a standalone executable 1047 instead of needing to refer to a dynamic library. The specific choice 1048 can be overridden using a command line flag: cmd/link -linkmode=internal or 1049 cmd/link -linkmode=external. 1050 1051 In an external link, cmd/link will create a temporary directory, write any 1052 host object files found in package archives to that directory (renamed 1053 to avoid conflicts), write the go.o file to that directory, and invoke 1054 the host linker. The default value for the host linker is $CC, split 1055 into fields, or else "gcc". The specific host linker command line can 1056 be overridden using command line flags: cmd/link -extld=clang 1057 -extldflags='-ggdb -O3'. If any package in a build includes a .cc or 1058 other file compiled by the C++ compiler, the go tool will use the 1059 -extld option to set the host linker to the C++ compiler. 1060 1061 These defaults mean that Go-aware build systems can ignore the linking 1062 changes and keep running plain 'cmd/link' and get reasonable results, but 1063 they can also control the linking details if desired. 1064 1065 */ 1066