Source file src/runtime/extern.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  Package runtime contains operations that interact with Go's runtime system,
     7  such as functions to control goroutines. It also includes the low-level type information
     8  used by the reflect package; see reflect's documentation for the programmable
     9  interface to the run-time type system.
    10  
    11  # Environment Variables
    12  
    13  The following environment variables ($name or %name%, depending on the host
    14  operating system) control the run-time behavior of Go programs. The meanings
    15  and use may change from release to release.
    16  
    17  The GOGC variable sets the initial garbage collection target percentage.
    18  A collection is triggered when the ratio of freshly allocated data to live data
    19  remaining after the previous collection reaches this percentage. The default
    20  is GOGC=100. Setting GOGC=off disables the garbage collector entirely.
    21  [runtime/debug.SetGCPercent] allows changing this percentage at run time.
    22  
    23  The GOMEMLIMIT variable sets a soft memory limit for the runtime. This memory limit
    24  includes the Go heap and all other memory managed by the runtime, and excludes
    25  external memory sources such as mappings of the binary itself, memory managed in
    26  other languages, and memory held by the operating system on behalf of the Go
    27  program. GOMEMLIMIT is a numeric value in bytes with an optional unit suffix.
    28  The supported suffixes include B, KiB, MiB, GiB, and TiB. These suffixes
    29  represent quantities of bytes as defined by the IEC 80000-13 standard. That is,
    30  they are based on powers of two: KiB means 2^10 bytes, MiB means 2^20 bytes,
    31  and so on. The default setting is math.MaxInt64, which effectively disables the
    32  memory limit. [runtime/debug.SetMemoryLimit] allows changing this limit at run
    33  time.
    34  
    35  The GODEBUG variable controls debugging variables within the runtime.
    36  It is a comma-separated list of name=val pairs setting these named variables:
    37  
    38  	allocfreetrace: setting allocfreetrace=1 causes every allocation to be
    39  	profiled and a stack trace printed on each object's allocation and free.
    40  
    41  	clobberfree: setting clobberfree=1 causes the garbage collector to
    42  	clobber the memory content of an object with bad content when it frees
    43  	the object.
    44  
    45  	cpu.*: cpu.all=off disables the use of all optional instruction set extensions.
    46  	cpu.extension=off disables use of instructions from the specified instruction set extension.
    47  	extension is the lower case name for the instruction set extension such as sse41 or avx
    48  	as listed in internal/cpu package. As an example cpu.avx=off disables runtime detection
    49  	and thereby use of AVX instructions.
    50  
    51  	cgocheck: setting cgocheck=0 disables all checks for packages
    52  	using cgo to incorrectly pass Go pointers to non-Go code.
    53  	Setting cgocheck=1 (the default) enables relatively cheap
    54  	checks that may miss some errors. A more complete, but slow,
    55  	cgocheck mode can be enabled using GOEXPERIMENT (which
    56  	requires a rebuild), see https://pkg.go.dev/internal/goexperiment for details.
    57  
    58  	dontfreezetheworld: by default, the start of a fatal panic or throw
    59  	"freezes the world", preempting all threads to stop all running
    60  	goroutines, which makes it possible to traceback all goroutines, and
    61  	keeps their state close to the point of panic. Setting
    62  	dontfreezetheworld=1 disables this preemption, allowing goroutines to
    63  	continue executing during panic processing. Note that goroutines that
    64  	naturally enter the scheduler will still stop. This can be useful when
    65  	debugging the runtime scheduler, as freezetheworld perturbs scheduler
    66  	state and thus may hide problems.
    67  
    68  	efence: setting efence=1 causes the allocator to run in a mode
    69  	where each object is allocated on a unique page and addresses are
    70  	never recycled.
    71  
    72  	gccheckmark: setting gccheckmark=1 enables verification of the
    73  	garbage collector's concurrent mark phase by performing a
    74  	second mark pass while the world is stopped.  If the second
    75  	pass finds a reachable object that was not found by concurrent
    76  	mark, the garbage collector will panic.
    77  
    78  	gcpacertrace: setting gcpacertrace=1 causes the garbage collector to
    79  	print information about the internal state of the concurrent pacer.
    80  
    81  	gcshrinkstackoff: setting gcshrinkstackoff=1 disables moving goroutines
    82  	onto smaller stacks. In this mode, a goroutine's stack can only grow.
    83  
    84  	gcstoptheworld: setting gcstoptheworld=1 disables concurrent garbage collection,
    85  	making every garbage collection a stop-the-world event. Setting gcstoptheworld=2
    86  	also disables concurrent sweeping after the garbage collection finishes.
    87  
    88  	gctrace: setting gctrace=1 causes the garbage collector to emit a single line to standard
    89  	error at each collection, summarizing the amount of memory collected and the
    90  	length of the pause. The format of this line is subject to change. Included in
    91  	the explanation below is also the relevant runtime/metrics metric for each field.
    92  	Currently, it is:
    93  		gc # @#s #%: #+#+# ms clock, #+#/#/#+# ms cpu, #->#-># MB, # MB goal, # MB stacks, #MB globals, # P
    94  	where the fields are as follows:
    95  		gc #         the GC number, incremented at each GC
    96  		@#s          time in seconds since program start
    97  		#%           percentage of time spent in GC since program start
    98  		#+...+#      wall-clock/CPU times for the phases of the GC
    99  		#->#-># MB   heap size at GC start, at GC end, and live heap, or /gc/scan/heap:bytes
   100  		# MB goal    goal heap size, or /gc/heap/goal:bytes
   101  		# MB stacks  estimated scannable stack size, or /gc/scan/stack:bytes
   102  		# MB globals scannable global size, or /gc/scan/globals:bytes
   103  		# P          number of processors used, or /sched/gomaxprocs:threads
   104  	The phases are stop-the-world (STW) sweep termination, concurrent
   105  	mark and scan, and STW mark termination. The CPU times
   106  	for mark/scan are broken down in to assist time (GC performed in
   107  	line with allocation), background GC time, and idle GC time.
   108  	If the line ends with "(forced)", this GC was forced by a
   109  	runtime.GC() call.
   110  
   111  	harddecommit: setting harddecommit=1 causes memory that is returned to the OS to
   112  	also have protections removed on it. This is the only mode of operation on Windows,
   113  	but is helpful in debugging scavenger-related issues on other platforms. Currently,
   114  	only supported on Linux.
   115  
   116  	inittrace: setting inittrace=1 causes the runtime to emit a single line to standard
   117  	error for each package with init work, summarizing the execution time and memory
   118  	allocation. No information is printed for inits executed as part of plugin loading
   119  	and for packages without both user defined and compiler generated init work.
   120  	The format of this line is subject to change. Currently, it is:
   121  		init # @#ms, # ms clock, # bytes, # allocs
   122  	where the fields are as follows:
   123  		init #      the package name
   124  		@# ms       time in milliseconds when the init started since program start
   125  		# clock     wall-clock time for package initialization work
   126  		# bytes     memory allocated on the heap
   127  		# allocs    number of heap allocations
   128  
   129  	madvdontneed: setting madvdontneed=0 will use MADV_FREE
   130  	instead of MADV_DONTNEED on Linux when returning memory to the
   131  	kernel. This is more efficient, but means RSS numbers will
   132  	drop only when the OS is under memory pressure. On the BSDs and
   133  	Illumos/Solaris, setting madvdontneed=1 will use MADV_DONTNEED instead
   134  	of MADV_FREE. This is less efficient, but causes RSS numbers to drop
   135  	more quickly.
   136  
   137  	memprofilerate: setting memprofilerate=X will update the value of runtime.MemProfileRate.
   138  	When set to 0 memory profiling is disabled.  Refer to the description of
   139  	MemProfileRate for the default value.
   140  
   141  	pagetrace: setting pagetrace=/path/to/file will write out a trace of page events
   142  	that can be viewed, analyzed, and visualized using the x/debug/cmd/pagetrace tool.
   143  	Build your program with GOEXPERIMENT=pagetrace to enable this functionality. Do not
   144  	enable this functionality if your program is a setuid binary as it introduces a security
   145  	risk in that scenario. Currently not supported on Windows, plan9 or js/wasm. Setting this
   146  	option for some applications can produce large traces, so use with care.
   147  
   148  	invalidptr: invalidptr=1 (the default) causes the garbage collector and stack
   149  	copier to crash the program if an invalid pointer value (for example, 1)
   150  	is found in a pointer-typed location. Setting invalidptr=0 disables this check.
   151  	This should only be used as a temporary workaround to diagnose buggy code.
   152  	The real fix is to not store integers in pointer-typed locations.
   153  
   154  	sbrk: setting sbrk=1 replaces the memory allocator and garbage collector
   155  	with a trivial allocator that obtains memory from the operating system and
   156  	never reclaims any memory.
   157  
   158  	scavtrace: setting scavtrace=1 causes the runtime to emit a single line to standard
   159  	error, roughly once per GC cycle, summarizing the amount of work done by the
   160  	scavenger as well as the total amount of memory returned to the operating system
   161  	and an estimate of physical memory utilization. The format of this line is subject
   162  	to change, but currently it is:
   163  		scav # KiB work (bg), # KiB work (eager), # KiB total, #% util
   164  	where the fields are as follows:
   165  		# KiB work (bg)    the amount of memory returned to the OS in the background since
   166  		                   the last line
   167  		# KiB work (eager) the amount of memory returned to the OS eagerly since the last line
   168  		# KiB now          the amount of address space currently returned to the OS
   169  		#% util            the fraction of all unscavenged heap memory which is in-use
   170  	If the line ends with "(forced)", then scavenging was forced by a
   171  	debug.FreeOSMemory() call.
   172  
   173  	scheddetail: setting schedtrace=X and scheddetail=1 causes the scheduler to emit
   174  	detailed multiline info every X milliseconds, describing state of the scheduler,
   175  	processors, threads and goroutines.
   176  
   177  	schedtrace: setting schedtrace=X causes the scheduler to emit a single line to standard
   178  	error every X milliseconds, summarizing the scheduler state.
   179  
   180  	tracebackancestors: setting tracebackancestors=N extends tracebacks with the stacks at
   181  	which goroutines were created, where N limits the number of ancestor goroutines to
   182  	report. This also extends the information returned by runtime.Stack. Ancestor's goroutine
   183  	IDs will refer to the ID of the goroutine at the time of creation; it's possible for this
   184  	ID to be reused for another goroutine. Setting N to 0 will report no ancestry information.
   185  
   186  	tracefpunwindoff: setting tracefpunwindoff=1 forces the execution tracer to
   187  	use the runtime's default stack unwinder instead of frame pointer unwinding.
   188  	This increases tracer overhead, but could be helpful as a workaround or for
   189  	debugging unexpected regressions caused by frame pointer unwinding.
   190  
   191  	asyncpreemptoff: asyncpreemptoff=1 disables signal-based
   192  	asynchronous goroutine preemption. This makes some loops
   193  	non-preemptible for long periods, which may delay GC and
   194  	goroutine scheduling. This is useful for debugging GC issues
   195  	because it also disables the conservative stack scanning used
   196  	for asynchronously preempted goroutines.
   197  
   198  The net and net/http packages also refer to debugging variables in GODEBUG.
   199  See the documentation for those packages for details.
   200  
   201  The GOMAXPROCS variable limits the number of operating system threads that
   202  can execute user-level Go code simultaneously. There is no limit to the number of threads
   203  that can be blocked in system calls on behalf of Go code; those do not count against
   204  the GOMAXPROCS limit. This package's GOMAXPROCS function queries and changes
   205  the limit.
   206  
   207  The GORACE variable configures the race detector, for programs built using -race.
   208  See https://golang.org/doc/articles/race_detector.html for details.
   209  
   210  The GOTRACEBACK variable controls the amount of output generated when a Go
   211  program fails due to an unrecovered panic or an unexpected runtime condition.
   212  By default, a failure prints a stack trace for the current goroutine,
   213  eliding functions internal to the run-time system, and then exits with exit code 2.
   214  The failure prints stack traces for all goroutines if there is no current goroutine
   215  or the failure is internal to the run-time.
   216  GOTRACEBACK=none omits the goroutine stack traces entirely.
   217  GOTRACEBACK=single (the default) behaves as described above.
   218  GOTRACEBACK=all adds stack traces for all user-created goroutines.
   219  GOTRACEBACK=system is like “all” but adds stack frames for run-time functions
   220  and shows goroutines created internally by the run-time.
   221  GOTRACEBACK=crash is like “system” but crashes in an operating system-specific
   222  manner instead of exiting. For example, on Unix systems, the crash raises
   223  SIGABRT to trigger a core dump.
   224  GOTRACEBACK=wer is like “crash” but doesn't disable Windows Error Reporting (WER).
   225  For historical reasons, the GOTRACEBACK settings 0, 1, and 2 are synonyms for
   226  none, all, and system, respectively.
   227  The runtime/debug package's SetTraceback function allows increasing the
   228  amount of output at run time, but it cannot reduce the amount below that
   229  specified by the environment variable.
   230  See https://golang.org/pkg/runtime/debug/#SetTraceback.
   231  
   232  The GOARCH, GOOS, GOPATH, and GOROOT environment variables complete
   233  the set of Go environment variables. They influence the building of Go programs
   234  (see https://golang.org/cmd/go and https://golang.org/pkg/go/build).
   235  GOARCH, GOOS, and GOROOT are recorded at compile time and made available by
   236  constants or functions in this package, but they do not influence the execution
   237  of the run-time system.
   238  
   239  # Security
   240  
   241  On Unix platforms, Go's runtime system behaves slightly differently when a
   242  binary is setuid/setgid or executed with setuid/setgid-like properties, in order
   243  to prevent dangerous behaviors. On Linux this is determined by checking for the
   244  AT_SECURE flag in the auxiliary vector, on the BSDs and Solaris/Illumos it is
   245  determined by checking the issetugid syscall, and on AIX it is determined by
   246  checking if the uid/gid match the effective uid/gid.
   247  
   248  When the runtime determines the binary is setuid/setgid-like, it does three main
   249  things:
   250    - The standard input/output file descriptors (0, 1, 2) are checked to be open.
   251      If any of them are closed, they are opened pointing at /dev/null.
   252    - The value of the GOTRACEBACK environment variable is set to 'none'.
   253    - When a signal is received that terminates the program, or the program
   254      encounters an unrecoverable panic that would otherwise override the value
   255      of GOTRACEBACK, the goroutine stack, registers, and other memory related
   256      information are omitted.
   257  */
   258  package runtime
   259  
   260  import (
   261  	"internal/goarch"
   262  	"internal/goos"
   263  )
   264  
   265  // Caller reports file and line number information about function invocations on
   266  // the calling goroutine's stack. The argument skip is the number of stack frames
   267  // to ascend, with 0 identifying the caller of Caller.  (For historical reasons the
   268  // meaning of skip differs between Caller and Callers.) The return values report the
   269  // program counter, file name, and line number within the file of the corresponding
   270  // call. The boolean ok is false if it was not possible to recover the information.
   271  func Caller(skip int) (pc uintptr, file string, line int, ok bool) {
   272  	rpc := make([]uintptr, 1)
   273  	n := callers(skip+1, rpc[:])
   274  	if n < 1 {
   275  		return
   276  	}
   277  	frame, _ := CallersFrames(rpc).Next()
   278  	return frame.PC, frame.File, frame.Line, frame.PC != 0
   279  }
   280  
   281  // Callers fills the slice pc with the return program counters of function invocations
   282  // on the calling goroutine's stack. The argument skip is the number of stack frames
   283  // to skip before recording in pc, with 0 identifying the frame for Callers itself and
   284  // 1 identifying the caller of Callers.
   285  // It returns the number of entries written to pc.
   286  //
   287  // To translate these PCs into symbolic information such as function
   288  // names and line numbers, use CallersFrames. CallersFrames accounts
   289  // for inlined functions and adjusts the return program counters into
   290  // call program counters. Iterating over the returned slice of PCs
   291  // directly is discouraged, as is using FuncForPC on any of the
   292  // returned PCs, since these cannot account for inlining or return
   293  // program counter adjustment.
   294  func Callers(skip int, pc []uintptr) int {
   295  	// runtime.callers uses pc.array==nil as a signal
   296  	// to print a stack trace. Pick off 0-length pc here
   297  	// so that we don't let a nil pc slice get to it.
   298  	if len(pc) == 0 {
   299  		return 0
   300  	}
   301  	return callers(skip, pc)
   302  }
   303  
   304  var defaultGOROOT string // set by cmd/link
   305  
   306  // GOROOT returns the root of the Go tree. It uses the
   307  // GOROOT environment variable, if set at process start,
   308  // or else the root used during the Go build.
   309  func GOROOT() string {
   310  	s := gogetenv("GOROOT")
   311  	if s != "" {
   312  		return s
   313  	}
   314  	return defaultGOROOT
   315  }
   316  
   317  // buildVersion is the Go tree's version string at build time.
   318  //
   319  // If any GOEXPERIMENTs are set to non-default values, it will include
   320  // "X:<GOEXPERIMENT>".
   321  //
   322  // This is set by the linker.
   323  //
   324  // This is accessed by "go version <binary>".
   325  var buildVersion string
   326  
   327  // Version returns the Go tree's version string.
   328  // It is either the commit hash and date at the time of the build or,
   329  // when possible, a release tag like "go1.3".
   330  func Version() string {
   331  	return buildVersion
   332  }
   333  
   334  // GOOS is the running program's operating system target:
   335  // one of darwin, freebsd, linux, and so on.
   336  // To view possible combinations of GOOS and GOARCH, run "go tool dist list".
   337  const GOOS string = goos.GOOS
   338  
   339  // GOARCH is the running program's architecture target:
   340  // one of 386, amd64, arm, s390x, and so on.
   341  const GOARCH string = goarch.GOARCH
   342  

View as plain text