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 The runtime/debug package's SetGCPercent function allows changing this 22 percentage at run time. See https://golang.org/pkg/runtime/debug/#SetGCPercent. 23 24 The GODEBUG variable controls debugging variables within the runtime. 25 It is a comma-separated list of name=val pairs setting these named variables: 26 27 allocfreetrace: setting allocfreetrace=1 causes every allocation to be 28 profiled and a stack trace printed on each object's allocation and free. 29 30 clobberfree: setting clobberfree=1 causes the garbage collector to 31 clobber the memory content of an object with bad content when it frees 32 the object. 33 34 cgocheck: setting cgocheck=0 disables all checks for packages 35 using cgo to incorrectly pass Go pointers to non-Go code. 36 Setting cgocheck=1 (the default) enables relatively cheap 37 checks that may miss some errors. Setting cgocheck=2 enables 38 expensive checks that should not miss any errors, but will 39 cause your program to run slower. 40 41 efence: setting efence=1 causes the allocator to run in a mode 42 where each object is allocated on a unique page and addresses are 43 never recycled. 44 45 gccheckmark: setting gccheckmark=1 enables verification of the 46 garbage collector's concurrent mark phase by performing a 47 second mark pass while the world is stopped. If the second 48 pass finds a reachable object that was not found by concurrent 49 mark, the garbage collector will panic. 50 51 gcpacertrace: setting gcpacertrace=1 causes the garbage collector to 52 print information about the internal state of the concurrent pacer. 53 54 gcshrinkstackoff: setting gcshrinkstackoff=1 disables moving goroutines 55 onto smaller stacks. In this mode, a goroutine's stack can only grow. 56 57 gcstoptheworld: setting gcstoptheworld=1 disables concurrent garbage collection, 58 making every garbage collection a stop-the-world event. Setting gcstoptheworld=2 59 also disables concurrent sweeping after the garbage collection finishes. 60 61 gctrace: setting gctrace=1 causes the garbage collector to emit a single line to standard 62 error at each collection, summarizing the amount of memory collected and the 63 length of the pause. The format of this line is subject to change. 64 Currently, it is: 65 gc # @#s #%: #+#+# ms clock, #+#/#/#+# ms cpu, #->#-># MB, # MB goal, # P 66 where the fields are as follows: 67 gc # the GC number, incremented at each GC 68 @#s time in seconds since program start 69 #% percentage of time spent in GC since program start 70 #+...+# wall-clock/CPU times for the phases of the GC 71 #->#-># MB heap size at GC start, at GC end, and live heap 72 # MB goal goal heap size 73 # P number of processors used 74 The phases are stop-the-world (STW) sweep termination, concurrent 75 mark and scan, and STW mark termination. The CPU times 76 for mark/scan are broken down in to assist time (GC performed in 77 line with allocation), background GC time, and idle GC time. 78 If the line ends with "(forced)", this GC was forced by a 79 runtime.GC() call. 80 81 Setting gctrace to any value > 0 also causes the garbage collector 82 to emit a summary when memory is released back to the system. 83 This process of returning memory to the system is called scavenging. 84 The format of this summary is subject to change. 85 Currently it is: 86 scvg#: # MB released printed only if non-zero 87 scvg#: inuse: # idle: # sys: # released: # consumed: # (MB) 88 where the fields are as follows: 89 scvg# the scavenge cycle number, incremented at each scavenge 90 inuse: # MB used or partially used spans 91 idle: # MB spans pending scavenging 92 sys: # MB mapped from the system 93 released: # MB released to the system 94 consumed: # MB allocated from the system 95 96 madvdontneed: setting madvdontneed=1 will use MADV_DONTNEED 97 instead of MADV_FREE on Linux when returning memory to the 98 kernel. This is less efficient, but causes RSS numbers to drop 99 more quickly. 100 101 memprofilerate: setting memprofilerate=X will update the value of runtime.MemProfileRate. 102 When set to 0 memory profiling is disabled. Refer to the description of 103 MemProfileRate for the default value. 104 105 invalidptr: defaults to invalidptr=1, causing the garbage collector and stack 106 copier to crash the program if an invalid pointer value (for example, 1) 107 is found in a pointer-typed location. Setting invalidptr=0 disables this check. 108 This should only be used as a temporary workaround to diagnose buggy code. 109 The real fix is to not store integers in pointer-typed locations. 110 111 sbrk: setting sbrk=1 replaces the memory allocator and garbage collector 112 with a trivial allocator that obtains memory from the operating system and 113 never reclaims any memory. 114 115 scavenge: scavenge=1 enables debugging mode of heap scavenger. 116 117 scheddetail: setting schedtrace=X and scheddetail=1 causes the scheduler to emit 118 detailed multiline info every X milliseconds, describing state of the scheduler, 119 processors, threads and goroutines. 120 121 schedtrace: setting schedtrace=X causes the scheduler to emit a single line to standard 122 error every X milliseconds, summarizing the scheduler state. 123 124 tracebackancestors: setting tracebackancestors=N extends tracebacks with the stacks at 125 which goroutines were created, where N limits the number of ancestor goroutines to 126 report. This also extends the information returned by runtime.Stack. Ancestor's goroutine 127 IDs will refer to the ID of the goroutine at the time of creation; it's possible for this 128 ID to be reused for another goroutine. Setting N to 0 will report no ancestry information. 129 130 The net, net/http, and crypto/tls packages also refer to debugging variables in GODEBUG. 131 See the documentation for those packages for details. 132 133 The GOMAXPROCS variable limits the number of operating system threads that 134 can execute user-level Go code simultaneously. There is no limit to the number of threads 135 that can be blocked in system calls on behalf of Go code; those do not count against 136 the GOMAXPROCS limit. This package's GOMAXPROCS function queries and changes 137 the limit. 138 139 The GORACE variable configures the race detector, for programs built using -race. 140 See https://golang.org/doc/articles/race_detector.html for details. 141 142 The GOTRACEBACK variable controls the amount of output generated when a Go 143 program fails due to an unrecovered panic or an unexpected runtime condition. 144 By default, a failure prints a stack trace for the current goroutine, 145 eliding functions internal to the run-time system, and then exits with exit code 2. 146 The failure prints stack traces for all goroutines if there is no current goroutine 147 or the failure is internal to the run-time. 148 GOTRACEBACK=none omits the goroutine stack traces entirely. 149 GOTRACEBACK=single (the default) behaves as described above. 150 GOTRACEBACK=all adds stack traces for all user-created goroutines. 151 GOTRACEBACK=system is like ``all'' but adds stack frames for run-time functions 152 and shows goroutines created internally by the run-time. 153 GOTRACEBACK=crash is like ``system'' but crashes in an operating system-specific 154 manner instead of exiting. For example, on Unix systems, the crash raises 155 SIGABRT to trigger a core dump. 156 For historical reasons, the GOTRACEBACK settings 0, 1, and 2 are synonyms for 157 none, all, and system, respectively. 158 The runtime/debug package's SetTraceback function allows increasing the 159 amount of output at run time, but it cannot reduce the amount below that 160 specified by the environment variable. 161 See https://golang.org/pkg/runtime/debug/#SetTraceback. 162 163 The GOARCH, GOOS, GOPATH, and GOROOT environment variables complete 164 the set of Go environment variables. They influence the building of Go programs 165 (see https://golang.org/cmd/go and https://golang.org/pkg/go/build). 166 GOARCH, GOOS, and GOROOT are recorded at compile time and made available by 167 constants or functions in this package, but they do not influence the execution 168 of the run-time system. 169 */ 170 package runtime 171 172 import "runtime/internal/sys" 173 174 // Caller reports file and line number information about function invocations on 175 // the calling goroutine's stack. The argument skip is the number of stack frames 176 // to ascend, with 0 identifying the caller of Caller. (For historical reasons the 177 // meaning of skip differs between Caller and Callers.) The return values report the 178 // program counter, file name, and line number within the file of the corresponding 179 // call. The boolean ok is false if it was not possible to recover the information. 180 func Caller(skip int) (pc uintptr, file string, line int, ok bool) { 181 rpc := make([]uintptr, 1) 182 n := callers(skip+1, rpc[:]) 183 if n < 1 { 184 return 185 } 186 frame, _ := CallersFrames(rpc).Next() 187 return frame.PC, frame.File, frame.Line, frame.PC != 0 188 } 189 190 // Callers fills the slice pc with the return program counters of function invocations 191 // on the calling goroutine's stack. The argument skip is the number of stack frames 192 // to skip before recording in pc, with 0 identifying the frame for Callers itself and 193 // 1 identifying the caller of Callers. 194 // It returns the number of entries written to pc. 195 // 196 // To translate these PCs into symbolic information such as function 197 // names and line numbers, use CallersFrames. CallersFrames accounts 198 // for inlined functions and adjusts the return program counters into 199 // call program counters. Iterating over the returned slice of PCs 200 // directly is discouraged, as is using FuncForPC on any of the 201 // returned PCs, since these cannot account for inlining or return 202 // program counter adjustment. 203 //go:noinline 204 func Callers(skip int, pc []uintptr) int { 205 // runtime.callers uses pc.array==nil as a signal 206 // to print a stack trace. Pick off 0-length pc here 207 // so that we don't let a nil pc slice get to it. 208 if len(pc) == 0 { 209 return 0 210 } 211 return callers(skip, pc) 212 } 213 214 // GOROOT returns the root of the Go tree. It uses the 215 // GOROOT environment variable, if set at process start, 216 // or else the root used during the Go build. 217 func GOROOT() string { 218 s := gogetenv("GOROOT") 219 if s != "" { 220 return s 221 } 222 return sys.DefaultGoroot 223 } 224 225 // Version returns the Go tree's version string. 226 // It is either the commit hash and date at the time of the build or, 227 // when possible, a release tag like "go1.3". 228 func Version() string { 229 return sys.TheVersion 230 } 231 232 // GOOS is the running program's operating system target: 233 // one of darwin, freebsd, linux, and so on. 234 // To view possible combinations of GOOS and GOARCH, run "go tool dist list". 235 const GOOS string = sys.GOOS 236 237 // GOARCH is the running program's architecture target: 238 // one of 386, amd64, arm, s390x, and so on. 239 const GOARCH string = sys.GOARCH 240