共计 6981 个字符,预计需要花费 18 分钟才能阅读完成。
syscall 函数
Syscall 函数的定义如下,传入 4 个参数,返回 3 个参数。
func syscall(fn, a1, a2, a3 uintptr) (r1, r2 uintptr, err Errno)
syscall 函数的作用是传入零碎调用的地址和参数,执行实现后返回。流程次要是零碎调用前执行 entersyscall,设置 g p 的状态,而后入参,执行后,写返回值而后执行 exitsyscall 设置 g p 的状态。
entersyscall 和 exitsyscall 在 g 的调用中细讲。
// func Syscall(trap int64, a1, a2, a3 uintptr) (r1, r2, err uintptr);
// Trap # in AX, args in DI SI DX R10 R8 R9, return in AX DX
// Note that this differs from "standard" ABI convention, which
// would pass 4th arg in CX, not R10.
// 4 个入参:PC param1 param2 param3
TEXT ·Syscall(SB),NOSPLIT,$0-56
// 调用 entersyscall 判断是执行条件是否满足 记录调度信息 切换 g p 的状态
CALL runtime·entersyscall(SB)
// 将参数存入寄存器中
MOVQ a1+8(FP), DI
MOVQ a2+16(FP), SI
MOVQ a3+24(FP), DX
MOVQ trap+0(FP), AX // syscall entry
SYSCALL
CMPQ AX, $0xfffffffffffff001
JLS ok
// 执行失败时 写返回值
MOVQ $-1, r1+32(FP)
MOVQ $0, r2+40(FP)
NEGQ AX
MOVQ AX, err+48(FP)
// 调用 exitsyscall 记录调度信息
CALL runtime·exitsyscall(SB)
RET
ok:
// 执行胜利时 写返回值
MOVQ AX, r1+32(FP)
MOVQ DX, r2+40(FP)
MOVQ $0, err+48(FP)
CALL runtime·exitsyscall(SB)
RET
TEXT ·RawSyscall(SB),NOSPLIT,$0-56
MOVQ a1+8(FP), DI
MOVQ a2+16(FP), SI
MOVQ a3+24(FP), DX
MOVQ trap+0(FP), AX // syscall entry
SYSCALL
JCC ok1
MOVQ $-1, r1+32(FP) // r1
MOVQ $0, r2+40(FP) // r2
MOVQ AX, err+48(FP) // errno
RET
ok1:
MOVQ AX, r1+32(FP) // r1
MOVQ DX, r2+40(FP) // r2
MOVQ $0, err+48(FP) // errno
RET
显著 SysCall 比 RawSyscall 多调用了两个办法,entersyscall 和 exitsyscall,减少这两个函数的调用,让调度器有机会去对行将要进入零碎调用的 goroutine 进行调整,不便调度。
entersyscall
// 零碎调用的时候调用该函数
// 进入零碎调用,G 将会进入_Gsyscall 状态,也就是会被临时挂起,直到零碎调用完结。// 此时 M 进入零碎调用,那么 P 也会放弃该 M。然而,此时 M 还指向 P,在 M 从零碎调用返回后还能找到 P
func entersyscall() {reentersyscall(getcallerpc(), getcallersp())
}
// Syscall 跟踪:// 在零碎调用开始时,咱们收回 traceGoSysCall 来捕捉堆栈跟踪。// 如果零碎调用未阻止,则咱们不会收回任何其余事件。// 如果零碎调用被阻止(即,从新获取了 P),则 retaker 会收回 traceGoSysBlock;// 当 syscall 返回时,咱们收回 traceGoSysExit,当 goroutine 开始运行时
//(可能立刻,如果 exitsyscallfast 返回 true),咱们收回 traceGoStart。// 为了确保在 traceGoSysBlock 之后严格收回 traceGoSysExit,// 咱们记得 syscalltick 的以后值以 m 为单位(_g_.m.syscalltick = _g_.m.p.ptr()。syscalltick),// 之后收回 traceGoSysBlock 的人将递增 p.syscalltick;// 咱们在收回 traceGoSysExit 之前期待增量。// 请留神,即便未启用跟踪,增量也会实现,// 因为能够在 syscall 的两头启用跟踪。咱们不心愿期待挂起。//go:nosplit
func reentersyscall(pc, sp uintptr) {_g_ := getg()
// 禁用抢占,因为在此性能期间 g 处于 Gsyscall 状态,但 g -> sched 可能不统一,请勿让 GC 察看它。_g_.m.locks++
// Entersyscall must not call any function that might split/grow the stack.
// (See details in comment above.)
// 捕捉可能产生的调用,办法是将堆栈爱护替换为会使任何堆栈查看失败的内容,并留下一个标记来告诉 newstack 终止。_g_.stackguard0 = stackPreempt
_g_.throwsplit = true
// Leave SP around for GC and traceback.
save(pc, sp)
_g_.syscallsp = sp
_g_.syscallpc = pc
// 让 G 进入_Gsyscall 状态,此时 G 曾经被挂起了,直到零碎调用完结,才会让 G 从新写进入 running
casgstatus(_g_, _Grunning, _Gsyscall)
if _g_.syscallsp < _g_.stack.lo || _g_.stack.hi < _g_.syscallsp {systemstack(func() {print("entersyscall inconsistent", hex(_g_.syscallsp), "[", hex(_g_.stack.lo), ",", hex(_g_.stack.hi), "]\n")
throw("entersyscall")
})
}
if trace.enabled {systemstack(traceGoSysCall)
// systemstack itself clobbers g.sched.{pc,sp} and we might
// need them later when the G is genuinely blocked in a
// syscall
save(pc, sp)
}
if atomic.Load(&sched.sysmonwait) != 0 {systemstack(entersyscall_sysmon)
save(pc, sp)
}
if _g_.m.p.ptr().runSafePointFn != 0 {
// runSafePointFn may stack split if run on this stack
systemstack(runSafePointFn)
save(pc, sp)
}
_g_.m.syscalltick = _g_.m.p.ptr().syscalltick
_g_.sysblocktraced = true
// 这里很要害:P 的 M 曾经陷入零碎调用,于是 P 忍痛放弃该 M
// 然而请留神:此时 M 还指向 P,在 M 从零碎调用返回后还能找到 P
pp := _g_.m.p.ptr()
pp.m = 0
_g_.m.oldp.set(pp)
_g_.m.p = 0
// P 的状态变为 Psyscall
atomic.Store(&pp.status, _Psyscall)
if sched.gcwaiting != 0 {systemstack(entersyscall_gcwait)
save(pc, sp)
}
_g_.m.locks--
}
该办法次要是为零碎调用前做了筹备工作:
- 批改 g 的状态为_Gsyscall
- 查看 sysmon 线程是否在执行,睡眠须要唤醒
- p 放弃 m,然而 m 仍旧持有 p 的指针,完结调用后优先选择 p
- 批改 p 的状态为_Psyscal
做好这些筹备工作便能够真正的执行零碎调用了。当该线程 m 长时间阻塞在零碎调用的时候,始终在运行的 sysmon 线程会检测到该 p 的状态,并将其剥离,驱动其余的 m(新建或获取)来调度执行该 p 上的工作, 这其中次要是在 retake 办法中实现的,该办法还解决了 goroutine 抢占调度,这里省略,前面介绍抢占调度在介绍:
exitsyscall
当零碎 Syscall 返回的时,会调用 exitsyscall 办法复原调度:
//go:nosplit
//go:nowritebarrierrec
//go:linkname exitsyscall
func exitsyscall() {_g_ := getg()
_g_.m.locks++ // see comment in entersyscall
if getcallersp() > _g_.syscallsp {throw("exitsyscall: syscall frame is no longer valid")
}
_g_.waitsince = 0
oldp := _g_.m.oldp.ptr()
_g_.m.oldp = 0
// 从新获取 p
if exitsyscallfast(oldp) {
if trace.enabled {if oldp != _g_.m.p.ptr() || _g_.m.syscalltick != _g_.m.p.ptr().syscalltick {systemstack(traceGoStart)
}
}
// There's a cpu for us, so we can run.
_g_.m.p.ptr().syscalltick++
// We need to cas the status and scan before resuming...
casgstatus(_g_, _Gsyscall, _Grunning)
// Garbage collector isn't running (since we are),
// so okay to clear syscallsp.
_g_.syscallsp = 0
_g_.m.locks--
if _g_.preempt {
// restore the preemption request in case we've cleared it in newstack
_g_.stackguard0 = stackPreempt
} else {
// otherwise restore the real _StackGuard, we've spoiled it in entersyscall/entersyscallblock
_g_.stackguard0 = _g_.stack.lo + _StackGuard
}
_g_.throwsplit = false
if sched.disable.user && !schedEnabled(_g_) {
// Scheduling of this goroutine is disabled.
Gosched()}
return
}
_g_.sysexitticks = 0
if trace.enabled {
// Wait till traceGoSysBlock event is emitted.
// This ensures consistency of the trace (the goroutine is started after it is blocked).
for oldp != nil && oldp.syscalltick == _g_.m.syscalltick {osyield()
}
// We can't trace syscall exit right now because we don't have a P.
// Tracing code can invoke write barriers that cannot run without a P.
// So instead we remember the syscall exit time and emit the event
// in execute when we have a P.
_g_.sysexitticks = cputicks()}
_g_.m.locks--
// 没有获取到 p,只能解绑以后 g,从新调度该 m 了
mcall(exitsyscall0)
// Scheduler returned, so we're allowed to run now.
// Delete the syscallsp information that we left for
// the garbage collector during the system call.
// Must wait until now because until gosched returns
// we don't know for sure that the garbage collector
// is not running.
_g_.syscallsp = 0
_g_.m.p.ptr().syscalltick++
_g_.throwsplit = false
}
exitsyscallfast
exitsyscall 会尝试从新绑定 p,优先选择之前 m 绑定的 p(进入零碎的调用的时候,p 只是单方面解绑了和 m 的关系,通过 m 仍旧能够找到 p):
//go:nosplit
func exitsyscallfast(oldp *p) bool {_g_ := getg()
// Freezetheworld sets stopwait but does not retake P's.
//stw,间接解绑 p,而后退出
if sched.stopwait == freezeStopWait {return false}
// Try to re-acquire the last P.
// 如果之前从属的 P 尚未被其余 M, 尝试绑定该 P
if oldp != nil && oldp.status == _Psyscall && atomic.Cas(&oldp.status, _Psyscall, _Pidle) {
// There's a cpu for us, so we can run.
wirep(oldp)
exitsyscallfast_reacquired()
return true
}
// 否则从闲暇 P 列表中取出一个来
// Try to get any other idle P.
if sched.pidle != 0 {
var ok bool
systemstack(func() {ok = exitsyscallfast_pidle()
if ok && trace.enabled {
if oldp != nil {
// Wait till traceGoSysBlock event is emitted.
// This ensures consistency of the trace (the goroutine is started after it is blocked).
for oldp.syscalltick == _g_.m.syscalltick {osyield()
}
}
traceGoSysExit(0)
}
})
if ok {return true}
}
return false
}
exitsyscall0
func exitsyscall0(gp *g) {_g_ := getg()
// 批改 g 状态为 _Grunable
casgstatus(gp, _Gsyscall, _Grunnable)
dropg() // 解绑
lock(&sched.lock)
var _p_ *p
// 尝试获取 p
if schedEnabled(_g_) {_p_ = pidleget()
}
if _p_ == nil {
// 未获取到 p,g 进入全局队列期待调度
globrunqput(gp)
} else if atomic.Load(&sched.sysmonwait) != 0 {atomic.Store(&sched.sysmonwait, 0)
notewakeup(&sched.sysmonnote)
}
unlock(&sched.lock)
// 获取到 p,绑定,而后执行
if _p_ != nil {acquirep(_p_)
execute(gp, false) // Never returns.
}
// // m 有绑定的 g,解绑 p 而后绑定的 g 来唤醒,执行
if _g_.m.lockedg != 0 {
// Wait until another thread schedules gp and so m again.
stoplockedm()
execute(gp, false) // Never returns.
}
// 关联 p 失败了,休眠,期待唤醒,在进行调度。stopm()
schedule() // Never returns.}
总结
上述便是 golang 零碎调用的整个流程,大抵如下:
- 业务调用封装好的零碎调用函数,编译器翻译到 Syscall
- 执行 entersyscall() 办法,批改 g,p 的状态,p 单方面解绑 m,并查看唤醒 sysmon 线程,检测零碎调用。
- 当 sysmon 线程检测到零碎调用阻塞工夫过长的时候,调用 retake,从新调度该 p,让 p 上可执行的得以执行,不浪费资源
- 零碎调用返回,进入 exitsyscall 办法,优先获取之前的 p,如果该 p 曾经被占有,从新获取闲暇的 p,绑定,而后继续执行该 g。当获取不到 p 的时候,调用 exitsyscall0,解绑 g,休眠,期待下次唤醒调度。
正文完