4 files changed, 45 insertions, 2 deletions
diff --git a/src/runtime/mgcpacer.go b/src/runtime/mgcpacer.go
index 094dcc701a..9cc7cf99db 100644
--- a/src/runtime/mgcpacer.go
+++ b/src/runtime/mgcpacer.go
@@ -47,6 +47,10 @@ const (
 
 	// defaultHeapMinimum is the value of heapMinimum for GOGC==100.
 	defaultHeapMinimum = 4 << 20
+
+	// scannableStackSizeSlack is the bytes of stack space allocated or freed
+	// that can accumulate on a P before updating gcController.stackSize.
+	scannableStackSizeSlack = 8 << 10
 )
 
 func init() {
@@ -166,6 +170,18 @@ type gcControllerState struct {
 	// Read and written atomically or with the world stopped.
 	heapScan uint64
 
+	// stackScan is a snapshot of scannableStackSize taken at each GC
+	// STW pause and is used in pacing decisions.
+	//
+	// Updated only while the world is stopped.
+	stackScan uint64
+
+	// scannableStackSize is the amount of allocated goroutine stack space in
+	// use by goroutines.
+	//
+	// Read and updated atomically.
+	scannableStackSize uint64
+
 	// heapMarked is the number of bytes marked by the previous
 	// GC. After mark termination, heapLive == heapMarked, but
 	// unlike heapLive, heapMarked does not change until the
@@ -276,6 +292,7 @@ func (c *gcControllerState) startCycle(markStartTime int64) {
 	c.fractionalMarkTime = 0
 	c.idleMarkTime = 0
 	c.markStartTime = markStartTime
+	c.stackScan = atomic.Load64(&c.scannableStackSize)
 
 	// Ensure that the heap goal is at least a little larger than
 	// the current live heap size. This may not be the case if GC
@@ -686,6 +703,18 @@ func (c *gcControllerState) update(dHeapLive, dHeapScan int64) {
 	}
 }
 
+func (c *gcControllerState) addScannableStack(pp *p, amount int64) {
+	if pp == nil {
+		atomic.Xadd64(&c.scannableStackSize, amount)
+		return
+	}
+	pp.scannableStackSizeDelta += amount
+	if pp.scannableStackSizeDelta >= scannableStackSizeSlack || pp.scannableStackSizeDelta <= -scannableStackSizeSlack {
+		atomic.Xadd64(&c.scannableStackSize, pp.scannableStackSizeDelta)
+		pp.scannableStackSizeDelta = 0
+	}
+}
+
 // commit sets the trigger ratio and updates everything
 // derived from it: the absolute trigger, the heap goal, mark pacing,
 // and sweep pacing.
diff --git a/src/runtime/proc.go b/src/runtime/proc.go
index bd7dacd440..615f53d31f 100644
--- a/src/runtime/proc.go
+++ b/src/runtime/proc.go
@@ -3623,8 +3623,10 @@ func goexit1() {
 // goexit continuation on g0.
 func goexit0(gp *g) {
 	_g_ := getg()
+	_p_ := _g_.m.p.ptr()
 
 	casgstatus(gp, _Grunning, _Gdead)
+	gcController.addScannableStack(_p_, -int64(gp.stack.hi-gp.stack.lo))
 	if isSystemGoroutine(gp, false) {
 		atomic.Xadd(&sched.ngsys, -1)
 	}
@@ -3655,7 +3657,7 @@ func goexit0(gp *g) {
 	dropg()
 
 	if GOARCH == "wasm" { // no threads yet on wasm
-		gfput(_g_.m.p.ptr(), gp)
+		gfput(_p_, gp)
 		schedule() // never returns
 	}
 
@@ -3663,7 +3665,7 @@ func goexit0(gp *g) {
 		print("invalid m->lockedInt = ", _g_.m.lockedInt, "\n")
 		throw("internal lockOSThread error")
 	}
-	gfput(_g_.m.p.ptr(), gp)
+	gfput(_p_, gp)
 	if locked {
 		// The goroutine may have locked this thread because
 		// it put it in an unusual kernel state. Kill it
@@ -4292,6 +4294,7 @@ func newproc1(fn *funcval, callergp *g, callerpc uintptr) *g {
 		newg.tracking = true
 	}
 	casgstatus(newg, _Gdead, _Grunnable)
+	gcController.addScannableStack(_p_, int64(newg.stack.hi-newg.stack.lo))
 
 	if _p_.goidcache == _p_.goidcacheend {
 		// Sched.goidgen is the last allocated id,
diff --git a/src/runtime/runtime2.go b/src/runtime/runtime2.go
index db1c6e307b..bfd857e8d5 100644
--- a/src/runtime/runtime2.go
+++ b/src/runtime/runtime2.go
@@ -734,6 +734,12 @@ type p struct {
 	// Race context used while executing timer functions.
 	timerRaceCtx uintptr
 
+	// scannableStackSizeDelta accumulates the amount of stack space held by
+	// live goroutines (i.e. those eligible for stack scanning).
+	// Flushed to gcController.scannableStackSize once scannableStackSizeSlack
+	// or -scannableStackSizeSlack is reached.
+	scannableStackSizeDelta int64
+
 	// preempt is set to indicate that this P should be enter the
 	// scheduler ASAP (regardless of what G is running on it).
 	preempt bool
diff --git a/src/runtime/stack.go b/src/runtime/stack.go
index 284c6b3b84..8ae9c1e698 100644
--- a/src/runtime/stack.go
+++ b/src/runtime/stack.go
@@ -852,6 +852,11 @@ func copystack(gp *g, newsize uintptr) {
 		throw("nil stackbase")
 	}
 	used := old.hi - gp.sched.sp
+	// Add just the difference to gcController.addScannableStack.
+	// g0 stacks never move, so this will never account for them.
+	// It's also fine if we have no P, addScannableStack can deal with
+	// that case.
+	gcController.addScannableStack(getg().m.p.ptr(), int64(newsize)-int64(old.hi-old.lo))
 
 	// allocate new stack
 	new := stackalloc(uint32(newsize))