runtime: refactor obtaining the system stack

The system stack is only needed when we're not on it. So we can directly
call task.SystemStack() without problems.

This also saves a tiny bit of binary size.

src/runtime/gc_stack_raw.go

@@ -20,7 +20,7 @@ func markStack() {
 
 	if !task.OnSystemStack() {
 		// Mark system stack.
-		markRoots(getSystemStackPointer(), stackTop)
+		markRoots(task.SystemStack(), stackTop)
 	}
 }
 

src/runtime/scheduler_none.go

@@ -73,9 +73,3 @@ func scheduler(returnAtDeadlock bool) {
 	// this code should be unreachable.
 	runtimePanic("unreachable: scheduler must not be called with the 'none' scheduler")
 }
-
-// getSystemStackPointer returns the current stack pointer of the system stack.
-// This is always the current stack pointer.
-func getSystemStackPointer() uintptr {
-	return getCurrentStackPointer()
-}

src/runtime/scheduler_tasks.go

@@ -1,17 +0,0 @@
-//go:build scheduler.tasks
-
-package runtime
-
-import "internal/task"
-
-// getSystemStackPointer returns the current stack pointer of the system stack.
-// This is not necessarily the same as the current stack pointer.
-func getSystemStackPointer() uintptr {
-	// TODO: this always returns the correct stack on Cortex-M, so don't bother
-	// comparing against 0.
-	sp := task.SystemStack()
-	if sp == 0 {
-		sp = getCurrentStackPointer()
-	}
-	return sp
-}