gc: use raw stack access whenever possible

The only architecture that actually needs special support for scanning
the stack is WebAssembly. All others allow raw access to the stack with
a small bit of assembly. Therefore, don't manually keep track of all
these objects on the stack manually and instead just use conservative
stack scanning.

This results in a massive code size decrease in the affected targets
(only tested linux/amd64 for code size) - sometimes around 33%. It also
allows for future improvements such as using proper stackful goroutines.

src/runtime/runtime_unix.go

@@ -37,6 +37,8 @@ type timespec struct {
 
 const CLOCK_MONOTONIC_RAW = 4
 
+var stackTop uintptr
+
 func postinit() {}
 
 // Entry point for Go. Initialize all packages and call main.main().
@@ -44,12 +46,22 @@ func postinit() {}
 func main() int {
 	preinit()
 
-	run()
+	// Obtain the initial stack pointer right before calling the run() function.
+	// The run function has been moved to a separate (non-inlined) function so
+	// that the correct stack pointer is read.
+	stackTop = getCurrentStackPointer()
+	runMain()
 
 	// For libc compatibility.
 	return 0
 }
 
+// Must be a separate function to get the correct stack pointer.
+//go:noinline
+func runMain() {
+	run()
+}
+
 func putchar(c byte) {
 	_putchar(int(c))
 }