stacker: make remaining stack computation more resilient

Sounds like it is plausible for the stack pointer to end up past the
computed stack limit depending on the implementation of stack growing
and the exact timing at which the stack limit is determined.

Supersedes #51

Co-authored-by: Jude Nelson <judecn@gmail.com>
Co-authored-by: Reagan Bohan <xbjfk.github@gmail.com>

Author: 3 people

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "stacker"
-version = "0.1.20"
+version = "0.1.21"
 edition = "2021"
 rust-version = "1.63"
 authors = ["Alex Crichton <alex@alexcrichton.com>", "Simonas Kazlauskas <stacker@kazlauskas.me>"]

src/lib.rs

@@ -91,7 +91,7 @@ pub fn grow<R, F: FnOnce() -> R>(stack_size: usize, callback: F) -> R {
 /// to determine whether a stack switch should be made or not.
 pub fn remaining_stack() -> Option<usize> {
     let current_ptr = current_stack_ptr();
-    get_stack_limit().map(|limit| current_ptr - limit)
+    get_stack_limit().map(|limit| current_ptr.saturating_sub(limit))
 }
 
 psm_stack_information!(