use a statically-allocated stack for Context::new benchmarking

Now it's an actual context creation benchmark, instead of benchmarking
our stack creation, which is swappable anyway.

benches/context_new.rs

@@ -1,14 +1,15 @@
 #![feature(test)]
 extern crate test;
 extern crate lwkt;
-use lwkt::{Context, StackSource};
+use lwkt::{Context, Stack};
 
-static mut ctx_slot: *mut Context<lwkt::os::Stack> = 0 as *mut Context<_>;
+static mut ctx_slot: *mut Context<SliceStack<'static>> = 0 as *mut Context<_>;
+static mut stack_buf: [u8; 1024] = [0; 1024];
 
 #[bench]
 fn context_new(b: &mut test::Bencher) {
   b.iter(|| unsafe {
-    let stack = lwkt::os::StackSource::get_stack(4 << 20);
+    let stack = SliceStack(&mut stack_buf);
 
     let mut ctx = Context::new(stack, move || {
       let ctx_ptr = ctx_slot;
@@ -22,3 +23,16 @@ fn context_new(b: &mut test::Bencher) {
     ctx.swap();
   })
 }
+
+struct SliceStack<'a>(&'a mut [u8]);
+impl<'a> lwkt::Stack for SliceStack<'a> {
+  fn top(&mut self) -> *mut u8 {
+    unsafe {
+      self.0.as_mut_ptr().offset(self.0.len() as isize)
+    }
+  }
+
+  fn limit(&self) -> *const u8 {
+    self.0.as_ptr()
+  }
+}