Fix fallback-split behavior: trim start of minimal bundle wrt start of original LR. (#85)

When a liverange starts at a *late* point of an instruction, and it
undergoes the fallback "split into all minimal pieces" transform, we end
up creating one minimal bundle that starts at the *early* point of the
instruction at the start of the original LR. This can create
impossible-to-allocate situations where a fixed-constraint LR overlaps
another constrained to the same register (e.g. at calls). We fix this by
ensuring the minimal bundle is trimmed only to the half of the
instruction that overlaps the original LR.

This is analogous to the third fix in #74, but on the other end (start
of LR rather than end of it).

Author: cfallin

src/ion/process.rs

@@ -800,6 +800,7 @@ impl<'a, F: Function> Env<'a, F> {
         for entry_idx in 0..self.bundles[bundle.index()].ranges.len() {
             // Iterate manually; don't borrow `self`.
             let entry = self.bundles[bundle.index()].ranges[entry_idx];
+            let lr_from = entry.range.from;
             let lr_to = entry.range.to;
 
             removed_lrs.insert(entry.index);
@@ -863,6 +864,7 @@ impl<'a, F: Function> Env<'a, F> {
 
                 // Otherwise, create a new LR.
                 let pos = ProgPoint::before(u.pos.inst());
+                let pos = std::cmp::max(lr_from, pos);
                 let cr = CodeRange { from: pos, to };
                 let lr = self.create_liverange(cr);
                 new_lrs.push((vreg, lr));