Rollup merge of #89532 - ecstatic-morse:maybe-live-locals-enum, r=oli-obk,tmiasko

Document behavior of  `MaybeLiveLocals` regarding enums and field-senstivity

This arose from a [discussion on Zulip](https://rust-lang.zulipchat.com/#narrow/stream/189540-t-compiler.2Fwg-mir-opt/topic/MaybeLiveLocals.20and.20Discriminants) where a new contributor attempted to implement a dead-store elimination pass using this analysis. They ran into a nasty hack around `SetDiscriminant` the effect of which is to lets handle assignments of literals to enum-typed locals (e.g. `x = Some(4)`) correctly. This took me a while to figure out.

Document this oddity, so the next person will have an easier time, and add a test to enshrine the current behavior.

r? ``@tmiasko``

3 files changed, 63 insertions, 0 deletions

diff --git a/compiler/rustc_mir_dataflow/src/impls/liveness.rs b/compiler/rustc_mir_dataflow/src/impls/liveness.rs
index 0039d3188d5..3e2548845e2 100644
--- a/compiler/rustc_mir_dataflow/src/impls/liveness.rs
+++ b/compiler/rustc_mir_dataflow/src/impls/liveness.rs
@@ -11,6 +11,37 @@ use crate::{AnalysisDomain, Backward, GenKill, GenKillAnalysis};
 /// exist. See [this `mir-dataflow` test][flow-test] for an example. You almost never want to use
 /// this analysis without also looking at the results of [`MaybeBorrowedLocals`].
 ///
+/// ## Field-(in)sensitivity
+///
+/// As the name suggests, this analysis is field insensitive. If a projection of a variable `x` is
+/// assigned to (e.g. `x.0 = 42`), it does not "define" `x` as far as liveness is concerned. In fact,
+/// such an assignment is currently marked as a "use" of `x` in an attempt to be maximally
+/// conservative.
+///
+/// ## Enums and `SetDiscriminant`
+///
+/// Assigning a literal value to an `enum` (e.g. `Option<i32>`), does not result in a simple
+/// assignment of the form `_1 = /*...*/` in the MIR. For example, the following assignment to `x`:
+///
+/// ```
+/// x = Some(4);
+/// ```
+///
+/// compiles to this MIR
+///
+/// ```
+/// ((_1 as Some).0: i32) = const 4_i32;
+/// discriminant(_1) = 1;
+/// ```
+///
+/// However, `MaybeLiveLocals` **does** mark `x` (`_1`) as "killed" after a statement like this.
+/// That's because it treats the `SetDiscriminant` operation as a definition of `x`, even though
+/// the writes that actually initialized the locals happened earlier.
+///
+/// This makes `MaybeLiveLocals` unsuitable for certain classes of optimization normally associated
+/// with a live variables analysis, notably dead-store elimination. It's a dirty hack, but it works
+/// okay for the generator state transform (currently the main consumuer of this analysis).
+///
 /// [`MaybeBorrowedLocals`]: super::MaybeBorrowedLocals
 /// [flow-test]: https://github.com/rust-lang/rust/blob/a08c47310c7d49cbdc5d7afb38408ba519967ecd/src/test/ui/mir-dataflow/liveness-ptr.rs
 /// [liveness]: https://en.wikipedia.org/wiki/Live_variable_analysis
diff --git a/src/test/ui/mir-dataflow/liveness-enum.rs b/src/test/ui/mir-dataflow/liveness-enum.rs
new file mode 100644
index 00000000000..5eb04ae8c8d
--- /dev/null
+++ b/src/test/ui/mir-dataflow/liveness-enum.rs
@@ -0,0 +1,22 @@
+#![feature(core_intrinsics, rustc_attrs)]
+
+use std::intrinsics::rustc_peek;
+
+#[rustc_mir(rustc_peek_liveness, stop_after_dataflow)]
+fn foo() -> Option<i32> {
+    let mut x = None;
+
+    // `x` is live here since it is used in the next statement...
+    rustc_peek(x);
+
+    dbg!(x);
+
+    // But not here, since it is overwritten below
+    rustc_peek(x); //~ ERROR rustc_peek: bit not set
+
+    x = Some(4);
+
+    x
+}
+
+fn main() {}
diff --git a/src/test/ui/mir-dataflow/liveness-enum.stderr b/src/test/ui/mir-dataflow/liveness-enum.stderr
new file mode 100644
index 00000000000..483944d731a
--- /dev/null
+++ b/src/test/ui/mir-dataflow/liveness-enum.stderr
@@ -0,0 +1,10 @@
+error: rustc_peek: bit not set
+  --> $DIR/liveness-enum.rs:15:5
+   |
+LL |     rustc_peek(x);
+   |     ^^^^^^^^^^^^^
+
+error: stop_after_dataflow ended compilation
+
+error: aborting due to 2 previous errors
+