Auto merge of #128265 - DianQK:instsimplify-before-inline, r=saethlin

 Perform instsimplify before inline to eliminate some trivial calls

I am currently working on #128081. In the current pipeline, we can get the following clone statements ([godbolt](https://rust.godbolt.org/z/931316fhP)):

```
    bb0: {
        StorageLive(_2);
        _2 = ((*_1).0: i32);
        StorageLive(_3);
        _3 = ((*_1).1: u64);
        _0 = Foo { a: move _2, b: move _3 };
        StorageDead(_3);
        StorageDead(_2);
        return;
    }
```

Analyzing such statements will be simple and fast. We don't need to consider branches or some interfering statements. However, this requires us to run `InstSimplify`, `ReferencePropagation`, and `SimplifyCFG` at least once. I can introduce a new pass, but I think the best place for it would be within `InstSimplify`.

I put `InstSimplify` before `Inline`, which takes some of the burden away from `Inline`.

r? `@saethlin`

Author: bors

compiler/rustc_mir_transform/src/instsimplify.rs

@@ -13,9 +13,25 @@ use rustc_target::spec::abi::Abi;
 use crate::simplify::simplify_duplicate_switch_targets;
 use crate::take_array;
 
-pub struct InstSimplify;
+pub enum InstSimplify {
+    BeforeInline,
+    AfterSimplifyCfg,
+}
+
+impl InstSimplify {
+    pub fn name(&self) -> &'static str {
+        match self {
+            InstSimplify::BeforeInline => "InstSimplify-before-inline",
+            InstSimplify::AfterSimplifyCfg => "InstSimplify-after-simplifycfg",
+        }
+    }
+}
 
 impl<'tcx> MirPass<'tcx> for InstSimplify {
+    fn name(&self) -> &'static str {
+        self.name()
+    }
+
     fn is_enabled(&self, sess: &rustc_session::Session) -> bool {
         sess.mir_opt_level() > 0
     }

compiler/rustc_mir_transform/src/lib.rs

@@ -571,6 +571,8 @@ fn run_optimization_passes<'tcx>(tcx: TyCtxt<'tcx>, body: &mut Body<'tcx>) {
             // Has to be done before inlining, otherwise actual call will be almost always inlined.
             // Also simple, so can just do first
             &lower_slice_len::LowerSliceLenCalls,
+            // Perform instsimplify before inline to eliminate some trivial calls (like clone shims).
+            &instsimplify::InstSimplify::BeforeInline,
             // Perform inlining, which may add a lot of code.
             &inline::Inline,
             // Code from other crates may have storage markers, so this needs to happen after inlining.
@@ -590,7 +592,8 @@ fn run_optimization_passes<'tcx>(tcx: TyCtxt<'tcx>, body: &mut Body<'tcx>) {
             &match_branches::MatchBranchSimplification,
             // inst combine is after MatchBranchSimplification to clean up Ne(_1, false)
             &multiple_return_terminators::MultipleReturnTerminators,
-            &instsimplify::InstSimplify,
+            // After simplifycfg, it allows us to discover new opportunities for peephole optimizations.
+            &instsimplify::InstSimplify::AfterSimplifyCfg,
             &simplify::SimplifyLocals::BeforeConstProp,
             &dead_store_elimination::DeadStoreElimination::Initial,
             &gvn::GVN,