Auto merge of #150309 - dianqk:ssa-range, r=<try>

[EXPERIMENT] New MIR Pass: SsaRangePropagation

Author: rust-bors[bot]

compiler/rustc_middle/src/mir/statement.rs

@@ -374,6 +374,12 @@ impl<'tcx> Place<'tcx> {
         self.projection.iter().any(|elem| elem.is_indirect())
     }
 
+    /// Returns `true` if the `Place` always refers to the same memory region
+    /// whatever the state of the program.
+    pub fn is_stable_offset(&self) -> bool {
+        self.projection.iter().all(|elem| elem.is_stable_offset())
+    }
+
     /// Returns `true` if this `Place`'s first projection is `Deref`.
     ///
     /// This is useful because for MIR phases `AnalysisPhase::PostCleanup` and later,

compiler/rustc_mir_transform/src/lib.rs

@@ -198,6 +198,7 @@ declare_passes! {
     mod single_use_consts : SingleUseConsts;
     mod sroa : ScalarReplacementOfAggregates;
     mod strip_debuginfo : StripDebugInfo;
+    mod ssa_range_prop: SsaRangePropagation;
     mod unreachable_enum_branching : UnreachableEnumBranching;
     mod unreachable_prop : UnreachablePropagation;
     mod validate : Validator;
@@ -743,6 +744,7 @@ pub(crate) fn run_optimization_passes<'tcx>(tcx: TyCtxt<'tcx>, body: &mut Body<'
             &simplify::SimplifyLocals::AfterGVN,
             &match_branches::MatchBranchSimplification,
             &dataflow_const_prop::DataflowConstProp,
+            &ssa_range_prop::SsaRangePropagation,
             &single_use_consts::SingleUseConsts,
             &o1(simplify_branches::SimplifyConstCondition::AfterConstProp),
             &jump_threading::JumpThreading,

compiler/rustc_mir_transform/src/ssa_range_prop.rs

@@ -0,0 +1,217 @@
+use rustc_abi::WrappingRange;
+use rustc_const_eval::interpret::Scalar;
+use rustc_data_structures::fx::FxHashMap;
+use rustc_data_structures::graph::dominators::Dominators;
+use rustc_index::bit_set::DenseBitSet;
+use rustc_middle::mir::visit::MutVisitor;
+use rustc_middle::mir::{BasicBlock, Body, Location, Operand, Place, TerminatorKind, *};
+use rustc_middle::ty::{TyCtxt, TypingEnv};
+use rustc_span::DUMMY_SP;
+
+use crate::ssa::SsaLocals;
+
+pub(super) struct SsaRangePropagation;
+
+impl<'tcx> crate::MirPass<'tcx> for SsaRangePropagation {
+    fn is_enabled(&self, sess: &rustc_session::Session) -> bool {
+        sess.mir_opt_level() > 1
+    }
+
+    fn run_pass(&self, tcx: TyCtxt<'tcx>, body: &mut Body<'tcx>) {
+        let typing_env = body.typing_env(tcx);
+        let ssa = SsaLocals::new(tcx, body, typing_env);
+        // Clone dominators because we need them while mutating the body.
+        let dominators = body.basic_blocks.dominators().clone();
+        let mut range_set =
+            RangeSet::new(tcx, typing_env, body, &ssa, &body.local_decls, dominators);
+
+        let reverse_postorder = body.basic_blocks.reverse_postorder().to_vec();
+        for bb in reverse_postorder {
+            let data = &mut body.basic_blocks.as_mut_preserves_cfg()[bb];
+            range_set.visit_basic_block_data(bb, data);
+        }
+    }
+
+    fn is_required(&self) -> bool {
+        false
+    }
+}
+
+struct RangeSet<'tcx, 'body, 'a> {
+    tcx: TyCtxt<'tcx>,
+    typing_env: TypingEnv<'tcx>,
+    ssa: &'a SsaLocals,
+    local_decls: &'body LocalDecls<'tcx>,
+    dominators: Dominators<BasicBlock>,
+    /// Known ranges at each locations.
+    ranges: FxHashMap<Place<'tcx>, Vec<(Location, WrappingRange)>>,
+    /// Determines if the basic block has a single unique predecessor.
+    unique_predecessors: DenseBitSet<BasicBlock>,
+}
+
+impl<'tcx, 'body, 'a> RangeSet<'tcx, 'body, 'a> {
+    fn new(
+        tcx: TyCtxt<'tcx>,
+        typing_env: TypingEnv<'tcx>,
+        body: &Body<'tcx>,
+        ssa: &'a SsaLocals,
+        local_decls: &'body LocalDecls<'tcx>,
+        dominators: Dominators<BasicBlock>,
+    ) -> Self {
+        let predecessors = body.basic_blocks.predecessors();
+        let mut unique_predecessors = DenseBitSet::new_empty(body.basic_blocks.len());
+        for (bb, _) in body.basic_blocks.iter_enumerated() {
+            if predecessors[bb].len() == 1 {
+                unique_predecessors.insert(bb);
+            }
+        }
+        RangeSet {
+            tcx,
+            typing_env,
+            ssa,
+            local_decls,
+            dominators,
+            ranges: FxHashMap::default(),
+            unique_predecessors,
+        }
+    }
+
+    /// Create a new known range at the location.
+    fn insert_range(&mut self, place: Place<'tcx>, location: Location, range: WrappingRange) {
+        self.ranges.entry(place).or_default().push((location, range));
+    }
+
+    /// Get the known range at the location.
+    fn get_range(&self, place: &Place<'tcx>, location: Location) -> Option<WrappingRange> {
+        let Some(ranges) = self.ranges.get(place) else {
+            return None;
+        };
+        // FIXME: This should use the intersection of all valid ranges.
+        let (_, range) =
+            ranges.iter().find(|(range_loc, _)| range_loc.dominates(location, &self.dominators))?;
+        Some(*range)
+    }
+
+    fn try_as_constant(
+        &mut self,
+        place: Place<'tcx>,
+        location: Location,
+    ) -> Option<ConstOperand<'tcx>> {
+        if let Some(range) = self.get_range(&place, location)
+            && range.start == range.end
+        {
+            let ty = place.ty(self.local_decls, self.tcx).ty;
+            let layout = self.tcx.layout_of(self.typing_env.as_query_input(ty)).ok()?;
+            let value = ConstValue::Scalar(Scalar::from_uint(range.start, layout.size));
+            let const_ = Const::Val(value, ty);
+            return Some(ConstOperand { span: DUMMY_SP, user_ty: None, const_ });
+        }
+        None
+    }
+
+    fn simplify_operand(
+        &mut self,
+        operand: &mut Operand<'tcx>,
+        location: Location,
+    ) -> Result<(), Option<Place<'tcx>>> {
+        let Some(place) = operand.place() else {
+            return Ok(());
+        };
+        let Some(const_) = self.try_as_constant(place, location) else {
+            if self.is_ssa(place) {
+                return Err(Some(place));
+            } else {
+                return Err(None);
+            }
+        };
+        *operand = Operand::Constant(Box::new(const_));
+        Ok(())
+    }
+
+    fn is_ssa(&self, place: Place<'tcx>) -> bool {
+        self.ssa.is_ssa(place.local) && place.is_stable_offset()
+    }
+}
+
+impl<'tcx> MutVisitor<'tcx> for RangeSet<'tcx, '_, '_> {
+    fn tcx(&self) -> TyCtxt<'tcx> {
+        self.tcx
+    }
+
+    fn visit_operand(&mut self, operand: &mut Operand<'tcx>, location: Location) {
+        let _ = self.simplify_operand(operand, location);
+    }
+
+    fn visit_statement(&mut self, statement: &mut Statement<'tcx>, location: Location) {
+        match &mut statement.kind {
+            StatementKind::Intrinsic(box NonDivergingIntrinsic::Assume(operand)) => {
+                if let Err(Some(place)) = self.simplify_operand(operand, location) {
+                    let successor = location.successor_within_block();
+                    let range = WrappingRange { start: 1, end: 1 };
+                    self.insert_range(place, successor, range);
+                }
+            }
+            _ => {}
+        }
+        self.super_statement(statement, location);
+    }
+
+    fn visit_terminator(&mut self, terminator: &mut Terminator<'tcx>, location: Location) {
+        match &mut terminator.kind {
+            TerminatorKind::Assert { cond, expected, target, .. } => {
+                if let Err(Some(place)) = self.simplify_operand(cond, location) {
+                    let successor = Location { block: *target, statement_index: 0 };
+                    if location.block != successor.block
+                        && self.unique_predecessors.contains(successor.block)
+                    {
+                        let val = *expected as u128;
+                        let range = WrappingRange { start: val, end: val };
+                        self.insert_range(place, successor, range);
+                    }
+                }
+            }
+            TerminatorKind::SwitchInt { discr, targets } => {
+                if let Err(Some(place)) = self.simplify_operand(discr, location)
+                    && targets.all_targets().len() < 8
+                {
+                    let mut distinct_targets: FxHashMap<BasicBlock, u8> = FxHashMap::default();
+                    for (_, target) in targets.iter() {
+                        let targets = distinct_targets.entry(target).or_default();
+                        if *targets == 0 {
+                            *targets = 1;
+                        } else {
+                            *targets = 2;
+                        }
+                    }
+                    for (val, target) in targets.iter() {
+                        if distinct_targets[&target] != 1 {
+                            continue;
+                        }
+                        let successor = Location { block: target, statement_index: 0 };
+                        if location.block != successor.block
+                            && self.unique_predecessors.contains(successor.block)
+                        {
+                            let range = WrappingRange { start: val, end: val };
+                            self.insert_range(place, successor, range);
+                        }
+                    }
+
+                    let otherwise = Location { block: targets.otherwise(), statement_index: 0 };
+                    if place.ty(self.local_decls, self.tcx).ty.is_bool()
+                        && let [val] = targets.all_values()
+                        && location.block != otherwise.block
+                        && self.unique_predecessors.contains(otherwise.block)
+                    {
+                        let range = if val.get() == 0 {
+                            WrappingRange { start: 1, end: 1 }
+                        } else {
+                            WrappingRange { start: 0, end: 0 }
+                        };
+                        self.insert_range(place, otherwise, range);
+                    }
+                }
+            }
+            _ => {}
+        }
+    }
+}

tests/mir-opt/range/ssa_range.on_assert.SsaRangePropagation.diff

@@ -0,0 +1,69 @@
+- // MIR for `on_assert` before SsaRangePropagation
++ // MIR for `on_assert` after SsaRangePropagation
+  
+  fn on_assert(_1: usize, _2: &[u8]) -> u8 {
+      debug i => _1;
+      debug v => _2;
+      let mut _0: u8;
+      let _3: ();
+      let mut _4: bool;
+      let mut _5: usize;
+      let mut _6: usize;
+      let mut _7: &[u8];
+      let mut _8: !;
+      let _9: usize;
+      let mut _10: usize;
+      let mut _11: bool;
+      scope 1 (inlined core::slice::<impl [u8]>::len) {
+          scope 2 (inlined std::ptr::metadata::<[u8]>) {
+          }
+      }
+  
+      bb0: {
+          StorageLive(_3);
+          nop;
+          StorageLive(_5);
+          _5 = copy _1;
+          nop;
+          StorageLive(_7);
+          _7 = &(*_2);
+          _6 = PtrMetadata(copy _2);
+          StorageDead(_7);
+          _4 = Lt(copy _1, copy _6);
+          switchInt(copy _4) -> [0: bb2, otherwise: bb1];
+      }
+  
+      bb1: {
+          nop;
+          StorageDead(_5);
+          _3 = const ();
+          nop;
+          StorageDead(_3);
+          StorageLive(_9);
+          _9 = copy _1;
+          _10 = copy _6;
+-         _11 = copy _4;
+-         assert(copy _4, "index out of bounds: the length is {} but the index is {}", copy _6, copy _1) -> [success: bb3, unwind unreachable];
++         _11 = const true;
++         assert(const true, "index out of bounds: the length is {} but the index is {}", copy _6, copy _1) -> [success: bb3, unwind unreachable];
+      }
+  
+      bb2: {
+          nop;
+          StorageDead(_5);
+          StorageLive(_8);
+          _8 = panic(const "assertion failed: i < v.len()") -> unwind unreachable;
+      }
+  
+      bb3: {
+          _0 = copy (*_2)[_1];
+          StorageDead(_9);
+          return;
+      }
+  }
+  
+  ALLOC0 (size: 29, align: 1) {
+      0x00 │ 61 73 73 65 72 74 69 6f 6e 20 66 61 69 6c 65 64 │ assertion failed
+      0x10 │ 3a 20 69 20 3c 20 76 2e 6c 65 6e 28 29          │ : i < v.len()
+  }
+  

tests/mir-opt/range/ssa_range.on_assume.SsaRangePropagation.diff

@@ -0,0 +1,56 @@
+- // MIR for `on_assume` before SsaRangePropagation
++ // MIR for `on_assume` after SsaRangePropagation
+  
+  fn on_assume(_1: usize, _2: &[u8]) -> u8 {
+      debug i => _1;
+      debug v => _2;
+      let mut _0: u8;
+      let _3: ();
+      let _4: ();
+      let mut _5: bool;
+      let mut _6: usize;
+      let mut _7: usize;
+      let mut _8: &[u8];
+      let _9: usize;
+      let mut _10: usize;
+      let mut _11: bool;
+      scope 1 (inlined core::slice::<impl [u8]>::len) {
+          scope 2 (inlined std::ptr::metadata::<[u8]>) {
+          }
+      }
+  
+      bb0: {
+          StorageLive(_3);
+          StorageLive(_4);
+          nop;
+          StorageLive(_6);
+          _6 = copy _1;
+          nop;
+          StorageLive(_8);
+          _8 = &(*_2);
+          _7 = PtrMetadata(copy _2);
+          StorageDead(_8);
+          _5 = Lt(copy _1, copy _7);
+          nop;
+          StorageDead(_6);
+          assume(copy _5);
+          nop;
+          StorageDead(_4);
+          _3 = const ();
+          StorageDead(_3);
+          StorageLive(_9);
+          _9 = copy _1;
+          _10 = copy _7;
+-         _11 = copy _5;
+-         assert(copy _5, "index out of bounds: the length is {} but the index is {}", copy _7, copy _1) -> [success: bb1, unwind unreachable];
++         _11 = const true;
++         assert(const true, "index out of bounds: the length is {} but the index is {}", copy _7, copy _1) -> [success: bb1, unwind unreachable];
+      }
+  
+      bb1: {
+          _0 = copy (*_2)[_1];
+          StorageDead(_9);
+          return;
+      }
+  }
+