Don't emit two assumes in transmutes when one is a subset of the other

For example, transmuting between `bool` and `Ordering` doesn't need two `assume`s because one range is a superset of the other.

Multiple are still used for things like `char` <-> `NonZero<u32>`, which overlap but where neither fully contains the other.

Author: scottmcm

compiler/rustc_abi/src/lib.rs

@@ -1376,6 +1376,28 @@ impl WrappingRange {
         }
     }
 
+    /// Returns `true` if all the values in `other` are contained in this range,
+    /// when the values are considered as having width `size`.
+    #[inline(always)]
+    pub fn contains_range(&self, other: Self, size: Size) -> bool {
+        if self.is_full_for(size) {
+            true
+        } else {
+            let trunc = |x| size.truncate(x);
+
+            let delta = self.start;
+            let max = trunc(self.end.wrapping_sub(delta));
+
+            let other_start = trunc(other.start.wrapping_sub(delta));
+            let other_end = trunc(other.end.wrapping_sub(delta));
+
+            // Having shifted both input ranges by `delta`, now we only need to check
+            // whether `0..=max` contains `other_start..=other_end`, which can only
+            // happen if the other doesn't wrap since `self` isn't everything.
+            (other_start <= other_end) && (other_end <= max)
+        }
+    }
+
     /// Returns `self` with replaced `start`
     #[inline(always)]
     fn with_start(mut self, start: u128) -> Self {

compiler/rustc_abi/src/tests.rs

@@ -5,3 +5,66 @@ fn align_constants() {
     assert_eq!(Align::ONE, Align::from_bytes(1).unwrap());
     assert_eq!(Align::EIGHT, Align::from_bytes(8).unwrap());
 }
+
+#[test]
+fn wrapping_range_contains_range() {
+    let size16 = Size::from_bytes(16);
+
+    let a = WrappingRange { start: 10, end: 20 };
+    assert!(a.contains_range(a, size16));
+    assert!(a.contains_range(WrappingRange { start: 11, end: 19 }, size16));
+    assert!(a.contains_range(WrappingRange { start: 10, end: 10 }, size16));
+    assert!(a.contains_range(WrappingRange { start: 20, end: 20 }, size16));
+    assert!(!a.contains_range(WrappingRange { start: 10, end: 21 }, size16));
+    assert!(!a.contains_range(WrappingRange { start: 9, end: 20 }, size16));
+    assert!(!a.contains_range(WrappingRange { start: 4, end: 6 }, size16));
+    assert!(!a.contains_range(WrappingRange { start: 24, end: 26 }, size16));
+
+    assert!(!a.contains_range(WrappingRange { start: 16, end: 14 }, size16));
+
+    let b = WrappingRange { start: 20, end: 10 };
+    assert!(b.contains_range(b, size16));
+    assert!(b.contains_range(WrappingRange { start: 20, end: 20 }, size16));
+    assert!(b.contains_range(WrappingRange { start: 10, end: 10 }, size16));
+    assert!(b.contains_range(WrappingRange { start: 0, end: 10 }, size16));
+    assert!(b.contains_range(WrappingRange { start: 20, end: 30 }, size16));
+    assert!(b.contains_range(WrappingRange { start: 20, end: 9 }, size16));
+    assert!(b.contains_range(WrappingRange { start: 21, end: 10 }, size16));
+    assert!(b.contains_range(WrappingRange { start: 999, end: 9999 }, size16));
+    assert!(b.contains_range(WrappingRange { start: 999, end: 9 }, size16));
+    assert!(!b.contains_range(WrappingRange { start: 19, end: 19 }, size16));
+    assert!(!b.contains_range(WrappingRange { start: 11, end: 11 }, size16));
+    assert!(!b.contains_range(WrappingRange { start: 19, end: 11 }, size16));
+    assert!(!b.contains_range(WrappingRange { start: 11, end: 19 }, size16));
+
+    let f = WrappingRange { start: 0, end: u128::MAX };
+    assert!(f.contains_range(WrappingRange { start: 10, end: 20 }, size16));
+    assert!(f.contains_range(WrappingRange { start: 20, end: 10 }, size16));
+
+    let g = WrappingRange { start: 2, end: 1 };
+    assert!(g.contains_range(WrappingRange { start: 10, end: 20 }, size16));
+    assert!(g.contains_range(WrappingRange { start: 20, end: 10 }, size16));
+
+    let size1 = Size::from_bytes(1);
+    let u8r = WrappingRange { start: 0, end: 255 };
+    let i8r = WrappingRange { start: 128, end: 127 };
+    assert!(u8r.contains_range(i8r, size1));
+    assert!(i8r.contains_range(u8r, size1));
+    assert!(!u8r.contains_range(i8r, size16));
+    assert!(i8r.contains_range(u8r, size16));
+
+    let boolr = WrappingRange { start: 0, end: 1 };
+    assert!(u8r.contains_range(boolr, size1));
+    assert!(i8r.contains_range(boolr, size1));
+    assert!(!boolr.contains_range(u8r, size1));
+    assert!(!boolr.contains_range(i8r, size1));
+
+    let cmpr = WrappingRange { start: 255, end: 1 };
+    assert!(u8r.contains_range(cmpr, size1));
+    assert!(i8r.contains_range(cmpr, size1));
+    assert!(!cmpr.contains_range(u8r, size1));
+    assert!(!cmpr.contains_range(i8r, size1));
+
+    assert!(!boolr.contains_range(cmpr, size1));
+    assert!(cmpr.contains_range(boolr, size1));
+}

compiler/rustc_codegen_ssa/src/mir/rvalue.rs

@@ -288,7 +288,7 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
         // valid ranges. For example, `char`s are passed as just `i32`, with no
         // way for LLVM to know that they're 0x10FFFF at most. Thus we assume
         // the range of the input value too, not just the output range.
-        assume_scalar_range(bx, imm, from_scalar, from_backend_ty);
+        assume_scalar_range(bx, imm, from_scalar, from_backend_ty, None);
 
         imm = match (from_scalar.primitive(), to_scalar.primitive()) {
             (Int(_, is_signed), Int(..)) => bx.intcast(imm, to_backend_ty, is_signed),
@@ -1064,7 +1064,7 @@ pub(super) fn transmute_scalar<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>>(
     // That said, last time we tried removing this, it didn't actually help
     // the rustc-perf results, so might as well keep doing it
     // <https://github.com/rust-lang/rust/pull/135610#issuecomment-2599275182>
-    assume_scalar_range(bx, imm, from_scalar, from_backend_ty);
+    assume_scalar_range(bx, imm, from_scalar, from_backend_ty, Some(&to_scalar));
 
     imm = match (from_scalar.primitive(), to_scalar.primitive()) {
         (Int(..) | Float(_), Int(..) | Float(_)) => bx.bitcast(imm, to_backend_ty),
@@ -1092,22 +1092,42 @@ pub(super) fn transmute_scalar<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>>(
     // since it's never passed to something with parameter metadata (especially
     // after MIR inlining) so the only way to tell the backend about the
     // constraint that the `transmute` introduced is to `assume` it.
-    assume_scalar_range(bx, imm, to_scalar, to_backend_ty);
+    assume_scalar_range(bx, imm, to_scalar, to_backend_ty, Some(&from_scalar));
 
     imm = bx.to_immediate_scalar(imm, to_scalar);
     imm
 }
 
+/// Emits an `assume` call that `imm`'s value is within the known range of `scalar`.
+///
+/// If `known` is `Some`, only emits the assume if it's more specific than
+/// whatever is already known from the range of *that* scalar.
 fn assume_scalar_range<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>>(
     bx: &mut Bx,
     imm: Bx::Value,
     scalar: abi::Scalar,
     backend_ty: Bx::Type,
+    known: Option<&abi::Scalar>,
 ) {
-    if matches!(bx.cx().sess().opts.optimize, OptLevel::No) || scalar.is_always_valid(bx.cx()) {
+    if matches!(bx.cx().sess().opts.optimize, OptLevel::No) {
         return;
     }
 
+    match (scalar, known) {
+        (abi::Scalar::Union { .. }, _) => return,
+        (_, None) => {
+            if scalar.is_always_valid(bx.cx()) {
+                return;
+            }
+        }
+        (abi::Scalar::Initialized { valid_range, .. }, Some(known)) => {
+            let known_range = known.valid_range(bx.cx());
+            if valid_range.contains_range(known_range, scalar.size(bx.cx())) {
+                return;
+            }
+        }
+    }
+
     match scalar.primitive() {
         abi::Primitive::Int(..) => {
             let range = scalar.valid_range(bx.cx());

tests/codegen-llvm/intrinsics/transmute-niched.rs

@@ -163,11 +163,8 @@ pub unsafe fn check_swap_pair(x: (char, NonZero<u32>)) -> (NonZero<u32>, char) {
 pub unsafe fn check_bool_from_ordering(x: std::cmp::Ordering) -> bool {
     // CHECK-NOT: icmp
     // CHECK-NOT: assume
-    // OPT: %0 = sub i8 %x, -1
-    // OPT: %1 = icmp ule i8 %0, 2
-    // OPT: call void @llvm.assume(i1 %1)
-    // OPT: %2 = icmp ule i8 %x, 1
-    // OPT: call void @llvm.assume(i1 %2)
+    // OPT: %0 = icmp ule i8 %x, 1
+    // OPT: call void @llvm.assume(i1 %0)
     // CHECK-NOT: icmp
     // CHECK-NOT: assume
     // CHECK: %[[R:.+]] = trunc{{( nuw)?}} i8 %x to i1
@@ -184,9 +181,6 @@ pub unsafe fn check_bool_to_ordering(x: bool) -> std::cmp::Ordering {
     // CHECK-NOT: assume
     // OPT: %0 = icmp ule i8 %_0, 1
     // OPT: call void @llvm.assume(i1 %0)
-    // OPT: %1 = sub i8 %_0, -1
-    // OPT: %2 = icmp ule i8 %1, 2
-    // OPT: call void @llvm.assume(i1 %2)
     // CHECK-NOT: icmp
     // CHECK-NOT: assume
     // CHECK: ret i8 %_0
@@ -221,3 +215,42 @@ pub unsafe fn check_ptr_to_nonnull(x: *const u8) -> NonNull<u8> {
 
     transmute(x)
 }
+
+#[repr(usize)]
+pub enum FourOrEight {
+    Four = 4,
+    Eight = 8,
+}
+
+// CHECK-LABEL: @check_nonnull_to_four_or_eight(
+#[no_mangle]
+pub unsafe fn check_nonnull_to_four_or_eight(x: NonNull<u8>) -> FourOrEight {
+    // CHECK: start
+    // CHECK-NEXT: %[[RET:.+]] = ptrtoint ptr %x to i64
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // OPT: %0 = sub i64 %[[RET]], 4
+    // OPT: %1 = icmp ule i64 %0, 4
+    // OPT: call void @llvm.assume(i1 %1)
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // CHECK: ret i64 %[[RET]]
+
+    transmute(x)
+}
+
+// CHECK-LABEL: @check_four_or_eight_to_nonnull(
+#[no_mangle]
+pub unsafe fn check_four_or_eight_to_nonnull(x: FourOrEight) -> NonNull<u8> {
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // OPT: %0 = sub i64 %x, 4
+    // OPT: %1 = icmp ule i64 %0, 4
+    // OPT: call void @llvm.assume(i1 %1)
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // CHECK: %[[RET:.+]] = getelementptr i8, ptr null, i64 %x
+    // CHECK-NEXT: ret ptr %[[RET]]
+
+    transmute(x)
+}