Auto merge of rust-lang#135610 - scottmcm:assume-less, r=<try>

Emit fewer `assume`s now that we have `range` metadata on parameters

We still need the `assume` for the *target* type's range, but we no longer need it for the *source* type's range.

The first stab at this regressed a test, but thanks to good advice in llvm/llvm-project#123278 (comment) the second commit here changes how we emit these range assertions to the form that LLVM apparently likes better (and, conveniently, is easier to emit too) which got that test passing again 🎉

Hopefully this means less crud for LLVM to churn through in `opt` builds...

compiler/rustc_codegen_ssa/src/mir/rvalue.rs

@@ -257,13 +257,11 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
                 if let OperandValueKind::Immediate(to_scalar) = cast_kind
                     && from_scalar.size(self.cx) == to_scalar.size(self.cx)
                 {
-                    let from_backend_ty = bx.backend_type(operand.layout);
                     let to_backend_ty = bx.backend_type(cast);
                     Some(OperandValue::Immediate(self.transmute_immediate(
                         bx,
                         imm,
                         from_scalar,
-                        from_backend_ty,
                         to_scalar,
                         to_backend_ty,
                     )))
@@ -279,13 +277,11 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
                     && in_a.size(self.cx) == out_a.size(self.cx)
                     && in_b.size(self.cx) == out_b.size(self.cx)
                 {
-                    let in_a_ibty = bx.scalar_pair_element_backend_type(operand.layout, 0, false);
-                    let in_b_ibty = bx.scalar_pair_element_backend_type(operand.layout, 1, false);
                     let out_a_ibty = bx.scalar_pair_element_backend_type(cast, 0, false);
                     let out_b_ibty = bx.scalar_pair_element_backend_type(cast, 1, false);
                     Some(OperandValue::Pair(
-                        self.transmute_immediate(bx, imm_a, in_a, in_a_ibty, out_a, out_a_ibty),
-                        self.transmute_immediate(bx, imm_b, in_b, in_b_ibty, out_b, out_b_ibty),
+                        self.transmute_immediate(bx, imm_a, in_a, out_a, out_a_ibty),
+                        self.transmute_immediate(bx, imm_b, in_b, out_b, out_b_ibty),
                     ))
                 } else {
                     None
@@ -309,12 +305,6 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
     ) -> Option<Bx::Value> {
         use abi::Primitive::*;
 
-        // When scalars are passed by value, there's no metadata recording their
-        // 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.
-        self.assume_scalar_range(bx, imm, from_scalar, from_backend_ty);
-
         imm = match (from_scalar.primitive(), to_scalar.primitive()) {
             (Int(_, is_signed), Int(..)) => bx.intcast(imm, to_backend_ty, is_signed),
             (Float(_), Float(_)) => {
@@ -356,7 +346,6 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
         bx: &mut Bx,
         mut imm: Bx::Value,
         from_scalar: abi::Scalar,
-        from_backend_ty: Bx::Type,
         to_scalar: abi::Scalar,
         to_backend_ty: Bx::Type,
     ) -> Bx::Value {
@@ -365,11 +354,13 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
         use abi::Primitive::*;
         imm = bx.from_immediate(imm);
 
-        // When scalars are passed by value, there's no metadata recording their
-        // 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.
-        self.assume_scalar_range(bx, imm, from_scalar, from_backend_ty);
+        // We used to `assume` the `from_scalar` here too, but that's no longer needed
+        // because if we have a scalar, we must already know its range. Either
+        // 1) It's a parameter with `range` parameter metadata,
+        // 2) It's something we `load`ed with `!range` metadata, or
+        // 3) It's something we transmuted and already `assume`d the range.
+        // And thus in all those cases another `assume` is just wasteful.
+        // (Case 1 didn't used to be covered, and thus the `assume` was needed.)
 
         imm = match (from_scalar.primitive(), to_scalar.primitive()) {
             (Int(..) | Float(_), Int(..) | Float(_)) => bx.bitcast(imm, to_backend_ty),
@@ -389,18 +380,19 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
                 bx.bitcast(int_imm, to_backend_ty)
             }
         };
-        self.assume_scalar_range(bx, imm, to_scalar, to_backend_ty);
+
+        // This `assume` remains important for cases like
+        //    transmute::<u32, NonZeroU32>(x) == 0
+        // 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.
+        self.assume_scalar_range(bx, imm, to_scalar);
+
         imm = bx.to_immediate_scalar(imm, to_scalar);
         imm
     }
 
-    fn assume_scalar_range(
-        &self,
-        bx: &mut Bx,
-        imm: Bx::Value,
-        scalar: abi::Scalar,
-        backend_ty: Bx::Type,
-    ) {
+    fn assume_scalar_range(&self, bx: &mut Bx, imm: Bx::Value, scalar: abi::Scalar) {
         if matches!(self.cx.sess().opts.optimize, OptLevel::No)
             // For now, the critical niches are all over `Int`eger values.
             // Should floating-point values or pointers ever get more complex
@@ -411,31 +403,8 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
             return;
         }
 
-        let abi::WrappingRange { start, end } = scalar.valid_range(self.cx);
-
-        if start <= end {
-            if start > 0 {
-                let low = bx.const_uint_big(backend_ty, start);
-                let cmp = bx.icmp(IntPredicate::IntUGE, imm, low);
-                bx.assume(cmp);
-            }
-
-            let type_max = scalar.size(self.cx).unsigned_int_max();
-            if end < type_max {
-                let high = bx.const_uint_big(backend_ty, end);
-                let cmp = bx.icmp(IntPredicate::IntULE, imm, high);
-                bx.assume(cmp);
-            }
-        } else {
-            let low = bx.const_uint_big(backend_ty, start);
-            let cmp_low = bx.icmp(IntPredicate::IntUGE, imm, low);
-
-            let high = bx.const_uint_big(backend_ty, end);
-            let cmp_high = bx.icmp(IntPredicate::IntULE, imm, high);
-
-            let or = bx.or(cmp_low, cmp_high);
-            bx.assume(or);
-        }
+        let range = scalar.valid_range(self.cx);
+        bx.assume_integer_range(imm, range);
     }
 
     pub(crate) fn codegen_rvalue_unsized(

compiler/rustc_codegen_ssa/src/traits/builder.rs

@@ -217,6 +217,28 @@ pub trait BuilderMethods<'a, 'tcx>:
         dest: PlaceRef<'tcx, Self::Value>,
     );
 
+    /// Emits an `assume` that the integer value `imm` is contained in `range`.
+    ///
+    /// This *always* emits the assumption, so you probably want to check the
+    /// optimization level and `Scalar::is_always_valid` before calling it.
+    fn assume_integer_range(&mut self, imm: Self::Value, range: WrappingRange) {
+        let WrappingRange { start, end } = range;
+        let backend_ty = self.cx().val_ty(imm);
+
+        // Perhaps one day we'll be able to use assume operand bundles for this,
+        // but for now this encoding with a single icmp+assume is best per
+        // </~https://github.com/llvm/llvm-project/issues/123278#issuecomment-2597440158>
+        let shifted = if start == 0 {
+            imm
+        } else {
+            let low = self.const_uint_big(backend_ty, start);
+            self.sub(imm, low)
+        };
+        let width = self.const_uint_big(backend_ty, u128::wrapping_sub(end, start));
+        let cmp = self.icmp(IntPredicate::IntULE, shifted, width);
+        self.assume(cmp);
+    }
+
     fn range_metadata(&mut self, load: Self::Value, range: WrappingRange);
     fn nonnull_metadata(&mut self, load: Self::Value);
 

tests/codegen/intrinsics/transmute-niched.rs

@@ -1,5 +1,6 @@
 //@ revisions: OPT DBG
 //@ [OPT] compile-flags: -C opt-level=3 -C no-prepopulate-passes
+//@ [OPT] min-llvm-version: 19 (for range parameter metadata)
 //@ [DBG] compile-flags: -C opt-level=0 -C no-prepopulate-passes
 //@ only-64bit (so I don't need to worry about usize)
 #![crate_type = "lib"]
@@ -17,26 +18,24 @@ pub enum SmallEnum {
 // CHECK-LABEL: @check_to_enum(
 #[no_mangle]
 pub unsafe fn check_to_enum(x: i8) -> SmallEnum {
-    // OPT: %0 = icmp uge i8 %x, 10
-    // OPT: call void @llvm.assume(i1 %0)
-    // OPT: %1 = icmp ule i8 %x, 12
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // OPT: %0 = sub i8 %x, 10
+    // OPT: %1 = icmp ule i8 %0, 2
     // OPT: call void @llvm.assume(i1 %1)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: ret i8 %x
 
     transmute(x)
 }
 
 // CHECK-LABEL: @check_from_enum(
+// OPT-SAME: range(i8 10, 13){{.+}}%x
 #[no_mangle]
 pub unsafe fn check_from_enum(x: SmallEnum) -> i8 {
-    // OPT: %0 = icmp uge i8 %x, 10
-    // OPT: call void @llvm.assume(i1 %0)
-    // OPT: %1 = icmp ule i8 %x, 12
-    // OPT: call void @llvm.assume(i1 %1)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: ret i8 %x
 
     transmute(x)
@@ -45,26 +44,24 @@ pub unsafe fn check_from_enum(x: SmallEnum) -> i8 {
 // CHECK-LABEL: @check_to_ordering(
 #[no_mangle]
 pub unsafe fn check_to_ordering(x: u8) -> std::cmp::Ordering {
-    // OPT: %0 = icmp uge i8 %x, -1
-    // OPT: %1 = icmp ule i8 %x, 1
-    // OPT: %2 = or i1 %0, %1
-    // OPT: call void @llvm.assume(i1 %2)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // 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)
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: ret i8 %x
 
     transmute(x)
 }
 
 // CHECK-LABEL: @check_from_ordering(
+// OPT-SAME: range(i8 -1, 2){{.+}}%x
 #[no_mangle]
 pub unsafe fn check_from_ordering(x: std::cmp::Ordering) -> u8 {
-    // OPT: %0 = icmp uge i8 %x, -1
-    // OPT: %1 = icmp ule i8 %x, 1
-    // OPT: %2 = or i1 %0, %1
-    // OPT: call void @llvm.assume(i1 %2)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: ret i8 %x
 
     transmute(x)
@@ -96,50 +93,50 @@ pub enum Minus100ToPlus100 {
 }
 
 // CHECK-LABEL: @check_enum_from_char(
+// OPT-SAME: range(i32 0, 1114112){{.+}}%x
 #[no_mangle]
 pub unsafe fn check_enum_from_char(x: char) -> Minus100ToPlus100 {
-    // OPT: %0 = icmp ule i32 %x, 1114111
-    // OPT: call void @llvm.assume(i1 %0)
-    // OPT: %1 = icmp uge i32 %x, -100
-    // OPT: %2 = icmp ule i32 %x, 100
-    // OPT: %3 = or i1 %1, %2
-    // OPT: call void @llvm.assume(i1 %3)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // OPT: %0 = sub i32 %x, -100
+    // OPT: %1 = icmp ule i32 %0, 200
+    // OPT: call void @llvm.assume(i1 %1)
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: ret i32 %x
 
     transmute(x)
 }
 
 // CHECK-LABEL: @check_enum_to_char(
+// OPT-SAME: range(i32 -100, 101){{.+}}%x
 #[no_mangle]
 pub unsafe fn check_enum_to_char(x: Minus100ToPlus100) -> char {
-    // OPT: %0 = icmp uge i32 %x, -100
-    // OPT: %1 = icmp ule i32 %x, 100
-    // OPT: %2 = or i1 %0, %1
-    // OPT: call void @llvm.assume(i1 %2)
-    // OPT: %3 = icmp ule i32 %x, 1114111
-    // OPT: call void @llvm.assume(i1 %3)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // OPT: %0 = icmp ule i32 %x, 1114111
+    // OPT: call void @llvm.assume(i1 %0)
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: ret i32 %x
 
     transmute(x)
 }
 
 // CHECK-LABEL: @check_swap_pair(
+// OPT-SAME: range(i32 0, 1114112){{.+}}%x.0
+// OPT-SAME: range(i32 1, 0){{.+}}%x.1
 #[no_mangle]
 pub unsafe fn check_swap_pair(x: (char, NonZero<u32>)) -> (NonZero<u32>, char) {
-    // OPT: %0 = icmp ule i32 %x.0, 1114111
-    // OPT: call void @llvm.assume(i1 %0)
-    // OPT: %1 = icmp uge i32 %x.0, 1
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // OPT: %0 = sub i32 %x.0, 1
+    // OPT: %1 = icmp ule i32 %0, -2
     // OPT: call void @llvm.assume(i1 %1)
-    // OPT: %2 = icmp uge i32 %x.1, 1
+    // OPT: %2 = icmp ule i32 %x.1, 1114111
     // OPT: call void @llvm.assume(i1 %2)
-    // OPT: %3 = icmp ule i32 %x.1, 1114111
-    // OPT: call void @llvm.assume(i1 %3)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: %[[P1:.+]] = insertvalue { i32, i32 } poison, i32 %x.0, 0
     // CHECK: %[[P2:.+]] = insertvalue { i32, i32 } %[[P1]], i32 %x.1, 1
     // CHECK: ret { i32, i32 } %[[P2]]
@@ -148,34 +145,33 @@ pub unsafe fn check_swap_pair(x: (char, NonZero<u32>)) -> (NonZero<u32>, char) {
 }
 
 // CHECK-LABEL: @check_bool_from_ordering(
+// OPT-SAME: range(i8 -1, 2){{.+}}%x
 #[no_mangle]
 pub unsafe fn check_bool_from_ordering(x: std::cmp::Ordering) -> bool {
-    // OPT: %0 = icmp uge i8 %x, -1
-    // OPT: %1 = icmp ule i8 %x, 1
-    // OPT: %2 = or i1 %0, %1
-    // OPT: call void @llvm.assume(i1 %2)
-    // OPT: %3 = icmp ule i8 %x, 1
-    // OPT: call void @llvm.assume(i1 %3)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // OPT: %0 = icmp ule i8 %x, 1
+    // OPT: call void @llvm.assume(i1 %0)
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: %[[R:.+]] = trunc i8 %x to i1
     // CHECK: ret i1 %[[R]]
 
     transmute(x)
 }
 
 // CHECK-LABEL: @check_bool_to_ordering(
+// OPT-SAME: i1 {{.+}} %x
 #[no_mangle]
 pub unsafe fn check_bool_to_ordering(x: bool) -> std::cmp::Ordering {
     // CHECK: %_0 = zext i1 %x to i8
-    // OPT: %0 = icmp ule i8 %_0, 1
-    // OPT: call void @llvm.assume(i1 %0)
-    // OPT: %1 = icmp uge i8 %_0, -1
-    // OPT: %2 = icmp ule i8 %_0, 1
-    // OPT: %3 = or i1 %1, %2
-    // OPT: call void @llvm.assume(i1 %3)
-    // DBG-NOT: icmp
-    // DBG-NOT: assume
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
+    // OPT: %0 = sub i8 %_0, -1
+    // OPT: %1 = icmp ule i8 %0, 2
+    // OPT: call void @llvm.assume(i1 %1)
+    // CHECK-NOT: icmp
+    // CHECK-NOT: assume
     // CHECK: ret i8 %_0
 
     transmute(x)

tests/codegen/issues/issue-119422.rs

@@ -2,6 +2,7 @@
 //! for `NonZero` integer types.
 //!
 //@ compile-flags: -O --edition=2021 -Zmerge-functions=disabled
+//@ min-llvm-version: 19 (for range parameter metadata)
 //@ only-64bit (because the LLVM type of i64 for usize shows up)
 #![crate_type = "lib"]