Auto merge of #142839 - oli-obk:denullarification, r=RalfJung,celinval

Stop backends from needing to support nullary intrinsics

And then remove our infrastructure special casing them. Further improvements can now be done to them by avoiding the intermediate ConstValue step, but let's leave that to follow up work

r? `@RalfJung`

Author: bors

compiler/rustc_codegen_cranelift/example/mini_core.rs

@@ -660,7 +660,7 @@ pub mod intrinsics {
     #[rustc_intrinsic]
     pub unsafe fn ctlz_nonzero<T>(x: T) -> u32;
     #[rustc_intrinsic]
-    pub fn needs_drop<T: ?::Sized>() -> bool;
+    pub const fn needs_drop<T: ?::Sized>() -> bool;
     #[rustc_intrinsic]
     pub fn bitreverse<T>(x: T) -> T;
     #[rustc_intrinsic]

compiler/rustc_codegen_cranelift/example/mini_core_hello_world.rs

@@ -1,4 +1,13 @@
-#![feature(no_core, lang_items, never_type, linkage, extern_types, thread_local, repr_simd)]
+#![feature(
+    no_core,
+    lang_items,
+    never_type,
+    linkage,
+    extern_types,
+    thread_local,
+    repr_simd,
+    rustc_private
+)]
 #![no_core]
 #![allow(dead_code, non_camel_case_types, internal_features)]
 
@@ -207,10 +216,14 @@ fn main() {
         assert_eq!(intrinsics::align_of::<u16>() as u8, 2);
         assert_eq!(intrinsics::align_of_val(&a) as u8, intrinsics::align_of::<&str>() as u8);
 
-        assert!(!intrinsics::needs_drop::<u8>());
-        assert!(!intrinsics::needs_drop::<[u8]>());
-        assert!(intrinsics::needs_drop::<NoisyDrop>());
-        assert!(intrinsics::needs_drop::<NoisyDropUnsized>());
+        let u8_needs_drop = const { intrinsics::needs_drop::<u8>() };
+        assert!(!u8_needs_drop);
+        let slice_needs_drop = const { intrinsics::needs_drop::<[u8]>() };
+        assert!(!slice_needs_drop);
+        let noisy_drop = const { intrinsics::needs_drop::<NoisyDrop>() };
+        assert!(noisy_drop);
+        let noisy_unsized_drop = const { intrinsics::needs_drop::<NoisyDropUnsized>() };
+        assert!(noisy_unsized_drop);
 
         Unique { pointer: NonNull(1 as *mut &str), _marker: PhantomData } as Unique<dyn SomeTrait>;
 

compiler/rustc_codegen_cranelift/src/intrinsics/mod.rs

@@ -812,21 +812,6 @@ fn codegen_regular_intrinsic_call<'tcx>(
             dest.write_cvalue(fx, val);
         }
 
-        sym::needs_drop | sym::type_id | sym::type_name | sym::variant_count => {
-            intrinsic_args!(fx, args => (); intrinsic);
-
-            let const_val = fx
-                .tcx
-                .const_eval_instance(
-                    ty::TypingEnv::fully_monomorphized(),
-                    instance,
-                    source_info.span,
-                )
-                .unwrap();
-            let val = crate::constant::codegen_const_value(fx, const_val, ret.layout().ty);
-            ret.write_cvalue(fx, val);
-        }
-
         sym::ptr_offset_from | sym::ptr_offset_from_unsigned => {
             intrinsic_args!(fx, args => (ptr, base); intrinsic);
             let ptr = ptr.load_scalar(fx);

compiler/rustc_codegen_gcc/example/mini_core_hello_world.rs

@@ -6,6 +6,7 @@
 )]
 #![no_core]
 #![allow(dead_code, internal_features, non_camel_case_types)]
+#![rustfmt::skip]
 
 extern crate mini_core;
 
@@ -197,10 +198,10 @@ fn main() {
         assert_eq!(intrinsics::align_of::<u16>() as u8, 2);
         assert_eq!(intrinsics::align_of_val(&a) as u8, intrinsics::align_of::<&str>() as u8);
 
-        assert!(!intrinsics::needs_drop::<u8>());
-        assert!(!intrinsics::needs_drop::<[u8]>());
-        assert!(intrinsics::needs_drop::<NoisyDrop>());
-        assert!(intrinsics::needs_drop::<NoisyDropUnsized>());
+        assert!(!const { intrinsics::needs_drop::<u8>() });
+        assert!(!const { intrinsics::needs_drop::<[u8]>() });
+        assert!(const { intrinsics::needs_drop::<NoisyDrop>() });
+        assert!(const { intrinsics::needs_drop::<NoisyDropUnsized>() });
 
         Unique {
             pointer: 0 as *const &str,

compiler/rustc_codegen_ssa/src/mir/intrinsic.rs

@@ -1,7 +1,7 @@
 use rustc_abi::WrappingRange;
-use rustc_middle::bug;
 use rustc_middle::mir::SourceInfo;
 use rustc_middle::ty::{self, Ty, TyCtxt};
+use rustc_middle::{bug, span_bug};
 use rustc_session::config::OptLevel;
 use rustc_span::sym;
 
@@ -98,6 +98,27 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
             discr.to_atomic_ordering()
         };
 
+        if args.is_empty() {
+            match name {
+                sym::abort
+                | sym::unreachable
+                | sym::cold_path
+                | sym::breakpoint
+                | sym::assert_zero_valid
+                | sym::assert_mem_uninitialized_valid
+                | sym::assert_inhabited
+                | sym::ub_checks
+                | sym::contract_checks
+                | sym::atomic_fence
+                | sym::atomic_singlethreadfence
+                | sym::caller_location => {}
+                _ => {
+                    span_bug!(span, "nullary intrinsic {name} must either be in a const block or explicitly opted out because it is inherently a runtime intrinsic
+");
+                }
+            }
+        }
+
         let llval = match name {
             sym::abort => {
                 bx.abort();
@@ -150,10 +171,6 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
                 }
                 value
             }
-            sym::needs_drop | sym::type_id | sym::type_name | sym::variant_count => {
-                let value = bx.tcx().const_eval_instance(bx.typing_env(), instance, span).unwrap();
-                OperandRef::from_const(bx, value, result.layout.ty).immediate_or_packed_pair(bx)
-            }
             sym::arith_offset => {
                 let ty = fn_args.type_at(0);
                 let layout = bx.layout_of(ty);

compiler/rustc_const_eval/messages.ftl

@@ -263,9 +263,6 @@ const_eval_non_const_try_block_from_output =
 const_eval_not_enough_caller_args =
     calling a function with fewer arguments than it requires
 
-const_eval_nullary_intrinsic_fail =
-    could not evaluate nullary intrinsic
-
 const_eval_offset_from_different_allocations =
     `{$name}` called on two different pointers that are not both derived from the same allocation
 const_eval_offset_from_out_of_bounds =

compiler/rustc_const_eval/src/const_eval/eval_queries.rs

@@ -20,7 +20,7 @@ use crate::const_eval::CheckAlignment;
 use crate::interpret::{
     CtfeValidationMode, GlobalId, Immediate, InternKind, InternResult, InterpCx, InterpErrorKind,
     InterpResult, MPlaceTy, MemoryKind, OpTy, RefTracking, StackPopCleanup, create_static_alloc,
-    eval_nullary_intrinsic, intern_const_alloc_recursive, interp_ok, throw_exhaust,
+    intern_const_alloc_recursive, interp_ok, throw_exhaust,
 };
 use crate::{CTRL_C_RECEIVED, errors};
 
@@ -280,34 +280,6 @@ pub fn eval_to_const_value_raw_provider<'tcx>(
     tcx: TyCtxt<'tcx>,
     key: ty::PseudoCanonicalInput<'tcx, GlobalId<'tcx>>,
 ) -> ::rustc_middle::mir::interpret::EvalToConstValueResult<'tcx> {
-    // We call `const_eval` for zero arg intrinsics, too, in order to cache their value.
-    // Catch such calls and evaluate them instead of trying to load a constant's MIR.
-    if let ty::InstanceKind::Intrinsic(def_id) = key.value.instance.def {
-        let ty = key.value.instance.ty(tcx, key.typing_env);
-        let ty::FnDef(_, args) = ty.kind() else {
-            bug!("intrinsic with type {:?}", ty);
-        };
-        return eval_nullary_intrinsic(tcx, key.typing_env, def_id, args).report_err().map_err(
-            |error| {
-                let span = tcx.def_span(def_id);
-
-                // FIXME(oli-obk): why don't we have any tests for this code path?
-                super::report(
-                    tcx,
-                    error.into_kind(),
-                    span,
-                    || (span, vec![]),
-                    |diag, span, _| {
-                        diag.span_label(
-                            span,
-                            crate::fluent_generated::const_eval_nullary_intrinsic_fail,
-                        );
-                    },
-                )
-            },
-        );
-    }
-
     tcx.eval_to_allocation_raw(key).map(|val| turn_into_const_value(tcx, val, key))
 }
 

compiler/rustc_const_eval/src/interpret/intrinsics.rs

@@ -6,19 +6,18 @@ use std::assert_matches::assert_matches;
 
 use rustc_abi::Size;
 use rustc_apfloat::ieee::{Double, Half, Quad, Single};
-use rustc_hir::def_id::DefId;
 use rustc_middle::mir::{self, BinOp, ConstValue, NonDivergingIntrinsic};
 use rustc_middle::ty::layout::{TyAndLayout, ValidityRequirement};
-use rustc_middle::ty::{GenericArgsRef, Ty, TyCtxt};
+use rustc_middle::ty::{Ty, TyCtxt};
 use rustc_middle::{bug, ty};
 use rustc_span::{Symbol, sym};
 use tracing::trace;
 
 use super::memory::MemoryKind;
 use super::util::ensure_monomorphic_enough;
 use super::{
-    Allocation, CheckInAllocMsg, ConstAllocation, GlobalId, ImmTy, InterpCx, InterpResult, Machine,
-    OpTy, PlaceTy, Pointer, PointerArithmetic, Provenance, Scalar, err_inval, err_ub_custom,
+    Allocation, CheckInAllocMsg, ConstAllocation, ImmTy, InterpCx, InterpResult, Machine, OpTy,
+    PlaceTy, Pointer, PointerArithmetic, Provenance, Scalar, err_inval, err_ub_custom,
     err_unsup_format, interp_ok, throw_inval, throw_ub_custom, throw_ub_format,
 };
 use crate::fluent_generated as fluent;
@@ -30,73 +29,6 @@ pub(crate) fn alloc_type_name<'tcx>(tcx: TyCtxt<'tcx>, ty: Ty<'tcx>) -> ConstAll
     tcx.mk_const_alloc(alloc)
 }
 
-/// The logic for all nullary intrinsics is implemented here. These intrinsics don't get evaluated
-/// inside an `InterpCx` and instead have their value computed directly from rustc internal info.
-pub(crate) fn eval_nullary_intrinsic<'tcx>(
-    tcx: TyCtxt<'tcx>,
-    typing_env: ty::TypingEnv<'tcx>,
-    def_id: DefId,
-    args: GenericArgsRef<'tcx>,
-) -> InterpResult<'tcx, ConstValue<'tcx>> {
-    let tp_ty = args.type_at(0);
-    let name = tcx.item_name(def_id);
-    interp_ok(match name {
-        sym::type_name => {
-            ensure_monomorphic_enough(tcx, tp_ty)?;
-            let alloc = alloc_type_name(tcx, tp_ty);
-            ConstValue::Slice { data: alloc, meta: alloc.inner().size().bytes() }
-        }
-        sym::needs_drop => {
-            ensure_monomorphic_enough(tcx, tp_ty)?;
-            ConstValue::from_bool(tp_ty.needs_drop(tcx, typing_env))
-        }
-        sym::type_id => {
-            ensure_monomorphic_enough(tcx, tp_ty)?;
-            ConstValue::from_u128(tcx.type_id_hash(tp_ty).as_u128())
-        }
-        sym::variant_count => match match tp_ty.kind() {
-            // Pattern types have the same number of variants as their base type.
-            // Even if we restrict e.g. which variants are valid, the variants are essentially just uninhabited.
-            // And `Result<(), !>` still has two variants according to `variant_count`.
-            ty::Pat(base, _) => *base,
-            _ => tp_ty,
-        }
-        .kind()
-        {
-            // Correctly handles non-monomorphic calls, so there is no need for ensure_monomorphic_enough.
-            ty::Adt(adt, _) => ConstValue::from_target_usize(adt.variants().len() as u64, &tcx),
-            ty::Alias(..) | ty::Param(_) | ty::Placeholder(_) | ty::Infer(_) => {
-                throw_inval!(TooGeneric)
-            }
-            ty::Pat(..) => unreachable!(),
-            ty::Bound(_, _) => bug!("bound ty during ctfe"),
-            ty::Bool
-            | ty::Char
-            | ty::Int(_)
-            | ty::Uint(_)
-            | ty::Float(_)
-            | ty::Foreign(_)
-            | ty::Str
-            | ty::Array(_, _)
-            | ty::Slice(_)
-            | ty::RawPtr(_, _)
-            | ty::Ref(_, _, _)
-            | ty::FnDef(_, _)
-            | ty::FnPtr(..)
-            | ty::Dynamic(_, _, _)
-            | ty::Closure(_, _)
-            | ty::CoroutineClosure(_, _)
-            | ty::Coroutine(_, _)
-            | ty::CoroutineWitness(..)
-            | ty::UnsafeBinder(_)
-            | ty::Never
-            | ty::Tuple(_)
-            | ty::Error(_) => ConstValue::from_target_usize(0u64, &tcx),
-        },
-        other => bug!("`{}` is not a zero arg intrinsic", other),
-    })
-}
-
 impl<'tcx, M: Machine<'tcx>> InterpCx<'tcx, M> {
     /// Returns `true` if emulation happened.
     /// Here we implement the intrinsics that are common to all Miri instances; individual machines can add their own
@@ -110,8 +42,77 @@ impl<'tcx, M: Machine<'tcx>> InterpCx<'tcx, M> {
     ) -> InterpResult<'tcx, bool> {
         let instance_args = instance.args;
         let intrinsic_name = self.tcx.item_name(instance.def_id());
+        let tcx = self.tcx.tcx;
 
         match intrinsic_name {
+            sym::type_name => {
+                let tp_ty = instance.args.type_at(0);
+                ensure_monomorphic_enough(tcx, tp_ty)?;
+                let alloc = alloc_type_name(tcx, tp_ty);
+                let val = ConstValue::Slice { data: alloc, meta: alloc.inner().size().bytes() };
+                let val = self.const_val_to_op(val, dest.layout.ty, Some(dest.layout))?;
+                self.copy_op(&val, dest)?;
+            }
+            sym::needs_drop => {
+                let tp_ty = instance.args.type_at(0);
+                ensure_monomorphic_enough(tcx, tp_ty)?;
+                let val = ConstValue::from_bool(tp_ty.needs_drop(tcx, self.typing_env));
+                let val = self.const_val_to_op(val, tcx.types.bool, Some(dest.layout))?;
+                self.copy_op(&val, dest)?;
+            }
+            sym::type_id => {
+                let tp_ty = instance.args.type_at(0);
+                ensure_monomorphic_enough(tcx, tp_ty)?;
+                let val = ConstValue::from_u128(tcx.type_id_hash(tp_ty).as_u128());
+                let val = self.const_val_to_op(val, dest.layout.ty, Some(dest.layout))?;
+                self.copy_op(&val, dest)?;
+            }
+            sym::variant_count => {
+                let tp_ty = instance.args.type_at(0);
+                let ty = match tp_ty.kind() {
+                    // Pattern types have the same number of variants as their base type.
+                    // Even if we restrict e.g. which variants are valid, the variants are essentially just uninhabited.
+                    // And `Result<(), !>` still has two variants according to `variant_count`.
+                    ty::Pat(base, _) => *base,
+                    _ => tp_ty,
+                };
+                let val = match ty.kind() {
+                    // Correctly handles non-monomorphic calls, so there is no need for ensure_monomorphic_enough.
+                    ty::Adt(adt, _) => {
+                        ConstValue::from_target_usize(adt.variants().len() as u64, &tcx)
+                    }
+                    ty::Alias(..) | ty::Param(_) | ty::Placeholder(_) | ty::Infer(_) => {
+                        throw_inval!(TooGeneric)
+                    }
+                    ty::Pat(..) => unreachable!(),
+                    ty::Bound(_, _) => bug!("bound ty during ctfe"),
+                    ty::Bool
+                    | ty::Char
+                    | ty::Int(_)
+                    | ty::Uint(_)
+                    | ty::Float(_)
+                    | ty::Foreign(_)
+                    | ty::Str
+                    | ty::Array(_, _)
+                    | ty::Slice(_)
+                    | ty::RawPtr(_, _)
+                    | ty::Ref(_, _, _)
+                    | ty::FnDef(_, _)
+                    | ty::FnPtr(..)
+                    | ty::Dynamic(_, _, _)
+                    | ty::Closure(_, _)
+                    | ty::CoroutineClosure(_, _)
+                    | ty::Coroutine(_, _)
+                    | ty::CoroutineWitness(..)
+                    | ty::UnsafeBinder(_)
+                    | ty::Never
+                    | ty::Tuple(_)
+                    | ty::Error(_) => ConstValue::from_target_usize(0u64, &tcx),
+                };
+                let val = self.const_val_to_op(val, dest.layout.ty, Some(dest.layout))?;
+                self.copy_op(&val, dest)?;
+            }
+
             sym::caller_location => {
                 let span = self.find_closest_untracked_caller_location();
                 let val = self.tcx.span_as_caller_location(span);
@@ -137,21 +138,6 @@ impl<'tcx, M: Machine<'tcx>> InterpCx<'tcx, M> {
                 self.write_scalar(Scalar::from_target_usize(result, self), dest)?;
             }
 
-            sym::needs_drop | sym::type_id | sym::type_name | sym::variant_count => {
-                let gid = GlobalId { instance, promoted: None };
-                let ty = self
-                    .tcx
-                    .fn_sig(instance.def_id())
-                    .instantiate(self.tcx.tcx, instance.args)
-                    .output()
-                    .no_bound_vars()
-                    .unwrap();
-                let val = self
-                    .ctfe_query(|tcx| tcx.const_eval_global_id(self.typing_env, gid, tcx.span))?;
-                let val = self.const_val_to_op(val, ty, Some(dest.layout))?;
-                self.copy_op(&val, dest)?;
-            }
-
             sym::fadd_algebraic
             | sym::fsub_algebraic
             | sym::fmul_algebraic

compiler/rustc_const_eval/src/interpret/mod.rs

@@ -29,7 +29,6 @@ pub use self::intern::{
     HasStaticRootDefId, InternKind, InternResult, intern_const_alloc_for_constprop,
     intern_const_alloc_recursive,
 };
-pub(crate) use self::intrinsics::eval_nullary_intrinsic;
 pub use self::machine::{AllocMap, Machine, MayLeak, ReturnAction, compile_time_machine};
 pub use self::memory::{AllocInfo, AllocKind, AllocRef, AllocRefMut, FnVal, Memory, MemoryKind};
 use self::operand::Operand;