Recreate DefIds when a cached query gets replayed

Author: oli-obk

compiler/rustc_middle/src/query/mod.rs

@@ -1718,6 +1718,7 @@ rustc_queries! {
         desc { |tcx| "computing visibility of `{}`", tcx.def_path_str(def_id) }
         separate_provide_extern
         feedable
+        cache_on_disk_if { def_id.is_local() }
     }
 
     query inhabited_predicate_adt(key: DefId) -> ty::inhabitedness::InhabitedPredicate<'tcx> {

compiler/rustc_middle/src/ty/context.rs

@@ -55,8 +55,9 @@ use rustc_hir::lang_items::LangItem;
 use rustc_hir::{HirId, Node, TraitCandidate};
 use rustc_index::IndexVec;
 use rustc_macros::HashStable;
-use rustc_query_system::dep_graph::DepNodeIndex;
+use rustc_query_system::dep_graph::{DepNodeIndex, TaskDepsRef};
 use rustc_query_system::ich::StableHashingContext;
+use rustc_query_system::query::DefIdInfo;
 use rustc_serialize::opaque::{FileEncodeResult, FileEncoder};
 use rustc_session::config::CrateType;
 use rustc_session::cstore::{CrateStoreDyn, Untracked};
@@ -1087,9 +1088,28 @@ impl<'tcx> TyCtxt<'tcx> {
 
         // This function modifies `self.definitions` using a side-effect.
         // We need to ensure that these side effects are re-run by the incr. comp. engine.
-        // Depending on the forever-red node will tell the graph that the calling query
-        // needs to be re-evaluated.
-        self.dep_graph.read_index(DepNodeIndex::FOREVER_RED_NODE);
+        tls::with_context_opt(|icx| {
+            let icx = icx.unwrap();
+            let Some(side_effects) = &icx.side_effects else {
+                assert!(matches!(icx.task_deps, TaskDepsRef::Ignore), "{:?}", icx.task_deps);
+                return;
+            };
+            let info = DefIdInfo { parent, data };
+            match icx.task_deps {
+                TaskDepsRef::Allow(_)
+                | TaskDepsRef::EvalAlways
+                | TaskDepsRef::Ignore
+                | TaskDepsRef::Forbid => {
+                    side_effects.lock().definitions.push(DefIdInfo { parent, data });
+                }
+                TaskDepsRef::Replay { prev_side_effects, created_def_ids } => {
+                    let prev_info = &prev_side_effects.definitions
+                        [created_def_ids.load(std::sync::atomic::Ordering::Relaxed)];
+                    created_def_ids.fetch_add(1, std::sync::atomic::Ordering::Relaxed);
+                    assert_eq!(*prev_info, info);
+                }
+            }
+        });
 
         let feed = self.feed_local_def_id(def_id);
         feed.def_kind(def_kind);

compiler/rustc_query_impl/src/plumbing.rs

@@ -21,8 +21,8 @@ use rustc_middle::ty::{self, print::with_no_queries, TyCtxt};
 use rustc_query_system::dep_graph::{DepNodeParams, HasDepContext};
 use rustc_query_system::ich::StableHashingContext;
 use rustc_query_system::query::{
-    force_query, QueryCache, QueryConfig, QueryContext, QueryJobId, QueryMap, QuerySideEffects,
-    QueryStackFrame,
+    force_query, DefIdInfo, QueryCache, QueryConfig, QueryContext, QueryJobId, QueryMap,
+    QuerySideEffects, QueryStackFrame,
 };
 use rustc_query_system::{LayoutOfDepth, QueryOverflow};
 use rustc_serialize::Decodable;
@@ -176,11 +176,15 @@ impl QueryContext for QueryCtxt<'_> {
     #[tracing::instrument(level = "trace", skip(self))]
     fn apply_side_effects(self, side_effects: QuerySideEffects) {
         let dcx = self.dep_context().sess().dcx();
-        let QuerySideEffects { diagnostics } = side_effects;
+        let QuerySideEffects { diagnostics, definitions } = side_effects;
 
         for diagnostic in diagnostics {
             dcx.emit_diagnostic(diagnostic);
         }
+
+        for DefIdInfo { parent, data } in definitions {
+            self.tcx.untracked().definitions.write().create_def(parent, data);
+        }
     }
 }
 

compiler/rustc_query_system/src/dep_graph/graph.rs

@@ -1,10 +1,16 @@
+use super::query::DepGraphQuery;
+use super::serialized::{GraphEncoder, SerializedDepGraph, SerializedDepNodeIndex};
+use super::{DepContext, DepKind, DepNode, Deps, HasDepContext, WorkProductId};
+use crate::dep_graph::edges::EdgesVec;
+use crate::ich::StableHashingContext;
+use crate::query::{QueryContext, QuerySideEffects};
 use rustc_data_structures::fingerprint::Fingerprint;
 use rustc_data_structures::fx::{FxHashMap, FxHashSet};
 use rustc_data_structures::profiling::{EventId, QueryInvocationId, SelfProfilerRef};
 use rustc_data_structures::sharded::{self, Sharded};
 use rustc_data_structures::stable_hasher::{HashStable, StableHasher};
 use rustc_data_structures::steal::Steal;
-use rustc_data_structures::sync::{AtomicU32, AtomicU64, Lock, Lrc};
+use rustc_data_structures::sync::{AtomicU32, AtomicU64, AtomicUsize, Lock, Lrc};
 use rustc_data_structures::unord::UnordMap;
 use rustc_index::IndexVec;
 use rustc_serialize::opaque::{FileEncodeResult, FileEncoder};
@@ -15,13 +21,6 @@ use std::hash::Hash;
 use std::marker::PhantomData;
 use std::sync::atomic::Ordering;
 
-use super::query::DepGraphQuery;
-use super::serialized::{GraphEncoder, SerializedDepGraph, SerializedDepNodeIndex};
-use super::{DepContext, DepKind, DepNode, Deps, HasDepContext, WorkProductId};
-use crate::dep_graph::edges::EdgesVec;
-use crate::ich::StableHashingContext;
-use crate::query::{QueryContext, QuerySideEffects};
-
 #[cfg(debug_assertions)]
 use {super::debug::EdgeFilter, std::env};
 
@@ -218,6 +217,15 @@ impl<D: Deps> DepGraph<D> {
         D::with_deps(TaskDepsRef::Ignore, op)
     }
 
+    pub(crate) fn with_replay<R>(
+        &self,
+        prev_side_effects: &QuerySideEffects,
+        created_def_ids: &AtomicUsize,
+        op: impl FnOnce() -> R,
+    ) -> R {
+        D::with_deps(TaskDepsRef::Replay { prev_side_effects, created_def_ids }, op)
+    }
+
     /// Used to wrap the deserialization of a query result from disk,
     /// This method enforces that no new `DepNodes` are created during
     /// query result deserialization.
@@ -272,6 +280,7 @@ impl<D: Deps> DepGraph<D> {
     }
 
     #[inline(always)]
+    /// A helper for `codegen_cranelift`.
     pub fn with_task<Ctxt: HasDepContext<Deps = D>, A: Debug, R>(
         &self,
         key: DepNode,
@@ -468,6 +477,12 @@ impl<D: Deps> DepGraph<D> {
                         return;
                     }
                     TaskDepsRef::Ignore => return,
+                    // We don't need to record dependencies when rerunning a query
+                    // because we have no disk cache entry to load. The dependencies
+                    // are preserved.
+                    // FIXME: assert that the dependencies don't change instead of
+                    // recording them.
+                    TaskDepsRef::Replay { .. } => return,
                     TaskDepsRef::Forbid => {
                         panic!("Illegal read of: {dep_node_index:?}")
                     }
@@ -573,6 +588,7 @@ impl<D: Deps> DepGraph<D> {
                     edges.push(DepNodeIndex::FOREVER_RED_NODE);
                 }
                 TaskDepsRef::Ignore => {}
+                TaskDepsRef::Replay { .. } => {}
                 TaskDepsRef::Forbid => {
                     panic!("Cannot summarize when dependencies are not recorded.")
                 }
@@ -1323,6 +1339,16 @@ pub enum TaskDepsRef<'a> {
     /// to ensure that the decoding process doesn't itself
     /// require the execution of any queries.
     Forbid,
+    /// Side effects from the previous run made available to
+    /// queries when they are reexecuted because their result was not
+    /// available in the cache. The query removes entries from the
+    /// side effect table. The table must be empty
+    Replay {
+        prev_side_effects: &'a QuerySideEffects,
+        /// Every new `DefId` is pushed here so we can check
+        /// that they match the cached ones.
+        created_def_ids: &'a AtomicUsize,
+    },
 }
 
 #[derive(Debug)]

compiler/rustc_query_system/src/query/mod.rs

@@ -23,7 +23,7 @@ use rustc_data_structures::stable_hasher::Hash64;
 use rustc_data_structures::sync::Lock;
 use rustc_errors::Diagnostic;
 use rustc_hir::def::DefKind;
-use rustc_span::def_id::DefId;
+use rustc_span::def_id::{DefId, LocalDefId};
 use rustc_span::Span;
 use thin_vec::ThinVec;
 
@@ -90,20 +90,30 @@ pub struct QuerySideEffects {
     /// These diagnostics will be re-emitted if we mark
     /// the query as green.
     pub diagnostics: ThinVec<Diagnostic>,
+    /// Stores any `DefId`s that were created during query execution.
+    /// These `DefId`s will be re-created when we mark the query as green.
+    pub definitions: ThinVec<DefIdInfo>,
+}
+
+#[derive(Debug, Clone, Encodable, Decodable, PartialEq)]
+pub struct DefIdInfo {
+    pub parent: LocalDefId,
+    pub data: rustc_hir::definitions::DefPathData,
 }
 
 impl QuerySideEffects {
     /// Returns true if there might be side effects.
     #[inline]
     pub fn maybe_any(&self) -> bool {
-        let QuerySideEffects { diagnostics } = self;
+        let QuerySideEffects { diagnostics, definitions } = self;
         // Use `has_capacity` so that the destructor for `self.diagnostics` can be skipped
         // if `maybe_any` is known to be false.
-        diagnostics.has_capacity()
+        diagnostics.has_capacity() || definitions.has_capacity()
     }
     pub fn append(&mut self, other: QuerySideEffects) {
-        let QuerySideEffects { diagnostics } = self;
+        let QuerySideEffects { diagnostics, definitions } = self;
         diagnostics.extend(other.diagnostics);
+        definitions.extend(other.definitions);
     }
 }
 

compiler/rustc_query_system/src/query/plumbing.rs

@@ -16,7 +16,7 @@ use rustc_data_structures::fingerprint::Fingerprint;
 use rustc_data_structures::fx::FxHashMap;
 use rustc_data_structures::sharded::Sharded;
 use rustc_data_structures::stack::ensure_sufficient_stack;
-use rustc_data_structures::sync::Lock;
+use rustc_data_structures::sync::{AtomicUsize, Lock};
 #[cfg(parallel_compiler)]
 use rustc_data_structures::{outline, sync};
 use rustc_errors::{DiagnosticBuilder, FatalError, StashKey};
@@ -502,7 +502,7 @@ where
         let dep_node =
             dep_node_opt.get_or_insert_with(|| query.construct_dep_node(*qcx.dep_context(), &key));
 
-        // The diagnostics for this query will be promoted to the current session during
+        // The side_effects for this query will be promoted to the current session during
         // `try_mark_green()`, so we can ignore them here.
         if let Some(ret) = qcx.start_query(job_id, false, None, || {
             try_load_from_disk_and_cache_in_memory(query, dep_graph_data, qcx, &key, dep_node)
@@ -620,8 +620,17 @@ where
     // recompute.
     let prof_timer = qcx.dep_context().profiler().query_provider();
 
+    let prev_side_effects = qcx.load_side_effects(prev_dep_node_index);
+    let created_def_ids = AtomicUsize::new(0);
     // The dep-graph for this computation is already in-place.
-    let result = qcx.dep_context().dep_graph().with_ignore(|| query.compute(qcx, *key));
+    let result =
+        qcx.dep_context()
+            .dep_graph()
+            .with_replay(&prev_side_effects, &created_def_ids, || query.compute(qcx, *key));
+
+    assert_eq!(created_def_ids.into_inner(), prev_side_effects.definitions.len());
+    // We already checked at `DefId` creation time, that the created `DefId`s have the same parent and `DefPathData`
+    // as the cached ones.
 
     prof_timer.finish_with_query_invocation_id(dep_node_index.into());