When the required discriminator value exceeds LLVM's limits, drop the debug info for the function instead of panicking.

khuey · khuey · commit f5b023bd9cf4 · 2024-11-19T05:19:09.000-08:00
The maximum discriminator value LLVM can currently encode is 2^12. If macro use
results in more than 2^12 calls to the same function attributed to the same
callsite, and those calls are MIR-inlined, we will require more than the maximum
discriminator value to completely represent the debug information. Once we reach
that point drop the debug info instead.
diff --git a/compiler/rustc_codegen_gcc/src/debuginfo.rs b/compiler/rustc_codegen_gcc/src/debuginfo.rs
@@ -113,15 +113,15 @@ fn make_mir_scope<'gcc, 'tcx>(
     let scope_data = &mir.source_scopes[scope];
     let parent_scope = if let Some(parent) = scope_data.parent_scope {
         make_mir_scope(cx, _instance, mir, variables, debug_context, instantiated, parent);
-        debug_context.scopes[parent]
+        debug_context.scopes[parent].unwrap()
     } else {
         // The root is the function itself.
         let file = cx.sess().source_map().lookup_source_file(mir.span.lo());
-        debug_context.scopes[scope] = DebugScope {
+        debug_context.scopes[scope] = Some(DebugScope {
             file_start_pos: file.start_pos,
             file_end_pos: file.end_position(),
-            ..debug_context.scopes[scope]
-        };
+            ..debug_context.scopes[scope].unwrap()
+        });
         instantiated.insert(scope);
         return;
     };
@@ -130,7 +130,7 @@ fn make_mir_scope<'gcc, 'tcx>(
         if !vars.contains(scope) && scope_data.inlined.is_none() {
             // Do not create a DIScope if there are no variables defined in this
             // MIR `SourceScope`, and it's not `inlined`, to avoid debuginfo bloat.
-            debug_context.scopes[scope] = parent_scope;
+            debug_context.scopes[scope] = Some(parent_scope);
             instantiated.insert(scope);
             return;
         }
@@ -157,12 +157,12 @@ fn make_mir_scope<'gcc, 'tcx>(
     // TODO(tempdragon): dbg_scope: Add support for scope extension here.
     inlined_at.or(p_inlined_at);
 
-    debug_context.scopes[scope] = DebugScope {
+    debug_context.scopes[scope] = Some(DebugScope {
         dbg_scope,
         inlined_at,
         file_start_pos: loc.file.start_pos,
         file_end_pos: loc.file.end_position(),
-    };
+    });
     instantiated.insert(scope);
 }
 
@@ -232,12 +232,12 @@ impl<'gcc, 'tcx> DebugInfoCodegenMethods<'tcx> for CodegenCx<'gcc, 'tcx> {
         }
 
         // Initialize fn debug context (including scopes).
-        let empty_scope = DebugScope {
+        let empty_scope = Some(DebugScope {
             dbg_scope: self.dbg_scope_fn(instance, fn_abi, Some(llfn)),
             inlined_at: None,
             file_start_pos: BytePos(0),
             file_end_pos: BytePos(0),
-        };
+        });
         let mut fn_debug_context = FunctionDebugContext {
             scopes: IndexVec::from_elem(empty_scope, mir.source_scopes.as_slice()),
             inlined_function_scopes: Default::default(),
diff --git a/compiler/rustc_codegen_llvm/src/debuginfo/create_scope_map.rs b/compiler/rustc_codegen_llvm/src/debuginfo/create_scope_map.rs
@@ -85,15 +85,23 @@ fn make_mir_scope<'ll, 'tcx>(
             discriminators,
             parent,
         );
-        debug_context.scopes[parent]
+        if let Some(parent_scope) = debug_context.scopes[parent] {
+            parent_scope
+        } else {
+            // If the parent scope could not be represented then no children
+            // can be either.
+            debug_context.scopes[scope] = None;
+            instantiated.insert(scope);
+            return;
+        }
     } else {
         // The root is the function itself.
         let file = cx.sess().source_map().lookup_source_file(mir.span.lo());
-        debug_context.scopes[scope] = DebugScope {
+        debug_context.scopes[scope] = Some(DebugScope {
             file_start_pos: file.start_pos,
             file_end_pos: file.end_position(),
-            ..debug_context.scopes[scope]
-        };
+            ..debug_context.scopes[scope].unwrap()
+        });
         instantiated.insert(scope);
         return;
     };
@@ -104,7 +112,7 @@ fn make_mir_scope<'ll, 'tcx>(
     {
         // Do not create a DIScope if there are no variables defined in this
         // MIR `SourceScope`, and it's not `inlined`, to avoid debuginfo bloat.
-        debug_context.scopes[scope] = parent_scope;
+        debug_context.scopes[scope] = Some(parent_scope);
         instantiated.insert(scope);
         return;
     }
@@ -137,21 +145,28 @@ fn make_mir_scope<'ll, 'tcx>(
         },
     };
 
-    let inlined_at = scope_data.inlined.map(|(_, callsite_span)| {
+    let mut debug_scope = Some(DebugScope {
+        dbg_scope,
+        inlined_at: parent_scope.inlined_at,
+        file_start_pos: loc.file.start_pos,
+        file_end_pos: loc.file.end_position(),
+    });
+
+    if let Some((_, callsite_span)) = scope_data.inlined {
         let callsite_span = hygiene::walk_chain_collapsed(callsite_span, mir.span);
         let callsite_scope = parent_scope.adjust_dbg_scope_for_span(cx, callsite_span);
         let loc = cx.dbg_loc(callsite_scope, parent_scope.inlined_at, callsite_span);
 
         // NB: In order to produce proper debug info for variables (particularly
-        // arguments) in multiply-inline functions, LLVM expects to see a single
+        // arguments) in multiply-inlined functions, LLVM expects to see a single
         // DILocalVariable with multiple different DILocations in the IR. While
         // the source information for each DILocation would be identical, their
         // inlinedAt attributes will be unique to the particular callsite.
         //
         // We generate DILocations here based on the callsite's location in the
         // source code. A single location in the source code usually can't
         // produce multiple distinct calls so this mostly works, until
-        // proc-macros get involved. A proc-macro can generate multiple calls
+        // macros get involved. A macro can generate multiple calls
         // at the same span, which breaks the assumption that we're going to
         // produce a unique DILocation for every scope we process here. We
         // have to explicitly add discriminators if we see inlines into the
@@ -160,24 +175,29 @@ fn make_mir_scope<'ll, 'tcx>(
         // Note further that we can't key this hashtable on the span itself,
         // because these spans could have distinct SyntaxContexts. We have
         // to key on exactly what we're giving to LLVM.
-        match discriminators.entry(callsite_span.lo()) {
+        let inlined_at = match discriminators.entry(callsite_span.lo()) {
             Entry::Occupied(mut o) => {
                 *o.get_mut() += 1;
                 unsafe { llvm::LLVMRustDILocationCloneWithBaseDiscriminator(loc, *o.get()) }
-                    .expect("Failed to encode discriminator in DILocation")
             }
             Entry::Vacant(v) => {
                 v.insert(0);
-                loc
+                Some(loc)
+            }
+        };
+        match inlined_at {
+            Some(inlined_at) => {
+                debug_scope.as_mut().unwrap().inlined_at = Some(inlined_at);
+            }
+            None => {
+                // LLVM has a maximum discriminator that it can encode (currently
+                // it uses 12 bits for 4096 possible values). If we exceed that
+                // there is little we can do but drop the debug info.
+                debug_scope = None;
             }
         }
-    });
+    }
 
-    debug_context.scopes[scope] = DebugScope {
-        dbg_scope,
-        inlined_at: inlined_at.or(parent_scope.inlined_at),
-        file_start_pos: loc.file.start_pos,
-        file_end_pos: loc.file.end_position(),
-    };
+    debug_context.scopes[scope] = debug_scope;
     instantiated.insert(scope);
 }
diff --git a/compiler/rustc_codegen_llvm/src/debuginfo/mod.rs b/compiler/rustc_codegen_llvm/src/debuginfo/mod.rs
@@ -294,12 +294,12 @@ impl<'ll, 'tcx> DebugInfoCodegenMethods<'tcx> for CodegenCx<'ll, 'tcx> {
         }
 
         // Initialize fn debug context (including scopes).
-        let empty_scope = DebugScope {
+        let empty_scope = Some(DebugScope {
             dbg_scope: self.dbg_scope_fn(instance, fn_abi, Some(llfn)),
             inlined_at: None,
             file_start_pos: BytePos(0),
             file_end_pos: BytePos(0),
-        };
+        });
         let mut fn_debug_context = FunctionDebugContext {
             scopes: IndexVec::from_elem(empty_scope, &mir.source_scopes),
             inlined_function_scopes: Default::default(),
diff --git a/compiler/rustc_codegen_ssa/src/mir/debuginfo.rs b/compiler/rustc_codegen_ssa/src/mir/debuginfo.rs
@@ -20,7 +20,9 @@ use crate::traits::*;
 
 pub struct FunctionDebugContext<'tcx, S, L> {
     /// Maps from source code to the corresponding debug info scope.
-    pub scopes: IndexVec<mir::SourceScope, DebugScope<S, L>>,
+    /// May be None if the backend is not capable of representing the scope for
+    /// some reason.
+    pub scopes: IndexVec<mir::SourceScope, Option<DebugScope<S, L>>>,
 
     /// Maps from an inlined function to its debug info declaration.
     pub inlined_function_scopes: FxHashMap<Instance<'tcx>, S>,
@@ -231,7 +233,7 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
         &self,
         source_info: mir::SourceInfo,
     ) -> Option<(Bx::DIScope, Option<Bx::DILocation>, Span)> {
-        let scope = &self.debug_context.as_ref()?.scopes[source_info.scope];
+        let scope = &self.debug_context.as_ref()?.scopes[source_info.scope]?;
         let span = hygiene::walk_chain_collapsed(source_info.span, self.mir.span);
         Some((scope.adjust_dbg_scope_for_span(self.cx, span), scope.inlined_at, span))
     }
diff --git a/tests/ui/debuginfo/debuginfo-inline-callsite-location-macro-2.rs b/tests/ui/debuginfo/debuginfo-inline-callsite-location-macro-2.rs
