gccrs: Add dead code scan on HIR

In order to find dead code we use a depth first search and keep liveness
variables, after type resolution. In this case, if a function is unused
and it calls another function the 2nd function is now unused since the
caller is not used etc. The algorithm is a depth first search.

	gcc/rust/
	* checks/lints/rust-lint-marklive-base.h: New.
	* checks/lints/rust-lint-marklive.cc: New.
	* checks/lints/rust-lint-marklive.h: New.
	* checks/lints/rust-lint-scan-deadcode.h: New.
This commit is contained in:
Thomas Young
2022-08-23 16:34:35 +01:00
committed by Arthur Cohen
parent ca246e573f
commit 520b52b24e
4 changed files with 789 additions and 0 deletions

View File

@@ -0,0 +1,45 @@
// Copyright (C) 2021-2022 Free Software Foundation, Inc.
// This file is part of GCC.
// GCC is free software; you can redistribute it and/or modify it under
// the terms of the GNU General Public License as published by the Free
// Software Foundation; either version 3, or (at your option) any later
// version.
// GCC is distributed in the hope that it will be useful, but WITHOUT ANY
// WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
// for more details.
// You should have received a copy of the GNU General Public License
// along with GCC; see the file COPYING3. If not see
// <http://www.gnu.org/licenses/>.
#ifndef RUST_HIR_LIVENESS_BASE
#define RUST_HIR_LIVENESS_BASE
#include "rust-diagnostics.h"
#include "rust-lint-marklive.h"
#include "rust-lint-marklive-base.h"
#include "rust-hir-visitor.h"
#include "rust-hir-map.h"
namespace Rust {
namespace Analysis {
class MarkLiveBase : public HIR::HIRFullVisitorBase
{
public:
virtual ~MarkLiveBase () {}
protected:
MarkLiveBase () : mappings (Analysis::Mappings::get ()) {}
Analysis::Mappings *mappings;
};
} // namespace Analysis
} // namespace Rust
#endif

View File

@@ -0,0 +1,282 @@
// Copyright (C) 2021-2022 Free Software Foundation, Inc.
// This file is part of GCC.
// GCC is free software; you can redistribute it and/or modify it under
// the terms of the GNU General Public License as published by the Free
// Software Foundation; either version 3, or (at your option) any later
// version.
// GCC is distributed in the hope that it will be useful, but WITHOUT ANY
// WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
// for more details.
// You should have received a copy of the GNU General Public License
// along with GCC; see the file COPYING3. If not see
// <http://www.gnu.org/licenses/>.
// The idea is that all reachable symbols are live, codes called
// from live codes are live, and everything else is dead.
#include "rust-lint-marklive.h"
#include "rust-hir-full.h"
#include "rust-name-resolver.h"
namespace Rust {
namespace Analysis {
// This class trys to find the live symbols which can be used as
// seeds in MarkLive
//
// 1. TODO: explicit live
// - Attribute like #[allow(dead_code)]
// - Attribute like #[lang=".."], it's not a intra-crate item.
// 2. TODO: foreign item
class FindEntryPoint : public MarkLiveBase
{
using Rust::Analysis::MarkLiveBase::visit;
public:
static std::vector<HirId> find (HIR::Crate &crate)
{
FindEntryPoint findEntryPoint;
for (auto it = crate.items.begin (); it != crate.items.end (); it++)
{
it->get ()->accept_vis (findEntryPoint);
}
return findEntryPoint.getEntryPoint ();
}
// TODO not only fn main can be a entry point.
void visit (HIR::Function &function) override
{
if (function.get_function_name () == "main")
{
entryPoints.push_back (function.get_mappings ().get_hirid ());
}
}
private:
FindEntryPoint () : MarkLiveBase () {}
std::vector<HirId> entryPoints;
std::vector<HirId> getEntryPoint () { return entryPoints; }
};
std::set<HirId>
MarkLive::Analysis (HIR::Crate &crate)
{
MarkLive marklive (FindEntryPoint::find (crate));
marklive.go (crate);
return marklive.liveSymbols;
}
// pop a live symbol from worklist every iteration,
// if it's a function then walk the function body, and
// 1. save all the live symbols in worklist which is
// visited first time
// 2. save all the live symbols in liveSymbols
void
MarkLive::go (HIR::Crate &crate)
{
while (!worklist.empty ())
{
HirId hirId = worklist.back ();
worklist.pop_back ();
scannedSymbols.emplace (hirId);
HIR::Item *item = mappings->lookup_hir_item (hirId);
liveSymbols.emplace (hirId);
if (item != nullptr)
{
item->accept_vis (*this);
}
else
{ // the item maybe inside a trait impl
HirId parent_impl_id = UNKNOWN_HIRID;
HIR::ImplItem *implItem
= mappings->lookup_hir_implitem (hirId, &parent_impl_id);
if (implItem != nullptr)
implItem->accept_vis (*this);
}
}
}
void
MarkLive::visit (HIR::PathInExpression &expr)
{
// We should iterate every path segment in order to mark the struct which
// is used in expression like Foo::bar(), we should mark the Foo alive.
expr.iterate_path_segments ([&] (HIR::PathExprSegment &seg) -> bool {
return visit_path_segment (seg);
});
// after iterate the path segments, we should mark functions and associated
// functions alive.
NodeId ast_node_id = expr.get_mappings ().get_nodeid ();
NodeId ref_node_id = UNKNOWN_NODEID;
find_ref_node_id (ast_node_id, ref_node_id);
// node back to HIR
HirId ref;
bool ok = mappings->lookup_node_to_hir (ref_node_id, &ref);
rust_assert (ok);
// it must resolve to some kind of HIR::Item or HIR::InheritImplItem
HIR::Item *resolved_item = mappings->lookup_hir_item (ref);
if (resolved_item != nullptr)
{
mark_hir_id (resolved_item->get_mappings ().get_hirid ());
}
else
{
HirId parent_impl_id = UNKNOWN_HIRID;
HIR::ImplItem *resolved_item
= mappings->lookup_hir_implitem (ref, &parent_impl_id);
if (resolved_item != nullptr)
{
mark_hir_id (resolved_item->get_impl_mappings ().get_hirid ());
}
}
}
void
MarkLive::visit (HIR::MethodCallExpr &expr)
{
expr.get_receiver ()->accept_vis (*this);
visit_path_segment (expr.get_method_name ());
for (auto &argument : expr.get_arguments ())
argument->accept_vis (*this);
// Trying to find the method definition and mark it alive.
NodeId ast_node_id = expr.get_mappings ().get_nodeid ();
NodeId ref_node_id = UNKNOWN_NODEID;
find_ref_node_id (ast_node_id, ref_node_id);
// node back to HIR
HirId ref;
bool ok = mappings->lookup_node_to_hir (ref_node_id, &ref);
rust_assert (ok);
mark_hir_id (ref);
}
bool
MarkLive::visit_path_segment (HIR::PathExprSegment seg)
{
NodeId ast_node_id = seg.get_mappings ().get_nodeid ();
NodeId ref_node_id = UNKNOWN_NODEID;
// There are two different kinds of segment for us.
// 1. function segment
// like the symbol "foo" in expression `foo()`.
// 2. type segment
// like the symbol "Foo" in expression `Foo{a: 1, b: 2}`
//
// We should mark them alive all and ignoring other kind of segments.
// If the segment we dont care then just return false is fine
if (!resolver->lookup_resolved_name (ast_node_id, &ref_node_id))
{
if (!resolver->lookup_resolved_type (ast_node_id, &ref_node_id))
return false;
}
HirId ref;
bool ok = mappings->lookup_node_to_hir (ref_node_id, &ref);
rust_assert (ok);
mark_hir_id (ref);
return true;
}
void
MarkLive::visit (HIR::FieldAccessExpr &expr)
{
// visit receiver at first
expr.get_receiver_expr ()->accept_vis (*this);
// resolve the receiver back to ADT type
TyTy::BaseType *receiver = nullptr;
if (!tyctx->lookup_type (
expr.get_receiver_expr ()->get_mappings ().get_hirid (), &receiver))
{
rust_error_at (expr.get_receiver_expr ()->get_locus (),
"unresolved type for receiver");
}
TyTy::ADTType *adt = nullptr;
if (receiver->get_kind () == TyTy::TypeKind::ADT)
{
adt = static_cast<TyTy::ADTType *> (receiver);
}
else if (receiver->get_kind () == TyTy::TypeKind::REF)
{
TyTy::ReferenceType *r = static_cast<TyTy::ReferenceType *> (receiver);
TyTy::BaseType *b = r->get_base ();
rust_assert (b->get_kind () == TyTy::TypeKind::ADT);
adt = static_cast<TyTy::ADTType *> (b);
}
rust_assert (adt != nullptr);
rust_assert (!adt->is_enum ());
rust_assert (adt->number_of_variants () == 1);
TyTy::VariantDef *variant = adt->get_variants ().at (0);
// get the field index
size_t index;
TyTy::StructFieldType *field;
bool ok = variant->lookup_field (expr.get_field_name (), &field, &index);
rust_assert (ok);
if (index >= variant->num_fields ())
{
rust_error_at (expr.get_receiver_expr ()->get_locus (),
"cannot access struct %s by index: %lu",
adt->get_name ().c_str (), (unsigned long) index);
return;
}
// get the field hir id
HirId field_id = field->get_ref ();
mark_hir_id (field_id);
}
void
MarkLive::visit (HIR::TupleIndexExpr &expr)
{
// TODO: unused tuple field detection
expr.get_tuple_expr ()->accept_vis (*this);
}
void
MarkLive::visit (HIR::TypeAlias &alias)
{
NodeId ast_node_id;
resolver->lookup_resolved_type (
alias.get_type_aliased ()->get_mappings ().get_nodeid (), &ast_node_id);
HirId hir_id;
bool ok = mappings->lookup_node_to_hir (ast_node_id, &hir_id);
rust_assert (ok);
mark_hir_id (hir_id);
}
void
MarkLive::mark_hir_id (HirId id)
{
if (scannedSymbols.find (id) == scannedSymbols.end ())
{
worklist.push_back (id);
}
liveSymbols.emplace (id);
}
void
MarkLive::find_ref_node_id (NodeId ast_node_id, NodeId &ref_node_id)
{
if (!resolver->lookup_resolved_name (ast_node_id, &ref_node_id))
{
bool ok = resolver->lookup_resolved_type (ast_node_id, &ref_node_id);
rust_assert (ok);
}
}
} // namespace Analysis
} // namespace Rust

View File

@@ -0,0 +1,308 @@
// Copyright (C) 2021-2022 Free Software Foundation, Inc.
// This file is part of GCC.
// GCC is free software; you can redistribute it and/or modify it under
// the terms of the GNU General Public License as published by the Free
// Software Foundation; either version 3, or (at your option) any later
// version.
// GCC is distributed in the hope that it will be useful, but WITHOUT ANY
// WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
// for more details.
// You should have received a copy of the GNU General Public License
// along with GCC; see the file COPYING3. If not see
// <http://www.gnu.org/licenses/>.
#ifndef RUST_HIR_LIVENESS
#define RUST_HIR_LIVENESS
#include "rust-hir-full-decls.h"
#include "rust-hir-map.h"
#include "rust-lint-marklive-base.h"
#include "rust-name-resolver.h"
namespace Rust {
namespace Analysis {
class MarkLive : public MarkLiveBase
{
using Rust::Analysis::MarkLiveBase::visit;
public:
static std::set<HirId> Analysis (HIR::Crate &crate);
void go (HIR::Crate &crate);
void visit (HIR::PathInExpression &expr) override;
void visit (HIR::FieldAccessExpr &expr) override;
void visit (HIR::TupleIndexExpr &expr) override;
void visit (HIR::MethodCallExpr &expr) override;
void visit (HIR::TypeAlias &alias) override;
void visit (HIR::BorrowExpr &expr) override
{
expr.get_expr ()->accept_vis (*this);
}
void visit (HIR::DereferenceExpr &expr) override
{
expr.get_expr ()->accept_vis (*this);
}
void visit (HIR::NegationExpr &expr) override
{
expr.get_expr ()->accept_vis (*this);
}
void visit (HIR::LazyBooleanExpr &expr) override
{
expr.get_lhs ()->accept_vis (*this);
expr.get_rhs ()->accept_vis (*this);
}
void visit (HIR::TypeCastExpr &expr) override
{
expr.get_expr ()->accept_vis (*this);
}
void visit (HIR::GroupedExpr &expr) override
{
expr.get_expr_in_parens ()->accept_vis (*this);
}
void visit (HIR::ArrayExpr &expr) override
{
expr.get_internal_elements ()->accept_vis (*this);
}
void visit (HIR::ArrayIndexExpr &expr) override
{
expr.get_array_expr ()->accept_vis (*this);
expr.get_index_expr ()->accept_vis (*this);
}
void visit (HIR::ArrayElemsValues &expr) override
{
for (auto &elem : expr.get_values ())
{
elem->accept_vis (*this);
}
}
void visit (HIR::TupleExpr &expr) override
{
for (auto &elem : expr.get_tuple_elems ())
{
elem->accept_vis (*this);
}
}
void visit (HIR::BlockExpr &expr) override
{
for (auto &s : expr.get_statements ())
{
s->accept_vis (*this);
}
if (expr.has_expr ())
{
expr.get_final_expr ()->accept_vis (*this);
}
}
void visit (HIR::UnsafeBlockExpr &expr) override
{
expr.get_block_expr ()->accept_vis (*this);
}
void visit (HIR::LoopExpr &expr) override
{
expr.get_loop_block ()->accept_vis (*this);
}
void visit (HIR::BreakExpr &expr) override
{
if (expr.has_break_expr ())
expr.get_expr ()->accept_vis (*this);
}
void visit (HIR::WhileLoopExpr &expr) override
{
expr.get_loop_block ()->accept_vis (*this);
expr.get_predicate_expr ()->accept_vis (*this);
}
void visit (HIR::Function &function) override
{
function.get_definition ()->accept_vis (*this);
}
void visit (HIR::ReturnExpr &expr) override
{
if (expr.has_return_expr ())
expr.get_expr ()->accept_vis (*this);
}
void visit (HIR::WhileLetLoopExpr &expr) override
{
expr.get_loop_block ()->accept_vis (*this);
expr.get_cond ()->accept_vis (*this);
}
void visit (HIR::ForLoopExpr &expr) override
{
expr.get_loop_block ()->accept_vis (*this);
expr.get_iterator_expr ()->accept_vis (*this);
}
void visit (HIR::ExprStmtWithoutBlock &stmt) override
{
stmt.get_expr ()->accept_vis (*this);
}
void visit (HIR::ExprStmtWithBlock &stmt) override
{
stmt.get_expr ()->accept_vis (*this);
}
void visit (HIR::CallExpr &expr) override
{
expr.get_fnexpr ()->accept_vis (*this);
for (auto &argument : expr.get_arguments ())
argument->accept_vis (*this);
}
void visit (HIR::ArithmeticOrLogicalExpr &expr) override
{
expr.visit_lhs (*this);
expr.visit_rhs (*this);
}
void visit (HIR::ComparisonExpr &expr) override
{
expr.get_lhs ()->accept_vis (*this);
expr.get_rhs ()->accept_vis (*this);
}
void visit (HIR::AssignmentExpr &expr) override
{
expr.visit_lhs (*this);
expr.visit_rhs (*this);
}
void visit (HIR::CompoundAssignmentExpr &expr) override
{
expr.visit_lhs (*this);
expr.visit_rhs (*this);
}
void visit (HIR::IfExpr &expr) override
{
expr.get_if_condition ()->accept_vis (*this);
expr.get_if_block ()->accept_vis (*this);
}
void visit (HIR::IfExprConseqElse &expr) override
{
expr.get_if_condition ()->accept_vis (*this);
expr.get_if_block ()->accept_vis (*this);
expr.get_else_block ()->accept_vis (*this);
}
void visit (HIR::MatchExpr &expr) override
{
expr.get_scrutinee_expr ()->accept_vis (*this);
std::vector<HIR::MatchCase> &cases = expr.get_match_cases ();
for (auto &&caz : cases)
{
auto case_arm = caz.get_arm ();
if (case_arm.has_match_arm_guard ())
case_arm.get_guard_expr ()->accept_vis (*this);
caz.get_expr ()->accept_vis (*this);
}
}
void visit (HIR::IfExprConseqIf &expr) override
{
expr.get_if_condition ()->accept_vis (*this);
expr.get_if_block ()->accept_vis (*this);
expr.get_conseq_if_expr ()->accept_vis (*this);
}
void visit (HIR::TraitItemFunc &item) override
{
item.get_block_expr ()->accept_vis (*this);
}
void visit (HIR::ImplBlock &impl) override
{
for (auto &&item : impl.get_impl_items ())
{
item->accept_vis (*this);
}
}
void visit (HIR::LetStmt &stmt) override
{
if (stmt.has_init_expr ())
{
stmt.get_init_expr ()->accept_vis (*this);
}
}
void visit (HIR::StructExprStruct &stct) override
{
stct.get_struct_name ().accept_vis (*this);
}
void visit (HIR::StructExprStructFields &stct) override
{
for (auto &field : stct.get_fields ())
{
field->accept_vis (*this);
}
stct.get_struct_name ().accept_vis (*this);
if (stct.has_struct_base ())
{
stct.struct_base->base_struct->accept_vis (*this);
}
}
virtual void visit (HIR::StructExprFieldIdentifierValue &field) override
{
field.get_value ()->accept_vis (*this);
}
void visit (HIR::StructExprStructBase &stct) override
{
stct.get_struct_base ()->base_struct->accept_vis (*this);
}
void visit (HIR::Module &module) override
{
for (auto &item : module.get_items ())
item->accept_vis (*this);
}
private:
std::vector<HirId> worklist;
std::set<HirId> liveSymbols;
std::set<HirId> scannedSymbols;
Analysis::Mappings *mappings;
Resolver::Resolver *resolver;
Resolver::TypeCheckContext *tyctx;
MarkLive (std::vector<HirId> worklist)
: worklist (worklist), mappings (Analysis::Mappings::get ()),
resolver (Resolver::Resolver::get ()),
tyctx (Resolver::TypeCheckContext::get ()){};
void mark_hir_id (HirId);
bool visit_path_segment (HIR::PathExprSegment);
void find_ref_node_id (NodeId ast_node_id, NodeId &ref_node_id);
};
} // namespace Analysis
} // namespace Rust
#endif

View File

@@ -0,0 +1,154 @@
// Copyright (C) 2021-2022 Free Software Foundation, Inc.
// This file is part of GCC.
// GCC is free software; you can redistribute it and/or modify it under
// the terms of the GNU General Public License as published by the Free
// Software Foundation; either version 3, or (at your option) any later
// version.
// GCC is distributed in the hope that it will be useful, but WITHOUT ANY
// WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
// for more details.
// You should have received a copy of the GNU General Public License
// along with GCC; see the file COPYING3. If not see
// <http://www.gnu.org/licenses/>.
#ifndef RUST_HIR_SCAN_DEADCODE
#define RUST_HIR_SCAN_DEADCODE
#include "rust-hir-full-decls.h"
#include "rust-hir-map.h"
#include "rust-lint-marklive.h"
#include "rust-name-resolver.h"
#include "rust-diagnostics.h"
namespace Rust {
namespace Analysis {
// Scan item symbols and warn the symbol if it is not in the live_symbols set.
// There are three kinds of item we should handle in this pass.
// 1. Function item
// 2. The function item in the impl block without trait
// 3. StructStruct, e.g., `Struct Foo{one: 1, two: 2}`. Furthermore, the unused
// struct fields will be warned too.
// 4. TupleStruct, e.g., `Struct Foo(i32, i32)`
class ScanDeadcode : public MarkLiveBase
{
using Rust::Analysis::MarkLiveBase::visit;
public:
static void Scan (HIR::Crate &crate)
{
std::set<HirId> live_symbols = Analysis::MarkLive::Analysis (crate);
ScanDeadcode sdc (live_symbols);
for (auto it = crate.items.begin (); it != crate.items.end (); it++)
{
it->get ()->accept_vis (sdc);
}
};
void visit (HIR::Function &function) override
{
HirId hirId = function.get_mappings ().get_hirid ();
if (should_warn (hirId))
{
if (mappings->is_impl_item (hirId))
{
HIR::ImplBlock *implBlock
= mappings->lookup_associated_impl (hirId);
if (!implBlock->has_trait_ref ())
{
rust_warning_at (function.get_locus (), 0,
"associated function is never used: %<%s%>",
function.get_function_name ().c_str ());
}
}
else
{
rust_warning_at (function.get_locus (), 0,
"function is never used: %<%s%>",
function.get_function_name ().c_str ());
}
}
}
void visit (HIR::StructStruct &stct) override
{
HirId hirId = stct.get_mappings ().get_hirid ();
if (should_warn (hirId))
{
bool name_starts_underscore = stct.get_identifier ().at (0) == '_';
if (!name_starts_underscore)
rust_warning_at (stct.get_locus (), 0,
"struct is never constructed: %<%s%>",
stct.get_identifier ().c_str ());
}
else
{
// only warn the unused fields when in unwarned struct.
for (auto &field : stct.get_fields ())
{
HirId field_hir_id = field.get_mappings ().get_hirid ();
if (should_warn (field_hir_id))
{
rust_warning_at (field.get_locus (), 0,
"field is never read: %<%s%>",
field.get_field_name ().c_str ());
}
}
}
}
void visit (HIR::TupleStruct &stct) override
{
// only warn tuple struct unconstructed, and ignoring unused field
HirId hirId = stct.get_mappings ().get_hirid ();
if (should_warn (hirId))
{
rust_warning_at (stct.get_locus (), 0,
"struct is never constructed: %<%s%>",
stct.get_identifier ().c_str ());
}
}
void visit (HIR::ImplBlock &blc) override
{
if (blc.has_impl_items ())
{
for (auto &implItem : blc.get_impl_items ())
{
implItem->accept_vis (*this);
}
}
}
void visit (HIR::Module &mod) override
{
for (auto &item : mod.get_items ())
item->accept_vis (*this);
}
private:
std::set<HirId> live_symbols;
Resolver::Resolver *resolver;
Analysis::Mappings *mappings;
ScanDeadcode (std::set<HirId> &live_symbols)
: live_symbols (live_symbols), resolver (Resolver::Resolver::get ()),
mappings (Analysis::Mappings::get ()){};
bool should_warn (HirId hirId)
{
// TODO: There are more condition to check if should warn, i.e visibility,
// attributes.
return live_symbols.find (hirId) == live_symbols.end ();
}
};
} // namespace Analysis
} // namespace Rust
#endif