-
Notifications
You must be signed in to change notification settings - Fork 14.3k
[clang][dataflow] Tighten checking for existence of a function body. #78163
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[clang][dataflow] Tighten checking for existence of a function body. #78163
Conversation
In various places, we would previously call `FunctionDecl::hasBody()` (which checks whether any redeclaration of the function has a body, not necessarily the one on which `hasBody()` is being called). This is bug-prone, as a recent bug in Crubit's nullability checker has shown ([fix](google/crubit@4b01ed0), [fix for the fix](google/crubit@e0c5d8d)). Instead, we now use `FunctionDecl::doesThisDeclarationHaveABody()` which, as the name implies, checks whether the specific redeclaration it is being called on has a body. Alternatively, I considered being more lenient and "canonicalizing" to the `FunctionDecl` that has the body if the `FunctionDecl` being passed is a different redeclaration. However, this also risks hiding bugs: A caller might inadverently perform the analysis for all redeclarations of a function and end up duplicating work without realizing it. By accepting only the redeclaration that contains the body, we prevent this. I've checked, and all clients that I'm aware of do currently pass in the redeclaration that contains the function body. Typically this is because they use the `ast_matchers::hasBody()` matcher which, unlike `FunctionDecl::hasBody()`, only matches for the redeclaration containing the body.
@llvm/pr-subscribers-clang @llvm/pr-subscribers-clang-analysis Author: None (martinboehme) ChangesIn various places, we would previously call This is bug-prone, as a recent bug in Crubit's nullability checker has shown Instead, we now use Alternatively, I considered being more lenient and "canonicalizing" to the I've checked, and all clients that I'm aware of do currently pass in the Full diff: https://github.com/llvm/llvm-project/pull/78163.diff 5 Files Affected:
diff --git a/clang/include/clang/Analysis/FlowSensitive/ControlFlowContext.h b/clang/include/clang/Analysis/FlowSensitive/ControlFlowContext.h
index 768387a121b920..405e93287a05d3 100644
--- a/clang/include/clang/Analysis/FlowSensitive/ControlFlowContext.h
+++ b/clang/include/clang/Analysis/FlowSensitive/ControlFlowContext.h
@@ -32,7 +32,8 @@ namespace dataflow {
class ControlFlowContext {
public:
/// Builds a ControlFlowContext from a `FunctionDecl`.
- /// `Func.hasBody()` must be true, and `Func.isTemplated()` must be false.
+ /// `Func.doesThisDeclarationHaveABody()` must be true, and
+ /// `Func.isTemplated()` must be false.
static llvm::Expected<ControlFlowContext> build(const FunctionDecl &Func);
/// Builds a ControlFlowContext from an AST node. `D` is the function in which
diff --git a/clang/include/clang/Analysis/FlowSensitive/DataflowEnvironment.h b/clang/include/clang/Analysis/FlowSensitive/DataflowEnvironment.h
index e8c27d6c12038b..1543f900e401d6 100644
--- a/clang/include/clang/Analysis/FlowSensitive/DataflowEnvironment.h
+++ b/clang/include/clang/Analysis/FlowSensitive/DataflowEnvironment.h
@@ -172,7 +172,8 @@ class Environment {
///
/// Requirements:
///
- /// The function must have a body.
+ /// The function must have a body, i.e.
+ /// `FunctionDecl::doesThisDecalarationHaveABody()` must be true.
void initialize();
/// Returns a new environment that is a copy of this one.
diff --git a/clang/lib/Analysis/FlowSensitive/ControlFlowContext.cpp b/clang/lib/Analysis/FlowSensitive/ControlFlowContext.cpp
index 56246066e4aa13..c9ebffe6f37801 100644
--- a/clang/lib/Analysis/FlowSensitive/ControlFlowContext.cpp
+++ b/clang/lib/Analysis/FlowSensitive/ControlFlowContext.cpp
@@ -69,7 +69,7 @@ static llvm::BitVector findReachableBlocks(const CFG &Cfg) {
llvm::Expected<ControlFlowContext>
ControlFlowContext::build(const FunctionDecl &Func) {
- if (!Func.hasBody())
+ if (!Func.doesThisDeclarationHaveABody())
return llvm::createStringError(
std::make_error_code(std::errc::invalid_argument),
"Cannot analyze function without a body");
diff --git a/clang/lib/Analysis/FlowSensitive/DataflowAnalysisContext.cpp b/clang/lib/Analysis/FlowSensitive/DataflowAnalysisContext.cpp
index fa114979c8e326..03670b1821e374 100644
--- a/clang/lib/Analysis/FlowSensitive/DataflowAnalysisContext.cpp
+++ b/clang/lib/Analysis/FlowSensitive/DataflowAnalysisContext.cpp
@@ -292,7 +292,7 @@ DataflowAnalysisContext::getControlFlowContext(const FunctionDecl *F) {
if (It != FunctionContexts.end())
return &It->second;
- if (F->hasBody()) {
+ if (F->doesThisDeclarationHaveABody()) {
auto CFCtx = ControlFlowContext::build(*F);
// FIXME: Handle errors.
assert(CFCtx);
diff --git a/clang/lib/Analysis/FlowSensitive/DataflowEnvironment.cpp b/clang/lib/Analysis/FlowSensitive/DataflowEnvironment.cpp
index 96fe6df88dbb9f..a50ee57a3c11b4 100644
--- a/clang/lib/Analysis/FlowSensitive/DataflowEnvironment.cpp
+++ b/clang/lib/Analysis/FlowSensitive/DataflowEnvironment.cpp
@@ -386,7 +386,7 @@ void Environment::initialize() {
return;
if (const auto *FuncDecl = dyn_cast<FunctionDecl>(DeclCtx)) {
- assert(FuncDecl->getBody() != nullptr);
+ assert(FuncDecl->doesThisDeclarationHaveABody());
initFieldsGlobalsAndFuncs(FuncDecl);
@@ -426,7 +426,7 @@ void Environment::initialize() {
// FIXME: Add support for resetting globals after function calls to enable
// the implementation of sound analyses.
void Environment::initFieldsGlobalsAndFuncs(const FunctionDecl *FuncDecl) {
- assert(FuncDecl->getBody() != nullptr);
+ assert(FuncDecl->doesThisDeclarationHaveABody());
FieldSet Fields;
llvm::DenseSet<const VarDecl *> Vars;
|
Build failure is merely a clang-format failure in a file not touched by this PR (clang/docs/HLSL/FunctionCalls.rst). |
…lvm#78163) In various places, we would previously call `FunctionDecl::hasBody()` (which checks whether any redeclaration of the function has a body, not necessarily the one on which `hasBody()` is being called). This is bug-prone, as a recent bug in Crubit's nullability checker has shown ([fix](google/crubit@4b01ed0), [fix for the fix](google/crubit@e0c5d8d)). Instead, we now use `FunctionDecl::doesThisDeclarationHaveABody()` which, as the name implies, checks whether the specific redeclaration it is being called on has a body. Alternatively, I considered being more lenient and "canonicalizing" to the `FunctionDecl` that has the body if the `FunctionDecl` being passed is a different redeclaration. However, this also risks hiding bugs: A caller might inadverently perform the analysis for all redeclarations of a function and end up duplicating work without realizing it. By accepting only the redeclaration that contains the body, we prevent this. I've checked, and all clients that I'm aware of do currently pass in the redeclaration that contains the function body. Typically this is because they use the `ast_matchers::hasBody()` matcher which, unlike `FunctionDecl::hasBody()`, only matches for the redeclaration containing the body.
In various places, we would previously call
FunctionDecl::hasBody()
(whichchecks whether any redeclaration of the function has a body, not necessarily the
one on which
hasBody()
is being called).This is bug-prone, as a recent bug in Crubit's nullability checker has shown
(fix, fix for the fix).
Instead, we now use
FunctionDecl::doesThisDeclarationHaveABody()
which, as thename implies, checks whether the specific redeclaration it is being called on
has a body.
Alternatively, I considered being more lenient and "canonicalizing" to the
FunctionDecl
that has the body if theFunctionDecl
being passed is adifferent redeclaration. However, this also risks hiding bugs: A caller might
inadverently perform the analysis for all redeclarations of a function and end
up duplicating work without realizing it. By accepting only the redeclaration
that contains the body, we prevent this.
I've checked, and all clients that I'm aware of do currently pass in the
redeclaration that contains the function body. Typically this is because they
use the
ast_matchers::hasBody()
matcher which, unlikeFunctionDecl::hasBody()
, only matches for the redeclaration containing thebody.