[clang] Reject constexpr-unknown values as constant expressions more consistently #129952
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
efriedma-quic
requested review from
cor3ntin,
shafik,
dtcxzyw,
frederick-vs-ja and
tbaederr
llvmbot
added
clang
|
@llvm/pr-subscribers-clang Author: Eli Friedman (efriedma-quic) ChangesPerform the check for constexpr-unknown values in the same place we perform checks for other values which don't count as constant expressions. While I'm here, also fix a rejects-valid with a reference that doesn't have an initializer. This diagnostic was also covering up some of the bugs here. The existing behavior with -fexperimental-new-constant-interpreter seems to be correct, but the diagnostics are slightly different; it would be helpful if someone could check on that as a followup. Followup to #128409. Fixes #129844. Fixes #129845. Full diff: https://github.com/llvm/llvm-project/pull/129952.diff 4 Files Affected:
diff --git a/clang/lib/AST/ExprConstant.cpp b/clang/lib/AST/ExprConstant.cpp
index d9a1e5bb42343..6b4afc099ad16 100644
--- a/clang/lib/AST/ExprConstant.cpp
+++ b/clang/lib/AST/ExprConstant.cpp
@@ -2419,6 +2419,18 @@ static bool CheckLValueConstantExpression(EvalInfo &Info, SourceLocation Loc,
LVal.getLValueCallIndex() == 0) &&
"have call index for global lvalue");
+ if (LVal.allowConstexprUnknown()) {
+ if (BaseVD) {
+ Info.FFDiag(Loc, diag::note_constexpr_var_init_non_constant, 1)
+ << BaseVD;
+ NoteLValueLocation(Info, Base);
+ } else {
+ Info.FFDiag(Loc);
+ }
+ return false;
+ }
+
+
if (Base.is<DynamicAllocLValue>()) {
Info.FFDiag(Loc, diag::note_constexpr_dynamic_alloc)
<< IsReferenceType << !Designator.Entries.empty();
@@ -3597,7 +3609,8 @@ static bool evaluateVarDeclInit(EvalInfo &Info, const Expr *E,
// expressions here; doing so would regress diagnostics for things like
// reading from a volatile constexpr variable.
if ((Info.getLangOpts().CPlusPlus && !VD->hasConstantInitialization() &&
- VD->mightBeUsableInConstantExpressions(Info.Ctx)) ||
+ VD->mightBeUsableInConstantExpressions(Info.Ctx) &&
+ !AllowConstexprUnknown) ||
((Info.getLangOpts().CPlusPlus || Info.getLangOpts().OpenCL) &&
!Info.getLangOpts().CPlusPlus11 && !VD->hasICEInitializer(Info.Ctx))) {
if (Init) {
@@ -16993,18 +17006,6 @@ bool Expr::EvaluateAsInitializer(APValue &Value, const ASTContext &Ctx,
if (!Info.discardCleanups())
llvm_unreachable("Unhandled cleanup; missing full expression marker?");
-
- if (Value.allowConstexprUnknown()) {
- assert(Value.isLValue() && "Expected an lvalue");
- auto Base = Value.getLValueBase();
- const auto *NewVD = Base.dyn_cast<const ValueDecl *>();
- if (!NewVD)
- NewVD = VD;
- Info.FFDiag(getExprLoc(), diag::note_constexpr_var_init_non_constant, 1)
- << NewVD;
- NoteLValueLocation(Info, Base);
- return false;
- }
}
return CheckConstantExpression(Info, DeclLoc, DeclTy, Value,
diff --git a/clang/test/CodeGenCXX/cxx23-p2280r4.cpp b/clang/test/CodeGenCXX/cxx23-p2280r4.cpp
index d5409be451df0..c91457e5c20de 100644
--- a/clang/test/CodeGenCXX/cxx23-p2280r4.cpp
+++ b/clang/test/CodeGenCXX/cxx23-p2280r4.cpp
@@ -13,3 +13,15 @@ int& test() {
auto &i = s;
return i;
}
+
+// CHECK: @_Z1fv(
+// CHECK-NEXT: entry:
+// CHECK-NEXT: getelementptr inbounds nuw %struct.B
+// CHECK-NEXT: getelementptr inbounds nuw %struct.A
+// CHECK-NEXT: [[X:%.*]] = load ptr, ptr @x, align 8
+// CHECK-NEXT: store ptr [[X]]
+int &ff();
+int &x = ff();
+struct A { int& x; };
+struct B { A x[20]; };
+B f() { return {x,x,x,x,x,x,x,x,x,x,x,x,x,x,x,x,x,x,x,x}; }
diff --git a/clang/test/SemaCXX/constant-expression-cxx11.cpp b/clang/test/SemaCXX/constant-expression-cxx11.cpp
index 76e2f81947051..c35f3a5632a05 100644
--- a/clang/test/SemaCXX/constant-expression-cxx11.cpp
+++ b/clang/test/SemaCXX/constant-expression-cxx11.cpp
@@ -1472,8 +1472,8 @@ namespace ConvertedConstantExpr {
enum class E {
em = m,
en = n, // expected-error {{enumerator value is not a constant expression}} cxx11_20-note {{initializer of 'n' is unknown}}
- eo = (m + // pre-cxx23-error {{not a constant expression}}
- n // cxx11_20-note {{initializer of 'n' is unknown}} cxx23-error {{not a constant expression}}
+ eo = (m + // expected-error {{not a constant expression}}
+ n // cxx11_20-note {{initializer of 'n' is unknown}}
),
eq = reinterpret_cast<long>((int*)0) // expected-error {{not a constant expression}} expected-note {{reinterpret_cast}}
};
diff --git a/clang/test/SemaCXX/constant-expression-p2280r4.cpp b/clang/test/SemaCXX/constant-expression-p2280r4.cpp
index 65e5e6b34b48f..731b0e9b03046 100644
--- a/clang/test/SemaCXX/constant-expression-p2280r4.cpp
+++ b/clang/test/SemaCXX/constant-expression-p2280r4.cpp
@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 -std=c++23 -verify %s
+// RUN: %clang_cc1 -std=c++23 -verify=expected,nointerpreter %s
// RUN: %clang_cc1 -std=c++23 -verify %s -fexperimental-new-constant-interpreter
using size_t = decltype(sizeof(0));
@@ -154,3 +154,26 @@ int g() {
static_assert(f(arr) == 5);
}
}
+
+namespace GH128409 {
+ int &ff();
+ int &x = ff(); // nointerpreter-note {{declared here}}
+ constinit int &z = x; // expected-error {{variable does not have a constant initializer}}
+ // expected-note@-1 {{required by 'constinit' specifier here}}
+ // nointerpreter-note@-2 {{initializer of 'x' is not a constant expression}}
+}
+
+namespace GH129845 {
+ int &ff();
+ int &x = ff(); // nointerpreter-note {{declared here}}
+ struct A { int& x; };
+ constexpr A g = {x}; // expected-error {{constexpr variable 'g' must be initialized by a constant expression}}
+ // nointerpreter-note@-1 {{initializer of 'x' is not a constant expression}}
+ const A* gg = &g;
+}
+
+namespace extern_reference_used_as_unknown {
+ extern int &x;
+ int y;
+ constinit int& g = (x,y); // expected-warning {{left operand of comma operator has no effect}}
+}
|
…consistently. Perform the check for constexpr-unknown values in the same place we perform checks for other values which don't count as constant expressions. While I'm here, also fix a rejects-valid with a reference that doesn't have an initializer. This diagnostic was also covering up some of the bugs here. The existing behavior with -fexperimental-new-constant-interpreter seems to be correct, but the diagnostics are slightly different; it would be helpful if someone could check on that as a followup. Followup to llvm#128409. Fixes llvm#129844. Fixes llvm#129845.
efriedma-quic
force-pushed
the
constexpr-unknown-check-return
branch
from
edcb9a0 to
9ac636e
Compare
tbaederr
reviewed
Mar 6, 2025
| namespace GH128409 { | ||
| int &ff(); | ||
| int &x = ff(); // nointerpreter-note {{declared here}} | ||
| constinit int &z = x; // expected-error {{variable does not have a constant initializer}} |
There was a problem hiding this comment.
If you end the comments in \, you don't need to add the @-2 stuff.
tbaederr
reviewed
Mar 6, 2025
| int &x = ff(); // nointerpreter-note {{declared here}} | ||
| constinit int &z = x; // expected-error {{variable does not have a constant initializer}} | ||
| // expected-note@-1 {{required by 'constinit' specifier here}} | ||
| // nointerpreter-note@-2 {{initializer of 'x' is not a constant expression}} |
There was a problem hiding this comment.
I don't want to block on this, but the diagnostic here is a little off... Even if the initializer of x was a constant expression, the declaration of z would still not be valid because x is not const (or constexpr).
In other words, if I saw this error in practice, I'd try to fix x's initializer, just to be greeted by another diagnostic.
There was a problem hiding this comment.
This is pre-existing though
There was a problem hiding this comment.
Both clang and gcc accept the following:
int y;
int &r1 = y;
constinit int &r2 = r1;
I think that's right, looking at the rules for "potentially-constant" variables?
There was a problem hiding this comment.
https://godbolt.org/z/81M8aYGPn
Yeah, not a fan of any of this. But this patch LGTM.
cor3ntin
approved these changes
Mar 6, 2025
| int &x = ff(); // nointerpreter-note {{declared here}} | ||
| constinit int &z = x; // expected-error {{variable does not have a constant initializer}} | ||
| // expected-note@-1 {{required by 'constinit' specifier here}} | ||
| // nointerpreter-note@-2 {{initializer of 'x' is not a constant expression}} |
There was a problem hiding this comment.
This is pre-existing though
|
/cherry-pick 42d49a7 |
|
Failed to cherry-pick: 42d49a7 https://github.com/llvm/llvm-project/actions/runs/13755084406 Please manually backport the fix and push it to your github fork. Once this is done, please create a pull request |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/154/builds/13059 Here is the relevant piece of the build log for the reference |
|
The failure above is real, ABI or 32-bitness issue I think. Looking at it now. |
DavidSpickett
added a commit
that referenced
this pull request
Mar 10, 2025
#129952 / 42d49a7 added this test which is failing on 32-bit ARM because the alignment chosen is 4 not 8. Which would make sense if this is a 32/64 bit difference https://lab.llvm.org/buildbot/#/builders/154/builds/13059 ``` <stdin>:34:30: note: scanning from here define dso_local void @_Z1fv(ptr dead_on_unwind noalias writable sret(%struct.B) align 4 %agg.result) #0 { ^ <stdin>:38:2: note: possible intended match here %0 = load ptr, ptr @x, align 4 ^ ``` The other test does not check alignment, so I'm assuming that it is not important here.
llvm-sync bot
pushed a commit
to arm/arm-toolchain
that referenced
this pull request
Mar 10, 2025
…e (#130589) llvm/llvm-project#129952 / 42d49a7 added this test which is failing on 32-bit ARM because the alignment chosen is 4 not 8. Which would make sense if this is a 32/64 bit difference https://lab.llvm.org/buildbot/#/builders/154/builds/13059 ``` <stdin>:34:30: note: scanning from here define dso_local void @_Z1fv(ptr dead_on_unwind noalias writable sret(%struct.B) align 4 %agg.result) #0 { ^ <stdin>:38:2: note: possible intended match here %0 = load ptr, ptr @x, align 4 ^ ``` The other test does not check alignment, so I'm assuming that it is not important here.
efriedma-quic
added a commit
to efriedma-quic/llvm-project
that referenced
this pull request
Mar 10, 2025
…consistently (llvm#129952) Perform the check for constexpr-unknown values in the same place we perform checks for other values which don't count as constant expressions. While I'm here, also fix a rejects-valid with a reference that doesn't have an initializer. This diagnostic was also covering up some of the bugs here. The existing behavior with -fexperimental-new-constant-interpreter seems to be correct, but the diagnostics are slightly different; it would be helpful if someone could check on that as a followup. Followup to llvm#128409. Fixes llvm#129844. Fixes llvm#129845.
efriedma-quic
pushed a commit
to efriedma-quic/llvm-project
that referenced
this pull request
Mar 10, 2025
) llvm#129952 / 42d49a7 added this test which is failing on 32-bit ARM because the alignment chosen is 4 not 8. Which would make sense if this is a 32/64 bit difference https://lab.llvm.org/buildbot/#/builders/154/builds/13059 ``` <stdin>:34:30: note: scanning from here define dso_local void @_Z1fv(ptr dead_on_unwind noalias writable sret(%struct.B) align 4 %agg.result) #0 { ^ <stdin>:38:2: note: possible intended match here %0 = load ptr, ptr @x, align 4 ^ ``` The other test does not check alignment, so I'm assuming that it is not important here.
llvm-sync bot
pushed a commit
to arm/arm-toolchain
that referenced
this pull request
Mar 11, 2025
llvm/llvm-project#129952 / 42d49a7 added this test which is failing on 32-bit ARM because the alignment chosen is 4 not 8. Which would make sense if this is a 32/64 bit difference https://lab.llvm.org/buildbot/#/builders/154/builds/13059 ``` <stdin>:34:30: note: scanning from here define dso_local void @_Z1fv(ptr dead_on_unwind noalias writable sret(%struct.B) align 4 %agg.result) #0 { ^ <stdin>:38:2: note: possible intended match here %0 = load ptr, ptr @x, align 4 ^ ``` The other test does not check alignment, so I'm assuming that it is not important here.
swift-ci
pushed a commit
to swiftlang/llvm-project
that referenced
this pull request
Mar 11, 2025
…consistently (llvm#129952) Perform the check for constexpr-unknown values in the same place we perform checks for other values which don't count as constant expressions. While I'm here, also fix a rejects-valid with a reference that doesn't have an initializer. This diagnostic was also covering up some of the bugs here. The existing behavior with -fexperimental-new-constant-interpreter seems to be correct, but the diagnostics are slightly different; it would be helpful if someone could check on that as a followup. Followup to llvm#128409. Fixes llvm#129844. Fixes llvm#129845.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Perform the check for constexpr-unknown values in the same place we perform checks for other values which don't count as constant expressions.
While I'm here, also fix a rejects-valid with a reference that doesn't have an initializer. This diagnostic was also covering up some of the bugs here.
The existing behavior with -fexperimental-new-constant-interpreter seems to be correct, but the diagnostics are slightly different; it would be helpful if someone could check on that as a followup.
Followup to #128409.
Fixes #129844. Fixes #129845.