[clang][driver] Expose a frontend option for trap-unreachable by ilovepi · Pull Request #174894 · llvm/llvm-project

ilovepi · 2026-01-08T00:42:11Z

We have several issues that list surprising behavior around UB. In many
cases, this causes undesirable control flow, such as execution falling
through to the next function (or whatever is in memory) instead of
remaining within the bounds of the procedure. #174844, #48943, #146791,
and #137741 all discuss a host of related issues. In #174844, it was
mentioned that we have backend support for this for Rust, and at least
one big class of these issues could be addressed by exposing the option
to clang.

This patch adds a new driver option that does just that. For now, we're
leaving this option off by default, though we expect only small
differences in code size or performance as a result if it were to be
enabled. There will be an RFC in the future when we have more confidence
this should be the default configuration.

Fixes #174844

ilovepi · 2026-01-08T00:42:32Z

[clang][driver] Expose a frontend option for trap-unreachable #174894 👈 (View in Graphite)
main

This stack of pull requests is managed by Graphite. Learn more about stacking.

llvmbot · 2026-01-08T00:42:45Z

@llvm/pr-subscribers-clang-codegen
@llvm/pr-subscribers-clang

@llvm/pr-subscribers-clang-driver

Author: Paul Kirth (ilovepi)

Changes

We have several issues that list surprising behavior around UB. In many
cases, this causes undesirable control flow, such as execution falling
through to the next function (or whatever is in memory) instead of
remaining within the bounds of the procedure. #174844, #48943, #146791,
and #137741 all discuss a host of related issues. In #174844, it was
mentioned that we have backend support for this for Rust, and at least
one big class of these issues could be addressed by exposing the option
to clang.

This patch adds a new driver option that does just that. For now, we're
leaving this option off by default, though we expect only small
differences in code size or performance as a result if it were to be
enabled. There will be an RFC in the future when we have more confidence
this should be the default configuration.

Fixes #174844

Full diff: https://github.com/llvm/llvm-project/pull/174894.diff

7 Files Affected:

(modified) clang/docs/ReleaseNotes.rst (+2)
(modified) clang/include/clang/Basic/CodeGenOptions.def (+3)
(modified) clang/include/clang/Options/Options.td (+6)
(modified) clang/lib/CodeGen/BackendUtil.cpp (+1)
(modified) clang/lib/Driver/ToolChains/Clang.cpp (+3)
(added) clang/test/CodeGen/X86/unreachable-trap.c (+18)
(modified) clang/test/Driver/clang_f_opts.c (+5)

diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index 932b9fad40c42..e5cbc3eecb25e 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -348,6 +348,8 @@ New Compiler Flags
 - New options for enabling allocation token instrumentation: ``-fsanitize=alloc-token``, ``-falloc-token-max=``, ``-fsanitize-alloc-token-fast-abi``, ``-fsanitize-alloc-token-extended``.
 - The ``-resource-dir`` option is now displayed in the list of options shown by ``--help``.
 - New option ``-fmatrix-memory-layout`` added to control the memory layout of Clang matrix types. (e.g. ``-fmatrix-memory-layout=column-major`` or ``-fmatrix-memory-layout=row-major``).
+- New option ``-ftrap-unreachable`` added to enable the existing backend option: TrapUnreachable.
+  This behavior is off by default (e.g. no change in the compiler's behavior) for now.
 
 Lanai Support
 ^^^^^^^^^^^^^^
diff --git a/clang/include/clang/Basic/CodeGenOptions.def b/clang/include/clang/Basic/CodeGenOptions.def
index a6b1acdcf5ea9..9e84e22ff8d46 100644
--- a/clang/include/clang/Basic/CodeGenOptions.def
+++ b/clang/include/clang/Basic/CodeGenOptions.def
@@ -428,6 +428,9 @@ CODEGENOPT(DirectAccessExternalData, 1, 0, Benign)
 /// paths that reach the end of a function without executing a required return.
 CODEGENOPT(StrictReturn, 1, 1, Benign)
 
+/// Whether we should use make unreachable trap or not.
+CODEGENOPT(TrapUnreachable, 1, 0, Benign)
+
 /// Whether emit pseudo probes for sample pgo profile collection.
 CODEGENOPT(PseudoProbeForProfiling, 1, 0, Benign)
 
diff --git a/clang/include/clang/Options/Options.td b/clang/include/clang/Options/Options.td
index f336542992993..33be1569b7e8c 100644
--- a/clang/include/clang/Options/Options.td
+++ b/clang/include/clang/Options/Options.td
@@ -4685,6 +4685,12 @@ defm strict_return : BoolFOption<"strict-return",
             " of a non-void function as unreachable">,
   PosFlag<SetTrue>>;
 
+defm trap_unreachable : BoolFOption<"trap-unreachable",
+  CodeGenOpts<"TrapUnreachable">, DefaultFalse,
+  PosFlag<SetTrue, [], [ClangOption, CC1Option],
+          "Treat unreachable instructions as traps">,
+  NegFlag<SetFalse>>;
+
 let Flags = [TargetSpecific] in {
 defm ptrauth_intrinsics : OptInCC1FFlag<"ptrauth-intrinsics", "Enable pointer authentication intrinsics">;
 defm ptrauth_calls : OptInCC1FFlag<"ptrauth-calls", "Enable signing and authentication of all indirect calls">;
diff --git a/clang/lib/CodeGen/BackendUtil.cpp b/clang/lib/CodeGen/BackendUtil.cpp
index c5ea8c9e8c3de..5f57705ffc77e 100644
--- a/clang/lib/CodeGen/BackendUtil.cpp
+++ b/clang/lib/CodeGen/BackendUtil.cpp
@@ -466,6 +466,7 @@ static bool initTargetOptions(const CompilerInstance &CI,
   Options.Hotpatch = CodeGenOpts.HotPatch;
   Options.JMCInstrument = CodeGenOpts.JMCInstrument;
   Options.XCOFFReadOnlyPointers = CodeGenOpts.XCOFFReadOnlyPointers;
+  Options.TrapUnreachable = CodeGenOpts.TrapUnreachable;
 
   switch (CodeGenOpts.getVecLib()) {
   case llvm::driver::VectorLibrary::NoLibrary:
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
index 699fc31f23946..f54af613e72b8 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -5808,6 +5808,9 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
   Args.addOptInFlag(CmdArgs, options::OPT_fsplit_stack,
                     options::OPT_fno_split_stack);
 
+  Args.addOptInFlag(CmdArgs, options::OPT_ftrap_unreachable,
+                     options::OPT_ftrap_unreachable);
+
   // -fprotect-parens=0 is default.
   if (Args.hasFlag(options::OPT_fprotect_parens,
                    options::OPT_fno_protect_parens, false))
diff --git a/clang/test/CodeGen/X86/unreachable-trap.c b/clang/test/CodeGen/X86/unreachable-trap.c
new file mode 100644
index 0000000000000..bea34e80dd873
--- /dev/null
+++ b/clang/test/CodeGen/X86/unreachable-trap.c
@@ -0,0 +1,18 @@
+// RUN: %clang_cc1 %s -O1 -triple=x86_64-unknown-linux-gnu -ftrap-unreachable -S  -o - 2>&1 | FileCheck %s --check-prefix=TRAP
+// RUN: %clang_cc1 %s -O1 -triple=x86_64-unknown-linux-gnu -S  -o - 2>&1 | FileCheck %s --check-prefix=NOTRAP
+
+// TRAP: ud2
+// NOTRAP-NOT: ud2
+
+[[noreturn]]
+void exit(int);
+
+#define NULL 0
+
+static void test(void) {
+    int *ptr = NULL;
+    *ptr = 0;
+    exit(0);
+}
+
+void foo() { test(); }
diff --git a/clang/test/Driver/clang_f_opts.c b/clang/test/Driver/clang_f_opts.c
index 5871f1580d6b7..aedbef393abcc 100644
--- a/clang/test/Driver/clang_f_opts.c
+++ b/clang/test/Driver/clang_f_opts.c
@@ -651,3 +651,8 @@
 // RUN: %clang -### --target=x86_64-pc-windows-msvc -fno-strict-aliasing %s 2>&1 | FileCheck -check-prefix=CHECK-NO-STRICT-ALIASING %s
 // CHECK-STRICT-ALIASING-NOT: -relaxed-aliasing
 // CHECK-NO-STRICT-ALIASING: -relaxed-aliasing
+
+// RUN: %clang -### -ftrap-unreachable %s 2>&1 | FileCheck %s -check-prefix=UNREACHABLE-TRAP
+// RUN: %clang -### -fno-trap-unreachable %s 2>&1 | FileCheck %s -check-prefix=NO-UNREACHABLE-TRAP
+// UNREACHABLE-TRAP: "-ftrap-unreachable"
+// NO-UNREACHABLE-TRAP-NOT: "-ftrap-unreachable"

llvmbot · 2026-01-08T00:42:45Z

@llvm/pr-subscribers-backend-x86

Author: Paul Kirth (ilovepi)

Changes

We have several issues that list surprising behavior around UB. In many
cases, this causes undesirable control flow, such as execution falling
through to the next function (or whatever is in memory) instead of
remaining within the bounds of the procedure. #174844, #48943, #146791,
and #137741 all discuss a host of related issues. In #174844, it was
mentioned that we have backend support for this for Rust, and at least
one big class of these issues could be addressed by exposing the option
to clang.

This patch adds a new driver option that does just that. For now, we're
leaving this option off by default, though we expect only small
differences in code size or performance as a result if it were to be
enabled. There will be an RFC in the future when we have more confidence
this should be the default configuration.

Fixes #174844

Full diff: https://github.com/llvm/llvm-project/pull/174894.diff

7 Files Affected:

(modified) clang/docs/ReleaseNotes.rst (+2)
(modified) clang/include/clang/Basic/CodeGenOptions.def (+3)
(modified) clang/include/clang/Options/Options.td (+6)
(modified) clang/lib/CodeGen/BackendUtil.cpp (+1)
(modified) clang/lib/Driver/ToolChains/Clang.cpp (+3)
(added) clang/test/CodeGen/X86/unreachable-trap.c (+18)
(modified) clang/test/Driver/clang_f_opts.c (+5)

diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index 932b9fad40c42..e5cbc3eecb25e 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -348,6 +348,8 @@ New Compiler Flags
 - New options for enabling allocation token instrumentation: ``-fsanitize=alloc-token``, ``-falloc-token-max=``, ``-fsanitize-alloc-token-fast-abi``, ``-fsanitize-alloc-token-extended``.
 - The ``-resource-dir`` option is now displayed in the list of options shown by ``--help``.
 - New option ``-fmatrix-memory-layout`` added to control the memory layout of Clang matrix types. (e.g. ``-fmatrix-memory-layout=column-major`` or ``-fmatrix-memory-layout=row-major``).
+- New option ``-ftrap-unreachable`` added to enable the existing backend option: TrapUnreachable.
+  This behavior is off by default (e.g. no change in the compiler's behavior) for now.
 
 Lanai Support
 ^^^^^^^^^^^^^^
diff --git a/clang/include/clang/Basic/CodeGenOptions.def b/clang/include/clang/Basic/CodeGenOptions.def
index a6b1acdcf5ea9..9e84e22ff8d46 100644
--- a/clang/include/clang/Basic/CodeGenOptions.def
+++ b/clang/include/clang/Basic/CodeGenOptions.def
@@ -428,6 +428,9 @@ CODEGENOPT(DirectAccessExternalData, 1, 0, Benign)
 /// paths that reach the end of a function without executing a required return.
 CODEGENOPT(StrictReturn, 1, 1, Benign)
 
+/// Whether we should use make unreachable trap or not.
+CODEGENOPT(TrapUnreachable, 1, 0, Benign)
+
 /// Whether emit pseudo probes for sample pgo profile collection.
 CODEGENOPT(PseudoProbeForProfiling, 1, 0, Benign)
 
diff --git a/clang/include/clang/Options/Options.td b/clang/include/clang/Options/Options.td
index f336542992993..33be1569b7e8c 100644
--- a/clang/include/clang/Options/Options.td
+++ b/clang/include/clang/Options/Options.td
@@ -4685,6 +4685,12 @@ defm strict_return : BoolFOption<"strict-return",
             " of a non-void function as unreachable">,
   PosFlag<SetTrue>>;
 
+defm trap_unreachable : BoolFOption<"trap-unreachable",
+  CodeGenOpts<"TrapUnreachable">, DefaultFalse,
+  PosFlag<SetTrue, [], [ClangOption, CC1Option],
+          "Treat unreachable instructions as traps">,
+  NegFlag<SetFalse>>;
+
 let Flags = [TargetSpecific] in {
 defm ptrauth_intrinsics : OptInCC1FFlag<"ptrauth-intrinsics", "Enable pointer authentication intrinsics">;
 defm ptrauth_calls : OptInCC1FFlag<"ptrauth-calls", "Enable signing and authentication of all indirect calls">;
diff --git a/clang/lib/CodeGen/BackendUtil.cpp b/clang/lib/CodeGen/BackendUtil.cpp
index c5ea8c9e8c3de..5f57705ffc77e 100644
--- a/clang/lib/CodeGen/BackendUtil.cpp
+++ b/clang/lib/CodeGen/BackendUtil.cpp
@@ -466,6 +466,7 @@ static bool initTargetOptions(const CompilerInstance &CI,
   Options.Hotpatch = CodeGenOpts.HotPatch;
   Options.JMCInstrument = CodeGenOpts.JMCInstrument;
   Options.XCOFFReadOnlyPointers = CodeGenOpts.XCOFFReadOnlyPointers;
+  Options.TrapUnreachable = CodeGenOpts.TrapUnreachable;
 
   switch (CodeGenOpts.getVecLib()) {
   case llvm::driver::VectorLibrary::NoLibrary:
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
index 699fc31f23946..f54af613e72b8 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -5808,6 +5808,9 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
   Args.addOptInFlag(CmdArgs, options::OPT_fsplit_stack,
                     options::OPT_fno_split_stack);
 
+  Args.addOptInFlag(CmdArgs, options::OPT_ftrap_unreachable,
+                     options::OPT_ftrap_unreachable);
+
   // -fprotect-parens=0 is default.
   if (Args.hasFlag(options::OPT_fprotect_parens,
                    options::OPT_fno_protect_parens, false))
diff --git a/clang/test/CodeGen/X86/unreachable-trap.c b/clang/test/CodeGen/X86/unreachable-trap.c
new file mode 100644
index 0000000000000..bea34e80dd873
--- /dev/null
+++ b/clang/test/CodeGen/X86/unreachable-trap.c
@@ -0,0 +1,18 @@
+// RUN: %clang_cc1 %s -O1 -triple=x86_64-unknown-linux-gnu -ftrap-unreachable -S  -o - 2>&1 | FileCheck %s --check-prefix=TRAP
+// RUN: %clang_cc1 %s -O1 -triple=x86_64-unknown-linux-gnu -S  -o - 2>&1 | FileCheck %s --check-prefix=NOTRAP
+
+// TRAP: ud2
+// NOTRAP-NOT: ud2
+
+[[noreturn]]
+void exit(int);
+
+#define NULL 0
+
+static void test(void) {
+    int *ptr = NULL;
+    *ptr = 0;
+    exit(0);
+}
+
+void foo() { test(); }
diff --git a/clang/test/Driver/clang_f_opts.c b/clang/test/Driver/clang_f_opts.c
index 5871f1580d6b7..aedbef393abcc 100644
--- a/clang/test/Driver/clang_f_opts.c
+++ b/clang/test/Driver/clang_f_opts.c
@@ -651,3 +651,8 @@
 // RUN: %clang -### --target=x86_64-pc-windows-msvc -fno-strict-aliasing %s 2>&1 | FileCheck -check-prefix=CHECK-NO-STRICT-ALIASING %s
 // CHECK-STRICT-ALIASING-NOT: -relaxed-aliasing
 // CHECK-NO-STRICT-ALIASING: -relaxed-aliasing
+
+// RUN: %clang -### -ftrap-unreachable %s 2>&1 | FileCheck %s -check-prefix=UNREACHABLE-TRAP
+// RUN: %clang -### -fno-trap-unreachable %s 2>&1 | FileCheck %s -check-prefix=NO-UNREACHABLE-TRAP
+// UNREACHABLE-TRAP: "-ftrap-unreachable"
+// NO-UNREACHABLE-TRAP-NOT: "-ftrap-unreachable"

github-actions · 2026-01-08T00:43:43Z

✅ With the latest revision this PR passed the C/C++ code formatter.

ilovepi · 2026-01-08T00:43:46Z

+// RUN: %clang_cc1 %s -O1 -triple=x86_64-unknown-linux-gnu -ftrap-unreachable -S  -o - 2>&1 | FileCheck %s --check-prefix=TRAP
+// RUN: %clang_cc1 %s -O1 -triple=x86_64-unknown-linux-gnu -S  -o - 2>&1 | FileCheck %s --check-prefix=NOTRAP


I don't really like testing the codegen option like this, but IDK another way to check that the right bit got set in the backend. Is this fine? Is there a better pattern I should follow for cases like this?

ilovepi · 2026-01-08T00:47:19Z

Also, I mentioned several issues in the description, but it wasn't immediately obvious which of those would actually be fixed by this. If folks know which of those (or additional) bugs should be closed with this landed let me know. Though I guess there's an argument to leave them open until we change the default?

nikic · 2026-01-08T09:51:51Z

  Options.Hotpatch = CodeGenOpts.HotPatch;
  Options.JMCInstrument = CodeGenOpts.JMCInstrument;
  Options.XCOFFReadOnlyPointers = CodeGenOpts.XCOFFReadOnlyPointers;
+  Options.TrapUnreachable = CodeGenOpts.TrapUnreachable;


Should we also enable NoTrapAfterNoreturn? (Alternatively we could have multiple modes like -ftrap-unreachable=all and -ftrap-unreachable=except-noreturn or so.)

That may be a good idea. Let me see how that looks.

That does seem better, so I uploaded the new version.

efriedma-quic · 2026-01-12T22:11:06Z

+    : Joined<["-"], "ftrap-unreachable=">,
+      Group<f_Group>,
+      Visibility<[ClangOption, CC1Option]>,
+      HelpText<"Treat unreachable instruction as traps.">,


I think we need to be careful with the documentation, so users don't expect the option to do more than it actually does. Even with trap-unreachable, the compiler will prune impossible codepaths... which can lead to wild jumps, in general. It's just less likely. Really, the key here is that trap-unreachable inserts traps in places where it's "free": it doesn't increase the number of dynamically executed instructions for well-defined code, just codesize.

Thanks for pointing that out. I'll come up w/ some more accurate description tomorrow.

I've updated the wording a bit. I also considered when it is supported and does not increase the dynamic instruction count for well-defined code. but that seemed a bit verbose. Happy to adjust the text if there is something specific you'd like to see instead.

efriedma-quic

How does this interact with LTO?

MaskRay · 2026-01-15T17:15:51Z

+    : Joined<["-"], "ftrap-unreachable=">,
+      Group<f_Group>,
+      Visibility<[ClangOption, CC1Option]>,
+      HelpText<"Replace ``llvm.unreachable`` instructions with traps, when it is supported and profitable.">,


We omit the trailing .. If more text is needed, consider adding it to https://clang.llvm.org/docs/UsersManual.html#controlling-code-generation near https://clang.llvm.org/docs/UsersManual.html#cmdoption-ftrap-function

done. Thanks for pointing that out.

shafik

Thank you for doing this legwork, I wish we could be more aggressive but this at least moves us a little further along. Once you land this it might be nice to leave comments in the more central issues that this is now an option.

ilovepi · 2026-02-12T18:38:53Z

How does this interact with LTO?

Thanks for pointing that out. I've added options to make sure the right backend options get passed to the linker. Ideally, we wouldn't be relying on target options, but for now its no worse than other options. Hopefully we can get a better mechanism for tracking these types of options in the IR.

github-actions · 2026-02-12T19:32:23Z

🪟 Windows x64 Test Results

55727 tests passed
2665 tests skipped

✅ The build succeeded and all tests passed.

We have several issues that list surprising behavior around UB. In many cases, this causes undesirable control flow, such as execution falling through to the next function (or whatever is in memory) instead of remaining within the bounds of the procedure. #174844, #48943, #146791, and #137741 all discuss a host of related issues. In #174844, it was mentioned that we have backend support for this for Rust, and at least one big class of these issues could be addressed by exposing the option to clang. This patch adds a new driver option that does just that. For now, we're leaving this option off by default, though we expect only small differences in code size or performance as a result if it were to be enabled. There will be an RFC in the future when we have more confidence this should be the default configuration. Fixes #174844

eleviant · 2026-05-18T19:07:12Z

A gentle ping on behalf of @ilovepi . Can we get some traction on this?

We are interested in getting this fixed.
There is an alternative version of the fix #196550 - in case a simpler solution would go easier.

MaskRay · 2026-06-17T03:34:26Z

No objection.

eleviant · 2026-06-24T17:48:58Z

@nikic Can you please take a look?

ilovepi commented Jan 8, 2026

View reviewed changes

ilovepi force-pushed the users/ilovepi/clang-trap-unreachable branch from 8d9929c to 7f7feb3 Compare January 8, 2026 00:44

nikic reviewed Jan 8, 2026

View reviewed changes

ilovepi requested review from MaskRay and shafik January 12, 2026 18:15

ilovepi force-pushed the users/ilovepi/clang-trap-unreachable branch from 7f7feb3 to 815adb8 Compare January 12, 2026 21:08

efriedma-quic reviewed Jan 12, 2026

View reviewed changes

ilovepi force-pushed the users/ilovepi/clang-trap-unreachable branch from 1f698eb to b864741 Compare January 13, 2026 21:18

efriedma-quic reviewed Jan 13, 2026

View reviewed changes

MaskRay reviewed Jan 15, 2026

View reviewed changes

shafik reviewed Jan 24, 2026

View reviewed changes

ilovepi force-pushed the users/ilovepi/clang-trap-unreachable branch from b864741 to aa02129 Compare February 12, 2026 18:34

ilovepi temporarily deployed to main-branch-only February 12, 2026 18:35 — with GitHub Actions Inactive

ilovepi temporarily deployed to main-branch-only February 12, 2026 18:42 — with GitHub Actions Inactive

ilovepi added 4 commits February 19, 2026 15:36

Also support NoTrapAfterNoreturn

88d5d93

Fix formatting

a19bb6d

Fix release notes and help text

4f73bd4

ilovepi added 4 commits February 19, 2026 15:36

Fix punctuation in option description

508bc63

Make sure options are passed to the linker in LTO

67fe6ce

Fix formatting

3971f9b

Specify linker and triple for LTO checks

946a068

ilovepi force-pushed the users/ilovepi/clang-trap-unreachable branch from 34cabe0 to 946a068 Compare February 20, 2026 21:37

efriedma-quic mentioned this pull request May 13, 2026

[clang][X86] Always emit trap for noreturn functions with -fms-kernel #196550

Open

efriedma-quic mentioned this pull request May 22, 2026

[clang-cl][MSVC ABI] Vtable slot points to 1-byte NOP stub instead of out-of-class template override when base is dllimport and declaration omits virtual/override #199115

Open

ViNN280801 mentioned this pull request May 22, 2026

Teuchos: Add explicit virtual/override to templated unit test helper macros to fix clang-cl vtable bug trilinos/Trilinos#15283

Merged

Merge branch 'main' into users/ilovepi/clang-trap-unreachable

6d2a201

ilovepi requested review from MaskRay, efriedma-quic and shafik June 16, 2026 23:00

ilovepi requested a review from nikic June 17, 2026 16:33

		// RUN: %clang_cc1 %s -O1 -triple=x86_64-unknown-linux-gnu -ftrap-unreachable -S -o - 2>&1 \| FileCheck %s --check-prefix=TRAP
		// RUN: %clang_cc1 %s -O1 -triple=x86_64-unknown-linux-gnu -S -o - 2>&1 \| FileCheck %s --check-prefix=NOTRAP

Uh oh!

Conversation

ilovepi commented Jan 8, 2026

Uh oh!

ilovepi commented Jan 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

llvmbot commented Jan 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

llvmbot commented Jan 8, 2026

Uh oh!

github-actions Bot commented Jan 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ilovepi commented Jan 8, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

efriedma-quic left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

shafik left a comment

Choose a reason for hiding this comment

Uh oh!

ilovepi commented Feb 12, 2026

Uh oh!

github-actions Bot commented Feb 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🪟 Windows x64 Test Results

Uh oh!

eleviant commented May 18, 2026

Uh oh!

MaskRay commented Jun 17, 2026

Uh oh!

eleviant commented Jun 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

ilovepi commented Jan 8, 2026 •

edited

Loading

llvmbot commented Jan 8, 2026 •

edited

Loading

github-actions Bot commented Jan 8, 2026 •

edited

Loading

github-actions Bot commented Feb 12, 2026 •

edited

Loading