Sanitizer: Support -fwrapv with -fsanitize=signed-integer-overflow

[JustinStitt]

Reasoning
Clang has a signed-integer-overflow sanitizer to catch arithmetic overflow; however, most of its instrumentation fails to apply when -fwrapv is enabled; this is by design.

The Linux kernel enables -fno-strict-overflow which implies -fwrapv. This means we are currently unable to detect signed-integer wrap-around. All the while, the root cause of many security vulnerabilities in the Linux kernel is arithmetic overflow.

To work around this and enhance the functionality of -fsanitize=signed-integer-overflow, let's instrument signed arithmetic even if the signed overflow behavior is defined.

Initially, I created a new sanitizer @ (pr/80089) but simply changing the SIO sanitizer itself may be the better approach as per @MaskRay 's review from that PR.

cc: @nickdesaulniers @kees @nathanchance @bwendling @MaskRay @vitalybuka

[github-actions]

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be
notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write
permissions for the repository. In which case you can instead tag reviewers by
name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review
by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate
is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

[llvmbot]

@llvm/pr-subscribers-clang

@llvm/pr-subscribers-clang-codegen

Author: Justin Stitt (JustinStitt)

Changes

Reasoning
Clang has a signed-integer-overflow sanitizer to catch arithmetic overflow; however, most of its instrumentation fails to apply when -fwrapv is enabled; this is by design.

The Linux kernel enables -fno-strict-overflow which implies -fwrapv. This means we are currently unable to detect signed-integer wrap-around. All the while, the root cause of many security vulnerabilities in the Linux kernel is arithmetic overflow.

To work around this and enhance the functionality of -fsanitize=signed-integer-overflow, let's instrument signed arithmetic even if the signed overflow behavior is defined.

Initially, I created a new sanitizer @ (pr/80089) but simply changing the SIO sanitizer itself may be the better approach as per @MaskRay 's review from that PR.

cc: @nickdesaulniers @kees @nathanchance @bwendling @MaskRay

---

Full diff: https://github.com/llvm/llvm-project/pull/82432.diff

4 Files Affected:

• (modified) clang/docs/ReleaseNotes.rst (+3)
• (modified) clang/docs/UndefinedBehaviorSanitizer.rst (+5-4)
• (modified) clang/lib/CodeGen/CGExprScalar.cpp (+12-4)
• (modified) clang/test/CodeGen/integer-overflow.c (+12)

diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index 5bca2c965c866b..685b19cabeb82c 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -399,6 +399,9 @@ Moved checkers
 Sanitizers
 ----------
 
+- ``-fsanitize=signed-integer-overflow`` now instruments signed arithmetic even
+  when ``-fwrapv`` is enabled. Previously, only division checks were enabled.
+
 Python Binding Changes
 ----------------------
 
diff --git a/clang/docs/UndefinedBehaviorSanitizer.rst b/clang/docs/UndefinedBehaviorSanitizer.rst
index b8ad3804f18903..8f58c92bd2a163 100644
--- a/clang/docs/UndefinedBehaviorSanitizer.rst
+++ b/clang/docs/UndefinedBehaviorSanitizer.rst
@@ -190,10 +190,11 @@ Available checks are:
   -  ``-fsanitize=signed-integer-overflow``: Signed integer overflow, where the
      result of a signed integer computation cannot be represented in its type.
      This includes all the checks covered by ``-ftrapv``, as well as checks for
-     signed division overflow (``INT_MIN/-1``), but not checks for
-     lossy implicit conversions performed before the computation
-     (see ``-fsanitize=implicit-conversion``). Both of these two issues are
-     handled by ``-fsanitize=implicit-conversion`` group of checks.
+     signed division overflow (``INT_MIN/-1``). Note that checks are still
+     added even when ``-fwrapv`` is enabled. This sanitizer does not check for
+     lossy implicit conversions performed before the computation (see
+     ``-fsanitize=implicit-conversion``). Both of these two issues are handled
+     by ``-fsanitize=implicit-conversion`` group of checks.
   -  ``-fsanitize=unreachable``: If control flow reaches an unreachable
      program point.
   -  ``-fsanitize=unsigned-integer-overflow``: Unsigned integer overflow, where
diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp
index 576734e460b9c1..7621d9bcdec991 100644
--- a/clang/lib/CodeGen/CGExprScalar.cpp
+++ b/clang/lib/CodeGen/CGExprScalar.cpp
@@ -723,7 +723,9 @@ class ScalarExprEmitter
     if (Ops.Ty->isSignedIntegerOrEnumerationType()) {
       switch (CGF.getLangOpts().getSignedOverflowBehavior()) {
       case LangOptions::SOB_Defined:
-        return Builder.CreateMul(Ops.LHS, Ops.RHS, "mul");
+        if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
+              return Builder.CreateMul(Ops.LHS, Ops.RHS, "mul");
+        [[fallthrough]];
       case LangOptions::SOB_Undefined:
         if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
           return Builder.CreateNSWMul(Ops.LHS, Ops.RHS, "mul");
@@ -2568,7 +2570,9 @@ llvm::Value *ScalarExprEmitter::EmitIncDecConsiderOverflowBehavior(
   StringRef Name = IsInc ? "inc" : "dec";
   switch (CGF.getLangOpts().getSignedOverflowBehavior()) {
   case LangOptions::SOB_Defined:
-    return Builder.CreateAdd(InVal, Amount, Name);
+    if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
+        return Builder.CreateAdd(InVal, Amount, Name);
+    [[fallthrough]];
   case LangOptions::SOB_Undefined:
     if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
       return Builder.CreateNSWAdd(InVal, Amount, Name);
@@ -3913,7 +3917,9 @@ Value *ScalarExprEmitter::EmitAdd(const BinOpInfo &op) {
   if (op.Ty->isSignedIntegerOrEnumerationType()) {
     switch (CGF.getLangOpts().getSignedOverflowBehavior()) {
     case LangOptions::SOB_Defined:
-      return Builder.CreateAdd(op.LHS, op.RHS, "add");
+      if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
+        return Builder.CreateAdd(op.LHS, op.RHS, "add");
+      [[fallthrough]];
     case LangOptions::SOB_Undefined:
       if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
         return Builder.CreateNSWAdd(op.LHS, op.RHS, "add");
@@ -4067,7 +4073,9 @@ Value *ScalarExprEmitter::EmitSub(const BinOpInfo &op) {
     if (op.Ty->isSignedIntegerOrEnumerationType()) {
       switch (CGF.getLangOpts().getSignedOverflowBehavior()) {
       case LangOptions::SOB_Defined:
-        return Builder.CreateSub(op.LHS, op.RHS, "sub");
+        if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
+          return Builder.CreateSub(op.LHS, op.RHS, "sub");
+        [[fallthrough]];
       case LangOptions::SOB_Undefined:
         if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
           return Builder.CreateNSWSub(op.LHS, op.RHS, "sub");
diff --git a/clang/test/CodeGen/integer-overflow.c b/clang/test/CodeGen/integer-overflow.c
index 9a3107c0b52926..b6d1a9fc4cf744 100644
--- a/clang/test/CodeGen/integer-overflow.c
+++ b/clang/test/CodeGen/integer-overflow.c
@@ -2,6 +2,7 @@
 // RUN: %clang_cc1 -triple x86_64-apple-darwin %s -emit-llvm -o - -fwrapv | FileCheck %s --check-prefix=WRAPV
 // RUN: %clang_cc1 -triple x86_64-apple-darwin %s -emit-llvm -o - -ftrapv | FileCheck %s --check-prefix=TRAPV
 // RUN: %clang_cc1 -triple x86_64-apple-darwin %s -emit-llvm -o - -fsanitize=signed-integer-overflow | FileCheck %s --check-prefix=CATCH_UB
+// RUN: %clang_cc1 -triple x86_64-apple-darwin %s -emit-llvm -o - -fsanitize=signed-integer-overflow -fwrapv | FileCheck %s --check-prefix=CATCH_WRAP
 // RUN: %clang_cc1 -triple x86_64-apple-darwin %s -emit-llvm -o - -ftrapv -ftrapv-handler foo | FileCheck %s --check-prefix=TRAPV_HANDLER
 
 
@@ -16,6 +17,7 @@ void test1(void) {
   // WRAPV: add i32
   // TRAPV: llvm.sadd.with.overflow.i32
   // CATCH_UB: llvm.sadd.with.overflow.i32
+  // CATCH_WRAP: llvm.sadd.with.overflow.i32
   // TRAPV_HANDLER: foo(
   f11G = a + b;
   
@@ -23,6 +25,7 @@ void test1(void) {
   // WRAPV: sub i32
   // TRAPV: llvm.ssub.with.overflow.i32
   // CATCH_UB: llvm.ssub.with.overflow.i32
+  // CATCH_WRAP: llvm.ssub.with.overflow.i32
   // TRAPV_HANDLER: foo(
   f11G = a - b;
   
@@ -30,6 +33,7 @@ void test1(void) {
   // WRAPV: mul i32
   // TRAPV: llvm.smul.with.overflow.i32
   // CATCH_UB: llvm.smul.with.overflow.i32
+  // CATCH_WRAP: llvm.smul.with.overflow.i32
   // TRAPV_HANDLER: foo(
   f11G = a * b;
 
@@ -37,6 +41,7 @@ void test1(void) {
   // WRAPV: sub i32 0, 
   // TRAPV: llvm.ssub.with.overflow.i32(i32 0
   // CATCH_UB: llvm.ssub.with.overflow.i32(i32 0
+  // CATCH_WRAP: llvm.ssub.with.overflow.i32(i32 0
   // TRAPV_HANDLER: foo(
   f11G = -a;
   
@@ -46,6 +51,7 @@ void test1(void) {
   // WRAPV: add i32 {{.*}}, 1
   // TRAPV: llvm.sadd.with.overflow.i32({{.*}}, i32 1)
   // CATCH_UB: llvm.sadd.with.overflow.i32({{.*}}, i32 1)
+  // CATCH_WRAP: llvm.sadd.with.overflow.i32({{.*}}, i32 1)
   // TRAPV_HANDLER: foo(
   ++a;
   
@@ -53,6 +59,7 @@ void test1(void) {
   // WRAPV: add i32 {{.*}}, -1
   // TRAPV: llvm.ssub.with.overflow.i32({{.*}}, i32 1)
   // CATCH_UB: llvm.ssub.with.overflow.i32({{.*}}, i32 1)
+  // CATCH_WRAP: llvm.ssub.with.overflow.i32({{.*}}, i32 1)
   // TRAPV_HANDLER: foo(
   --a;
   
@@ -70,6 +77,7 @@ void test1(void) {
   // WRAPV: add i8 {{.*}}, 1
   // TRAPV: add i8 {{.*}}, 1
   // CATCH_UB: add i8 {{.*}}, 1
+  // CATCH_WRAP: add i8 {{.*}}, 1
   ++PR9350_char_inc;
 
   // PR9350: char pre-decrement never overflows.
@@ -78,6 +86,7 @@ void test1(void) {
   // WRAPV: add i8 {{.*}}, -1
   // TRAPV: add i8 {{.*}}, -1
   // CATCH_UB: add i8 {{.*}}, -1
+  // CATCH_WRAP: add i8 {{.*}}, -1
   --PR9350_char_dec;
 
   // PR9350: short pre-increment never overflows.
@@ -86,6 +95,7 @@ void test1(void) {
   // WRAPV: add i16 {{.*}}, 1
   // TRAPV: add i16 {{.*}}, 1
   // CATCH_UB: add i16 {{.*}}, 1
+  // CATCH_WRAP: add i16 {{.*}}, 1
   ++PR9350_short_inc;
 
   // PR9350: short pre-decrement never overflows.
@@ -94,6 +104,7 @@ void test1(void) {
   // WRAPV: add i16 {{.*}}, -1
   // TRAPV: add i16 {{.*}}, -1
   // CATCH_UB: add i16 {{.*}}, -1
+  // CATCH_WRAP: add i16 {{.*}}, -1
   --PR9350_short_dec;
 
   // PR24256: don't instrument __builtin_frame_address.
@@ -102,4 +113,5 @@ void test1(void) {
   // WRAPV:    call ptr @llvm.frameaddress.p0(i32 0)
   // TRAPV:    call ptr @llvm.frameaddress.p0(i32 0)
   // CATCH_UB: call ptr @llvm.frameaddress.p0(i32 0)
+  // CATCH_WRAP: call ptr @llvm.frameaddress.p0(i32 0)
 }

[github-actions]

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

[JustinStitt]

⚠️ C/C++ code formatter, clang-format found issues in your code. ⚠️

You can test this locally with the following command:
View the diff from clang-format here.

Fixed in e5e92e6

[MaskRay]

Thanks! This does look simpler than -fsanitize=signed-integer-wrap

[MaskRay]

@efriedma-quic @rjmccall

[vitalybuka]

LGTM

[kees]

I've tested this and it's working as expected with the Linux kernel -next tree with CONFIG_UBSAN_SIGNED_WRAP=y. It passes test_ubsan.ko installation and trigger other expected build time and run time warnings. Yay!

[kees]

Working as expected for me!

[github-actions]

@JustinStitt Congratulations on having your first Pull Request (PR) merged into the LLVM Project!

Your changes will be combined with recent changes from other authors, then tested
by our build bots. If there is a problem with a build, you may recieve a report in an email or a comment on this PR.

Please check whether problems have been caused by your change specifically, as
the builds can include changes from many authors. It is not uncommon for your
change to be included in a build that fails due to someone else's changes, or
infrastructure issues.

How to do this, and the rest of the post-merge process, is covered in detail here.

If your change does cause a problem, it may be reverted, or you can revert it yourself.
This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again.

If you don't get any reports, no action is required from you. Your changes are working as expected, well done!