[UBSan] Support src:*=sanitize for multiple ignorelists. (llvm#141640)

See: llvm#139128 and
llvm#140529 for the background.

The introduction of these new tests (ubsan-src-ignorelist-category.test)
`-fsanitize-ignorelist=%t/src.ignorelist
-fsanitize-ignorelist=%t/src.ignorelist.contradict9` in this PR will not
lead to failures in the previous implementation (without this PR). This
is because the existing logic distinguishes between Sections in
different ignorelists, even if their names are identical. The order of
these Sections is preserved using a `vector`.

Author: qinkunbao

clang/include/clang/Basic/SanitizerSpecialCaseList.h

@@ -19,6 +19,7 @@
 #include "llvm/ADT/StringRef.h"
 #include "llvm/Support/SpecialCaseList.h"
 #include <memory>
+#include <utility>
 #include <vector>
 
 namespace llvm {
@@ -44,20 +45,24 @@ class SanitizerSpecialCaseList : public llvm::SpecialCaseList {
                  StringRef Category = StringRef()) const;
 
   // Query ignorelisted entries if any bit in Mask matches the entry's section.
-  // Return 0 if not found. If found, return the line number (starts with 1).
-  unsigned inSectionBlame(SanitizerMask Mask, StringRef Prefix, StringRef Query,
-                          StringRef Category = StringRef()) const;
+  // Return NotFound (0,0) if not found. If found, return the file index number
+  // and the line number (FileIdx, LineNo) (FileIdx starts with 1 and LineNo
+  // starts with 0).
+  std::pair<unsigned, unsigned>
+  inSectionBlame(SanitizerMask Mask, StringRef Prefix, StringRef Query,
+                 StringRef Category = StringRef()) const;
 
 protected:
   // Initialize SanitizerSections.
   void createSanitizerSections();
 
   struct SanitizerSection {
-    SanitizerSection(SanitizerMask SM, SectionEntries &E)
-        : Mask(SM), Entries(E) {};
+    SanitizerSection(SanitizerMask SM, SectionEntries &E, unsigned idx)
+        : Mask(SM), Entries(E), FileIdx(idx) {};
 
     SanitizerMask Mask;
     SectionEntries &Entries;
+    unsigned FileIdx;
   };
 
   std::vector<SanitizerSection> SanitizerSections;

clang/lib/Basic/NoSanitizeList.cpp

@@ -44,13 +44,11 @@ bool NoSanitizeList::containsFunction(SanitizerMask Mask,
 
 bool NoSanitizeList::containsFile(SanitizerMask Mask, StringRef FileName,
                                   StringRef Category) const {
-  unsigned NoSanLine = SSCL->inSectionBlame(Mask, "src", FileName, Category);
-  if (NoSanLine == 0)
+  auto NoSan = SSCL->inSectionBlame(Mask, "src", FileName, Category);
+  if (NoSan == llvm::SpecialCaseList::NotFound)
     return false;
-  unsigned SanLine = SSCL->inSectionBlame(Mask, "src", FileName, "sanitize");
-  // If we have two cases such as `src:a.cpp=sanitize` and `src:a.cpp`, the
-  // current entry override the previous entry.
-  return !SanLine || NoSanLine > SanLine;
+  auto San = SSCL->inSectionBlame(Mask, "src", FileName, "sanitize");
+  return San == llvm::SpecialCaseList::NotFound || NoSan > San;
 }
 
 bool NoSanitizeList::containsMainFile(SanitizerMask Mask, StringRef FileName,

clang/lib/Basic/SanitizerSpecialCaseList.cpp

@@ -50,27 +50,27 @@ void SanitizerSpecialCaseList::createSanitizerSections() {
 #undef SANITIZER
 #undef SANITIZER_GROUP
 
-    SanitizerSections.emplace_back(Mask, S.Entries);
+    SanitizerSections.emplace_back(Mask, S.Entries, S.FileIdx);
   }
 }
 
 bool SanitizerSpecialCaseList::inSection(SanitizerMask Mask, StringRef Prefix,
                                          StringRef Query,
                                          StringRef Category) const {
-  return inSectionBlame(Mask, Prefix, Query, Category);
+  return inSectionBlame(Mask, Prefix, Query, Category) != NotFound;
 }
 
-unsigned SanitizerSpecialCaseList::inSectionBlame(SanitizerMask Mask,
-                                                  StringRef Prefix,
-                                                  StringRef Query,
-                                                  StringRef Category) const {
+std::pair<unsigned, unsigned>
+SanitizerSpecialCaseList::inSectionBlame(SanitizerMask Mask, StringRef Prefix,
+                                         StringRef Query,
+                                         StringRef Category) const {
   for (const auto &S : llvm::reverse(SanitizerSections)) {
     if (S.Mask & Mask) {
-      unsigned lineNum =
+      unsigned LineNum =
           SpecialCaseList::inSectionBlame(S.Entries, Prefix, Query, Category);
-      if (lineNum > 0)
-        return lineNum;
+      if (LineNum > 0)
+        return {S.FileIdx, LineNum};
     }
   }
-  return 0;
+  return NotFound;
 }

clang/test/CodeGen/ubsan-src-ignorelist-category.test

@@ -12,6 +12,9 @@
 // RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow -fsanitize-ignorelist=%t/src.ignorelist.contradict6 -emit-llvm %t/test1.c -o - | FileCheck %s --check-prefixes=CHECK1,SANITIZE
 // RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow -fsanitize-ignorelist=%t/src.ignorelist.contradict7 -emit-llvm %t/test1.c -o - | FileCheck %s --check-prefixes=CHECK1,IGNORE
 // RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow -fsanitize-ignorelist=%t/src.ignorelist.contradict8 -emit-llvm %t/test1.c -o - | FileCheck %s --check-prefixes=CHECK1,SANITIZE
+// RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow -fsanitize-ignorelist=%t/src.ignorelist -fsanitize-ignorelist=%t/src.ignorelist.contradict9 -emit-llvm %t/test1.c -o - | FileCheck %s --check-prefixes=CHECK1,IGNORE
+// RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow -fsanitize-ignorelist=%t/src.ignorelist.contradict9 -fsanitize-ignorelist=%t/src.ignorelist -emit-llvm %t/test1.c -o - | FileCheck %s --check-prefixes=CHECK1,SANITIZE
+
 
 // Verify ubsan only emits checks for files in the allowlist
 
@@ -77,6 +80,14 @@ src:*/tes*1.c=sanitize
 src:*/te*t1.c
 src:*/t*st1.c=sanitize
 
+//--- src.ignorelist.contradict9
+src:*
+src:*/test1.c=sanitize
+src:*/test1.c
+src:*/test1.c=sanitize
+src:*/te*t1.c
+src:*/test*.c
+
 
 //--- test1.c
 // CHECK1-LABEL: define dso_local i32 @add