scanf: parse the 'm' length modifier, and check that the right arguments

are used with that and the 'a' length modifier.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@148029 91177308-0d34-0410-b5e6-96231b3b80d8

include/clang/Analysis/Analyses/FormatString.h

@@ -71,7 +71,8 @@ public:
     AsSizeT,      // 'z'
     AsPtrDiff,    // 't'
     AsLongDouble, // 'L'
-    AsAllocate,    // for '%as', GNU extension to C90 scanf
+    AsAllocate,   // for '%as', GNU extension to C90 scanf
+    AsMAllocate,  // for '%ms', GNU extension to scanf
     AsWideChar = AsLong // for '%ls', only makes sense for printf
   };
 

lib/Analysis/FormatString.cpp

@@ -212,6 +212,13 @@ clang::analyze_format_string::ParseLengthModifier(FormatSpecifier &FS,
         --I;
       }
       return false;
+    case 'm':
+      if (IsScanf) {
+        lmKind = LengthModifier::AsMAllocate;
+        ++I;
+        break;
+      }
+      return false;
   }
   LengthModifier lm(lmPosition, lmKind);
   FS.setLengthModifier(lm);
@@ -409,6 +416,8 @@ analyze_format_string::LengthModifier::toString() const {
     return "L";
   case AsAllocate:
     return "a";
+  case AsMAllocate:
+    return "m";
   case None:
     return "";
   }
@@ -555,6 +564,18 @@ bool FormatSpecifier::hasValidLengthModifier() const {
         default:
           return false;
       }
+
+    case LengthModifier::AsMAllocate:
+      switch (CS.getKind()) {
+        case ConversionSpecifier::cArg:
+        case ConversionSpecifier::CArg:
+        case ConversionSpecifier::sArg:
+        case ConversionSpecifier::SArg:
+        case ConversionSpecifier::ScanListArg:
+          return true;
+        default:
+          return false;
+      }
   }
   return false;
 }

lib/Analysis/PrintfFormatString.cpp

@@ -273,6 +273,7 @@ ArgTypeResult PrintfSpecifier::getArgType(ASTContext &Ctx) const {
       case LengthModifier::AsPtrDiff:
         return ArgTypeResult(Ctx.getPointerDiffType(), "ptrdiff_t");
       case LengthModifier::AsAllocate:
+      case LengthModifier::AsMAllocate:
         return ArgTypeResult::Invalid();
     }
 
@@ -294,6 +295,7 @@ ArgTypeResult PrintfSpecifier::getArgType(ASTContext &Ctx) const {
         // version of ptrdiff_t?
         return ArgTypeResult();
       case LengthModifier::AsAllocate:
+      case LengthModifier::AsMAllocate:
         return ArgTypeResult::Invalid();
     }
 

lib/Analysis/ScanfFormatString.cpp

@@ -220,6 +220,7 @@ ScanfArgTypeResult ScanfSpecifier::getArgType(ASTContext &Ctx) const {
           return ScanfArgTypeResult(Ctx.getPointerDiffType(), "ptrdiff_t *");
         case LengthModifier::AsLongDouble: return ScanfArgTypeResult::Invalid();
         case LengthModifier::AsAllocate: return ScanfArgTypeResult::Invalid();
+        case LengthModifier::AsMAllocate: return ScanfArgTypeResult::Invalid();
       }
 
     // Unsigned int.
@@ -243,6 +244,7 @@ ScanfArgTypeResult ScanfSpecifier::getArgType(ASTContext &Ctx) const {
           return ScanfArgTypeResult();
         case LengthModifier::AsLongDouble: return ScanfArgTypeResult::Invalid();
         case LengthModifier::AsAllocate: return ScanfArgTypeResult::Invalid();
+        case LengthModifier::AsMAllocate: return ScanfArgTypeResult::Invalid();
       }
 
     // Float.
@@ -271,15 +273,24 @@ ScanfArgTypeResult ScanfSpecifier::getArgType(ASTContext &Ctx) const {
         case LengthModifier::None: return ScanfArgTypeResult::CStrTy;
         case LengthModifier::AsLong:
           return ScanfArgTypeResult(ScanfArgTypeResult::WCStrTy, "wchar_t *");
+        case LengthModifier::AsAllocate:
+        case LengthModifier::AsMAllocate:
+          return ScanfArgTypeResult(ArgTypeResult::CStrTy);
         default:
           return ScanfArgTypeResult::Invalid();
       }
     case ConversionSpecifier::CArg:
     case ConversionSpecifier::SArg:
       // FIXME: Mac OS X specific?
-      if (LM.getKind() == LengthModifier::None)
-        return ScanfArgTypeResult(ScanfArgTypeResult::WCStrTy, "wchar_t *");
-      return ScanfArgTypeResult::Invalid();
+      switch (LM.getKind()) {
+        case LengthModifier::None:
+          return ScanfArgTypeResult(ScanfArgTypeResult::WCStrTy, "wchar_t *");
+        case LengthModifier::AsAllocate:
+        case LengthModifier::AsMAllocate:
+          return ScanfArgTypeResult(ArgTypeResult::WCStrTy, "wchar_t **");
+        default:
+          return ScanfArgTypeResult::Invalid();
+      }
 
     // Pointer.
     case ConversionSpecifier::pArg:

test/Sema/format-strings-c90.c

@@ -1,4 +1,4 @@
-/* RUN: %clang_cc1 -fsyntax-only -verify -pedantic -std=c89 %s
+/* RUN: %clang_cc1 -fsyntax-only -verify -triple i386-apple-darwin9 -pedantic -std=c89 %s
  */
 
 int scanf(const char * restrict, ...);
@@ -19,4 +19,9 @@ void foo(char **sp, float *fp, int *ip) {
   printf("%afoo", 1.0);
 
   scanf("%da", ip);
+
+  /* Test argument type check for the 'a' length modifier. */
+  scanf("%as", fp); /* expected-warning{{conversion specifies type 'char **' but the argument has type 'float *'}} */
+  scanf("%aS", fp); /* expected-warning{{conversion specifies type 'wchar_t **' (aka 'int **') but the argument has type 'float *'}} */
+  scanf("%a[abc]", fp); /* expected-warning{{conversion specifies type 'char **' but the argument has type 'float *'}} */
 }

test/Sema/format-strings-scanf.c

@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 -fsyntax-only -verify -Wformat-nonliteral %s
+// RUN: %clang_cc1 -fsyntax-only -verify -triple i386-apple-darwin9 -Wformat-nonliteral %s
 
 #include <stdarg.h>
 typedef __typeof(sizeof(int)) size_t;
@@ -73,11 +73,27 @@ void test_scanlist(int *ip, char *sp) {
   scanf("%h[abc]", sp); // expected-warning{{length modifier 'h' results in undefined behavior or no effect with '[' conversion specifier}}
 }
 
-void test_alloc_extension(char **sp, wchar_t **lsp) {
+void test_alloc_extension(char **sp, wchar_t **lsp, float *fp) {
   /* Make sure "%a" gets parsed as a conversion specifier for float,
    * even when followed by an 's', 'S' or '[', which would cause it to be
    * parsed as a length modifier in C90. */
   scanf("%as", sp); // expected-warning{{conversion specifies type 'float *' but the argument has type 'char **'}}
   scanf("%aS", lsp); // expected-warning{{conversion specifies type 'float *' but the argument has type 'wchar_t **'}}
   scanf("%a[bcd]", sp); // expected-warning{{conversion specifies type 'float *' but the argument has type 'char **'}}
+
+  // Test that the 'm' length modifier is only allowed with s, S, c, C or [.
+  // TODO: Warn that 'm' is an extension.
+  scanf("%ms", sp); // No warning.
+  scanf("%mS", lsp); // No warning.
+  scanf("%mc", sp); // No warning.
+  scanf("%mC", lsp); // No warning.
+  scanf("%m[abc]", sp); // No warning.
+  scanf("%md", sp); // expected-warning{{length modifier 'm' results in undefined behavior or no effect with 'd' conversion specifier}}
+
+  // Test argument type check for the 'm' length modifier.
+  scanf("%ms", fp); // expected-warning{{conversion specifies type 'char **' but the argument has type 'float *'}}
+  scanf("%mS", fp); // expected-warning{{conversion specifies type 'wchar_t **' (aka 'int **') but the argument has type 'float *'}}
+  scanf("%mc", fp); // expected-warning{{conversion specifies type 'char **' but the argument has type 'float *'}}
+  scanf("%mC", fp); // expected-warning{{conversion specifies type 'wchar_t **' (aka 'int **') but the argument has type 'float *'}}
+  scanf("%m[abc]", fp); // expected-warning{{conversion specifies type 'char **' but the argument has type 'float *'}}
 }