Fix a heap buffer overflow issue in CPDF_CMap::GetNextChar

Add a check to make sure offset is less than the size of string in the function of GetNextChar().

BUG=471651
R=tsepez@chromium.org

Review URL: https://codereview.chromium.org/1067073003

7 files changed, 73 insertions, 96 deletions

diff --git a/core/include/fpdfapi/fpdf_resource.h b/core/include/fpdfapi/fpdf_resource.h
index 54e1c97ee7..3050985567 100644
--- a/core/include/fpdfapi/fpdf_resource.h
+++ b/core/include/fpdfapi/fpdf_resource.h
@@ -164,11 +164,13 @@ public:
         return m_Font.GetFace();
     }
 
-
-
-    virtual FX_DWORD		GetNextChar(FX_LPCSTR pString, int& offset) const
+    virtual FX_DWORD		GetNextChar(FX_LPCSTR pString, int nStrLen, int& offset) const
     {
-        return (FX_BYTE)pString[offset++];
+        if (offset < 0 || nStrLen < 1) {
+            return 0;
+        }
+        FX_BYTE ch = offset < nStrLen ? pString[offset++] : pString[nStrLen-1];
+        return static_cast<FX_DWORD>(ch);
     }
 
     virtual int				CountChar(FX_LPCSTR pString, int size) const
@@ -512,99 +514,74 @@ public:
 
     virtual ~CPDF_CIDFont();
 
-    FX_BOOL					LoadGB2312();
-    virtual int				GlyphFromCharCode(FX_DWORD charcode, FX_BOOL *pVertGlyph = NULL);
-    virtual int				GetCharWidthF(FX_DWORD charcode, int level = 0);
-    virtual void			GetCharBBox(FX_DWORD charcode, FX_RECT& rect, int level = 0);
-
-    FX_WORD					CIDFromCharCode(FX_DWORD charcode) const;
+    FX_BOOL                 LoadGB2312();
+    virtual int             GlyphFromCharCode(FX_DWORD charcode, FX_BOOL *pVertGlyph = NULL);
+    virtual int             GetCharWidthF(FX_DWORD charcode, int level = 0);
+    virtual void            GetCharBBox(FX_DWORD charcode, FX_RECT& rect, int level = 0);
+    FX_WORD                 CIDFromCharCode(FX_DWORD charcode) const;
 
-    FX_BOOL					IsTrueType()
+    FX_BOOL                 IsTrueType()
     {
         return !m_bType1;
     }
 
+    virtual FX_DWORD        GetNextChar(const FX_LPCSTR pString, int nStrLen, int& offset) const override;
+    virtual int             CountChar(const FX_LPCSTR pString, int size) const;
+    virtual int             AppendChar(FX_LPSTR str, FX_DWORD charcode) const;
+    virtual int             GetCharSize(FX_DWORD charcode) const;
 
-    virtual FX_DWORD		GetNextChar(const FX_CHAR* pString, int& offset) const;
-    virtual int				CountChar(const FX_CHAR* pString, int size) const;
-    virtual int				AppendChar(FX_LPSTR str, FX_DWORD charcode) const;
-    virtual int				GetCharSize(FX_DWORD charcode) const;
-
-
-    int						GetCharset() const
+    int                     GetCharset() const
     {
         return m_Charset;
     }
 
-    FX_LPCBYTE				GetCIDTransform(FX_WORD CID) const;
-
-
+    FX_LPCBYTE              GetCIDTransform(FX_WORD CID) const;
+    virtual FX_BOOL         IsVertWriting() const;
+    short                   GetVertWidth(FX_WORD CID) const;
+    void                    GetVertOrigin(FX_WORD CID, short& vx, short& vy) const;
+    virtual FX_BOOL         IsUnicodeCompatible() const;
+    virtual FX_BOOL         IsFontStyleFromCharCode(FX_DWORD charcode) const;
 
-    virtual FX_BOOL			IsVertWriting() const;
-
-    short					GetVertWidth(FX_WORD CID) const;
-
-    void					GetVertOrigin(FX_WORD CID, short& vx, short& vy) const;
-
-    virtual FX_BOOL			IsUnicodeCompatible() const;
-    virtual FX_BOOL			IsFontStyleFromCharCode(FX_DWORD charcode) const;
 protected:
-    friend class			CPDF_Font;
-    virtual FX_BOOL			_Load();
-    virtual FX_WCHAR		_UnicodeFromCharCode(FX_DWORD charcode) const;
-    virtual FX_DWORD		_CharCodeFromUnicode(FX_WCHAR Unicode) const;
-    int				GetGlyphIndex(FX_DWORD unicodeb, FX_BOOL *pVertGlyph);
-
-    CPDF_CMap*				m_pCMap;
-
-    CPDF_CMap*				m_pAllocatedCMap;
-
-    CPDF_CID2UnicodeMap*	m_pCID2UnicodeMap;
-
-    int						m_Charset;
-
-    FX_BOOL					m_bType1;
-
-    CPDF_StreamAcc*			m_pCIDToGIDMap;
-    FX_BOOL					m_bCIDIsGID;
-
-
-
-    FX_WORD					m_DefaultWidth;
-
-    FX_WORD*				m_pAnsiWidths;
-
-    FX_SMALL_RECT			m_CharBBox[256];
-
-    CFX_DWordArray			m_WidthList;
-
-    short					m_DefaultVY;
-
-    short					m_DefaultW1;
-
-    CFX_DWordArray			m_VertMetrics;
-
-
-    void					LoadMetricsArray(CPDF_Array* pArray, CFX_DWordArray& result, int nElements);
-
-    void					LoadSubstFont();
+    friend class            CPDF_Font;
+    virtual FX_BOOL         _Load();
+    virtual FX_WCHAR        _UnicodeFromCharCode(FX_DWORD charcode) const;
+    virtual FX_DWORD        _CharCodeFromUnicode(FX_WCHAR Unicode) const;
+    int                     GetGlyphIndex(FX_DWORD unicodeb, FX_BOOL *pVertGlyph);
+
+    CPDF_CMap*              m_pCMap;
+    CPDF_CMap*              m_pAllocatedCMap;
+    CPDF_CID2UnicodeMap*    m_pCID2UnicodeMap;
+    int                     m_Charset;
+    FX_BOOL                 m_bType1;
+    CPDF_StreamAcc*         m_pCIDToGIDMap;
+    FX_BOOL                 m_bCIDIsGID;
+    FX_WORD                 m_DefaultWidth;
+    FX_WORD*                m_pAnsiWidths;
+    FX_SMALL_RECT           m_CharBBox[256];
+    CFX_DWordArray          m_WidthList;
+    short                   m_DefaultVY;
+    short                   m_DefaultW1;
+    CFX_DWordArray          m_VertMetrics;
+
+    void                    LoadMetricsArray(CPDF_Array* pArray, CFX_DWordArray& result, int nElements);
+    void                    LoadSubstFont();
 
     FX_BOOL					m_bAdobeCourierStd;
-
-    CFX_CTTGSUBTable*			m_pTTGSUBTable;
+    CFX_CTTGSUBTable*       m_pTTGSUBTable;
 };
-#define PDFCS_DEVICEGRAY		1
-
-#define PDFCS_DEVICERGB			2
-#define PDFCS_DEVICECMYK		3
-#define PDFCS_CALGRAY			4
-#define PDFCS_CALRGB			5
-#define PDFCS_LAB				6
-#define PDFCS_ICCBASED			7
-#define PDFCS_SEPARATION		8
-#define PDFCS_DEVICEN			9
-#define PDFCS_INDEXED			10
-#define PDFCS_PATTERN			11
+
+#define PDFCS_DEVICEGRAY        1
+#define PDFCS_DEVICERGB         2
+#define PDFCS_DEVICECMYK        3
+#define PDFCS_CALGRAY           4
+#define PDFCS_CALRGB            5
+#define PDFCS_LAB               6
+#define PDFCS_ICCBASED          7
+#define PDFCS_SEPARATION        8
+#define PDFCS_DEVICEN           9
+#define PDFCS_INDEXED           10
+#define PDFCS_PATTERN           11
 class CPDF_ColorSpace : public CFX_Object
 {
 public:
diff --git a/core/src/fpdfapi/fpdf_font/font_int.h b/core/src/fpdfapi/fpdf_font/font_int.h
index 6048ba981d..43f4e302af 100644
--- a/core/src/fpdfapi/fpdf_font/font_int.h
+++ b/core/src/fpdfapi/fpdf_font/font_int.h
@@ -95,8 +95,8 @@ public:
     FX_WORD					CIDFromCharCode(FX_DWORD charcode) const;
     FX_DWORD				CharCodeFromCID(FX_WORD CID) const;
     int						GetCharSize(FX_DWORD charcode) const;
-    FX_DWORD				GetNextChar(const FX_CHAR* pString, int& offset) const;
-    int						CountChar(const FX_CHAR* pString, int size) const;
+    FX_DWORD				GetNextChar(FX_LPCSTR pString, int nStrLen, int& offset) const;
+    int						CountChar(FX_LPCSTR pString, int size) const;
     int						AppendChar(FX_LPSTR str, FX_DWORD charcode) const;
     typedef enum {OneByte, TwoBytes, MixedTwoBytes, MixedFourBytes} CodingScheme;
 protected:
diff --git a/core/src/fpdfapi/fpdf_font/fpdf_font.cpp b/core/src/fpdfapi/fpdf_font/fpdf_font.cpp
index dd646ca113..41bb95a89d 100644
--- a/core/src/fpdfapi/fpdf_font/fpdf_font.cpp
+++ b/core/src/fpdfapi/fpdf_font/fpdf_font.cpp
@@ -213,7 +213,7 @@ CFX_WideString CPDF_Font::DecodeString(const CFX_ByteString& str) const
     FX_LPCSTR src_buf = str;
     int src_pos = 0;
     while (src_pos < src_len) {
-        FX_DWORD charcode = GetNextChar(src_buf, src_pos);
+        FX_DWORD charcode = GetNextChar(src_buf, src_len, src_pos);
         CFX_WideString unicode = UnicodeFromCharCode(charcode);
         if (!unicode.IsEmpty()) {
             result += unicode;
@@ -379,7 +379,7 @@ int CPDF_Font::GetStringWidth(FX_LPCSTR pString, int size)
     int offset = 0;
     int width = 0;
     while (offset < size) {
-        FX_DWORD charcode = GetNextChar(pString, offset);
+        FX_DWORD charcode = GetNextChar(pString, size, offset);
         width += GetCharWidthF(charcode);
     }
     return width;
diff --git a/core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp b/core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp
index 1ce91f9ed6..0d193e2189 100644
--- a/core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp
+++ b/core/src/fpdfapi/fpdf_font/fpdf_font_cid.cpp
@@ -491,7 +491,7 @@ static int _CheckCodeRange(FX_LPBYTE codes, int size, _CMap_CodeRange* pRanges,
     }
     return 0;
 }
-FX_DWORD CPDF_CMap::GetNextChar(FX_LPCSTR pString, int& offset) const
+FX_DWORD CPDF_CMap::GetNextChar(FX_LPCSTR pString, int nStrLen, int& offset) const
 {
     switch (m_CodingScheme) {
         case OneByte:
@@ -524,7 +524,7 @@ FX_DWORD CPDF_CMap::GetNextChar(FX_LPCSTR pString, int& offset) const
                         }
                         return charcode;
                     }
-                    if (char_size == 4) {
+                    if (char_size == 4 || offset == nStrLen) {
                         return 0;
                     }
                     codes[char_size ++] = ((FX_LPBYTE)pString)[offset++];
@@ -576,7 +576,7 @@ int CPDF_CMap::CountChar(FX_LPCSTR pString, int size) const
         case MixedFourBytes: {
                 int count = 0, offset = 0;
                 while (offset < size) {
-                    GetNextChar(pString, offset);
+                    GetNextChar(pString, size, offset);
                     count ++;
                 }
                 return count;
@@ -1317,9 +1317,9 @@ int CPDF_CIDFont::GlyphFromCharCode(FX_DWORD charcode, FX_BOOL *pVertGlyph)
     FX_LPCBYTE pdata = m_pCIDToGIDMap->GetData() + byte_pos;
     return pdata[0] * 256 + pdata[1];
 }
-FX_DWORD CPDF_CIDFont::GetNextChar(FX_LPCSTR pString, int& offset) const
+FX_DWORD CPDF_CIDFont::GetNextChar(FX_LPCSTR pString, int nStrLen, int& offset) const
 {
-    return m_pCMap->GetNextChar(pString, offset);
+    return m_pCMap->GetNextChar(pString, nStrLen, offset);
 }
 int CPDF_CIDFont::GetCharSize(FX_DWORD charcode) const
 {
diff --git a/core/src/fpdfapi/fpdf_page/fpdf_page.cpp b/core/src/fpdfapi/fpdf_page/fpdf_page.cpp
index 590a01aa6d..ccdfb9fcbb 100644
--- a/core/src/fpdfapi/fpdf_page/fpdf_page.cpp
+++ b/core/src/fpdfapi/fpdf_page/fpdf_page.cpp
@@ -247,7 +247,7 @@ void CPDF_TextObject::SetSegments(const CFX_ByteString* pStrs, FX_FLOAT* pKernin
             FX_LPCSTR segment = pStrs[i];
             int offset = 0, len = pStrs[i].GetLength();
             while (offset < len) {
-                m_pCharCodes[index++] = pFont->GetNextChar(segment, offset);
+                m_pCharCodes[index++] = pFont->GetNextChar(segment, len, offset);
             }
             if (i != nsegs - 1) {
                 m_pCharPos[index - 1] = pKerning[i];
@@ -256,7 +256,7 @@ void CPDF_TextObject::SetSegments(const CFX_ByteString* pStrs, FX_FLOAT* pKernin
         }
     } else {
         int offset = 0;
-        m_pCharCodes = (FX_DWORD*)(FX_UINTPTR)pFont->GetNextChar(pStrs[0], offset);
+        m_pCharCodes = (FX_DWORD*)(FX_UINTPTR)pFont->GetNextChar(pStrs[0], pStrs[0].GetLength(), offset);
     }
 }
 void CPDF_TextObject::SetText(const CFX_ByteString& str)
diff --git a/core/src/fpdfapi/fpdf_render/fpdf_render_text.cpp b/core/src/fpdfapi/fpdf_render/fpdf_render_text.cpp
index f99f7cec38..0ea7ea1672 100644
--- a/core/src/fpdfapi/fpdf_render/fpdf_render_text.cpp
+++ b/core/src/fpdfapi/fpdf_render/fpdf_render_text.cpp
@@ -613,7 +613,7 @@ void CPDF_TextRenderer::DrawTextString(CFX_RenderDevice* pDevice, FX_FLOAT origi
     FX_DWORD* pCharCodes;
     FX_FLOAT* pCharPos;
     if (nChars == 1) {
-        charcode = pFont->GetNextChar(str, offset);
+        charcode = pFont->GetNextChar(str, str.GetLength(), offset);
         pCharCodes = (FX_DWORD*)(FX_UINTPTR)charcode;
         pCharPos = NULL;
     } else {
@@ -621,7 +621,7 @@ void CPDF_TextRenderer::DrawTextString(CFX_RenderDevice* pDevice, FX_FLOAT origi
         pCharPos = FX_Alloc(FX_FLOAT, nChars - 1);
         FX_FLOAT cur_pos = 0;
         for (int i = 0; i < nChars; i ++) {
-            pCharCodes[i] = pFont->GetNextChar(str, offset);
+            pCharCodes[i] = pFont->GetNextChar(str, str.GetLength(), offset);
             if (i) {
                 pCharPos[i - 1] = cur_pos;
             }
diff --git a/core/src/fpdftext/fpdf_text.cpp b/core/src/fpdftext/fpdf_text.cpp
index a4a124d5e6..d6d6de9a83 100644
--- a/core/src/fpdftext/fpdf_text.cpp
+++ b/core/src/fpdftext/fpdf_text.cpp
@@ -181,7 +181,7 @@ CTextBaseLine* CTextPage::InsertTextBox(CTextBaseLine* pBaseLine, FX_FLOAT basey
     FX_LPCSTR pStr = str;
     int len = str.GetLength(), offset = 0;
     while (offset < len) {
-        FX_DWORD ch = pFont->GetNextChar(pStr, offset);
+        FX_DWORD ch = pFont->GetNextChar(pStr, len, offset);
         CFX_WideString unicode_str = pFont->UnicodeFromCharCode(ch);
         if (unicode_str.IsEmpty()) {
             text += (FX_WCHAR)ch;