2 files changed, 48 insertions, 43 deletions

diff --git a/ChangeLog b/ChangeLog
index e0a5cb3..2f2416d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,6 @@
+Fri 23 Aug 2019 09:34:35 AM CEST
+        Fix invalid read for malformed DVI files (#5846). -CG
+
 Mon 29 Jul 2019 05:57:04 PM CEST
         Fix build issue with exiv2-0.27 (#5820). -AS (via Gentoo)
 
diff --git a/src/plugins/dvi_extractor.c b/src/plugins/dvi_extractor.c
index 268b48c..e3aa450 100644
--- a/src/plugins/dvi_extractor.c
+++ b/src/plugins/dvi_extractor.c
@@ -1,6 +1,6 @@
 /*
      This file is part of libextractor.
-     Copyright (C) 2002, 2003, 2004, 2012, 2017 Vidyut Samanta and Christian Grothoff
+     Copyright (C) 2002, 2003, 2004, 2012, 2017, 2019 Vidyut Samanta and Christian Grothoff
 
      libextractor is free software; you can redistribute it and/or modify
      it under the terms of the GNU General Public License as published
@@ -182,6 +182,8 @@ EXTRACTOR_dvi_extract_method (struct EXTRACTOR_ExtractContext *ec)
   size = ec->get_size (ec->cls);
   if (size > 16 * 1024 * 1024)
     return; /* too large */
+  if (klen + 15 > size)
+    return; /* malformed klen */
   if (NULL == (data = malloc ((size_t) size)))
     return; /* out of memory */
   memcpy (data, buf, iret);
@@ -189,16 +191,16 @@ EXTRACTOR_dvi_extract_method (struct EXTRACTOR_ExtractContext *ec)
   while (off < size)
     {
       if (0 >= (iret = ec->read (ec->cls, &buf, 16 * 1024)))
-        {
-          free (data);
-          return;
-        }
+        {
+          free (data);
+          return;
+        }
       memcpy (&data[off], buf, iret);
       off += iret;
     }
   pos = size - 1;
   while ( (223 == data[pos]) &&
-          (pos > 0) )
+          (pos > 0) )
     pos--;
   if ( (2 != data[pos]) ||
        (pos < 40) )
@@ -225,9 +227,9 @@ EXTRACTOR_dvi_extract_method (struct EXTRACTOR_ExtractContext *ec)
         break;
       if ( (pos + 45 > size) ||
            (pos + 45 < pos) )
-        goto CLEANUP;
+        goto CLEANUP;
       if (data[pos] != 139)     /* expect 'bop' */
-        goto CLEANUP;
+        goto CLEANUP;
       pageCount++;
       opos = pos;
       pos = getIntAt (&data[opos + 41]);
@@ -238,24 +240,24 @@ EXTRACTOR_dvi_extract_method (struct EXTRACTOR_ExtractContext *ec)
     }
   /* ok, now we believe it's a dvi... */
   snprintf (pages,
-            sizeof (pages),
-            "%u", 
-            pageCount);
+            sizeof (pages),
+            "%u", 
+            pageCount);
   if (0 != ec->proc (ec->cls, 
-                     "dvi",
-                     EXTRACTOR_METATYPE_PAGE_COUNT,
-                     EXTRACTOR_METAFORMAT_UTF8,
-                     "text/plain",
-                     pages,
-                     strlen (pages) + 1))
+                     "dvi",
+                     EXTRACTOR_METATYPE_PAGE_COUNT,
+                     EXTRACTOR_METAFORMAT_UTF8,
+                     "text/plain",
+                     pages,
+                     strlen (pages) + 1))
     goto CLEANUP;
   if (0 != ec->proc (ec->cls, 
-                     "dvi",
-                     EXTRACTOR_METATYPE_MIMETYPE,
-                     EXTRACTOR_METAFORMAT_UTF8,
-                     "text/plain",
-                     "application/x-dvi",
-                     strlen ("application/x-dvi") + 1))
+                     "dvi",
+                     EXTRACTOR_METATYPE_MIMETYPE,
+                     EXTRACTOR_METAFORMAT_UTF8,
+                     "text/plain",
+                     "application/x-dvi",
+                     strlen ("application/x-dvi") + 1))
     goto CLEANUP;
   {
     char comment[klen + 1];
@@ -263,18 +265,18 @@ EXTRACTOR_dvi_extract_method (struct EXTRACTOR_ExtractContext *ec)
     comment[klen] = '\0';
     memcpy (comment, &data[15], klen);
     if (0 != ec->proc (ec->cls, 
-                       "dvi",
-                       EXTRACTOR_METATYPE_COMMENT,
-                       EXTRACTOR_METAFORMAT_C_STRING,
-                       "text/plain",
-                       comment,
-                       klen + 1))
+                       "dvi",
+                       EXTRACTOR_METATYPE_COMMENT,
+                       EXTRACTOR_METAFORMAT_C_STRING,
+                       "text/plain",
+                       comment,
+                       klen + 1))
       goto CLEANUP;
   }
   /* try to find PDF/ps special */
   pos = opos;
   while ( (size >= 100) &&
-          (pos < size - 100) )
+          (pos < size - 100) )
     {
       switch (data[pos])
         {
@@ -284,34 +286,34 @@ EXTRACTOR_dvi_extract_method (struct EXTRACTOR_ExtractContext *ec)
           break;
         case 239:              /* zzz1 */
           len = data[pos + 1];
-          if (pos + 2 + len < size)
-            if (0 != parseZZZ ((const char *) data, pos + 2, len, ec->proc, ec->cls))
-              goto CLEANUP;
+          if ( (pos + 2 + len < size) &&
+               (0 != parseZZZ ((const char *) data, pos + 2, len, ec->proc, ec->cls)) )
+              goto CLEANUP;
           pos += len + 2;
           break;
         case 240:              /* zzz2 */
           len = getShortAt (&data[pos + 1]);
-          if (pos + 3 + len < size)
-            if (0 != parseZZZ ((const char *) data, pos + 3, len, ec->proc, ec->cls))
-              goto CLEANUP;
+          if ( (pos + 3 + len < size) &&
+               (0 != parseZZZ ((const char *) data, pos + 3, len, ec->proc, ec->cls)) )
+            goto CLEANUP;
           pos += len + 3;
           break;
         case 241:              /* zzz3, who uses that? */
           len = (getShortAt (&data[pos + 1])) + 65536 * data[pos + 3];
-          if (pos + 4 + len < size)
-            if (0 != parseZZZ ((const char *) data, pos + 4, len, ec->proc, ec->cls))
-              goto CLEANUP;
+          if ( (pos + 4 + len < size) &&
+               (0 != parseZZZ ((const char *) data, pos + 4, len, ec->proc, ec->cls)) )
+              goto CLEANUP;
           pos += len + 4;
           break;
         case 242:              /* zzz4, hurray! */
           len = getIntAt (&data[pos + 1]);
-          if (pos + 1 + len < size)
-            if (0 != parseZZZ ((const char *) data, pos + 5, len, ec->proc, ec->cls))
-              goto CLEANUP;
+          if ( (pos + 1 + len < size) &&
+               (0 != parseZZZ ((const char *) data, pos + 5, len, ec->proc, ec->cls)) )
+            goto CLEANUP;
           pos += len + 5;
           break;
         default:               /* unsupported opcode, abort scan */
-          goto CLEANUP;
+          goto CLEANUP;
         }
     }
  CLEANUP: