<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<!DOCTYPE bugzilla SYSTEM "http://bugs.gentoo.org/bugzilla.dtd">

<bugzilla version="2.22.7"
          urlbase="http://bugs.gentoo.org/"
          maintainer="bugzilla@gentoo.org"
>

    <bug>
          <bug_id>201163</bug_id>
          <alias>CVE-2007-6203</alias>
          <creation_ts>2007-12-04 01:22 0000</creation_ts>
          <short_desc>www-servers/apache &lt; 2.2.6-r6 413 Request Entity Too Large XSS (CVE-2007-6203)</short_desc>
          <delta_ts>2008-03-11 21:49:29 0000</delta_ts>
          <reporter_accessible>1</reporter_accessible>
          <cclist_accessible>1</cclist_accessible>
          <classification_id>1</classification_id>
          <classification>Unclassified</classification>
          <product>Gentoo Security</product>
          <component>Vulnerabilities</component>
          <version>unspecified</version>
          <rep_platform>All</rep_platform>
          <op_sys>Linux</op_sys>
          <bug_status>RESOLVED</bug_status>
          <resolution>FIXED</resolution>
          <bug_file_loc>http://secunia.com/advisories/27906</bug_file_loc>
          <status_whiteboard>A4 [glsa]</status_whiteboard>
          
          <priority>P2</priority>
          <bug_severity>minor</bug_severity>
          <target_milestone>---</target_milestone>
          <dependson>204838</dependson>
          
          
          <everconfirmed>1</everconfirmed>
          <reporter>rbu@gentoo.org</reporter>
          <assigned_to>security@gentoo.org</assigned_to>
          <cc>apache-bugs@gentoo.org</cc>
    
    <cc>bernd@linx.net</cc>
    
    <cc>chainsaw@gentoo.org</cc>
    
    <cc>lkundrak@v3.sk</cc>

      

      
          <long_desc isprivate="0">
            <who>rbu@gentoo.org</who>
            <bug_when>2007-12-04 01:22:48 0000</bug_when>
            <thetext>CVE-2007-6203 (http://nvd.nist.gov/nvd.cfm?cvename=CVE-2007-6203):
  Apache HTTP Server 2.0.x and 2.2.x does not sanitize the HTTP Method
  specifier header from an HTTP request when it is reflected back in a &quot;413
  Request Entity Too Large&quot; error message, which might allow cross-site
  scripting (XSS) style attacks using web client components that can send
  arbitrary headers in requests, as demonstrated via an HTTP request containing
  an invalid Content-length value, a similar issue to CVE-2006-3918.</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>rbu@gentoo.org</who>
            <bug_when>2007-12-04 01:24:37 0000</bug_when>
            <thetext>Apache herd, please advise.</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>lkundrak@v3.sk</who>
            <bug_when>2007-12-04 08:14:04 0000</bug_when>
            <thetext>This has no security impact.
How do you trick user into sending garbage before actual request method name?</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>rbu@gentoo.org</who>
            <bug_when>2007-12-04 15:50:16 0000</bug_when>
            <thetext>According to the advisory, flash movies can generate such requests:

From http://archives.neohapsis.com/archives/bugtraq/2006-07/0425.html :

var req:LoadVars=new LoadVars();
req.addRequestHeader(&quot;Foo&quot;,&quot;Bar&quot;);
req.send(&quot;http://www.vuln.site/some/page.cgi?p1=v1&amp;p2=v2&quot;,
         &quot;_blank&quot;,&quot;GET&quot;);

So if tricking a user to load a malicious flash movie, an attacker could redirect a user to a defaced URL on a remote server.</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>lkundrak@v3.sk</who>
            <bug_when>2007-12-05 18:56:52 0000</bug_when>
            <thetext>If I understend correctly, you&apos;d have to control the &quot;GET&quot; part of this:

&gt; req.send(&quot;http://www.vuln.site/some/page.cgi?p1=v1&amp;p2=v2&quot;,
&gt;          &quot;_blank&quot;,&quot;GET&quot;);

I am not able to check if it is possible, but the advisory sounds like it isn&apos;t:

&gt; The reason why we didn&apos;t consider this vulnerability a security risk is
&gt; because the attacker needs to force the victim&apos;s browser to submit a malformed
&gt; HTTP method.

...

&gt; However, in this case we need to spoof the HTTP METHOD to a specially-crafted
&gt; value.</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>hollow@gentoo.org</who>
            <bug_when>2007-12-15 14:34:45 0000</bug_when>
            <thetext>fixed in 2.2.6-r6, but please dot stabilize this version now, since it is the first unmasked USE_EXPAND version of apache and still needs some testing. i don&apos;t think this is a problem since the vuln is not even acknowledged upstream but fixed in their svn branch anyway.</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>hollow@gentoo.org</who>
            <bug_when>2008-01-07 23:05:28 0000</bug_when>
            <thetext>2.2.6-r7 is ready for stabilization, see #204838</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>hollow@gentoo.org</who>
            <bug_when>2008-01-10 16:18:47 0000</bug_when>
            <thetext>this one is ready</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>py@gentoo.org</who>
            <bug_when>2008-01-12 21:39:12 0000</bug_when>
            <thetext>time for glsa decision. I&apos;ll vote YES just because of the crash issue (bug #204410)</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>jaervosz@gentoo.org</who>
            <bug_when>2008-01-13 14:05:05 0000</bug_when>
            <thetext>Voting YES and filing.</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who>py@gentoo.org</who>
            <bug_when>2008-03-11 21:49:29 0000</bug_when>
            <thetext>GLSA 200803-19</thetext>
          </long_desc>
      
    </bug>

</bugzilla>