False positives for CVE-2016-3720 for several jackson libraries

[albuch]

CVE-2016-3270 is falsely reported for the following libraries and probably all other com.fasterxml.jackson libraries for all versions (e.g. [version] = 2.5.0, 2.5.4, 2.6.5)

• jackson-annotations-[version].jar
• jackson-core-[version].jar
• jackson-databind-[version].jar
• jackson-datatype-jdk8-[version].jar
• jackson-datatype-jsr310-[version].jar

Am I right that we can't use the CPE as an suppression rule here as that would exclude the vulnerable library jackson-dataformat-xml as well?

[awhitford]

Is the CVE limited to just jackson-dataformat-xml? Because that is what I thought at first, but then the CVE mentions cpe:/a:fasterxml:jackson:- which seems pretty broad.

Note that the issue seems limited to XML, yet I only use Jackson for JSON.

[albuch]

As far as I understand the original Fedora bug report only the extension jackson-dataformat-xml is affected which must be be put explicitly on the classpath.

[awhitford]

In that case, isn't the CVE metadata incorrect? As long as it says cpe:/a:fasterxml:jackson:-, then it is going to get hits on the artifacts that you mentioned. I don't see why the metadata can not be more specific to the dataformat-xml artifact. (Whom do you complain to about CVE metadata?)

[albuch]

That is correct Anthony. Though I'm not sure where to file a fix for CVE metadata, and if it would be fixed at all. I guess RedHat could be the contact here as the responsible CVE Numbering Authority.

[albuch]

FYI, I've sent an email to RedHat security team. We'll see if I'm mistaken or if they can fix it.

[awhitford]

I raised a ticket for jackson-dataformat-xml and they responded saying that it has already been fixed for Jackson 2.7.4.

I also sent an email to [email protected] (after reading the FAQ) and they forwarded the issue to [email protected].

[jeremylong]

After thinking about this some more I believe we could use the following suppression rules to suppress the false positive for all except the Jackson-dataformat-xml:

<?xml version="1.0" encoding="UTF-8"?>
<suppressions xmlns="https://jeremylong.github.io/DependencyCheck/dependency-suppression.1.1.xsd">
   <suppress>
      <notes><![CDATA[
      file name: jackson-core-2.6.5.jar
      ]]></notes>
      <gav regex="true">com\.fasterxml\.jackson\.core:jackson.*</gav>
      <cve>CVE-2016-3720</cve>
   </suppress>
   <suppress>
      <notes><![CDATA[
      file name: jackson-core-2.6.5.jar
      ]]></notes>
      <gav regex="true">com\.fasterxml\.jackson\.dataformat:jackson(?!\-dataformat\-xml).*</gav>
      <cve>CVE-2016-3720</cve>
   </suppress>
</suppressions>

[susenj]

@awhitford , did you get any update over you e-mail? I am safely assuming that the vulnerability exists even with Jackson 2.7.4.

Does anyone know the tentative timeline of the next Jackson version?

[jeremylong]

The fix for this FP has been implemented in 1.4.5-SNAPSHOT and will be included in the next release.

[samisalonen]

Jackson artifacts had different Maven coordinates earlier. I added the following suppression to handle those old artifacts with Dependency-Check 1.4.5.

   <suppress>
      <notes><![CDATA[
      https://github.com/jeremylong/DependencyCheck/issues/517
      ]]></notes>
      <gav regex="true">org\.codehaus\.jackson:jackson-.*</gav>
      <cve>CVE-2016-3720</cve>
   </suppress>

I am not aware which old artifact(s) corresponds to vulnerable com.fasterxml.jackson.dataformat:jackson-dataformat-xml (if any), so this potentially suppresses too much.

[jeremylong]

Thanks - I did a little research and updated the regex to include the alternative group.

[albuch]

Hi @jeremylong,
another package that is falsely reported is

ehcache-2.10.3.jar/rest-management-private-classpath/META-INF/maven/com.fasterxml.jackson.jaxrs/jackson-jaxrs-base/pom.xml (com.fasterxml.jackson.jaxrs:jackson-jaxrs-base:2.3.3, cpe:/a:fasterxml:jackson:2.3.3) : CVE-2016-3720

I've locally updated the rule to exclude all sub-groups except dataformat:

    <suppress base="true">
        <notes><![CDATA[
            This CVE only affects jackson-dataformat-xml. See issue #517.
      ]]></notes>
        <gav regex="true">(org\.codehaus\.jackson|com\.fasterxml\.jackson\.(?!dataformat)[^:]+):jackson.*</gav>
        <cve>CVE-2016-3720</cve>
    </suppress>

[cowtowncoder]

Apologies for commenting on a closed ticket, but since I had to send an update request for another related Jackson XML CVE I thought update might make sense.

Jackson side issue for CVE-3720 is: FasterXML/jackson-dataformat-xml#190 and it was indeed fixed in 2.7.4 and only affects jackson-dataformat-xml.

There is a newer but similar CVE, 7051: http://www.cvedetails.com/cve/CVE-2016-7051/ which:

• Affects jackson-dataformat-xml (only)
• Fixed in 2.7.8 (and 2.8.4)
• Jackson side issue: Disable SUPPORT_DTD for XMLInputFactory unless explicitly overridden [CVE-2016-7051] FasterXML/jackson-dataformat-xml#211

It also looks like description at mitre.org is too general as well, leading to false alarms:

FasterXML/jackson-core#371

Similar to issue here.

I suspect it might make sense to file a new issue, but thought maybe starting here would make most sense.

[nhenneaux]

It still fails with v2.0.1. Should I create a new ticket?

jackson-jaxrs-base-2.8.9.jar (com.fasterxml.jackson.jaxrs:jackson-jaxrs-base:2.8.9, cpe:/a:fasterxml:jackson:2.8.9) : CVE-2016-7051
jackson-datatype-joda-2.8.9.jar (com.fasterxml.jackson.datatype:jackson-datatype-joda:2.8.9, cpe:/a:fasterxml:jackson:2.8.9) : CVE-2016-7051

[Stephan202]

@nhenneaux see #792.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.