Bug 1082842 (CVE-2018-7454)

Summary: VUL-1: CVE-2018-7454: xpdf: A NULL pointer dereference in XFAForm::scanFields in XFAForm.cc allows attackers to launch denial of service via a specific pdf file
Product: [Novell Products] SUSE Security Incidents Reporter: Karol Babioch <karol>
Component: IncidentsAssignee: Peter Simons <peter.simons>
Status: NEW --- QA Contact: Security Team bot <security-team>
Severity: Normal    
Priority: P4 - Low CC: smash_bz
Version: unspecified   
Target Milestone: ---   
Hardware: Other   
OS: Other   
URL: https://smash.suse.de/issue/200814/
Whiteboard: CVSSv3:SUSE:CVE-2018-7454:3.3:(AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:L) maint:planned:update
Found By: Security Response Team Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Bug Depends on:    
Bug Blocks: 1133493    

Description Karol Babioch 2018-02-26 13:08:49 UTC 
CVE-2018-7454

A NULL pointer dereference in XFAForm::scanFields in XFAForm.cc in xpdf 4.00
allows attackers to launch denial of service via a specific pdf file, as
demonstrated by pdftohtml.

References:
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2018-7454
http://www.cvedetails.com/cve/CVE-2018-7454/
Comment 1 Peter Simons 2018-06-21 09:14:54 UTC 
No fix for this issue was published. According to the web forum, upstream tried to remedy this issue by

> adding a test for nComps==0 when reading the SIZ segment

..., but this turned out to be insufficient. After some back and forth, the solution apparently is to disable 0 page PDF files altogether:

> Zero-page PDF files are kind of an odd case. I'll add a test in XFAForm::load that bails out early if nPages==0.

I looked at the code of XFAForm::load() in the 4.00 release, trying to re-implement the fix, but unfortunately there is no obvious way to write that test. (nPages is an attribute deep inside of a XFAForm object, so the object has to be constructed already before I can access it, which seems too late for the test to be meaningful).

Upstream was asked whether they'll release the fix -- or a new version that contains it --, and they said:

> This will be in the next release. I don't have a schedule for that yet.

That was over a year ago.