On Mon, Apr 30, 2012, Welling, Conrad Gerhart wrote: > > > Our client application uses cURL-7.19.7 + openssl-0.9.8l to transfer > text files via HTTPS, using server-only authentication. The only call > the cURL source makes to an openssl d2i_*_bio() or d2i_*_fp() function > is d2i_PKCS12_fp(). Unfortunately, it is not absolutely clear - yet - > that a PKCS#12-related function would have anything to do with a SSL > hand-shake sequence, which is the only time I surmise the CVE-2012-2110 > vulnerability applies in this scenario. Now, I DO surmise - although > not having found it definitively stated anywhere - that the HTTPS server > does follow the ServerHello message by sending it's certificate in ASN.1 > representation in the Certificate message, and, not in any other > representation. And, since I understand that the code for parsing ASN.1 > is the heart of the vulnerability, I'm very much inclined to act as if > CVE-2012-2110 applies to our application. I do, however, want clarity > regarding this issue: Can someone make this glaringly clear for me? > > >
There is a subtle point here which I hope to clarify. An application is vulnerable if it calls d2i_*_fp or d2i_*_bio on untrusted data: that is an attacker has a means to inject data into these functions. Normally d2i_PKCS12_fp reads in a file from trusted configuration so this cannot happen. The advisory states that SSL/TLS code is not affected. There is a reason for this. Although the SSL/TLS code reads in ASN1 DER format data in several places it uses memory based functions such as d2i_X509 which are NOT affected. Steve. -- Dr Stephen N. Henson. OpenSSL project core developer. Commercial tech support now available see: http://www.openssl.org ______________________________________________________________________ OpenSSL Project http://www.openssl.org User Support Mailing List openssl-users@openssl.org Automated List Manager majord...@openssl.org
