On 08/13/2011 12:30 PM, Sean Leonard wrote:
> On 7/29/2011 2:21 AM, Sean Leonard wrote:
>> What is the procedure to validate an arbitrary extended key usage (EKU)
>> with NSS?
>>
>> Suppose that one has an application built on NSS, where certificates can
>> be used if they have the extended key usage (EKU) 1.2.3.4.5.999999.
>>
>> The API calls CERT_VerifyCertificate and CERT_PKIXVerifyCert expect a
>> SECCertificateUsage argument to be passed to them. 
No, there's no way to dynamically extend SECCertificateUsage currently.
>> This argument can be
>> one or possibly more of a very small set of enumerated items such as SSL
>> client and SSL server, defined in certt.h. There is evidence that
>> CERT_PKIXVerifyCert has some code floating around in it (namely in
>> functions such as cert_NssCertificateUsageToPkixKUAndEKU and
>> PKIX_EkuChecker_Create) where arbitrary EKUs were at least considered,
>> but it also looks like these code paths were never hooked up to the main
>> CERT_PKIXVerifyCert algorithm.
Without refreshing my memory by looking at the code, I suspect your
evaluation is correct. I don't think a patch corrects this issue would
be looked on unfavorably. I think we just haven't needed to dynamically
add additional EKU's yet. The CERT_PKIXVerifyCert has the ability to
take on new parameter types without changing it's signature, so it could
be a fixable problem...

NOTE: our current usages don't always map to EKU, some certificates get
implicit usages.
>>
>> If validating an arbitrary EKU is not possible, what is the procedure to
>> validate a certificate for "any usage", so that the application can then
>> whittle down the chains as it sees fit after candidate chain(s) are
>> returned?
certificateUsageCheckAllUsages will verify the cert against all the
possible usages and return the usages it found. It doesn't look like we
have a generic 'Any' usage.

bob
>>
>> Sean
>
> *Bump*
>
> Anybody?
>
> -Sean


-- 
dev-tech-crypto mailing list
dev-tech-crypto@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-tech-crypto

Reply via email to