Issue report: x509.Name fails to process the TCG OIDs (2.23.133.2.*) about asn1crypto HOT 8 CLOSED

The code at #270 should fix this

from asn1crypto.

Hi,
I think (really not sure!) the problem is here:
https://github.com/wbond/asn1crypto/blob/master/asn1crypto/x509.py#L1028-L1031

  else:
     value = DirectoryString(
                    name=encoding_name,
                    value=encoding_class(attribute_value)
                )

The build function assumes, with some exceptions, assumes that everything is a DirectoryString. When this code is changed to use the type from NameTypeAndValue._oid_specs it works as expected.

from asn1crypto.

Hi,

Regarding the document of TCG EK Credential Profile below.
https://trustedcomputinggroup.org/wp-content/uploads/TCG-EK-Credential-Profile-V-2.5-R2_published.pdf

In section 3.2, there is a table mentioned the type of subject alt name for TPM Manufacturer, model, and version should be GeneralName and directoryName.

Does that mean we should fix this as the comment 1 mentioned?

Thanks.

from asn1crypto.

from asn1crypto.

Yes, it looks like switching it to use the data from NameTypeAndValue._oid_specs is probably the correct move.

from asn1crypto.

@joseph-lo-nvtw I looked at https://trustedcomputinggroup.org/wp-content/uploads/TCG-EK-Credential-Profile-V-2.5-R2_published.pdf and it still specifies UTF8String for tcg-at-tpmManufacturer (e.g. 2.23.133.2.1), etc.

from asn1crypto.

Yes, the final raw binary string in ASN1 is UTF8String.

But regarding Table 1 in section 3.2, the RFC5280 type of Subject Alternative Name is directoryName.
Which data type should we use here?
https://github.com/wbond/asn1crypto/blob/master/asn1crypto/x509.py#L689

With UTF8String, it doesn't work correctly. The x509.Name.build logic doesn't accept UTF8String data type.

Thanks for your reply.

from asn1crypto.

Hi @wbond ,

Thanks for the fix. I have verified it.

from asn1crypto.