[krbdev.mit.edu #8506] PKINIT fails with PKCS#11 middlware that implements PKCS#1 V2.1

[Tom Yu via RT]

As for "how did this ever work?": it seems that some PKCS #1 signature
verifiers are lenient (which can lead to vulnerabilities), so omitting
the (nominally required) NULL value from the AlgorithmIdentifier
encoding still produces a signature that they will accept.  OpenSSL
master seems to do a round trip through DER, with the parameters field
optional (as in the ASN.1 module for PKCS #1).

This one smart card seems to verify the encoding.  I don't know how many
other smart cards verify the DigestInfo encoding.