Decode the WOLFSSL_CERT_NAME_ALL subject components

[Frauschi]

Description

certNameSubject[] is indexed by id - 3 and, without WOLFSSL_CERT_NAME_ALL, covers DN tag ids 3 through 18. The four entries WOLFSSL_CERT_NAME_ALL appends are labelled name, givenName, initials and dnQualifier - ids 41, 42, 43 and 46 - but appending them puts them at positions that make them ids 19, 20, 21 and 22.

Two consequences, in opposite directions:

• The four components are dropped. GetRDN() finds no entry for the real ids, leaves typeStr NULL and never calls SetSubject(), so subjectN, subjectGN, subjectI and subjectDNQ are never populated.
• Four unrelated components are decoded under the wrong name. 2.5.4.19 through 2.5.4.22 - physicalDeliveryOfficeName, telephoneNumber, teletexTerminalIdentifier, x121Address - match those entries instead, landing in the wrong DecodedCert field, under the wrong tag string, with the wrong X509_NAME NID.

Affects WOLFSSL_CERT_NAME_ALL builds: --enable-all and the openvpn, nginx, haproxy, hitch, bind and rsyslog recipes.

Fix

Drop the four misplaced entries and give the ids their own branches in GetRDN() and SetSubject(), beside the ASN_X500_UNIQUE_ID branch already there for a component outside the contiguous range. The tag-string macros are already defined in asn.h under the same guard. Without WOLFSSL_CERT_NAME_ALL the table already stopped at id 18, so ids 19 to 22 go back to being unrecognised there, as they were.

Testing

Adds test_wc_CertNameAllSubject() to the asn group: generates a self-signed cert carrying all four RDNs, re-parses it, and asserts each lands in the matching DecodedCert field and appears in dCert.subject. Nothing covered these paths before - no cert under certs/ carries any of the four OIDs, and no test referenced the fields.

make check green under --enable-all, and the new test fails when asn.c is restored from master, so it pins the indexing rather than passing vacuously.

Note for the reviewer

The subject string uses /N= and /GN=, matching the legacy asn_orig.c decoder, while RebuildFullName() regenerates it from obj->sName ("name", "givenName"). That divergence is pre-existing and mirrors the SN vs surname split in ssl_asn1.c, but this change makes it reachable in template builds for the first time. Left alone since aligning them is a user-visible output change; happy to do it separately.

[github-actions]

MemBrowse Memory Report

gcc-arm-cortex-m4-openssl-compat

• FLASH: .rodata -144 B, .text +256 B (+0.0%, 784,116 B / 1,048,576 B, total: 75% used)

linuxkm-pie

• Data: __patchable_function_entries +8 B (+0.0%, 27,080 B)
  No memory changes detected for:
• gcc-arm-cortex-m0plus
• gcc-arm-cortex-m3
• gcc-arm-cortex-m4
• gcc-arm-cortex-m4-baremetal
• gcc-arm-cortex-m4-crypto-only
• gcc-arm-cortex-m4-dtls13
• gcc-arm-cortex-m4-min-ecc
• gcc-arm-cortex-m4-pkcs7
• gcc-arm-cortex-m4-pq
• gcc-arm-cortex-m4-rsa-only
• gcc-arm-cortex-m4-sp-math
• gcc-arm-cortex-m4-tls12
• gcc-arm-cortex-m4-tls13
• gcc-arm-cortex-m7
• gcc-arm-cortex-m7-pq
• gcc-arm-cortex-m7-tls13
• linuxkm-standard
• stm32-sim-stm32h753