Reputation: 3496
I am debugging a TLS handshake between a client and a server. Both certificates have been signed by the same CA via openssl
.
The process fails at this function in mbedtls
:
/*
* Compare two X.509 Names (aka rdnSequence).
*
* See RFC 5280 section 7.1, though we don't implement the whole algorithm:
* we sometimes return unequal when the full algorithm would return equal,
* but never the other way. (In particular, we don't do Unicode normalisation
* or space folding.)
*
* Return 0 if equal, -1 otherwise.
*/
static int x509_name_cmp( const mbedtls_x509_name *a, const mbedtls_x509_name *b )
{
/* Avoid recursion, it might not be optimised by the compiler */
while( a != NULL || b != NULL )
{
if( a == NULL || b == NULL )
return( -1 );
/* type */
if( a->oid.tag != b->oid.tag ||
a->oid.len != b->oid.len ||
memcmp( a->oid.p, b->oid.p, b->oid.len ) != 0 )
{
// *** FAIL ***
return( -1 );
}
/* value */
if( x509_string_cmp( &a->val, &b->val ) != 0 )
return( -1 );
/* structure of the list of sets */
if( a->next_merged != b->next_merged )
return( -1 );
a = a->next;
b = b->next;
}
/* a == NULL == b */
return( 0 );
}
I am wondering about mbedtls_x509_name
structure and its oid.p
field:
/**
* Type-length-value structure that allows for ASN1 using DER.
*/
typedef struct mbedtls_asn1_buf
{
int tag; /**< ASN1 type, e.g. MBEDTLS_ASN1_UTF8_STRING. */
size_t len; /**< ASN1 length, in octets. */
unsigned char *p; /**< ASN1 data, e.g. in ASCII. */
}
mbedtls_asn1_buf;
In my case, both tag
are 6 (MBEDTLS_ASN1_OID
), len
is 3 and I have the following buffers: 0x55 0x40 0x70
and 0x55 0x40 0xA0
What do these values mean (i.e. which information from the certificate do they match) ? Why the last byte is different ?
Upvotes: 3
Views: 522