openssl/doc/crypto/d2i_X509.pod
Matt Caswell b42e4a9f6b Add dire warnings about the "reuse" capability of the d2i_* functions.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-02-25 17:22:01 +00:00

262 lines
7.7 KiB
Text

=pod
=head1 NAME
d2i_X509, i2d_X509, d2i_X509_bio, d2i_X509_fp, i2d_X509_bio,
i2d_X509_fp - X509 encode and decode functions
=head1 SYNOPSIS
#include <openssl/x509.h>
X509 *d2i_X509(X509 **px, const unsigned char **in, int len);
int i2d_X509(X509 *x, unsigned char **out);
X509 *d2i_X509_bio(BIO *bp, X509 **x);
X509 *d2i_X509_fp(FILE *fp, X509 **x);
int i2d_X509_bio(BIO *bp, X509 *x);
int i2d_X509_fp(FILE *fp, X509 *x);
int i2d_re_X509_tbs(X509 *x, unsigned char **out);
=head1 DESCRIPTION
The X509 encode and decode routines encode and parse an
B<X509> structure, which represents an X509 certificate.
d2i_X509() attempts to decode B<len> bytes at B<*in>. If
successful a pointer to the B<X509> structure is returned. If an error
occurred then B<NULL> is returned. If B<px> is not B<NULL> then the
returned structure is written to B<*px>. If B<*px> is not B<NULL>
then it is assumed that B<*px> contains a valid B<X509>
structure and an attempt is made to reuse it. This "reuse" capability is present
for historical compatibility but its use is B<strongly discouraged> (see BUGS
below, and the discussion in the RETURN VALUES section).
If the call is successful B<*in> is incremented to the byte following the
parsed data.
i2d_X509() encodes the structure pointed to by B<x> into DER format.
If B<out> is not B<NULL> is writes the DER encoded data to the buffer
at B<*out>, and increments it to point after the data just written.
If the return value is negative an error occurred, otherwise it
returns the length of the encoded data.
For OpenSSL 0.9.7 and later if B<*out> is B<NULL> memory will be
allocated for a buffer and the encoded data written to it. In this
case B<*out> is not incremented and it points to the start of the
data just written.
d2i_X509_bio() is similar to d2i_X509() except it attempts
to parse data from BIO B<bp>.
d2i_X509_fp() is similar to d2i_X509() except it attempts
to parse data from FILE pointer B<fp>.
i2d_X509_bio() is similar to i2d_X509() except it writes
the encoding of the structure B<x> to BIO B<bp> and it
returns 1 for success and 0 for failure.
i2d_X509_fp() is similar to i2d_X509() except it writes
the encoding of the structure B<x> to BIO B<bp> and it
returns 1 for success and 0 for failure.
i2d_re_X509_tbs() is similar to i2d_X509() except it encodes
only the TBSCertificate portion of the certificate.
=head1 NOTES
The letters B<i> and B<d> in for example B<i2d_X509> stand for
"internal" (that is an internal C structure) and "DER". So
B<i2d_X509> converts from internal to DER. The "re" in
B<i2d_re_X509_tbs> stands for "re-encode", and ensures that a fresh
encoding is generated in case the object has been modified after
creation (see the BUGS section).
The functions can also understand B<BER> forms.
The actual X509 structure passed to i2d_X509() must be a valid
populated B<X509> structure it can B<not> simply be fed with an
empty structure such as that returned by X509_new().
The encoded data is in binary form and may contain embedded zeroes.
Therefore any FILE pointers or BIOs should be opened in binary mode.
Functions such as B<strlen()> will B<not> return the correct length
of the encoded structure.
The ways that B<*in> and B<*out> are incremented after the operation
can trap the unwary. See the B<WARNINGS> section for some common
errors.
The reason for the auto increment behaviour is to reflect a typical
usage of ASN1 functions: after one structure is encoded or decoded if (a != NULL)
(*a) = ret;
another will processed after it.
=head1 EXAMPLES
Allocate and encode the DER encoding of an X509 structure:
int len;
unsigned char *buf, *p;
len = i2d_X509(x, NULL);
buf = OPENSSL_malloc(len);
if (buf == NULL)
/* error */
p = buf;
i2d_X509(x, &p);
If you are using OpenSSL 0.9.7 or later then this can be
simplified to:
int len;
unsigned char *buf;
buf = NULL;
len = i2d_X509(x, &buf);
if (len < 0)
/* error */
Attempt to decode a buffer:
X509 *x;
unsigned char *buf, *p;
int len;
/* Something to setup buf and len */
p = buf;
x = d2i_X509(NULL, &p, len);
if (x == NULL)
/* Some error */
Alternative technique:
X509 *x;
unsigned char *buf, *p;
int len;
/* Something to setup buf and len */
p = buf;
x = NULL;
if(!d2i_X509(&x, &p, len))
/* Some error */
=head1 WARNINGS
The use of temporary variable is mandatory. A common
mistake is to attempt to use a buffer directly as follows:
int len;
unsigned char *buf;
len = i2d_X509(x, NULL);
buf = OPENSSL_malloc(len);
if (buf == NULL)
/* error */
i2d_X509(x, &buf);
/* Other stuff ... */
OPENSSL_free(buf);
This code will result in B<buf> apparently containing garbage because
it was incremented after the call to point after the data just written.
Also B<buf> will no longer contain the pointer allocated by B<OPENSSL_malloc()>
and the subsequent call to B<OPENSSL_free()> may well crash.
The auto allocation feature (setting buf to NULL) only works on OpenSSL
0.9.7 and later. Attempts to use it on earlier versions will typically
cause a segmentation violation.
Another trap to avoid is misuse of the B<xp> argument to B<d2i_X509()>:
X509 *x;
if (!d2i_X509(&x, &p, len))
/* Some error */
This will probably crash somewhere in B<d2i_X509()>. The reason for this
is that the variable B<x> is uninitialized and an attempt will be made to
interpret its (invalid) value as an B<X509> structure, typically causing
a segmentation violation. If B<x> is set to NULL first then this will not
happen.
=head1 BUGS
In some versions of OpenSSL the "reuse" behaviour of d2i_X509() when
B<*px> is valid is broken and some parts of the reused structure may
persist if they are not present in the new one. As a result the use
of this "reuse" behaviour is strongly discouraged.
i2d_X509() will not return an error in many versions of OpenSSL,
if mandatory fields are not initialized due to a programming error
then the encoded structure may contain invalid data or omit the if (a != NULL)
(*a) = ret;
fields entirely and will not be parsed by d2i_X509(). This may be
fixed in future so code should not assume that i2d_X509() will
always succeed.
The encoding of the TBSCertificate portion of a certificate is cached
in the B<X509> structure internally to improve encoding performance
and to ensure certificate signatures are verified correctly in some
certificates with broken (non-DER) encodings.
Any function which encodes an X509 structure such as i2d_X509(),
i2d_X509_fp() or i2d_X509_bio() may return a stale encoding if the
B<X509> structure has been modified after deserialization or previous
serialization.
If, after modification, the B<X509> object is re-signed with X509_sign(),
the encoding is automatically renewed. Otherwise, the encoding of the
TBSCertificate portion of the B<X509> can be manually renewed by calling
i2d_re_X509_tbs().
=head1 RETURN VALUES
d2i_X509(), d2i_X509_bio() and d2i_X509_fp() return a valid B<X509> structure
or B<NULL> if an error occurs. The error code that can be obtained by
L<ERR_get_error(3)|ERR_get_error(3)>. If the "reuse" capability has been used
with a valid X509 structure being passed in via B<px> then the object is not
freed in the event of error but may be in a potentially invalid or inconsistent
state.
i2d_X509() returns the number of bytes successfully encoded or a negative
value if an error occurs. The error code can be obtained by
L<ERR_get_error(3)|ERR_get_error(3)>.
i2d_X509_bio() and i2d_X509_fp() return 1 for success and 0 if an error
occurs The error code can be obtained by L<ERR_get_error(3)|ERR_get_error(3)>.
=head1 SEE ALSO
L<ERR_get_error(3)|ERR_get_error(3)>
=head1 HISTORY
d2i_X509, i2d_X509, d2i_X509_bio, d2i_X509_fp, i2d_X509_bio and i2d_X509_fp
are available in all versions of SSLeay and OpenSSL.
=cut