Extend scope of TC in ECDH-param extraction from CRT
Previously, TinyCrypt was only used for ECDHE-ECDSA/RSA ciphersuites.
This commit is a step towards using it for _all_ ciphersuites involving
ECDHE (specifically: ECDHE, ECDHE-PSK, static ECDH), extending the scope
of the use of TinyCrypt in the assembly of the PMS.
The SSL context maintains a set of 'out pointers' indicating the
address at which to write the header fields of the next outgoing
record. Some of these addresses have a static offset from the
beginning of the record header, while other offsets can vary
depending on the active record encryption mechanism: For example,
if an explicit IV is in use, there's an offset between the end
of the record header and the beginning of the encrypted data to
allow the explicit IV to be placed in between; also, if the DTLS
Connection ID (CID) feature is in use, the CID is part of the
record header, shifting all subsequent information (length, IV, data)
to the back.
When setting up an SSL context, the out pointers are initialized
according to the identity transform + no CID, and it is important
to keep them up to date whenever the record encryption mechanism
changes, which is done by the helper function ssl_update_out_pointers().
During context deserialization, updating the out pointers according
to the deserialized record transform went missing, leaving the out
pointers the initial state. When attemping to encrypt a record in
this state, this lead to failure if either a CID or an explicit IV
was in use. This wasn't caught in the tests by the bad luck that
they didn't use CID, _and_ used the default ciphersuite based on
ChaChaPoly, which doesn't have an explicit IV. Changing either of
this would have made the existing tests fail.
This commit fixes the bug by adding a call to ssl_update_out_pointers()
to ssl_context_load() implementing context deserialization.
Extending test coverage is left for a separate commit.
The NO_INLINE annotation of tls_prf_sha256() and tls_prf_sha384() from
the last commit surprisingly had an influence on ARMC5 compilation in
that tls_prf_generic() was no longer automatically inlined into
tls_prf_sha256() if only the latter was enabled (and is the point
where tls_prf_generic() is called). This commit forces inlining
of tls_prf_generic() in this case.
Usually, compilers are clever enough to pick the best inlining
strategy, but in this instance, it appears that compiling on ARMC6,
the compilers inlines xxx_prf_yyy() and xxx_calc_finished_yyy()
even though it really shouldn't. Forbid inlining through the use
of __attribute__((noinline)).
Somehow, at least ARMC5 isn't able to recognize this automatically.
Since some of the arguments to ssl_populate_transform() are compile-
time constants in reduced configurations, inlining leads to slightly
shorter code.
This saves a few bytes in configurations where only one hash
is enabled, and configurations allowing multiple hashes probably
don't care about code-size anyway.
This function is called on client-only once the ciphersuite has
been chosen and it it is known which digest the client will need
for the handshake transcript throughout the handshake, and causes
all other unneeded handshake transcripts to be discontinued.
(On the server, we cannot call this function because we don't know
which hash the client will those in its CertificateVerify message).
However, the benefit of this call is marginal, since transcript hash
computation is negligible compared to asymmetric crypto, and moreover
the handshake transcript contexts for the unused digests are still
stored in the SSL handshake parameter structure and not freed until
the end of the handshake.
Finally, if we're running on a _really_ constrained client, there
will be only one hash function enabled anyway, and in this case
the checksum optimization has no effect.
This commit therefore removes checksum optimization altogether,
saving some code on constrained systems.
Now function mbedtls_ssl_set_hostname is compile-time configurable
in config.h with define MBEDTLS_X509_REMOVE_HOSTNAME_VERIFICATION.
This affects to many x509 API's. See config.h for details.
Previously, MBEDTLS_KEY_EXCHANGE_ECDH[E]_XXX_ENABLED would imply
that MBEDTLS_ECDH_C is set, but with the introduction of tinyCrypt
as an alternative ECDH implementation, this is no longer the case.
The postprocessing code for the server-side incoming client key
exchange and the client-side outgoing client key exchange both
contain the same code-paths for building the premaster secret
depending on the chosen ciphersuite (e.g., for ECDHE-PSK,
concatenating the ECDHE secret with the chosen PSK).
This commit moves this common code to ssl_tls.c, allowing
client- and server-side to share it.
This commit implements the record checking API
mbedtls_ssl_check_record()
on top of the restructured incoming record stack.
Specifically, it makes use of the fact that the core processing routines
ssl_parse_record_header()
mbedtls_ssl_decrypt_buf()
now operate on instances of the SSL record structure mbedtls_record
instead of the previous mbedtls_ssl_context::in_xxx fields.
After the rewrite of incoming record processing to use the internal
SSL record structure mbedtls_record (which contains the data_offset
field to indicate where the IV resides), this field is no longer
necessary.
Note: This is an API break.
ssl_get_next_record() updates the legacy in_xxx fields in two places,
once before record decryption and once after. Now that record decryption
doesn't use or affect the in_xxx fields anymore, setting up the these
legacy fields can entirely be moved to the end of ssl_get_next_record(),
which is what this comit does.
This commit solely moves existing code, but doesn't yet simplify the
now partially redundant settings of the in_xxx fields. This will be
done in a separate commit.
Multiple record attributes such as content type and payload length
may change during record decryption, and the legacy in_xxx fields
in the SSL context therefore need to be updated after the record
decryption routine ssl_decrypt_buf() has been called.
After the previous commit has made ssl_prepare_record_content()
independent of the in_xxx fields, setting them can be moved
outside of ssl_prepare_record_content(), which is what this
commit does.
Previously, ssl_update_in_pointers() ensured that the in_xxx pointers
in the SSL context are set to their default state so that the record
header parsing function ssl_parse_record_header() could make use of them.
By now, the latter is independent of these pointers, so they don't need
to be setup before calling ssl_parse_record_header() anymore.
However, other parts of the messaging stack might still depend on it
(to be studied), and hence this commit does not yet reomve
ssl_update_in_pointers() entirely.
The stack maintains pointers mbedtls_ssl_context::in_xxx pointing to
various parts of the [D]TLS record header. Originally, these fields
were determined and set in ssl_parse_record_header(). By now,
ssl_parse_record_header() has been modularized to setup an instance
of the internal SSL record structure mbedtls_record, and to derive
the old in_xxx fields from that.
This commit takes a further step towards removing the in_xxx fields
by deriving them from the established record structure _outside_ of
ssl_parse_record_header() after the latter has succeeded.
One exception is the handling of possible client reconnects,
which happens in the case then ssl_parse_record_header() returns
MBEDTLS_ERR_SSL_UNEXPECTED_RECORD; since ssl_check_client_reconnect()
so far uses the in_xxx fields, they need to be derived from the
record structure beforehand.
This commit makes a first step towards modularizing the incoming record
processing by having it operate on instances of the structure mbedtls_record
representing SSL records.
So far, only record encryption/decryption operate in terms of record
instances, but the rest of the parsing doesn't. In particular,
ssl_parse_record_header() operates directly on the fixed input buffer,
setting the various ssl->in_xxx pointers and fields, and only directly
before/after calling ssl_decrypt_buf() these fields a converted to/from
mbedtls_record instances.
This commit does not yet remove the ssl->in_xxx fields, but makes a step
towards extending the lifetime of mbedtls_record structure representing
incoming records, by modifying ssl_parse_record_header() to setup an
instance of mbedtls_record, and setting the ssl->in_xxx fields from that
instance. The instance so-constructed isn't used further so far, and in
particular it is not yet consolidated with the instance set up for use
in ssl_decrypt_record(). That's for a later commit.
Previously, ssl_parse_record_header() did not check whether the current
datagram is large enough to hold a record of the advertised size. This
could lead to records being silently skipped over or backed up on the
basis of an invalid record length. Concretely, the following would happen:
1) In the case of a record from an old epoch, the record would be
'skipped over' by setting next_record_offset according to the advertised
but non-validated length, and only in the subsequent mbedtls_ssl_fetch_input()
it would be noticed in an assertion failure if the record length is too
large for the current incoming datagram.
While not critical, this is fragile, and also contrary to the intend
that MBEDTLS_ERR_SSL_INTERNAL_ERROR should never be trigger-able by
external input.
2) In the case of a future record being buffered, it might be that we
backup a record before we have validated its length, hence copying
parts of the input buffer that don't belong to the current record.
This is a bug, and it's by luck that it doesn't seem to have critical
consequences.
This commit fixes this by modifying ssl_parse_record_header() to check that
the current incoming datagram is large enough to hold a record of the
advertised length, returning MBEDTLS_ERR_SSL_INVALID_RECORD otherwise.
We don't send alerts on other instances of ill-formed records,
so why should we do it here? If we want to keep it, the alerts
should rather be sent ssl_get_next_record().
As explained in the previous commit, if mbedtls_ssl_fetch_input()
is called multiple times, all but the first call are equivalent to
bounds checks in the incoming datagram.
In DTLS, if mbedtls_ssl_fetch_input() is called multiple times without
resetting the input buffer in between, the non-initial calls are functionally
equivalent to mere bounds checks ensuring that the incoming datagram is
large enough to hold the requested data. In the interest of code-size
and modularity (removing a call to a non-const function which is logically
const in this instance), this commit replaces such a call to
mbedtls_ssl_fetch_input() by an explicit bounds check in
ssl_parse_record_header().
Previously, `ssl_handle_possible_reconnect()` was part of
`ssl_parse_record_header()`, which was required to return a non-zero error
code to indicate a record which should not be further processed because it
was invalid, unexpected, duplicate, .... In this case, some error codes
would lead to some actions to be taken, e.g. `MBEDTLS_ERR_SSL_EARLY_MESSAGE`
to potential buffering of the record, but eventually, the record would be
dropped regardless of the precise value of the error code. The error code
`MBEDTLS_ERR_SSL_HELLO_VERIFY_REQUIRED` returned from
`ssl_handle_possible_reconnect()` did not receive any special treatment and
lead to silent dopping of the record - in particular, it was never returned
to the user.
In the new logic this commit introduces, `ssl_handle_possible_reconnect()` is
part of `ssl_check_client_reconnect()` which is triggered _after_
`ssl_parse_record_header()` found an unexpected record, which is already in
the code-path eventually dropping the record; we want to leave this code-path
only if a valid cookie has been found and we want to reset, but do nothing
otherwise. That's why `ssl_handle_possible_reconnect()` now returns `0` unless
a valid cookie has been found or a fatal error occurred.
Availability of sufficient incoming data should be checked when
it is needed, which is in mbedtls_ssl_fetch_input(), and this
function has the necessary bounds checks in place.
mbedtls_ssl_decrypt_buf() asserts that the passed transform is not NULL,
but the function is only invoked in a single place, and this invocation
is clearly visible to be within a branch ensuring that the incoming
transform isn't NULL. Remove the assertion for the benefit of code-size.
The previous code performed architectural maximum record length checks
both before and after record decryption. Since MBEDTLS_SSL_IN_CONTENT_LEN
bounds the maximum length of the record plaintext, it suffices to check
only once after (potential) decryption.
This must not be confused with the internal check that the record
length is small enough to make the record fit into the internal input
buffer; this is done in mbedtls_ssl_fetch_input().
The check is in terms of the internal input buffer length and is
hence likely to be originally intended to protect against overflow
of the input buffer when fetching data from the underlying
transport in mbedtls_ssl_fetch_input(). For locality of reasoning,
it's better to perform such a check close to where it's needed,
and in fact, mbedtls_ssl_fetch_input() _does_ contain an equivalent
bounds check, too, rendering the bounds check in question redundant.
Breaking into a series of statements makes things easier when stepping through
the code in a debugger.
Previous comments we stating the opposite or what the code tested for (what we
want vs what we're erroring out on) which was confusing.
Also expand a bit on the reasons for these restrictions.
ssl_get_next_record() may pend fatal alerts in response to receiving
invalid records. Previously, however, those were never actually sent
because there was no code-path checking for pending alerts.
This commit adds a call to ssl_send_pending_fatal_alert() after
the invocation of ssl_get_next_record() to fix this.
Modelled after the config-checking header from session s11n.
The list of relevant config flags was established by manually checking the
fields serialized in the format, and which config.h flags they depend on.
This probably deserves double-checking by reviewers.
Since the type of cid_len is unsigned but shorter than int, it gets
"promoted" to int (which is also the type of the result), unless we make the
other operand an unsigned int which then forces the expression to unsigned int
as well.
