This is enabled by default as we generally enable things by default unless
there's a reason not to (experimental, deprecated, security risk).
We need a compile-time option because, even though the functions themselves
can be easily garbage-collected by the linker, implementing them will require
saving 64 bytes of Client/ServerHello.random values after the handshake, that
would otherwise not be needed, and people who don't need this feature
shouldn't have to pay the price of increased RAM usage.
This commit modifies `scripts/baremetal.sh` to print the total code-size
of the SSL, X.509 and Crypto libraries are runs of
- ./scripts/baremetal.sh --rom --gcc
- ./scripts/baremetal.sh --rom --armc5
- ./scripts/baremetal.sh --rom --armc6
This eases quick investigation of the effect of changes on code-size.
This commit handles occurrences of case 2 and 3 in the following list:
1. Some DTLS-specific code with no TLS-specific code (most frequent)
2. Some specific code for each protocol
3. Some TLS-specific code with no DTLS-specific code (least frequent)
Case 3 previously had a weird structure in that the TLS-specific code was
always present, but the if structure was conditional on DTLS being enabled.
This is changed by this commit to a more logical structure where both the code
and the test are conditional on TLS being enabled.
Case 2 doesn't require any change in the code structure in general. However,
there is one occurrence where the if/else structure is simplified to assigning
the result of a boolean operation, and one occurrence where I also noticed a
useless use of `ssl_ep_len()` in a TLS-specific branch, that I turned to the
constant 0 as it makes more sense.
Case 1 will be handled in the next commit, as it can easily be handled in an
automated way - only cases 2 and 3 (sometimes) required manual intervention.
The list of occurrences for cases 2 and 3 was established manually by looking
for occurrences of '= MBEDTLS_SSL_TRANSPORT_' in the code and manually
checking if there was a TLS-specific branch.
New sizes (see previous commit for the measuring script):
```
both
text data bss dec hex filename
1820 0 4 1824 720 debug.o (ex library/libmbedtls.a)
0 0 0 0 0 net_sockets.o (ex library/libmbedtls.a)
548 0 0 548 224 ssl_cache.o (ex library/libmbedtls.a)
11155 0 596 11751 2de7 ssl_ciphersuites.o (ex library/libmbedtls.a)
17156 0 0 17156 4304 ssl_cli.o (ex library/libmbedtls.a)
460 0 0 460 1cc ssl_cookie.o (ex library/libmbedtls.a)
17649 0 0 17649 44f1 ssl_srv.o (ex library/libmbedtls.a)
800 0 0 800 320 ssl_ticket.o (ex library/libmbedtls.a)
39286 60 0 39346 99b2 ssl_tls.o (ex library/libmbedtls.a)
88874 60 600 89534 15dbe (TOTALS)
DTLS-only
text data bss dec hex filename
1820 0 4 1824 720 debug.o (ex library/libmbedtls.a)
0 0 0 0 0 net_sockets.o (ex library/libmbedtls.a)
548 0 0 548 224 ssl_cache.o (ex library/libmbedtls.a)
11155 0 596 11751 2de7 ssl_ciphersuites.o (ex library/libmbedtls.a)
17068 0 0 17068 42ac ssl_cli.o (ex library/libmbedtls.a)
460 0 0 460 1cc ssl_cookie.o (ex library/libmbedtls.a)
17553 0 0 17553 4491 ssl_srv.o (ex library/libmbedtls.a)
800 0 0 800 320 ssl_ticket.o (ex library/libmbedtls.a)
38499 60 0 38559 969f ssl_tls.o (ex library/libmbedtls.a)
87903 60 600 88563 159f3 (TOTALS)
TLS-only
text data bss dec hex filename
1820 0 4 1824 720 debug.o (ex library/libmbedtls.a)
0 0 0 0 0 net_sockets.o (ex library/libmbedtls.a)
548 0 0 548 224 ssl_cache.o (ex library/libmbedtls.a)
11155 0 596 11751 2de7 ssl_ciphersuites.o (ex library/libmbedtls.a)
14912 0 0 14912 3a40 ssl_cli.o (ex library/libmbedtls.a)
460 0 0 460 1cc ssl_cookie.o (ex library/libmbedtls.a)
15868 0 0 15868 3dfc ssl_srv.o (ex library/libmbedtls.a)
800 0 0 800 320 ssl_ticket.o (ex library/libmbedtls.a)
27619 60 0 27679 6c1f ssl_tls.o (ex library/libmbedtls.a)
73182 60 600 73842 12072 (TOTALS)
```
And use those tools in a few places. For now the purpose is just to validate
those tools before using them in all occurrences of transport-specific code.
The effect of these changes was measured with the following script:
```
set -eu
build() {
printf "\n$1\n"
CC=arm-none-eabi-gcc CFLAGS='-Werror -Os -march=armv6-m -mthumb' \
AR=arm-none-eabi-ar LD=arm-none-eabi-ld make clean lib >/dev/null
arm-none-eabi-size -t library/libmbedtls.a
}
git checkout -- include/mbedtls/config.h
scripts/config.pl unset MBEDTLS_NET_C
scripts/config.pl unset MBEDTLS_TIMING_C
scripts/config.pl unset MBEDTLS_FS_IO
scripts/config.pl unset MBEDTLS_ENTROPY_NV_SEED
scripts/config.pl set MBEDTLS_NO_PLATFORM_ENTROPY
build "both"
scripts/config.pl unset MBEDTLS_SSL_PROTO_TLS
build "DTLS-only"
scripts/config.pl set MBEDTLS_SSL_PROTO_TLS
scripts/config.pl unset MBEDTLS_SSL_PROTO_DTLS
scripts/config.pl unset MBEDTLS_SSL_DTLS_HELLO_VERIFY
scripts/config.pl unset MBEDTLS_SSL_DTLS_ANTI_REPLAY
scripts/config.pl unset MBEDTLS_SSL_DTLS_BADMAC_LIMIT
scripts/config.pl unset MBEDTLS_SSL_DTLS_CLIENT_PORT_REUSE
build "TLS-only"
git checkout -- include/mbedtls/config.h
```
The output of the script is as follows:
```
both
text data bss dec hex filename
1820 0 4 1824 720 debug.o (ex library/libmbedtls.a)
0 0 0 0 0 net_sockets.o (ex library/libmbedtls.a)
548 0 0 548 224 ssl_cache.o (ex library/libmbedtls.a)
11155 0 596 11751 2de7 ssl_ciphersuites.o (ex library/libmbedtls.a)
17160 0 0 17160 4308 ssl_cli.o (ex library/libmbedtls.a)
460 0 0 460 1cc ssl_cookie.o (ex library/libmbedtls.a)
17637 0 0 17637 44e5 ssl_srv.o (ex library/libmbedtls.a)
800 0 0 800 320 ssl_ticket.o (ex library/libmbedtls.a)
39322 60 0 39382 99d6 ssl_tls.o (ex library/libmbedtls.a)
88902 60 600 89562 15dda (TOTALS)
DTLS-only
text data bss dec hex filename
1820 0 4 1824 720 debug.o (ex library/libmbedtls.a)
0 0 0 0 0 net_sockets.o (ex library/libmbedtls.a)
548 0 0 548 224 ssl_cache.o (ex library/libmbedtls.a)
11155 0 596 11751 2de7 ssl_ciphersuites.o (ex library/libmbedtls.a)
17072 0 0 17072 42b0 ssl_cli.o (ex library/libmbedtls.a)
460 0 0 460 1cc ssl_cookie.o (ex library/libmbedtls.a)
17565 0 0 17565 449d ssl_srv.o (ex library/libmbedtls.a)
800 0 0 800 320 ssl_ticket.o (ex library/libmbedtls.a)
38953 60 0 39013 9865 ssl_tls.o (ex library/libmbedtls.a)
88373 60 600 89033 15bc9 (TOTALS)
TLS-only
text data bss dec hex filename
1820 0 4 1824 720 debug.o (ex library/libmbedtls.a)
0 0 0 0 0 net_sockets.o (ex library/libmbedtls.a)
548 0 0 548 224 ssl_cache.o (ex library/libmbedtls.a)
11155 0 596 11751 2de7 ssl_ciphersuites.o (ex library/libmbedtls.a)
14916 0 0 14916 3a44 ssl_cli.o (ex library/libmbedtls.a)
460 0 0 460 1cc ssl_cookie.o (ex library/libmbedtls.a)
15852 0 0 15852 3dec ssl_srv.o (ex library/libmbedtls.a)
800 0 0 800 320 ssl_ticket.o (ex library/libmbedtls.a)
27623 60 0 27683 6c23 ssl_tls.o (ex library/libmbedtls.a)
73174 60 600 73834 1206a (TOTALS)
```
It can be seen that a DTLS-only build is now starting to be a bit smaller than
a dual-mode build, which is the purpose of the new build option.
For now the option has no effect.
Adapted existing example config files. The fact that I needed to do this
highlights that this is a slightly incompatible change: existing users need to
update their existing custom configs (if standalone as opposed to based on the
default config) in order to still get the same behaviour.
The alternative would be to have a negative config option (eg NO_TLS or
DTLS_ONLY) but this doesn't fit as nicely with the existing options, so
hopefully the minor incompatibility is acceptable.
I don't think it's worth adding a new component to all.sh:
- builds with both DTLS and TLS are done in the default (and full) config
- TLS-only builds are done with eg config-suite-b.h in test-ref-configs
- a DTLS-only build is done with config-thread.h in test-ref-configs
- builds with none of them (and SSL_TLS_C enabled) are forbidden
Context: During a handshake, the SSL/TLS handshake logic constructs
an instance of ::mbedtls_ssl_session representing the SSL session
being established. This structure contains information such as the
session's master secret, the peer certificate, or the session ticket
issues by the server (if applicable).
During a renegotiation, the new session is constructed aside the existing
one and destroys and replaces the latter only when the renegotiation is
complete. While conceptually clear, this means that during the renegotiation,
large pieces of information such as the peer's CRT or the session ticket
exist twice in memory, even though the original versions are removed
eventually.
This commit removes the simultaneous presence of two peer CRT chains
in memory during renegotiation, in the following way:
- Unlike in the case of SessionTickets handled in the previous commit,
we cannot simply free the peer's CRT chain from the previous handshake
before parsing the new one, as we need to verify that the peer's end-CRT
hasn't changed to mitigate the 'Triple Handshake Attack'.
- Instead, we perform a binary comparison of the original peer end-CRT
with the one presented during renegotiation, and if it succeeds, we
avoid re-parsing CRT by moving the corresponding CRT pointer from the
old to the new session structure.
- The remaining CRTs in the peer's chain are not affected by the triple
handshake attack protection, and for them we may employ the canonical
approach of freeing them before parsing the remainder of the new chain.
Note that this commit intends to not change any observable behavior
of the stack. In particular:
- The peer's CRT chain is still verified during renegotiation.
- The tail of the peer's CRT chain may change during renegotiation.
Context: During a handshake, the SSL/TLS handshake logic constructs
an instance of ::mbedtls_ssl_session representing the SSL session
being established. This structure contains information such as the
session's master secret, the peer certificate, or the session ticket
issues by the server (if applicable).
During a renegotiation, the new session is constructed aside the existing
one and destroys and replaces the latter only when the renegotiation is
complete. While conceptually clear, this means that during the renegotiation,
large pieces of information such as the peer's CRT or the session ticket
exist twice in memory, even though the original versions are removed
eventually.
This commit starts removing this memory inefficiency by freeing the old
session's SessionTicket before the one for the new session is allocated.
The existing test `x509parse_crt()` for X.509 CRT parsing
so far used the generic parsing API `mbedtls_x509_crt_parse()`
capable of parsing both PEM encoded and DER encoded certficates,
but was actually only used with DER encoded input data. Moreover,
as the purpose of the test is the testing of the core DER X.509 parsing
functionality, not the PEM vs. DER dispatch (which is now already tested
in the various `x509_crt_info()` tests), the call can be replaced with a
direct call to `mbedtls_x509_parse_crt_der()`.
This commit does that, and further adds to the test an analogous
call to the new API `mbedtls_x509_parse_crt_der_nocopy()` to test
copyless parsing of X.509 certificates.
Context:
The existing API `mbedtls_x509_parse_crt_der()` for parsing DER
encoded X.509 CRTs unconditionally makes creates a copy of the
input buffer in RAM. While this comes at the benefit of easy use,
-- specifically: allowing the user to free or re-use the input
buffer right after the call -- it creates a significant memory
overhead, as the CRT is duplicated in memory (at least temporarily).
This might not be tolerable a resource constrained device.
As a remedy, this commit adds a new X.509 API call
`mbedtls_x509_parse_crt_der_nocopy()`
which has the same signature as `mbedtls_x509_parse_crt_der()`
and almost the same semantics, with one difference: The input
buffer must persist and be unmodified for the lifetime of the
established instance of `mbedtls_x509_crt`, that is, until
`mbedtls_x509_crt_free()` is called.
To prevent dropping the same message over and over again, the UDP proxy
test application programs/test/udp_proxy _logically_ maintains a mapping
from records to the number of times the record has already been dropped,
and stops dropping once a configurable threshold (currently 2) is passed.
However, the actual implementation deviates from this logical view
in two crucial respects:
- To keep the implementation simple and independent of
implementations of suitable map interfaces, it only counts how
many times a record of a given _size_ has been dropped, and
stops dropping further records of that size once the configurable
threshold is passed. Of course, this is not fail-proof, but a
good enough approximation for the proxy, and it allows to use
an inefficient but simple array for the required map.
- The implementation mixes datagram lengths and record lengths:
When deciding whether it is allowed to drop a datagram, it
uses the total datagram size as a lookup index into the map
counting the number of times a package has been dropped. However,
when updating this map, the UDP proxy traverses the datagram
record by record, and updates the mapping at the level of record
lengths.
Apart from this inconsistency, the introduction of the Connection ID
feature leads to yet another problem: The CID length is not part of
the record header but dynamically negotiated during (potentially
encrypted!) handshakes, and it is hence impossible for a passive traffic
analyzer (in this case our UDP proxy) to reliably parse record headers;
especially, it isn't possible to reliably infer the length of a record,
nor to dissect a datagram into records.
The previous implementation of the UDP proxy was not CID-aware and
assumed that the record length would always reside at offsets 11, 12
in the DTLS record header, which would allow it to iterate through
the datagram record by record. As mentioned, this is no longer possible
for CID-based records, and the current implementation can run into
a buffer overflow in this case (because it doesn't validate that
the record length is not larger than what remains in the datagram).
This commit removes the inconsistency in datagram vs. record length
and resolves the buffer overflow issue by not attempting any dissection
of datagrams into records, and instead only counting how often _datagrams_
of a particular size have been dropped.
There is only one practical situation where this makes a difference:
If datagram packing is used by default but disabled on retransmission
(which OpenSSL has been seen to do), it can happen that we drop a
datagram in its initial transmission, then also drop some of its records
when they retransmitted one-by-one afterwards, yet still keeping the
drop-counter at 1 instead of 2. However, even in this situation, we'll
correctly count the number of droppings from that point on and eventually
stop dropping, because the peer will not fall back to using packing
and hence use stable record lengths.
