mbedtls_md_setup() allocates a hash-specific context and then, if
requested, an extra HMAC context. If the second allocation failed, the
hash context was not freed.
Fix this by ensuring that the mbedtls_md_context_t object is always in
a consistent state, in particular, that the md_info field is always
set. For robustness, ensure that the object is in a consistent state
even on errors (other than BAD_INPUT_DATA if the object was not in a
consistent state on entry).
Fix#3486
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The tests are supposed to be failing now (in all.sh component
test_memsan_constant_flow), but they don't as apparently MemSan doesn't
complain when the src argument of memcpy() is uninitialized, see
https://github.com/google/sanitizers/issues/1296
The next commit will add an option to test constant flow with valgrind, which
will hopefully correctly flag the current non-constant-flow implementation.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
This paves the way for a constant-flow implementation of HMAC checking, by
making sure that the comparison happens at a constant address. The missing
step is obviously to copy the HMAC from the secret offset to this temporary
buffer with constant flow, which will be done in the next few commits.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
As a result, the copyright of contributors other than Arm is now
acknowledged, and the years of publishing are no longer tracked in the
source files.
Also remove the now-redundant lines declaring that the files are part of
MbedTLS.
This commit was generated using the following script:
# ========================
#!/bin/sh
# Find files
find '(' -path './.git' -o -path './3rdparty' ')' -prune -o -type f -print | xargs sed -bi '
# Replace copyright attribution line
s/Copyright.*Arm.*/Copyright The Mbed TLS Contributors/I
# Remove redundant declaration and the preceding line
$!N
/This file is part of Mbed TLS/Id
P
D
'
# ========================
Signed-off-by: Bence Szépkúti <bence.szepkuti@arm.com>
RFC5280 does not state that the `revocationDate` should be checked.
In addition, when no time source is available (i.e., when MBEDTLS_HAVE_TIME_DATE is not defined), `mbedtls_x509_time_is_past` always returns 0. This results in the CRL not being checked at all.
https://tools.ietf.org/html/rfc5280
Signed-off-by: Raoul Strackx <raoul.strackx@fortanix.com>
Description referred to mbedtls_ssl_sent_t callback,
but the callback is named mbedtls_ssl_send_t.
Signed-off-by: Christopher Moynihan <christophm@gmail.com>
Three tests were guarded by `MBEDTLS_KEY_EXCHANGE_ECJPAKE`,
not `MBEDTLS_KEY_EXCHANGE_ECJPAKE_ENABLED`, as it should be.
Curiously, the guard still functioned as intended, perhaps
because `MBEDTLS_KEY_EXCHANGE_ECJPAKE` is a prefix of
`MBEDTLS_KEY_EXCHANGE_ECJPAKE_ENABLED`.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
The debug output for supported ciphersuites has been changed
from `deadbeef` to `0xdeadbeef` in a previous commit, but the
test script `ssl-opt.sh` grepping for lines in the debug log
to determine test success/failure hadn't been adjusted accordingly.
This commit fixes this.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
The client previously reproted the offered ciphersuites through
their numerical identifier only, while the server reported them
through their name.
This commit modifies the debug output on client and server to
both use the format `ID (NAME)` for the ciphersuites.
The example application programs/ssl/ssl_client2 allows the
configuration of a client CRT through the parameters
- crt_file, key_file
However, password protected key files are not supported.
This commit adds a new command line option
- key_pwd
which allow to specify a password for the key file specified
in the key_file parameter.
The example application programs/ssl/ssl_server2 allows the
configuration of up to two CRTs through the command line
parameters
- crt_file, key_file
- crt_file2, key_file2.
However, password protected key files are not supported.
This commit adds command line options
- key_pwd
- key_pwd2
which allow to specify passwords for the key files specified
in key_file and key_file2, respectively.
* development: (55 commits)
Log change as bugfix
Add changelog entry
Clarify updates to the persistent state in storage
With multiple applicable transparent drivers, the order is unspecified
Minor clarifications
Give some examples of purpsoses of pure-software transparent driver
Fix typos
Add a link to the PSA API specification
Explain locations vs lifetimes
Initialize key pointer in ecdh to NULL
Add buffer zeroization when ecp_write_key fails
Simplified key slot deletion
Style fixes
Use arc4random_buf instead of rand on NetBSD
Apply review feedback
Update open question section about public key storage
Remove the paragraph about declaring application needs
Change driver persistent data to a callback interface
Rework and expand key management in opaque drivers
Fix typos and copypasta
...
Mention explicitly that only DNS names are supported so far, and while at it
explain where the name is searched.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
There is little point in leaving the order in which drivers are
considered unspecified. This gives flexibility to the implementation
for a process that is generally performed at build time, not in a
constrained environment. Having a well-defined order is especially
useful with fallback.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
It's ok if they map to the same function names and an error otherwise.
It's an error to have multiple opaque drivers for the same location.
If multiple transparent drivers apply, which one applies is unspecified.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The existing description of the syntax of capabilities also describes
the semantics of each property, but the semantics of the capability as
a whole is not immediately clear. Add a subsection that explains
precisely when a capability is applicable.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Inferring entry points from algorithms is nice in that it makes
capability specifications shorter and less redundant, but that's not
really important. It also makes capabilities more fragile: if the core
starts supporting new mechanisms based on the same algorithm (for
example, adding hash-and-sign when only sign-the-hash existed before),
a driver only supporting the old mechanisms would fail at build time.
So make entry points mandatory.
This has the benefit of making the semantics of capabilities easier to
describe.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
