Commit Graph

4407 Commits

Author SHA1 Message Date
SimonB
295dfa24e7 Clarifies documentation on reported memory statistics 2016-05-23 14:29:29 +01:00
SimonB
4a3b023172 Adds line numbering in errors for test helpers
Adds to the 'generate_code.pl' tool, support to insert line numbers
before test suite helper code.
2016-05-23 14:29:29 +01:00
SimonB
20273ddc4c Adds reporting of file/line no. in failed tests
Tests in tests/suites will now report the file and line number of
failed test assertions.
2016-05-23 14:29:29 +01:00
SimonB
525b792823 Fixes off by 1 error reported in line number errors 2016-05-23 14:29:29 +01:00
SimonB
ab07135165 Adds line number substitution in test cases
Expanded generate_code.pl to substitute !LINE_NO! in test cases.
2016-05-23 14:29:29 +01:00
SimonB
99cff58958 Fixes memory leak in memory_buffer_alloc.c debug
Debug symbols were being leaked in memory_buffer_alloc.c
2016-05-23 14:29:29 +01:00
Nicholas Wilson
1b666554c9 Silence a clang-analyze warning
The check is already effectively performed later in the function, but
implicitly, so Clang's analysis fail to notice the functions are in
fact safe.  Pulling the check up to the top helps Clang to verify the
behaviour.
2016-05-23 14:29:29 +01:00
Nicholas Wilson
f0021645b0 Refactor slightly to silence a clang-analyze warning
Since the buffer is used in a few places, it seems Clang isn't clever
enough to realise that the first byte is never touched.  So, even though
the function has a correct null check for ssl->handshake, Clang
complains.  Pulling the handshake type out into its own variable is
enough for Clang's analysis to kick in though.
2016-05-23 14:29:28 +01:00
Nicholas Wilson
b47fd5e8c9 Remove a dead store to silence clang-analyze 2016-05-23 14:29:28 +01:00
Nicholas Wilson
e735303026 Shut up a few clang-analyze warnings about use of uninitialized variables
The functions are all safe, Clang just isn't clever enough to realise
it.
2016-05-23 14:29:28 +01:00
Nicholas Wilson
91c68a5e15 Shut up a clang-analyzer warning
The function appears to be safe, since grow() is called with sensible
arguments in previous functions.  Ideally Clang would be clever enough to
realise this.  Even if N has size MBEDTLS_MPI_MAX_LIMBS, which will
cause the grow to fail, the affected lines in montmul won't be reached.
Having this sanity check can hardly hurt though.
2016-05-23 14:29:28 +01:00
Attila Molnar
2791ba1429 Fix handle leak in mbedtls_platform_entropy_poll() on Windows on error 2016-05-23 14:29:28 +01:00
Alexey Skalozub
c8404607ea Move K inside MBEDTLS_SHA512_PROCESS_ALT block
It is used only by `mbedtls_sha512_process()`, and in case `MBEDTLS_SHA512_PROCESS_ALT` is defined, it still cannot be reused because of `static` declaration.
2016-05-23 14:29:28 +01:00
James Cowgill
c788b4cb5a Fix segfault on x32 by using better register constraints in bn_mul.h
On x32, pointers are only 4-bytes wide and need to be loaded using the "movl"
instruction instead of "movq" to avoid loading garbage into the register.

The MULADDC routines for x86-64 are adjusted to work on x32 as well by getting
gcc to load all the registers for us in advance (and storing them later) by
using better register constraints. The b, c, D and S constraints correspond to
the rbx, rcx, rdi and rsi registers respectively.
2016-05-23 14:29:28 +01:00
James Cowgill
5c1e24ca05 Fix build errors on x32 by using the generic 'add' instruction
On x32 systems, pointers are 4-bytes wide and are therefore stored in %e?x
registers (instead of %r?x registers). These registers must be accessed using
"addl" instead of "addq", however the GNU assembler will acccept the generic
"add" instruction and determine the correct opcode based on the registers
passed to it.
2016-05-23 14:29:28 +01:00
Nicholas Wilson
b19bac4d82 Allow test suites to be run on Windows
For a start, they don't even compile with Visual Studio due to strcasecmp
being missing.  Secondly, on Windows Perl scripts aren't executable and have
to be run using the Perl interpreter directly; thankfully CMake is able to
find cygwin Perl straight away without problems.
2016-05-23 14:29:28 +01:00
Janos Follath
c6dab2b029 Fix non compliance SSLv3 in server extension handling.
The server code parses the client hello extensions even when the
protocol is SSLv3 and this behaviour is non compliant with rfc6101.
Also the server sends extensions in the server hello and omitting
them may prevent interoperability problems.
2016-05-23 14:27:02 +01:00
Janos Follath
00efff7469 Add a test for SSLv3 with extensions, server side
This test verifies if the server parses or sends extensions when
the protocol is SSLv3.
2016-05-06 13:57:19 +01:00
Attila Molnar
0b98d2f086 Fix minor doc issue 2016-05-02 11:06:47 +02:00
Simon Butcher
45732c7cac Update ChangeLog for bug #429 in ssl_fork_server 2016-04-29 00:12:53 +01:00
Janos Follath
582a461a49 Improves and makes pretty the ssl_fork_server output 2016-04-29 00:12:35 +01:00
Janos Follath
fe049db8ef Fix issue #429 in ssl_fork_server.c 2016-04-29 00:12:19 +01:00
Simon Butcher
b2d5dd105d Fixes X509 sample app and SSL test suite
Fixes the X.509 cert_app and the SSL test suite for the non-default
configs which don't build with if MBEDTLS_PLATFORM_C isn't defined.
2016-04-27 13:35:37 +01:00
Simon Butcher
d3138c35c6 Fixes SSL sample apps for non-default configs
Fixes the SSL sample applications to build for the non-default configs
which don't build if MBEDTLS_PLATFORM_C isn't defined.
2016-04-27 01:26:50 +01:00
Simon Butcher
80aea30aa5 Fixes syntax and naming for check-names.sh
Some macros were failing checks by check-names.sh
2016-04-27 00:28:14 +01:00
Simon Butcher
3fe6cd3a2d Fixes time() abstraction for custom configs
Added platform abstraction of time() to ChangeLog, version features, and fixed the build for dynamic configuration.
2016-04-26 19:51:29 +01:00
SimonB
d5800b7761 Abstracts away time()/stdlib.h into platform
Substitutes time() into a configurable platform interface to allow it to be
easily substituted.
2016-04-26 14:49:59 +01:00
SimonB
1594210a49 Adds better support to debug generated code
The commit adds to the generate_code.pl script support to add #line directives
to generated code to allow build breaks to be more easily found from the
generated code.
2016-04-26 14:46:56 +01:00
Simon Butcher
a543d11d3a Fixes mbedtls_mpi_zeroize() function name in ChangeLog 2016-04-26 12:51:37 +01:00
Simon Butcher
d7e9ad7d83 Updates ChangeLog with faster MPI zeroize fix
Added optimised mbedtls_mpi_zeroise() credit to ChangeLog.
2016-04-25 16:07:12 +01:00
Alexey Skalozub
e17a8da17e Rename MPI zeroize function to mbedtls_mpi_zeroize
Avoid naming confusion
2016-04-25 16:01:07 +01:00
Alexey Skalozub
3d53f41638 Faster mbedtls_zeroize for MPI
Writes in `sizeof(mbedtls_mpi_uint)` units perform faster than plain chars, also eliminates multiplication by `ciL`
2016-04-25 16:00:50 +01:00
Janos Follath
8a3170571e Fix bug in ssl_write_supported_elliptic_curves_ext
Passing invalid curves to mbedtls_ssl_conf_curves potentially could caused a
crash later in ssl_write_supported_elliptic_curves_ext. #373
2016-04-22 00:41:54 +01:00
Janos Follath
55abc21521 Fix ci break in builds without platform.h 2016-04-19 15:15:53 +01:00
Janos Follath
1ed9f99ef3 Fix null pointer dereference in the RSA module.
Introduced null pointer checks in mbedtls_rsa_rsaes_pkcs1_v15_encrypt
2016-04-19 10:16:31 +01:00
SimonB
8ca7bc42d0 Adds verbose mode to the test suites
Added a verbose option to the generated test suites which can list the
dependencies not met for skipped test cases.
Also clarifies internal interfaces between the main_test.function and test code,
and fixed a bug on calculating available tests in run-test-suites.pl.
2016-04-17 23:24:50 +01:00
SimonB
098a3b5025 Makes basic-build-test.sh test the full config
Previously the test worked on the default configuration which missed deprecated
or legacy features. This change tests the full configuration and all available
tests.
2016-04-16 21:56:59 +01:00
SimonB
2e23c82753 Adds check to avoid overwriting files
Adds check to avoid accidental overwriting of config.h or the yotta module, as
well as a force option to override any changes.
2016-04-16 21:54:39 +01:00
Simon Butcher
3f5c875654 Adds test for odd bit length RSA key size
Also tidy up ChangeLog following review.
2016-04-15 19:06:59 +01:00
Janos Follath
10c575be3e Fix odd bitlength RSA key generation
Fix issue that caused a hang up when generating RSA keys of odd
bitlength.
2016-04-15 18:49:13 +01:00
Manuel Pégourié-Gonnard
0e4d9afa61 Use Freescale's RNGA on the K64F 2016-04-15 10:40:31 +01:00
Simon Butcher
6eaf3659ea Fixes Travis post-mortem script dump following review
Changes made:
 * Added copyright and project statement
 * Limited size of each file to dump to 1Mbyte
 * Changed name of script
2016-04-10 15:11:27 +01:00
Simon Butcher
371a9e0ae9 Merge branch 'iotssl-683-travis-failing-intermittently' 2016-04-10 14:24:31 +01:00
SimonB
73883c12bc Fixes formatting of spacing in config.pl 2016-04-04 13:49:10 +01:00
SimonB
ba9dd1ec13 Adds to footprint.sh MBEDTLS_NO_PLATFORM_ENTROPY
For baremetal builds MBEDTLS_NO_PLATFORM_ENTROPY must now be set to avoid the
build failing.

Fixes #449.
2016-04-03 15:06:52 +01:00
SimonB
60f2cf93f5 Adds option to config.pl to force config changes
The script config.pl fails when setting a #define symbol if the symbol isn't
already in the configuration header. This adds an option '--force' to append
the symbol to the end of the file if it isn't already present.

Also clarified usage, and added copyright to the config.pl.
2016-04-03 14:16:08 +01:00
Simon Butcher
a7ffc8f739 Update the yotta module version number
The minor version must rise to allow other software with dependencies on mbed
TLS to be dependent on the next version following the 2016 Q1 release.
2016-03-23 16:22:24 +00:00
Simon Butcher
3527514e94 Update the yotta module number
Changed the yotta module number to 2.2.3
2016-03-23 15:38:37 +00:00
Simon Butcher
f2d18a29da Merge branch 'development' 2016-03-21 22:55:56 +00:00
Simon Butcher
cd0ee5e499 Fixes following review of 'iotssl-682-selftest-ci-break' 2016-03-21 22:54:37 +00:00