yuzu-emu/mbedtls
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
4559 commits 1 branch 0 tags 94 MiB
Commit graph

2439 commits

Author SHA1 Message Date
Manuel Pégourié-Gonnard
58fb49531d Fix potential buffer overflow in mpi_read_string() 
Found by Guido Vranken.

Two possible integer overflows (during << 2 or addition in BITS_TO_LIMB())
could result in far too few memory to be allocated, then overflowing the
buffer in the subsequent for loop.

Both integer overflows happen when slen is close to or greater than
SIZE_T_MAX >> 2 (ie 2^30 on a 32 bit system).

Note: one could also avoid those overflows by changing BITS_TO_LIMB(s << 2) to
CHARS_TO_LIMB(s >> 1) but the solution implemented looks more robust with
respect to future code changes.
 2015-09-28 15:59:54 +02:00
Simon Butcher
ed9976634f Added bounds checking for TLS extensions 
IOTSSL-478 - Added checks to prevent buffer overflows.
 2015-09-28 02:14:30 +01:00
Simon Butcher
89f77623b8 Added max length checking of hostname 2015-09-27 22:50:49 +01:00
Tillmann Karras
588ad50c5a Fix a fairly common typo in comments 2015-09-25 04:27:22 +02:00
Manuel Pégourié-Gonnard
faee44ded1 Avoid false positives in bounds check 
The size of the buffer already accounts for the extra data before the actual
message, so the allowed length is SSL_MAX_CONTENT_LEN starting from _msg
 2015-09-24 22:19:58 +02:00
Manuel Pégourié-Gonnard
d0d8cb36a4 Cache ClientHello extension 
This extension is quite costly to generate, and we don't want to re-do it
again when the server performs a DTLS HelloVerify. So, cache the result the
first time and re-use if/when we build a new ClientHello.

Note: re-send due to timeouts are different, as the whole message is cached
already, so they don't need any special support.
 2015-09-17 14:16:30 +02:00
Manuel Pégourié-Gonnard
77c0646ef2 Add cache for EC J-PAKE client extension 
Not used yet, just add the variables and cleanup code.
 2015-09-17 13:59:49 +02:00
Manuel Pégourié-Gonnard
8cea8ad8b8 Bump version to 2.1.1 2015-09-17 11:58:45 +02:00
Simon Butcher
ac58c53ab1 Merge remote-tracking branch 'origin/development' 2015-09-16 23:25:25 +01:00
Manuel Pégourié-Gonnard
0f1660ab4f Implement key exchange messages and PMS derivation 
This completes the first working version. No interop testing done yet.
 2015-09-16 22:58:30 +02:00
Manuel Pégourié-Gonnard
25dbeb002d Skip certificate-related messages with ECJPAKE 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
0a1324aaa1 Add client-side extension parsing 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
55c7f99112 Add server writing of the extension 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
bf57be690e Add server extension parsing 
Only accept EC J-PAKE ciphersuite if extension was present and OK (single flag
for both), and ignore extension if we have no password.
 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
e511b4e7cb Ignore ECJPAKE suite if not configured on server 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
c1b46d0242 Fix bug in server parsing point formats extension 
This bug becomes noticeable when the extension following the "supported point
formats" extension has a number starting with 0x01, which is the case of the
EC J-PAKE extension, which explains what I noticed the bug now.

This will be immediately backported to the stable branches,
see the corresponding commits for impact analysis.
 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
60884a1597 Improve debug formatting of ciphersuites 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
eef142d753 Depend on ECJPAKE key exchange, not module 
This is more consistent, as it doesn't make any sense for a user to be able to
set up an EC J-PAKE password with TLS if the corresponding key exchange is
disabled.

Arguably this is what we should de for other key exchanges as well instead of
depending on ECDH_C etc, but this is an independent issue, so let's just do
the right thing with the new key exchange and fix the other ones later. (This
is a marginal issue anyway, since people who disable all ECDH key exchange are
likely to also disable ECDH_C in order to minimize footprint.)
 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
ddf97a6c92 Skip ECJPAKE suite in ClientHello if no pw set up 
When we don't have a password, we want to skip the costly process of
generating the extension. So for consistency don't offer the ciphersuite
without the extension.
 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
538cb7b0b4 Add the ECJPAKE ciphersuite 2015-09-16 22:58:29 +02:00
Manuel Pégourié-Gonnard
557535d8c4 Add ECJPAKE key exchange 2015-09-16 22:58:29 +02:00
Simon Butcher
7dd82f8fd5 Merge branch 'development' with bugfix branch 
Conflicts:
	ChangeLog
 2015-09-16 16:21:38 +01:00
Simon Butcher
5793e7ef01 Merge 'development' into iotssl-411-port-reuse 
Conflicts:
	ChangeLog
 2015-09-16 15:25:53 +01:00
Manuel Pégourié-Gonnard
294139b57a Add client extension writing 2015-09-16 16:10:48 +02:00
Manuel Pégourié-Gonnard
b813accf84 Add mbedtls_ecjpake_check(), tells if set up 
This will be used in SSL to avoid the computation-heavy processing of EC
J-PAKE hello extensions in case we don't have an EC J-PAKE password
 2015-09-16 16:10:48 +02:00
Manuel Pégourié-Gonnard
7002f4a560 Add mbedtls_ssl_set_hs_ecjpake_password() 2015-09-16 16:10:48 +02:00
Manuel Pégourié-Gonnard
f7022d1131 Fix bug in server parsing point formats extension 
There is only one length byte but for some reason we skipped two, resulting in
reading one byte past the end of the extension. Fortunately, even if that
extension is at the very end of the ClientHello, it can't be at the end of the
buffer since the ClientHello length is at most SSL_MAX_CONTENT_LEN and the
buffer has some more room after that for MAC and so on. So there is no
buffer overread.

Possible consequences are:
- nothing, if the next byte is 0x00, which is a comment first byte for other
  extensions, which is why the bug remained unnoticed
- using a point format that was not offered by the peer if next byte is 0x01.
  In that case the peer will reject our ServerKeyExchange message and the
handshake will fail.
- thinking that we don't have a common point format even if we do, which will
  cause us to immediately abort the handshake.
None of these are a security issue.

The same bug was fixed client-side in fd35af15
 2015-09-16 11:32:18 +02:00
Manuel Pégourié-Gonnard
76cfd3f97f Add EC J-PAKE context in handshake structure 2015-09-15 18:24:08 +02:00
Manuel Pégourié-Gonnard
f472179d44 Adjust dependencies for EC extensions 
The Thread spec says we need those for EC J-PAKE too.
However, we won't be using the information, so we can skip the parsing
functions in an EC J-PAKE only config; keep the writing functions in order to
comply with the spec.
 2015-09-15 18:22:00 +02:00
Manuel Pégourié-Gonnard
ea5370d4a2 Don't allow reconnect during handshake 
Especially for resumed handshake, it's entirely possible for an epoch=0
ClientHello to be retransmitted or arrive so late that the server is already
at epoch=1. There is no good way to detect whether it's that or a reconnect.

However:
- a late ClientHello seems more likely that client going down and then up
  again in the middle of a handshake
- even if that's the case, we'll time out on that handshake soon enough
- we don't want to break handshake flows that used to work
So the safest option is to not treat that as a reconnect.
 2015-09-15 15:17:54 +02:00
Simon Butcher
49641ad799 Merge pull request #275 from embedthis/fix-1 
FIX: compiler warning with recvfrom on 64-bit
 2015-09-14 19:59:28 +01:00
Simon Butcher
d0bf6a3891 Update ssl_tls.c 
Clarification in comments
 2015-09-11 17:34:49 +01:00
Simon Butcher
74ca8d07ad Update ssl_tls.c 
Clarification in comments to ssl_handle_possible_reconnect()
 2015-09-11 17:22:40 +01:00
Simon Butcher
0789aed39d Update ssl_tls.c 
Typo
 2015-09-11 17:15:17 +01:00
Embedthis Software
a25cab8bea FIX: compiler warning with recvfrom on 64-bit 2015-09-09 08:49:48 -07:00
Manuel Pégourié-Gonnard
ddfe5d20d1 Tune dependencies 
Don't depend on srv.c in config.h, but add explicit checks. This is more
in line with other options that only make sense server-side, and also it
allows to test full config minus srv.c more easily.
 2015-09-09 12:46:16 +02:00
Manuel Pégourié-Gonnard
2ed05a049a Fix typos 2015-09-09 11:52:28 +02:00
Manuel Pégourié-Gonnard
ab05d23b29 Update generated file 2015-09-09 11:50:00 +02:00
Manuel Pégourié-Gonnard
62c74bb78a Stop wasting resources 
Use a custom function that minimally parses the message an creates a reply
without the overhead of a full SSL context.

Also fix dependencies: needs DTLS_HELLO_VERIFY for the cookie types, and let's
also depend on SRV_C as is doesn't make sense on client.
 2015-09-09 11:22:52 +02:00
Nicholas Wilson
2088e2ebd9 fix const-ness of argument to mbedtls_ssl_conf_cert_profile 
Otherwise, it's impossible to pass in a pointer to
mbedtls_x509_crt_profile_next!
 2015-09-08 16:53:18 +01:00
Simon Butcher
e5a21b4493 Merge pull request #282 from ARMmbed/iotssl-469-rsa-crt-restricted 
Add counter-measure against RSA-CRT attack
 2015-09-08 13:05:51 +01:00
Manuel Pégourié-Gonnard
5f50104c52 Add counter-measure against RSA-CRT attack 
https://securityblog.redhat.com/2015/09/02/factoring-rsa-keys-with-tls-perfect-forward-secrecy/
 2015-09-08 13:39:29 +02:00
Manuel Pégourié-Gonnard
3f09b6d4c2 Fix API 2015-09-08 11:58:14 +02:00
Manuel Pégourié-Gonnard
be619c1264 Clean up error codes 2015-09-08 11:21:21 +02:00
Manuel Pégourié-Gonnard
11331fc25b First working dirty version 
- uses too much resources
- wrong API
 2015-09-08 10:39:06 +02:00
Manuel Pégourié-Gonnard
9650205df7 Start detecting epoch 0 ClientHellos 2015-09-08 10:39:06 +02:00
Manuel Pégourié-Gonnard
d9802af1d0 Add tests for round 2 
Also move one check earlier as it makes more sense
 2015-09-07 12:43:11 +02:00
Manuel Pégourié-Gonnard
3059095e86 Complete tests for reading round one 
Also change the code to forbid public keys being 0
 2015-09-07 12:43:11 +02:00
Manuel Pégourié-Gonnard
d0d8a935b2 Blind operations on the secret 
I'm not sure this is necessary, because it is only multiplied by xm2 which is
already random and secret, but OTOH, xm2 is related to a public value, so
let's add blinding with a random value that's only use for blinding, just to
be extra sure.
 2015-09-07 12:43:11 +02:00
Manuel Pégourié-Gonnard
c907081a20 Polish the source 2015-09-07 12:43:11 +02:00
Manuel Pégourié-Gonnard
f7368c983a Polish API and documentation 2015-09-07 12:43:11 +02:00
Manuel Pégourié-Gonnard
e1927101fb Unify round two 2015-09-07 12:43:11 +02:00
Manuel Pégourié-Gonnard
d8204a7bea Provide symmetric API for the first round 2015-09-07 12:43:11 +02:00
Manuel Pégourié-Gonnard
e2d3a4e1b4 Unify loading of test vectors in tests 2015-09-07 12:43:11 +02:00
Manuel Pégourié-Gonnard
ce4567614b Rename variable to prepare for cli/srv unification 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
6b798b9dae Tune up some comments 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
e0ad57b0b3 Replace explicit IDs with table look-ups 
That's a first step towards merging symmetric version of different functions
 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
5f18829609 Add derive_pms, completing first working version 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
6449391852 Store our role in the context 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
614bd5e919 Add write_client_params 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
ec0eece2ba Add read_client_params 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
bed9e41761 Add writing of server params 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
8d31e80da4 Improve testing strategy 
- reference handshake tests that we get the right values (not much now, but
  much more later when we get to deriving the PMS)
- random handshake in addition tests our generate/write functions against our
  read functions, that are tested by the reference handshake, and will be
further tested in the test suite later against invalid inputs
 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
1a7c5ef42b Optimize some case of mbedtls_ecp_muladd() 
Those are used by EC-JPAKE
 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
cb7cd03412 Add first draft or read_server_params 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
23dcbe3f16 Add support for passphrase in the context 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
4e8bc78ad9 Add context-using functions for Hello extensions 
Also re-order functions in the header so that they appear in the order they're
use, ie free() last.
 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
7af8bc1007 Start introducing mbedtls_ecjpake_context 2015-09-07 12:43:10 +02:00
Manuel Pégourié-Gonnard
3aed1851b3 Re-order functions. 
Use the same order for all read-write pair of functions
 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
9028c5af9a Improve const correctness of read() functions 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
082767ff0c Add ecjpake_kkpp_read/write 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
4f2cd95e1d Fix potential memory leaks 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
b1b250b68c Add ecjpake_kkp_read/write() 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
967cd7192d Add test vector for ZKP verification 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
6029a85572 Add ecjpake_zpk_read() 
Not really tested yet
 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
c618195bc4 Fix base point in ecjpake_write_zkp() 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
d9a3f47ecd Add mbedtls_ecp_gen_keypair_base() 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
8489f17277 First draft of ecjpake_write_zkp() 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
3dbf2fbb89 Implement hashing function for ZKP 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
4d8685b4ff Add skeleton for EC J-PAKE module 2015-09-07 12:43:09 +02:00
Manuel Pégourié-Gonnard
7f2f062a5d Fix possible client crash on API misuse 2015-09-07 12:27:24 +02:00
Manuel Pégourié-Gonnard
aac5502553 Bump version to 2.1.0 2015-09-04 14:33:31 +02:00
Manuel Pégourié-Gonnard
37ff14062e Change main license to Apache 2.0 2015-09-04 14:21:07 +02:00
Simon Butcher
52754594b6 Merging iotssl-457-badtail with development branch 2015-09-03 13:06:01 +01:00
Manuel Pégourié-Gonnard
fdbdd72b8b Skip to trusted certs early in the chain 
This helps in the case where an intermediate certificate is directly trusted.
In that case we want to ignore what comes after it in the chain, not only for
performance but also to avoid false negatives (eg an old root being no longer
trusted while the newer intermediate is directly trusted).

closes #220
 2015-09-01 17:24:42 +02:00
Manuel Pégourié-Gonnard
4d04cdcd12 Fix RSA mutex fix 
Once the mutex is acquired, we must goto cleanup rather that return.
Since cleanup adjusts the return value, adjust that in test cases.

Also, at cleanup we don't want to overwrite 'ret', or we'll loose track of
errors.

see #257
 2015-08-31 09:31:55 +02:00
Manuel Pégourié-Gonnard
1385a289f4 Fix possible mutex lock/unlock mismatch 
fixes #257
 2015-08-27 11:30:58 +02:00
Manuel Pégourié-Gonnard
e578b1c79a Relax timing_self_test for windows idiosyncrasies 2015-08-18 20:11:48 +02:00
Manuel Pégourié-Gonnard
c98204e68f Fix missing break in switch for SSL presets 
closes #235
 2015-08-11 04:21:01 +02:00
Manuel Pégourié-Gonnard
91bbfb6fb7 Make timing selftest less sensitive 
- allow up to 12.5% security/error margin
- use larger delays
- this avoid the security/error margin being too low

The test used to fail about 1 out of 6 times on some buildbots VMs, but never
failed on the physical machines used for development.
 2015-08-10 14:33:12 +02:00
Manuel Pégourié-Gonnard
ed46c436c0 Fix error when loading libmbedtls.so 2015-08-10 10:17:32 +02:00
Manuel Pégourié-Gonnard
111ce9f735 Fix build error with shared libraries for windows 2015-08-07 12:07:16 +02:00
Manuel Pégourié-Gonnard
8018c28600 Add -static-libgcc for Windows dll builds 2015-08-07 11:55:56 +02:00
Manuel Pégourié-Gonnard
620ee19823 Fix return of x509_self_test without SHA-1 
No being able to run the test is not a failure
 2015-08-07 10:57:47 +02:00
Manuel Pégourié-Gonnard
d1004f02e6 Fix printed output of some selftests 2015-08-07 10:57:41 +02:00
Manuel Pégourié-Gonnard
0a8857435c DTLS: treat bad MAC on Finished as an error 
This is not required nor recommended by the protocol, and it's a layering
violation, but it's a know flaw in the protocol that you can't detect a PSK
auth error in any other way, so it is probably the right thing to do.

closes #227
 2015-08-04 12:11:17 +02:00
Manuel Pégourié-Gonnard
052d10c9d5 Accept a trailing space at end of PEM lines 
With certs being copy-pasted from webmails and all, this will probably become
more and more common.

closes #226
 2015-07-31 11:11:26 +02:00
Manuel Pégourié-Gonnard
6fb8187279 Update date in copyright line 2015-07-28 17:11:58 +02:00
Simon Butcher
10a6f02f83 Merge branch 'development' into IOTSSL-442-hello-noext 
Conflicts:
	ChangeLog
 2015-07-27 13:45:40 +01:00
Manuel Pégourié-Gonnard
6f42417ba8 Fix typo in that broke installation in cmake 
closes #221
 2015-07-24 16:55:22 +02:00
Manuel Pégourié-Gonnard
4cc8c63226 Add test for extensionless ClientHello 2015-07-23 12:24:03 +02:00
Manuel Pégourié-Gonnard
a6e5bd5654 Fix bug with extension-less ServerHello 
https://tls.mbed.org/discussions/bug-report-issues/server-hello-parsing-bug

in_hslen include the length of the handshake header. (We might want to change
that in the future, as it is a bit annoying.)
 2015-07-23 12:23:19 +02:00
Manuel Pégourié-Gonnard
cb0d212c97 Fix level of some debug messages 2015-07-22 11:52:11 +02:00
Manuel Pégourié-Gonnard
b076116e14 Fix one debug message 2015-07-22 11:39:23 +02:00
Manuel Pégourié-Gonnard
1bab7d7064 Fix blank line in comments 2015-07-13 09:06:18 +01:00
Paul Bakker
4cb87f409d Prepare for 2.0.0 release 2015-07-10 14:09:43 +01:00
Manuel Pégourié-Gonnard
fc2ccfe72c Fix missing comma with ENTROPY_HARDWARE_ALT 2015-07-10 11:15:50 +01:00
Manuel Pégourié-Gonnard
a4f055fe0c Some windows environments don't have _snprint_s 
Do an alternative version for them.
That happens for example with our windows buildbot with mingw32-make.
 2015-07-08 17:35:37 +02:00
Manuel Pégourié-Gonnard
20af64dc2c Still need to #define inline for MSVC 
I only tested with VS2015 earlier, but previous versions apparently still
don't know that standard C99 keyword though it's documented on MSDN...
 2015-07-07 23:21:30 +02:00
Manuel Pégourié-Gonnard
e540b49a3f Add one more debug message 2015-07-07 12:44:38 +02:00
Manuel Pégourié-Gonnard
b4b19f395f Add a debug message 2015-07-07 11:41:21 +02:00
Manuel Pégourié-Gonnard
001f2b6246 Use xxx_clone() instead of memcpy() in SSL 2015-07-06 16:54:51 +02:00
Manuel Pégourié-Gonnard
c0bf01e8d2 Undo overzealous renaming of internal variables 
The rename script couldn't know it was a local variable with the same name as
on of the global functions
 2015-07-06 16:26:23 +02:00
Manuel Pégourié-Gonnard
052a6c9cfe Add mbedtls_md_clone() 2015-07-06 16:06:02 +02:00
Manuel Pégourié-Gonnard
16d412f465 Add md/shaXXX_clone() API 
Will be used in the SSL/TLS modules
 2015-07-06 15:48:34 +02:00
Manuel Pégourié-Gonnard
b9d64e5bbe Fix missing calls to md/shaxxx_free() 2015-07-06 14:18:56 +02:00
Manuel Pégourié-Gonnard
71d296a15d Add missing calls to _free() in md_wrap 
We can't just assume xxx_free() will only zeroize, because of alternative
implementations
 2015-07-06 11:36:25 +02:00
Manuel Pégourié-Gonnard
5791109707 Make the hardclock test optional 
Known to fail on VMs (such as the buildbots), see eg
http://blog.badtrace.com/post/rdtsc-x86-instruction-to-detect-vms/
 2015-07-01 19:22:12 +02:00
Manuel Pégourié-Gonnard
9bd0afdb22 Add guards for closed socket in net.c 
This is particularly problematic when calling FD_SET( -1, ... ), but let's
check it in all functions.

This was introduced with the new API and the fact the net_free() now sets the
internal fd to -1 in order to mark it as closed: now using this information.
 2015-07-01 19:03:27 +02:00
Manuel Pégourié-Gonnard
2505528be4 Rm obsolete defines for inline wiht MSVC 
The "inline" keyword is supported since Visual Studio 2005 according to MSDN,
and we require Visual Studio 2010 or higher.
 2015-07-01 17:22:36 +02:00
Manuel Pégourié-Gonnard
636741b176 Remove obsolete hacks for uin32_t 
We now require support for stdint.h from the compiler.
 2015-07-01 17:13:05 +02:00
Manuel Pégourié-Gonnard
9de64f5af1 Fix MSVC warnings in library and programs 2015-07-01 16:56:08 +02:00
Manuel Pégourié-Gonnard
acecb653d5 Fix mbedtls_net_usleep() on Windows 
For some reason select() doesn't seem to work.
 2015-07-01 12:00:56 +02:00
Manuel Pégourié-Gonnard
abc729e664 Simplify net_accept() with UDP sockets 
This is made possible by the new API where net_accept() gets a pointer to
bind_ctx, so it can update it.
 2015-07-01 01:28:24 +02:00
Manuel Pégourié-Gonnard
db2468d7aa Update old comment 2015-06-30 17:19:48 +02:00
Manuel Pégourié-Gonnard
3d7d00ad23 Rename mbedtls_net_close() to mbedtls_net_free() 
close() may be more meaningful, but free() is symmetric with _init(), and more
consistent with all other modules
 2015-06-30 16:50:37 +02:00
Manuel Pégourié-Gonnard
91895853ac Move from naked int to a structure in net.c 
Provides more flexibility for future changes/extensions.
 2015-06-30 15:56:25 +02:00
Manuel Pégourié-Gonnard
16a17a496c Fix net_accept() for UDP sockets on Windows 
On Windows, recvfrom() returns an error code if the destination buffer is too
small to hold the next datagram.
 2015-06-30 11:31:10 +02:00
Manuel Pégourié-Gonnard
a16e7c468c Rename a debug function 2015-06-29 20:14:19 +02:00
Manuel Pégourié-Gonnard
80d627a5ae Remove now useless function 2015-06-29 20:12:51 +02:00
Manuel Pégourié-Gonnard
b74c245a20 Rework debug to not need dynamic alloc 
But introduces dependency on variadic macros
 2015-06-29 20:08:23 +02:00
Manuel Pégourié-Gonnard
fa67ebaebb Fix X.509 keysize check with multiple CAs 
Assume we have two trusted CAs with the same name, the first uses ECDSA 256
bits, the second RSA 2048; cert is signed by the second. If we do the keysize
check before we checked the key types match, we'll raise the badkey flags when
checking the EC-256 CA and it will remain up even when we finally find the
correct CA. So, move the check for the key size after signature verification,
which implicitly checks the key type.
 2015-06-27 14:41:38 +02:00
Manuel Pégourié-Gonnard
f659d2cd40 Tune up Windows snprintf() support 
When we build with Visual Studio in debug mode, the invalid parameter handler
aborts the application (and offers to debug it) when n is 0. We want to
just return -1 instead (as calls with n == 0 are expected and happen in our
tests).
 2015-06-26 17:45:00 +02:00
Manuel Pégourié-Gonnard
e1d34d1707 cmake: add shortcut 'lib' for all libraries 
Name chosen to match the existing make target.
 2015-06-25 14:53:13 +02:00
Manuel Pégourié-Gonnard
574ae18088 Fix stupid typo that broke make SHARED=1 2015-06-25 14:30:51 +02:00
Manuel Pégourié-Gonnard
147be4f315 Rm old variable from (c)make files 2015-06-25 11:57:13 +02:00
Manuel Pégourié-Gonnard
752c501126 One soversion per library 2015-06-25 11:56:17 +02:00
Manuel Pégourié-Gonnard
c7781addcb Split library in CMake 2015-06-25 10:59:57 +02:00
Manuel Pégourié-Gonnard
ba2c8763b6 cmake: adjust libraries linking 2015-06-25 10:59:57 +02:00
Manuel Pégourié-Gonnard
216a1831de Fix whitespace in CMakeLists.txt 
- all spaces no tabs
- indent with 4 spaces everywhere
 2015-06-25 10:59:57 +02:00
Manuel Pégourié-Gonnard
cde2aba0af Do no test net_usleep in timing_selftest 
Timing belongs in libcrypto (due to havege depending on it)
while net.c was put in libtls (only test ssl servers use it)
 2015-06-25 10:59:57 +02:00
Manuel Pégourié-Gonnard
53585eeb17 Remove test DHM params from certs.c 
certs.c belongs to the X.509 library, while DHM belongs to the crypto lib.
 2015-06-25 10:59:57 +02:00
Manuel Pégourié-Gonnard
0761733c1b Fix potential NULL dereference 
We document that either of recv or recv_timeout may be NULL, but for TLS we
always used recv... Thanks Coverity for catching that.
(Not remotely trigerrable: local configuration.)

Also made me notice net_recv_timeout didn't do its job properly.
 2015-06-25 10:59:57 +02:00
Manuel Pégourié-Gonnard
dba460f2f3 Add SSL "assertion" to help static analysis 2015-06-25 10:59:57 +02:00
Manuel Pégourié-Gonnard
1cf7b30dc8 Rewrite test to make Coverity happier 
With the default config, it noticed the accept_comp was always 0, so the rest
of the test was dead code.
 2015-06-25 10:59:56 +02:00
Manuel Pégourié-Gonnard
5c59a4fea5 Split libs with make + general make cleanups 2015-06-25 10:59:56 +02:00
Manuel Pégourié-Gonnard
21dcc1e748 fixup 2015-06-25 10:59:56 +02:00
Manuel Pégourié-Gonnard
463e09d64b Prepare library split 2015-06-24 12:05:33 +02:00
Manuel Pégourié-Gonnard
bae389b4d4 Fix uninitialized access 
Found using Codenomicon Defensics.
 2015-06-24 10:47:33 +02:00
Manuel Pégourié-Gonnard
bcc030849a Avoid fclose( NULL ) 
Found by Coverity Scan.
 2015-06-24 00:09:29 +02:00
Manuel Pégourié-Gonnard
fd474233c8 Change SSL debug API in the library 2015-06-23 18:44:11 +02:00
Manuel Pégourié-Gonnard
79c4e3ee59 Rm obsolete comments 2015-06-23 18:44:10 +02:00
Manuel Pégourié-Gonnard
b86145e6cd Avoid potential NULL dereference. 
May happen with a faulty configuration (eg no allowed curve but trying to use
ECDHE key exchange), but not trigger able remotely.

(Found with Clang's scan-build.)
 2015-06-23 18:44:10 +02:00
Manuel Pégourié-Gonnard
14bf7063b9 Add SSL "assertions" to help static analyzers 
scan-build was reporting NULL dereferences
 2015-06-23 18:44:10 +02:00
Manuel Pégourié-Gonnard
b9c93d0d0a Fix earlier incomplete change in RSA PMS reading 
Probably a bad merge from the 1.3 branch
 2015-06-23 18:43:53 +02:00
Manuel Pégourié-Gonnard
19389753c8 Avoid dead stores (makes scan-build happier) 2015-06-23 13:46:44 +02:00
Manuel Pégourié-Gonnard
c0d749418b Make 'port' a string in NET module 
- avoids dependency on snprintf
- allows using "smtps" instead of "456" if desired
 2015-06-23 13:09:11 +02:00
Manuel Pégourié-Gonnard
d23f593737 Avoid static buffer in debug module 
Caused issues in threading situations
 2015-06-23 13:09:11 +02:00
Manuel Pégourié-Gonnard
96fb685e31 Some more init calls 2015-06-23 13:09:11 +02:00
Manuel Pégourié-Gonnard
496f24e949 Deduplicate SHA-2 wrappers 2015-06-23 13:09:11 +02:00
Manuel Pégourié-Gonnard
ab5932192a Call init functions in MD alloc wrappers 
When someone defines MBEDTLS_MD5_ALT for example, the init function may need
to do more that just zeroizing the context
 2015-06-23 13:09:11 +02:00
Manuel Pégourié-Gonnard
1cd10adc7c Update prototype of x509write_set_key_usage() 
Allow for future support of decipherOnly and encipherOnly. Some work will be
required to ensure we still write only one byte when only one is needed.
 2015-06-23 13:09:10 +02:00
Manuel Pégourié-Gonnard
655a964539 Adapt check_key_usage to new weird bits 2015-06-23 13:09:10 +02:00
Manuel Pégourié-Gonnard
9a702255f4 Add parsing/printing for new X.509 keyUsage flags 2015-06-23 13:09:10 +02:00
Manuel Pégourié-Gonnard
b80d16d171 Fix return convention of x509_wildcard_verify() 2015-06-23 13:09:10 +02:00
Manuel Pégourié-Gonnard
07894338a0 Rename M255 to Curve25519 2015-06-23 13:09:10 +02:00
Manuel Pégourié-Gonnard
7320eb46d4 Remove references to some Montgomery curves 
After all it looks like those won't become standard.
 2015-06-23 13:09:10 +02:00
Manuel Pégourié-Gonnard
9386664543 Move from inttypes.h to stdint.h 
Some toolchains do not have inttypes.h, and we only need stdint.h which is a
subset of it.
 2015-06-22 23:41:26 +02:00
Manuel Pégourié-Gonnard
e7e89844d6 Fix and document corner-cases of time checking 2015-06-22 23:41:24 +02:00
Manuel Pégourié-Gonnard
57e10d71be Fix potential NULL dereference. 
Introduced when moving from gmtime_r() to gmtime().
Found with fbinfer.
 2015-06-22 23:40:44 +02:00
Manuel Pégourié-Gonnard
f9b85d96a9 Fix potential resource leak in X.509 parse dir 
Found with fbinfer.
 2015-06-22 18:39:57 +02:00
Manuel Pégourié-Gonnard
bcf13bab5d Fix issue with MemSan and entropy 
Due to the recent change about entropy sources strength, it is no longer
acceptable to just disable the platform source. So, instead "fix" it so that
it is clear to MemSan that memory is initialized.

I tried __attribute__((no_sanitize_memory)) and MemSan's blacklist file, but
couldn't seem to get them to work.
 2015-06-22 18:25:41 +02:00
Manuel Pégourié-Gonnard
cdc26ae099 Add mbedtls_ssl_set_hs_authmode 
While at it, fix the following:
- on server with RSA_PSK, we don't want to set flags (client auth happens via
  the PSK, no cert is expected).
- use safer tests (eg == OPTIONAL vs != REQUIRED)
 2015-06-22 14:52:40 +02:00
Manuel Pégourié-Gonnard
9dbaf400ef Rationalize other snprintf() uses 2015-06-22 14:42:04 +02:00
Manuel Pégourié-Gonnard
1685368408 Rationalize snprintf() usage in X.509 modules 2015-06-22 14:42:04 +02:00
Manuel Pégourié-Gonnard
6c0c8e0d3d Include fixed snprintf for Windows in platform.c 
Use _WIN32 to detect it rather that _MSC_VER as it turns out MSYS2 uses the
broken MS version by default too.
 2015-06-22 14:42:04 +02:00
Manuel Pégourié-Gonnard
f9cbd73191 Update generated files 2015-06-22 14:40:56 +02:00
Manuel Pégourié-Gonnard
7580ba475d Add a concept of entropy source strength. 
The main goal is, we want and error if cycle counter is the only source.
 2015-06-22 14:40:56 +02:00
Manuel Pégourié-Gonnard
3f77dfbd52 Add MBEDTLS_ENTROPY_HARDWARE_ALT 
Makes it easier for an external module to plug its hardware entropy collector.
 2015-06-22 14:40:56 +02:00
Manuel Pégourié-Gonnard
bf82ff0209 Fix entropy thresholds 2015-06-22 14:40:56 +02:00
Manuel Pégourié-Gonnard
60c793bdc9 Split HAVE_TIME into HAVE_TIME + HAVE_TIME_DATE 
First one means we have time() but it may not return the actual wall clock
time, second means it does.
 2015-06-22 14:40:56 +02:00
Manuel Pégourié-Gonnard
c0696c216b Rename mbedtls_mpi_msb to mbedtls_mpi_bitlen 2015-06-18 16:49:37 +02:00
Manuel Pégourié-Gonnard
097c7bb05b Rename relevant global symbols from size to bitlen 
Just applying rename.pl with this file:

mbedtls_cipher_get_key_size mbedtls_cipher_get_key_bitlen
mbedtls_pk_get_size mbedtls_pk_get_bitlen
MBEDTLS_BLOWFISH_MIN_KEY MBEDTLS_BLOWFISH_MIN_KEY_BITS
MBEDTLS_BLOWFISH_MAX_KEY MBEDTLS_BLOWFISH_MAX_KEY_BITS
 2015-06-18 16:43:38 +02:00
Manuel Pégourié-Gonnard
fb317c5221 Rename parameter in a x509 helper 2015-06-18 16:41:13 +02:00
Manuel Pégourié-Gonnard
39a48f4934 Internal renamings in PK 
+ an unrelated comment in SSL
 2015-06-18 16:06:55 +02:00
Manuel Pégourié-Gonnard
12ad798c87 Rename ssl_session.length to id_len 2015-06-18 15:50:37 +02:00
Manuel Pégourié-Gonnard
898e0aa210 Rename key_length in cipher_info 2015-06-18 15:31:10 +02:00
Manuel Pégourié-Gonnard
b8186a5e54 Rename len to bitlen in function parameters 
Clarify a few comments too.
 2015-06-18 14:58:58 +02:00
Manuel Pégourié-Gonnard
b31c5f68b1 Add SSL presets. 
No need to use a separate profile as in X.509, everything we need is already
in ssl_config. Just load appropriate values.
 2015-06-17 14:59:27 +02:00
Manuel Pégourié-Gonnard
7bfc122703 Implement sig_hashes 2015-06-17 14:34:48 +02:00
Manuel Pégourié-Gonnard
36a8b575a9 Create API for mbedtls_ssl_conf_sig_hashes(). 
Not implemented yet.
 2015-06-17 14:27:39 +02:00
Manuel Pégourié-Gonnard
9d412d872c Small internal changes in curve checking 
- switch from is_acceptable to the more usual check
- add NULL check just in case user screwed up config
 2015-06-17 14:27:39 +02:00
Manuel Pégourié-Gonnard
a83e4e2bf5 Extra check in verify_with_profile() 
This could happen if someone doesn't set the SSL configuration properly. In
that case we don't want to segfault...
 2015-06-17 14:27:38 +02:00
Manuel Pégourié-Gonnard
b541da6ef3 Fix define for ssl_conf_curves() 
This is a security feature, it shouldn't be optional.
 2015-06-17 14:27:38 +02:00
Manuel Pégourié-Gonnard
6e3ee3ad43 Add mbedtls_ssl_conf_cert_profile() 2015-06-17 14:27:38 +02:00
Manuel Pégourié-Gonnard
cbb1f6e5cb Implement cert profile checking 2015-06-17 14:27:38 +02:00
Manuel Pégourié-Gonnard
f8ea856296 Change data structure of profiles to bitfields 
- allows to express 'none' or 'all' more easily than lists
- more compact and easier to declare statically
- easier to check too

Only drawback: if we ever have more than 32 curves, we'll need an ABI change to
make that field a uint64_t.
 2015-06-17 14:27:38 +02:00
Manuel Pégourié-Gonnard
88db5da117 Add pre-defined profiles for cert verification 2015-06-17 14:27:38 +02:00
Manuel Pégourié-Gonnard
9505164ef4 Create cert profile API (unimplemented yet) 2015-06-17 14:27:38 +02:00
Manuel Pégourié-Gonnard
bd990d6629 Add ssl_conf_dhm_min_bitlen() 2015-06-17 11:37:04 +02:00