aboutsummaryrefslogtreecommitdiffstats
path: root/src/lib
Commit message (Collapse)AuthorAgeFilesLines
* Merge GH #1215 Deprecate package transformJack Lloyd2017-09-191-0/+2
|\
| * Deprecate package transformJack Lloyd2017-09-191-0/+2
| | | | | | | | | | Unclear what's up with #825 but in any case this function is funky and I don't care enough to try to debug/fix it...
* | Add basic tests for const time utilsJack Lloyd2017-09-191-28/+10
|/ | | | Remove CT::min and CT::max which were unused and it turns out, broken.
* Merge GH #872 Add ability for TLS servers to prohibit renegotiationJack Lloyd2017-09-193-2/+16
|\
| * Fix logic of renegotiation checkJack Lloyd2017-02-203-4/+6
| | | | | | | | | | | | Turning the policy off broke the server entirely. Expose the new flag to Text_Policy
| * Add TLS::Policy::allow_client_initiated_renegotiationJack Lloyd2017-02-193-2/+14
| | | | | | | | Parallel of the server policy flag.
* | Merge GH #1205 Support large blocks in OCB modeJack Lloyd2017-09-192-62/+134
|\ \
| * | Change wide block OCBJack Lloyd2017-09-151-29/+0
| | | | | | | | | | | | | | | Ted Krovetz confirmed there were bugs in the reference code for blocks > 128 bits so these values should be the correct ones.
| * | Add alternate form for matching OCB ref codeJack Lloyd2017-09-111-4/+30
| | |
| * | Support larger block sizes in OCBJack Lloyd2017-09-102-59/+134
| | | | | | | | | | | | | | | | | | This doesn't match the draft-3 test vectors and may be bogus. [ci skip]
* | | Merge GH #1213 Improve ARM processor detection for systems without getauxvalJack Lloyd2017-09-188-502/+670
|\ \ \
| * | | Add ARM feature detection for systems without getauxvalJack Lloyd2017-09-172-15/+129
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | For iOS use sysctl to get the product name and use a static table of minimum versions. For everything else (not Linux/Android or iOS) try probe functions. Only for Aarch64 to keep things simple.
| * | | Split up cpuid.cpp into arch-specific filesJack Lloyd2017-09-178-502/+556
| | | | | | | | | | | | | | | | | | | | As more archs and detection methods are used, the file was getting a little hard to follow.
* | | | Avoid int vs size_t warning in OpenSSL codeJack Lloyd2017-09-171-1/+1
| | | |
* | | | Avoid old style casts in Ed25519Jack Lloyd2017-09-172-270/+262
| | | |
* | | | Fix Doxygen comment problemsJack Lloyd2017-09-171-4/+4
| | | |
* | | | Move socket classes to anon namespace.Jack Lloyd2017-09-171-172/+181
| | | | | | | | | | | | | | | | A little cleaner than in the function.
* | | | Fix Wshadow warningJack Lloyd2017-09-171-5/+5
|/ / /
* | | Simplify code in get_processor_timestampJack Lloyd2017-09-161-21/+12
| | |
* | | Fix an off by 8Jack Lloyd2017-09-161-1/+1
| | |
* | | Revert SIMDing of xor_buf in f39aa7bbJack Lloyd2017-09-161-29/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | Forgot this causes all kinds of issues when the SIMD ISA is not part of the base instruction set (needing -maltivec, etc). It just happens to work on x86-64 because SSE2 is part of the base ISA there. Instead just unroll by 16x this seems to be enough that GCC figures it out.
* | | Use constant_time_compare instead of same_memJack Lloyd2017-09-1622-23/+36
| | | | | | | | | | | | New name, same great operation
* | | De-inline xor_buf, add SIMD and unrollingJack Lloyd2017-09-163-34/+133
| | | | | | | | | | | | Improves CBC and OCB performance with AES-NI quite noticably
* | | Small cleanup in CBC modeJack Lloyd2017-09-162-21/+23
| | |
* | | Small cleanup in bcryptJack Lloyd2017-09-161-3/+3
| | |
* | | Add missing override annotationsJack Lloyd2017-09-161-6/+6
| | |
* | | Small simplification in CAST-128Jack Lloyd2017-09-161-41/+41
| | |
* | | Merge duplicated logic for clearing CPUID bitsJack Lloyd2017-09-162-0/+42
| | |
* | | Minor simplifications in SIMD_32 constructorJack Lloyd2017-09-161-2/+5
| | |
* | | reenable 8 byte gcm tagsAtanas Filyanov2017-09-151-1/+2
| | |
* | | Allow GCM tag size in the range 96..128 bitsAtanas Filyanov2017-09-141-1/+2
| | |
* | | Merge GH #1208 Fix decoding of RFC 5915 ECC keysJack Lloyd2017-09-131-3/+0
|\ \ \
| * | | Remove inner/outer OID check in EC_PrivateKey ctorRené Korthaus2017-09-131-3/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | RFC 5915 mandates that the OID of an ECDSA ECPrivateKey shall be id-ecPublicKey with the named curve as the parameters. ECPrivateKey may contain the named curve OID, too, which is compared to id-ecPublicKey. Such keys could not be loaded. We remove this check and add a test vector from strongswan.
* | | | allow 12 byte GCM tags and add test vectorsAtanas Filyanov2017-09-131-1/+1
|/ / /
* | | Merge GH #1202 Fix CFB IV carryover closes #1200Jack Lloyd2017-09-112-47/+44
|\ \ \ | |/ / |/| |
| * | Address CFB carryover bugJack Lloyd2017-09-102-47/+44
| | | | | | | | | | | | | | | | | | Test data generated by 1.10 so hopefully no further issues here. GH #1200
* | | Change SM2 encryption to match updated standard.Jack Lloyd2017-09-101-5/+5
| | | | | | | | | | | | | | | | | | | | | | | | Unfortunately it seems the SM2 format changed between 2010 and 2012, now the ciphertext is C1 || C3 || C2. Unfortunate no matter how you slice it, but at least it's easy to convert from one form to another.
* | | Merge GH #1199 Run Windows CI build via Python scriptJack Lloyd2017-09-102-4/+1
|\ \ \
| * | | Script the Windows CI buildJack Lloyd2017-09-082-4/+1
| |/ /
* | | Merge GH #1203 Add Win32 filesystem API access for MinGWJack Lloyd2017-09-101-1/+55
|\ \ \
| * | | Make it actually workJack Lloyd2017-09-101-2/+6
| | | | | | | | | | | | | | | | | | | | Weirdly Wine at least does not set FILE_ATTRIBUTE_NORMAL on the files, instead sets FILE_ATTRIBUTE_ARCHIVE. ?? I have no idea what that's about.
| * | | Well, it compilesJack Lloyd2017-09-101-1/+51
| |/ /
* / / Fix bad comparison in Winsock codeJack Lloyd2017-09-101-1/+1
|/ / | | | | | | [ci skip]
* | Fix loading of plaintext PKCS#8 private keysRené Korthaus2017-09-081-5/+10
| | | | | | | | | | We fixed this in the C++ API in GH #381, but apparently not in ffi. Also adds the missing tests.
* | Avoid throwing in deallocateJack Lloyd2017-09-071-4/+4
| | | | | | | | | | | | | | Could end up causing a throw during a destructor leading to a crash, if the application created a very large secure_vector. Flagged by Coverity.
* | Fix invalid comparison, potential segfault on invalid typeJack Lloyd2017-09-071-1/+2
| | | | | | | | Found by Coverity
* | Bump version for SM2Jack Lloyd2017-09-071-1/+1
| | | | | | | | | | | | To reflect support for arbitrary hashes [ci skip]
* | Merge GH #1189 Correct return value of botan_pk_op_verify_finishJack Lloyd2017-09-073-2/+4
|\ \
| * | Correct return value of botan_pk_op_verify_finishJack Lloyd2017-09-063-2/+4
| | | | | | | | | | | | | | | | | | | | | This function changed behavior in 0d403a3 see also GH #1187 Add new return code BOTAN_FFI_INVALID_VERIFIER and use it for both signature and bcrypt verification functions.
* | | Add support for computing SM2 ZA field to FFIJack Lloyd2017-09-062-0/+42
| | | | | | | | | | | | This is a contribution from Ribose Inc.
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-17 05:42:08 +0000