- 29 Apr, 2015 1 commit
-
-
Adam Langley authored
“ECDHE-PSK-WITH-AES-128-GCM-SHA256” doesn't follow the standard naming for OpenSSL: it was “-WITH-” in it and has a hyphen between “AES” and “128”. This change fixes that. Change-Id: Ie504624857f227fb18835a99cec7c3363beeed96
-
- 24 Apr, 2015 2 commits
-
-
Adam Langley authored
SSL_get0_chain_certs calls a ctrl function with SSL_CTRL_GET_CHAIN_CERTS. The switch failed to set a positive return value and so the call always appeared to fail. (Imported from upstream's https://boringssl-review.googlesource.com/#/c/4521/) Change-Id: Ia69c404c528b0cb01c7ff5e56ca8a8415265fa73
-
Adam Langley authored
Upstream settled in this API, and it's also the one that we expect internally and that third_party code will expect. This is an import of upstream's 5f0efe06e199a1bd96f161eb45f3dd76924cdc2a. Change-Id: Ib4c7054a382dccdd23919407742bd037b9653a4b
-
- 23 Apr, 2015 1 commit
-
-
Adam Langley authored
When |BN_dec2bn| and |BN_hex2bn| were merged (way back in the initial BoringSSL change), the neg flag was set too soon and could be cleared by |BN_add_word|. This is an import of upstream's c85573cc. The unittest change isn't included here because bn_test.c has changed significantly in upstream and BoringSSL unittests aren't run in the Android environment. Bug: 20523350 Change-Id: Iaf8efe2fe3419218437f5ebb9a15f73559860a0f
-
- 21 Apr, 2015 1 commit
-
-
Adam Langley authored
In https://android-review.googlesource.com/#/c/147551/, I missed the OPENSSL_EXPORT tag thus the .so doesn't expose it as a dynamic symbol. BUG=20419899 Change-Id: I849888cf9a3383570b352911867e983b547e6742
-
- 20 Apr, 2015 1 commit
-
-
Adam Langley authored
BoringSSL always uses uncompressed points. This function aborts if another form is requested or does nothing if uncompressed points are requested. Bug: 20419899 Change-Id: Ib5d0707c4e9eaee67e46a0d73d41be70ce0a9353
-
- 19 Apr, 2015 1 commit
-
-
Adam Langley authored
It appears that the version of Clang in Android doesn't support the .arch_extension directive. This change removes the .arch and .arch_extension lines (because they are triggering errors) and adds a -march option on the command line instead. The aarch64+Clang build is still broken with this change, but it's broken in binder rather than BoringSSL with it. Change-Id: I32c557bdfde4df66d26794ccdd650356f2bbaf8f
-
- 18 Apr, 2015 5 commits
-
-
Kenny Root authored
-
Adam Langley authored
OS X builds with NO_ASM and was getting both generic.c and x86_64-gcc.c. This change updates the latter so that it's excluded in NO_ASM builds. Change-Id: Idf801013db636bd4d8f6c4588102c241fdb9fbf6
-
Kenny Root authored
Clang doesn't currently understand the ".arch armv8-a+crypto" syntax and instead requires the specification of ".arch_extension" to enable the crypto functionality. Change-Id: Ib7bab562055e7c7925a47030044479ec172d0a34
-
Adam Langley authored
The immediate in this operation is too large for ARM. GCC will automatically rewrite it to use bic (where bic does an AND NOT). Clang, however doesn't, and reasonably throws an error. This change switches to using bic in the source file, thus making both happy. Change-Id: I117083f4f70c199e5d2f933c0a0516a6f4059a92
-
Adam Langley authored
Clang will warn that the argument to Speed is unused. Change-Id: Ic3a6a7e2f4638e8ad5435332077791db015d4779
-
- 08 Apr, 2015 1 commit
-
-
Adam Langley authored
Keystore has added support for it so these functions are needed again. Change-Id: Id3bf3dd10e182fe7a9b1c51bd3184ecac4cfde8b
-
- 02 Apr, 2015 1 commit
-
-
Adam Langley authored
Change-Id: I2fcb67fea859220e7e7bbbdb5dd910fb847c5600
-
- 25 Mar, 2015 1 commit
-
-
Adam Langley authored
Change-Id: I19ed78acc67bd0ad8b905ce0ac628b39da8bb161
-
- 16 Mar, 2015 3 commits
-
-
Kenny Root authored
-
Adam Langley authored
MIPS64 confusingly sets __mips__, but it's not a 32-bit platform. This change updates the defines in base.h to recognise MIPS64 based on both __mips__ and __LP64__ being defined. BUG: 19665578 Change-Id: I6290b6def9f999a01b500c918d1488a22fd57460
-
Adam Langley authored
The assert was supposed to be *added* in fcf25833 but instead replaced the check. Change-Id: I74d67f49a4375f6b9f1272cbfcf87a438b3b6f29
-
- 06 Mar, 2015 3 commits
-
-
Adam Langley authored
This is backport of BoringSSL's d216b71f909fe56255813dab0a8d052534bdcb91 and https://boringssl-review.googlesource.com/3810 and should allow asm on x86-64 to build correctly. Change-Id: Id321768930182951223dbf90c4c910e24d9b6798
-
Adam Langley authored
This is a backport of e023ad2d83bb8a804e989d38e91e0685db179830 from BoringSSL. Change-Id: I77b074221f44081229d48f67ee791d5165b2598d
-
Adam Langley authored
system/keymaster is using them now. Change-Id: I396e7001e6edf443ed2726d68d21704c7e557748
-
- 05 Mar, 2015 2 commits
-
-
Kenny Root authored
gmtime_s first appeared in MSVCR80, but libmingwex has a helper function that tries to find the symbol or falls back to an internal implementation. Change-Id: I5bc27e1cfcc208eb9ea1159d47791fcc90bc7794
-
Kenny Root authored
There are various problems with assembly on both the Mac and Windows builds, so just disable it for now. Change-Id: I6163e97089c17f70ef9361279319badd22c354b2
-
- 04 Mar, 2015 1 commit
-
-
Kenny Root authored
On Windows this doesn't matter since the filesystems are case- insensitive, but building BoringSSL on Linux with MinGW has case-sensitive filesystems. Change-Id: Iefd319cfda89d2d1f8d43cea39c68295bfa65c83
-
- 27 Feb, 2015 5 commits
-
-
Kenny Root authored
-
Adam Langley authored
(This is a backport of upstream BoringSSL's 4e581b5378d7ef435c9abe39ad5c2a334bd7b6e9.) Change-Id: If799791f28cd37ce810c0065939cf1942771c7f7
-
Kenny Root authored
Change-Id: If7c25984fadcb98ea9fb26983e04b1be3b6a2ecd
-
Kenny Root authored
-
Adam Langley authored
(This is a cherry-pick of upstream BoringSSL's 50073e8c5e81d8151718e05ec54b7c213372b878.) Change-Id: Id29ea889055dbecfbba4fc4a9e01af0c49b8073e
-
- 25 Feb, 2015 2 commits
-
-
Adam Langley authored
(I got this wrong when reading the OpenSSL code.) (Cherry-picked form upstream BoringSSL's 7c21925a10d451ed13ab201e0161dea40b974397) Change-Id: Icedaa45c9d17e7c2b95fa5be1f7c0bf41cac0880
-
Adam Langley authored
It was a mistake to remove this in the first place. (Cherry-picked from upstream BoringSSL's 93531bd70f48bc63ad7d4aedf32d69c8095170bd.) Change-Id: Iee35482bf11dd8813622e36b0c0eb8f91538007a
-
- 13 Feb, 2015 2 commits
-
-
Adam Langley authored
Without this, the linker says: (sha512-armv4.o): requires unsupported dynamic reloc R_ARM_REL32; recompile with -fPIC This is (I believe) a very misleading error message. The R_ARM_REL32 relocation type is the correct type for position independent code. But unless the target symbol is hidden then the linker doesn't know that it's not going to be overridden by a different ELF module. Change-Id: I9bb5f6b5f85c8de1ad5b6426cf27826976ce8248
-
Adam Langley authored
Without this the conscrypt build fails in `make checkbuild` because GCC emits calls to __memset_chk and friends, but the NDK doesn't provide them. Change-Id: I37d783f0e0e7d53a671fc03056e6450a4f5d1c95
-
- 09 Feb, 2015 2 commits
-
-
Dan Albert authored
The manifest change was reverted, so this isn't needed anymore. This reverts commit 54d4b126. Change-Id: I4d43b2362262d3a22d5b68d9a210ac6d1d150387
-
Dan Albert authored
All these modules are defined in external/openssl as well, which is breaking the build. Disable everything in the makefile until someone that owns the project has a chance to clean it up. Change-Id: I732739999e89bb832005842a5776d1540fb6bdbb
-
- 31 Jan, 2015 1 commit
-
-
Adam Langley authored
-