openbsd/src/lib/libcrypto/bn/arch/aarch64, branch OPENBSD_7_4A mirror of https://github.com/libressl/openbsd.git
https://git.lua4.win/openbsd/atom?h=OPENBSD_7_42023-07-24T10:21:29+00:00Mark read/write registers as non-overlapping for bn_qwmulw_{addqw_,}addw()2023-07-24T10:21:29+00:00jsing2023-07-24T10:21:29+00:00urn:sha1:4c10fa62f800291a02f5e14311eec1a14d18057b
This does not cause an issue currently, however if called differently to
their current usage, it can lead to an input being overwritten and
incorrect results being generated.
Provide optimised bn_clzw() for aarch64.2023-06-21T07:56:43+00:00jsing2023-06-21T07:56:43+00:00urn:sha1:ad88bf2f67997bb60ec7841577bd3253b872853eOptimise bn_mul2_mulw_addtw() for aarch64.2023-06-17T15:40:46+00:00jsing2023-06-17T15:40:46+00:00urn:sha1:c4cf90a6f9bdc4c2b5bf83876448ace95db0376a
This provides significant performance gains for bn_sqr_comba4() and
bn_sqr_comba8().
Optimise quad word primitives on aarch64.2023-06-12T16:42:11+00:00jsing2023-06-12T16:42:11+00:00urn:sha1:2270ac93c092d155258184da66cd035546f23c1a
This provides a performance gain across most BN operations.
Provide optimised bn_mulw_{addw,addw_addw,addtw}() for aarch64.2023-05-28T17:42:30+00:00jsing2023-05-28T17:42:30+00:00urn:sha1:847bb6ade26d4856fd05909787fd82194b263387
This results in bn_mul_comba4() and bn_mul_comba8() requiring ~30% less
instructions than they did previously.
Provide optimised bn_addw_addw()/bn_subw_subw() for aarch64.2023-05-28T17:22:04+00:00jsing2023-05-28T17:22:04+00:00urn:sha1:3b16059743e4e2f0b44353e1ae9dd45339145f76Tweak indent and use named registers.2023-04-17T12:51:09+00:00jsing2023-04-17T12:51:09+00:00urn:sha1:d8dee4a3ad7e9f5d1075d150faf3862f903b10f9
No functional change.
Use separate lines instead of semicolons.2023-02-25T15:39:40+00:00bcook2023-02-25T15:39:40+00:00urn:sha1:090d81850efecfea5c1d751b14720459c4ca3fc9
macOS aarch64 assembly dialect treats ; as comment instead of a newline
ok tb@, jsing@
Provide optimised versions of bn_addw() and bn_subw() for aarch64.2023-02-17T05:46:57+00:00jsing2023-02-17T05:46:57+00:00urn:sha1:e8f247b8ef41ae77dd7ee9d5bc450d4e9c4af4f5Rename bn_umul_hilo() to bn_mulw().2023-02-16T10:41:03+00:00jsing2023-02-16T10:41:03+00:00urn:sha1:b11f50c06cb0a98a91ab4cbfb49715c95ee07486
This keeps the naming consistent with the other bignum primitives that have
been recently introduced. Also, use 1/0 intead of h/l (e.g. a1 instead of
ah), as this keeps consistency with other primitives and allows for naming
that works with double word, triple word and quadruple word inputs/outputs.
Discussed with tb@