openbsd/src/lib/libcrypto/cryptlib.c, branch tb_20230715A mirror of https://github.com/libressl/openbsd.git
https://git.lua4.win/openbsd/atom?h=tb_202307152023-07-15T19:29:46+00:00This commit was manufactured by cvs2git to create tag 'tb_20230715'.2023-07-15T19:29:46+00:00cvs2svnadmin@example.com2023-07-15T19:29:46+00:00urn:sha1:72cc860132e5b7971e495ba621dcd0713b5dd801Hide symbols in crypto.h2023-07-08T08:28:23+00:00beck2023-07-08T08:28:23+00:00urn:sha1:842e06ad9bdf75858b4eca096b3f27979b2101fd
ok jsing@
Recent discussions about abort() potentially leaving key material in2019-01-26T11:30:32+00:00deraadt2019-01-26T11:30:32+00:00urn:sha1:22c5663efc9813a7ec74f1d59c835b34aceb6062
core files (which can depend upon various file layouts) have resonated
with my hate for this function outside a purely debugging context. I
also dislike how the report goes to stderr which may get lost or ignored.
Increase the noise (with syslog_r) and use _exit(1) to gaurantee termination.
ok jsing
Store and return the locking callbacks, restoring previous behaviour.2018-11-24T04:11:47+00:00jsing2018-11-24T04:11:47+00:00urn:sha1:8e6972c56a2f404ee1ab7aef845a882f2db938d8
The previous code meant that a caller could set the locking callback, after
which CRYPTO_get_locking_callback() would return non-NULL. Some existing
code depends on this behaviour, specifically to identify if lock handling
has been configured. As such, always returning NULL from
CRYPTO_get_locking_callback() can result in unexpected application
behaviour.
ok bcook@
Add back a few missing compatibility stubs2018-11-11T16:32:28+00:00bcook2018-11-11T16:32:28+00:00urn:sha1:8a99f7096d5d13dfe9e814a08cb0a206eb8d0bbc
ok beck@
Add automatic threading initialization for libcrypto.2018-11-11T06:41:28+00:00bcook2018-11-11T06:41:28+00:00urn:sha1:3c41ae7ec959c3c608931951c8655a0e1df18696
This implements automatic thread support initialization in libcrypto.
This does not remove any functions from the ABI, but does turn them into
no-ops. Stub implementations of pthread_mutex_(init|lock|unlock) are
provided for ramdisks.
This does not implement the new OpenSSL 1.1 thread API internally,
keeping the original CRYTPO_lock / CRYPTO_add_lock functions for library
locking. For -portable, crypto_lock.c can be reimplemented with
OS-specific primitives as needed.
ok beck@, tb@, looks sane guenther@
Stop calling OPENSSL_init() internally, since it is a no-op. Also place2017-04-29T21:48:44+00:00jsing2017-04-29T21:48:44+00:00urn:sha1:1de47e5fe8ef45dc6638449b6a0decdcf07cc840
it under #ifndef LIBRESSL_INTERNAL.
ok beck@
Send the function codes from the error functions to the bit bucket,2017-01-29T17:49:23+00:00beck2017-01-29T17:49:23+00:00urn:sha1:d1f47bd292f36094480caa49ada36b99a69c59b0
as was done earlier in libssl. Thanks inoguchi@ for noticing
libssl had more reacharounds into this.
ok jsing@ inoguchi@
Replace all uses of magic numbers when operating on OPENSSL_ia32_P[] by2016-11-04T17:30:30+00:00miod2016-11-04T17:30:30+00:00urn:sha1:1a12fc8399638223feca8f853e2ac2cc22eeb471
meaningful constants in a private header file, so that reviewers can actually
get a chance to figure out what the code is attempting to do without knowing
all cpuid bits.
While there, turn it from an array of two 32-bit ints into a properly aligned
64-bit int.
Use of OPENSSL_ia32_P is now restricted to the assembler parts. C code will
now always use OPENSSL_cpu_caps() and check for the proper bits in the
whole 64-bit word it returns.
i386 tests and ok jsing@
Remove I386_ONLY define. It was only used to prefer a2016-11-04T13:56:05+00:00miod2016-11-04T13:56:05+00:00urn:sha1:098856ba6c9340c7f8061be62dd042763fe233a3
faster-on-genuine-80386-but-slower-on-80486-onwards innstruction sequence in
the SHA512 code, and had not been enabled in years, if at all.
ok tom@ bcook@