diff options
author | fanquake <fanquake@gmail.com> | 2021-12-18 11:34:47 +0800 |
---|---|---|
committer | fanquake <fanquake@gmail.com> | 2021-12-18 14:47:44 +0800 |
commit | c06cda3e48e9826043ebc5790a7bb505bfbf368c (patch) | |
tree | 77f0286bae3271a764e5db3f108e26395ed173ad /src/secp256k1/README.md | |
parent | 97b2fc03f666ccbadcd3a497303fb6577842d11f (diff) | |
parent | 314195c8be3bd7db0d5817c4fb3aa85c84363ce9 (diff) | |
download | bitcoin-c06cda3e48e9826043ebc5790a7bb505bfbf368c.tar.xz |
Merge bitcoin/bitcoin#23383: Update libsecp256k1 subtree to current master
314195c8be3bd7db0d5817c4fb3aa85c84363ce9 Remove unnecessary cast in CKey::SignSchnorr (Pieter Wuille)
a1f76cdb22e3278a48d63dd23c1fe3308daedd8c Remove --disable-openssl-tests for libsecp256k1 configure (Pieter Wuille)
86dbc4d075decb82fbba837aaa283cf0561897ad Squashed 'src/secp256k1/' changes from be8d9c262f..0559fc6e41 (Pieter Wuille)
Pull request description:
The motivation for this bump is getting rid of a cast in `CKey::SignSchnorr`; the `aux_rand` argument isn't modified by the `secp256k1_schnorrsig_sign` function, but was marked as non-`const` anyway. This is fixed now (bitcoin-core/secp256k1#966), and the cast is removed in this PR.
There are a few other relevant changes:
* (bitcoin-core/secp256k1#956): replaces a runtime-computed table with a precomputed one; this adds arouns 1 MiB to the binary size, but is a step towards significantly simplifying the API. If 1 MiB is too much, it can be reduced by 2 or 4 (or more) for a slight verification performance reduction.
* (bitcoin-core/secp256k1#983): removes (test/bench only) OpenSSL support entirely, removing the need to pass `--disable-openssl-tests` (see #23314).
* (bitcoin-core/secp256k1#810): mild performance increase for 64-bit non-x86 platforms.
* (bitcoin-core/secp256k1#1002): Make aux_rnd32==NULL behave identical to 0x0000..00 (which impacts BIP341/BIP342 signing in Bitcoin Core, making it more strictly BIP340 compliant, though not in a manner that affects security).
ACKs for top commit:
fanquake:
ACK 314195c8be3bd7db0d5817c4fb3aa85c84363ce9 - this includes a nice simplification to the lilbsecp build system (and thus our build system), and fixes issues like #22854. Did a Guix build on x86 (above), as well as a build on arm64 (except for the arm64 host):
Tree-SHA512: 0e048390fc148fbbdf5b98d9cce8c71067564e7d69d97b68347808a9bc45a04f4fc653c392c880d79d5d8b9cf282195520955581ac4f1595f6a948080cf5949d
Diffstat (limited to 'src/secp256k1/README.md')
-rw-r--r-- | src/secp256k1/README.md | 23 |
1 files changed, 13 insertions, 10 deletions
diff --git a/src/secp256k1/README.md b/src/secp256k1/README.md index 182c29d9ce..5fc07dd4fa 100644 --- a/src/secp256k1/README.md +++ b/src/secp256k1/README.md @@ -66,18 +66,9 @@ libsecp256k1 is built using autotools: $ ./autogen.sh $ ./configure $ make - $ make check + $ make check # run the test suite $ sudo make install # optional -Exhaustive tests ------------ - - $ ./exhaustive_tests - -With valgrind, you might need to increase the max stack size: - - $ valgrind --max-stackframe=2500000 ./exhaustive_tests - Test coverage ----------- @@ -100,6 +91,18 @@ To create a HTML report with coloured and annotated source code: $ mkdir -p coverage $ gcovr --exclude 'src/bench*' --html --html-details -o coverage/coverage.html +Benchmark +------------ +If configured with `--enable-benchmark` (which is the default), binaries for benchmarking the libsecp256k1 functions will be present in the root directory after the build. + +To print the benchmark result to the command line: + + $ ./bench_name + +To create a CSV file for the benchmark result : + + $ ./bench_name | sed '2d;s/ \{1,\}//g' > bench_name.csv + Reporting a vulnerability ------------ |