af6abcb3d0097a7f7892fb8b54a4c6363e5c2c7f Make bench support selecting which benchmarks to run (Pieter Wuille)
9f56bdf5b9ba2e22e77c6adaaeb8302398732df3 Merge bench_schnorrsig into bench (Pieter Wuille)
3208557ae1062f7fcce25d5f2c5a29dc34a18895 Merge bench_recover into bench (Pieter Wuille)
855e18d8a809b98a622ab55765792aca132ea640 Merge bench_ecdh into bench (Pieter Wuille)
2a7be678a660d58742b1e767444c0fe75fa22592 Combine bench_sign and bench_verify into single bench (Pieter Wuille)
Pull request description:
This combines `bench_verify`, `bench_sign`, `bench_ecdh`, `bench_recovery`, and `bench_schnorrsig` into a single `bench` binary.
I don't think there is a good reason to have this many binaries, and it complicates build config and CI.
ACKs for top commit:
real-or-random:
ACK af6abcb3d0097a7f7892fb8b54a4c6363e5c2c7f diff looks good, command line options work, valgrind is happy
siv2r:
tACK af6abcb, the command-line options work as expected on my ubuntu machine. The diff looks good.
Tree-SHA512: 39c099b508c70136eaac8a429591b4250a8c22e423fa922d83928ea9273d8f2c1136317427563b28f249c02cf83d1c73ea787c6d26aa88545236241641965705
5324f8942dd322448fae6c9b225ecac2854fa7e2 Make aux_rnd32==NULL behave identical to 0x0000..00. (Pieter Wuille)
Pull request description:
BIP340's default signing algorithm always requires an aux_rnd argument, but permits using an all-zero one when no randomness is available.
Make secp256k1_schnorrsig_sign follow this even when aux_rnd32==NULL, by treating the same as if an all-zero byte array was provided as input.
ACKs for top commit:
junderw:
ACK 5324f89
elichai:
ACK 5324f8942dd322448fae6c9b225ecac2854fa7e2
jonasnick:
ACK 5324f8942dd322448fae6c9b225ecac2854fa7e2
real-or-random:
utACK 5324f8942dd322448fae6c9b225ecac2854fa7e2
Tree-SHA512: caa1d5a0eacea3239d8aaace5284eedcd850058bbe759768e626233a010199db6c637618aedccfb51fe94ec8d28f45bc0c441be77e2e12fa2a393b9cc3a5d3ae
BIP340's default signing algorithm always requires an aux_rnd argument,
but permits using an all-zero one when no randomness is available.
Make secp256k1_schnorrsig_sign follow this even when aux_rnd32==NULL,
by treating the same as if an all-zero byte array was provided as
input.
2888640132eb64ed30a8a208931f27447c3e0366 VERIFY_CHECK precondition for secp256k1_fe_set_int. (Russell O'Connor)
d49011f54c2b31807158bdf06364f331558cccc7 Make _set_fe_int( . , 0 ) set magnitude to 0 (Tim Ruffing)
Pull request description:
Also set the magnitude to 0 when setting the value to 0.
ACKs for top commit:
real-or-random:
ACK 2888640132eb64ed30a8a208931f27447c3e0366
jonasnick:
ACK 2888640132eb64ed30a8a208931f27447c3e0366
Tree-SHA512: 6ec9b3485380503b11c00f30bfa79f92ba3facb93ee4f3df582b881c4e19fb8ae8b5acd5aeb6326497c290cd0904230d0356f33bd136ca577d2f25616279e090
23e2f66726f930ac01d5075106aa16a4073442b4 bench: don't return 1 in have_flag() if argc = 1 (Jonas Nick)
96b1ad2ea9f9d9419e566b95162487c48902c3eb bench_ecmult: improve clarity of output (Jonas Nick)
Pull request description:
Previously "ecmult{,_multi} xg" meant multiplication with (x - 1) random points
and base point G. Now
- xP means multiplication with x random points and
- xP & G means multiplication with x random points and G
ACKs for top commit:
siv2r:
tACK 23e2f66
real-or-random:
ACK 23e2f66726f930ac01d5075106aa16a4073442b4
Tree-SHA512: 0218aaa0baa4c2f92a7b98c97b8cc3b596e3da44d7f38ab4bdd707a4bdb96bb071b953fc6106cd34977a562278e4eaa860a3a7fa64c323c5117945e7a3107162
This makes the semantic of have_flag more clear and fixes a bug
that was introduced in
2fe1b50df16c9f41ea77b151634d734b930eeddd
Add ecmult_gen, ecmult_const and ecmult to benchmark
where the behavior introduced by this commit was already assumed. If
bench_ecmult was called without arguments, have_flag("simple") returned 1 and no
scratch space was allocated which led to very wrong output.
Previously "ecmult{,_multi} xg" meant multiplication with (x - 1) random points
and base point G. Now
- ecmult_{,multi_}xp means multiplication with x random points and
- ecmult_{,multi_}xp_g means multiplication with x random points and G
b4b130678db31a7cabc2cde091bc4acbca92b7a3 create csv file from the benchmark output (siv2r)
26a255beb673217c839dcc51790d9a484f9a292d Shared benchmark format for command line and CSV outputs (siv2r)
Pull request description:
ACKs for top commit:
real-or-random:
ACK b4b130678db31a7cabc2cde091bc4acbca92b7a3
jonasnick:
ACK b4b130678db31a7cabc2cde091bc4acbca92b7a3
Tree-SHA512: 1eebbdd7701ad21d9647434ff05f23827be217d47870bb05a2fdb12447abc365fc6e56306f344e05d8d2ec1ff5532562131b3876261733e4412117357c5c65f8
044d95630556dda5492a70af056bc277f0b79ebc Fix G.y parity in sage code (Pieter Wuille)
Pull request description:
I'm not sure if `EllipticCurve.lift_x` has well-defined Y coordinate or not, but at least my current version of Sage computes the wrong G. Fix this.
ACKs for top commit:
real-or-random:
ACK 044d95630556dda5492a70af056bc277f0b79ebc
Tree-SHA512: afb919af29027da2bb3c58628924f9740672d3c347ad39cc663c9c399b1aa8536256fd3fd4e1e54457e38344704d47f281d82488da413f4e6e67e191decc960f
1. add `print_output_table_header_row` func to print the table header for benchmark output
2. modify the following benchmarks to include the table header
- bench_ecdh.c
- bench_ecmult.c
- bench_internal.c
- bench_recover.c
- bench_schnorrsig.c
- bench_sign.c
- bench_verify.c
b53e0cd61fce0bcef178f317537c91efc9afd04d Avoid overly-wide multiplications (Peter Dettman)
Pull request description:
Speeds up bench_ecdh, bench_sign, bench_verify relative to master by 5+% at -O3, haswell.
ACKs for top commit:
sipa:
ACK b53e0cd61fce0bcef178f317537c91efc9afd04d
real-or-random:
ACK b53e0cd61fce0bcef178f317537c91efc9afd04d I've inspected the diff and run the tests without asm for a CPU day
Tree-SHA512: 4f79c98371a3dc9da013632210c8db979f910b222291999dfaa0c31849a77eb427361e4ab9206cbfee73c30a8933178784d6cb8e747e8dca6b227eb77fbea2a2
9be7b0f08340a063d961547b5d2663405f3fc162 Avoid computing out-of-bounds pointer. (Tim Ruffing)
Pull request description:
This is a pedantic case of UB.
Spotted in #879.
ACKs for top commit:
elichai:
ACK 9be7b0f08340a063d961547b5d2663405f3fc162
practicalswift:
cr ACK 9be7b0f08340a063d961547b5d2663405f3fc162
sipa:
ACK 9be7b0f08340a063d961547b5d2663405f3fc162
Tree-SHA512: a9d028c4cdb37ad0d5fcf0d2f678eef732a653d37155a69a20272c6b283c28e083172485d7a37dc4a7c6100b22a6f5b6a92e729239031be228cc511842ee35e8
bc08599e776aff33c834ef829843ec5f629d1f39 Remove OpenSSL testing support (Pieter Wuille)
Pull request description:
This removes the ability to test against OpenSSL, as well as the OpenSSL verification benchmark.
The motivation is that OpenSSL 3 is deprecating part of the API used here (see #869), and I'm not sure it's worth maintaining. We do lose the fact that this is the only test that verifies randomly-generated cases against an independent implementation. On the other hand, there are tons of existing fixed tests now that test all kinds of edge cases already.
ACKs for top commit:
elichai:
tACK bc08599
real-or-random:
ACK bc08599e776aff33c834ef829843ec5f629d1f39
jonasnick:
ACK bc08599e776aff33c834ef829843ec5f629d1f39
Tree-SHA512: 632e6d3cf7bbc5828f5ca1f0f2a92c80bcb681bbcd4320c352b4a86fd521e410c852ccebcfc30fadc8fbf86649267a9e521f53e0f78072a8cd74d8726da28973
189f6bcfef6578b89e21f937b24060f74bd18f00 Fix unused parameter warnings when building without VERIFY (Jonas Nick)
Pull request description:
This commit makes `./configure --enable-coverage && make check` free of warnings.
ACKs for top commit:
practicalswift:
cr ACK 189f6bcfef6578b89e21f937b24060f74bd18f00
elichai:
utACK 189f6bcfef6578b89e21f937b24060f74bd18f00
siv2r:
Tested ACK 189f6bc
Tree-SHA512: 727fe0e40ff61f404780b32dfa4102a58bed9d922e61bd17ddaaf1243b0c06edd9697ff4763b5e92d033e7db3778193bee07d85cfa3b9c46d45e5fec3f568009
d43993724deb5fdc1d2162f7423f8e8398103dd5 tests: remove `secp256k1_fe_verify` from tests.c and modify `secp256k1_fe_from_storage` to call `secp256k1_fe_verify` (siv2r)
Pull request description:
ACKs for top commit:
roconnor-blockstream:
utACK d439937 diff looks correct, I also didn't run the tests locally.
real-or-random:
utACK d43993724deb5fdc1d2162f7423f8e8398103dd5 diff looks correct, I didn't run the tests locally
jonasnick:
ACK d43993724deb5fdc1d2162f7423f8e8398103dd5 ran tests with `--enable-coverage`
Tree-SHA512: c3c9ecf8e9b7dfdcd1144ddcf8bcc637996c699dbd0fc6223e6186d082908728468fa276b09c6f344e036ca05f54432dde6366a83eb39f915a334164faadd556
1. secp256k1_fe_verify is removed from tests since, it throws an error if VERIFY is not defined during compilation.
(Ex: ./configure --enable-coverage)
2. `secp256k1_fe_from_storage` calls `secp256k1_fe_verify` in the VERIFY build to check for invalid field element.
72713872a8597884918bcf1edbc12f5c969ca680 Add missing static to secp256k1_schnorrsig_sign_internal (Elichai Turkel)
Pull request description:
This function isn't used outside of this module so it should be declared static
ACKs for top commit:
real-or-random:
ACK 72713872a8597884918bcf1edbc12f5c969ca680
jonasnick:
ACK 72713872a8597884918bcf1edbc12f5c969ca680
Tree-SHA512: 6107a2c84c3e11ffd68de22a5288d989a3c71c2ec1ee4827c88f6165fc27ef8339d0f6740928540e8ccd03aff49a2a96149bf698ccebe6d6d8ad6e23e38e8838
adec5a16383f1704d80d7c767b2a65d9221cee08 Add missing null check for ctx and input keys in the public API (Elichai Turkel)
f4edfc758142d6e100ca5d086126bf532b8a7020 Improve consistency for NULL arguments in the public interface (Elichai Turkel)
Pull request description:
I went over the public API and added missing explanations on when a pointer can be null and when it cannot,
and added some missing checks for null ctx and null pubkey pointers.
Open questions IMHO:
1. Can `secp256k1_context_create` return NULL? right now it could return null if you replaced the callbacks at compile time to ones that do return(unlike the default ones which never return).
2. Related to the first, should we document that the callbacks should never return? (in the tests we use returning callbacks but we can violate our own API) right now we say the following:
> After this callback returns, anything may happen, including crashing.
Is this enough to document answer `no` for the first question and just saying that if the callback returned then you violated the API so `secp256k1_context_create` can return NULL even though it is promised not to?
Right now we AFAICT we never check if it returns null
Another nit I'm not sure about is wording `(does nothing if NULL)`/`(ignored if NULL)`/`(can be NULL)`
More missing docs:
1. Documenting the `data` argument to the default nonce functions
ACKs for top commit:
ariard:
ACK adec5a16
jonasnick:
ACK adec5a16383f1704d80d7c767b2a65d9221cee08
Tree-SHA512: 6fe785776b7e451e9e8cae944987f927b1eb2e2d404dfcb1b0ceb0a30bda4ce16469708920269417e5ada09739723a430e270dea1868fe7d12ccd5699dde5976
This header contains a static array that replaces the ecmult_context pre_g and pre_g_128 tables.
The gen_ecmult_static_pre_g program generates this header file.
clang 7 to 11 (and maybe earlier versions) warn about recid being
potentially unitiliazed in "CHECK(recid >= 0 [...]", which was mitigated
in commit 3d2cf6c5bd35b0d72716b47bdd7e3892388aafc4 by initializing recid
to make clang happy but VG_UNDEF'ing the variable after initializiation
in order to ensure valgrind's memcheck analysis will still be sound and
complain if recid is not actually written to when creating a signature.
However, it turns out that at least for binaries produced by clang 11
(but not clang 7), valgrind complains about a branch on unitialized data
in the recid variable in that line before *and* after the aforementioned
commit. While the complaint after the commit could be spurious (clang
knows that recid is initialized, so it's fine to access it even though
the access is stupid), the complaint before the commit indicates a real
problem: it might be the case that clang is performing a wrong
optimization that leads to a situation where recid is really not
guaranteed to be initialized when it's accessed. As a result, clang
warns about this and generates code that just accesses the variable.
I'm not going to bother with this further because this is fixed in
clang 12 and the problem is just in our test code, not in the tested
code.
This commit rewrites the code in a way that groups the signing together
with the CHECK such that it's very easy to figure out for clang that
recid will be initialized properly. This seems to circument the issue.
aeece4459977b69962bcd1e1ee8845c18c74ff8f gen_context: Don't use any ASM (Tim Ruffing)
Pull request description:
See https://github.com/bitcoin/bitcoin/issues/22441 , we need to wait for the testing results there.
ACKs for top commit:
sipa:
utACK aeece4459977b69962bcd1e1ee8845c18c74ff8f
jonasnick:
ACK aeece4459977b69962bcd1e1ee8845c18c74ff8f
Tree-SHA512: 52ff90f3dedda90124140de1c2c1c065a2f9374930d6b988d35c37f5eeae97f7d557b7ab0cf99d22add5a76ff8a3e06226572e43949e12d1048cb323d1b3d92b
unsigned char foo[4] = "abcd" is not valid C++ because the string
literal "abcd" does not fit into foo due to the terminating NUL
character. This is valid in C, it will just omit the NUL character.
Fixes#962.
