libsecp256k1's API changed, so update key.cpp to use it.
Libsecp256k1 now has explicit context objects, which makes it completely thread-safe.
In turn, keep an explicit context object in key.cpp, which is explicitly initialized
destroyed. This is not really pretty now, but it's more efficient than the static
initialized object in key.cpp (which made for example bitcoin-tx slow, as for most of
its calls, libsecp256k1 wasn't actually needed).
This also brings in the new blinding support in libsecp256k1. By passing in a random
seed, temporary variables during the elliptic curve computations are altered, in such
a way that if an attacker does not know the blind, observing the internal operations
leaks less information about the keys used. This was implemented by Greg Maxwell.
This was added a while ago for testing purposes, but was never intended to be
used. Remove it until upstream libsecp256k1 decides that verification is
stable/ready.
Add a sanity check to prevent cosmic rays from flipping a bit in the
generated public key, or bugs in the elliptic curve code. This is
simply done by signing a (randomized) message, and verifying the
result.
- add a small wrapper in util around RAND_bytes() and replace with
GetRandBytes() in the code to log errors from calling RAND_bytes()
- remove OpenSSL header rand.h where no longer needed
Previously if bitcoind is linked with an OpenSSL which is compiled
without EC support, this is seen as an assertion failure "pKey !=
NULL" at key.cpp:134, which occurs after several seconds. It is an
esoteric piece of knowledge to interpret this as "oops, I linked
with the wrong OpenSSL", and because of the delay it may not even
be noticed.
The new output is
: OpenSSL appears to lack support for elliptic curve cryptography. For
more information, visit
https://en.bitcoin.it/wiki/OpenSSL_and_EC_Libraries
: Initialization sanity check failed. Bitcoin Core is shutting down.
which occurs immediately after attempted startup.
This also blocks in an InitSanityCheck() function which currently only
checks for EC support but should eventually do more. See #4081.
a81cd968 introduced a malleability breaker for signatures
(using an even value for S). In e0e14e43 this was changed to
the lower of two potential values, rather than the even one.
Only the signing code was changed though, the (for now unused)
verification code wasn't adapted.
There were quite a few places where assert() was used with side effects,
making operation with NDEBUG non-functional. This commit fixes all the
cases I know about, but also adds an #error on NDEBUG because the code
is untested without assertions and may still have vulnerabilities if
used without assert.
Use misc methods of avoiding unnecesary header includes.
Replace int typedefs with int##_t from stdint.h.
Replace PRI64[xdu] with PRI[xdu]64 from inttypes.h.
Normalize QT_VERSION ifs where possible.
Resolve some indirect dependencies as direct ones.
Remove extern declarations from .cpp files.
To fix a minor malleability found by Sergio Lerner (reported here:
https://bitcointalk.org/index.php?topic=8392.msg1245898#msg1245898)
The problem is that if (R,S) is a valid ECDSA signature for a given
message and public key, (R,-S) is also valid. Modulo N (the order
of the secp256k1 curve), this means that both (R,S) and (R,N-S) are
valid. Given that N is odd, S and N-S have a different lowest bit.
We solve the problem by forcing signatures to have an even S value,
excluding one of the alternatives.
This commit just changes the signing code to always produce even S
values, and adds a verification mode to check it. This code is not
enabled anywhere yet. Existing tests in key_tests.cpp verify that
the produced signatures are still valid.
Pavel Janík