32 Commits

Author SHA1 Message Date
Gregory Maxwell
f735446c4d Convert the rest of the codebase to C89.
Update build system to enforce -std=c89 -pedantic.
2015-01-25 17:44:10 +00:00
Pieter Wuille
fcc48c4561 Remove the non-storage cmov 2015-01-25 00:56:22 -04:00
Pieter Wuille
ff889f7d90 Field storage type 2015-01-25 00:11:20 -04:00
Gregory Maxwell
25b35c7ecb Convert field code to strict C89 (+ long long, +__int128)
This makes the software more portable to embedded systems
 and static analysis tools.

Sadly, it can't result in identical binaries because C99 mixed
 declarations seem to make GCC emit superfluous stack-pointer
 updates. The compiler is also somewhat dependent on the
 declaration order.
2015-01-23 05:48:27 +00:00
Pieter Wuille
19f3e76002 Remove unused secp256k1_fe_inner_{start, stop} functions 2015-01-22 22:44:52 -05:00
Gregory Maxwell
7688e341c5 Add magnitude limits to secp256k1_fe_verify to ensure that it's own tests function correctly. 2015-01-02 07:52:27 -08:00
Peter Dettman
49ee0dbe16 Add _normalizes_to_zero_var variant 2014-12-20 14:38:29 +01:00
Peter Dettman
eed599dd72 Add _fe_normalizes_to_zero method 2014-12-20 14:38:24 +01:00
Pieter Wuille
d7174edf5f Weak normalization for secp256k1_fe_equal 2014-12-20 14:38:20 +01:00
Pieter Wuille
0295f0a33d weak normalization 2014-12-20 14:38:07 +01:00
Pieter Wuille
1ba4a60a51 Configure options reorganization 2014-12-13 15:04:28 +01:00
Pieter Wuille
6a9901e15b
Merge pull request #137
39bd94d Variable time normalize (Pieter Wuille)
2014-12-07 14:35:23 +01:00
Pieter Wuille
39bd94d86d Variable time normalize 2014-12-06 18:18:28 +01:00
Pieter Wuille
b2c9681c6f Make {mul,sqr}_inner use the same argument order as {mul,sqr} 2014-12-04 13:54:36 +01:00
Pieter Wuille
efb7d4b299 Use constant-time conditional moves instead of byte slicing 2014-12-03 02:41:55 +01:00
Pieter Wuille
be82e92fc4 Require that r and b are different for field multiplication.
Suggested by Peter Dettman, this prepares for slightly faster muitiplication
which writes results immediately to r before finishing reading b.
2014-12-01 13:40:34 +01:00
Pieter Wuille
f24041d6aa Switch all EC/ECDSA logic from num to scalar 2014-11-30 23:38:01 +01:00
Pieter Wuille
d907ebc0e3 Add bounds checking to field element setters 2014-11-26 15:21:31 +01:00
Gregory Maxwell
71712b27e5 Switch to C89 comments in prep for making the whole codebase C89 compatible.
This should be whitespace/comment only changes and should produce the same
object code.
2014-11-15 07:33:07 -08:00
Gregory Maxwell
2cad067a36 Correct function prototypes and avoid unused parameter warnings. 2014-11-12 15:47:06 -08:00
Gregory Maxwell
a4a43d7543 Reorder static to comply with C99 and switch to the inline macro. 2014-11-12 13:07:55 -08:00
Gregory Maxwell
8563713a4f Add non-null and unused-result warnings for the external API.
GCC (and clang) supports extensions to annotate functions so that their
 results must be used and so that their arguments can't be statically
 provable to be null. If a caller violates these requirements they
 get a warning, so this helps them write correct code.

I deployed this in libopus a couple years ago with good success, and
 the implementation here is basically copied straight from that.

One consideration is that the non-null annotation teaches the optimizer
 and will actually compile out runtime non-nullness checks as dead-code.
 Since this is usually not whats wanted, the non-null annotations are
 disabled when compiling the library itself.

The commit also removes some dead inclusions of assert.h and introduces
 compatibility macros for restrict and inline in preparation for some
 portability improvements.
2014-11-12 12:23:09 -08:00
Peter Dettman
137e77afb4 Address 'constant-time' TODOs in field impls 2014-11-04 14:57:04 +07:00
Pieter Wuille
7a8e385d74 Fix interaction between magnitudes and negation
Magnitude m means values are allowed to be up to 2 * 0xFFF...FFF * m,
while the argument passed to secp256k1_fe_negate didn't take the 2 into
account. Fix this.
2014-11-02 01:25:07 -08:00
Peter Dettman
7d681ac636 Add verification to 32bit field
- implement _fe_verify for 10x26
- fe is normalized after _fe_clear
- a few corresponding changes in 64bit field
2014-11-01 16:58:42 +07:00
Pieter Wuille
1c7fa133a6 Add VERIFY_CHECK/DEBUG_CHECK and use CHECK macros more 2014-08-26 01:23:51 +02:00
Gregory Maxwell
2f6c801911
Try to not leave secret data on the stack or heap.
This makes a basic effort and has not been audited.
Doesn't appear to have a measurable performance impact on bench.

It also adds a secp256k1_num_free to secp256k1_ecdsa_pubkey_create.
2014-08-14 07:06:36 -07:00
Pieter Wuille
21f81a8469 Correct secp256k1_fe_verify and use it everywhere 2014-07-17 21:07:40 +02:00
Peter Dettman
59447da38d Test demonstrating discrepancy in sqr output 2014-07-17 21:07:38 +02:00
Peter Dettman
42822baaa8 Improve normalization performance for 32bit
- Uses a similar approach to the latest 64bit _normalize.
- Add one useful optimization back into the 64bit _normalize too.

Performance of 'bench' improved by around 0.5% for the 32bit field (but tested on a 64-bit machine).
2014-06-23 12:12:58 +07:00
Peter Dettman
9037707720 Rewrite 5x52 normalize method to be faster 2014-06-03 18:59:21 +07:00
Pieter Wuille
11ab562203 Move implementations from impl/*.h to *_impl.h 2014-03-12 18:40:02 +01:00