Fix the unittest with Emscripten

Split main into two helper functions since having a few ML-KEM key blobs
on the stack makes Emscripten's stack explode, leading to inscrutable
silent failures unles

Fix the unittest with Emscripten

Split main into two helper functions since having a few ML-KEM key blobs
on the stack makes Emscripten's stack explode, leading to inscrutable
silent failures unles ASAN is enabled. Go figure.

show more ...

Overhaul ML-KEM regress once more

Implement a file parser that drives a state machine to extract the test data
from the .txt files and manages the parsed data. Comments and empty lines are
ignored.

Overhaul ML-KEM regress once more

Implement a file parser that drives a state machine to extract the test data
from the .txt files and manages the parsed data. Comments and empty lines are
ignored. The code currently assumes that instruction lines are at the start
of the file (which isn't generally true) and only supports two line types for
now. This is good enough for all the ML-KEM tests but should be easy enough
to extend.

Once all data for a test case is parsed in the expected order, a test handler
is called which can retrieve the test data via a simple API and throw warnings
and errors with information on the test case line number, etc.

Merge the tests into three programs: one parsing the .txt files and running
the corresponding test cases, a unit test and the iteration tests. Deduplicate
the actual test code and let the caller pass in an object containing the
API functions, private keys and arrays that need to be different between the
768 version and the 1024 version. This way we don't have two sets of half a
dozen .c files differing only in 3 or 4 occurrences of 768 and 1024.

All this will also make it a lot easier to hook these tests into portable.

show more ...

Rework and fix the mlkem tests

Make proper use of CBB and CBS. If a CBS ever owns data, you're holding
it wrong. Ditch gross macros, sscanf, and globals. The use of fgets is
annoying here, so replac

Rework and fix the mlkem tests

Make proper use of CBB and CBS. If a CBS ever owns data, you're holding
it wrong. Ditch gross macros, sscanf, and globals. The use of fgets is
annoying here, so replace it with getline, which be provided by portable
if needed.

Most importantly, make the tests actually signal failure rather than
only printing an error. Fix the state machines in a few of them. Some
tests didn't parse the .txt file at all. Others mostly did but didn't
actually test what they were supposed to be testing. Such failures
were hidden by the way the tests were written.

This basically needed a complete revamp. It still isn't pretty and much
of it could be deduplicated, but I only have so much time alotted on this
blue planet.

show more ...

Normalize top of files

RCS marker, KNF for comment, fix and sort includes as usual.

Add ML-KEM 1024 from BoringSSL

Changes include conversion from C++, basic KNF, then adaptation to
use our sha3 functions for sha3 and shake instead of the BorinSSL
version. This Adds units tests to

Add ML-KEM 1024 from BoringSSL

Changes include conversion from C++, basic KNF, then adaptation to
use our sha3 functions for sha3 and shake instead of the BorinSSL
version. This Adds units tests to run against BoringSSL and NIST test
vectors.

The future public API is the same as Boring's - but is not yet exposed
pending making bytestring.h public (which will happen separately) and
a minor bump

Currently this will just ensure we build and run regress.

ok tb@ to get it into the tree and massage from there.

show more ...

Add ML-KEM 768 from BoringSSL

Changes include conversion from C++, basic KNF, then adaptation to
use our sha3 functions for sha3 and shake instead of the BorinSSL
version. This Adds units tests to r

Add ML-KEM 768 from BoringSSL

Changes include conversion from C++, basic KNF, then adaptation to
use our sha3 functions for sha3 and shake instead of the BorinSSL
version. This Adds units tests to run against BoringSSL and NIST test
vectors.

The future public API is the same as Boring's - but is not yet exposed
pending making bytesring.h public (which will happen separately) and
a minor bump

Currently this will just ensure we build and run regress.

ok tb@ to get it into the tree and massage from there.

show more ...