Fixed-base scalar mul makes use of the add_incomplete and add
instructions internally. The full-width and short signed share
some common logic, which is captured in chip::mul_fixed.rs.
The signed short variant introduces additional logic to handle
the scalar's sign. This is done in the submodule mul_fixed::short.
A scalar used in fixed-base scalar mul needs to be decomposed into
windows to use with the fixed-base window table. Both full-width
and short signed scalars share some logic (captured in the function
decompose_scalar_fixed()).
A short signed scalar introduces additional logic: its magnitude is
decomposed, and its sign is separately witnessed. This is handled
in the submodule witness_scalar_fixed::short.
This uses the complete addition instruction internally. The module
is split up into mul::incomplete.rs and mul::complete.rs, where
mul::incomplete handles the incomplete additions used in the
starting rounds of the variable-base scalar mul algorithm, and
mul::complete handles the complete additions in the final rounds.
Incomplete additions are broken into "hi" and "lo" halves and
processed on the same rows across different columns. This is an
optimization to make full use of the advice columns in this
instruction.
This has three const generic parameters: PATH_LENGTH, K, MAX_WORDS.
PATH_LENGTH is the length of the Merkle path being hashed. K and
MAX_WORDS parameterize the internal Sinsemilla instance used in
hashing the path.
These instructions were not making any assignments; instead, they
were calling through to witness_message_piece_field().
This PR also renames the witness_message_piece_field() instruction
to witness_message_piece().
Also introduce a "strict" mode for the full-length lookup, where
"true" requires the field element to be within num_words * K bits,
whereas "false" does not.
hash_piece() is an internal API, which means its caller hash_message()
is working in the same region. We rely on the caller to have already
assigned each piece's initial x_a at the correct offset before making
the call to hash_piece().
Co-authored-by: Jack Grigg <jack@electriccoin.co>
This toggles the assignment of q_s2 on the last row of each piece.
We assign q_s2 = 2 on the last row of the final piece, and q_s2 = 0
on the last row of other pieces.
This allows us to process the final_piece in the main loop together
with the other pieces.
Co-authored-by: Jack Grigg <jack@electriccoin.co>
Also, GeneratorTable::configure() was not being called in the main
SinsemillaChip::configure(), which meant the lookup argument had
not been activated. This has now been fixed.
Co-authored-by: Jack Grigg <jack@electriccoin.co>
witness_message() witnesses a full message given a bitstring.
The other two APIs, witness_message_piece_bitstring() and
witness_message_piece_field(), both witness a message piece, i.e.
part of a message that fits within a single base field element.
witness_message_piece_bitstring() takes in a bitstring, while
witness_message_piece_field() takes in a field element. In the
latter case, the number of words encoded must be specified.
This defines a Sinsemilla message in terms of pieces and subpieces.
This is useful when decomposing field elements and packing them
into K-bit messages.
SinsemillaInstructions has two const generic parameters: K, which
is the number of bits in each word of the hash, and MAX_WORDS,
which is the maximum number of words the hash can process.
For Orchard, K = 10, MAX_WORDS = 253.