feat(compress): move no-delimiter sequence transfer to Rust

Port the no-delimiter external-sequence block scanner, match splitting, repcode updates, validation, and sequence/literal storage behind a narrow C shim. Preserve trailing bytes as literals when the sequence buffer ends before the requested block, and remove the obsolete C validation wrapper.

Test Plan:

- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (242 passed)

- cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression [--benches|--tests] (clean)

- cargo +nightly fmt --manifest-path rust/Cargo.toml (clean)

- make -B -C lib -j2 lib (passed)

- make -C tests -j2 test-zstream (84 deterministic; 5960 and 8307 randomized passed)

- make -C tests -j2 test-fuzzer (existing test 56 CCtx-reuse mismatch; also fails at 37eb75758)

- make -C tests -j2 test-zstd (existing Rust CLI --patch-from gap)
This commit is contained in:
2026-07-18 07:49:11 +02:00
parent 37eb757585
commit 7ee445c431
2 changed files with 385 additions and 124 deletions
+367
View File
@@ -784,6 +784,183 @@ pub unsafe extern "C" fn ZSTD_rust_transferSequencesWBlockDelim(
block_size
}
/// Transfers externally produced sequences through a block-size boundary when
/// the producer did not provide explicit delimiters. A block may end in the
/// middle of a match; in that case the sequence position is retained for the
/// next call. This is the Rust leaf for
/// `ZSTD_transferSequences_noDelim()`.
#[allow(clippy::too_many_arguments)]
#[no_mangle]
pub unsafe extern "C" fn ZSTD_rust_transferSequencesNoDelim(
seq_store: *mut SeqStore_t,
seq_pos: *mut ZSTD_SequencePosition,
in_seqs: *const ZSTD_Sequence,
in_seqs_size: usize,
src: *const u8,
block_size: usize,
prev_repcodes: *const u32,
next_repcodes: *mut u32,
dict_size: u32,
validate_sequences: c_int,
min_match: u32,
window_log: u32,
use_sequence_producer: c_int,
) -> usize {
if seq_store.is_null()
|| seq_pos.is_null()
|| in_seqs.is_null()
|| src.is_null()
|| prev_repcodes.is_null()
|| next_repcodes.is_null()
|| block_size > u32::MAX as usize
{
return ERROR(ZstdErrorCode::ExternalSequencesInvalid);
}
let seq_store = unsafe { &mut *seq_store };
let seq_pos = unsafe { &mut *seq_pos };
let start_idx = seq_pos.idx as usize;
if start_idx > in_seqs_size {
return ERROR(ZstdErrorCode::ExternalSequencesInvalid);
}
let block_size_u32 = block_size as u32;
let mut start_pos_in_sequence = seq_pos.posInSequence;
let mut end_pos_in_sequence = start_pos_in_sequence.wrapping_add(block_size_u32);
let mut source_offset = 0usize;
let mut bytes_adjustment = 0u32;
let mut final_match_split = false;
let mut updated_repcodes = [0u32; ZSTD_REP_NUM];
unsafe {
ptr::copy_nonoverlapping(prev_repcodes, updated_repcodes.as_mut_ptr(), ZSTD_REP_NUM);
}
let mut idx = start_idx;
while end_pos_in_sequence != 0 && idx < in_seqs_size && !final_match_split {
let current_sequence = unsafe { *in_seqs.add(idx) };
let mut lit_length = current_sequence.litLength;
let mut match_length = current_sequence.matchLength;
let raw_offset = current_sequence.offset;
let current_size = current_sequence
.litLength
.wrapping_add(current_sequence.matchLength);
if end_pos_in_sequence >= current_size {
if start_pos_in_sequence >= lit_length {
start_pos_in_sequence = start_pos_in_sequence.wrapping_sub(lit_length);
lit_length = 0;
match_length = match_length.wrapping_sub(start_pos_in_sequence);
} else {
lit_length = lit_length.wrapping_sub(start_pos_in_sequence);
}
end_pos_in_sequence = end_pos_in_sequence.wrapping_sub(current_size);
start_pos_in_sequence = 0;
} else {
/* The block ends inside this sequence. */
if end_pos_in_sequence > lit_length {
lit_length = if start_pos_in_sequence >= lit_length {
0
} else {
lit_length.wrapping_sub(start_pos_in_sequence)
};
let first_half_match_length = end_pos_in_sequence
.wrapping_sub(start_pos_in_sequence)
.wrapping_sub(lit_length);
if match_length > block_size_u32 && first_half_match_length >= min_match {
let second_half_match_length = current_size.wrapping_sub(end_pos_in_sequence);
let mut first_half_match_length = first_half_match_length;
if second_half_match_length < min_match {
let adjustment = min_match.wrapping_sub(second_half_match_length);
end_pos_in_sequence = end_pos_in_sequence.wrapping_sub(adjustment);
bytes_adjustment = adjustment;
first_half_match_length = first_half_match_length.wrapping_sub(adjustment);
}
match_length = first_half_match_length;
final_match_split = true;
} else {
bytes_adjustment = end_pos_in_sequence.wrapping_sub(current_sequence.litLength);
end_pos_in_sequence = current_sequence.litLength;
break;
}
} else {
/* The block ends inside the literals; leave them as literals. */
break;
}
}
let ll0 = lit_length == 0;
let off_base = unsafe {
ZSTD_rust_finalizeOffBase(raw_offset, updated_repcodes.as_ptr(), u32::from(ll0))
};
update_rep(&mut updated_repcodes, off_base, ll0);
if validate_sequences != 0 {
let sequence_size = lit_length.wrapping_add(match_length) as usize;
seq_pos.posInSrc = seq_pos.posInSrc.wrapping_add(sequence_size);
let validation = ZSTD_rust_validateSequence(
off_base,
match_length,
min_match,
seq_pos.posInSrc,
window_log,
dict_size as usize,
use_sequence_producer,
);
if ERR_isError(validation) {
return validation;
}
}
if idx - start_idx >= seq_store.maxNbSeq {
return ERROR(ZstdErrorCode::ExternalSequencesInvalid);
}
let stored = unsafe {
store_external_sequence(
seq_store,
src,
source_offset,
block_size,
lit_length as usize,
off_base,
match_length as usize,
)
};
if !stored {
return ERROR(ZstdErrorCode::ExternalSequencesInvalid);
}
let sequence_size = lit_length.wrapping_add(match_length) as usize;
source_offset = match source_offset.checked_add(sequence_size) {
Some(offset) => offset,
None => return ERROR(ZstdErrorCode::ExternalSequencesInvalid),
};
if !final_match_split {
idx += 1;
}
}
seq_pos.idx = idx as u32;
seq_pos.posInSequence = end_pos_in_sequence;
unsafe {
ptr::copy_nonoverlapping(updated_repcodes.as_ptr(), next_repcodes, ZSTD_REP_NUM);
}
let consumed_size = match block_size.checked_sub(bytes_adjustment as usize) {
Some(size) => size,
None => return ERROR(ZstdErrorCode::ExternalSequencesInvalid),
};
if source_offset > consumed_size {
return ERROR(ZstdErrorCode::ExternalSequencesInvalid);
}
if source_offset != consumed_size {
let last_literal_size = consumed_size - source_offset;
unsafe {
ZSTD_rust_storeLastLiterals(seq_store, src.add(source_offset), last_literal_size);
}
seq_pos.posInSrc = seq_pos.posInSrc.wrapping_add(last_literal_size);
}
consumed_size
}
/// Finds the next explicit block delimiter and returns the represented size.
///
/// The scan is half-open at `inSeqsSize`: a delimiter at the final element is
@@ -2318,6 +2495,22 @@ mod tests {
}
}
fn empty_external_seq_store(sequences: &mut [SeqDef], literals: &mut [u8]) -> SeqStore_t {
SeqStore_t {
sequencesStart: sequences.as_mut_ptr(),
sequences: sequences.as_mut_ptr(),
litStart: literals.as_mut_ptr(),
lit: literals.as_mut_ptr(),
llCode: std::ptr::null_mut(),
mlCode: std::ptr::null_mut(),
ofCode: std::ptr::null_mut(),
maxNbSeq: sequences.len(),
maxNbLit: literals.len(),
longLengthType: 0,
longLengthPos: 0,
}
}
#[test]
fn c_leaf_layouts_match_supported_abis() {
assert_eq!(size_of::<SeqDef>(), 8);
@@ -2605,6 +2798,180 @@ mod tests {
assert_eq!(literals, [9]);
}
#[test]
fn transfer_sequences_no_delim_splits_a_match_at_the_block_boundary() {
let input = [ZSTD_Sequence {
offset: 50,
litLength: 0,
matchLength: 10,
rep: 0,
}];
let source = [1u8, 2, 3, 4, 5, 6];
let mut output = [SeqDef::default(); 1];
let mut literals = [0u8; 1];
let mut seq_store = empty_external_seq_store(&mut output, &mut literals);
let mut position = ZSTD_SequencePosition::default();
let previous_repcodes = [7u32, 20, 30];
let mut next_repcodes = [0u32; ZSTD_REP_NUM];
let result = unsafe {
ZSTD_rust_transferSequencesNoDelim(
&mut seq_store,
&mut position,
input.as_ptr(),
input.len(),
source.as_ptr(),
source.len(),
previous_repcodes.as_ptr(),
next_repcodes.as_mut_ptr(),
0,
0,
3,
10,
0,
)
};
assert_eq!(result, source.len());
assert_eq!(position.idx, 0);
assert_eq!(position.posInSequence, source.len() as u32);
assert_eq!(position.posInSrc, 0);
assert_eq!(output[0].offBase, 50 + ZSTD_REP_NUM as u32);
assert_eq!(output[0].litLength, 0);
assert_eq!(output[0].mlBase, source.len() as u16 - MINMATCH as u16);
assert_eq!(next_repcodes, [50, 7, 20]);
}
#[test]
fn transfer_sequences_no_delim_shortens_a_match_to_keep_the_tail_valid() {
let input = [ZSTD_Sequence {
offset: 50,
litLength: 0,
matchLength: 10,
rep: 0,
}];
let source = [0u8; 8];
let mut output = [SeqDef::default(); 1];
let mut literals = [0u8; 1];
let mut seq_store = empty_external_seq_store(&mut output, &mut literals);
let mut position = ZSTD_SequencePosition::default();
let previous_repcodes = [7u32, 20, 30];
let mut next_repcodes = [0u32; ZSTD_REP_NUM];
let result = unsafe {
ZSTD_rust_transferSequencesNoDelim(
&mut seq_store,
&mut position,
input.as_ptr(),
input.len(),
source.as_ptr(),
source.len(),
previous_repcodes.as_ptr(),
next_repcodes.as_mut_ptr(),
0,
0,
3,
10,
0,
)
};
assert_eq!(result, 7);
assert_eq!(position.idx, 0);
assert_eq!(position.posInSequence, 7);
assert_eq!(output[0].offBase, 50 + ZSTD_REP_NUM as u32);
assert_eq!(output[0].mlBase, 7 - MINMATCH as u16);
assert_eq!(next_repcodes, [50, 7, 20]);
}
#[test]
fn transfer_sequences_no_delim_leaves_a_literal_prefix_unsequenced() {
let input = [ZSTD_Sequence {
offset: 50,
litLength: 4,
matchLength: 6,
rep: 0,
}];
let source = [1u8, 2, 3, 4, 5, 6];
let mut output = [SeqDef::default(); 1];
let mut literals = [0u8; 2];
let mut seq_store = empty_external_seq_store(&mut output, &mut literals);
let mut position = ZSTD_SequencePosition::default();
let previous_repcodes = [7u32, 20, 30];
let mut next_repcodes = [0u32; ZSTD_REP_NUM];
let result = unsafe {
ZSTD_rust_transferSequencesNoDelim(
&mut seq_store,
&mut position,
input.as_ptr(),
input.len(),
source.as_ptr(),
2,
previous_repcodes.as_ptr(),
next_repcodes.as_mut_ptr(),
0,
0,
3,
10,
0,
)
};
assert_eq!(result, 2);
assert_eq!(position.idx, 0);
assert_eq!(position.posInSequence, 2);
assert_eq!(position.posInSrc, 2);
assert_eq!(seq_store.sequences, seq_store.sequencesStart);
assert_eq!(literals, [1, 2]);
assert_eq!(next_repcodes, previous_repcodes);
}
#[test]
fn transfer_sequences_no_delim_treats_bytes_after_the_last_sequence_as_literals() {
let input = [ZSTD_Sequence {
offset: 50,
litLength: 2,
matchLength: 3,
rep: 0,
}];
let source = [10u8, 11, 12, 13, 14, 15, 16];
let mut output = [SeqDef::default(); 1];
let mut literals = [0u8; 4];
let mut seq_store = empty_external_seq_store(&mut output, &mut literals);
let mut position = ZSTD_SequencePosition::default();
let previous_repcodes = [7u32, 20, 30];
let mut next_repcodes = [0u32; ZSTD_REP_NUM];
let result = unsafe {
ZSTD_rust_transferSequencesNoDelim(
&mut seq_store,
&mut position,
input.as_ptr(),
input.len(),
source.as_ptr(),
source.len(),
previous_repcodes.as_ptr(),
next_repcodes.as_mut_ptr(),
0,
0,
3,
10,
0,
)
};
assert_eq!(result, source.len());
assert_eq!(position.idx, 1);
assert_eq!(position.posInSequence, 2);
assert_eq!(position.posInSrc, 2);
assert_eq!(output[0].offBase, 50 + ZSTD_REP_NUM as u32);
assert_eq!(output[0].litLength, 2);
assert_eq!(output[0].mlBase, 0);
assert_eq!(literals, [10, 11, 15, 16]);
assert_eq!(next_repcodes, [50, 7, 20]);
}
#[test]
fn transfer_sequences_rejects_invalid_match_lengths_and_block_mismatches() {
let invalid_match = [