feat(cli): move shared decompression scheduling into Rust

Move the shared-destination branch of FIO_decompressMultipleFilenames into a Rust-owned scheduler. Preserve source order, destination propagation, cursor advancement, success counting, aggregate errors, and C-owned source/format/resource behavior including --rm and output close handling.

Test Plan:
- ulimit -v 41943040; CARGO_BUILD_JOBS=1 cargo test --manifest-path rust/Cargo.toml decompression_multiple_shared_destination -- --test-threads=1
- ulimit -v 41943040; make -B -C programs -j1 zstd
- ulimit -v 41943040; make -B -C tests -j1 test-cli-tests
This commit is contained in:
2026-07-19 11:10:09 +02:00
parent 4dc6761e60
commit 802cb70184
2 changed files with 220 additions and 4 deletions
+55 -4
View File
@@ -1179,6 +1179,35 @@ typedef char FIO_rust_compress_multiple_separate_projection_size[
int FIO_rust_compressMultipleSeparateFilenames(
const FIO_rust_compress_multiple_separate_projection_t* projection);
typedef int (*FIO_rust_decompress_multiple_file_fn)(
void* opaque, const char* outFileName, const char* srcFileName);
typedef struct {
void* fCtx;
const char** srcNamesTable;
const char* outFileName;
void* opaque;
FIO_rust_decompress_multiple_file_fn decompressFile;
} FIO_rust_decompress_multiple_projection_t;
typedef char FIO_rust_decompress_multiple_fctx_offset[
(offsetof(FIO_rust_decompress_multiple_projection_t, fCtx) == 0) ? 1 : -1];
typedef char FIO_rust_decompress_multiple_input_names_offset[
(offsetof(FIO_rust_decompress_multiple_projection_t, srcNamesTable)
== sizeof(void*)) ? 1 : -1];
typedef char FIO_rust_decompress_multiple_output_name_offset[
(offsetof(FIO_rust_decompress_multiple_projection_t, outFileName)
== 2 * sizeof(void*)) ? 1 : -1];
typedef char FIO_rust_decompress_multiple_opaque_offset[
(offsetof(FIO_rust_decompress_multiple_projection_t, opaque)
== 3 * sizeof(void*)) ? 1 : -1];
typedef char FIO_rust_decompress_multiple_callback_offset[
(offsetof(FIO_rust_decompress_multiple_projection_t, decompressFile)
== 4 * sizeof(void*)) ? 1 : -1];
typedef char FIO_rust_decompress_multiple_projection_size[
(sizeof(FIO_rust_decompress_multiple_projection_t)
== 4 * sizeof(void*) + sizeof(FIO_rust_decompress_multiple_file_fn)) ? 1 : -1];
int FIO_rust_decompressMultipleFilenames(
const FIO_rust_decompress_multiple_projection_t* projection);
enum {
FIO_RUST_ZSTD_OK = 0,
FIO_RUST_ZSTD_COMPRESS_ERROR = 1,
@@ -3838,6 +3867,23 @@ static int FIO_decompressSrcFile(FIO_ctx_t* const fCtx, FIO_prefs_t* const prefs
return result;
}
typedef struct {
FIO_ctx_t* fCtx;
FIO_prefs_t* prefs;
dRess_t* ress;
} FIO_rust_decompress_multiple_context_t;
/* Keep source opening, format dispatch, diagnostics, and --rm in C. */
static int FIO_rust_decompressMultipleFileCallback(void* opaque,
const char* outFileName,
const char* srcFileName)
{
FIO_rust_decompress_multiple_context_t* const context =
(FIO_rust_decompress_multiple_context_t*)opaque;
return FIO_decompressSrcFile(
context->fCtx, context->prefs, *context->ress, outFileName, srcFileName);
}
int FIO_decompressFilename(FIO_ctx_t* const fCtx, FIO_prefs_t* const prefs,
@@ -3921,10 +3967,15 @@ FIO_decompressMultipleFilenames(FIO_ctx_t* const fCtx,
if (dstFile == 0) EXM_THROW(19, "cannot open %s", outFileName);
AIO_WritePool_setFile(ress.writeCtx, dstFile);
}
for (; fCtx->currFileIdx < fCtx->nbFilesTotal; fCtx->currFileIdx++) {
status = FIO_decompressSrcFile(fCtx, prefs, ress, outFileName, srcNamesTable[fCtx->currFileIdx]);
if (!status) fCtx->nbFilesProcessed++;
error |= status;
{
FIO_rust_decompress_multiple_context_t callbackContext = {
fCtx, prefs, &ress
};
FIO_rust_decompress_multiple_projection_t projection = {
fCtx, srcNamesTable, outFileName,
&callbackContext, FIO_rust_decompressMultipleFileCallback
};
error = FIO_rust_decompressMultipleFilenames(&projection);
}
if ((!prefs->testMode) && (AIO_WritePool_closeFile(ress.writeCtx)))
EXM_THROW(72, "Write error : %s : cannot properly close output file",
+165
View File
@@ -332,6 +332,42 @@ const _: () = {
);
};
pub type FIO_rust_decompress_multiple_file_fn =
unsafe extern "C" fn(*mut c_void, *const c_char, *const c_char) -> c_int;
/// Rust owns only the shared-destination file iteration and aggregate error
/// handling. C retains source opening, format dispatch, diagnostics, and
/// source removal through the per-source callback.
#[repr(C)]
pub struct FIO_rust_decompress_multiple_projection_t {
pub f_ctx: *mut c_void,
pub src_names_table: *const *const c_char,
pub out_file_name: *const c_char,
pub opaque: *mut c_void,
pub decompress_file: Option<FIO_rust_decompress_multiple_file_fn>,
}
const _: () = {
assert!(std::mem::offset_of!(FIO_rust_decompress_multiple_projection_t, f_ctx) == 0);
assert!(
std::mem::offset_of!(FIO_rust_decompress_multiple_projection_t, src_names_table)
== size_of::<usize>()
);
assert!(
std::mem::offset_of!(FIO_rust_decompress_multiple_projection_t, out_file_name)
== 2 * size_of::<usize>()
);
assert!(
std::mem::offset_of!(FIO_rust_decompress_multiple_projection_t, opaque)
== 3 * size_of::<usize>()
);
assert!(
std::mem::offset_of!(FIO_rust_decompress_multiple_projection_t, decompress_file)
== 4 * size_of::<usize>()
);
assert!(size_of::<FIO_rust_decompress_multiple_file_fn>() == size_of::<usize>());
assert!(size_of::<FIO_rust_decompress_multiple_projection_t>() == 5 * size_of::<usize>());
};
pub const FIO_RUST_ZSTD_OK: c_int = 0;
pub const FIO_RUST_ZSTD_COMPRESS_ERROR: c_int = 1;
pub const FIO_RUST_ZSTD_INCOMPLETE_INPUT: c_int = 2;
@@ -2272,6 +2308,42 @@ pub unsafe extern "C" fn FIO_rust_compressMultipleSeparateFilenames(
error
}
/// Iterate files that share one already-open decompression destination. The
/// C callback retains source opening, format dispatch, diagnostics, and
/// `--rm`; Rust preserves the non-short-circuiting loop and counter updates.
#[no_mangle]
pub unsafe extern "C" fn FIO_rust_decompressMultipleFilenames(
projection: *const FIO_rust_decompress_multiple_projection_t,
) -> c_int {
assert!(!projection.is_null());
let projection = unsafe { &*projection };
assert!(!projection.f_ctx.is_null());
assert!(!projection.src_names_table.is_null());
assert!(!projection.out_file_name.is_null());
let decompress_file = projection
.decompress_file
.expect("shared-destination decompression callback is required");
let f_ctx = projection.f_ctx.cast::<FIO_rust_compression_context_t>();
let mut error = 0;
while unsafe { (*f_ctx).currFileIdx < (*f_ctx).nbFilesTotal } {
let file_index = unsafe { (*f_ctx).currFileIdx as usize };
let src_file_name = unsafe { *projection.src_names_table.add(file_index) };
let status =
unsafe { decompress_file(projection.opaque, projection.out_file_name, src_file_name) };
if status == 0 {
unsafe {
(*f_ctx).nbFilesProcessed = (*f_ctx).nbFilesProcessed.wrapping_add(1);
}
}
error |= status;
unsafe {
(*f_ctx).currFileIdx = (*f_ctx).currFileIdx.wrapping_add(1);
}
}
error
}
#[inline]
fn zstd_adapt_level(projection: &FIO_rust_zstd_adapt_projection_t, slower: bool) -> c_int {
if slower {
@@ -4573,6 +4645,29 @@ mod tests {
status
}
#[derive(Default)]
struct MultipleDecompressionState {
sources: Vec<*const c_char>,
destinations: Vec<*const c_char>,
statuses: Vec<c_int>,
}
unsafe extern "C" fn record_multiple_decompression_file(
opaque: *mut c_void,
destination: *const c_char,
source: *const c_char,
) -> c_int {
let state = unsafe { &mut *opaque.cast::<MultipleDecompressionState>() };
let status = state
.statuses
.get(state.sources.len())
.copied()
.unwrap_or(0);
state.sources.push(source);
state.destinations.push(destination);
status
}
#[test]
fn compression_multiple_separate_destinations_preserves_order_and_counters() {
let source_names = [c"zero".as_ptr(), c"one".as_ptr(), c"two".as_ptr()];
@@ -4707,6 +4802,76 @@ mod tests {
assert_eq!(context.nbFilesProcessed, 4);
}
#[test]
fn decompression_multiple_shared_destination_preserves_order_and_counters() {
let source_names = [c"zero".as_ptr(), c"one".as_ptr(), c"two".as_ptr()];
let destination = c"archive";
let mut context = FIO_rust_compression_context_t {
nbFilesTotal: 3,
hasStdinInput: 0,
hasStdoutOutput: 0,
currFileIdx: 1,
nbFilesProcessed: 5,
totalBytesInput: 0,
totalBytesOutput: 0,
};
let mut state = MultipleDecompressionState {
statuses: vec![0, 0],
..MultipleDecompressionState::default()
};
let projection = FIO_rust_decompress_multiple_projection_t {
f_ctx: (&mut context as *mut FIO_rust_compression_context_t).cast(),
src_names_table: source_names.as_ptr(),
out_file_name: destination.as_ptr(),
opaque: (&mut state as *mut MultipleDecompressionState).cast(),
decompress_file: Some(record_multiple_decompression_file),
};
assert_eq!(
unsafe { FIO_rust_decompressMultipleFilenames(&projection) },
0
);
assert_eq!(state.sources, vec![source_names[1], source_names[2]]);
assert_eq!(state.destinations, vec![destination.as_ptr(); 2]);
assert_eq!(context.currFileIdx, 3);
assert_eq!(context.nbFilesProcessed, 7);
}
#[test]
fn decompression_multiple_shared_destination_accumulates_errors_without_short_circuiting() {
let source_names = [c"one".as_ptr(), c"two".as_ptr(), c"three".as_ptr()];
let destination = c"archive";
let mut context = FIO_rust_compression_context_t {
nbFilesTotal: 3,
hasStdinInput: 0,
hasStdoutOutput: 0,
currFileIdx: 0,
nbFilesProcessed: 4,
totalBytesInput: 0,
totalBytesOutput: 0,
};
let mut state = MultipleDecompressionState {
statuses: vec![1, 4, 2],
..MultipleDecompressionState::default()
};
let projection = FIO_rust_decompress_multiple_projection_t {
f_ctx: (&mut context as *mut FIO_rust_compression_context_t).cast(),
src_names_table: source_names.as_ptr(),
out_file_name: destination.as_ptr(),
opaque: (&mut state as *mut MultipleDecompressionState).cast(),
decompress_file: Some(record_multiple_decompression_file),
};
assert_eq!(
unsafe { FIO_rust_decompressMultipleFilenames(&projection) },
1 | 4 | 2
);
assert_eq!(state.sources, source_names);
assert_eq!(state.destinations, vec![destination.as_ptr(); 3]);
assert_eq!(context.currFileIdx, 3);
assert_eq!(context.nbFilesProcessed, 4);
}
struct ZstdProjectionState {
input: [u8; 5],
input_pos: usize,