block_buffer_encoder.c | block_buffer_encoder.c | |||
---|---|---|---|---|
skipping to change at line 13 | skipping to change at line 13 | |||
/// \file block_buffer_encoder.c | /// \file block_buffer_encoder.c | |||
/// \brief Single-call .xz Block encoder | /// \brief Single-call .xz Block encoder | |||
// | // | |||
// Author: Lasse Collin | // Author: Lasse Collin | |||
// | // | |||
// This file has been put into the public domain. | // This file has been put into the public domain. | |||
// You can do whatever you want with this file. | // You can do whatever you want with this file. | |||
// | // | |||
/////////////////////////////////////////////////////////////////////////// //// | /////////////////////////////////////////////////////////////////////////// //// | |||
#include "block_buffer_encoder.h" | ||||
#include "block_encoder.h" | #include "block_encoder.h" | |||
#include "filter_encoder.h" | #include "filter_encoder.h" | |||
#include "lzma2_encoder.h" | #include "lzma2_encoder.h" | |||
#include "check.h" | #include "check.h" | |||
/// Estimate the maximum size of the Block Header and Check fields for | /// Estimate the maximum size of the Block Header and Check fields for | |||
/// a Block that uses LZMA2 uncompressed chunks. We could use | /// a Block that uses LZMA2 uncompressed chunks. We could use | |||
/// lzma_block_header_size() but this is simpler. | /// lzma_block_header_size() but this is simpler. | |||
/// | /// | |||
/// Block Header Size + Block Flags + Compressed Size | /// Block Header Size + Block Flags + Compressed Size | |||
/// + Uncompressed Size + Filter Flags for LZMA2 + CRC32 + Check | /// + Uncompressed Size + Filter Flags for LZMA2 + CRC32 + Check | |||
/// and round up to the next multiple of four to take Header Padding | /// and round up to the next multiple of four to take Header Padding | |||
/// into account. | /// into account. | |||
#define HEADERS_BOUND ((1 + 1 + 2 * LZMA_VLI_BYTES_MAX + 3 + 4 \ | #define HEADERS_BOUND ((1 + 1 + 2 * LZMA_VLI_BYTES_MAX + 3 + 4 \ | |||
+ LZMA_CHECK_SIZE_MAX + 3) & ~3) | + LZMA_CHECK_SIZE_MAX + 3) & ~3) | |||
static lzma_vli | static uint64_t | |||
lzma2_bound(lzma_vli uncompressed_size) | lzma2_bound(uint64_t uncompressed_size) | |||
{ | { | |||
// Prevent integer overflow in overhead calculation. | // Prevent integer overflow in overhead calculation. | |||
if (uncompressed_size > COMPRESSED_SIZE_MAX) | if (uncompressed_size > COMPRESSED_SIZE_MAX) | |||
return 0; | return 0; | |||
// Calculate the exact overhead of the LZMA2 headers: Round | // Calculate the exact overhead of the LZMA2 headers: Round | |||
// uncompressed_size up to the next multiple of LZMA2_CHUNK_MAX, | // uncompressed_size up to the next multiple of LZMA2_CHUNK_MAX, | |||
// multiply by the size of per-chunk header, and add one byte for | // multiply by the size of per-chunk header, and add one byte for | |||
// the end marker. | // the end marker. | |||
const lzma_vli overhead = ((uncompressed_size + LZMA2_CHUNK_MAX - 1) | const uint64_t overhead = ((uncompressed_size + LZMA2_CHUNK_MAX - 1) | |||
/ LZMA2_CHUNK_MAX) | / LZMA2_CHUNK_MAX) | |||
* LZMA2_HEADER_UNCOMPRESSED + 1; | * LZMA2_HEADER_UNCOMPRESSED + 1; | |||
// Catch the possible integer overflow. | // Catch the possible integer overflow. | |||
if (COMPRESSED_SIZE_MAX - overhead < uncompressed_size) | if (COMPRESSED_SIZE_MAX - overhead < uncompressed_size) | |||
return 0; | return 0; | |||
return uncompressed_size + overhead; | return uncompressed_size + overhead; | |||
} | } | |||
extern LZMA_API(size_t) | extern uint64_t | |||
lzma_block_buffer_bound(size_t uncompressed_size) | lzma_block_buffer_bound64(uint64_t uncompressed_size) | |||
{ | { | |||
// For now, if the data doesn't compress, we always use uncompressed | // If the data doesn't compress, we always use uncompressed | |||
// chunks of LZMA2. In future we may use Subblock filter too, but | // LZMA2 chunks. | |||
// but for simplicity we probably will still use the same bound | uint64_t lzma2_size = lzma2_bound(uncompressed_size); | |||
// calculation even though Subblock filter would have slightly less | ||||
// overhead. | ||||
lzma_vli lzma2_size = lzma2_bound(uncompressed_size); | ||||
if (lzma2_size == 0) | if (lzma2_size == 0) | |||
return 0; | return 0; | |||
// Take Block Padding into account. | // Take Block Padding into account. | |||
lzma2_size = (lzma2_size + 3) & ~LZMA_VLI_C(3); | lzma2_size = (lzma2_size + 3) & ~UINT64_C(3); | |||
#if SIZE_MAX < LZMA_VLI_MAX | // No risk of integer overflow because lzma2_bound() already takes | |||
// Catch the possible integer overflow on 32-bit systems. There's no | ||||
// overflow on 64-bit systems, because lzma2_bound() already takes | ||||
// into account the size of the headers in the Block. | // into account the size of the headers in the Block. | |||
if (SIZE_MAX - HEADERS_BOUND < lzma2_size) | return HEADERS_BOUND + lzma2_size; | |||
} | ||||
extern LZMA_API(size_t) | ||||
lzma_block_buffer_bound(size_t uncompressed_size) | ||||
{ | ||||
uint64_t ret = lzma_block_buffer_bound64(uncompressed_size); | ||||
#if SIZE_MAX < UINT64_MAX | ||||
// Catch the possible integer overflow on 32-bit systems. | ||||
if (ret > SIZE_MAX) | ||||
return 0; | return 0; | |||
#endif | #endif | |||
return HEADERS_BOUND + lzma2_size; | return ret; | |||
} | } | |||
static lzma_ret | static lzma_ret | |||
block_encode_uncompressed(lzma_block *block, const uint8_t *in, size_t in_s ize, | block_encode_uncompressed(lzma_block *block, const uint8_t *in, size_t in_s ize, | |||
uint8_t *out, size_t *out_pos, size_t out_size) | uint8_t *out, size_t *out_pos, size_t out_size) | |||
{ | { | |||
// TODO: Figure out if the last filter is LZMA2 or Subblock and use | ||||
// that filter to encode the uncompressed chunks. | ||||
// Use LZMA2 uncompressed chunks. We wouldn't need a dictionary at | // Use LZMA2 uncompressed chunks. We wouldn't need a dictionary at | |||
// all, but LZMA2 always requires a dictionary, so use the minimum | // all, but LZMA2 always requires a dictionary, so use the minimum | |||
// value to minimize memory usage of the decoder. | // value to minimize memory usage of the decoder. | |||
lzma_options_lzma lzma2 = { | lzma_options_lzma lzma2 = { | |||
.dict_size = LZMA_DICT_SIZE_MIN, | .dict_size = LZMA_DICT_SIZE_MIN, | |||
}; | }; | |||
lzma_filter filters[2]; | lzma_filter filters[2]; | |||
filters[0].id = LZMA_FILTER_LZMA2; | filters[0].id = LZMA_FILTER_LZMA2; | |||
filters[0].options = &lzma2; | filters[0].options = &lzma2; | |||
skipping to change at line 158 | skipping to change at line 161 | |||
} | } | |||
// End marker | // End marker | |||
out[(*out_pos)++] = 0x00; | out[(*out_pos)++] = 0x00; | |||
assert(*out_pos <= out_size); | assert(*out_pos <= out_size); | |||
return LZMA_OK; | return LZMA_OK; | |||
} | } | |||
static lzma_ret | static lzma_ret | |||
block_encode_normal(lzma_block *block, lzma_allocator *allocator, | block_encode_normal(lzma_block *block, const lzma_allocator *allocator, | |||
const uint8_t *in, size_t in_size, | const uint8_t *in, size_t in_size, | |||
uint8_t *out, size_t *out_pos, size_t out_size) | uint8_t *out, size_t *out_pos, size_t out_size) | |||
{ | { | |||
// Find out the size of the Block Header. | // Find out the size of the Block Header. | |||
block->compressed_size = lzma2_bound(in_size); | ||||
if (block->compressed_size == 0) | ||||
return LZMA_DATA_ERROR; | ||||
block->uncompressed_size = in_size; | ||||
return_if_error(lzma_block_header_size(block)); | return_if_error(lzma_block_header_size(block)); | |||
// Reserve space for the Block Header and skip it for now. | // Reserve space for the Block Header and skip it for now. | |||
if (out_size - *out_pos <= block->header_size) | if (out_size - *out_pos <= block->header_size) | |||
return LZMA_BUF_ERROR; | return LZMA_BUF_ERROR; | |||
const size_t out_start = *out_pos; | const size_t out_start = *out_pos; | |||
*out_pos += block->header_size; | *out_pos += block->header_size; | |||
// Limit out_size so that we stop encoding if the output would grow | // Limit out_size so that we stop encoding if the output would grow | |||
skipping to change at line 218 | skipping to change at line 216 | |||
ret = LZMA_BUF_ERROR; | ret = LZMA_BUF_ERROR; | |||
} | } | |||
// Reset *out_pos if something went wrong. | // Reset *out_pos if something went wrong. | |||
if (ret != LZMA_OK) | if (ret != LZMA_OK) | |||
*out_pos = out_start; | *out_pos = out_start; | |||
return ret; | return ret; | |||
} | } | |||
extern LZMA_API(lzma_ret) | static lzma_ret | |||
lzma_block_buffer_encode(lzma_block *block, lzma_allocator *allocator, | block_buffer_encode(lzma_block *block, const lzma_allocator *allocator, | |||
const uint8_t *in, size_t in_size, | const uint8_t *in, size_t in_size, | |||
uint8_t *out, size_t *out_pos, size_t out_size) | uint8_t *out, size_t *out_pos, size_t out_size, | |||
bool try_to_compress) | ||||
{ | { | |||
// Validate the arguments. | // Validate the arguments. | |||
if (block == NULL || (in == NULL && in_size != 0) || out == NULL | if (block == NULL || (in == NULL && in_size != 0) || out == NULL | |||
|| out_pos == NULL || *out_pos > out_size) | || out_pos == NULL || *out_pos > out_size) | |||
return LZMA_PROG_ERROR; | return LZMA_PROG_ERROR; | |||
// The contents of the structure may depend on the version so | // The contents of the structure may depend on the version so | |||
// check the version before validating the contents of *block. | // check the version before validating the contents of *block. | |||
if (block->version != 0) | if (block->version > 1) | |||
return LZMA_OPTIONS_ERROR; | return LZMA_OPTIONS_ERROR; | |||
if ((unsigned int)(block->check) > LZMA_CHECK_ID_MAX | if ((unsigned int)(block->check) > LZMA_CHECK_ID_MAX | |||
|| block->filters == NULL) | || (try_to_compress && block->filters == NULL)) | |||
return LZMA_PROG_ERROR; | return LZMA_PROG_ERROR; | |||
if (!lzma_check_is_supported(block->check)) | if (!lzma_check_is_supported(block->check)) | |||
return LZMA_UNSUPPORTED_CHECK; | return LZMA_UNSUPPORTED_CHECK; | |||
// Size of a Block has to be a multiple of four, so limit the size | // Size of a Block has to be a multiple of four, so limit the size | |||
// here already. This way we don't need to check it again when addin g | // here already. This way we don't need to check it again when addin g | |||
// Block Padding. | // Block Padding. | |||
out_size -= (out_size - *out_pos) & 3; | out_size -= (out_size - *out_pos) & 3; | |||
// Get the size of the Check field. | // Get the size of the Check field. | |||
const size_t check_size = lzma_check_size(block->check); | const size_t check_size = lzma_check_size(block->check); | |||
assert(check_size != UINT32_MAX); | assert(check_size != UINT32_MAX); | |||
// Reserve space for the Check field. | // Reserve space for the Check field. | |||
if (out_size - *out_pos <= check_size) | if (out_size - *out_pos <= check_size) | |||
return LZMA_BUF_ERROR; | return LZMA_BUF_ERROR; | |||
out_size -= check_size; | out_size -= check_size; | |||
// Initialize block->uncompressed_size and calculate the worst-case | ||||
// value for block->compressed_size. | ||||
block->uncompressed_size = in_size; | ||||
block->compressed_size = lzma2_bound(in_size); | ||||
if (block->compressed_size == 0) | ||||
return LZMA_DATA_ERROR; | ||||
// Do the actual compression. | // Do the actual compression. | |||
const lzma_ret ret = block_encode_normal(block, allocator, | lzma_ret ret = LZMA_BUF_ERROR; | |||
in, in_size, out, out_pos, out_size); | if (try_to_compress) | |||
ret = block_encode_normal(block, allocator, | ||||
in, in_size, out, out_pos, out_size); | ||||
if (ret != LZMA_OK) { | if (ret != LZMA_OK) { | |||
// If the error was something else than output buffer | // If the error was something else than output buffer | |||
// becoming full, return the error now. | // becoming full, return the error now. | |||
if (ret != LZMA_BUF_ERROR) | if (ret != LZMA_BUF_ERROR) | |||
return ret; | return ret; | |||
// The data was uncompressible (at least with the options | // The data was uncompressible (at least with the options | |||
// given to us) or the output buffer was too small. Use the | // given to us) or the output buffer was too small. Use the | |||
// uncompressed chunks of LZMA2 to wrap the data into a vali d | // uncompressed chunks of LZMA2 to wrap the data into a vali d | |||
// Block. If we haven't been given enough output space, even | // Block. If we haven't been given enough output space, even | |||
skipping to change at line 300 | skipping to change at line 309 | |||
lzma_check_update(&check, block->check, in, in_size); | lzma_check_update(&check, block->check, in, in_size); | |||
lzma_check_finish(&check, block->check); | lzma_check_finish(&check, block->check); | |||
memcpy(block->raw_check, check.buffer.u8, check_size); | memcpy(block->raw_check, check.buffer.u8, check_size); | |||
memcpy(out + *out_pos, check.buffer.u8, check_size); | memcpy(out + *out_pos, check.buffer.u8, check_size); | |||
*out_pos += check_size; | *out_pos += check_size; | |||
} | } | |||
return LZMA_OK; | return LZMA_OK; | |||
} | } | |||
extern LZMA_API(lzma_ret) | ||||
lzma_block_buffer_encode(lzma_block *block, const lzma_allocator *allocator | ||||
, | ||||
const uint8_t *in, size_t in_size, | ||||
uint8_t *out, size_t *out_pos, size_t out_size) | ||||
{ | ||||
return block_buffer_encode(block, allocator, | ||||
in, in_size, out, out_pos, out_size, true); | ||||
} | ||||
extern LZMA_API(lzma_ret) | ||||
lzma_block_uncomp_encode(lzma_block *block, | ||||
const uint8_t *in, size_t in_size, | ||||
uint8_t *out, size_t *out_pos, size_t out_size) | ||||
{ | ||||
// It won't allocate any memory from heap so no need | ||||
// for lzma_allocator. | ||||
return block_buffer_encode(block, NULL, | ||||
in, in_size, out, out_pos, out_size, false); | ||||
} | ||||
End of changes. 19 change blocks. | ||||
33 lines changed or deleted | 42 lines changed or added | |||
This html diff was produced by rfcdiff 1.41. The latest version is available from http://tools.ietf.org/tools/rfcdiff/ |