lzma_block_unpadded_size(const lzma_block *block)
{
// Validate the values that we are interested in i.e. all but
// Uncompressed Size and the filters.
//
// NOTE: This function is used for validation too, so it is
// essential that these checks are always done even if
// Compressed Size is unknown.
if (block == NULL || block->version != 0
|| block->header_size < LZMA_BLOCK_HEADER_SIZE_MIN
|| block->header_size > LZMA_BLOCK_HEADER_SIZE_MAX
|| (block->header_size & 3)
|| !lzma_vli_is_valid(block->compressed_size)
|| block->compressed_size == 0
|| (unsigned int)(block->check) > LZMA_CHECK_ID_MAX)
return 0;
// If Compressed Size is unknown, return that we cannot know
// size of the Block either.
if (block->compressed_size == LZMA_VLI_UNKNOWN)
return LZMA_VLI_UNKNOWN;
// Calculate Unpadded Size and validate it.
const lzma_vli unpadded_size = block->compressed_size
+ block->header_size
+ lzma_check_size(block->check);
assert(unpadded_size >= UNPADDED_SIZE_MIN);
if (unpadded_size > UNPADDED_SIZE_MAX)
return 0;
return unpadded_size;
}
/// \brief Parse the Check field and put it into check_value[]
///
/// \return False on success, true on error.
static bool
parse_check_value(file_pair *pair, const lzma_index_iter *iter)
{
// Don't read anything from the file if there is no integrity Check.
if (iter->stream.flags->check == LZMA_CHECK_NONE) {
snprintf(check_value, sizeof(check_value), "---");
return false;
}
// Locate and read the Check field.
const uint32_t size = lzma_check_size(iter->stream.flags->check);
const off_t offset = iter->block.compressed_file_offset
+ iter->block.total_size - size;
io_buf buf;
if (io_pread(pair, &buf, size, offset))
return true;
// CRC32 and CRC64 are in little endian. Guess that all the future
// 32-bit and 64-bit Check values are little endian too. It shouldn't
// be a too big problem if this guess is wrong.
if (size == 4)
snprintf(check_value, sizeof(check_value),
"%08" PRIx32, conv32le(buf.u32[0]));
else if (size == 8)
snprintf(check_value, sizeof(check_value),
"%016" PRIx64, conv64le(buf.u64[0]));
else
for (size_t i = 0; i < size; ++i)
snprintf(check_value + i * 2, 3, "%02x", buf.u8[i]);
return false;
}
lzma_block_compressed_size(lzma_block *block, lzma_vli unpadded_size)
{
// Validate everything but Uncompressed Size and filters.
if (lzma_block_unpadded_size(block) == 0)
return LZMA_PROG_ERROR;
const uint32_t container_size = block->header_size
+ lzma_check_size(block->check);
// Validate that Compressed Size will be greater than zero.
if (unpadded_size <= container_size)
return LZMA_DATA_ERROR;
// Calculate what Compressed Size is supposed to be.
// If Compressed Size was present in Block Header,
// compare that the new value matches it.
const lzma_vli compressed_size = unpadded_size - container_size;
if (block->compressed_size != LZMA_VLI_UNKNOWN
&& block->compressed_size != compressed_size)
return LZMA_DATA_ERROR;
block->compressed_size = compressed_size;
return LZMA_OK;
}
Handle<Value> lzmaCheckSize(const Arguments& args) {
HandleScope scope;
Local<Integer> arg = Local<Integer>::Cast(args[0]);
return scope.Close(Integer::NewFromUnsigned(lzma_check_size((lzma_check) arg->Value())));
}
static bool
print_info_adv(xz_file_info *xfi, file_pair *pair)
{
// Print the overall information.
print_adv_helper(lzma_index_stream_count(xfi->idx),
lzma_index_block_count(xfi->idx),
lzma_index_file_size(xfi->idx),
lzma_index_uncompressed_size(xfi->idx),
lzma_index_checks(xfi->idx),
xfi->stream_padding);
// Size of the biggest Check. This is used to calculate the width
// of the CheckVal field. The table would get insanely wide if
// we always reserved space for 64-byte Check (128 chars as hex).
uint32_t check_max = 0;
// Print information about the Streams.
//
// TRANSLATORS: The second line is column headings. All except
// Check are right aligned; Check is left aligned. Test with
// "xz -lv foo.xz".
puts(_(" Streams:\n Stream Blocks"
" CompOffset UncompOffset"
" CompSize UncompSize Ratio"
" Check Padding"));
lzma_index_iter iter;
lzma_index_iter_init(&iter, xfi->idx);
while (!lzma_index_iter_next(&iter, LZMA_INDEX_ITER_STREAM)) {
const char *cols1[4] = {
uint64_to_str(iter.stream.number, 0),
uint64_to_str(iter.stream.block_count, 1),
uint64_to_str(iter.stream.compressed_offset, 2),
uint64_to_str(iter.stream.uncompressed_offset, 3),
};
printf(" %*s %*s %*s %*s ",
tuklib_mbstr_fw(cols1[0], 6), cols1[0],
tuklib_mbstr_fw(cols1[1], 9), cols1[1],
tuklib_mbstr_fw(cols1[2], 15), cols1[2],
tuklib_mbstr_fw(cols1[3], 15), cols1[3]);
const char *cols2[5] = {
uint64_to_str(iter.stream.compressed_size, 0),
uint64_to_str(iter.stream.uncompressed_size, 1),
get_ratio(iter.stream.compressed_size,
iter.stream.uncompressed_size),
_(check_names[iter.stream.flags->check]),
uint64_to_str(iter.stream.padding, 2),
};
printf("%*s %*s %*s %-*s %*s\n",
tuklib_mbstr_fw(cols2[0], 15), cols2[0],
tuklib_mbstr_fw(cols2[1], 15), cols2[1],
tuklib_mbstr_fw(cols2[2], 5), cols2[2],
tuklib_mbstr_fw(cols2[3], 10), cols2[3],
tuklib_mbstr_fw(cols2[4], 7), cols2[4]);
// Update the maximum Check size.
if (lzma_check_size(iter.stream.flags->check) > check_max)
check_max = lzma_check_size(iter.stream.flags->check);
}
// Cache the verbosity level to a local variable.
const bool detailed = message_verbosity_get() >= V_DEBUG;
// Information collected from Block Headers
block_header_info bhi;
// Print information about the Blocks but only if there is
// at least one Block.
if (lzma_index_block_count(xfi->idx) > 0) {
// Calculate the width of the CheckVal field.
const int checkval_width = my_max(8, 2 * check_max);
// TRANSLATORS: The second line is column headings. All
// except Check are right aligned; Check is left aligned.
printf(_(" Blocks:\n Stream Block"
" CompOffset UncompOffset"
" TotalSize UncompSize Ratio Check"));
if (detailed) {
// TRANSLATORS: These are additional column headings
// for the most verbose listing mode. CheckVal
// (Check value), Flags, and Filters are left aligned.
// Header (Block Header Size), CompSize, and MemUsage
// are right aligned. %*s is replaced with 0-120
// spaces to make the CheckVal column wide enough.
// Test with "xz -lvv foo.xz".
printf(_(" CheckVal %*s Header Flags "
"CompSize MemUsage Filters"),
checkval_width - 8, "");
}
putchar('\n');
lzma_index_iter_init(&iter, xfi->idx);
// Iterate over the Blocks.
while (!lzma_index_iter_next(&iter, LZMA_INDEX_ITER_BLOCK)) {
if (detailed && parse_details(pair, &iter, &bhi, xfi))
return true;
const char *cols1[4] = {
uint64_to_str(iter.stream.number, 0),
uint64_to_str(
iter.block.number_in_stream, 1),
uint64_to_str(
iter.block.compressed_file_offset, 2),
uint64_to_str(
iter.block.uncompressed_file_offset, 3)
};
printf(" %*s %*s %*s %*s ",
tuklib_mbstr_fw(cols1[0], 6), cols1[0],
tuklib_mbstr_fw(cols1[1], 9), cols1[1],
tuklib_mbstr_fw(cols1[2], 15), cols1[2],
tuklib_mbstr_fw(cols1[3], 15), cols1[3]);
const char *cols2[4] = {
uint64_to_str(iter.block.total_size, 0),
uint64_to_str(iter.block.uncompressed_size,
1),
get_ratio(iter.block.total_size,
iter.block.uncompressed_size),
_(check_names[iter.stream.flags->check])
};
printf("%*s %*s %*s %-*s",
tuklib_mbstr_fw(cols2[0], 15), cols2[0],
tuklib_mbstr_fw(cols2[1], 15), cols2[1],
tuklib_mbstr_fw(cols2[2], 5), cols2[2],
tuklib_mbstr_fw(cols2[3], detailed ? 11 : 1),
cols2[3]);
if (detailed) {
const lzma_vli compressed_size
= iter.block.unpadded_size
- bhi.header_size
- lzma_check_size(
iter.stream.flags->check);
const char *cols3[6] = {
check_value,
uint64_to_str(bhi.header_size, 0),
bhi.flags,
uint64_to_str(compressed_size, 1),
uint64_to_str(
round_up_to_mib(bhi.memusage),
2),
bhi.filter_chain
};
// Show MiB for memory usage, because it
// is the only size which is not in bytes.
printf("%-*s %*s %-5s %*s %*s MiB %s",
checkval_width, cols3[0],
tuklib_mbstr_fw(cols3[1], 6), cols3[1],
cols3[2],
tuklib_mbstr_fw(cols3[3], 15),
cols3[3],
tuklib_mbstr_fw(cols3[4], 7), cols3[4],
cols3[5]);
}
putchar('\n');
}
}
if (detailed) {
printf(_(" Memory needed: %s MiB\n"), uint64_to_str(
round_up_to_mib(xfi->memusage_max), 0));
printf(_(" Sizes in headers: %s\n"),
xfi->all_have_sizes ? _("Yes") : _("No"));
printf(_(" Minimum XZ Utils version: %s\n"),
xz_ver_to_str(xfi->min_version));
}
return false;
}
/// \brief Parse the Block Header
///
/// The result is stored into *bhi. The caller takes care of initializing it.
///
/// \return False on success, true on error.
static bool
parse_block_header(file_pair *pair, const lzma_index_iter *iter,
block_header_info *bhi, xz_file_info *xfi)
{
#if IO_BUFFER_SIZE < LZMA_BLOCK_HEADER_SIZE_MAX
# error IO_BUFFER_SIZE < LZMA_BLOCK_HEADER_SIZE_MAX
#endif
// Get the whole Block Header with one read, but don't read past
// the end of the Block (or even its Check field).
const uint32_t size = my_min(iter->block.total_size
- lzma_check_size(iter->stream.flags->check),
LZMA_BLOCK_HEADER_SIZE_MAX);
io_buf buf;
if (io_pread(pair, &buf, size, iter->block.compressed_file_offset))
return true;
// Zero would mean Index Indicator and thus not a valid Block.
if (buf.u8[0] == 0)
goto data_error;
// Initialize the block structure and decode Block Header Size.
lzma_filter filters[LZMA_FILTERS_MAX + 1];
lzma_block block;
block.version = 0;
block.check = iter->stream.flags->check;
block.filters = filters;
block.header_size = lzma_block_header_size_decode(buf.u8[0]);
if (block.header_size > size)
goto data_error;
// Decode the Block Header.
switch (lzma_block_header_decode(&block, NULL, buf.u8)) {
case LZMA_OK:
break;
case LZMA_OPTIONS_ERROR:
message_error("%s: %s", pair->src_name,
message_strm(LZMA_OPTIONS_ERROR));
return true;
case LZMA_DATA_ERROR:
goto data_error;
default:
message_bug();
}
// Check the Block Flags. These must be done before calling
// lzma_block_compressed_size(), because it overwrites
// block.compressed_size.
bhi->flags[0] = block.compressed_size != LZMA_VLI_UNKNOWN
? 'c' : '-';
bhi->flags[1] = block.uncompressed_size != LZMA_VLI_UNKNOWN
? 'u' : '-';
bhi->flags[2] = '\0';
// Collect information if all Blocks have both Compressed Size
// and Uncompressed Size fields. They can be useful e.g. for
// multi-threaded decompression so it can be useful to know it.
xfi->all_have_sizes &= block.compressed_size != LZMA_VLI_UNKNOWN
&& block.uncompressed_size != LZMA_VLI_UNKNOWN;
// Validate or set block.compressed_size.
switch (lzma_block_compressed_size(&block,
iter->block.unpadded_size)) {
case LZMA_OK:
// Validate also block.uncompressed_size if it is present.
// If it isn't present, there's no need to set it since
// we aren't going to actually decompress the Block; if
// we were decompressing, then we should set it so that
// the Block decoder could validate the Uncompressed Size
// that was stored in the Index.
if (block.uncompressed_size == LZMA_VLI_UNKNOWN
|| block.uncompressed_size
== iter->block.uncompressed_size)
break;
// If the above fails, the file is corrupt so
// LZMA_DATA_ERROR is a good error code.
// Fall through
case LZMA_DATA_ERROR:
// Free the memory allocated by lzma_block_header_decode().
for (size_t i = 0; filters[i].id != LZMA_VLI_UNKNOWN; ++i)
free(filters[i].options);
goto data_error;
default:
message_bug();
}
// Copy the known sizes.
bhi->header_size = block.header_size;
bhi->compressed_size = block.compressed_size;
// Calculate the decoder memory usage and update the maximum
// memory usage of this Block.
bhi->memusage = lzma_raw_decoder_memusage(filters);
if (xfi->memusage_max < bhi->memusage)
xfi->memusage_max = bhi->memusage;
// Determine the minimum XZ Utils version that supports this Block.
//
// Currently the only thing that 5.0.0 doesn't support is empty
// LZMA2 Block. This decoder bug was fixed in 5.0.2.
{
size_t i = 0;
while (filters[i + 1].id != LZMA_VLI_UNKNOWN)
++i;
if (filters[i].id == LZMA_FILTER_LZMA2
&& iter->block.uncompressed_size == 0
&& xfi->min_version < 50000022U)
xfi->min_version = 50000022U;
}
// Convert the filter chain to human readable form.
message_filters_to_str(bhi->filter_chain, filters, false);
// Free the memory allocated by lzma_block_header_decode().
for (size_t i = 0; filters[i].id != LZMA_VLI_UNKNOWN; ++i)
free(filters[i].options);
return false;
data_error:
// Show the error message.
message_error("%s: %s", pair->src_name,
message_strm(LZMA_DATA_ERROR));
return true;
}
lzma_block_buffer_encode(lzma_block *block, lzma_allocator *allocator,
const uint8_t *in, size_t in_size,
uint8_t *out, size_t *out_pos, size_t out_size)
{
size_t check_size;
lzma_ret ret;
size_t i;
// Validate the arguments.
if (block == NULL || (in == NULL && in_size != 0) || out == NULL
|| out_pos == NULL || *out_pos > out_size)
return LZMA_PROG_ERROR;
// The contents of the structure may depend on the version so
// check the version before validating the contents of *block.
if (block->version != 0)
return LZMA_OPTIONS_ERROR;
if ((unsigned int)(block->check) > LZMA_CHECK_ID_MAX
|| block->filters == NULL)
return LZMA_PROG_ERROR;
if (!lzma_check_is_supported(block->check))
return LZMA_UNSUPPORTED_CHECK;
// 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 adding
// Block Padding.
out_size -= (out_size - *out_pos) & 3;
// Get the size of the Check field.
check_size = lzma_check_size(block->check);
assert(check_size != UINT32_MAX);
// Reserve space for the Check field.
if (out_size - *out_pos <= check_size)
return LZMA_BUF_ERROR;
out_size -= check_size;
// Do the actual compression.
ret = block_encode_normal(block, allocator,
in, in_size, out, out_pos, out_size);
if (ret != LZMA_OK) {
// If the error was something else than output buffer
// becoming full, return the error now.
if (ret != LZMA_BUF_ERROR)
return ret;
// The data was uncompressible (at least with the options
// given to us) or the output buffer was too small. Use the
// uncompressed chunks of LZMA2 to wrap the data into a valid
// Block. If we haven't been given enough output space, even
// this may fail.
return_if_error(block_encode_uncompressed(block, in, in_size,
out, out_pos, out_size));
}
assert(*out_pos <= out_size);
// Block Padding. No buffer overflow here, because we already adjusted
// out_size so that (out_size - out_start) is a multiple of four.
// Thus, if the buffer is full, the loop body can never run.
for (i = (size_t)(block->compressed_size); i & 3; ++i) {
assert(*out_pos < out_size);
out[(*out_pos)++] = 0x00;
}
// If there's no Check field, we are done now.
if (check_size > 0) {
// Calculate the integrity check. We reserved space for
// the Check field earlier so we don't need to check for
// available output space here.
lzma_check_state check;
lzma_check_init(&check, block->check);
lzma_check_update(&check, block->check, in, in_size);
lzma_check_finish(&check, block->check);
memcpy(block->raw_check, check.buffer.u8, check_size);
memcpy(out + *out_pos, check.buffer.u8, check_size);
*out_pos += check_size;
}
return LZMA_OK;
}
/// \brief Parse the Block Header
///
/// The result is stored into *bhi. The caller takes care of initializing it.
///
/// \return False on success, true on error.
static bool
parse_block_header(file_pair *pair, const lzma_index_iter *iter,
block_header_info *bhi, xz_file_info *xfi)
{
#if IO_BUFFER_SIZE < LZMA_BLOCK_HEADER_SIZE_MAX
# error IO_BUFFER_SIZE < LZMA_BLOCK_HEADER_SIZE_MAX
#endif
// Get the whole Block Header with one read, but don't read past
// the end of the Block (or even its Check field).
const uint32_t size = my_min(iter->block.total_size
- lzma_check_size(iter->stream.flags->check),
LZMA_BLOCK_HEADER_SIZE_MAX);
io_buf buf;
if (io_pread(pair, &buf, size, iter->block.compressed_file_offset))
return true;
// Zero would mean Index Indicator and thus not a valid Block.
if (buf.u8[0] == 0)
goto data_error;
lzma_block block;
lzma_filter filters[LZMA_FILTERS_MAX + 1];
// Initialize the pointers so that they can be passed to free().
for (size_t i = 0; i < ARRAY_SIZE(filters); ++i)
filters[i].options = NULL;
// Initialize the block structure and decode Block Header Size.
block.version = 0;
block.check = iter->stream.flags->check;
block.filters = filters;
block.header_size = lzma_block_header_size_decode(buf.u8[0]);
if (block.header_size > size)
goto data_error;
// Decode the Block Header.
switch (lzma_block_header_decode(&block, NULL, buf.u8)) {
case LZMA_OK:
break;
case LZMA_OPTIONS_ERROR:
message_error("%s: %s", pair->src_name,
message_strm(LZMA_OPTIONS_ERROR));
return true;
case LZMA_DATA_ERROR:
goto data_error;
default:
message_bug();
}
// Check the Block Flags. These must be done before calling
// lzma_block_compressed_size(), because it overwrites
// block.compressed_size.
bhi->flags[0] = block.compressed_size != LZMA_VLI_UNKNOWN
? 'c' : '-';
bhi->flags[1] = block.uncompressed_size != LZMA_VLI_UNKNOWN
? 'u' : '-';
bhi->flags[2] = '\0';
// Collect information if all Blocks have both Compressed Size
// and Uncompressed Size fields. They can be useful e.g. for
// multi-threaded decompression so it can be useful to know it.
xfi->all_have_sizes &= block.compressed_size != LZMA_VLI_UNKNOWN
&& block.uncompressed_size != LZMA_VLI_UNKNOWN;
// Validate or set block.compressed_size.
switch (lzma_block_compressed_size(&block,
iter->block.unpadded_size)) {
case LZMA_OK:
break;
case LZMA_DATA_ERROR:
goto data_error;
default:
message_bug();
}
// Copy the known sizes.
bhi->header_size = block.header_size;
bhi->compressed_size = block.compressed_size;
// Calculate the decoder memory usage and update the maximum
// memory usage of this Block.
bhi->memusage = lzma_raw_decoder_memusage(filters);
if (xfi->memusage_max < bhi->memusage)
xfi->memusage_max = bhi->memusage;
// Convert the filter chain to human readable form.
message_filters_to_str(bhi->filter_chain, filters, false);
// Free the memory allocated by lzma_block_header_decode().
for (size_t i = 0; filters[i].id != LZMA_VLI_UNKNOWN; ++i)
free(filters[i].options);
return false;
data_error:
// Show the error message.
message_error("%s: %s", pair->src_name,
message_strm(LZMA_DATA_ERROR));
// Free the memory allocated by lzma_block_header_decode().
// This is truly needed only if we get here after a succcessful
// call to lzma_block_header_decode() but it doesn't hurt to
// always do it.
for (size_t i = 0; filters[i].id != LZMA_VLI_UNKNOWN; ++i)
free(filters[i].options);
return true;
}
