Merge pull request #7 from tukaani-project/tuktest_index_hash

Tuktest index hash
This commit is contained in:
Jia Tan 2023-01-07 00:10:50 +08:00 committed by GitHub
commit 6fd39664de
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
13 changed files with 414 additions and 8 deletions

1
.gitignore vendored
View File

@ -62,6 +62,7 @@ coverage
/tests/test_filter_flags
/tests/test_hardware
/tests/test_index
/tests/test_index_hash
/test/test_lzip_decoder
/tests/test_memlimit
/tests/test_stream_flags

View File

@ -801,6 +801,7 @@ if(BUILD_TESTING)
test_filter_flags
test_hardware
test_index
test_index_hash
test_memlimit
test_stream_flags
test_vli
@ -812,6 +813,7 @@ if(BUILD_TESTING)
target_include_directories("${TEST}" PRIVATE
src/common
src/liblzma/api
src/liblzma
lib
)

View File

@ -10,6 +10,7 @@
//
///////////////////////////////////////////////////////////////////////////////
#include "common.h"
#include "index.h"
#include "stream_flags_common.h"

View File

@ -2,6 +2,13 @@
//
/// \file index.h
/// \brief Handling of Index
/// \note This header file does not include common.h or lzma.h because
/// this file is needed by both liblzma internally and by the
/// tests. Including common.h will include and define many things
/// the tests do not need and prevents issues with header file
/// include order. This way, if lzma.h or common.h are not
/// included before this file it will break on every OS instead
/// of causing more subtle errors.
//
// Author: Lasse Collin
//
@ -13,8 +20,6 @@
#ifndef LZMA_INDEX_H
#define LZMA_INDEX_H
#include "common.h"
/// Minimum Unpadded Size
#define UNPADDED_SIZE_MIN LZMA_VLI_C(5)
@ -22,6 +27,9 @@
/// Maximum Unpadded Size
#define UNPADDED_SIZE_MAX (LZMA_VLI_MAX & ~LZMA_VLI_C(3))
/// Index Indicator based on xz specification
#define INDEX_INDICATOR 0
/// Get the size of the Index Padding field. This is needed by Index encoder
/// and decoder, but applications should have no use for this.

View File

@ -80,7 +80,7 @@ index_decode(void *coder_ptr, const lzma_allocator *allocator,
// format". One could argue that the application should
// verify the Index Indicator before trying to decode the
// Index, but well, I suppose it is simpler this way.
if (in[(*in_pos)++] != 0x00)
if (in[(*in_pos)++] != INDEX_INDICATOR)
return LZMA_DATA_ERROR;
coder->sequence = SEQ_COUNT;

View File

@ -13,6 +13,7 @@
#ifndef LZMA_INDEX_DECODER_H
#define LZMA_INDEX_DECODER_H
#include "common.h"
#include "index.h"

View File

@ -65,7 +65,7 @@ index_encode(void *coder_ptr,
while (*out_pos < out_size)
switch (coder->sequence) {
case SEQ_INDICATOR:
out[*out_pos] = 0x00;
out[*out_pos] = INDEX_INDICATOR;
++*out_pos;
coder->sequence = SEQ_COUNT;
break;

View File

@ -145,7 +145,7 @@ lzma_index_hash_append(lzma_index_hash *index_hash, lzma_vli unpadded_size,
lzma_vli uncompressed_size)
{
// Validate the arguments.
if (index_hash->sequence != SEQ_BLOCK
if (index_hash == NULL || index_hash->sequence != SEQ_BLOCK
|| unpadded_size < UNPADDED_SIZE_MIN
|| unpadded_size > UNPADDED_SIZE_MAX
|| uncompressed_size > LZMA_VLI_MAX)
@ -190,7 +190,7 @@ lzma_index_hash_decode(lzma_index_hash *index_hash, const uint8_t *in,
switch (index_hash->sequence) {
case SEQ_BLOCK:
// Check the Index Indicator is present.
if (in[(*in_pos)++] != 0x00)
if (in[(*in_pos)++] != INDEX_INDICATOR)
return LZMA_DATA_ERROR;
index_hash->sequence = SEQ_COUNT;

View File

@ -10,6 +10,7 @@
//
///////////////////////////////////////////////////////////////////////////////
#include "common.h"
#include "index.h"

View File

@ -12,6 +12,7 @@
#include "stream_decoder.h"
#include "block_decoder.h"
#include "index.h"
typedef struct {
@ -164,7 +165,7 @@ stream_decode(void *coder_ptr, const lzma_allocator *allocator,
if (coder->pos == 0) {
// Detect if it's Index.
if (in[*in_pos] == 0x00) {
if (in[*in_pos] == INDEX_INDICATOR) {
coder->sequence = SEQ_INDEX;
break;
}

View File

@ -887,7 +887,7 @@ decode_block_header(struct lzma_stream_coder *coder,
if (coder->pos == 0) {
// Detect if it's Index.
if (in[*in_pos] == 0x00)
if (in[*in_pos] == INDEX_INDICATOR)
return LZMA_INDEX_DETECTED;
// Calculate the size of the Block Header. Note that

View File

@ -26,6 +26,7 @@ EXTRA_DIST = \
AM_CPPFLAGS = \
-I$(top_srcdir)/src/common \
-I$(top_srcdir)/src/liblzma/api \
-I$(top_srcdir)/src/liblzma \
-I$(top_builddir)/lib
LDADD = $(top_builddir)/src/liblzma/liblzma.la
@ -44,6 +45,7 @@ check_PROGRAMS = \
test_filter_flags \
test_block_header \
test_index \
test_index_hash \
test_bcj_exact_size \
test_memlimit \
test_lzip_decoder \
@ -56,6 +58,7 @@ TESTS = \
test_filter_flags \
test_block_header \
test_index \
test_index_hash \
test_bcj_exact_size \
test_memlimit \
test_lzip_decoder \

388
tests/test_index_hash.c Normal file
View File

@ -0,0 +1,388 @@
///////////////////////////////////////////////////////////////////////////////
//
/// \file test_index_hash.c
/// \brief Tests src/liblzma/common/index_hash.c API functions
///
/// \note No test included for lzma_index_hash_end since it
/// would be trivial unless tested for memory leaks
/// with something like valgrind
//
// Author: Jia Tan
//
// This file has been put into the public domain.
// You can do whatever you want with this file.
//
///////////////////////////////////////////////////////////////////////////////
#include "tests.h"
// Needed for UNPADDED_SIZE_MIN and UNPADDED_SIZE_MAX macro definitions
// and index_size and vli_ceil4 helper functions
#include "common/index.h"
static void
test_lzma_index_hash_init(void)
{
#ifndef HAVE_DECODERS
assert_skip("Decoder support disabled");
#else
// First test with NULL index_hash.
// This should create a fresh index_hash.
lzma_index_hash *index_hash = lzma_index_hash_init(NULL, NULL);
assert_true(index_hash != NULL);
// Next test with non-NULL index_hash.
lzma_index_hash *second_hash = lzma_index_hash_init(index_hash, NULL);
// It should not create a new index_hash pointer.
// Instead it must just re-init the first index_hash.
assert_true(index_hash == second_hash);
lzma_index_hash_end(index_hash, NULL);
#endif
}
static void
test_lzma_index_hash_append(void)
{
#ifndef HAVE_DECODERS
assert_skip("Decoder support disabled");
#else
// Test all invalid parameters
assert_lzma_ret(lzma_index_hash_append(NULL, 0, 0),
LZMA_PROG_ERROR);
// Test NULL index_hash
assert_lzma_ret(lzma_index_hash_append(NULL, UNPADDED_SIZE_MIN,
LZMA_VLI_MAX), LZMA_PROG_ERROR);
// Test with invalid Unpadded Size
lzma_index_hash *index_hash = lzma_index_hash_init(NULL, NULL);
assert_true(index_hash != NULL);
assert_lzma_ret(lzma_index_hash_append(index_hash,
UNPADDED_SIZE_MIN - 1, LZMA_VLI_MAX),
LZMA_PROG_ERROR);
// Test with invalid Uncompressed Size
assert_lzma_ret(lzma_index_hash_append(index_hash,
UNPADDED_SIZE_MIN, LZMA_VLI_MAX + 1),
LZMA_PROG_ERROR);
// First append a Record describing a small Block.
// This should succeed.
assert_lzma_ret(lzma_index_hash_append(index_hash,
UNPADDED_SIZE_MIN, 1), LZMA_OK);
// Append another small Record.
assert_lzma_ret(lzma_index_hash_append(index_hash,
UNPADDED_SIZE_MIN, 1), LZMA_OK);
// Append a Record that would cause the compressed size to grow
// too big
assert_lzma_ret(lzma_index_hash_append(index_hash,
UNPADDED_SIZE_MAX, 1), LZMA_DATA_ERROR);
lzma_index_hash_end(index_hash, NULL);
#endif
}
#ifdef HAVE_DECODERS
// Fill an index_hash with unpadded and uncompressed VLIs
// by calling lzma_index_hash_append
static void
fill_index_hash(lzma_index_hash *index_hash, const lzma_vli *unpadded_sizes,
const lzma_vli *uncomp_sizes, uint32_t block_count)
{
for (uint32_t i = 0; i < block_count; ++i)
assert_lzma_ret(lzma_index_hash_append(index_hash,
unpadded_sizes[i], uncomp_sizes[i]), LZMA_OK);
}
#ifdef HAVE_ENCODERS
// Set the contents of buf to the expected Index based on the
// .xz specification. This needs the unpadded and uncompressed VLIs
// to correctly create the Index.
static void
generate_index(uint8_t *buf, const lzma_vli *unpadded_sizes,
const lzma_vli *uncomp_sizes, uint32_t block_count,
size_t index_max_size)
{
size_t in_pos = 0;
size_t out_pos = 0;
// First set Index Indicator
buf[out_pos++] = INDEX_INDICATOR;
// Next write out Number of Records
assert_lzma_ret(lzma_vli_encode(block_count, &in_pos, buf,
&out_pos, index_max_size), LZMA_STREAM_END);
// Next write out each Record.
// A Record consists of Unpadded Size and Uncompressed Size
// written next to each other as VLIs.
for (uint32_t i = 0; i < block_count; ++i) {
in_pos = 0;
assert_lzma_ret(lzma_vli_encode(unpadded_sizes[i], &in_pos,
buf, &out_pos, index_max_size), LZMA_STREAM_END);
in_pos = 0;
assert_lzma_ret(lzma_vli_encode(uncomp_sizes[i], &in_pos,
buf, &out_pos, index_max_size), LZMA_STREAM_END);
}
// Add Index Padding
lzma_vli rounded_out_pos = vli_ceil4(out_pos);
memzero(buf + out_pos, rounded_out_pos - out_pos);
out_pos = rounded_out_pos;
// Add the CRC32
write32le(buf + out_pos, lzma_crc32(buf, out_pos, 0));
out_pos += 4;
assert_uint_eq(out_pos, index_max_size);
}
#endif
#endif
static void
test_lzma_index_hash_decode(void)
{
#if !defined(HAVE_ENCODERS) || !defined(HAVE_DECODERS)
assert_skip("Encoder or decoder support disabled");
#else
lzma_index_hash *index_hash = lzma_index_hash_init(NULL, NULL);
assert_true(index_hash != NULL);
size_t in_pos = 0;
// Six valid values for the Unpadded Size fields in an Index
const lzma_vli unpadded_sizes[6] = {
UNPADDED_SIZE_MIN,
1000,
4000,
8000,
16000,
32000
};
// Six valid values for the Uncompressed Size fields in an Index
const lzma_vli uncomp_sizes[6] = {
1,
500,
8000,
20,
1,
500
};
// Add two Records to an index_hash
fill_index_hash(index_hash, unpadded_sizes, uncomp_sizes, 2);
const lzma_vli size_two_records = lzma_index_hash_size(index_hash);
assert_uint(size_two_records, >, 0);
uint8_t *index_two_records = tuktest_malloc(size_two_records);
generate_index(index_two_records, unpadded_sizes, uncomp_sizes, 2,
size_two_records);
// First test for basic buffer size error
in_pos = size_two_records + 1;
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_two_records, &in_pos,
size_two_records), LZMA_BUF_ERROR);
// Next test for invalid Index Indicator
in_pos = 0;
index_two_records[0] ^= 1;
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_two_records, &in_pos,
size_two_records), LZMA_DATA_ERROR);
index_two_records[0] ^= 1;
// Next verify the index_hash as expected
in_pos = 0;
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_two_records, &in_pos,
size_two_records), LZMA_STREAM_END);
// Next test an index_hash with three Records
index_hash = lzma_index_hash_init(index_hash, NULL);
fill_index_hash(index_hash, unpadded_sizes, uncomp_sizes, 3);
const lzma_vli size_three_records = lzma_index_hash_size(
index_hash);
assert_uint(size_three_records, >, 0);
uint8_t *index_three_records = tuktest_malloc(size_three_records);
generate_index(index_three_records, unpadded_sizes, uncomp_sizes,
3, size_three_records);
in_pos = 0;
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_three_records, &in_pos,
size_three_records), LZMA_STREAM_END);
// Next test an index_hash with five Records
index_hash = lzma_index_hash_init(index_hash, NULL);
fill_index_hash(index_hash, unpadded_sizes, uncomp_sizes, 5);
const lzma_vli size_five_records = lzma_index_hash_size(
index_hash);
assert_uint(size_five_records, >, 0);
uint8_t *index_five_records = tuktest_malloc(size_five_records);
generate_index(index_five_records, unpadded_sizes, uncomp_sizes, 5,
size_five_records);
// Instead of testing all input at once, give input
// one byte at a time
in_pos = 0;
for (lzma_vli i = 0; i < size_five_records - 1; ++i) {
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_five_records, &in_pos, in_pos + 1),
LZMA_OK);
}
// Last byte should return LZMA_STREAM_END
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_five_records, &in_pos,
in_pos + 1), LZMA_STREAM_END);
// Next test if the index_hash is given an incorrect Unpadded
// Size. Should detect and report LZMA_DATA_ERROR
index_hash = lzma_index_hash_init(index_hash, NULL);
fill_index_hash(index_hash, unpadded_sizes, uncomp_sizes, 5);
// The sixth Record will have an invalid Unpadded Size
assert_lzma_ret(lzma_index_hash_append(index_hash,
unpadded_sizes[5] + 1,
uncomp_sizes[5]), LZMA_OK);
const lzma_vli size_six_records = lzma_index_hash_size(
index_hash);
assert_uint(size_six_records, >, 0);
uint8_t *index_six_records = tuktest_malloc(size_six_records);
generate_index(index_six_records, unpadded_sizes, uncomp_sizes, 6,
size_six_records);
in_pos = 0;
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_six_records, &in_pos,
size_six_records), LZMA_DATA_ERROR);
// Next test if the Index is corrupt (invalid CRC32).
// Should detect and report LZMA_DATA_ERROR
index_hash = lzma_index_hash_init(index_hash, NULL);
fill_index_hash(index_hash, unpadded_sizes, uncomp_sizes, 2);
index_two_records[size_two_records - 1] ^= 1;
in_pos = 0;
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_two_records, &in_pos,
size_two_records), LZMA_DATA_ERROR);
// Next test with Index and index_hash struct not matching
// a Record
index_hash = lzma_index_hash_init(index_hash, NULL);
fill_index_hash(index_hash, unpadded_sizes, uncomp_sizes, 2);
// Recalculate Index with invalid Unpadded Size
const lzma_vli unpadded_sizes_invalid[2] = {
unpadded_sizes[0],
unpadded_sizes[1] + 1
};
generate_index(index_two_records, unpadded_sizes_invalid,
uncomp_sizes, 2, size_two_records);
in_pos = 0;
assert_lzma_ret(lzma_index_hash_decode(index_hash,
index_two_records, &in_pos,
size_two_records), LZMA_DATA_ERROR);
lzma_index_hash_end(index_hash, NULL);
#endif
}
static void
test_lzma_index_hash_size(void)
{
#ifndef HAVE_DECODERS
assert_skip("Decoder support disabled");
#else
lzma_index_hash *index_hash = lzma_index_hash_init(NULL, NULL);
assert_true(index_hash != NULL);
// First test empty index_hash
// Expected size should be:
// Index Indicator - 1 byte
// Number of Records - 1 byte
// List of Records - 0 bytes
// Index Padding - 2 bytes
// CRC32 - 4 bytes
// Total - 8 bytes
assert_uint_eq(lzma_index_hash_size(index_hash), 8);
// Append a Record describing a small Block to the index_hash
assert_lzma_ret(lzma_index_hash_append(index_hash,
UNPADDED_SIZE_MIN, 1), LZMA_OK);
// Expected size should be:
// Index Indicator - 1 byte
// Number of Records - 1 byte
// List of Records - 2 bytes
// Index Padding - 0 bytes
// CRC32 - 4 bytes
// Total - 8 bytes
lzma_vli expected_size = 8;
assert_uint_eq(lzma_index_hash_size(index_hash), expected_size);
// Append additional small Record
assert_lzma_ret(lzma_index_hash_append(index_hash,
UNPADDED_SIZE_MIN, 1), LZMA_OK);
// Expected size should be:
// Index Indicator - 1 byte
// Number of Records - 1 byte
// List of Records - 4 bytes
// Index Padding - 2 bytes
// CRC32 - 4 bytes
// Total - 12 bytes
expected_size = 12;
assert_uint_eq(lzma_index_hash_size(index_hash), expected_size);
// Append a larger Record to the index_hash (3 bytes for each VLI)
const lzma_vli three_byte_vli = 0x10000;
assert_lzma_ret(lzma_index_hash_append(index_hash,
three_byte_vli, three_byte_vli), LZMA_OK);
// Expected size should be:
// Index Indicator - 1 byte
// Number of Records - 1 byte
// List of Records - 10 bytes
// Index Padding - 0 bytes
// CRC32 - 4 bytes
// Total - 16 bytes
expected_size = 16;
assert_uint_eq(lzma_index_hash_size(index_hash), expected_size);
lzma_index_hash_end(index_hash, NULL);
#endif
}
extern int
main(int argc, char **argv)
{
tuktest_start(argc, argv);
tuktest_run(test_lzma_index_hash_init);
tuktest_run(test_lzma_index_hash_append);
tuktest_run(test_lzma_index_hash_decode);
tuktest_run(test_lzma_index_hash_size);
return tuktest_end();
}