TheCloud/dwarfs
1
0
Fork 0
mirror of https://github.com/mhx/dwarfs.git synced 2025-09-15 15:26:19 -04:00
Code Issues Packages Projects Releases Wiki Activity

Fix segmenter for the case when segmenting is disable

Browse Source
This commit is contained in:
Marcus Holland-Moritz 2023-09-05 11:14:40 +02:00
parent bad2c5e057
commit b101694d06
1 changed files with 29 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
src/dwarfs/segmenter.cpp
View File

@ -178,6 +178,7 @@ class alignas(64) bloom_filter {
explicit bloom_filter(size_t size) explicit bloom_filter(size_t size)
: index_mask_{(std::max(size, value_mask + 1) >> index_shift) - 1} : index_mask_{(std::max(size, value_mask + 1) >> index_shift) - 1}
, size_{std::max(size, value_mask + 1)} { , size_{std::max(size, value_mask + 1)} {
if (size > 0) {
if (size & (size - 1)) { if (size & (size - 1)) {
throw std::runtime_error("size must be a power of two"); throw std::runtime_error("size must be a power of two");
} }
@ -188,8 +189,13 @@ class alignas(64) bloom_filter {
} }
clear(); clear();
} }
}
~bloom_filter() { boost::alignment::aligned_free(bits_); } ~bloom_filter() {
if (bits_) {
boost::alignment::aligned_free(bits_);
}
}
void add(size_t ix) { void add(size_t ix) {
auto bits = bits_; auto bits = bits_;
@ -225,7 +231,7 @@ class alignas(64) bloom_filter {
bits_type* begin() { return bits_; } bits_type* begin() { return bits_; }
bits_type* end() { return bits_ + (size_ >> index_shift); } bits_type* end() { return bits_ + (size_ >> index_shift); }
bits_type* bits_; bits_type* bits_{nullptr};
size_t const index_mask_; size_t const index_mask_;
size_t const size_; size_t const size_;
}; };
@ -682,12 +688,16 @@ class segmenter_ final : public segmenter::impl, private SegmentingPolicy {
void segment_and_add_data(chunkable& chkable, size_t size_in_frames); void segment_and_add_data(chunkable& chkable, size_t size_in_frames);
size_t bloom_filter_size(const segmenter::config& cfg) const { size_t bloom_filter_size(const segmenter::config& cfg) const {
if constexpr (is_segmentation_enabled()) {
auto hash_count = auto hash_count =
std::bit_ceil(std::max<size_t>(1, cfg.max_active_blocks) * std::bit_ceil(std::max<size_t>(1, cfg.max_active_blocks) *
(block_size_in_frames(cfg) / window_step(cfg))); (block_size_in_frames(cfg) / window_step(cfg)));
return (static_cast<size_t>(1) << cfg.bloom_filter_size) * hash_count; return (static_cast<size_t>(1) << cfg.bloom_filter_size) * hash_count;
} }
return 0;
}
static size_t window_size(const segmenter::config& cfg) { static size_t window_size(const segmenter::config& cfg) {
return cfg.blockhash_window_size > 0 return cfg.blockhash_window_size > 0
? static_cast<size_t>(1) << cfg.blockhash_window_size ? static_cast<size_t>(1) << cfg.blockhash_window_size
@ -972,10 +982,12 @@ void segmenter_<LoggerPolicy, SegmentingPolicy>::append_to_block(
blocks_.pop_front(); blocks_.pop_front();
} }
if constexpr (is_segmentation_enabled()) {
global_filter_.clear(); global_filter_.clear();
for (auto const& b : blocks_) { for (auto const& b : blocks_) {
global_filter_.merge(b.filter()); global_filter_.merge(b.filter());
} }
}
add_new_block(blocks_, LOG_GET_LOGGER, repeating_sequence_hash_values_, add_new_block(blocks_, LOG_GET_LOGGER, repeating_sequence_hash_values_,
repeating_collisions_, blkmgr_->get_logical_block(), repeating_collisions_, blkmgr_->get_logical_block(),