@nxtedition/rocksdb 13.1.5 → 13.3.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (237) hide show
  1. package/binding.cc +62 -15
  2. package/deps/rocksdb/rocksdb/{TARGETS → BUCK} +27 -0
  3. package/deps/rocksdb/rocksdb/CMakeLists.txt +3 -1
  4. package/deps/rocksdb/rocksdb/Makefile +2 -2
  5. package/deps/rocksdb/rocksdb/cache/cache.cc +3 -1
  6. package/deps/rocksdb/rocksdb/db/arena_wrapped_db_iter.h +2 -0
  7. package/deps/rocksdb/rocksdb/db/attribute_group_iterator_impl.h +34 -9
  8. package/deps/rocksdb/rocksdb/db/blob/blob_source.cc +7 -6
  9. package/deps/rocksdb/rocksdb/db/blob/blob_source.h +5 -1
  10. package/deps/rocksdb/rocksdb/db/blob/blob_source_test.cc +22 -14
  11. package/deps/rocksdb/rocksdb/db/blob/db_blob_basic_test.cc +149 -0
  12. package/deps/rocksdb/rocksdb/db/builder.cc +13 -24
  13. package/deps/rocksdb/rocksdb/db/coalescing_iterator.h +35 -10
  14. package/deps/rocksdb/rocksdb/db/column_family.cc +21 -10
  15. package/deps/rocksdb/rocksdb/db/column_family.h +15 -8
  16. package/deps/rocksdb/rocksdb/db/column_family_test.cc +98 -7
  17. package/deps/rocksdb/rocksdb/db/compaction/compaction.cc +126 -16
  18. package/deps/rocksdb/rocksdb/db/compaction/compaction.h +51 -5
  19. package/deps/rocksdb/rocksdb/db/compaction/compaction_iterator.cc +2 -2
  20. package/deps/rocksdb/rocksdb/db/compaction/compaction_iterator.h +2 -8
  21. package/deps/rocksdb/rocksdb/db/compaction/compaction_iterator_test.cc +24 -0
  22. package/deps/rocksdb/rocksdb/db/compaction/compaction_job.cc +52 -22
  23. package/deps/rocksdb/rocksdb/db/compaction/compaction_job.h +9 -7
  24. package/deps/rocksdb/rocksdb/db/compaction/compaction_job_test.cc +36 -9
  25. package/deps/rocksdb/rocksdb/db/compaction/compaction_outputs.h +6 -0
  26. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker.cc +30 -17
  27. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker.h +26 -23
  28. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker_fifo.cc +43 -33
  29. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker_fifo.h +6 -5
  30. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker_level.cc +19 -9
  31. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker_level.h +6 -5
  32. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker_test.cc +632 -411
  33. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker_universal.cc +171 -51
  34. package/deps/rocksdb/rocksdb/db/compaction/compaction_picker_universal.h +7 -5
  35. package/deps/rocksdb/rocksdb/db/compaction/compaction_service_job.cc +37 -10
  36. package/deps/rocksdb/rocksdb/db/compaction/compaction_service_test.cc +51 -11
  37. package/deps/rocksdb/rocksdb/db/compaction/subcompaction_state.cc +10 -3
  38. package/deps/rocksdb/rocksdb/db/compaction/tiered_compaction_test.cc +350 -154
  39. package/deps/rocksdb/rocksdb/db/convenience.cc +1 -1
  40. package/deps/rocksdb/rocksdb/db/db_block_cache_test.cc +62 -27
  41. package/deps/rocksdb/rocksdb/db/db_bloom_filter_test.cc +68 -1
  42. package/deps/rocksdb/rocksdb/db/db_compaction_test.cc +91 -0
  43. package/deps/rocksdb/rocksdb/db/db_impl/db_impl.cc +134 -70
  44. package/deps/rocksdb/rocksdb/db/db_impl/db_impl.h +71 -23
  45. package/deps/rocksdb/rocksdb/db/db_impl/db_impl_compaction_flush.cc +43 -16
  46. package/deps/rocksdb/rocksdb/db/db_impl/db_impl_debug.cc +47 -33
  47. package/deps/rocksdb/rocksdb/db/db_impl/db_impl_files.cc +27 -19
  48. package/deps/rocksdb/rocksdb/db/db_impl/db_impl_open.cc +38 -25
  49. package/deps/rocksdb/rocksdb/db/db_impl/db_impl_readonly.cc +3 -3
  50. package/deps/rocksdb/rocksdb/db/db_impl/db_impl_secondary.cc +7 -4
  51. package/deps/rocksdb/rocksdb/db/db_impl/db_impl_write.cc +258 -42
  52. package/deps/rocksdb/rocksdb/db/db_io_failure_test.cc +161 -9
  53. package/deps/rocksdb/rocksdb/db/db_iter.cc +118 -86
  54. package/deps/rocksdb/rocksdb/db/db_iter.h +44 -17
  55. package/deps/rocksdb/rocksdb/db/db_options_test.cc +27 -6
  56. package/deps/rocksdb/rocksdb/db/db_test.cc +48 -16
  57. package/deps/rocksdb/rocksdb/db/db_test2.cc +60 -15
  58. package/deps/rocksdb/rocksdb/db/db_test_util.cc +97 -44
  59. package/deps/rocksdb/rocksdb/db/db_test_util.h +7 -1
  60. package/deps/rocksdb/rocksdb/db/dbformat.cc +15 -5
  61. package/deps/rocksdb/rocksdb/db/dbformat.h +137 -55
  62. package/deps/rocksdb/rocksdb/db/event_helpers.cc +1 -0
  63. package/deps/rocksdb/rocksdb/db/experimental.cc +54 -0
  64. package/deps/rocksdb/rocksdb/db/external_sst_file_basic_test.cc +663 -8
  65. package/deps/rocksdb/rocksdb/db/external_sst_file_ingestion_job.cc +152 -91
  66. package/deps/rocksdb/rocksdb/db/external_sst_file_ingestion_job.h +134 -11
  67. package/deps/rocksdb/rocksdb/db/external_sst_file_test.cc +55 -9
  68. package/deps/rocksdb/rocksdb/db/flush_job.cc +52 -29
  69. package/deps/rocksdb/rocksdb/db/flush_job.h +5 -3
  70. package/deps/rocksdb/rocksdb/db/flush_job_test.cc +18 -12
  71. package/deps/rocksdb/rocksdb/db/forward_iterator.cc +23 -29
  72. package/deps/rocksdb/rocksdb/db/import_column_family_job.cc +3 -2
  73. package/deps/rocksdb/rocksdb/db/import_column_family_test.cc +2 -0
  74. package/deps/rocksdb/rocksdb/db/internal_stats.cc +9 -6
  75. package/deps/rocksdb/rocksdb/db/internal_stats.h +54 -0
  76. package/deps/rocksdb/rocksdb/db/job_context.h +1 -1
  77. package/deps/rocksdb/rocksdb/db/log_reader.cc +6 -7
  78. package/deps/rocksdb/rocksdb/db/manifest_ops.cc +47 -0
  79. package/deps/rocksdb/rocksdb/db/manifest_ops.h +20 -0
  80. package/deps/rocksdb/rocksdb/db/memtable.cc +165 -64
  81. package/deps/rocksdb/rocksdb/db/memtable.h +422 -243
  82. package/deps/rocksdb/rocksdb/db/memtable_list.cc +99 -68
  83. package/deps/rocksdb/rocksdb/db/memtable_list.h +63 -38
  84. package/deps/rocksdb/rocksdb/db/memtable_list_test.cc +28 -25
  85. package/deps/rocksdb/rocksdb/db/multi_cf_iterator_impl.h +118 -60
  86. package/deps/rocksdb/rocksdb/db/multi_cf_iterator_test.cc +344 -89
  87. package/deps/rocksdb/rocksdb/db/range_tombstone_fragmenter.h +2 -3
  88. package/deps/rocksdb/rocksdb/db/repair.cc +15 -14
  89. package/deps/rocksdb/rocksdb/db/repair_test.cc +0 -13
  90. package/deps/rocksdb/rocksdb/db/snapshot_checker.h +7 -0
  91. package/deps/rocksdb/rocksdb/db/table_cache.cc +62 -65
  92. package/deps/rocksdb/rocksdb/db/table_cache.h +70 -76
  93. package/deps/rocksdb/rocksdb/db/table_cache_sync_and_async.h +5 -6
  94. package/deps/rocksdb/rocksdb/db/table_properties_collector_test.cc +1 -1
  95. package/deps/rocksdb/rocksdb/db/transaction_log_impl.cc +8 -7
  96. package/deps/rocksdb/rocksdb/db/version_builder.cc +17 -19
  97. package/deps/rocksdb/rocksdb/db/version_builder.h +13 -12
  98. package/deps/rocksdb/rocksdb/db/version_edit.h +30 -0
  99. package/deps/rocksdb/rocksdb/db/version_edit_handler.cc +3 -5
  100. package/deps/rocksdb/rocksdb/db/version_set.cc +89 -129
  101. package/deps/rocksdb/rocksdb/db/version_set.h +12 -4
  102. package/deps/rocksdb/rocksdb/db/version_set_sync_and_async.h +1 -2
  103. package/deps/rocksdb/rocksdb/db/version_set_test.cc +12 -8
  104. package/deps/rocksdb/rocksdb/db/wide/wide_column_serialization.cc +0 -15
  105. package/deps/rocksdb/rocksdb/db/wide/wide_column_serialization.h +0 -2
  106. package/deps/rocksdb/rocksdb/db/wide/wide_column_serialization_test.cc +9 -7
  107. package/deps/rocksdb/rocksdb/db/wide/wide_columns_helper.cc +0 -8
  108. package/deps/rocksdb/rocksdb/db/wide/wide_columns_helper.h +28 -2
  109. package/deps/rocksdb/rocksdb/db/write_batch.cc +32 -10
  110. package/deps/rocksdb/rocksdb/db/write_batch_internal.h +9 -0
  111. package/deps/rocksdb/rocksdb/db/write_batch_test.cc +2 -1
  112. package/deps/rocksdb/rocksdb/db/write_thread.cc +3 -1
  113. package/deps/rocksdb/rocksdb/db/write_thread.h +6 -2
  114. package/deps/rocksdb/rocksdb/db_stress_tool/batched_ops_stress.cc +15 -0
  115. package/deps/rocksdb/rocksdb/db_stress_tool/cf_consistency_stress.cc +7 -0
  116. package/deps/rocksdb/rocksdb/db_stress_tool/db_stress_common.h +4 -0
  117. package/deps/rocksdb/rocksdb/db_stress_tool/db_stress_gflags.cc +18 -2
  118. package/deps/rocksdb/rocksdb/db_stress_tool/db_stress_test_base.cc +100 -22
  119. package/deps/rocksdb/rocksdb/db_stress_tool/db_stress_test_base.h +15 -4
  120. package/deps/rocksdb/rocksdb/db_stress_tool/multi_ops_txns_stress.cc +34 -8
  121. package/deps/rocksdb/rocksdb/db_stress_tool/no_batched_ops_stress.cc +223 -78
  122. package/deps/rocksdb/rocksdb/env/file_system.cc +6 -1
  123. package/deps/rocksdb/rocksdb/env/fs_posix.cc +53 -0
  124. package/deps/rocksdb/rocksdb/env/io_posix.cc +63 -17
  125. package/deps/rocksdb/rocksdb/env/io_posix.h +30 -1
  126. package/deps/rocksdb/rocksdb/file/file_prefetch_buffer.cc +132 -48
  127. package/deps/rocksdb/rocksdb/file/file_prefetch_buffer.h +92 -24
  128. package/deps/rocksdb/rocksdb/file/prefetch_test.cc +727 -109
  129. package/deps/rocksdb/rocksdb/file/random_access_file_reader.cc +3 -4
  130. package/deps/rocksdb/rocksdb/file/random_access_file_reader.h +1 -1
  131. package/deps/rocksdb/rocksdb/file/writable_file_writer.cc +8 -0
  132. package/deps/rocksdb/rocksdb/include/rocksdb/attribute_groups.h +20 -1
  133. package/deps/rocksdb/rocksdb/include/rocksdb/compaction_job_stats.h +9 -0
  134. package/deps/rocksdb/rocksdb/include/rocksdb/configurable.h +9 -5
  135. package/deps/rocksdb/rocksdb/include/rocksdb/convenience.h +2 -0
  136. package/deps/rocksdb/rocksdb/include/rocksdb/db.h +10 -2
  137. package/deps/rocksdb/rocksdb/include/rocksdb/env.h +1 -0
  138. package/deps/rocksdb/rocksdb/include/rocksdb/experimental.h +7 -0
  139. package/deps/rocksdb/rocksdb/include/rocksdb/file_system.h +34 -37
  140. package/deps/rocksdb/rocksdb/include/rocksdb/iterator_base.h +21 -0
  141. package/deps/rocksdb/rocksdb/include/rocksdb/options.h +56 -28
  142. package/deps/rocksdb/rocksdb/include/rocksdb/sst_file_writer.h +3 -0
  143. package/deps/rocksdb/rocksdb/include/rocksdb/table.h +36 -28
  144. package/deps/rocksdb/rocksdb/include/rocksdb/table_properties.h +11 -0
  145. package/deps/rocksdb/rocksdb/include/rocksdb/thread_status.h +1 -0
  146. package/deps/rocksdb/rocksdb/include/rocksdb/utilities/options_type.h +84 -60
  147. package/deps/rocksdb/rocksdb/include/rocksdb/utilities/secondary_index.h +102 -0
  148. package/deps/rocksdb/rocksdb/include/rocksdb/utilities/table_properties_collectors.h +89 -2
  149. package/deps/rocksdb/rocksdb/include/rocksdb/utilities/transaction.h +32 -0
  150. package/deps/rocksdb/rocksdb/include/rocksdb/utilities/transaction_db.h +30 -1
  151. package/deps/rocksdb/rocksdb/include/rocksdb/utilities/write_batch_with_index.h +23 -2
  152. package/deps/rocksdb/rocksdb/include/rocksdb/version.h +2 -2
  153. package/deps/rocksdb/rocksdb/include/rocksdb/write_batch.h +2 -0
  154. package/deps/rocksdb/rocksdb/memtable/inlineskiplist.h +79 -21
  155. package/deps/rocksdb/rocksdb/memtable/skiplist.h +41 -18
  156. package/deps/rocksdb/rocksdb/memtable/skiplistrep.cc +1 -5
  157. package/deps/rocksdb/rocksdb/memtable/wbwi_memtable.cc +169 -0
  158. package/deps/rocksdb/rocksdb/memtable/wbwi_memtable.h +400 -0
  159. package/deps/rocksdb/rocksdb/monitoring/thread_status_util_debug.cc +2 -0
  160. package/deps/rocksdb/rocksdb/options/cf_options.cc +137 -82
  161. package/deps/rocksdb/rocksdb/options/cf_options.h +18 -6
  162. package/deps/rocksdb/rocksdb/options/configurable.cc +31 -17
  163. package/deps/rocksdb/rocksdb/options/configurable_helper.h +7 -6
  164. package/deps/rocksdb/rocksdb/options/options_helper.cc +10 -8
  165. package/deps/rocksdb/rocksdb/options/options_parser.cc +74 -54
  166. package/deps/rocksdb/rocksdb/options/options_settable_test.cc +89 -0
  167. package/deps/rocksdb/rocksdb/options/options_test.cc +112 -26
  168. package/deps/rocksdb/rocksdb/port/port.h +5 -9
  169. package/deps/rocksdb/rocksdb/src.mk +8 -0
  170. package/deps/rocksdb/rocksdb/table/adaptive/adaptive_table_factory.h +4 -0
  171. package/deps/rocksdb/rocksdb/table/block_based/block.h +1 -7
  172. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_builder.cc +2 -0
  173. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_factory.cc +62 -80
  174. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_factory.h +13 -3
  175. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_iterator.cc +16 -5
  176. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_iterator.h +38 -7
  177. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_reader.cc +12 -4
  178. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_reader.h +4 -1
  179. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_reader_sync_and_async.h +4 -1
  180. package/deps/rocksdb/rocksdb/table/block_based/block_based_table_reader_test.cc +204 -1
  181. package/deps/rocksdb/rocksdb/table/block_based/data_block_hash_index_test.cc +3 -3
  182. package/deps/rocksdb/rocksdb/table/block_fetcher_test.cc +2 -1
  183. package/deps/rocksdb/rocksdb/table/cuckoo/cuckoo_table_factory.h +4 -0
  184. package/deps/rocksdb/rocksdb/table/format.cc +3 -3
  185. package/deps/rocksdb/rocksdb/table/meta_blocks.cc +4 -1
  186. package/deps/rocksdb/rocksdb/table/mock_table.cc +0 -50
  187. package/deps/rocksdb/rocksdb/table/mock_table.h +53 -0
  188. package/deps/rocksdb/rocksdb/table/plain/plain_table_factory.h +4 -0
  189. package/deps/rocksdb/rocksdb/table/sst_file_dumper.cc +1 -1
  190. package/deps/rocksdb/rocksdb/table/sst_file_writer.cc +10 -5
  191. package/deps/rocksdb/rocksdb/table/table_builder.h +3 -1
  192. package/deps/rocksdb/rocksdb/table/table_properties.cc +181 -0
  193. package/deps/rocksdb/rocksdb/table/table_reader_bench.cc +5 -5
  194. package/deps/rocksdb/rocksdb/table/table_test.cc +71 -64
  195. package/deps/rocksdb/rocksdb/tools/block_cache_analyzer/block_cache_pysim.py +45 -45
  196. package/deps/rocksdb/rocksdb/tools/block_cache_analyzer/block_cache_pysim_test.py +35 -35
  197. package/deps/rocksdb/rocksdb/tools/block_cache_analyzer/block_cache_trace_analyzer_plot.py +43 -43
  198. package/deps/rocksdb/rocksdb/tools/db_bench_tool.cc +41 -4
  199. package/deps/rocksdb/rocksdb/tools/ldb_cmd.cc +1 -0
  200. package/deps/rocksdb/rocksdb/tools/sst_dump_test.cc +1 -1
  201. package/deps/rocksdb/rocksdb/unreleased_history/add.sh +13 -0
  202. package/deps/rocksdb/rocksdb/util/aligned_buffer.h +24 -5
  203. package/deps/rocksdb/rocksdb/util/compaction_job_stats_impl.cc +7 -0
  204. package/deps/rocksdb/rocksdb/util/file_checksum_helper.cc +0 -52
  205. package/deps/rocksdb/rocksdb/util/file_checksum_helper.h +1 -10
  206. package/deps/rocksdb/rocksdb/util/file_reader_writer_test.cc +92 -0
  207. package/deps/rocksdb/rocksdb/util/thread_operation.h +1 -0
  208. package/deps/rocksdb/rocksdb/util/udt_util.cc +50 -4
  209. package/deps/rocksdb/rocksdb/util/udt_util.h +24 -11
  210. package/deps/rocksdb/rocksdb/util/udt_util_test.cc +26 -13
  211. package/deps/rocksdb/rocksdb/utilities/memory/memory_test.cc +1 -16
  212. package/deps/rocksdb/rocksdb/utilities/options/options_util_test.cc +2 -0
  213. package/deps/rocksdb/rocksdb/utilities/secondary_index/faiss_ivf_index.cc +214 -0
  214. package/deps/rocksdb/rocksdb/utilities/secondary_index/faiss_ivf_index.h +60 -0
  215. package/deps/rocksdb/rocksdb/utilities/secondary_index/faiss_ivf_index_test.cc +124 -0
  216. package/deps/rocksdb/rocksdb/utilities/secondary_index/secondary_index_mixin.h +441 -0
  217. package/deps/rocksdb/rocksdb/utilities/table_properties_collectors/compact_for_tiering_collector.cc +34 -3
  218. package/deps/rocksdb/rocksdb/utilities/table_properties_collectors/compact_for_tiering_collector.h +7 -2
  219. package/deps/rocksdb/rocksdb/utilities/transactions/optimistic_transaction_test.cc +437 -0
  220. package/deps/rocksdb/rocksdb/utilities/transactions/pessimistic_transaction.cc +34 -11
  221. package/deps/rocksdb/rocksdb/utilities/transactions/pessimistic_transaction.h +14 -7
  222. package/deps/rocksdb/rocksdb/utilities/transactions/pessimistic_transaction_db.cc +7 -1
  223. package/deps/rocksdb/rocksdb/utilities/transactions/snapshot_checker.cc +17 -0
  224. package/deps/rocksdb/rocksdb/utilities/transactions/transaction_base.cc +69 -0
  225. package/deps/rocksdb/rocksdb/utilities/transactions/transaction_base.h +20 -0
  226. package/deps/rocksdb/rocksdb/utilities/transactions/transaction_test.cc +1290 -0
  227. package/deps/rocksdb/rocksdb/utilities/transactions/write_committed_transaction_ts_test.cc +324 -0
  228. package/deps/rocksdb/rocksdb/utilities/transactions/write_prepared_txn.cc +18 -1
  229. package/deps/rocksdb/rocksdb/utilities/transactions/write_prepared_txn.h +8 -1
  230. package/deps/rocksdb/rocksdb/utilities/write_batch_with_index/write_batch_with_index.cc +57 -12
  231. package/deps/rocksdb/rocksdb/utilities/write_batch_with_index/write_batch_with_index_internal.cc +32 -3
  232. package/deps/rocksdb/rocksdb/utilities/write_batch_with_index/write_batch_with_index_internal.h +33 -2
  233. package/deps/rocksdb/rocksdb/utilities/write_batch_with_index/write_batch_with_index_test.cc +721 -9
  234. package/deps/rocksdb/rocksdb.gyp +2 -0
  235. package/package.json +1 -1
  236. package/prebuilds/darwin-arm64/@nxtedition+rocksdb.node +0 -0
  237. package/prebuilds/linux-x64/@nxtedition+rocksdb.node +0 -0
@@ -15,7 +15,9 @@
15
15
  #include <sstream>
16
16
  #include <string>
17
17
 
18
+ #include "file/random_access_file_reader.h"
18
19
  #include "file/readahead_file_info.h"
20
+ #include "file_util.h"
19
21
  #include "monitoring/statistics_impl.h"
20
22
  #include "port/port.h"
21
23
  #include "rocksdb/env.h"
@@ -149,6 +151,9 @@ enum class FilePrefetchBufferUsage {
149
151
  //
150
152
  // If num_buffers_ == 1, it's a sequential read flow. Read API will be called on
151
153
  // that one buffer whenever the data is requested and is not in the buffer.
154
+ // When reusing the file system allocated buffer, overlap_buf_ is used if the
155
+ // main buffer only contains part of the requested data. It is returned to
156
+ // the caller after the remaining data is fetched.
152
157
  // If num_buffers_ > 1, then the data is prefetched asynchronosuly in the
153
158
  // buffers whenever the data is consumed from the buffers and that buffer is
154
159
  // freed.
@@ -206,10 +211,15 @@ class FilePrefetchBuffer {
206
211
  assert((num_file_reads_ >= num_file_reads_for_auto_readahead_ + 1) ||
207
212
  (num_file_reads_ == 0));
208
213
 
209
- // If num_buffers_ > 1, data is asynchronously filled in the
210
- // queue. As result, data can be overlapping in two buffers. It copies the
211
- // data to overlap_buf_ in order to to return continuous buffer.
212
- if (num_buffers_ > 1) {
214
+ // overlap_buf_ is used whenever the main buffer only has part of the
215
+ // requested data. The relevant data is copied into overlap_buf_ and the
216
+ // remaining data is copied in later to satisfy the user's request. This is
217
+ // used in both the synchronous (num_buffers_ = 1) and asynchronous
218
+ // (num_buffers_ > 1) cases. In the asynchronous case, the requested data
219
+ // may be spread out over 2 buffers.
220
+ if (num_buffers_ > 1 ||
221
+ (fs_ != nullptr &&
222
+ CheckFSFeatureSupport(fs_, FSSupportedOps::kFSBuffer))) {
213
223
  overlap_buf_ = new BufferInfo();
214
224
  }
215
225
 
@@ -379,12 +389,21 @@ class FilePrefetchBuffer {
379
389
  void PrefetchAsyncCallback(FSReadRequest& req, void* cb_arg);
380
390
 
381
391
  void TEST_GetBufferOffsetandSize(
382
- std::vector<std::pair<uint64_t, size_t>>& buffer_info) {
392
+ std::vector<std::tuple<uint64_t, size_t, bool>>& buffer_info) {
383
393
  for (size_t i = 0; i < bufs_.size(); i++) {
384
- buffer_info[i].first = bufs_[i]->offset_;
385
- buffer_info[i].second = bufs_[i]->async_read_in_progress_
386
- ? bufs_[i]->async_req_len_
387
- : bufs_[i]->CurrentSize();
394
+ std::get<0>(buffer_info[i]) = bufs_[i]->offset_;
395
+ std::get<1>(buffer_info[i]) = bufs_[i]->async_read_in_progress_
396
+ ? bufs_[i]->async_req_len_
397
+ : bufs_[i]->CurrentSize();
398
+ std::get<2>(buffer_info[i]) = bufs_[i]->async_read_in_progress_;
399
+ }
400
+ }
401
+
402
+ void TEST_GetOverlapBufferOffsetandSize(
403
+ std::pair<uint64_t, size_t>& buffer_info) {
404
+ if (overlap_buf_ != nullptr) {
405
+ buffer_info.first = overlap_buf_->offset_;
406
+ buffer_info.second = overlap_buf_->CurrentSize();
388
407
  }
389
408
  }
390
409
 
@@ -394,7 +413,7 @@ class FilePrefetchBuffer {
394
413
  // required.
395
414
  void PrepareBufferForRead(BufferInfo* buf, size_t alignment, uint64_t offset,
396
415
  size_t roundup_len, bool refit_tail,
397
- uint64_t& aligned_useful_len);
416
+ bool use_fs_buffer, uint64_t& aligned_useful_len);
398
417
 
399
418
  void AbortOutdatedIO(uint64_t offset);
400
419
 
@@ -411,14 +430,16 @@ class FilePrefetchBuffer {
411
430
 
412
431
  Status Read(BufferInfo* buf, const IOOptions& opts,
413
432
  RandomAccessFileReader* reader, uint64_t read_len,
414
- uint64_t aligned_useful_len, uint64_t start_offset);
433
+ uint64_t aligned_useful_len, uint64_t start_offset,
434
+ bool use_fs_buffer);
415
435
 
416
436
  Status ReadAsync(BufferInfo* buf, const IOOptions& opts,
417
437
  RandomAccessFileReader* reader, uint64_t read_len,
418
438
  uint64_t start_offset);
419
439
 
420
440
  // Copy the data from src to overlap_buf_.
421
- void CopyDataToBuffer(BufferInfo* src, uint64_t& offset, size_t& length);
441
+ void CopyDataToOverlapBuffer(BufferInfo* src, uint64_t& offset,
442
+ size_t& length);
422
443
 
423
444
  bool IsBlockSequential(const size_t& offset) {
424
445
  return (prev_len_ == 0 || (prev_offset_ + prev_len_ == offset));
@@ -465,6 +486,49 @@ class FilePrefetchBuffer {
465
486
  return true;
466
487
  }
467
488
 
489
+ // Whether we reuse the file system provided buffer
490
+ // Until we also handle the async read case, only enable this optimization
491
+ // for the synchronous case when num_buffers_ = 1.
492
+ bool UseFSBuffer(RandomAccessFileReader* reader) {
493
+ return reader->file() != nullptr && !reader->use_direct_io() &&
494
+ fs_ != nullptr &&
495
+ CheckFSFeatureSupport(fs_, FSSupportedOps::kFSBuffer) &&
496
+ num_buffers_ == 1;
497
+ }
498
+
499
+ // When we are reusing the file system provided buffer, we are not concerned
500
+ // with alignment. However, quite a bit of prefetch code incorporates
501
+ // alignment, so we can put in 1 to keep the code simpler.
502
+ size_t GetRequiredBufferAlignment(RandomAccessFileReader* reader) {
503
+ if (UseFSBuffer(reader)) {
504
+ return 1;
505
+ }
506
+ return reader->file()->GetRequiredBufferAlignment();
507
+ }
508
+
509
+ // Reuses the file system allocated buffer to avoid an extra copy
510
+ IOStatus FSBufferDirectRead(RandomAccessFileReader* reader, BufferInfo* buf,
511
+ const IOOptions& opts, uint64_t offset, size_t n,
512
+ Slice& result) {
513
+ FSReadRequest read_req;
514
+ read_req.offset = offset;
515
+ read_req.len = n;
516
+ read_req.scratch = nullptr;
517
+ IOStatus s = reader->MultiRead(opts, &read_req, 1, nullptr);
518
+ if (!s.ok()) {
519
+ return s;
520
+ }
521
+ s = read_req.status;
522
+ if (!s.ok()) {
523
+ return s;
524
+ }
525
+ buf->buffer_.SetBuffer(read_req.result, std::move(read_req.fs_scratch));
526
+ buf->offset_ = offset;
527
+ buf->initial_end_offset_ = offset + read_req.result.size();
528
+ result = read_req.result;
529
+ return s;
530
+ }
531
+
468
532
  void DestroyAndClearIOHandle(BufferInfo* buf) {
469
533
  if (buf->io_handle_ != nullptr && buf->del_fn_ != nullptr) {
470
534
  buf->del_fn_(buf->io_handle_);
@@ -474,24 +538,28 @@ class FilePrefetchBuffer {
474
538
  buf->async_read_in_progress_ = false;
475
539
  }
476
540
 
477
- Status HandleOverlappingData(const IOOptions& opts,
478
- RandomAccessFileReader* reader, uint64_t offset,
479
- size_t length, size_t readahead_size,
480
- bool& copy_to_third_buffer, uint64_t& tmp_offset,
481
- size_t& tmp_length);
541
+ void HandleOverlappingSyncData(uint64_t offset, size_t length,
542
+ uint64_t& tmp_offset, size_t& tmp_length,
543
+ bool& use_overlap_buffer);
544
+
545
+ Status HandleOverlappingAsyncData(const IOOptions& opts,
546
+ RandomAccessFileReader* reader,
547
+ uint64_t offset, size_t length,
548
+ size_t readahead_size,
549
+ bool& copy_to_third_buffer,
550
+ uint64_t& tmp_offset, size_t& tmp_length);
482
551
 
483
552
  bool TryReadFromCacheUntracked(const IOOptions& opts,
484
553
  RandomAccessFileReader* reader,
485
554
  uint64_t offset, size_t n, Slice* result,
486
- Status* s,
487
- bool for_compaction = false);
555
+ Status* s, bool for_compaction = false);
488
556
 
489
557
  void ReadAheadSizeTuning(BufferInfo* buf, bool read_curr_block,
490
- bool refit_tail, uint64_t prev_buf_end_offset,
491
- size_t alignment, size_t length,
492
- size_t readahead_size, uint64_t& offset,
493
- uint64_t& end_offset, size_t& read_len,
494
- uint64_t& aligned_useful_len);
558
+ bool refit_tail, bool use_fs_buffer,
559
+ uint64_t prev_buf_end_offset, size_t alignment,
560
+ size_t length, size_t readahead_size,
561
+ uint64_t& offset, uint64_t& end_offset,
562
+ size_t& read_len, uint64_t& aligned_useful_len);
495
563
 
496
564
  void UpdateStats(bool found_in_buffer, size_t length_found) {
497
565
  if (found_in_buffer) {