From: Andrew Kryczka <andrewkr@fb.com>
Date: Mon, 12 Oct 2020 19:42:58 +0000 (-0700)
Subject: Fix bug in pinned partitioned indexes with some reads bypassing block cache
X-Git-Tag: v6.13.2~4
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=6676d31d3f6378d84e4f56516a9b3ced3bfeb763;p=rocksdb.git

Fix bug in pinned partitioned indexes with some reads bypassing block cache

Backports part of 75d3b6fdf0aa1007c4d26382f65be0adf4519a37.
---

diff --git a/HISTORY.md b/HISTORY.md
index b2d855db2..c4cd886d6 100644
--- a/HISTORY.md
+++ b/HISTORY.md
@@ -3,6 +3,7 @@
 ### Bug Fixes
 * Since 6.12, memtable lookup should report unrecognized value_type as corruption (#7121).
 * Fixed a bug in the following combination of features: indexes with user keys (`format_version >= 3`), indexes are partitioned (`index_type == kTwoLevelIndexSearch`), and some index partitions are pinned in memory (`BlockBasedTableOptions::pin_l0_filter_and_index_blocks_in_cache`). The bug could cause keys to be truncated when read from the index leading to wrong read results or other unexpected behavior.
+* Fixed a bug when indexes are partitioned (`index_type == kTwoLevelIndexSearch`), some index partitions are pinned in memory (`BlockBasedTableOptions::pin_l0_filter_and_index_blocks_in_cache`), and partitions reads could be mixed between block cache and directly from the file (e.g., with `enable_index_compression == 1` and `mmap_read == 1`, partitions that were stored uncompressed due to poor compression ratio would be read directly from the file via mmap, while partitions that were stored compressed would be read from block cache). The bug could cause index partitions to be mistakenly considered empty during reads leading to wrong read results.
 
 ## 6.13 (09/24/2020)
 ### Bug fixes
diff --git a/table/block_based/partitioned_index_reader.cc b/table/block_based/partitioned_index_reader.cc
index 89c16f65b..d60e93de2 100644
--- a/table/block_based/partitioned_index_reader.cc
+++ b/table/block_based/partitioned_index_reader.cc
@@ -173,7 +173,7 @@ Status PartitionIndexReader::CacheDependencies(const ReadOptions& ro,
       return s;
     }
     if (block.GetValue() != nullptr) {
-      if (block.IsCached()) {
+      if (block.IsCached() || block.GetOwnValue()) {
         if (pin) {
           partition_map_[handle.offset()] = std::move(block);
         }